agora inbox for [email protected]
help / color / mirror / Atom feedpgbench failed when -f option contains a char '@'
272+ messages / 5 participants
[nested] [flat]
* pgbench failed when -f option contains a char '@'
@ 2020-12-18 06:22 Wang, Shenhao <[email protected]>
2020-12-18 08:59 ` Re: pgbench failed when -f option contains a char '@' Heikki Linnakangas <[email protected]>
2020-12-20 13:31 ` Re: pgbench failed when -f option contains a char '@' Fabien COELHO <[email protected]>
0 siblings, 2 replies; 272+ messages in thread
From: Wang, Shenhao @ 2020-12-18 06:22 UTC (permalink / raw)
To: [email protected] <[email protected]>
Hi, hackers
pgbench use -f filename[@weight] to receive a sql script file with a weight,
but if I create a file contains char'@', like [email protected], specify this file without weigth,
pgbench will failed with error:
pgbench: fatal: invalid weight specification: @2.sql
This action may be unfriendly, because the char '@' is a valid character on Linux
and Windows.
I have created a patch to modify this action. The patch is attached.
Thoughts?
Regards
Shenhao Wang
Attachments:
[application/octet-stream] 0001-Improve-pgbench-when-f-option-contain-char-in-filepa.patch (2.9K, ../../1c4e81550d214741827a03292222db8d@G08CNEXMBPEKD06.g08.fujitsu.local/2-0001-Improve-pgbench-when-f-option-contain-char-in-filepa.patch)
download | inline diff:
From 7ad5a2aabab2393406f632e2559474af8af1382c Mon Sep 17 00:00:00 2001
From: Shenhao Wang <[email protected]>
Date: Fri, 18 Dec 2020 11:36:08 +0800
Subject: [PATCH] Improve pgbench when -f option contain char '@' in filepath
---
src/bin/pgbench/pgbench.c | 43 +++++++++++++++-----
src/bin/pgbench/t/001_pgbench_with_server.pl | 4 +-
2 files changed, 34 insertions(+), 13 deletions(-)
diff --git a/src/bin/pgbench/pgbench.c b/src/bin/pgbench/pgbench.c
index 3057665bbe..ceebb38409 100644
--- a/src/bin/pgbench/pgbench.c
+++ b/src/bin/pgbench/pgbench.c
@@ -4925,24 +4925,37 @@ read_file_contents(FILE *fd)
}
/*
- * Given a file name, read it and add its script to the list.
+ * Given a file name, and open it
* "-" means to read stdin.
* NB: filename must be storage that won't disappear.
*/
-static void
-process_file(const char *filename, int weight)
+static FILE*
+open_file(const char * filename, bool error_exit)
{
FILE *fd;
- char *buf;
-
- /* Slurp the file contents into "buf" */
if (strcmp(filename, "-") == 0)
fd = stdin;
- else if ((fd = fopen(filename, "r")) == NULL)
+ else if ((fd = fopen(filename, "r")) == NULL && error_exit)
{
pg_log_fatal("could not open file \"%s\": %m", filename);
exit(1);
}
+ return fd;
+}
+
+/*
+ * Given a FILE ptr, read it and add its script to the list.
+ * if FILE ptr is null, open file with filename
+ * NB: filename must be storage that won't disappear.
+ */
+static void
+process_file(FILE *fd, int weight, const char *filename)
+{
+ char *buf;
+
+ /* Slurp the file contents into "buf" */
+ if (fd == NULL)
+ fd = open_file(filename, true);
buf = read_file_contents(fd);
@@ -5633,10 +5646,18 @@ main(int argc, char **argv)
internal_script_used = true;
break;
case 'f':
- weight = parseScriptWeight(optarg, &script);
- process_file(script, weight);
- benchmarking_option_set = true;
- break;
+ {
+ FILE *fd = open_file(optarg, false);
+ if (fd != NULL)
+ process_file(fd, 1, optarg);
+ else
+ {
+ weight = parseScriptWeight(optarg, &script);
+ process_file(NULL, weight, script);
+ }
+ benchmarking_option_set = true;
+ break;
+ }
case 'D':
{
char *p;
diff --git a/src/bin/pgbench/t/001_pgbench_with_server.pl b/src/bin/pgbench/t/001_pgbench_with_server.pl
index 61b671d54f..ecb8024ef0 100644
--- a/src/bin/pgbench/t/001_pgbench_with_server.pl
+++ b/src/bin/pgbench/t/001_pgbench_with_server.pl
@@ -233,14 +233,14 @@ pgbench(
'-n -t 10 -c 1 -M simple',
0,
[
- qr{type: .*/001_pgbench_custom_script_3},
+ qr{type: .*/001_pgbench_\@custom_script_3},
qr{processed: 10/10},
qr{mode: simple}
],
[qr{^$}],
'pgbench custom script',
{
- '001_pgbench_custom_script_3' => q{-- select only variant
+ '001_pgbench_@custom_script_3' => q{-- select only variant
\set aid random(1, :scale * 100000)
BEGIN;
SELECT abalance::INTEGER AS balance
--
2.26.2
^ permalink raw reply [nested|flat] 272+ messages in thread
* Re: pgbench failed when -f option contains a char '@'
2020-12-18 06:22 pgbench failed when -f option contains a char '@' Wang, Shenhao <[email protected]>
@ 2020-12-18 08:59 ` Heikki Linnakangas <[email protected]>
2020-12-18 15:10 ` Re: pgbench failed when -f option contains a char '@' Tom Lane <[email protected]>
1 sibling, 1 reply; 272+ messages in thread
From: Heikki Linnakangas @ 2020-12-18 08:59 UTC (permalink / raw)
To: Wang, Shenhao <[email protected]>; [email protected] <[email protected]>
On 18/12/2020 08:22, Wang, Shenhao wrote:
> Hi, hackers
>
> pgbench use -f filename[@weight] to receive a sql script file with a weight,
> but if I create a file contains char'@', like [email protected], specify this file without weigth,
> pgbench will failed with error:
> pgbench: fatal: invalid weight specification: @2.sql
>
> This action may be unfriendly, because the char '@' is a valid character on Linux
> and Windows.
>
> I have created a patch to modify this action. The patch is attached.
This patch changes it to first check if the file "[email protected]" exists, and
if it doesn't, only then it tries to interpret it as a weight, as
filename "a" and weight "2.sql". That stilll doesn't fix the underlying
ambiguity, though. If you have a file called "script" and "script@1",
this makes it impossible to specify "script" with weight 1, because "-f
script@1" will now always open the file "script@1".
I think we should just leave this as it is. The user can simply rename
the file.
Or maybe one change would be worthwhile here: First check if the part
after the @ contains only digits. If doesn't, then assume it's part of
the filename rather than a weight. That would fix this for cases like
"[email protected]", although not for "foo@1".
- Heikki
^ permalink raw reply [nested|flat] 272+ messages in thread
* Re: pgbench failed when -f option contains a char '@'
2020-12-18 06:22 pgbench failed when -f option contains a char '@' Wang, Shenhao <[email protected]>
2020-12-18 08:59 ` Re: pgbench failed when -f option contains a char '@' Heikki Linnakangas <[email protected]>
@ 2020-12-18 15:10 ` Tom Lane <[email protected]>
2020-12-20 13:43 ` Re: pgbench failed when -f option contains a char '@' Fabien COELHO <[email protected]>
0 siblings, 1 reply; 272+ messages in thread
From: Tom Lane @ 2020-12-18 15:10 UTC (permalink / raw)
To: Heikki Linnakangas <[email protected]>; +Cc: Wang, Shenhao <[email protected]>; [email protected] <[email protected]>
Heikki Linnakangas <[email protected]> writes:
> I think we should just leave this as it is. The user can simply rename
> the file.
Yeah. The assumption when we defined the script-weight syntax was that
there's no particular reason to use "@" in a script file name, and
I don't see why that's a bad assumption.
> Or maybe one change would be worthwhile here: First check if the part
> after the @ contains only digits. If doesn't, then assume it's part of
> the filename rather than a weight. That would fix this for cases like
> "[email protected]", although not for "foo@1".
I do not like introducing ambiguity of that sort. Not being entirely
clear on which script file is going to be read seems like a recipe
for security issues.
regards, tom lane
^ permalink raw reply [nested|flat] 272+ messages in thread
* Re: pgbench failed when -f option contains a char '@'
2020-12-18 06:22 pgbench failed when -f option contains a char '@' Wang, Shenhao <[email protected]>
2020-12-18 08:59 ` Re: pgbench failed when -f option contains a char '@' Heikki Linnakangas <[email protected]>
2020-12-18 15:10 ` Re: pgbench failed when -f option contains a char '@' Tom Lane <[email protected]>
@ 2020-12-20 13:43 ` Fabien COELHO <[email protected]>
2020-12-20 18:38 ` Re: pgbench failed when -f option contains a char '@' Tom Lane <[email protected]>
0 siblings, 1 reply; 272+ messages in thread
From: Fabien COELHO @ 2020-12-20 13:43 UTC (permalink / raw)
To: Tom Lane <[email protected]>; +Cc: Heikki Linnakangas <[email protected]>; Wang, Shenhao <[email protected]>; [email protected] <[email protected]>
Hello Tom,
>> I think we should just leave this as it is. The user can simply rename
>> the file.
>
> Yeah. The assumption when we defined the script-weight syntax was that
> there's no particular reason to use "@" in a script file name, and
> I don't see why that's a bad assumption.
The "parser" looks for the last @ in the argument, so the simple
workaround is to append "@1".
I suggest the attached doc update, or anything in better English.
--
Fabien.
Attachments:
[text/x-diff] pgbench-weight-doc-1.patch (684B, ../../alpine.DEB.2.22.394.2012201436420.17354@pseudo/2-pgbench-weight-doc-1.patch)
download | inline diff:
diff --git a/doc/src/sgml/ref/pgbench.sgml b/doc/src/sgml/ref/pgbench.sgml
index 7180fedd65..bba3cf05b0 100644
--- a/doc/src/sgml/ref/pgbench.sgml
+++ b/doc/src/sgml/ref/pgbench.sgml
@@ -461,6 +461,8 @@ pgbench <optional> <replaceable>options</replaceable> </optional> <replaceable>d
the list of executed scripts.
An optional integer weight after <literal>@</literal> allows to adjust the
probability of drawing the test.
+ If the filename includes a <literal>@</literal> character, append a weight so
+ that there is no ambiguity: <literal>filen@me@1</literal>.
See below for details.
</para>
</listitem>
^ permalink raw reply [nested|flat] 272+ messages in thread
* Re: pgbench failed when -f option contains a char '@'
2020-12-18 06:22 pgbench failed when -f option contains a char '@' Wang, Shenhao <[email protected]>
2020-12-18 08:59 ` Re: pgbench failed when -f option contains a char '@' Heikki Linnakangas <[email protected]>
2020-12-18 15:10 ` Re: pgbench failed when -f option contains a char '@' Tom Lane <[email protected]>
2020-12-20 13:43 ` Re: pgbench failed when -f option contains a char '@' Fabien COELHO <[email protected]>
@ 2020-12-20 18:38 ` Tom Lane <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Tom Lane @ 2020-12-20 18:38 UTC (permalink / raw)
To: Fabien COELHO <[email protected]>; +Cc: Heikki Linnakangas <[email protected]>; Wang, Shenhao <[email protected]>; [email protected] <[email protected]>
Fabien COELHO <[email protected]> writes:
> The "parser" looks for the last @ in the argument, so the simple
> workaround is to append "@1".
> I suggest the attached doc update, or anything in better English.
Agreed, done.
regards, tom lane
^ permalink raw reply [nested|flat] 272+ messages in thread
* Re: pgbench failed when -f option contains a char '@'
2020-12-18 06:22 pgbench failed when -f option contains a char '@' Wang, Shenhao <[email protected]>
@ 2020-12-20 13:31 ` Fabien COELHO <[email protected]>
1 sibling, 0 replies; 272+ messages in thread
From: Fabien COELHO @ 2020-12-20 13:31 UTC (permalink / raw)
To: Wang, Shenhao <[email protected]>; +Cc: [email protected] <[email protected]>
Hello,
> pgbench use -f filename[@weight] to receive a sql script file with a weight,
ISTM that I thought of this: "pgbench -f filen@me@1" does work.
sh> touch foo@bla
sh> pgbench -f foo@bla@1
pgbench: fatal: empty command list for script "foo@bla"
The documentation could point this out, though.
--
Fabien.
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
* [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup.
@ 2024-05-18 01:41 Thomas Munro <[email protected]>
0 siblings, 0 replies; 272+ messages in thread
From: Thomas Munro @ 2024-05-18 01:41 UTC (permalink / raw)
When backend processes were launched in EXEC_BACKEND builds, they would
run LocalProcessControlFile() to read in pg_control and extract several
important settings.
This happens too early to acquire ControlFileLock, and the postmaster is
also not allowed to acquire ControlFileLock, so it can't safely take a
copy to give to the child.
Instead, pass down the "proto-controlfile" that was read by the
postmaster in LocalProcessControlFile(). Introduce functions
ExportProtoControlFile() and ImportProtoControlFile() to allow that.
Subprocesses will extract information from that, and then later attach
to the current control file in shared memory.
Reported-by: Melanie Plageman <[email protected]> per Windows CI failure
Discussion: https://postgr.es/m/CAAKRu_YNGwEYrorQYza_W8tU%2B%3DtoXRHG8HpyHC-KDbZqA_ZVSA%40mail.gmail.com
---
src/backend/access/transam/xlog.c | 46 +++++++++++++++++++++++--
src/backend/postmaster/launch_backend.c | 21 +++++++----
src/include/access/xlog.h | 5 +++
3 files changed, 64 insertions(+), 8 deletions(-)
diff --git a/src/backend/access/transam/xlog.c b/src/backend/access/transam/xlog.c
index 13ec6225b85..e52517eb9c1 100644
--- a/src/backend/access/transam/xlog.c
+++ b/src/backend/access/transam/xlog.c
@@ -575,6 +575,10 @@ static WALInsertLockPadded *WALInsertLocks = NULL;
*/
static ControlFileData *ControlFile = NULL;
+#ifdef EXEC_BACKEND
+static ControlFileData *ProtoControlFile = NULL;
+#endif
+
/*
* Calculate the amount of space left on the page after 'endptr'. Beware
* multiple evaluation!
@@ -692,6 +696,7 @@ static bool PerformRecoveryXLogAction(void);
static void InitControlFile(uint64 sysidentifier, uint32 data_checksum_version);
static void WriteControlFile(void);
static void ReadControlFile(void);
+static void ScanControlFile(void);
static void UpdateControlFile(void);
static char *str_time(pg_time_t tnow, char *buf, size_t bufsize);
@@ -4385,9 +4390,7 @@ WriteControlFile(void)
static void
ReadControlFile(void)
{
- pg_crc32c crc;
int fd;
- char wal_segsz_str[20];
int r;
/*
@@ -4420,6 +4423,15 @@ ReadControlFile(void)
close(fd);
+ ScanControlFile();
+}
+
+static void
+ScanControlFile(void)
+{
+ static char wal_segsz_str[20];
+ pg_crc32c crc;
+
/*
* Check for expected pg_control format version. If this is wrong, the
* CRC check will likely fail because we'll be checking the wrong number
@@ -4941,8 +4953,33 @@ LocalProcessControlFile(bool reset)
Assert(reset || ControlFile == NULL);
ControlFile = palloc_object(ControlFileData);
ReadControlFile();
+
+#ifdef EXEC_BACKEND
+ /* We need to be able to give this to subprocesses. */
+ ProtoControlFile = ControlFile;
+#endif
}
+#ifdef EXEC_BACKEND
+void
+ExportProtoControlFile(ControlFileData *copy)
+{
+ *copy = *ProtoControlFile;
+}
+
+/*
+ * Like LocalProcessControlFile(), but used early in EXEC_BACKEND children's
+ * startup. This receives the same file that the postmaster first read.
+ */
+void
+ImportProtoControlFile(const ControlFileData *copy)
+{
+ ControlFile = palloc(sizeof(ControlFileData));
+ *ControlFile = *copy;
+ ScanControlFile();
+}
+#endif
+
/*
* Get the wal_level from the control file. For a standby, this value should be
* considered as its active wal_level, because it may be different from what
@@ -5061,7 +5098,12 @@ XLOGShmemInit(void)
if (localControlFile)
{
memcpy(ControlFile, localControlFile, sizeof(ControlFileData));
+#ifdef EXEC_BACKEND
+ /* We still hold a reference to give to subprocesses. */
+ Assert(ProtoControlFile == localControlFile);
+#else
pfree(localControlFile);
+#endif
}
/*
diff --git a/src/backend/postmaster/launch_backend.c b/src/backend/postmaster/launch_backend.c
index 45690b11c99..e08a405f949 100644
--- a/src/backend/postmaster/launch_backend.c
+++ b/src/backend/postmaster/launch_backend.c
@@ -33,6 +33,9 @@
#include <unistd.h>
+#include "access/xlog.h"
+#include "catalog/pg_control.h"
+#include "common/file_utils.h"
#include "libpq/libpq-be.h"
#include "miscadmin.h"
#include "postmaster/autovacuum.h"
@@ -133,6 +136,14 @@ typedef struct
int MyPMChildSlot;
+ /*
+ * A copy of the ControlFileData from early in Postmaster startup. We
+ * need to access its contents it at a phase of initialization before we
+ * are allowed to acquire LWLocks, so we can't just use shared memory or
+ * read the file from disk.
+ */
+ ControlFileData proto_controlfile;
+
/*
* These are only used by backend processes, but are here because passing
* a socket needs some special handling on Windows. 'client_sock' is an
@@ -659,12 +670,6 @@ SubPostmasterMain(int argc, char *argv[])
*/
checkDataDir();
- /*
- * (re-)read control file, as it contains config. The postmaster will
- * already have read this, but this process doesn't know about that.
- */
- LocalProcessControlFile(false);
-
/*
* Reload any libraries that were preloaded by the postmaster. Since we
* exec'd this process, those libraries didn't come along with us; but we
@@ -752,6 +757,8 @@ save_backend_variables(BackendParameters *param,
param->MaxBackends = MaxBackends;
param->num_pmchild_slots = num_pmchild_slots;
+ ExportProtoControlFile(¶m->proto_controlfile);
+
#ifdef WIN32
param->PostmasterHandle = PostmasterHandle;
if (!write_duplicated_handle(¶m->initial_signal_pipe,
@@ -1026,6 +1033,8 @@ restore_backend_variables(BackendParameters *param)
strlcpy(pkglib_path, param->pkglib_path, MAXPGPATH);
+ ImportProtoControlFile(¶m->proto_controlfile);
+
/*
* We need to restore fd.c's counts of externally-opened FDs; to avoid
* confusion, be sure to do this after restoring max_safe_fds. (Note:
diff --git a/src/include/access/xlog.h b/src/include/access/xlog.h
index fdfb572467b..f1f3ad4e96e 100644
--- a/src/include/access/xlog.h
+++ b/src/include/access/xlog.h
@@ -208,6 +208,7 @@ typedef enum WALAvailability
struct XLogRecData;
struct XLogReaderState;
+struct ControlFileData;
extern XLogRecPtr XLogInsertRecord(struct XLogRecData *rdata,
XLogRecPtr fpw_lsn,
@@ -250,6 +251,10 @@ extern void XLOGShmemInit(void);
extern void BootStrapXLOG(uint32 data_checksum_version);
extern void InitializeWalConsistencyChecking(void);
extern void LocalProcessControlFile(bool reset);
+#ifdef EXEC_BACKEND
+extern void ExportProtoControlFile(struct ControlFileData *copy);
+extern void ImportProtoControlFile(const struct ControlFileData *copy);
+#endif
extern WalLevel GetActiveWalLevelOnStandby(void);
extern void StartupXLOG(void);
extern void ShutdownXLOG(int code, Datum arg);
--
2.47.3
--dhbc6bswyy6qufwn--
^ permalink raw reply [nested|flat] 272+ messages in thread
end of thread, other threads:[~2024-05-18 01:41 UTC | newest]
Thread overview: 272+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2020-12-18 06:22 pgbench failed when -f option contains a char '@' Wang, Shenhao <[email protected]>
2020-12-18 08:59 ` Heikki Linnakangas <[email protected]>
2020-12-18 15:10 ` Tom Lane <[email protected]>
2020-12-20 13:43 ` Fabien COELHO <[email protected]>
2020-12-20 18:38 ` Tom Lane <[email protected]>
2020-12-20 13:31 ` Fabien COELHO <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
2024-05-18 01:41 [PATCH v2] Fix pg_control corruption in EXEC_BACKEND startup. Thomas Munro <[email protected]>
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox