public inbox for [email protected]
help / color / mirror / Atom feedFrom: Peter Geoghegan <[email protected]>
To: Tomas Vondra <[email protected]>
Cc: Andres Freund <[email protected]>
Cc: Thomas Munro <[email protected]>
Cc: Nazir Bilal Yavuz <[email protected]>
Cc: Robert Haas <[email protected]>
Cc: Melanie Plageman <[email protected]>
Cc: PostgreSQL Hackers <[email protected]>
Cc: Georgios <[email protected]>
Cc: Konstantin Knizhnik <[email protected]>
Cc: Dilip Kumar <[email protected]>
Subject: Re: index prefetching
Date: Tue, 13 Jan 2026 15:36:28 -0500
Message-ID: <CAH2-Wz=6a7fGz2rALDX+xiFDuEaGQWpZ49xEaBUDKiPH8gcL+Q@mail.gmail.com> (raw)
In-Reply-To: <CAH2-WzmH7pVQ0-mYAxb82aWbz29_BiBPq2wV5p7+1o2sRFqDRQ@mail.gmail.com>
References: <[email protected]>
<5pltwb73d7cynsxo2yb54ygjk7haviatkrx43mnzihc6kkield@ahnstpgof46i>
<CA+hUKGKL3MRvEftAE+kwBuL2PLg2CwUoHEMr=-KSvsWN8pHq9w@mail.gmail.com>
<[email protected]>
<e33gafg4p7iwvo24ytrxuw43nafm5xm3jefpdspnarcbkfurs7@3jbgdiinxem5>
<[email protected]>
<CAH2-Wz=DfvzasnzLv43cu36Q1Ca8Qi70_JjZ7SRbNhDwwgvirg@mail.gmail.com>
<qdl4fojnbfcnm2k7b4zpvgd6gwzwdgtbl5c7shpimrb76dbyy6@scdnspus3ejh>
<bpdeohyqvltb77viyft4bza4xc4peed3jcoep74d2ih6ynqlke@wbnhcwmq3ril>
<CAH2-WznFdjY_OB2S7_BY4iAyeffK+XrE2qsX6aghgP63VocRfQ@mail.gmail.com>
<4zeu5yb73byiquvf3eefsunnrydyqfxy3eup66jrliutrtd4xl@5iifjey4n5m5>
<[email protected]>
<CAH2-WzkAC_gWQgc3MV3MWBx2EUaE4TfPk8XQPkJgs1R_qNvo=A@mail.gmail.com>
<CAH2-WznLRnUZHC4kmR+J3MgGpN0iTUaZ-4xeSHPiaYR=y3C8Og@mail.gmail.com>
<CAH2-WzkHK0++yn3_nmWL2rWzXiu8Qc=2=NPQDG4nfVjvYkRDrQ@mail.gmail.com>
<CAH2-Wzk9=x=a2TbcqYcX+XXmDHQr5=1v9m4Z_v8a-KwF1Zoz0A@mail.gmail.com>
<CAH2-WzmYqhacBH161peAWb5eF=Ja7CFAQ+0jSEMq=qnfLVTOOg@mail.gmail.com>
<CAH2-Wzmm+mXuv_r+eYgcwOKZzNSmbYoEmxi5mdKMUtkmYa_qXw@mail.gmail.com>
<CAH2-WzkC_EdGB_tdEQr63BiOuJbcvZiV=xO+OJYux4wOaxPD8w@mail.gmail.com>
<CAH2-Wzn8whv-RE4E5L2WQB_Ha_jy0UReJq1yypu83gDYHeXq5g@mail.gmail.com>
<CAH2-Wzm7-QuDOs6TcqfhhDsGEZCuHtn=D-SriOTnTZ_fiXNBvA@mail.gmail.com>
<CAH2-WzmH7pVQ0-mYAxb82aWbz29_BiBPq2wV5p7+1o2sRFqDRQ@mail.gmail.com>
On Wed, Jan 7, 2026 at 1:50 PM Peter Geoghegan <[email protected]> wrote:
> v6 focusses on simplifying the batch management code in
> heapam_batch_getnext_tid. Importantly, heapam_batch_getnext_tid no
> longer uses a loop to process items from the currently loaded batch/to
> load the next batch. The control flow in heapam_batch_getnext_tid is a
> lot simpler in general compared to v5.
The batch stopped applying again. Attached is v7.
Changes:
* Much improved read stream callback code, with comments that explain
exactly what's going on at each step.
* We once again support prefetching with index-only scans (for any
heap fetches that might be required).
We must cache visibility information at the level of whole batches for
this. Otherwise, the read stream have a different idea about which
heap pages are considered all visible from other code. The
corresponding code that actually reads heap tuples expects to be able
to get buffers from the read stream that precisely match what it
believes are required for any heap fetches. If they don't both agree
about visibility info, chaos ensues.
We avoid per-tuple visibility lookups in v7, preferring to do
everything (every VM lookup for every TID) up front for each batch.
This is simpler, and I suspect it's somewhat faster on average.
* Fixed several bugs involving scrollable cursors + index prefetching.
> I still haven't had time to produce an implementation of the "heap
> buffer locking minimization" optimization that's clean enough to
> present to the list.
Still haven't done this. Our new thinking on this is that it'd be best
to get prefetching in better shape before proceeding with the heap
buffer locking optimization.
--
Peter Geoghegan
Attachments:
[application/octet-stream] v7-0004-Make-hash-index-AM-use-amgetbatch-interface.patch (37.1K, 2-v7-0004-Make-hash-index-AM-use-amgetbatch-interface.patch)
download | inline diff:
From 7d69161eb80df8cea1ee74357bf43e0622e2534b Mon Sep 17 00:00:00 2001
From: Peter Geoghegan <[email protected]>
Date: Tue, 25 Nov 2025 18:03:15 -0500
Subject: [PATCH v7 4/4] Make hash index AM use amgetbatch interface.
Replace hashgettuple with hashgetbatch, a function that implements the
new amgetbatch interface. Plain index scans of hash indexes now return
matching items in batches consisting of all of the matches from a given
bucket or overflow page. This gives the core executor the ability to
perform optimizations like index prefetching during hash index scans.
Note that this makes hash index scans use the dropPin optimization,
since that is required to use the amgetbatch interface. This won't
avoid making hash index vacuuming wait for a cleanup lock when an index
scan holds onto a conflicting pin, since such index scans still need to
hold onto a conflicting bucket page pin for the duration of the scan.
In other words, use of the dropPin optimization won't benefit hash index
scans in the same way that it benefits nbtree index scans following
commit 2ed5b87f. However, there is still some value in keeping the
number of buffer pins held at any one time to a minimum.
Index prefetching tends to hold open as many as several dozen batches
with certain workloads (workloads where the stream position has to get
quite far ahead of the read position in order to maintain the
appropriate prefetch distance on the heapam side). Guaranteeing that
open batches won't hold buffer pins on index pages (at least in the
common case where the dropPin optimization is safe to use) thereby
simplifies resource management during index prefetching.
Also add Valgrind buffer lock instrumentation to hash, bringing it in
line with nbtree following commit 4a70f829. This is another requirement
when using the amgetbatch interface.
Author: Peter Geoghegan <[email protected]>
Discussion: https://postgr.es/m/CAH2-WzmYqhacBH161peAWb5eF=Ja7CFAQ+0jSEMq=qnfLVTOOg@mail.gmail.com
---
src/include/access/hash.h | 73 +-----
src/backend/access/hash/hash.c | 122 ++++------
src/backend/access/hash/hashpage.c | 26 +--
src/backend/access/hash/hashsearch.c | 331 +++++++++++----------------
src/backend/access/hash/hashutil.c | 100 ++++----
src/tools/pgindent/typedefs.list | 2 -
6 files changed, 256 insertions(+), 398 deletions(-)
diff --git a/src/include/access/hash.h b/src/include/access/hash.h
index a8702f0e5..1b80fd7ed 100644
--- a/src/include/access/hash.h
+++ b/src/include/access/hash.h
@@ -100,58 +100,6 @@ typedef HashPageOpaqueData *HashPageOpaque;
*/
#define HASHO_PAGE_ID 0xFF80
-typedef struct HashScanPosItem /* what we remember about each match */
-{
- ItemPointerData heapTid; /* TID of referenced heap item */
- OffsetNumber indexOffset; /* index item's location within page */
-} HashScanPosItem;
-
-typedef struct HashScanPosData
-{
- Buffer buf; /* if valid, the buffer is pinned */
- BlockNumber currPage; /* current hash index page */
- BlockNumber nextPage; /* next overflow page */
- BlockNumber prevPage; /* prev overflow or bucket page */
-
- /*
- * The items array is always ordered in index order (ie, increasing
- * indexoffset). When scanning backwards it is convenient to fill the
- * array back-to-front, so we start at the last slot and fill downwards.
- * Hence we need both a first-valid-entry and a last-valid-entry counter.
- * itemIndex is a cursor showing which entry was last returned to caller.
- */
- int firstItem; /* first valid index in items[] */
- int lastItem; /* last valid index in items[] */
- int itemIndex; /* current index in items[] */
-
- HashScanPosItem items[MaxIndexTuplesPerPage]; /* MUST BE LAST */
-} HashScanPosData;
-
-#define HashScanPosIsPinned(scanpos) \
-( \
- AssertMacro(BlockNumberIsValid((scanpos).currPage) || \
- !BufferIsValid((scanpos).buf)), \
- BufferIsValid((scanpos).buf) \
-)
-
-#define HashScanPosIsValid(scanpos) \
-( \
- AssertMacro(BlockNumberIsValid((scanpos).currPage) || \
- !BufferIsValid((scanpos).buf)), \
- BlockNumberIsValid((scanpos).currPage) \
-)
-
-#define HashScanPosInvalidate(scanpos) \
- do { \
- (scanpos).buf = InvalidBuffer; \
- (scanpos).currPage = InvalidBlockNumber; \
- (scanpos).nextPage = InvalidBlockNumber; \
- (scanpos).prevPage = InvalidBlockNumber; \
- (scanpos).firstItem = 0; \
- (scanpos).lastItem = 0; \
- (scanpos).itemIndex = 0; \
- } while (0)
-
/*
* HashScanOpaqueData is private state for a hash index scan.
*/
@@ -178,15 +126,6 @@ typedef struct HashScanOpaqueData
* referred only when hashso_buc_populated is true.
*/
bool hashso_buc_split;
- /* info about killed items if any (killedItems is NULL if never used) */
- int *killedItems; /* currPos.items indexes of killed items */
- int numKilled; /* number of currently stored items */
-
- /*
- * Identify all the matching items on a page and save them in
- * HashScanPosData
- */
- HashScanPosData currPos; /* current position data */
} HashScanOpaqueData;
typedef HashScanOpaqueData *HashScanOpaque;
@@ -368,11 +307,14 @@ extern bool hashinsert(Relation rel, Datum *values, bool *isnull,
IndexUniqueCheck checkUnique,
bool indexUnchanged,
struct IndexInfo *indexInfo);
-extern bool hashgettuple(IndexScanDesc scan, ScanDirection dir);
+extern BatchIndexScan hashgetbatch(IndexScanDesc scan,
+ BatchIndexScan priorbatch,
+ ScanDirection dir);
extern int64 hashgetbitmap(IndexScanDesc scan, TIDBitmap *tbm);
extern IndexScanDesc hashbeginscan(Relation rel, int nkeys, int norderbys);
extern void hashrescan(IndexScanDesc scan, ScanKey scankey, int nscankeys,
ScanKey orderbys, int norderbys);
+extern void hashfreebatch(IndexScanDesc scan, BatchIndexScan batch);
extern void hashendscan(IndexScanDesc scan);
extern IndexBulkDeleteResult *hashbulkdelete(IndexVacuumInfo *info,
IndexBulkDeleteResult *stats,
@@ -445,8 +387,9 @@ extern void _hash_finish_split(Relation rel, Buffer metabuf, Buffer obuf,
uint32 lowmask);
/* hashsearch.c */
-extern bool _hash_next(IndexScanDesc scan, ScanDirection dir);
-extern bool _hash_first(IndexScanDesc scan, ScanDirection dir);
+extern BatchIndexScan _hash_next(IndexScanDesc scan, ScanDirection dir,
+ BatchIndexScan priorbatch);
+extern BatchIndexScan _hash_first(IndexScanDesc scan, ScanDirection dir);
/* hashsort.c */
typedef struct HSpool HSpool; /* opaque struct in hashsort.c */
@@ -476,7 +419,7 @@ extern BlockNumber _hash_get_oldblock_from_newbucket(Relation rel, Bucket new_bu
extern BlockNumber _hash_get_newblock_from_oldbucket(Relation rel, Bucket old_bucket);
extern Bucket _hash_get_newbucket_from_oldbucket(Relation rel, Bucket old_bucket,
uint32 lowmask, uint32 maxbucket);
-extern void _hash_kill_items(IndexScanDesc scan);
+extern void _hash_kill_items(IndexScanDesc scan, BatchIndexScan batch);
/* hash.c */
extern void hashbucketcleanup(Relation rel, Bucket cur_bucket,
diff --git a/src/backend/access/hash/hash.c b/src/backend/access/hash/hash.c
index 6a20b67f6..361357428 100644
--- a/src/backend/access/hash/hash.c
+++ b/src/backend/access/hash/hash.c
@@ -101,9 +101,9 @@ hashhandler(PG_FUNCTION_ARGS)
.amadjustmembers = hashadjustmembers,
.ambeginscan = hashbeginscan,
.amrescan = hashrescan,
- .amgettuple = hashgettuple,
- .amgetbatch = NULL,
- .amfreebatch = NULL,
+ .amgettuple = NULL,
+ .amgetbatch = hashgetbatch,
+ .amfreebatch = hashfreebatch,
.amgetbitmap = hashgetbitmap,
.amendscan = hashendscan,
.amposreset = NULL,
@@ -286,53 +286,27 @@ hashinsert(Relation rel, Datum *values, bool *isnull,
/*
- * hashgettuple() -- Get the next tuple in the scan.
+ * hashgetbatch() -- Get the first or next batch of tuples in the scan
*/
-bool
-hashgettuple(IndexScanDesc scan, ScanDirection dir)
+BatchIndexScan
+hashgetbatch(IndexScanDesc scan, BatchIndexScan priorbatch, ScanDirection dir)
{
HashScanOpaque so = (HashScanOpaque) scan->opaque;
- bool res;
+ Relation rel = scan->indexRelation;
/* Hash indexes are always lossy since we store only the hash code */
scan->xs_recheck = true;
- /*
- * If we've already initialized this scan, we can just advance it in the
- * appropriate direction. If we haven't done so yet, we call a routine to
- * get the first item in the scan.
- */
- if (!HashScanPosIsValid(so->currPos))
- res = _hash_first(scan, dir);
- else
+ if (priorbatch == NULL)
{
- /*
- * Check to see if we should kill the previously-fetched tuple.
- */
- if (scan->kill_prior_tuple)
- {
- /*
- * Yes, so remember it for later. (We'll deal with all such tuples
- * at once right after leaving the index page or at end of scan.)
- * In case if caller reverses the indexscan direction it is quite
- * possible that the same item might get entered multiple times.
- * But, we don't detect that; instead, we just forget any excess
- * entries.
- */
- if (so->killedItems == NULL)
- so->killedItems = palloc_array(int, MaxIndexTuplesPerPage);
+ _hash_dropscanbuf(rel, so);
- if (so->numKilled < MaxIndexTuplesPerPage)
- so->killedItems[so->numKilled++] = so->currPos.itemIndex;
- }
-
- /*
- * Now continue the scan.
- */
- res = _hash_next(scan, dir);
+ /* Initialize the scan, and return first batch of matching items */
+ return _hash_first(scan, dir);
}
- return res;
+ /* Return batch positioned after caller's batch (in direction 'dir') */
+ return _hash_next(scan, dir, priorbatch);
}
@@ -342,26 +316,23 @@ hashgettuple(IndexScanDesc scan, ScanDirection dir)
int64
hashgetbitmap(IndexScanDesc scan, TIDBitmap *tbm)
{
- HashScanOpaque so = (HashScanOpaque) scan->opaque;
- bool res;
+ BatchIndexScan batch;
int64 ntids = 0;
- HashScanPosItem *currItem;
+ int itemIndex;
- res = _hash_first(scan, ForwardScanDirection);
+ batch = _hash_first(scan, ForwardScanDirection);
- while (res)
+ while (batch != NULL)
{
- currItem = &so->currPos.items[so->currPos.itemIndex];
+ for (itemIndex = batch->firstItem;
+ itemIndex <= batch->lastItem;
+ itemIndex++)
+ {
+ tbm_add_tuples(tbm, &batch->items[itemIndex].heapTid, 1, true);
+ ntids++;
+ }
- /*
- * _hash_first and _hash_next handle eliminate dead index entries
- * whenever scan->ignore_killed_tuples is true. Therefore, there's
- * nothing to do here except add the results to the TIDBitmap.
- */
- tbm_add_tuples(tbm, &(currItem->heapTid), 1, true);
- ntids++;
-
- res = _hash_next(scan, ForwardScanDirection);
+ batch = _hash_next(scan, ForwardScanDirection, batch);
}
return ntids;
@@ -383,17 +354,14 @@ hashbeginscan(Relation rel, int nkeys, int norderbys)
scan = RelationGetIndexScan(rel, nkeys, norderbys);
so = (HashScanOpaque) palloc_object(HashScanOpaqueData);
- HashScanPosInvalidate(so->currPos);
so->hashso_bucket_buf = InvalidBuffer;
so->hashso_split_bucket_buf = InvalidBuffer;
so->hashso_buc_populated = false;
so->hashso_buc_split = false;
- so->killedItems = NULL;
- so->numKilled = 0;
-
scan->opaque = so;
+ scan->maxitemsbatch = MaxIndexTuplesPerPage;
return scan;
}
@@ -408,18 +376,8 @@ hashrescan(IndexScanDesc scan, ScanKey scankey, int nscankeys,
HashScanOpaque so = (HashScanOpaque) scan->opaque;
Relation rel = scan->indexRelation;
- if (HashScanPosIsValid(so->currPos))
- {
- /* Before leaving current page, deal with any killed items */
- if (so->numKilled > 0)
- _hash_kill_items(scan);
- }
-
_hash_dropscanbuf(rel, so);
- /* set position invalid (this will cause _hash_first call) */
- HashScanPosInvalidate(so->currPos);
-
/* Update scan key, if a new one is given */
if (scankey && scan->numberOfKeys > 0)
memcpy(scan->keyData, scankey, scan->numberOfKeys * sizeof(ScanKeyData));
@@ -428,6 +386,25 @@ hashrescan(IndexScanDesc scan, ScanKey scankey, int nscankeys,
so->hashso_buc_split = false;
}
+/*
+ * hashfreebatch() -- Free batch resources, including its buffer pin
+ */
+void
+hashfreebatch(IndexScanDesc scan, BatchIndexScan batch)
+{
+ if (batch->numKilled > 0)
+ _hash_kill_items(scan, batch);
+
+ if (!scan->dropPin)
+ {
+ /* indexam_util_batch_unlock didn't unpin page earlier, do it now */
+ ReleaseBuffer(batch->buf);
+ batch->buf = InvalidBuffer;
+ }
+
+ indexam_util_batch_release(scan, batch);
+}
+
/*
* hashendscan() -- close down a scan
*/
@@ -437,17 +414,8 @@ hashendscan(IndexScanDesc scan)
HashScanOpaque so = (HashScanOpaque) scan->opaque;
Relation rel = scan->indexRelation;
- if (HashScanPosIsValid(so->currPos))
- {
- /* Before leaving current page, deal with any killed items */
- if (so->numKilled > 0)
- _hash_kill_items(scan);
- }
-
_hash_dropscanbuf(rel, so);
- if (so->killedItems != NULL)
- pfree(so->killedItems);
pfree(so);
scan->opaque = NULL;
}
diff --git a/src/backend/access/hash/hashpage.c b/src/backend/access/hash/hashpage.c
index 8e220a3ae..9b6911905 100644
--- a/src/backend/access/hash/hashpage.c
+++ b/src/backend/access/hash/hashpage.c
@@ -35,6 +35,7 @@
#include "port/pg_bitutils.h"
#include "storage/predicate.h"
#include "storage/smgr.h"
+#include "utils/memdebug.h"
#include "utils/rel.h"
static bool _hash_alloc_buckets(Relation rel, BlockNumber firstblock,
@@ -79,6 +80,9 @@ _hash_getbuf(Relation rel, BlockNumber blkno, int access, int flags)
if (access != HASH_NOLOCK)
LockBuffer(buf, access);
+ if (!RelationUsesLocalBuffers(rel))
+ VALGRIND_MAKE_MEM_DEFINED(BufferGetPage(buf), BLCKSZ);
+
/* ref count and lock type are correct */
_hash_checkpage(rel, buf, flags);
@@ -108,6 +112,9 @@ _hash_getbuf_with_condlock_cleanup(Relation rel, BlockNumber blkno, int flags)
return InvalidBuffer;
}
+ if (!RelationUsesLocalBuffers(rel))
+ VALGRIND_MAKE_MEM_DEFINED(BufferGetPage(buf), BLCKSZ);
+
/* ref count and lock type are correct */
_hash_checkpage(rel, buf, flags);
@@ -280,31 +287,24 @@ _hash_dropbuf(Relation rel, Buffer buf)
}
/*
- * _hash_dropscanbuf() -- release buffers used in scan.
+ * _hash_dropscanbuf() -- release buffers owned by scan.
*
- * This routine unpins the buffers used during scan on which we
- * hold no lock.
+ * This routine unpins the buffers for the primary bucket page and for the
+ * bucket page of a bucket being split as needed.
*/
void
_hash_dropscanbuf(Relation rel, HashScanOpaque so)
{
/* release pin we hold on primary bucket page */
- if (BufferIsValid(so->hashso_bucket_buf) &&
- so->hashso_bucket_buf != so->currPos.buf)
+ if (BufferIsValid(so->hashso_bucket_buf))
_hash_dropbuf(rel, so->hashso_bucket_buf);
so->hashso_bucket_buf = InvalidBuffer;
- /* release pin we hold on primary bucket page of bucket being split */
- if (BufferIsValid(so->hashso_split_bucket_buf) &&
- so->hashso_split_bucket_buf != so->currPos.buf)
+ /* release pin held on primary bucket page of bucket being split */
+ if (BufferIsValid(so->hashso_split_bucket_buf))
_hash_dropbuf(rel, so->hashso_split_bucket_buf);
so->hashso_split_bucket_buf = InvalidBuffer;
- /* release any pin we still hold */
- if (BufferIsValid(so->currPos.buf))
- _hash_dropbuf(rel, so->currPos.buf);
- so->currPos.buf = InvalidBuffer;
-
/* reset split scan */
so->hashso_buc_populated = false;
so->hashso_buc_split = false;
diff --git a/src/backend/access/hash/hashsearch.c b/src/backend/access/hash/hashsearch.c
index 89d1c5bc6..c72705dc3 100644
--- a/src/backend/access/hash/hashsearch.c
+++ b/src/backend/access/hash/hashsearch.c
@@ -22,105 +22,83 @@
#include "storage/predicate.h"
#include "utils/rel.h"
-static bool _hash_readpage(IndexScanDesc scan, Buffer *bufP,
- ScanDirection dir);
+static bool _hash_readpage(IndexScanDesc scan, Buffer buf, ScanDirection dir,
+ BatchIndexScan batch);
static int _hash_load_qualified_items(IndexScanDesc scan, Page page,
- OffsetNumber offnum, ScanDirection dir);
-static inline void _hash_saveitem(HashScanOpaque so, int itemIndex,
+ OffsetNumber offnum, ScanDirection dir,
+ BatchIndexScan batch);
+static inline void _hash_saveitem(BatchIndexScan batch, int itemIndex,
OffsetNumber offnum, IndexTuple itup);
static void _hash_readnext(IndexScanDesc scan, Buffer *bufp,
Page *pagep, HashPageOpaque *opaquep);
/*
- * _hash_next() -- Get the next item in a scan.
+ * _hash_next() -- Get the next batch of items in a scan.
*
- * On entry, so->currPos describes the current page, which may
- * be pinned but not locked, and so->currPos.itemIndex identifies
- * which item was previously returned.
+ * On entry, priorbatch describes the current page batch with items
+ * already returned.
*
- * On successful exit, scan->xs_heaptid is set to the TID of the next
- * heap tuple. so->currPos is updated as needed.
+ * On successful exit, returns a batch containing matching items from
+ * next page. Otherwise returns NULL, indicating that there are no
+ * further matches. No locks are ever held when we return.
*
- * On failure exit (no more tuples), we return false with pin
- * held on bucket page but no pins or locks held on overflow
- * page.
+ * Retains pins according to the same rules as _hash_first.
*/
-bool
-_hash_next(IndexScanDesc scan, ScanDirection dir)
+BatchIndexScan
+_hash_next(IndexScanDesc scan, ScanDirection dir, BatchIndexScan priorbatch)
{
Relation rel = scan->indexRelation;
HashScanOpaque so = (HashScanOpaque) scan->opaque;
- HashScanPosItem *currItem;
BlockNumber blkno;
Buffer buf;
- bool end_of_scan = false;
+ BatchIndexScan batch;
/*
- * Advance to the next tuple on the current page; or if done, try to read
- * data from the next or previous page based on the scan direction. Before
- * moving to the next or previous page make sure that we deal with all the
- * killed items.
+ * Determine which page to read next based on scan direction and details
+ * taken from the prior batch
*/
if (ScanDirectionIsForward(dir))
{
- if (++so->currPos.itemIndex > so->currPos.lastItem)
- {
- if (so->numKilled > 0)
- _hash_kill_items(scan);
-
- blkno = so->currPos.nextPage;
- if (BlockNumberIsValid(blkno))
- {
- buf = _hash_getbuf(rel, blkno, HASH_READ, LH_OVERFLOW_PAGE);
- if (!_hash_readpage(scan, &buf, dir))
- end_of_scan = true;
- }
- else
- end_of_scan = true;
- }
+ blkno = priorbatch->nextPage;
+ if (!BlockNumberIsValid(blkno) || !priorbatch->moreRight)
+ return NULL;
}
else
{
- if (--so->currPos.itemIndex < so->currPos.firstItem)
- {
- if (so->numKilled > 0)
- _hash_kill_items(scan);
-
- blkno = so->currPos.prevPage;
- if (BlockNumberIsValid(blkno))
- {
- buf = _hash_getbuf(rel, blkno, HASH_READ,
- LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
-
- /*
- * We always maintain the pin on bucket page for whole scan
- * operation, so releasing the additional pin we have acquired
- * here.
- */
- if (buf == so->hashso_bucket_buf ||
- buf == so->hashso_split_bucket_buf)
- _hash_dropbuf(rel, buf);
-
- if (!_hash_readpage(scan, &buf, dir))
- end_of_scan = true;
- }
- else
- end_of_scan = true;
- }
+ blkno = priorbatch->prevPage;
+ if (!BlockNumberIsValid(blkno) || !priorbatch->moreLeft)
+ return NULL;
}
- if (end_of_scan)
+ /* Allocate space for next batch */
+ batch = indexam_util_batch_alloc(scan);
+
+ /* Get the buffer for next batch */
+ if (ScanDirectionIsForward(dir))
+ buf = _hash_getbuf(rel, blkno, HASH_READ, LH_OVERFLOW_PAGE);
+ else
{
- _hash_dropscanbuf(rel, so);
- HashScanPosInvalidate(so->currPos);
- return false;
+ buf = _hash_getbuf(rel, blkno, HASH_READ,
+ LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
+
+ /*
+ * We always maintain the pin on bucket page for whole scan operation,
+ * so releasing the additional pin we have acquired here.
+ */
+ if (buf == so->hashso_bucket_buf ||
+ buf == so->hashso_split_bucket_buf)
+ _hash_dropbuf(rel, buf);
}
- /* OK, itemIndex says what to return */
- currItem = &so->currPos.items[so->currPos.itemIndex];
- scan->xs_heaptid = currItem->heapTid;
+ /* Read the next page and load items into allocated batch */
+ if (!_hash_readpage(scan, buf, dir, batch))
+ {
+ indexam_util_batch_release(scan, batch);
+ return NULL;
+ }
- return true;
+ /* Return the batch containing matched items from next page */
+ return batch;
}
/*
@@ -270,22 +248,21 @@ _hash_readprev(IndexScanDesc scan,
}
/*
- * _hash_first() -- Find the first item in a scan.
+ * _hash_first() -- Find the first batch of items in a scan.
*
- * We find the first item (or, if backward scan, the last item) in the
- * index that satisfies the qualification associated with the scan
- * descriptor.
+ * We find the first batch of items (or, if backward scan, the last
+ * batch) in the index that satisfies the qualification associated with
+ * the scan descriptor.
*
- * On successful exit, if the page containing current index tuple is an
- * overflow page, both pin and lock are released whereas if it is a bucket
- * page then it is pinned but not locked and data about the matching
- * tuple(s) on the page has been loaded into so->currPos,
- * scan->xs_heaptid is set to the heap TID of the current tuple.
+ * On successful exit, returns a batch containing matching items.
+ * Otherwise returns NULL, indicating that there are no further matches.
+ * No locks are ever held when we return.
*
- * On failure exit (no more tuples), we return false, with pin held on
- * bucket page but no pins or locks held on overflow page.
+ * We always retain our own pin on the bucket page. When we return a
+ * batch with a bucket page, it will retain its own reference pin iff
+ * indexam_util_batch_release determined that table AM requires one.
*/
-bool
+BatchIndexScan
_hash_first(IndexScanDesc scan, ScanDirection dir)
{
Relation rel = scan->indexRelation;
@@ -296,7 +273,7 @@ _hash_first(IndexScanDesc scan, ScanDirection dir)
Buffer buf;
Page page;
HashPageOpaque opaque;
- HashScanPosItem *currItem;
+ BatchIndexScan batch;
pgstat_count_index_scan(rel);
if (scan->instrument)
@@ -326,7 +303,7 @@ _hash_first(IndexScanDesc scan, ScanDirection dir)
* items in the index.
*/
if (cur->sk_flags & SK_ISNULL)
- return false;
+ return NULL;
/*
* Okay to compute the hash key. We want to do this before acquiring any
@@ -419,191 +396,159 @@ _hash_first(IndexScanDesc scan, ScanDirection dir)
_hash_readnext(scan, &buf, &page, &opaque);
}
- /* remember which buffer we have pinned, if any */
- Assert(BufferIsInvalid(so->currPos.buf));
- so->currPos.buf = buf;
+ /* Allocate space for first batch */
+ batch = indexam_util_batch_alloc(scan);
- /* Now find all the tuples satisfying the qualification from a page */
- if (!_hash_readpage(scan, &buf, dir))
- return false;
+ /* Read the first page and load items into allocated batch */
+ if (!_hash_readpage(scan, buf, dir, batch))
+ {
+ indexam_util_batch_release(scan, batch);
+ return NULL;
+ }
- /* OK, itemIndex says what to return */
- currItem = &so->currPos.items[so->currPos.itemIndex];
- scan->xs_heaptid = currItem->heapTid;
-
- /* if we're here, _hash_readpage found a valid tuples */
- return true;
+ /* Return the batch containing matched items */
+ return batch;
}
/*
- * _hash_readpage() -- Load data from current index page into so->currPos
+ * _hash_readpage() -- Load data from current index page into batch
*
* We scan all the items in the current index page and save them into
- * so->currPos if it satisfies the qualification. If no matching items
+ * the batch if they satisfy the qualification. If no matching items
* are found in the current page, we move to the next or previous page
* in a bucket chain as indicated by the direction.
*
* Return true if any matching items are found else return false.
*/
static bool
-_hash_readpage(IndexScanDesc scan, Buffer *bufP, ScanDirection dir)
+_hash_readpage(IndexScanDesc scan, Buffer buf, ScanDirection dir,
+ BatchIndexScan batch)
{
Relation rel = scan->indexRelation;
HashScanOpaque so = (HashScanOpaque) scan->opaque;
- Buffer buf;
Page page;
HashPageOpaque opaque;
OffsetNumber offnum;
uint16 itemIndex;
- buf = *bufP;
Assert(BufferIsValid(buf));
_hash_checkpage(rel, buf, LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
page = BufferGetPage(buf);
opaque = HashPageGetOpaque(page);
- so->currPos.buf = buf;
- so->currPos.currPage = BufferGetBlockNumber(buf);
+ batch->buf = buf;
+ batch->currPage = BufferGetBlockNumber(buf);
+ batch->dir = dir;
if (ScanDirectionIsForward(dir))
{
- BlockNumber prev_blkno = InvalidBlockNumber;
-
for (;;)
{
/* new page, locate starting position by binary search */
offnum = _hash_binsearch(page, so->hashso_sk_hash);
- itemIndex = _hash_load_qualified_items(scan, page, offnum, dir);
+ itemIndex = _hash_load_qualified_items(scan, page, offnum, dir,
+ batch);
if (itemIndex != 0)
break;
/*
- * Could not find any matching tuples in the current page, move to
- * the next page. Before leaving the current page, deal with any
- * killed items.
+ * Could not find any matching tuples in the current page, try to
+ * move to the next page
*/
- if (so->numKilled > 0)
- _hash_kill_items(scan);
-
- /*
- * If this is a primary bucket page, hasho_prevblkno is not a real
- * block number.
- */
- if (so->currPos.buf == so->hashso_bucket_buf ||
- so->currPos.buf == so->hashso_split_bucket_buf)
- prev_blkno = InvalidBlockNumber;
- else
- prev_blkno = opaque->hasho_prevblkno;
-
_hash_readnext(scan, &buf, &page, &opaque);
- if (BufferIsValid(buf))
+ if (!BufferIsValid(buf))
{
- so->currPos.buf = buf;
- so->currPos.currPage = BufferGetBlockNumber(buf);
- }
- else
- {
- /*
- * Remember next and previous block numbers for scrollable
- * cursors to know the start position and return false
- * indicating that no more matching tuples were found. Also,
- * don't reset currPage or lsn, because we expect
- * _hash_kill_items to be called for the old page after this
- * function returns.
- */
- so->currPos.prevPage = prev_blkno;
- so->currPos.nextPage = InvalidBlockNumber;
- so->currPos.buf = buf;
+ batch->buf = InvalidBuffer;
return false;
}
+
+ batch->buf = buf;
+ batch->currPage = BufferGetBlockNumber(buf);
}
- so->currPos.firstItem = 0;
- so->currPos.lastItem = itemIndex - 1;
- so->currPos.itemIndex = 0;
+ batch->firstItem = 0;
+ batch->lastItem = itemIndex - 1;
}
else
{
- BlockNumber next_blkno = InvalidBlockNumber;
-
for (;;)
{
/* new page, locate starting position by binary search */
offnum = _hash_binsearch_last(page, so->hashso_sk_hash);
- itemIndex = _hash_load_qualified_items(scan, page, offnum, dir);
+ itemIndex = _hash_load_qualified_items(scan, page, offnum, dir,
+ batch);
if (itemIndex != MaxIndexTuplesPerPage)
break;
/*
- * Could not find any matching tuples in the current page, move to
- * the previous page. Before leaving the current page, deal with
- * any killed items.
+ * Could not find any matching tuples in the current page, try to
+ * move to the previous page
*/
- if (so->numKilled > 0)
- _hash_kill_items(scan);
-
- if (so->currPos.buf == so->hashso_bucket_buf ||
- so->currPos.buf == so->hashso_split_bucket_buf)
- next_blkno = opaque->hasho_nextblkno;
-
_hash_readprev(scan, &buf, &page, &opaque);
- if (BufferIsValid(buf))
+ if (!BufferIsValid(buf))
{
- so->currPos.buf = buf;
- so->currPos.currPage = BufferGetBlockNumber(buf);
- }
- else
- {
- /*
- * Remember next and previous block numbers for scrollable
- * cursors to know the start position and return false
- * indicating that no more matching tuples were found. Also,
- * don't reset currPage or lsn, because we expect
- * _hash_kill_items to be called for the old page after this
- * function returns.
- */
- so->currPos.prevPage = InvalidBlockNumber;
- so->currPos.nextPage = next_blkno;
- so->currPos.buf = buf;
+ batch->buf = InvalidBuffer;
return false;
}
+
+ batch->buf = buf;
+ batch->currPage = BufferGetBlockNumber(buf);
}
- so->currPos.firstItem = itemIndex;
- so->currPos.lastItem = MaxIndexTuplesPerPage - 1;
- so->currPos.itemIndex = MaxIndexTuplesPerPage - 1;
+ batch->firstItem = itemIndex;
+ batch->lastItem = MaxIndexTuplesPerPage - 1;
}
- if (so->currPos.buf == so->hashso_bucket_buf ||
- so->currPos.buf == so->hashso_split_bucket_buf)
+ /*
+ * Saved at least one match in batch.items[]. Prepare for hashgetbatch to
+ * return it by initializing remaining uninitialized fields.
+ */
+ if (batch->buf == so->hashso_bucket_buf ||
+ batch->buf == so->hashso_split_bucket_buf)
{
- so->currPos.prevPage = InvalidBlockNumber;
- so->currPos.nextPage = opaque->hasho_nextblkno;
- LockBuffer(so->currPos.buf, BUFFER_LOCK_UNLOCK);
+ /*
+ * Batch's buffer is either the primary bucket, or a bucket being
+ * populated due to a split.
+ *
+ * Increment local reference count so that batch gets an independent
+ * buffer reference that can be released (by hashfreebatch) before the
+ * hashso_bucket_buf/hashso_split_bucket_buf references are released.
+ */
+ IncrBufferRefCount(batch->buf);
+
+ /* Can only use opaque->hasho_nextblkno */
+ batch->prevPage = InvalidBlockNumber;
+ batch->nextPage = opaque->hasho_nextblkno;
}
else
{
- so->currPos.prevPage = opaque->hasho_prevblkno;
- so->currPos.nextPage = opaque->hasho_nextblkno;
- _hash_relbuf(rel, so->currPos.buf);
- so->currPos.buf = InvalidBuffer;
+ /* Can use opaque->hasho_prevblkno and opaque->hasho_nextblkno */
+ batch->prevPage = opaque->hasho_prevblkno;
+ batch->nextPage = opaque->hasho_nextblkno;
}
- Assert(so->currPos.firstItem <= so->currPos.lastItem);
+ batch->moreLeft = BlockNumberIsValid(batch->prevPage);
+ batch->moreRight = BlockNumberIsValid(batch->nextPage);
+
+ /* Unlock (and likely unpin) buffer, per amgetbatch contract */
+ indexam_util_batch_unlock(scan, batch);
+
+ Assert(batch->firstItem <= batch->lastItem);
return true;
}
/*
* Load all the qualified items from a current index page
- * into so->currPos. Helper function for _hash_readpage.
+ * into batch. Helper function for _hash_readpage.
*/
static int
_hash_load_qualified_items(IndexScanDesc scan, Page page,
- OffsetNumber offnum, ScanDirection dir)
+ OffsetNumber offnum, ScanDirection dir,
+ BatchIndexScan batch)
{
HashScanOpaque so = (HashScanOpaque) scan->opaque;
IndexTuple itup;
@@ -640,7 +585,7 @@ _hash_load_qualified_items(IndexScanDesc scan, Page page,
_hash_checkqual(scan, itup))
{
/* tuple is qualified, so remember it */
- _hash_saveitem(so, itemIndex, offnum, itup);
+ _hash_saveitem(batch, itemIndex, offnum, itup);
itemIndex++;
}
else
@@ -687,7 +632,7 @@ _hash_load_qualified_items(IndexScanDesc scan, Page page,
{
itemIndex--;
/* tuple is qualified, so remember it */
- _hash_saveitem(so, itemIndex, offnum, itup);
+ _hash_saveitem(batch, itemIndex, offnum, itup);
}
else
{
@@ -706,13 +651,15 @@ _hash_load_qualified_items(IndexScanDesc scan, Page page,
}
}
-/* Save an index item into so->currPos.items[itemIndex] */
+/* Save an index item into batch->items[itemIndex] */
static inline void
-_hash_saveitem(HashScanOpaque so, int itemIndex,
+_hash_saveitem(BatchIndexScan batch, int itemIndex,
OffsetNumber offnum, IndexTuple itup)
{
- HashScanPosItem *currItem = &so->currPos.items[itemIndex];
+ BatchMatchingItem *currItem = &batch->items[itemIndex];
currItem->heapTid = itup->t_tid;
currItem->indexOffset = offnum;
+ currItem->tupleOffset = 0;
+ currItem->allVisible = false;
}
diff --git a/src/backend/access/hash/hashutil.c b/src/backend/access/hash/hashutil.c
index cf7f0b901..f99105d3b 100644
--- a/src/backend/access/hash/hashutil.c
+++ b/src/backend/access/hash/hashutil.c
@@ -510,81 +510,84 @@ _hash_get_newbucket_from_oldbucket(Relation rel, Bucket old_bucket,
* _hash_kill_items - set LP_DEAD state for items an indexscan caller has
* told us were killed.
*
- * scan->opaque, referenced locally through so, contains information about the
- * current page and killed tuples thereon (generally, this should only be
- * called if so->numKilled > 0).
+ * The batch parameter contains information about the current page and killed
+ * tuples thereon (this should only be called if batch->numKilled > 0).
*
- * The caller does not have a lock on the page and may or may not have the
- * page pinned in a buffer. Note that read-lock is sufficient for setting
- * LP_DEAD status (which is only a hint).
+ * Caller should not have a lock on the batch position's page, but must hold a
+ * buffer pin when !dropPin. When we return, it still won't be locked. It'll
+ * continue to hold whatever pins were held before calling here.
*
- * The caller must have pin on bucket buffer, but may or may not have pin
- * on overflow buffer, as indicated by HashScanPosIsPinned(so->currPos).
- *
- * We match items by heap TID before assuming they are the right ones to
- * delete.
- *
- * There are never any scans active in a bucket at the time VACUUM begins,
- * because VACUUM takes a cleanup lock on the primary bucket page and scans
- * hold a pin. A scan can begin after VACUUM leaves the primary bucket page
- * but before it finishes the entire bucket, but it can never pass VACUUM,
- * because VACUUM always locks the next page before releasing the lock on
- * the previous one. Therefore, we don't have to worry about accidentally
- * killing a TID that has been reused for an unrelated tuple.
+ * We match items by heap TID before assuming they are the right ones to set
+ * LP_DEAD. If the scan is one that holds a buffer pin on the target page
+ * continuously from initially reading the items until applying this function
+ * (if it is a !dropPin scan), VACUUM cannot have deleted any items on the
+ * page, so the page's TIDs can't have been recycled by now. There's no risk
+ * that we'll confuse a new index tuple that happens to use a recycled TID
+ * with a now-removed tuple with the same TID (that used to be on this same
+ * page). We can't rely on that during scans that drop buffer pins eagerly
+ * (i.e. dropPin scans), though, so we must condition setting LP_DEAD bits on
+ * the page LSN having not changed since back when _hash_readpage saw the page.
+ * We totally give up on setting LP_DEAD bits when the page LSN changed.
*/
void
-_hash_kill_items(IndexScanDesc scan)
+_hash_kill_items(IndexScanDesc scan, BatchIndexScan batch)
{
- HashScanOpaque so = (HashScanOpaque) scan->opaque;
Relation rel = scan->indexRelation;
- BlockNumber blkno;
Buffer buf;
Page page;
HashPageOpaque opaque;
OffsetNumber offnum,
maxoff;
- int numKilled = so->numKilled;
- int i;
bool killedsomething = false;
- bool havePin = false;
- Assert(so->numKilled > 0);
- Assert(so->killedItems != NULL);
- Assert(HashScanPosIsValid(so->currPos));
+ Assert(batch->numKilled > 0);
+ Assert(batch->killedItems != NULL);
+ Assert(BlockNumberIsValid(batch->currPage));
- /*
- * Always reset the scan state, so we don't look for same items on other
- * pages.
- */
- so->numKilled = 0;
-
- blkno = so->currPos.currPage;
- if (HashScanPosIsPinned(so->currPos))
+ if (!scan->dropPin)
{
/*
- * We already have pin on this buffer, so, all we need to do is
- * acquire lock on it.
+ * We have held the pin on this page since we read the index tuples,
+ * so all we need to do is lock it. The pin will have prevented
+ * concurrent VACUUMs from recycling any of the TIDs on the page.
*/
- havePin = true;
- buf = so->currPos.buf;
+ buf = batch->buf;
LockBuffer(buf, BUFFER_LOCK_SHARE);
}
else
- buf = _hash_getbuf(rel, blkno, HASH_READ, LH_OVERFLOW_PAGE);
+ {
+ XLogRecPtr latestlsn;
+
+ Assert(RelationNeedsWAL(rel));
+ buf = _hash_getbuf(rel, batch->currPage, HASH_READ,
+ LH_BUCKET_PAGE | LH_OVERFLOW_PAGE);
+
+ latestlsn = BufferGetLSNAtomic(buf);
+ Assert(batch->lsn <= latestlsn);
+ if (batch->lsn != latestlsn)
+ {
+ /* Modified, give up on hinting */
+ _hash_relbuf(rel, buf);
+ return;
+ }
+
+ /* Unmodified, hinting is safe */
+ }
page = BufferGetPage(buf);
opaque = HashPageGetOpaque(page);
maxoff = PageGetMaxOffsetNumber(page);
- for (i = 0; i < numKilled; i++)
+ /* Iterate through batch->killedItems[] in index page order */
+ for (int i = 0; i < batch->numKilled; i++)
{
- int itemIndex = so->killedItems[i];
- HashScanPosItem *currItem = &so->currPos.items[itemIndex];
+ int itemIndex = batch->killedItems[i];
+ BatchMatchingItem *currItem = &batch->items[itemIndex];
offnum = currItem->indexOffset;
- Assert(itemIndex >= so->currPos.firstItem &&
- itemIndex <= so->currPos.lastItem);
+ Assert(itemIndex >= batch->firstItem &&
+ itemIndex <= batch->lastItem);
while (offnum <= maxoff)
{
@@ -613,9 +616,8 @@ _hash_kill_items(IndexScanDesc scan)
MarkBufferDirtyHint(buf, true);
}
- if (so->hashso_bucket_buf == so->currPos.buf ||
- havePin)
- LockBuffer(so->currPos.buf, BUFFER_LOCK_UNLOCK);
+ if (!scan->dropPin)
+ LockBuffer(buf, BUFFER_LOCK_UNLOCK);
else
_hash_relbuf(rel, buf);
}
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index d25ba3412..3173b3330 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -1195,8 +1195,6 @@ HashPageStat
HashPath
HashScanOpaque
HashScanOpaqueData
-HashScanPosData
-HashScanPosItem
HashSkewBucket
HashState
HashValueFunc
--
2.51.0
[application/octet-stream] v7-0002-Add-prefetching-to-index-scans-using-batch-interf.patch (27.5K, 3-v7-0002-Add-prefetching-to-index-scans-using-batch-interf.patch)
download | inline diff:
From 02ab6053d9ea64b46a69a55fe197526d0c95956e Mon Sep 17 00:00:00 2001
From: Peter Geoghegan <[email protected]>
Date: Sat, 15 Nov 2025 14:03:58 -0500
Subject: [PATCH v7 2/4] Add prefetching to index scans using batch interfaces.
This commit implements I/O prefetching for index scans, made possible by
the recent addition of batching interfaces to both the table AM and
index AM APIs.
The amgetbatch index AM interface provides batches of TIDs (rather than
one at a time) from a single index leaf page, and allows multiple
batches to be held in memory/pinned simultaneously. This gives the
table AM the freedom to readahead within an index scan, which is crucial
for I/O prefetching with certain workloads (workloads that would
otherwise be unable to keep a sufficiently high prefetch distance for
heap block I/O). Prefetching is implemented using a read stream under
the control of the table AM.
XXX When the batch ring buffer reaches capacity, the stream pauses until
the scan catches up and frees some batches. We need a more principled
approach here. Essentially, we need infrastructure that allows a read
stream call back to tell the read stream to temporarily yield without it
fully ending/resetting the read stream.
Author: Tomas Vondra <[email protected]>
Author: Peter Geoghegan <[email protected]>
Reviewed-By: Andres Freund <[email protected]>
Reviewed-By: Thomas Munro <[email protected]>
Discussion: https://postgr.es/m/[email protected]
---
src/include/access/relscan.h | 41 ++-
src/include/optimizer/cost.h | 1 +
src/backend/access/heap/heapam_handler.c | 330 +++++++++++++++++-
src/backend/access/index/indexam.c | 10 +-
src/backend/access/index/indexbatch.c | 17 +-
src/backend/optimizer/path/costsize.c | 1 +
src/backend/storage/aio/read_stream.c | 14 +-
src/backend/utils/misc/guc_parameters.dat | 7 +
src/backend/utils/misc/postgresql.conf.sample | 1 +
src/test/regress/expected/sysviews.out | 3 +-
10 files changed, 418 insertions(+), 7 deletions(-)
diff --git a/src/include/access/relscan.h b/src/include/access/relscan.h
index ca1207be6..a517abe08 100644
--- a/src/include/access/relscan.h
+++ b/src/include/access/relscan.h
@@ -20,6 +20,7 @@
#include "nodes/tidbitmap.h"
#include "port/atomics.h"
#include "storage/buf.h"
+#include "storage/read_stream.h"
#include "storage/relfilelocator.h"
#include "storage/spin.h"
#include "utils/relcache.h"
@@ -124,6 +125,7 @@ typedef struct ParallelBlockTableScanWorkerData *ParallelBlockTableScanWorker;
typedef struct IndexFetchTableData
{
Relation rel;
+ ReadStream *rs;
} IndexFetchTableData;
/*
@@ -221,8 +223,14 @@ typedef struct BatchIndexScanData *BatchIndexScan;
* Maximum number of batches (leaf pages) we can keep in memory. We need a
* minimum of two, since we'll only consider releasing one batch when another
* is read.
+ *
+ * The choice of 64 batches is arbitrary. It's about 1MB of data with 8KB
+ * pages (512kB for pages, and then a bit of overhead). We should not really
+ * need this many batches in most cases, though. The read stream looks ahead
+ * just enough to queue enough IOs, adjusting the distance (TIDs, but
+ * ultimately the number of future batches) to meet that.
*/
-#define INDEX_SCAN_MAX_BATCHES 2
+#define INDEX_SCAN_MAX_BATCHES 64
#define INDEX_SCAN_CACHE_BATCHES 2
#define INDEX_SCAN_BATCH_COUNT(scan) \
((scan)->ringbuf->nextBatch - (scan)->ringbuf->headBatch)
@@ -270,15 +278,46 @@ typedef struct BatchIndexScanData *BatchIndexScan;
* matches in. However, table AMs are free to fetch table tuples in whatever
* order is most convenient/efficient -- provided that such reordering cannot
* affect the order that table_index_getnext_slot later returns tuples in.
+ *
+ * This data structure also provides table AMs with a way to read ahead of the
+ * current read position by _multiple_ batches/index pages. The further out
+ * the table AM reads ahead like this, the further it can see into the future.
+ * That way the table AM is able to reorder work as aggressively as desired.
+ * For example, index scans sometimes need to readahead by as many as a few
+ * dozen amgetbatch batches in order to maintain an optimal I/O prefetch
+ * distance (distance for reading table blocks/fetching table tuples).
*/
typedef struct BatchRingBuffer
{
+ bool reset;
+
+ /*
+ * Did we disable prefetching/use of a read stream because it didn't pay
+ * for itself?
+ */
+ bool prefetchingLockedIn;
+ bool disabled;
+
+ /*
+ * During prefetching, currentPrefetchBlock is the table AM block number
+ * that was returned by our read stream callback most recently. Used to
+ * suppress duplicate successive read stream block requests.
+ *
+ * Prefetching can still perform non-successive requests for the same
+ * block number (in general we're prefetching in exactly the same order
+ * that the scan will return table AM TIDs in). We need to avoid
+ * duplicate successive requests because table AMs expect to be able to
+ * hang on to buffer pins across table_index_fetch_tuple calls.
+ */
+ BlockNumber currentPrefetchBlock;
+
/* Current scan direction, for the currently loaded batches */
ScanDirection direction;
/* current positions in batches[] for scan */
BatchRingItemPos scanPos; /* scan's read position */
BatchRingItemPos markPos; /* mark/restore position */
+ BatchRingItemPos prefetchPos; /* prefetching position */
BatchIndexScan markBatch;
diff --git a/src/include/optimizer/cost.h b/src/include/optimizer/cost.h
index 07b8bfa63..31236ceac 100644
--- a/src/include/optimizer/cost.h
+++ b/src/include/optimizer/cost.h
@@ -51,6 +51,7 @@ extern PGDLLIMPORT Cost disable_cost;
extern PGDLLIMPORT int max_parallel_workers_per_gather;
extern PGDLLIMPORT bool enable_seqscan;
extern PGDLLIMPORT bool enable_indexscan;
+extern PGDLLIMPORT bool enable_indexscan_prefetch;
extern PGDLLIMPORT bool enable_indexonlyscan;
extern PGDLLIMPORT bool enable_bitmapscan;
extern PGDLLIMPORT bool enable_tidscan;
diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index 03b4fada9..deef73069 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -37,6 +37,7 @@
#include "commands/progress.h"
#include "executor/executor.h"
#include "miscadmin.h"
+#include "optimizer/cost.h"
#include "pgstat.h"
#include "storage/bufmgr.h"
#include "storage/bufpage.h"
@@ -60,6 +61,9 @@ static BlockNumber heapam_scan_get_blocks_done(HeapScanDesc hscan);
static bool BitmapHeapScanNextBlock(TableScanDesc scan,
bool *recheck,
uint64 *lossy_pages, uint64 *exact_pages);
+static BlockNumber heapam_getnext_stream(ReadStream *stream,
+ void *callback_private_data,
+ void *per_buffer_data);
/* ------------------------------------------------------------------------
@@ -85,6 +89,7 @@ heapam_index_fetch_begin(Relation rel)
IndexFetchHeapData *hscan = palloc_object(IndexFetchHeapData);
hscan->xs_base.rel = rel;
+ hscan->xs_base.rs = NULL;
hscan->xs_cbuf = InvalidBuffer;
hscan->xs_blk = InvalidBlockNumber;
hscan->vmbuf = InvalidBuffer;
@@ -97,6 +102,9 @@ heapam_index_fetch_reset(IndexFetchTableData *scan)
{
IndexFetchHeapData *hscan = (IndexFetchHeapData *) scan;
+ if (scan->rs)
+ read_stream_reset(scan->rs);
+
/* deliberately don't drop VM buffer pin here */
if (BufferIsValid(hscan->xs_cbuf))
{
@@ -113,6 +121,9 @@ heapam_index_fetch_end(IndexFetchTableData *scan)
heapam_index_fetch_reset(scan);
+ if (scan->rs)
+ read_stream_end(scan->rs);
+
if (hscan->vmbuf != InvalidBuffer)
{
ReleaseBuffer(hscan->vmbuf);
@@ -150,7 +161,13 @@ heapam_index_fetch_tuple(struct IndexFetchTableData *scan,
* When using a read stream, the stream will already know which block
* number comes next (though an assertion will verify a match below)
*/
- hscan->xs_cbuf = ReadBuffer(hscan->xs_base.rel, hscan->xs_blk);
+ if (scan->rs)
+ hscan->xs_cbuf = read_stream_next_buffer(scan->rs, NULL);
+ else
+ hscan->xs_cbuf = ReadBuffer(hscan->xs_base.rel, hscan->xs_blk);
+
+ Assert(BufferIsValid(hscan->xs_cbuf));
+ Assert(BufferGetBlockNumber(hscan->xs_cbuf) == ItemPointerGetBlockNumber(tid));
/*
* Prune page when it is pinned for the first time
@@ -248,6 +265,15 @@ heapam_batch_return_tid(IndexScanDesc scan, BatchIndexScan scanBatch,
/*
* heap_batch_resolve_visibility
* Obtain visibility information for every TID from caller's batch.
+ *
+ * heapam_batch_getnext_tid must reliably agree with heapam_getnext_stream
+ * about which heap blocks/TIDs will require a heap fetch (and which TIDs
+ * won't due to pointing to an all-visible heap page). Otherwise we risk
+ * allowing the read stream to return unexpected heap buffers/pages.
+ *
+ * Caching visibility information up front avoids that problem. If a VM bit
+ * is concurrently set (or unset), it can't matter, since everybody will have
+ * works off of this immutable local cache.
*/
static void
heap_batch_resolve_visibility(IndexScanDesc scan, BatchIndexScan batch)
@@ -377,6 +403,19 @@ heap_batch_getnext(IndexScanDesc scan, BatchIndexScan priorbatch,
DEBUG_LOG("batch_getnext headBatch %d nextBatch %d batch %p",
ringbuf->headBatch, ringbuf->nextBatch, batch);
+
+ /* Delay initializing stream until reading from scan's second batch */
+ if (!scan->xs_heapfetch->rs && !ringbuf->disabled && priorbatch &&
+ enable_indexscan_prefetch)
+ {
+ Assert(INDEX_SCAN_POS_INVALID(&ringbuf->prefetchPos));
+ Assert(ringbuf->currentPrefetchBlock == InvalidBlockNumber);
+
+ scan->xs_heapfetch->rs =
+ read_stream_begin_relation(READ_STREAM_DEFAULT, NULL,
+ scan->heapRelation, MAIN_FORKNUM,
+ heapam_getnext_stream, scan, 0);
+ }
}
/* xs_hitup is not supported by amgetbatch scans */
@@ -411,9 +450,53 @@ heapam_batch_getnext_tid(IndexScanDesc scan, ScanDirection direction)
/* Initialize direction on first call */
if (ringbuf->direction == NoMovementScanDirection)
ringbuf->direction = direction;
+ else if (unlikely(ringbuf->disabled && scan->xs_heapfetch->rs))
+ {
+ /*
+ * Handle cancelling the use of the read stream for prefetching
+ */
+ batch_reset_pos(&ringbuf->prefetchPos);
+ ringbuf->currentPrefetchBlock = InvalidBlockNumber;
+ read_stream_reset(scan->xs_heapfetch->rs);
+ scan->xs_heapfetch->rs = NULL;
+ }
+ else if (unlikely(ringbuf->reset))
+ {
+ ringbuf->reset = false;
+
+ /*
+ * Need to reset the stream position, it might be too far behind.
+ * Ultimately we want to set it to scanPos, but we can't do that yet -
+ * scanPos still point sat the old batch, so just reset it and we'll
+ * init it to scanPos later in the callback.
+ */
+ batch_reset_pos(&ringbuf->prefetchPos);
+ ringbuf->currentPrefetchBlock = InvalidBlockNumber;
+
+ if (scan->xs_heapfetch->rs)
+ read_stream_reset(scan->xs_heapfetch->rs);
+ }
+
+ /*
+ * XXX Shouldn't this also update the ringbuf->direction? If we get to the
+ * next block hangling direction change, then we will remember it (because
+ * heapam_batch_rewind will store it). But if we return in the next block,
+ * won't we forget about it?
+ *
+ * XXX It's a bit weird we handle the direction change in two places.
+ * Would be good to explain why that's necessary.
+ *
+ * XXX How come this doesn't need to do heapam_batch_rewind too? Could
+ * there be some future batches already loaded?
+ */
if (unlikely(ringbuf->direction != direction))
{
+ if (scan->xs_heapfetch->rs)
+ read_stream_reset(scan->xs_heapfetch->rs);
+ batch_reset_pos(&ringbuf->prefetchPos);
+ ringbuf->currentPrefetchBlock = InvalidBlockNumber;
+
/* We may change direction after reading the last batch. */
scan->finished = false;
}
@@ -497,6 +580,251 @@ heapam_batch_getnext_tid(IndexScanDesc scan, ScanDirection direction)
return heapam_batch_return_tid(scan, scanBatch, scanPos);
}
+/*
+ * Controls when we cancel use of a read stream to do prefetching
+ */
+#define INDEX_SCAN_MIN_DISTANCE_NBATCHES 20
+#define INDEX_SCAN_MIN_TUPLE_DISTANCE 7
+
+/*
+ * heapam_getnext_stream
+ * return the next block to pass to the read stream
+ *
+ * The initial batch is always loaded by heapam_batch_getnext_tid. We don't
+ * get called until the first read_stream_next_buffer() call, when a heap
+ * block is requested from the scan's stream for the first time.
+ *
+ * The position of the read_stream is stored in prefetchPos. It is typical for
+ * prefetchPos to consistently stay ahead of the scanPos position that's used to
+ * track the next TID to be returned to the scan by heapam_batch_getnext_tid
+ * after the first time we get called. However, that isn't a precondition.
+ * There is a strict postcondition, though: when we return we'll always leave
+ * scanPos <= prefetchPos (except in cases where we return InvalidBlockNumber).
+ */
+static BlockNumber
+heapam_getnext_stream(ReadStream *stream, void *callback_private_data,
+ void *per_buffer_data)
+{
+ IndexScanDesc scan = (IndexScanDesc) callback_private_data;
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+ BatchRingItemPos *scanPos = &ringbuf->scanPos;
+ BatchRingItemPos *prefetchPos = &ringbuf->prefetchPos;
+ ScanDirection direction = ringbuf->direction;
+ BatchIndexScan prefetchBatch;
+ bool fromReadPos = false;
+
+ Assert(!scan->finished && !ringbuf->disabled);
+
+ /*
+ * scanPos must always be valid when we're called -- there has to be at
+ * least one batch, loaded, for scanBatch. prefetchPos might not yet be
+ * valid, in which case it'll be initialized using scanPos.
+ */
+ Assert(INDEX_SCAN_BATCH_COUNT(scan) > 0);
+ batch_assert_pos_valid(scan, scanPos);
+
+ /*
+ * It is possible for the scan's direction to change, but that's handled
+ * elsewhere. We don't know how to deal with any variation in scan
+ * direction here. We assume that all loaded and newly requested batches
+ * must use the same scan direction.
+ */
+ Assert(direction != NoMovementScanDirection);
+
+ /*
+ * If the stream position has not been initialized yet, initialize it
+ * using the current read position.
+ *
+ * We do this now (rather than doing it when the read stream is created)
+ * to avoid incorrectly returning the scan's IndexFetchHeapData.xs_blk as
+ * it was at the time of read stream creation. Note that the scan might
+ * have to hold onto its existing not-managed-by-read-stream buffer pin
+ * after the read stream is created; there'll often be a few more heap
+ * TIDs that point to the same pinned heap page from before.
+ */
+ if (INDEX_SCAN_POS_INVALID(prefetchPos))
+ {
+ Assert(ringbuf->currentPrefetchBlock == InvalidBlockNumber);
+
+ *prefetchPos = *scanPos;
+ fromReadPos = true;
+ }
+
+ prefetchBatch = INDEX_SCAN_BATCH(scan, prefetchPos->batch);
+ for (;;)
+ {
+ BatchMatchingItem *item;
+ ItemPointer tid;
+
+ if (fromReadPos)
+ {
+ /*
+ * Don't increment item when prefetchPos was just initialized
+ * using scanPos. We return the scanPos item's heap block
+ * directly on the first call here.
+ */
+ fromReadPos = false;
+ }
+ else if (!heap_batchpos_advance(prefetchBatch, prefetchPos, direction))
+ {
+ /*
+ * Ran out of items from prefetchBatch. Try to advance it to next
+ * batch.
+ */
+ if (INDEX_SCAN_BATCH_LOADED(scan, prefetchPos->batch + 1))
+ {
+ /*
+ * The next batch was already loaded for us.
+ *
+ * Typically, prefetchPos is ahead of scanPos for the entire
+ * duration of the scan (at least after we're first called).
+ * However, prefetchPos can sometimes fall behind scanPos.
+ * That's why we need to handle already-loaded batches here.
+ *
+ * This happens when some blocks are skipped and not returned
+ * to the read_stream. An example is an index scan on a
+ * correlated index, with many duplicate blocks are skipped,
+ * or an IOS where all-visible blocks are skipped.
+ */
+ prefetchBatch = INDEX_SCAN_BATCH(scan, prefetchPos->batch + 1);
+ }
+ else
+ {
+ /*
+ * If we already used the maximum number of batch slots
+ * available, it's pointless to try loading another one. This
+ * can happen for various reasons, e.g. for index-only scans
+ * on all-visible table, or skipping duplicate blocks on
+ * perfectly correlated indexes, etc.
+ *
+ * We could enlarge the array to allow more batches, but
+ * that's futile, we can always construct a case using more
+ * memory. Not only it would risk OOM, it'd also be
+ * inefficient because this happens early in the scan (so it'd
+ * interfere with LIMIT queries).
+ */
+ if (INDEX_SCAN_BATCH_FULL(scan))
+ {
+ DEBUG_LOG("batch_getnext: ran out of space for batches");
+ scan->ringbuf->reset = true;
+ break;
+ }
+
+ prefetchBatch = heap_batch_getnext(scan, prefetchBatch, direction);
+ if (!prefetchBatch)
+ {
+ /*
+ * Failed to load next batch, so all the batches that the
+ * scan will ever require (barring a change in scan
+ * direction) are now loaded
+ */
+ scan->finished = true;
+ break;
+ }
+
+ /*
+ * Consider disabling prefetching when we can't keep a
+ * sufficiently large "index tuple distance" between scanPos
+ * and prefetchPos.
+ *
+ * Only consider doing this when we're not on the scan's
+ * initial batch, when scanPos and prefetchPos share the same
+ * batch.
+ */
+ if (!ringbuf->prefetchingLockedIn)
+ {
+ int itemdiff;
+
+ if (prefetchPos->batch <= INDEX_SCAN_MIN_DISTANCE_NBATCHES)
+ {
+ /*
+ * Too early to check if prefetching should be
+ * disabled
+ */
+ }
+ else if (scanPos->batch == prefetchPos->batch)
+ {
+ if (ScanDirectionIsForward(direction))
+ itemdiff = prefetchPos->item - scanPos->item;
+ else
+ {
+ BatchIndexScan scanBatch =
+ INDEX_SCAN_BATCH(scan, scanPos->batch);
+
+ itemdiff = (scanPos->item - scanBatch->firstItem) -
+ (prefetchPos->item - scanBatch->firstItem);
+ }
+
+ if (itemdiff < INDEX_SCAN_MIN_TUPLE_DISTANCE)
+ {
+ ringbuf->disabled = true;
+ return InvalidBlockNumber;
+ }
+ else
+ {
+ ringbuf->prefetchingLockedIn = true;
+ }
+ }
+ else
+ ringbuf->prefetchingLockedIn = true;
+ }
+ }
+
+ /* Position prefetchPos to the start of new prefetchBatch */
+ heap_batchpos_newbatch(prefetchBatch, prefetchPos, direction);
+ Assert(INDEX_SCAN_BATCH(scan, prefetchPos->batch) == prefetchBatch);
+ }
+
+ /*
+ * We advanced the position. Either return the block for the TID, or
+ * skip it (and then try advancing again).
+ */
+ Assert(prefetchBatch->dir == direction);
+ Assert(scanPos->batch < prefetchPos->batch ||
+ (scanPos->batch == prefetchPos->batch &&
+ ScanDirectionIsForward(direction) ?
+ scanPos->item <= prefetchPos->item :
+ scanPos->item >= prefetchPos->item));
+
+ /*
+ * The block may be "skipped" for two reasons. First, the caller may
+ * define a "prefetch" callback that tells us to skip items (IOS does
+ * this to skip all-visible pages). Second, currentPrefetchBlock is
+ * used to skip duplicate block numbers (a sequence of TIDS for the
+ * same block).
+ */
+ batch_assert_pos_valid(scan, prefetchPos);
+ item = &prefetchBatch->items[prefetchPos->item];
+ tid = &item->heapTid;
+
+ DEBUG_LOG("heapam_getnext_stream: item %d, TID (%u,%u)",
+ prefetchPos->item,
+ ItemPointerGetBlockNumber(tid),
+ ItemPointerGetOffsetNumber(tid));
+
+ /*
+ * For index-only scans, determine if the page is all-visible now. If
+ * it is, we won't need the block and can skip it too.
+ */
+ if (scan->xs_want_itup && item->allVisible)
+ continue;
+
+ /* same block as before, don't need to read it */
+ if (ringbuf->currentPrefetchBlock == ItemPointerGetBlockNumber(tid))
+ {
+ DEBUG_LOG("heapam_getnext_stream: skip block (currentPrefetchBlock)");
+ continue;
+ }
+
+ ringbuf->currentPrefetchBlock = ItemPointerGetBlockNumber(tid);
+
+ return ringbuf->currentPrefetchBlock;
+ }
+
+ /* no more items in this scan */
+ return InvalidBlockNumber;
+}
+
/* ----------------
* index_fetch_heap - get the scan's next heap tuple
*
diff --git a/src/backend/access/index/indexam.c b/src/backend/access/index/indexam.c
index 9eaecd943..d15a2ee81 100644
--- a/src/backend/access/index/indexam.c
+++ b/src/backend/access/index/indexam.c
@@ -467,7 +467,15 @@ index_restrpos(IndexScanDesc scan)
CHECK_SCAN_PROCEDURE(amgetbatch);
CHECK_SCAN_PROCEDURE(amposreset);
- /* release resources (like buffer pins) from table accesses */
+ /*
+ * release resources (like buffer pins) from table accesses
+ *
+ * XXX: Currently, the distance is always remembered across any
+ * read_stream_reset calls (to work around the scan->ringbuf->reset
+ * behavior of resetting the stream to deal with running out of batches).
+ * We probably _should_ be forgetting the distance when we reset the
+ * stream here (through our table_index_fetch_reset call), though.
+ */
if (scan->xs_heapfetch)
table_index_fetch_reset(scan->xs_heapfetch);
diff --git a/src/backend/access/index/indexbatch.c b/src/backend/access/index/indexbatch.c
index 86a1c6f56..4df92e7af 100644
--- a/src/backend/access/index/indexbatch.c
+++ b/src/backend/access/index/indexbatch.c
@@ -74,11 +74,16 @@ index_batch_init(IndexScanDesc scan)
(!scan->xs_want_itup && IsMVCCSnapshot(scan->xs_snapshot) &&
RelationNeedsWAL(scan->indexRelation));
scan->finished = false;
+ scan->ringbuf->reset = false;
+ scan->ringbuf->prefetchingLockedIn = false;
+ scan->ringbuf->disabled = false;
+ scan->ringbuf->currentPrefetchBlock = InvalidBlockNumber;
scan->ringbuf->direction = NoMovementScanDirection;
/* positions in the ring buffer of batches */
batch_reset_pos(&scan->ringbuf->scanPos);
batch_reset_pos(&scan->ringbuf->markPos);
+ batch_reset_pos(&scan->ringbuf->prefetchPos);
scan->ringbuf->markBatch = NULL;
scan->ringbuf->headBatch = 0; /* initial head batch */
@@ -107,9 +112,12 @@ index_batch_reset(IndexScanDesc scan, bool complete)
batch_assert_batches_valid(scan);
batch_debug_print_batches("index_batch_reset", scan);
Assert(scan->xs_heapfetch);
+ if (scan->xs_heapfetch->rs)
+ read_stream_reset(scan->xs_heapfetch->rs);
/* reset the positions */
batch_reset_pos(&ringbuf->scanPos);
+ batch_reset_pos(&ringbuf->prefetchPos);
/*
* With "complete" reset, make sure to also free the marked batch, either
@@ -153,6 +161,8 @@ index_batch_reset(IndexScanDesc scan, bool complete)
ringbuf->nextBatch = 0; /* initial batch is empty */
scan->finished = false;
+ ringbuf->reset = false;
+ ringbuf->currentPrefetchBlock = InvalidBlockNumber;
batch_assert_batches_valid(scan);
}
@@ -213,9 +223,13 @@ index_batch_restore_pos(IndexScanDesc scan)
{
BatchRingBuffer *ringbuf = scan->ringbuf;
BatchRingItemPos *markPos = &ringbuf->markPos;
- BatchRingItemPos *scanPos = &ringbuf->scanPos ;
BatchIndexScan markBatch = ringbuf->markBatch;
+ /*
+ * XXX Disable this optimization when I/O prefetching is in use, at least
+ * until the possible interactions with prefetchPos are fully understood.
+ */
+#if 0
if (scanPos->batch == markPos->batch &&
scanPos->batch == ringbuf->headBatch)
{
@@ -226,6 +240,7 @@ index_batch_restore_pos(IndexScanDesc scan)
scanPos->item = markPos->item;
return;
}
+#endif
/*
* Call amposreset to let index AM know to invalidate any private state
diff --git a/src/backend/optimizer/path/costsize.c b/src/backend/optimizer/path/costsize.c
index 16bf1f61a..23e7c0a2f 100644
--- a/src/backend/optimizer/path/costsize.c
+++ b/src/backend/optimizer/path/costsize.c
@@ -144,6 +144,7 @@ int max_parallel_workers_per_gather = 2;
bool enable_seqscan = true;
bool enable_indexscan = true;
+bool enable_indexscan_prefetch = true;
bool enable_indexonlyscan = true;
bool enable_bitmapscan = true;
bool enable_tidscan = true;
diff --git a/src/backend/storage/aio/read_stream.c b/src/backend/storage/aio/read_stream.c
index 88717c2ff..7463651e0 100644
--- a/src/backend/storage/aio/read_stream.c
+++ b/src/backend/storage/aio/read_stream.c
@@ -99,6 +99,7 @@ struct ReadStream
int16 forwarded_buffers;
int16 pinned_buffers;
int16 distance;
+ int16 distance_old;
int16 initialized_buffers;
int read_buffers_flags;
bool sync_mode; /* using io_method=sync */
@@ -464,6 +465,7 @@ read_stream_look_ahead(ReadStream *stream)
if (blocknum == InvalidBlockNumber)
{
/* End of stream. */
+ stream->distance_old = stream->distance;
stream->distance = 0;
break;
}
@@ -862,6 +864,7 @@ read_stream_next_buffer(ReadStream *stream, void **per_buffer_data)
else
{
/* No more blocks, end of stream. */
+ stream->distance_old = stream->distance;
stream->distance = 0;
stream->oldest_buffer_index = stream->next_buffer_index;
stream->pinned_buffers = 0;
@@ -1046,6 +1049,9 @@ read_stream_reset(ReadStream *stream)
int16 index;
Buffer buffer;
+ /* remember the old distance (if we reset before end of the stream) */
+ stream->distance_old = Max(stream->distance, stream->distance_old);
+
/* Stop looking ahead. */
stream->distance = 0;
@@ -1078,8 +1084,12 @@ read_stream_reset(ReadStream *stream)
Assert(stream->pinned_buffers == 0);
Assert(stream->ios_in_progress == 0);
- /* Start off assuming data is cached. */
- stream->distance = 1;
+ /*
+ * Restore the old distance, if we have one. Otherwise start assuming data
+ * is cached.
+ */
+ stream->distance = Max(1, stream->distance_old);
+ stream->distance_old = 0;
}
/*
diff --git a/src/backend/utils/misc/guc_parameters.dat b/src/backend/utils/misc/guc_parameters.dat
index 7c60b1255..a99aa41db 100644
--- a/src/backend/utils/misc/guc_parameters.dat
+++ b/src/backend/utils/misc/guc_parameters.dat
@@ -891,6 +891,13 @@
boot_val => 'true',
},
+{ name => 'enable_indexscan_prefetch', type => 'bool', context => 'PGC_USERSET', group => 'QUERY_TUNING_METHOD',
+ short_desc => 'Enables prefetching for index scans and index-only-scans.',
+ flags => 'GUC_EXPLAIN',
+ variable => 'enable_indexscan_prefetch',
+ boot_val => 'true',
+},
+
{ name => 'enable_material', type => 'bool', context => 'PGC_USERSET', group => 'QUERY_TUNING_METHOD',
short_desc => 'Enables the planner\'s use of materialization.',
flags => 'GUC_EXPLAIN',
diff --git a/src/backend/utils/misc/postgresql.conf.sample b/src/backend/utils/misc/postgresql.conf.sample
index dc9e2255f..da50ae15f 100644
--- a/src/backend/utils/misc/postgresql.conf.sample
+++ b/src/backend/utils/misc/postgresql.conf.sample
@@ -412,6 +412,7 @@
#enable_incremental_sort = on
#enable_indexscan = on
#enable_indexonlyscan = on
+#enable_indexscan_prefetch = on
#enable_material = on
#enable_memoize = on
#enable_mergejoin = on
diff --git a/src/test/regress/expected/sysviews.out b/src/test/regress/expected/sysviews.out
index 3dd63fd88..b5628736b 100644
--- a/src/test/regress/expected/sysviews.out
+++ b/src/test/regress/expected/sysviews.out
@@ -159,6 +159,7 @@ select name, setting from pg_settings where name like 'enable%';
enable_incremental_sort | on
enable_indexonlyscan | on
enable_indexscan | on
+ enable_indexscan_prefetch | on
enable_material | on
enable_memoize | on
enable_mergejoin | on
@@ -173,7 +174,7 @@ select name, setting from pg_settings where name like 'enable%';
enable_seqscan | on
enable_sort | on
enable_tidscan | on
-(25 rows)
+(26 rows)
-- There are always wait event descriptions for various types. InjectionPoint
-- may be present or absent, depending on history since last postmaster start.
--
2.51.0
[application/octet-stream] v7-0003-bufmgr-aio-Prototype-for-not-waiting-for-already-.patch (6.9K, 4-v7-0003-bufmgr-aio-Prototype-for-not-waiting-for-already-.patch)
download | inline diff:
From fe184933383163d89d859fe7cf759704626c1154 Mon Sep 17 00:00:00 2001
From: Andres Freund <[email protected]>
Date: Fri, 15 Aug 2025 11:01:52 -0400
Subject: [PATCH v7 3/4] bufmgr: aio: Prototype for not waiting for
already-in-progress IO
Author:
Reviewed-by:
Discussion: https://postgr.es/m/zljergweqti7x67lg5ije2rzjusie37nslsnkjkkby4laqqbfw@3p3zu522yykv
Backpatch:
---
src/include/storage/bufmgr.h | 1 +
src/backend/storage/buffer/bufmgr.c | 150 ++++++++++++++++++++++++++--
2 files changed, 142 insertions(+), 9 deletions(-)
diff --git a/src/include/storage/bufmgr.h b/src/include/storage/bufmgr.h
index 715ae96f0..0b6fa848a 100644
--- a/src/include/storage/bufmgr.h
+++ b/src/include/storage/bufmgr.h
@@ -147,6 +147,7 @@ struct ReadBuffersOperation
int flags;
int16 nblocks;
int16 nblocks_done;
+ bool foreign_io;
PgAioWaitRef io_wref;
PgAioReturn io_return;
};
diff --git a/src/backend/storage/buffer/bufmgr.c b/src/backend/storage/buffer/bufmgr.c
index a036c2aa2..d3dd64808 100644
--- a/src/backend/storage/buffer/bufmgr.c
+++ b/src/backend/storage/buffer/bufmgr.c
@@ -1632,6 +1632,46 @@ ReadBuffersCanStartIOOnce(Buffer buffer, bool nowait)
return StartBufferIO(GetBufferDescriptor(buffer - 1), true, nowait);
}
+/*
+ * Check if the buffer is already undergoing read AIO. If it is, assign the
+ * IO's wait reference to operation->io_wref, thereby allowing the caller to
+ * wait for that IO.
+ */
+static inline bool
+ReadBuffersIOAlreadyInProgress(ReadBuffersOperation *operation, Buffer buffer)
+{
+ BufferDesc *desc;
+ uint32 buf_state;
+ PgAioWaitRef iow;
+
+ pgaio_wref_clear(&iow);
+
+ if (BufferIsLocal(buffer))
+ {
+ desc = GetLocalBufferDescriptor(-buffer - 1);
+ buf_state = pg_atomic_read_u32(&desc->state);
+ if ((buf_state & BM_IO_IN_PROGRESS) && !(buf_state & BM_VALID))
+ iow = desc->io_wref;
+ }
+ else
+ {
+ desc = GetBufferDescriptor(buffer - 1);
+ buf_state = LockBufHdr(desc);
+
+ if ((buf_state & BM_IO_IN_PROGRESS) && !(buf_state & BM_VALID))
+ iow = desc->io_wref;
+ UnlockBufHdr(desc);
+ }
+
+ if (pgaio_wref_valid(&iow))
+ {
+ operation->io_wref = iow;
+ return true;
+ }
+
+ return false;
+}
+
/*
* Helper for AsyncReadBuffers that tries to get the buffer ready for IO.
*/
@@ -1764,7 +1804,7 @@ WaitReadBuffers(ReadBuffersOperation *operation)
*
* we first check if we already know the IO is complete.
*/
- if (aio_ret->result.status == PGAIO_RS_UNKNOWN &&
+ if ((operation->foreign_io || aio_ret->result.status == PGAIO_RS_UNKNOWN) &&
!pgaio_wref_check_done(&operation->io_wref))
{
instr_time io_start = pgstat_prepare_io_time(track_io_timing);
@@ -1783,11 +1823,66 @@ WaitReadBuffers(ReadBuffersOperation *operation)
Assert(pgaio_wref_check_done(&operation->io_wref));
}
- /*
- * We now are sure the IO completed. Check the results. This
- * includes reporting on errors if there were any.
- */
- ProcessReadBuffersResult(operation);
+ if (unlikely(operation->foreign_io))
+ {
+ Buffer buffer = operation->buffers[operation->nblocks_done];
+ BufferDesc *desc;
+ uint32 buf_state;
+
+ if (BufferIsLocal(buffer))
+ {
+ desc = GetLocalBufferDescriptor(-buffer - 1);
+ buf_state = pg_atomic_read_u32(&desc->state);
+ }
+ else
+ {
+ desc = GetBufferDescriptor(buffer - 1);
+ buf_state = LockBufHdr(desc);
+ UnlockBufHdr(desc);
+ }
+
+ if (buf_state & BM_VALID)
+ {
+ operation->nblocks_done += 1;
+ Assert(operation->nblocks_done <= operation->nblocks);
+
+ /*
+ * Report and track this as a 'hit' for this backend, even
+ * though it must have started out as a miss in
+ * PinBufferForBlock(). The other backend (or ourselves,
+ * as part of a read started earlier) will track this as a
+ * 'read'.
+ */
+ TRACE_POSTGRESQL_BUFFER_READ_DONE(operation->forknum,
+ operation->blocknum + operation->nblocks_done,
+ operation->smgr->smgr_rlocator.locator.spcOid,
+ operation->smgr->smgr_rlocator.locator.dbOid,
+ operation->smgr->smgr_rlocator.locator.relNumber,
+ operation->smgr->smgr_rlocator.backend,
+ true);
+
+ if (BufferIsLocal(buffer))
+ pgBufferUsage.local_blks_hit += 1;
+ else
+ pgBufferUsage.shared_blks_hit += 1;
+
+ if (operation->rel)
+ pgstat_count_buffer_hit(operation->rel);
+
+ pgstat_count_io_op(io_object, io_context, IOOP_HIT, 1, 0);
+
+ if (VacuumCostActive)
+ VacuumCostBalance += VacuumCostPageHit;
+ }
+ }
+ else
+ {
+ /*
+ * We now are sure the IO completed. Check the results. This
+ * includes reporting on errors if there were any.
+ */
+ ProcessReadBuffersResult(operation);
+ }
}
/*
@@ -1873,6 +1968,43 @@ AsyncReadBuffers(ReadBuffersOperation *operation, int *nblocks_progress)
io_object = IOOBJECT_RELATION;
}
+ /*
+ * If AIO is in progress, be it in this backend or another backend, we
+ * just associate the wait reference with the operation and wait in
+ * WaitReadBuffers(). This turns out to be important for performance in
+ * two workloads:
+ *
+ * 1) A read stream that has to read the same block multiple times within
+ * the readahead distance. This can happen e.g. for the table accesses of
+ * an index scan.
+ *
+ * 2) Concurrent scans by multiple backends on the same relation.
+ *
+ * If we were to synchronously wait for the in-progress IO, we'd not be
+ * able to keep enough I/O in flight.
+ *
+ * If we do find there is ongoing I/O for the buffer, we set up a 1-block
+ * ReadBuffersOperation that WaitReadBuffers then can wait on.
+ *
+ * It's possible that another backend starts IO on the buffer between this
+ * check and the ReadBuffersCanStartIO(nowait = false) below. In that case
+ * we will synchronously wait for the IO below, but the window for that is
+ * small enough that it won't happen often enough to have a significant
+ * performance impact.
+ */
+ if (ReadBuffersIOAlreadyInProgress(operation, buffers[nblocks_done]))
+ {
+ *nblocks_progress = 1;
+ operation->foreign_io = true;
+
+ CheckReadBuffersOperation(operation, false);
+
+
+ return true;
+ }
+
+ operation->foreign_io = false;
+
/*
* If zero_damaged_pages is enabled, add the READ_BUFFERS_ZERO_ON_ERROR
* flag. The reason for that is that, hopefully, zero_damaged_pages isn't
@@ -1930,9 +2062,9 @@ AsyncReadBuffers(ReadBuffersOperation *operation, int *nblocks_progress)
/*
* Check if we can start IO on the first to-be-read buffer.
*
- * If an I/O is already in progress in another backend, we want to wait
- * for the outcome: either done, or something went wrong and we will
- * retry.
+ * If a synchronous I/O is in progress in another backend (it can't be
+ * this backend), we want to wait for the outcome: either done, or
+ * something went wrong and we will retry.
*/
if (!ReadBuffersCanStartIO(buffers[nblocks_done], false))
{
--
2.51.0
[application/octet-stream] v7-0001-Add-batching-interfaces-used-by-heapam-and-nbtree.patch (201.6K, 5-v7-0001-Add-batching-interfaces-used-by-heapam-and-nbtree.patch)
download | inline diff:
From a62617bf8ef00aeb70d193d820128a86f763f651 Mon Sep 17 00:00:00 2001
From: Peter Geoghegan <[email protected]>
Date: Tue, 9 Sep 2025 19:50:03 -0400
Subject: [PATCH v7 1/4] Add batching interfaces used by heapam and nbtree.
Add a new amgetbatch index AM interface that allows index access methods
to implement plain/ordered index scans that return index entries in
per-leaf-page batches, rather than one at a time. This enables a
variety of optimizations on the table AM side, most notably I/O
prefetching of heap tuples during ordered index scans. It will also
enable an optimization that has heapam avoid repeatedly locking and
unlocking the same heap page's buffer.
Index access methods that support plain index scans must now implement
either the amgetbatch interface OR the amgettuple interface. The
amgettuple interface will still be used by index AMs that require direct
control over the progress of index scans (e.g., GiST with KNN ordered
scans).
This commit also adds a new table AM interface callback, called by the
core executor through the new table_index_getnext_slot shim function.
This allows the table AM to directly manage the progress of index scans
rather than having individual TIDs passed in by the caller. The
amgetbatch interface is tightly coupled with the new approach to ordered
index scans added to the table AM. The table AM can apply knowledge of
which TIDs will be returned to scan in the near future to optimize and
batch table AM block accesses, and to perform I/O prefetching. These
optimizations are left as work for later commits.
Batches returned from amgetbatch are guaranteed to be associated with an
index page containing at least one matching tuple. The amgetbatch
interface may hold buffer pins as interlocks against concurrent TID
recycling by VACUUM. This extends/generalizes the mechanism added to
nbtree by commit 2ed5b87f to all index AMs that add support for the new
amgetbatch interface.
Author: Tomas Vondra <[email protected]>
Author: Peter Geoghegan <[email protected]>
Reviewed-By: Andres Freund <[email protected]>
Reviewed-By: Thomas Munro <[email protected]>
Discussion: https://postgr.es/m/[email protected]
Discussion: https://postgr.es/m/efac3238-6f34-41ea-a393-26cc0441b506%40vondra.me
Discussion: https://postgr.es/m/CAH2-Wzk9%3Dx%3Da2TbcqYcX%2BXXmDHQr5%3D1v9m4Z_v8a-KwF1Zoz0A%40mail.gmail.com
---
src/include/access/amapi.h | 22 +-
src/include/access/genam.h | 24 +-
src/include/access/heapam.h | 3 +
src/include/access/nbtree.h | 176 ++----
src/include/access/relscan.h | 252 +++++++-
src/include/access/tableam.h | 41 ++
src/include/executor/instrument_node.h | 6 +
src/include/nodes/execnodes.h | 2 -
src/include/nodes/pathnodes.h | 2 +-
src/backend/access/brin/brin.c | 5 +-
src/backend/access/gin/ginget.c | 6 +-
src/backend/access/gin/ginutil.c | 5 +-
src/backend/access/gist/gist.c | 5 +-
src/backend/access/hash/hash.c | 5 +-
src/backend/access/heap/heapam_handler.c | 563 ++++++++++++++++-
src/backend/access/index/Makefile | 3 +-
src/backend/access/index/genam.c | 15 +-
src/backend/access/index/indexam.c | 135 ++--
src/backend/access/index/indexbatch.c | 589 ++++++++++++++++++
src/backend/access/index/meson.build | 1 +
src/backend/access/nbtree/README | 6 +-
src/backend/access/nbtree/nbtpage.c | 3 +
src/backend/access/nbtree/nbtreadpage.c | 195 +++---
src/backend/access/nbtree/nbtree.c | 299 +++------
src/backend/access/nbtree/nbtsearch.c | 510 ++++++---------
src/backend/access/nbtree/nbtutils.c | 93 +--
src/backend/access/spgist/spgutils.c | 5 +-
src/backend/commands/explain.c | 23 +-
src/backend/commands/indexcmds.c | 2 +-
src/backend/executor/execAmi.c | 2 +-
src/backend/executor/execIndexing.c | 6 +-
src/backend/executor/execReplication.c | 8 +-
src/backend/executor/nodeBitmapIndexscan.c | 1 +
src/backend/executor/nodeIndexonlyscan.c | 108 +---
src/backend/executor/nodeIndexscan.c | 13 +-
src/backend/optimizer/path/indxpath.c | 2 +-
src/backend/optimizer/util/plancat.c | 6 +-
src/backend/replication/logical/relation.c | 3 +-
src/backend/utils/adt/amutils.c | 8 +-
src/backend/utils/adt/selfuncs.c | 68 +-
contrib/bloom/blutils.c | 5 +-
doc/src/sgml/indexam.sgml | 311 +++++++--
doc/src/sgml/ref/create_table.sgml | 13 +-
.../modules/dummy_index_am/dummy_index_am.c | 5 +-
src/tools/pgindent/typedefs.list | 4 -
45 files changed, 2319 insertions(+), 1240 deletions(-)
create mode 100644 src/backend/access/index/indexbatch.c
diff --git a/src/include/access/amapi.h b/src/include/access/amapi.h
index ecfbd017d..962c70f43 100644
--- a/src/include/access/amapi.h
+++ b/src/include/access/amapi.h
@@ -198,6 +198,15 @@ typedef void (*amrescan_function) (IndexScanDesc scan,
typedef bool (*amgettuple_function) (IndexScanDesc scan,
ScanDirection direction);
+/* next batch of valid tuples */
+typedef BatchIndexScan (*amgetbatch_function) (IndexScanDesc scan,
+ BatchIndexScan priorbatch,
+ ScanDirection direction);
+
+/* release batch of valid tuples */
+typedef void (*amfreebatch_function) (IndexScanDesc scan,
+ BatchIndexScan batch);
+
/* fetch all valid tuples */
typedef int64 (*amgetbitmap_function) (IndexScanDesc scan,
TIDBitmap *tbm);
@@ -205,11 +214,9 @@ typedef int64 (*amgetbitmap_function) (IndexScanDesc scan,
/* end index scan */
typedef void (*amendscan_function) (IndexScanDesc scan);
-/* mark current scan position */
-typedef void (*ammarkpos_function) (IndexScanDesc scan);
-
-/* restore marked scan position */
-typedef void (*amrestrpos_function) (IndexScanDesc scan);
+/* invalidate index AM state that independently tracks scan's position */
+typedef void (*amposreset_function) (IndexScanDesc scan,
+ BatchIndexScan batch);
/*
* Callback function signatures - for parallel index scans.
@@ -309,10 +316,11 @@ typedef struct IndexAmRoutine
ambeginscan_function ambeginscan;
amrescan_function amrescan;
amgettuple_function amgettuple; /* can be NULL */
+ amgetbatch_function amgetbatch; /* can be NULL */
+ amfreebatch_function amfreebatch; /* can be NULL */
amgetbitmap_function amgetbitmap; /* can be NULL */
amendscan_function amendscan;
- ammarkpos_function ammarkpos; /* can be NULL */
- amrestrpos_function amrestrpos; /* can be NULL */
+ amposreset_function amposreset; /* can be NULL */
/* interface functions to support parallel index scans */
amestimateparallelscan_function amestimateparallelscan; /* can be NULL */
diff --git a/src/include/access/genam.h b/src/include/access/genam.h
index 4c0429cc6..01f059fc4 100644
--- a/src/include/access/genam.h
+++ b/src/include/access/genam.h
@@ -94,6 +94,7 @@ typedef bool (*IndexBulkDeleteCallback) (ItemPointer itemptr, void *state);
/* struct definitions appear in relscan.h */
typedef struct IndexScanDescData *IndexScanDesc;
+typedef struct BatchIndexScanData *BatchIndexScan;
typedef struct SysScanDescData *SysScanDesc;
typedef struct ParallelIndexScanDescData *ParallelIndexScanDesc;
@@ -154,6 +155,7 @@ extern void index_insert_cleanup(Relation indexRelation,
extern IndexScanDesc index_beginscan(Relation heapRelation,
Relation indexRelation,
+ bool xs_want_itup,
Snapshot snapshot,
IndexScanInstrumentation *instrument,
int nkeys, int norderbys);
@@ -180,14 +182,12 @@ extern void index_parallelscan_initialize(Relation heapRelation,
extern void index_parallelrescan(IndexScanDesc scan);
extern IndexScanDesc index_beginscan_parallel(Relation heaprel,
Relation indexrel,
+ bool xs_want_itup,
IndexScanInstrumentation *instrument,
int nkeys, int norderbys,
ParallelIndexScanDesc pscan);
extern ItemPointer index_getnext_tid(IndexScanDesc scan,
ScanDirection direction);
-extern bool index_fetch_heap(IndexScanDesc scan, TupleTableSlot *slot);
-extern bool index_getnext_slot(IndexScanDesc scan, ScanDirection direction,
- TupleTableSlot *slot);
extern int64 index_getbitmap(IndexScanDesc scan, TIDBitmap *bitmap);
extern IndexBulkDeleteResult *index_bulk_delete(IndexVacuumInfo *info,
@@ -251,4 +251,22 @@ extern void systable_inplace_update_begin(Relation relation,
extern void systable_inplace_update_finish(void *state, HeapTuple tuple);
extern void systable_inplace_update_cancel(void *state);
+/*
+ * amgetbatch utilities called by indexam.c (in indexbatch.c)
+ */
+extern void index_batch_init(IndexScanDesc scan);
+extern void batch_free(IndexScanDesc scan, BatchIndexScan batch);
+extern void index_batch_reset(IndexScanDesc scan, bool complete);
+extern void index_batch_mark_pos(IndexScanDesc scan);
+extern void index_batch_restore_pos(IndexScanDesc scan);
+extern void index_batch_kill_item(IndexScanDesc scan);
+extern void index_batch_end(IndexScanDesc scan);
+
+/*
+ * amgetbatch utilities called by index AMs (in indexbatch.c)
+ */
+extern void indexam_util_batch_unlock(IndexScanDesc scan, BatchIndexScan batch);
+extern BatchIndexScan indexam_util_batch_alloc(IndexScanDesc scan);
+extern void indexam_util_batch_release(IndexScanDesc scan, BatchIndexScan batch);
+
#endif /* GENAM_H */
diff --git a/src/include/access/heapam.h b/src/include/access/heapam.h
index 3c0961ab3..494f36ecc 100644
--- a/src/include/access/heapam.h
+++ b/src/include/access/heapam.h
@@ -117,7 +117,10 @@ typedef struct IndexFetchHeapData
IndexFetchTableData xs_base; /* AM independent part of the descriptor */
Buffer xs_cbuf; /* current heap buffer in scan, if any */
+ BlockNumber xs_blk; /* xs_cbuf's block number, if any */
/* NB: if xs_cbuf is not InvalidBuffer, we hold a pin on that buffer */
+
+ Buffer vmbuf; /* visibility map buffer */
} IndexFetchHeapData;
/* Result codes for HeapTupleSatisfiesVacuum */
diff --git a/src/include/access/nbtree.h b/src/include/access/nbtree.h
index 772248596..e2f210f07 100644
--- a/src/include/access/nbtree.h
+++ b/src/include/access/nbtree.h
@@ -924,112 +924,6 @@ typedef struct BTVacuumPostingData
typedef BTVacuumPostingData *BTVacuumPosting;
-/*
- * BTScanOpaqueData is the btree-private state needed for an indexscan.
- * This consists of preprocessed scan keys (see _bt_preprocess_keys() for
- * details of the preprocessing), information about the current location
- * of the scan, and information about the marked location, if any. (We use
- * BTScanPosData to represent the data needed for each of current and marked
- * locations.) In addition we can remember some known-killed index entries
- * that must be marked before we can move off the current page.
- *
- * Index scans work a page at a time: we pin and read-lock the page, identify
- * all the matching items on the page and save them in BTScanPosData, then
- * release the read-lock while returning the items to the caller for
- * processing. This approach minimizes lock/unlock traffic. We must always
- * drop the lock to make it okay for caller to process the returned items.
- * Whether or not we can also release the pin during this window will vary.
- * We drop the pin (when so->dropPin) to avoid blocking progress by VACUUM
- * (see nbtree/README section about making concurrent TID recycling safe).
- * We'll always release both the lock and the pin on the current page before
- * moving on to its sibling page.
- *
- * If we are doing an index-only scan, we save the entire IndexTuple for each
- * matched item, otherwise only its heap TID and offset. The IndexTuples go
- * into a separate workspace array; each BTScanPosItem stores its tuple's
- * offset within that array. Posting list tuples store a "base" tuple once,
- * allowing the same key to be returned for each TID in the posting list
- * tuple.
- */
-
-typedef struct BTScanPosItem /* what we remember about each match */
-{
- ItemPointerData heapTid; /* TID of referenced heap item */
- OffsetNumber indexOffset; /* index item's location within page */
- LocationIndex tupleOffset; /* IndexTuple's offset in workspace, if any */
-} BTScanPosItem;
-
-typedef struct BTScanPosData
-{
- Buffer buf; /* currPage buf (invalid means unpinned) */
-
- /* page details as of the saved position's call to _bt_readpage */
- BlockNumber currPage; /* page referenced by items array */
- BlockNumber prevPage; /* currPage's left link */
- BlockNumber nextPage; /* currPage's right link */
- XLogRecPtr lsn; /* currPage's LSN (when so->dropPin) */
-
- /* scan direction for the saved position's call to _bt_readpage */
- ScanDirection dir;
-
- /*
- * If we are doing an index-only scan, nextTupleOffset is the first free
- * location in the associated tuple storage workspace.
- */
- int nextTupleOffset;
-
- /*
- * moreLeft and moreRight track whether we think there may be matching
- * index entries to the left and right of the current page, respectively.
- */
- bool moreLeft;
- bool moreRight;
-
- /*
- * The items array is always ordered in index order (ie, increasing
- * indexoffset). When scanning backwards it is convenient to fill the
- * array back-to-front, so we start at the last slot and fill downwards.
- * Hence we need both a first-valid-entry and a last-valid-entry counter.
- * itemIndex is a cursor showing which entry was last returned to caller.
- */
- int firstItem; /* first valid index in items[] */
- int lastItem; /* last valid index in items[] */
- int itemIndex; /* current index in items[] */
-
- BTScanPosItem items[MaxTIDsPerBTreePage]; /* MUST BE LAST */
-} BTScanPosData;
-
-typedef BTScanPosData *BTScanPos;
-
-#define BTScanPosIsPinned(scanpos) \
-( \
- AssertMacro(BlockNumberIsValid((scanpos).currPage) || \
- !BufferIsValid((scanpos).buf)), \
- BufferIsValid((scanpos).buf) \
-)
-#define BTScanPosUnpin(scanpos) \
- do { \
- ReleaseBuffer((scanpos).buf); \
- (scanpos).buf = InvalidBuffer; \
- } while (0)
-#define BTScanPosUnpinIfPinned(scanpos) \
- do { \
- if (BTScanPosIsPinned(scanpos)) \
- BTScanPosUnpin(scanpos); \
- } while (0)
-
-#define BTScanPosIsValid(scanpos) \
-( \
- AssertMacro(BlockNumberIsValid((scanpos).currPage) || \
- !BufferIsValid((scanpos).buf)), \
- BlockNumberIsValid((scanpos).currPage) \
-)
-#define BTScanPosInvalidate(scanpos) \
- do { \
- (scanpos).buf = InvalidBuffer; \
- (scanpos).currPage = InvalidBlockNumber; \
- } while (0)
-
/* We need one of these for each equality-type SK_SEARCHARRAY scan key */
typedef struct BTArrayKeyInfo
{
@@ -1050,6 +944,30 @@ typedef struct BTArrayKeyInfo
ScanKey high_compare; /* array's < or <= upper bound */
} BTArrayKeyInfo;
+/*
+ * BTScanOpaqueData is the btree-private state needed for an indexscan.
+ * This consists of preprocessed scan keys (see _bt_preprocess_keys() for
+ * details of the preprocessing), and information about the current array
+ * keys. There are assumptions about how the current array keys track the
+ * progress of the index scan through the index's key space (see _bt_readpage
+ * and _bt_advance_array_keys), but we don't actually track anything about the
+ * current scan position in this opaque struct. That is tracked externally,
+ * by implementing a ring buffer of "batches", where each batch represents the
+ * items returned by btgetbatch within a single leaf page.
+ *
+ * Index scans work a page at a time, as required by the amgetbatch contract:
+ * we pin and read-lock the page, identify all the matching items on the page
+ * and return them in a newly allocated batch. We then release the read-lock
+ * using amgetbatch utility routines. This approach minimizes lock/unlock
+ * traffic. _bt_next is passed priorbatch, which contains details of which
+ * page is next in line to be read (priorbatch is provided as an argument to
+ * btgetbatch by core code).
+ *
+ * If we are doing an index-only scan, we save the entire IndexTuple for each
+ * matched item, otherwise only its heap TID and offset. This is also per the
+ * amgetbatch contract. Posting list tuples store a "base" tuple once,
+ * allowing the same key to be returned for each TID in the posting list.
+ */
typedef struct BTScanOpaqueData
{
/* these fields are set by _bt_preprocess_keys(): */
@@ -1066,32 +984,6 @@ typedef struct BTScanOpaqueData
BTArrayKeyInfo *arrayKeys; /* info about each equality-type array key */
FmgrInfo *orderProcs; /* ORDER procs for required equality keys */
MemoryContext arrayContext; /* scan-lifespan context for array data */
-
- /* info about killed items if any (killedItems is NULL if never used) */
- int *killedItems; /* currPos.items indexes of killed items */
- int numKilled; /* number of currently stored items */
- bool dropPin; /* drop leaf pin before btgettuple returns? */
-
- /*
- * If we are doing an index-only scan, these are the tuple storage
- * workspaces for the currPos and markPos respectively. Each is of size
- * BLCKSZ, so it can hold as much as a full page's worth of tuples.
- */
- char *currTuples; /* tuple storage for currPos */
- char *markTuples; /* tuple storage for markPos */
-
- /*
- * If the marked position is on the same page as current position, we
- * don't use markPos, but just keep the marked itemIndex in markItemIndex
- * (all the rest of currPos is valid for the mark position). Hence, to
- * determine if there is a mark, first look at markItemIndex, then at
- * markPos.
- */
- int markItemIndex; /* itemIndex, or -1 if not valid */
-
- /* keep these last in struct for efficiency */
- BTScanPosData currPos; /* current position data */
- BTScanPosData markPos; /* marked position, if any */
} BTScanOpaqueData;
typedef BTScanOpaqueData *BTScanOpaque;
@@ -1160,14 +1052,16 @@ extern bool btinsert(Relation rel, Datum *values, bool *isnull,
extern IndexScanDesc btbeginscan(Relation rel, int nkeys, int norderbys);
extern Size btestimateparallelscan(Relation rel, int nkeys, int norderbys);
extern void btinitparallelscan(void *target);
-extern bool btgettuple(IndexScanDesc scan, ScanDirection dir);
+extern BatchIndexScan btgetbatch(IndexScanDesc scan,
+ BatchIndexScan priorbatch,
+ ScanDirection dir);
extern int64 btgetbitmap(IndexScanDesc scan, TIDBitmap *tbm);
extern void btrescan(IndexScanDesc scan, ScanKey scankey, int nscankeys,
ScanKey orderbys, int norderbys);
+extern void btfreebatch(IndexScanDesc scan, BatchIndexScan batch);
extern void btparallelrescan(IndexScanDesc scan);
extern void btendscan(IndexScanDesc scan);
-extern void btmarkpos(IndexScanDesc scan);
-extern void btrestrpos(IndexScanDesc scan);
+extern void btposreset(IndexScanDesc scan, BatchIndexScan markbatch);
extern IndexBulkDeleteResult *btbulkdelete(IndexVacuumInfo *info,
IndexBulkDeleteResult *stats,
IndexBulkDeleteCallback callback,
@@ -1271,8 +1165,9 @@ extern void _bt_preprocess_keys(IndexScanDesc scan);
/*
* prototypes for functions in nbtreadpage.c
*/
-extern bool _bt_readpage(IndexScanDesc scan, ScanDirection dir,
- OffsetNumber offnum, bool firstpage);
+extern bool _bt_readpage(IndexScanDesc scan, BatchIndexScan newbatch,
+ ScanDirection dir, OffsetNumber offnum,
+ bool firstpage);
extern void _bt_start_array_keys(IndexScanDesc scan, ScanDirection dir);
extern int _bt_binsrch_array_skey(FmgrInfo *orderproc,
bool cur_elem_trig, ScanDirection dir,
@@ -1287,8 +1182,9 @@ extern BTStack _bt_search(Relation rel, Relation heaprel, BTScanInsert key,
Buffer *bufP, int access);
extern OffsetNumber _bt_binsrch_insert(Relation rel, BTInsertState insertstate);
extern int32 _bt_compare(Relation rel, BTScanInsert key, Page page, OffsetNumber offnum);
-extern bool _bt_first(IndexScanDesc scan, ScanDirection dir);
-extern bool _bt_next(IndexScanDesc scan, ScanDirection dir);
+extern BatchIndexScan _bt_first(IndexScanDesc scan, ScanDirection dir);
+extern BatchIndexScan _bt_next(IndexScanDesc scan, ScanDirection dir,
+ BatchIndexScan priorbatch);
extern Buffer _bt_get_endpoint(Relation rel, uint32 level, bool rightmost);
/*
@@ -1296,7 +1192,7 @@ extern Buffer _bt_get_endpoint(Relation rel, uint32 level, bool rightmost);
*/
extern BTScanInsert _bt_mkscankey(Relation rel, IndexTuple itup);
extern void _bt_freestack(BTStack stack);
-extern void _bt_killitems(IndexScanDesc scan);
+extern void _bt_killitems(IndexScanDesc scan, BatchIndexScan batch);
extern BTCycleId _bt_vacuum_cycleid(Relation rel);
extern BTCycleId _bt_start_vacuum(Relation rel);
extern void _bt_end_vacuum(Relation rel);
diff --git a/src/include/access/relscan.h b/src/include/access/relscan.h
index ce340c076..ca1207be6 100644
--- a/src/include/access/relscan.h
+++ b/src/include/access/relscan.h
@@ -16,8 +16,10 @@
#include "access/htup_details.h"
#include "access/itup.h"
+#include "access/sdir.h"
#include "nodes/tidbitmap.h"
#include "port/atomics.h"
+#include "storage/buf.h"
#include "storage/relfilelocator.h"
#include "storage/spin.h"
#include "utils/relcache.h"
@@ -124,6 +126,179 @@ typedef struct IndexFetchTableData
Relation rel;
} IndexFetchTableData;
+/*
+ * Location of a BatchMatchingItem that appears in a BatchIndexScan returned
+ * by (and subsequently passed to) an amgetbatch routine
+ */
+typedef struct BatchRingItemPos
+{
+ /* BatchRingBuffer.batches[]-wise index to relevant BatchIndexScan */
+ int batch;
+
+ /* BatchIndexScan.items[]-wise index to relevant BatchMatchingItem */
+ int item;
+} BatchRingItemPos;
+
+static inline void
+batch_reset_pos(BatchRingItemPos *pos)
+{
+ pos->batch = -1;
+ pos->item = -1;
+}
+
+/*
+ * Matching item returned by amgetbatch (in returned BatchIndexScan) during an
+ * index scan. Used by table AM to locate relevant matching table tuple.
+ */
+typedef struct BatchMatchingItem
+{
+ ItemPointerData heapTid; /* TID of referenced heap item */
+ OffsetNumber indexOffset; /* index item's location within page */
+ LocationIndex tupleOffset; /* IndexTuple's offset in workspace, if any */
+ bool allVisible; /* TID points to all-visible page */
+} BatchMatchingItem;
+
+/*
+ * Data about one batch of items returned by (and passed to) amgetbatch during
+ * index scans
+ */
+typedef struct BatchIndexScanData
+{
+ /*
+ * Information output by amgetbatch index AMs upon returning a batch with
+ * one or more matching items, describing details of the index page where
+ * matches were located.
+ *
+ * Used in the next amgetbatch call to determine which index page to read
+ * next (or to determine if there's no further matches in current scan
+ * direction).
+ */
+ BlockNumber currPage; /* Index page with matching items */
+ BlockNumber prevPage; /* currPage's left link */
+ BlockNumber nextPage; /* currPage's right link */
+
+ Buffer buf; /* currPage buf (invalid means unpinned) */
+ XLogRecPtr lsn; /* currPage's LSN (when dropPin) */
+
+ /* scan direction when the index page was read */
+ ScanDirection dir;
+
+ /*
+ * moreLeft and moreRight track whether we think there may be matching
+ * index entries to the left and right of the current page, respectively
+ */
+ bool moreLeft;
+ bool moreRight;
+
+ /*
+ * The items array is always ordered in index order (ie, increasing
+ * indexoffset). When scanning backwards it is convenient to fill the
+ * array back-to-front, so we start at the last slot and fill downwards.
+ * Hence we need both a first-valid-entry and a last-valid-entry counter.
+ */
+ int firstItem; /* first valid index in items[] */
+ int lastItem; /* last valid index in items[] */
+
+ /* info about killed items if any (killedItems is NULL if never used) */
+ int *killedItems; /* indexes of killed items */
+ int numKilled; /* number of currently stored items */
+
+ /*
+ * Matching items state for this batch.
+ *
+ * If we are doing an index-only scan, these are the tuple storage
+ * workspaces for the matching tuples (tuples referenced by items[]). Each
+ * is of size BLCKSZ, so it can hold as much as a full page's worth of
+ * tuples.
+ */
+ char *currTuples; /* tuple storage for items[] */
+ BatchMatchingItem items[FLEXIBLE_ARRAY_MEMBER];
+} BatchIndexScanData;
+
+typedef struct BatchIndexScanData *BatchIndexScan;
+
+/*
+ * Maximum number of batches (leaf pages) we can keep in memory. We need a
+ * minimum of two, since we'll only consider releasing one batch when another
+ * is read.
+ */
+#define INDEX_SCAN_MAX_BATCHES 2
+#define INDEX_SCAN_CACHE_BATCHES 2
+#define INDEX_SCAN_BATCH_COUNT(scan) \
+ ((scan)->ringbuf->nextBatch - (scan)->ringbuf->headBatch)
+
+/* Did we already load batch with the requested index? */
+#define INDEX_SCAN_BATCH_LOADED(scan, idx) \
+ ((idx) >= (scan)->ringbuf->headBatch && (idx) < (scan)->ringbuf->nextBatch)
+
+/* Have we loaded the maximum number of batches? */
+#define INDEX_SCAN_BATCH_FULL(scan) \
+ (INDEX_SCAN_BATCH_COUNT(scan) == INDEX_SCAN_MAX_BATCHES)
+
+/* Return batch for the provided index */
+#define INDEX_SCAN_BATCH(scan, idx) \
+( \
+ AssertMacro(INDEX_SCAN_BATCH_LOADED(scan, idx)), \
+ ((scan)->ringbuf->batches[(idx) % INDEX_SCAN_MAX_BATCHES]) \
+)
+
+/* Append given batch to scan's batch ring buffer */
+#define INDEX_SCAN_BATCH_APPEND(scan, batch) \
+ do { \
+ BatchRingBuffer *mringbuf = (scan)->ringbuf; \
+ int nextBatch = mringbuf->nextBatch; \
+ mringbuf->batches[nextBatch % INDEX_SCAN_MAX_BATCHES] = (batch); \
+ mringbuf->nextBatch++; \
+ } while(0)
+
+/* Is the position invalid/undefined? */
+#define INDEX_SCAN_POS_INVALID(pos) \
+ (((pos)->batch == -1) && ((pos)->item == -1))
+
+#ifdef INDEXAM_DEBUG
+#define DEBUG_LOG(...) elog(AmRegularBackendProcess() ? NOTICE : DEBUG2, __VA_ARGS__)
+#else
+#define DEBUG_LOG(...)
+#endif
+
+/*
+ * State used by table AMs to manage an index scan that uses the amgetbatch
+ * interface. Scans use a ring buffer of batches returned by amgetbatch.
+ *
+ * Batches are kept in the order that they were returned in by amgetbatch,
+ * since that is the same order that table_index_getnext_slot will return
+ * matches in. However, table AMs are free to fetch table tuples in whatever
+ * order is most convenient/efficient -- provided that such reordering cannot
+ * affect the order that table_index_getnext_slot later returns tuples in.
+ */
+typedef struct BatchRingBuffer
+{
+ /* Current scan direction, for the currently loaded batches */
+ ScanDirection direction;
+
+ /* current positions in batches[] for scan */
+ BatchRingItemPos scanPos; /* scan's read position */
+ BatchRingItemPos markPos; /* mark/restore position */
+
+ BatchIndexScan markBatch;
+
+ /*
+ * Array of batches returned by the AM. The array has a capacity (but can
+ * be resized if needed). The headBatch is an index of the batch we're
+ * currently reading from (this needs to be translated by modulo
+ * INDEX_SCAN_MAX_BATCHES into index in the batches array).
+ */
+ int headBatch; /* head batch slot */
+ int nextBatch; /* next empty batch slot */
+
+ /* Array of pointers to cached recyclable batches */
+ BatchIndexScan cache[INDEX_SCAN_CACHE_BATCHES];
+
+ /* Array of pointers to ring buffer batches */
+ BatchIndexScan batches[INDEX_SCAN_MAX_BATCHES];
+
+} BatchRingBuffer;
+
struct IndexScanInstrumentation;
/*
@@ -141,6 +316,13 @@ typedef struct IndexScanDescData
int numberOfOrderBys; /* number of ordering operators */
struct ScanKeyData *keyData; /* array of index qualifier descriptors */
struct ScanKeyData *orderByData; /* array of ordering op descriptors */
+
+ /* index access method's private state */
+ void *opaque; /* access-method-specific info */
+
+ /* table access method's private amgetbatch state */
+ BatchRingBuffer *ringbuf; /* amgetbatch related state */
+
bool xs_want_itup; /* caller requests index tuples */
bool xs_temp_snap; /* unregister snapshot at scan end? */
@@ -149,9 +331,13 @@ typedef struct IndexScanDescData
bool ignore_killed_tuples; /* do not return killed entries */
bool xactStartedInRecovery; /* prevents killing/seeing killed
* tuples */
+ /* amgetbatch can safely drop pins on returned batch's index page? */
+ bool dropPin;
- /* index access method's private state */
- void *opaque; /* access-method-specific info */
+ /*
+ * Did we read the final batch in this scan direction?
+ */
+ bool finished;
/*
* Instrumentation counters maintained by all index AMs during both
@@ -176,6 +362,8 @@ typedef struct IndexScanDescData
IndexFetchTableData *xs_heapfetch;
bool xs_recheck; /* T means scan keys must be rechecked */
+ bool xs_visible; /* T means the heap page is all-visible */
+ uint16 maxitemsbatch; /* set by ambeginscan when amgetbatch used */
/*
* When fetching with an ordering operator, the values of the ORDER BY
@@ -215,4 +403,64 @@ typedef struct SysScanDescData
struct TupleTableSlot *slot;
} SysScanDescData;
+/*
+ * Check that a position (batch,item) is valid with respect to the batches we
+ * have currently loaded.
+ */
+static inline void
+batch_assert_pos_valid(IndexScanDescData *scan, BatchRingItemPos *pos)
+{
+#ifdef USE_ASSERT_CHECKING
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+ BatchIndexScan batch = INDEX_SCAN_BATCH(scan, pos->batch);
+
+ /* make sure the position is valid for currently loaded batches */
+ Assert(pos->batch >= ringbuf->headBatch);
+ Assert(pos->batch < ringbuf->nextBatch);
+ Assert(pos->item >= batch->firstItem);
+ Assert(pos->item <= batch->lastItem);
+#endif
+}
+
+/*
+ * Check a single batch is valid.
+ */
+static inline void
+batch_assert_batch_valid(IndexScanDescData *scan, BatchIndexScan batch)
+{
+ /* batch must have one or more matching items returned by index AM */
+ Assert(batch->firstItem >= 0 && batch->firstItem <= batch->lastItem);
+ Assert(batch->items != NULL);
+
+ /*
+ * The number of killed items must be valid, and there must be an array of
+ * indexes if there are items.
+ */
+ Assert(batch->numKilled >= 0);
+ Assert(!(batch->numKilled > 0 && batch->killedItems == NULL));
+}
+
+static inline void
+batch_assert_batches_valid(IndexScanDescData *scan)
+{
+#ifdef USE_ASSERT_CHECKING
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+
+ /* we should have batches initialized */
+ Assert(ringbuf != NULL);
+
+ /* The head/next indexes should define a valid range */
+ Assert(ringbuf->headBatch >= 0 &&
+ ringbuf->headBatch <= ringbuf->nextBatch);
+
+ /* Check all current batches */
+ for (int i = ringbuf->headBatch; i < ringbuf->nextBatch; i++)
+ {
+ BatchIndexScan batch = INDEX_SCAN_BATCH(scan, i);
+
+ batch_assert_batch_valid(scan, batch);
+ }
+#endif
+}
+
#endif /* RELSCAN_H */
diff --git a/src/include/access/tableam.h b/src/include/access/tableam.h
index e2ec5289d..98e337972 100644
--- a/src/include/access/tableam.h
+++ b/src/include/access/tableam.h
@@ -433,11 +433,29 @@ typedef struct TableAmRoutine
*/
void (*index_fetch_end) (struct IndexFetchTableData *data);
+ /*
+ * Fetch the next tuple from an index scan into slot, scanning in the
+ * specified direction, and return true if a tuple was found, false
+ * otherwise.
+ *
+ * This callback allows the table AM to directly manage the scan process,
+ * including interfacing with the index AM. The caller simply specifies
+ * the direction of the scan; the table AM takes care of retrieving TIDs
+ * from the index, performing visibility checks, and returning tuples in
+ * the slot.
+ */
+ bool (*index_getnext_slot) (IndexScanDesc scan,
+ ScanDirection direction,
+ TupleTableSlot *slot);
+
/*
* Fetch tuple at `tid` into `slot`, after doing a visibility test
* according to `snapshot`. If a tuple was found and passed the visibility
* test, return true, false otherwise.
*
+ * This is a lower-level callback that takes a TID from the caller.
+ * Callers should favor the index_getnext_slot callback whenever possible.
+ *
* Note that AMs that do not necessarily update indexes when indexed
* columns do not change, need to return the current/correct version of
* the tuple that is visible to the snapshot, even if the tid points to an
@@ -1188,6 +1206,26 @@ table_index_fetch_end(struct IndexFetchTableData *scan)
scan->rel->rd_tableam->index_fetch_end(scan);
}
+/*
+ * Fetch the next tuple from an index scan into `slot`, scanning in the
+ * specified direction. Returns true if a tuple was found, false otherwise.
+ *
+ * The index scan should have been started via table_index_fetch_begin().
+ * Callers must check scan->xs_recheck and recheck scan keys if required.
+ *
+ * Index-only scan callers (that pass xs_want_itup=true to index_beginscan)
+ * can consume index tuple results by examining IndexScanDescData fields such
+ * as xs_itup and xs_hitup. The table AM won't usually fetch a heap tuple
+ * into the provided slot in the case of xs_want_itup=true callers.
+ */
+static inline bool
+table_index_getnext_slot(IndexScanDesc iscan, ScanDirection direction,
+ TupleTableSlot *slot)
+{
+ return iscan->heapRelation->rd_tableam->index_getnext_slot(iscan,
+ direction, slot);
+}
+
/*
* Fetches, as part of an index scan, tuple at `tid` into `slot`, after doing
* a visibility test according to `snapshot`. If a tuple was found and passed
@@ -1211,6 +1249,9 @@ table_index_fetch_end(struct IndexFetchTableData *scan)
* entry (like heap's HOT). Whereas table_tuple_fetch_row_version() only
* evaluates the tuple exactly at `tid`. Outside of index entry ->table tuple
* lookups, table_tuple_fetch_row_version() is what's usually needed.
+ *
+ * This is a lower-level interface that takes a TID from the caller. Callers
+ * should favor the table_index_getnext_slot interface whenever possible.
*/
static inline bool
table_index_fetch_tuple(struct IndexFetchTableData *scan,
diff --git a/src/include/executor/instrument_node.h b/src/include/executor/instrument_node.h
index 8847d7f94..b5b8f509a 100644
--- a/src/include/executor/instrument_node.h
+++ b/src/include/executor/instrument_node.h
@@ -48,6 +48,12 @@ typedef struct IndexScanInstrumentation
{
/* Index search count (incremented with pgstat_count_index_scan call) */
uint64 nsearches;
+
+ /*
+ * heap blocks fetched counts (incremented by index_getnext_slot calls
+ * within table AMs, though only during index-only scans)
+ */
+ uint64 nheapfetches;
} IndexScanInstrumentation;
/*
diff --git a/src/include/nodes/execnodes.h b/src/include/nodes/execnodes.h
index f8053d9e5..793b1a3c6 100644
--- a/src/include/nodes/execnodes.h
+++ b/src/include/nodes/execnodes.h
@@ -1753,7 +1753,6 @@ typedef struct IndexScanState
* Instrument local index scan instrumentation
* SharedInfo parallel worker instrumentation (no leader entry)
* TableSlot slot for holding tuples fetched from the table
- * VMBuffer buffer in use for visibility map testing, if any
* PscanLen size of parallel index-only scan descriptor
* NameCStringAttNums attnums of name typed columns to pad to NAMEDATALEN
* NameCStringCount number of elements in the NameCStringAttNums array
@@ -1776,7 +1775,6 @@ typedef struct IndexOnlyScanState
IndexScanInstrumentation ioss_Instrument;
SharedIndexScanInstrumentation *ioss_SharedInfo;
TupleTableSlot *ioss_TableSlot;
- Buffer ioss_VMBuffer;
Size ioss_PscanLen;
AttrNumber *ioss_NameCStringAttNums;
int ioss_NameCStringCount;
diff --git a/src/include/nodes/pathnodes.h b/src/include/nodes/pathnodes.h
index 449885b93..3c81602b4 100644
--- a/src/include/nodes/pathnodes.h
+++ b/src/include/nodes/pathnodes.h
@@ -1344,7 +1344,7 @@ typedef struct IndexOptInfo
/* does AM have amgetbitmap interface? */
bool amhasgetbitmap;
bool amcanparallel;
- /* does AM have ammarkpos interface? */
+ /* is AM prepared for us to restore a mark? */
bool amcanmarkpos;
/* AM's cost estimator */
/* Rather than include amapi.h here, we declare amcostestimate like this */
diff --git a/src/backend/access/brin/brin.c b/src/backend/access/brin/brin.c
index 6887e4214..d5d01b877 100644
--- a/src/backend/access/brin/brin.c
+++ b/src/backend/access/brin/brin.c
@@ -294,10 +294,11 @@ brinhandler(PG_FUNCTION_ARGS)
.ambeginscan = brinbeginscan,
.amrescan = brinrescan,
.amgettuple = NULL,
+ .amgetbatch = NULL,
+ .amfreebatch = NULL,
.amgetbitmap = bringetbitmap,
.amendscan = brinendscan,
- .ammarkpos = NULL,
- .amrestrpos = NULL,
+ .amposreset = NULL,
.amestimateparallelscan = NULL,
.aminitparallelscan = NULL,
.amparallelrescan = NULL,
diff --git a/src/backend/access/gin/ginget.c b/src/backend/access/gin/ginget.c
index 6b148e69a..8f7033d62 100644
--- a/src/backend/access/gin/ginget.c
+++ b/src/backend/access/gin/ginget.c
@@ -1953,9 +1953,9 @@ gingetbitmap(IndexScanDesc scan, TIDBitmap *tbm)
* into the main index, and so we might visit it a second time during the
* main scan. This is okay because we'll just re-set the same bit in the
* bitmap. (The possibility of duplicate visits is a major reason why GIN
- * can't support the amgettuple API, however.) Note that it would not do
- * to scan the main index before the pending list, since concurrent
- * cleanup could then make us miss entries entirely.
+ * can't support either the amgettuple or amgetbatch API.) Note that it
+ * would not do to scan the main index before the pending list, since
+ * concurrent cleanup could then make us miss entries entirely.
*/
scanPendingInsert(scan, tbm, &ntids);
diff --git a/src/backend/access/gin/ginutil.c b/src/backend/access/gin/ginutil.c
index d205093e2..1263dc180 100644
--- a/src/backend/access/gin/ginutil.c
+++ b/src/backend/access/gin/ginutil.c
@@ -82,10 +82,11 @@ ginhandler(PG_FUNCTION_ARGS)
.ambeginscan = ginbeginscan,
.amrescan = ginrescan,
.amgettuple = NULL,
+ .amgetbatch = NULL,
+ .amfreebatch = NULL,
.amgetbitmap = gingetbitmap,
.amendscan = ginendscan,
- .ammarkpos = NULL,
- .amrestrpos = NULL,
+ .amposreset = NULL,
.amestimateparallelscan = NULL,
.aminitparallelscan = NULL,
.amparallelrescan = NULL,
diff --git a/src/backend/access/gist/gist.c b/src/backend/access/gist/gist.c
index d5944205d..9c219cadf 100644
--- a/src/backend/access/gist/gist.c
+++ b/src/backend/access/gist/gist.c
@@ -103,10 +103,11 @@ gisthandler(PG_FUNCTION_ARGS)
.ambeginscan = gistbeginscan,
.amrescan = gistrescan,
.amgettuple = gistgettuple,
+ .amgetbatch = NULL,
+ .amfreebatch = NULL,
.amgetbitmap = gistgetbitmap,
.amendscan = gistendscan,
- .ammarkpos = NULL,
- .amrestrpos = NULL,
+ .amposreset = NULL,
.amestimateparallelscan = NULL,
.aminitparallelscan = NULL,
.amparallelrescan = NULL,
diff --git a/src/backend/access/hash/hash.c b/src/backend/access/hash/hash.c
index e88ddb32a..6a20b67f6 100644
--- a/src/backend/access/hash/hash.c
+++ b/src/backend/access/hash/hash.c
@@ -102,10 +102,11 @@ hashhandler(PG_FUNCTION_ARGS)
.ambeginscan = hashbeginscan,
.amrescan = hashrescan,
.amgettuple = hashgettuple,
+ .amgetbatch = NULL,
+ .amfreebatch = NULL,
.amgetbitmap = hashgetbitmap,
.amendscan = hashendscan,
- .ammarkpos = NULL,
- .amrestrpos = NULL,
+ .amposreset = NULL,
.amestimateparallelscan = NULL,
.aminitparallelscan = NULL,
.amparallelrescan = NULL,
diff --git a/src/backend/access/heap/heapam_handler.c b/src/backend/access/heap/heapam_handler.c
index cbef73e5d..03b4fada9 100644
--- a/src/backend/access/heap/heapam_handler.c
+++ b/src/backend/access/heap/heapam_handler.c
@@ -19,6 +19,7 @@
*/
#include "postgres.h"
+#include "access/amapi.h"
#include "access/genam.h"
#include "access/heapam.h"
#include "access/heaptoast.h"
@@ -81,10 +82,12 @@ heapam_slot_callbacks(Relation relation)
static IndexFetchTableData *
heapam_index_fetch_begin(Relation rel)
{
- IndexFetchHeapData *hscan = palloc0_object(IndexFetchHeapData);
+ IndexFetchHeapData *hscan = palloc_object(IndexFetchHeapData);
hscan->xs_base.rel = rel;
hscan->xs_cbuf = InvalidBuffer;
+ hscan->xs_blk = InvalidBlockNumber;
+ hscan->vmbuf = InvalidBuffer;
return &hscan->xs_base;
}
@@ -94,10 +97,12 @@ heapam_index_fetch_reset(IndexFetchTableData *scan)
{
IndexFetchHeapData *hscan = (IndexFetchHeapData *) scan;
+ /* deliberately don't drop VM buffer pin here */
if (BufferIsValid(hscan->xs_cbuf))
{
ReleaseBuffer(hscan->xs_cbuf);
hscan->xs_cbuf = InvalidBuffer;
+ hscan->xs_blk = InvalidBlockNumber;
}
}
@@ -108,6 +113,12 @@ heapam_index_fetch_end(IndexFetchTableData *scan)
heapam_index_fetch_reset(scan);
+ if (hscan->vmbuf != InvalidBuffer)
+ {
+ ReleaseBuffer(hscan->vmbuf);
+ hscan->vmbuf = InvalidBuffer;
+ }
+
pfree(hscan);
}
@@ -125,22 +136,32 @@ heapam_index_fetch_tuple(struct IndexFetchTableData *scan,
Assert(TTS_IS_BUFFERTUPLE(slot));
/* We can skip the buffer-switching logic if we're in mid-HOT chain. */
- if (!*call_again)
+ if (hscan->xs_blk != ItemPointerGetBlockNumber(tid))
{
- /* Switch to correct buffer if we don't have it already */
- Buffer prev_buf = hscan->xs_cbuf;
+ Assert(!*call_again);
- hscan->xs_cbuf = ReleaseAndReadBuffer(hscan->xs_cbuf,
- hscan->xs_base.rel,
- ItemPointerGetBlockNumber(tid));
+ /* Remember this buffer's block number for next time */
+ hscan->xs_blk = ItemPointerGetBlockNumber(tid);
+
+ if (BufferIsValid(hscan->xs_cbuf))
+ ReleaseBuffer(hscan->xs_cbuf);
/*
- * Prune page, but only if we weren't already on this page
+ * When using a read stream, the stream will already know which block
+ * number comes next (though an assertion will verify a match below)
*/
- if (prev_buf != hscan->xs_cbuf)
- heap_page_prune_opt(hscan->xs_base.rel, hscan->xs_cbuf);
+ hscan->xs_cbuf = ReadBuffer(hscan->xs_base.rel, hscan->xs_blk);
+
+ /*
+ * Prune page when it is pinned for the first time
+ */
+ heap_page_prune_opt(hscan->xs_base.rel, hscan->xs_cbuf);
}
+ /* Assert that the TID's block number's buffer is now pinned */
+ Assert(BufferIsValid(hscan->xs_cbuf));
+ Assert(BufferGetBlockNumber(hscan->xs_cbuf) == hscan->xs_blk);
+
/* Obtain share-lock on the buffer so we can examine visibility */
LockBuffer(hscan->xs_cbuf, BUFFER_LOCK_SHARE);
got_heap_tuple = heap_hot_search_buffer(tid,
@@ -173,6 +194,521 @@ heapam_index_fetch_tuple(struct IndexFetchTableData *scan,
return got_heap_tuple;
}
+static pg_noinline void
+heapam_batch_rewind(IndexScanDesc scan, BatchRingBuffer *ringbuf,
+ ScanDirection direction)
+{
+ /*
+ * Handle a change in the scan's direction.
+ *
+ * Release future batches properly, to make it look like the current batch
+ * is the only one we loaded.
+ */
+ while (ringbuf->nextBatch > ringbuf->headBatch + 1)
+ {
+ /* release "later" batches in reverse order */
+ BatchIndexScan fbatch;
+
+ fbatch = INDEX_SCAN_BATCH(scan, ringbuf->nextBatch - 1);
+ batch_free(scan, fbatch);
+ ringbuf->nextBatch--;
+ }
+
+ /*
+ * Remember the new direction, and make sure the scan is not marked as
+ * "finished" (we might have already read the last batch, but now we need
+ * to start over).
+ */
+ ringbuf->direction = direction;
+ scan->finished = false;
+}
+
+static inline ItemPointer
+heapam_batch_return_tid(IndexScanDesc scan, BatchIndexScan scanBatch,
+ BatchRingItemPos *scanPos)
+{
+ batch_assert_pos_valid(scan, scanPos);
+
+ pgstat_count_index_tuples(scan->indexRelation, 1);
+
+ /* set the TID / itup for the scan */
+ scan->xs_heaptid = scanBatch->items[scanPos->item].heapTid;
+
+ /* plain index scans will have flags left set to 0 */
+ scan->xs_visible = scanBatch->items[scanPos->item].allVisible;
+
+ if (scan->xs_want_itup)
+ scan->xs_itup =
+ (IndexTuple) (scanBatch->currTuples +
+ scanBatch->items[scanPos->item].tupleOffset);
+
+ return &scan->xs_heaptid;
+}
+
+/*
+ * heap_batch_resolve_visibility
+ * Obtain visibility information for every TID from caller's batch.
+ */
+static void
+heap_batch_resolve_visibility(IndexScanDesc scan, BatchIndexScan batch)
+{
+ IndexFetchHeapData *hscan = (IndexFetchHeapData *) scan->xs_heapfetch;
+
+ for (int i = batch->firstItem; i <= batch->lastItem; i++)
+ {
+ BatchMatchingItem *item = &batch->items[i];
+ ItemPointer tid = &item->heapTid;
+
+ if (VM_ALL_VISIBLE(scan->heapRelation,
+ ItemPointerGetBlockNumber(tid),
+ &hscan->vmbuf))
+ {
+ item->allVisible = true;
+ }
+ }
+}
+
+/*
+ * heap_batchpos_advance
+ * Advance position to its next item in the batch.
+ *
+ * Advance to the next item within the provided batch (or to the previous item,
+ * when scanning backwards).
+ *
+ * Returns true if the position could be advanced. Returns false when there
+ * are no more items in the batch in the given direction.
+ */
+static inline bool
+heap_batchpos_advance(BatchIndexScan batch, BatchRingItemPos *pos,
+ ScanDirection direction)
+{
+ Assert(!INDEX_SCAN_POS_INVALID(pos));
+
+ if (ScanDirectionIsForward(direction))
+ {
+ if (++pos->item > batch->lastItem)
+ return false;
+ }
+ else /* ScanDirectionIsBackward */
+ {
+ if (--pos->item < batch->firstItem)
+ return false;
+ }
+
+ /* Advanced within batch */
+ return true;
+}
+
+/*
+ * heap_batchpos_newbatch
+ * Advance batch position start of its new batch.
+ *
+ * Sets the given position to the fist item in the given scan direction (or to
+ * the last item, when scanning backwards). Also advances/increments batch
+ * offset from position such that it points to newBatchForPos.
+ */
+static inline void
+heap_batchpos_newbatch(BatchIndexScan newBatchForPos, BatchRingItemPos *pos,
+ ScanDirection direction)
+{
+ Assert(newBatchForPos->dir == direction);
+
+ /* Next batch successfully loaded */
+ pos->batch++;
+ if (ScanDirectionIsForward(direction))
+ pos->item = newBatchForPos->firstItem;
+ else
+ pos->item = newBatchForPos->lastItem;
+
+ Assert(!INDEX_SCAN_POS_INVALID(pos));
+}
+
+/* ----------------
+ * heap_batch_getnext - get the next batch of TIDs from a scan
+ *
+ * Called when we need to load the next batch of index entries to process in
+ * the given direction.
+ *
+ * Returns the next batch to be processed by the index scan, or NULL when
+ * there are no more matches in the given scan direction. Also appends the
+ * returned batch to the end of the scan's batch ring buffer.
+ * ----------------
+ */
+static BatchIndexScan
+heap_batch_getnext(IndexScanDesc scan, BatchIndexScan priorbatch,
+ ScanDirection direction)
+{
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+ BatchIndexScan batch = NULL;
+
+ /* XXX: we should assert that a snapshot is pushed or registered */
+ Assert(TransactionIdIsValid(RecentXmin));
+
+ /*
+ * When caller provides a priorbatch it had better be for the last valid
+ * batch currently in the batch ring buffer (otherwise appending a new
+ * batch would result batches that aren't in scan order, which is wrong).
+ */
+ Assert(!INDEX_SCAN_BATCH_FULL(scan));
+ Assert(!priorbatch ||
+ (INDEX_SCAN_BATCH_COUNT(scan) > 0 &&
+ INDEX_SCAN_BATCH(scan, ringbuf->nextBatch - 1) == priorbatch));
+
+ if (scan->finished)
+ return NULL;
+
+ batch = scan->indexRelation->rd_indam->amgetbatch(scan, priorbatch,
+ direction);
+ if (batch != NULL)
+ {
+ /* We got the batch from the AM -- append it */
+ Assert(batch->dir == direction);
+
+ if (scan->xs_want_itup)
+ {
+ /*
+ * Index-only scan. Eagerly fetch visibility info from visibility
+ * map for all batch item TIDs.
+ */
+ heap_batch_resolve_visibility(scan, batch);
+ }
+
+ INDEX_SCAN_BATCH_APPEND(scan, batch);
+
+ DEBUG_LOG("batch_getnext headBatch %d nextBatch %d batch %p",
+ ringbuf->headBatch, ringbuf->nextBatch, batch);
+ }
+
+ /* xs_hitup is not supported by amgetbatch scans */
+ Assert(!scan->xs_hitup);
+
+ batch_assert_batches_valid(scan);
+
+ return batch;
+}
+
+/* ----------------
+ * heapam_batch_getnext_tid - get next TID from batch ring buffer
+ *
+ * This function implements heapam's version of getting the next TID from an
+ * index scan that uses the amgetbatch interface. It is implemented using
+ * various indexbatch.c utility routines.
+ * ----------------
+ */
+static ItemPointer
+heapam_batch_getnext_tid(IndexScanDesc scan, ScanDirection direction)
+{
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+ BatchRingItemPos *scanPos = &ringbuf->scanPos;
+ BatchIndexScan scanBatch = NULL;
+
+ /* shouldn't get here without batching */
+ batch_assert_batches_valid(scan);
+
+ /* xs_hitup is not supported by amgetbatch scans */
+ Assert(!scan->xs_hitup);
+
+ /* Initialize direction on first call */
+ if (ringbuf->direction == NoMovementScanDirection)
+ ringbuf->direction = direction;
+
+ if (unlikely(ringbuf->direction != direction))
+ {
+ /* We may change direction after reading the last batch. */
+ scan->finished = false;
+ }
+
+ /*
+ * Try advancing the position in the current batch. If that doesn't
+ * succeed, it means we don't have more items in it, and we need to
+ * advance to the next one (in the new scan direction).
+ */
+ if (INDEX_SCAN_BATCH_LOADED(scan, scanPos->batch))
+ {
+ scanBatch = INDEX_SCAN_BATCH(scan, scanPos->batch);
+
+ if (heap_batchpos_advance(scanBatch, scanPos, direction))
+ return heapam_batch_return_tid(scan, scanBatch, scanPos);
+ }
+
+ if (unlikely(ringbuf->direction != direction))
+ {
+ /* XXX shouldn't heapam_batch_rewind update the scanPos too? */
+ heapam_batch_rewind(scan, ringbuf, direction);
+
+ /*
+ * XXX It seems a bit weird to update just the batch part of the
+ * scanPos. Doesn't it make it rather wrong, with the item still set
+ * from the original batch? The next code block sets item too. So it
+ * seems we're doing this only to "fake" the batch, and then the next
+ * block will advance batch and reset the item. It's confusing, worth
+ * documenting. Maybe we should set item=-1?
+ */
+ scanPos->batch = ringbuf->nextBatch - 1;
+ }
+
+ /*
+ * Ran out of items from scanBatch. Try to advance it to next batch.
+ */
+ if (INDEX_SCAN_BATCH_LOADED(scan, scanPos->batch + 1))
+ {
+ /* Next batch already loaded by heapam_getnext_stream */
+ scanBatch = INDEX_SCAN_BATCH(scan, scanPos->batch + 1);
+ }
+ else if ((scanBatch = heap_batch_getnext(scan, scanBatch, direction)) != NULL)
+ {
+ /* Called amgetbatch again, loading new scanBatch into ring buffer */
+ }
+ else
+ {
+ /*
+ * There are no more batches to be loaded in the current scan
+ * direction. Defensively reset the read position.
+ */
+ batch_reset_pos(scanPos);
+ scan->finished = true;
+
+ return NULL;
+ }
+
+ /* Position scanPos to the start of new scanBatch */
+ heap_batchpos_newbatch(scanBatch, scanPos, direction);
+ Assert(INDEX_SCAN_BATCH(scan, scanPos->batch) == scanBatch);
+
+ /* Free now-unneeded older batch/prior scanBatch */
+ if (scanPos->batch != ringbuf->headBatch)
+ {
+ BatchIndexScan headBatch = INDEX_SCAN_BATCH(scan,
+ ringbuf->headBatch);
+
+ /* Free the head batch (except when it's markBatch) */
+ batch_free(scan, headBatch);
+
+ /*
+ * In any case, remove the batch from the ring buffer, even if we kept
+ * it for mark/restore
+ */
+ ringbuf->headBatch++;
+
+ /* we can't skip any batches */
+ Assert(ringbuf->headBatch == scanPos->batch);
+ }
+
+ return heapam_batch_return_tid(scan, scanBatch, scanPos);
+}
+
+/* ----------------
+ * index_fetch_heap - get the scan's next heap tuple
+ *
+ * The result is a visible heap tuple associated with the index TID most
+ * recently fetched by our caller in scan->xs_heaptid, or NULL if no more
+ * matching tuples exist. (There can be more than one matching tuple because
+ * of HOT chains, although when using an MVCC snapshot it should be impossible
+ * for more than one such tuple to exist.)
+ *
+ * On success, the buffer containing the heap tup is pinned. The pin must be
+ * dropped elsewhere.
+ * ----------------
+ */
+static bool
+index_fetch_heap(IndexScanDesc scan, TupleTableSlot *slot)
+{
+ bool all_dead = false;
+ bool found;
+
+ found = heapam_index_fetch_tuple(scan->xs_heapfetch, &scan->xs_heaptid,
+ scan->xs_snapshot, slot,
+ &scan->xs_heap_continue, &all_dead);
+
+ if (found)
+ pgstat_count_heap_fetch(scan->indexRelation);
+
+ /*
+ * If we scanned a whole HOT chain and found only dead tuples, remember it
+ * for later. We do not do this when in recovery because it may violate
+ * MVCC to do so. See comments in RelationGetIndexScan().
+ */
+ if (!scan->xactStartedInRecovery)
+ {
+ if (scan->ringbuf)
+ {
+ if (all_dead)
+ index_batch_kill_item(scan);
+ }
+ else
+ {
+ /*
+ * Tell amgettuple-based index AM to kill its entry for that TID
+ * (this will take effect in the next call, in index_getnext_tid)
+ */
+ scan->kill_prior_tuple = all_dead;
+ }
+ }
+
+ return found;
+}
+
+/* ----------------
+ * heapam_index_getnext_slot - get the next tuple from a scan
+ *
+ * The result is true if a tuple satisfying the scan keys and the snapshot was
+ * found, false otherwise. The tuple is stored in the specified slot.
+ *
+ * On success, resources (like buffer pins) are likely to be held, and will be
+ * dropped by a future call here (or by a later call to index_endscan).
+ *
+ * Note: caller must check scan->xs_recheck, and perform rechecking of the
+ * scan keys if required. We do not do that here because we don't have
+ * enough information to do it efficiently in the general case.
+ * ----------------
+ */
+static bool
+heapam_index_getnext_slot(IndexScanDesc scan, ScanDirection direction,
+ TupleTableSlot *slot)
+{
+ IndexFetchHeapData *hscan = (IndexFetchHeapData *) scan->xs_heapfetch;
+ ItemPointer tid = NULL;
+
+ for (;;)
+ {
+ if (!scan->xs_heap_continue)
+ {
+ /*
+ * Scans that use an amgetbatch index AM are managed by heapam's
+ * index scan manager. This gives heapam the ability to read heap
+ * tuples in a flexible order that is attuned to both costs and
+ * benefits on the heapam and table AM side.
+ *
+ * Scans that use an amgettuple index AM simply call through to
+ * index_getnext_tid to get the next TID returned by index AM. The
+ * progress of the scan will be under the control of index AM (we
+ * just pass it through a direction to get the next tuple in), so
+ * we cannot reorder any work.
+ */
+ if (scan->ringbuf != NULL)
+ tid = heapam_batch_getnext_tid(scan, direction);
+ else
+ {
+ tid = index_getnext_tid(scan, direction);
+
+ /*
+ * make sure to set the xs_visible flag, just like it's done
+ * in heapam_batch_getnext_tid
+ */
+ if ((tid != NULL) && (scan->xs_want_itup))
+ scan->xs_visible = VM_ALL_VISIBLE(scan->heapRelation,
+ ItemPointerGetBlockNumber(tid),
+ &hscan->vmbuf);
+ }
+
+ /* If we're out of index entries, we're done */
+ if (tid == NULL)
+ break;
+ }
+
+ /*
+ * Fetch the next (or only) visible heap tuple for this index entry.
+ * If we don't find anything, loop around and grab the next TID from
+ * the index.
+ */
+ Assert(ItemPointerIsValid(&scan->xs_heaptid));
+ if (!scan->xs_want_itup)
+ {
+ /* Plain index scan */
+ if (index_fetch_heap(scan, slot))
+ return true;
+ }
+ else
+ {
+ /*
+ * Index-only scan.
+ *
+ * We can skip the heap fetch if the TID references a heap page on
+ * which all tuples are known visible to everybody. In any case,
+ * we'll use the index tuple not the heap tuple as the data
+ * source.
+ *
+ * Note on Memory Ordering Effects: visibilitymap_get_status does
+ * not lock the visibility map buffer, and therefore the result we
+ * read here could be slightly stale. However, it can't be stale
+ * enough to matter.
+ *
+ * We need to detect clearing a VM bit due to an insert right
+ * away, because the tuple is present in the index page but not
+ * visible. The reading of the TID by this scan (using a shared
+ * lock on the index buffer) is serialized with the insert of the
+ * TID into the index (using an exclusive lock on the index
+ * buffer). Because the VM bit is cleared before updating the
+ * index, and locking/unlocking of the index page acts as a full
+ * memory barrier, we are sure to see the cleared bit if we see a
+ * recently-inserted TID.
+ *
+ * Deletes do not update the index page (only VACUUM will clear
+ * out the TID), so the clearing of the VM bit by a delete is not
+ * serialized with this test below, and we may see a value that is
+ * significantly stale. However, we don't care about the delete
+ * right away, because the tuple is still visible until the
+ * deleting transaction commits or the statement ends (if it's our
+ * transaction). In either case, the lock on the VM buffer will
+ * have been released (acting as a write barrier) after clearing
+ * the bit. And for us to have a snapshot that includes the
+ * deleting transaction (making the tuple invisible), we must have
+ * acquired ProcArrayLock after that time, acting as a read
+ * barrier.
+ *
+ * It's worth going through this complexity to avoid needing to
+ * lock the VM buffer, which could cause significant contention.
+ */
+ if (!scan->xs_visible)
+ {
+ /*
+ * Rats, we have to visit the heap to check visibility.
+ */
+ if (scan->instrument)
+ scan->instrument->nheapfetches++;
+
+ if (!index_fetch_heap(scan, slot))
+ continue; /* no visible tuple, try next index entry */
+
+ ExecClearTuple(slot);
+
+ /*
+ * Only MVCC snapshots are supported with standard index-only
+ * scans, so there should be no need to keep following the HOT
+ * chain once a visible entry has been found. Other callers
+ * (currently only selfuncs.c) use SnapshotNonVacuumable, and
+ * want us to assume that just having one visible tuple in the
+ * hot chain is always good enough.
+ */
+ Assert(!(scan->xs_heap_continue &&
+ IsMVCCSnapshot(scan->xs_snapshot)));
+
+ /*
+ * Note: at this point we are holding a pin on the heap page,
+ * as recorded in IndexFetchHeapData.xs_cbuf. We could
+ * release that pin now, but it's not clear whether it's a win
+ * to do so. The next index entry might require a visit to
+ * the same heap page.
+ */
+ }
+ else
+ {
+ /*
+ * We didn't access the heap, so we'll need to take a
+ * predicate lock explicitly, as if we had. For now we do
+ * that at page level.
+ */
+ PredicateLockPage(hscan->xs_base.rel,
+ ItemPointerGetBlockNumber(tid),
+ scan->xs_snapshot);
+ }
+
+ return true;
+ }
+ }
+
+ return false;
+}
/* ------------------------------------------------------------------------
* Callbacks for non-modifying operations on individual tuples for heap AM
@@ -753,7 +1289,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
tableScan = NULL;
heapScan = NULL;
- indexScan = index_beginscan(OldHeap, OldIndex, SnapshotAny, NULL, 0, 0);
+ indexScan = index_beginscan(OldHeap, OldIndex, false, SnapshotAny,
+ NULL, 0, 0);
index_rescan(indexScan, NULL, 0, NULL, 0);
}
else
@@ -790,7 +1327,8 @@ heapam_relation_copy_for_cluster(Relation OldHeap, Relation NewHeap,
if (indexScan != NULL)
{
- if (!index_getnext_slot(indexScan, ForwardScanDirection, slot))
+ if (!heapam_index_getnext_slot(indexScan, ForwardScanDirection,
+ slot))
break;
/* Since we used no scan keys, should never need to recheck */
@@ -2647,6 +3185,7 @@ static const TableAmRoutine heapam_methods = {
.index_fetch_begin = heapam_index_fetch_begin,
.index_fetch_reset = heapam_index_fetch_reset,
.index_fetch_end = heapam_index_fetch_end,
+ .index_getnext_slot = heapam_index_getnext_slot,
.index_fetch_tuple = heapam_index_fetch_tuple,
.tuple_insert = heapam_tuple_insert,
diff --git a/src/backend/access/index/Makefile b/src/backend/access/index/Makefile
index 6f2e3061a..e6d681b40 100644
--- a/src/backend/access/index/Makefile
+++ b/src/backend/access/index/Makefile
@@ -16,6 +16,7 @@ OBJS = \
amapi.o \
amvalidate.o \
genam.o \
- indexam.o
+ indexam.o \
+ indexbatch.o
include $(top_srcdir)/src/backend/common.mk
diff --git a/src/backend/access/index/genam.c b/src/backend/access/index/genam.c
index a29be6f46..095761a8f 100644
--- a/src/backend/access/index/genam.c
+++ b/src/backend/access/index/genam.c
@@ -89,6 +89,8 @@ RelationGetIndexScan(Relation indexRelation, int nkeys, int norderbys)
scan->xs_snapshot = InvalidSnapshot; /* caller must initialize this */
scan->numberOfKeys = nkeys;
scan->numberOfOrderBys = norderbys;
+ scan->ringbuf = NULL; /* set later for amgetbatch callers */
+ scan->xs_want_itup = false; /* caller must initialize this */
/*
* We allocate key workspace here, but it won't get filled until amrescan.
@@ -102,8 +104,6 @@ RelationGetIndexScan(Relation indexRelation, int nkeys, int norderbys)
else
scan->orderByData = NULL;
- scan->xs_want_itup = false; /* may be set later */
-
/*
* During recovery we ignore killed tuples and don't bother to kill them
* either. We do this because the xmin on the primary node could easily be
@@ -115,6 +115,8 @@ RelationGetIndexScan(Relation indexRelation, int nkeys, int norderbys)
* should not be altered by index AMs.
*/
scan->kill_prior_tuple = false;
+ scan->dropPin = true; /* for now */
+ scan->finished = false;
scan->xactStartedInRecovery = TransactionStartedDuringRecovery();
scan->ignore_killed_tuples = !scan->xactStartedInRecovery;
@@ -446,7 +448,7 @@ systable_beginscan(Relation heapRelation,
elog(ERROR, "column is not in index");
}
- sysscan->iscan = index_beginscan(heapRelation, irel,
+ sysscan->iscan = index_beginscan(heapRelation, irel, false,
snapshot, NULL, nkeys, 0);
index_rescan(sysscan->iscan, idxkey, nkeys, NULL, 0);
sysscan->scan = NULL;
@@ -517,7 +519,8 @@ systable_getnext(SysScanDesc sysscan)
if (sysscan->irel)
{
- if (index_getnext_slot(sysscan->iscan, ForwardScanDirection, sysscan->slot))
+ if (table_index_getnext_slot(sysscan->iscan, ForwardScanDirection,
+ sysscan->slot))
{
bool shouldFree;
@@ -707,7 +710,7 @@ systable_beginscan_ordered(Relation heapRelation,
elog(ERROR, "column is not in index");
}
- sysscan->iscan = index_beginscan(heapRelation, indexRelation,
+ sysscan->iscan = index_beginscan(heapRelation, indexRelation, false,
snapshot, NULL, nkeys, 0);
index_rescan(sysscan->iscan, idxkey, nkeys, NULL, 0);
sysscan->scan = NULL;
@@ -734,7 +737,7 @@ systable_getnext_ordered(SysScanDesc sysscan, ScanDirection direction)
HeapTuple htup = NULL;
Assert(sysscan->irel);
- if (index_getnext_slot(sysscan->iscan, direction, sysscan->slot))
+ if (table_index_getnext_slot(sysscan->iscan, direction, sysscan->slot))
htup = ExecFetchSlotHeapTuple(sysscan->slot, false, NULL);
/* See notes in systable_getnext */
diff --git a/src/backend/access/index/indexam.c b/src/backend/access/index/indexam.c
index 4ed0508c6..9eaecd943 100644
--- a/src/backend/access/index/indexam.c
+++ b/src/backend/access/index/indexam.c
@@ -24,9 +24,7 @@
* index_parallelscan_initialize - initialize parallel scan
* index_parallelrescan - (re)start a parallel scan of an index
* index_beginscan_parallel - join parallel index scan
- * index_getnext_tid - get the next TID from a scan
- * index_fetch_heap - get the scan's next heap tuple
- * index_getnext_slot - get the next tuple from a scan
+ * index_getnext_tid - amgettuple table AM helper routine
* index_getbitmap - get all tuples from a scan
* index_bulk_delete - bulk deletion of index tuples
* index_vacuum_cleanup - post-deletion cleanup of an index
@@ -255,6 +253,7 @@ index_insert_cleanup(Relation indexRelation,
IndexScanDesc
index_beginscan(Relation heapRelation,
Relation indexRelation,
+ bool xs_want_itup,
Snapshot snapshot,
IndexScanInstrumentation *instrument,
int nkeys, int norderbys)
@@ -282,6 +281,10 @@ index_beginscan(Relation heapRelation,
scan->heapRelation = heapRelation;
scan->xs_snapshot = snapshot;
scan->instrument = instrument;
+ scan->xs_want_itup = xs_want_itup;
+
+ if (indexRelation->rd_indam->amgetbatch != NULL)
+ index_batch_init(scan);
/* prepare to fetch index matches from table */
scan->xs_heapfetch = table_index_fetch_begin(heapRelation);
@@ -380,6 +383,15 @@ index_rescan(IndexScanDesc scan,
scan->kill_prior_tuple = false; /* for safety */
scan->xs_heap_continue = false;
+ /*
+ * ringbuf shouldn't be marked finished (must make sure that
+ * index_batch_reset doesn't see this, since indexam_util_batch_release
+ * will be affected)
+ */
+ scan->finished = false;
+
+ index_batch_reset(scan, true);
+
scan->indexRelation->rd_indam->amrescan(scan, keys, nkeys,
orderbys, norderbys);
}
@@ -394,6 +406,9 @@ index_endscan(IndexScanDesc scan)
SCAN_CHECKS;
CHECK_SCAN_PROCEDURE(amendscan);
+ /* Cleanup batching, so that the AM can release pins and so on. */
+ index_batch_end(scan);
+
/* Release resources (like buffer pins) from table accesses */
if (scan->xs_heapfetch)
{
@@ -422,9 +437,10 @@ void
index_markpos(IndexScanDesc scan)
{
SCAN_CHECKS;
- CHECK_SCAN_PROCEDURE(ammarkpos);
+ CHECK_SCAN_PROCEDURE(amposreset);
- scan->indexRelation->rd_indam->ammarkpos(scan);
+ /* Only amgetbatch index AMs support mark and restore */
+ index_batch_mark_pos(scan);
}
/* ----------------
@@ -448,7 +464,8 @@ index_restrpos(IndexScanDesc scan)
Assert(IsMVCCSnapshot(scan->xs_snapshot));
SCAN_CHECKS;
- CHECK_SCAN_PROCEDURE(amrestrpos);
+ CHECK_SCAN_PROCEDURE(amgetbatch);
+ CHECK_SCAN_PROCEDURE(amposreset);
/* release resources (like buffer pins) from table accesses */
if (scan->xs_heapfetch)
@@ -457,7 +474,7 @@ index_restrpos(IndexScanDesc scan)
scan->kill_prior_tuple = false; /* for safety */
scan->xs_heap_continue = false;
- scan->indexRelation->rd_indam->amrestrpos(scan);
+ index_batch_restore_pos(scan);
}
/*
@@ -579,6 +596,8 @@ index_parallelrescan(IndexScanDesc scan)
if (scan->xs_heapfetch)
table_index_fetch_reset(scan->xs_heapfetch);
+ index_batch_reset(scan, true);
+
/* amparallelrescan is optional; assume no-op if not provided by AM */
if (scan->indexRelation->rd_indam->amparallelrescan != NULL)
scan->indexRelation->rd_indam->amparallelrescan(scan);
@@ -591,6 +610,7 @@ index_parallelrescan(IndexScanDesc scan)
*/
IndexScanDesc
index_beginscan_parallel(Relation heaprel, Relation indexrel,
+ bool xs_want_itup,
IndexScanInstrumentation *instrument,
int nkeys, int norderbys,
ParallelIndexScanDesc pscan)
@@ -613,6 +633,10 @@ index_beginscan_parallel(Relation heaprel, Relation indexrel,
scan->heapRelation = heaprel;
scan->xs_snapshot = snapshot;
scan->instrument = instrument;
+ scan->xs_want_itup = xs_want_itup;
+
+ if (indexrel->rd_indam->amgetbatch != NULL)
+ index_batch_init(scan);
/* prepare to fetch index matches from table */
scan->xs_heapfetch = table_index_fetch_begin(heaprel);
@@ -621,10 +645,14 @@ index_beginscan_parallel(Relation heaprel, Relation indexrel,
}
/* ----------------
- * index_getnext_tid - get the next TID from a scan
+ * index_getnext_tid - amgettuple interface
*
* The result is the next TID satisfying the scan keys,
* or NULL if no more matching tuples exist.
+ *
+ * This should only be called by table AM's index_getnext_slot implementation,
+ * and only given an index AM that supports the single-tuple amgettuple
+ * interface.
* ----------------
*/
ItemPointer
@@ -667,97 +695,6 @@ index_getnext_tid(IndexScanDesc scan, ScanDirection direction)
return &scan->xs_heaptid;
}
-/* ----------------
- * index_fetch_heap - get the scan's next heap tuple
- *
- * The result is a visible heap tuple associated with the index TID most
- * recently fetched by index_getnext_tid, or NULL if no more matching tuples
- * exist. (There can be more than one matching tuple because of HOT chains,
- * although when using an MVCC snapshot it should be impossible for more than
- * one such tuple to exist.)
- *
- * On success, the buffer containing the heap tup is pinned (the pin will be
- * dropped in a future index_getnext_tid, index_fetch_heap or index_endscan
- * call).
- *
- * Note: caller must check scan->xs_recheck, and perform rechecking of the
- * scan keys if required. We do not do that here because we don't have
- * enough information to do it efficiently in the general case.
- * ----------------
- */
-bool
-index_fetch_heap(IndexScanDesc scan, TupleTableSlot *slot)
-{
- bool all_dead = false;
- bool found;
-
- found = table_index_fetch_tuple(scan->xs_heapfetch, &scan->xs_heaptid,
- scan->xs_snapshot, slot,
- &scan->xs_heap_continue, &all_dead);
-
- if (found)
- pgstat_count_heap_fetch(scan->indexRelation);
-
- /*
- * If we scanned a whole HOT chain and found only dead tuples, tell index
- * AM to kill its entry for that TID (this will take effect in the next
- * amgettuple call, in index_getnext_tid). We do not do this when in
- * recovery because it may violate MVCC to do so. See comments in
- * RelationGetIndexScan().
- */
- if (!scan->xactStartedInRecovery)
- scan->kill_prior_tuple = all_dead;
-
- return found;
-}
-
-/* ----------------
- * index_getnext_slot - get the next tuple from a scan
- *
- * The result is true if a tuple satisfying the scan keys and the snapshot was
- * found, false otherwise. The tuple is stored in the specified slot.
- *
- * On success, resources (like buffer pins) are likely to be held, and will be
- * dropped by a future index_getnext_tid, index_fetch_heap or index_endscan
- * call).
- *
- * Note: caller must check scan->xs_recheck, and perform rechecking of the
- * scan keys if required. We do not do that here because we don't have
- * enough information to do it efficiently in the general case.
- * ----------------
- */
-bool
-index_getnext_slot(IndexScanDesc scan, ScanDirection direction, TupleTableSlot *slot)
-{
- for (;;)
- {
- if (!scan->xs_heap_continue)
- {
- ItemPointer tid;
-
- /* Time to fetch the next TID from the index */
- tid = index_getnext_tid(scan, direction);
-
- /* If we're out of index entries, we're done */
- if (tid == NULL)
- break;
-
- Assert(ItemPointerEquals(tid, &scan->xs_heaptid));
- }
-
- /*
- * Fetch the next (or only) visible heap tuple for this index entry.
- * If we don't find anything, loop around and grab the next TID from
- * the index.
- */
- Assert(ItemPointerIsValid(&scan->xs_heaptid));
- if (index_fetch_heap(scan, slot))
- return true;
- }
-
- return false;
-}
-
/* ----------------
* index_getbitmap - get all tuples at once from an index scan
*
diff --git a/src/backend/access/index/indexbatch.c b/src/backend/access/index/indexbatch.c
new file mode 100644
index 000000000..86a1c6f56
--- /dev/null
+++ b/src/backend/access/index/indexbatch.c
@@ -0,0 +1,589 @@
+/*-------------------------------------------------------------------------
+ *
+ * indexbatch.c
+ * amgetbatch implementation routines
+ *
+ * Portions Copyright (c) 1996-2025, PostgreSQL Global Development Group
+ * Portions Copyright (c) 1994, Regents of the University of California
+ *
+ *
+ * IDENTIFICATION
+ * src/backend/access/index/indexbatch.c
+ *
+ * INTERFACE ROUTINES
+ * index_batch_init - Initialize fields needed by batching
+ * index_batch_reset - reset a batch
+ * index_batch_mark_pos - set a mark from current batch position
+ * index_batch_restore_pos - restore mark to current batch position
+ * index_batch_kill_item - record dead index tuple
+ * index_batch_end - end batch
+ *
+ * indexam_util_batch_unlock - unlock batch's buffer lock
+ * indexam_util_batch_alloc - allocate another batch
+ * indexam_util_batch_release - release allocated batch
+ *
+ *-------------------------------------------------------------------------
+ */
+
+#include "postgres.h"
+
+#include "access/amapi.h"
+#include "access/tableam.h"
+#include "common/int.h"
+#include "lib/qunique.h"
+#include "pgstat.h"
+#include "utils/memdebug.h"
+
+static int batch_compare_int(const void *va, const void *vb);
+static void batch_debug_print_batches(const char *label, IndexScanDesc scan);
+
+/*
+ * index_batch_init
+ * Initialize various fields and arrays needed by batching.
+ *
+ * Sets up the batch ring buffer structure and its initial read position.
+ * Also determines whether the scan will eagerly drop index page pins.
+ *
+ * Only call here when all of the index related fields in 'scan' were already
+ * initialized.
+ */
+void
+index_batch_init(IndexScanDesc scan)
+{
+ /* Both amgetbatch and amfreebatch must be present together */
+ Assert(scan->indexRelation->rd_indam->amgetbatch != NULL);
+ Assert(scan->indexRelation->rd_indam->amfreebatch != NULL);
+
+ scan->ringbuf = palloc_object(BatchRingBuffer);
+
+ /*
+ * We prefer to eagerly drop leaf page pins before amgetbatch returns.
+ * This avoids making VACUUM wait to acquire a cleanup lock on the page.
+ *
+ * We cannot safely drop leaf page pins during index-only scans due to a
+ * race condition involving VACUUM setting pages all-visible in the VM.
+ * It's also unsafe for plain index scans that use a non-MVCC snapshot.
+ *
+ * When we drop pins eagerly, the mechanism that marks index tuples as
+ * LP_DEAD has to deal with concurrent TID recycling races. The scheme
+ * used to detect unsafe TID recycling won't work when scanning unlogged
+ * relations (since it involves saving an affected page's LSN). Opt out
+ * of eager pin dropping during unlogged relation scans for now.
+ */
+ scan->dropPin =
+ (!scan->xs_want_itup && IsMVCCSnapshot(scan->xs_snapshot) &&
+ RelationNeedsWAL(scan->indexRelation));
+ scan->finished = false;
+ scan->ringbuf->direction = NoMovementScanDirection;
+
+ /* positions in the ring buffer of batches */
+ batch_reset_pos(&scan->ringbuf->scanPos);
+ batch_reset_pos(&scan->ringbuf->markPos);
+
+ scan->ringbuf->markBatch = NULL;
+ scan->ringbuf->headBatch = 0; /* initial head batch */
+ scan->ringbuf->nextBatch = 0; /* initial batch starts empty */
+ memset(&scan->ringbuf->cache, 0, sizeof(scan->ringbuf->cache));
+}
+
+/* ----------------
+ * index_batch_reset - reset batch ring buffer and read position
+ *
+ * Resets all loaded batches in the ring buffer, and resets the read position
+ * to the initial state (or just initialize ring buffer state). When
+ * 'complete' is true, also frees the scan's marked batch (if any), which is
+ * useful when ending an amgetbatch-based index scan.
+ * ----------------
+ */
+void
+index_batch_reset(IndexScanDesc scan, bool complete)
+{
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+
+ /* bail out if batching not enabled */
+ if (!ringbuf)
+ return;
+
+ batch_assert_batches_valid(scan);
+ batch_debug_print_batches("index_batch_reset", scan);
+ Assert(scan->xs_heapfetch);
+
+ /* reset the positions */
+ batch_reset_pos(&ringbuf->scanPos);
+
+ /*
+ * With "complete" reset, make sure to also free the marked batch, either
+ * by just forgetting it (if it's still in the ring buffer), or by
+ * explicitly freeing it.
+ */
+ if (complete && unlikely(ringbuf->markBatch != NULL))
+ {
+ BatchRingItemPos *markPos = &ringbuf->markPos;
+ BatchIndexScan markBatch = ringbuf->markBatch;
+
+ /* always reset the position, forget the marked batch */
+ ringbuf->markBatch = NULL;
+
+ /*
+ * If we've already moved past the marked batch (it's not loaded into
+ * the ring buffer), free it explicitly now. Otherwise, it'll be
+ * freed along with the other loaded batches.
+ */
+ if (!INDEX_SCAN_BATCH_LOADED(scan, markPos->batch))
+ batch_free(scan, markBatch);
+
+ batch_reset_pos(&ringbuf->markPos);
+ }
+
+ /* now release all other currently loaded batches */
+ while (ringbuf->headBatch < ringbuf->nextBatch)
+ {
+ BatchIndexScan batch = INDEX_SCAN_BATCH(scan, ringbuf->headBatch);
+
+ DEBUG_LOG("freeing batch %d %p", ringbuf->headBatch, batch);
+
+ batch_free(scan, batch);
+
+ /* update the valid range, so that asserts / debugging works */
+ ringbuf->headBatch++;
+ }
+
+ /* reset relevant batch state fields */
+ ringbuf->headBatch = 0; /* initial batch */
+ ringbuf->nextBatch = 0; /* initial batch is empty */
+
+ scan->finished = false;
+
+ batch_assert_batches_valid(scan);
+}
+
+/* ----------------
+ * index_batch_mark_pos - mark current position in scan for restoration
+ *
+ * Saves the current read position and associated batch so that the scan can
+ * be restored to this point later, via a call to index_batch_restore_pos.
+ * The marked batch is retained and not freed until a new mark is set or the
+ * scan ends (or until the mark is restored).
+ * ----------------
+ */
+void
+index_batch_mark_pos(IndexScanDesc scan)
+{
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+ BatchRingItemPos *markPos = &ringbuf->markPos;
+ BatchIndexScan markBatch = ringbuf->markBatch;
+
+ /*
+ * Free the previous mark batch (if any), but only if the batch is no
+ * longer loaded into the ring buffer
+ */
+ if (markBatch && !INDEX_SCAN_BATCH_LOADED(scan, markPos->batch))
+ {
+ ringbuf->markBatch = NULL;
+ batch_free(scan, markBatch);
+ }
+
+ /* copy the scan's position */
+ ringbuf->markPos = ringbuf->scanPos;
+ ringbuf->markBatch = INDEX_SCAN_BATCH(scan, ringbuf->markPos.batch);
+
+ /* scanPos/markPos must be valid */
+ batch_assert_pos_valid(scan, &ringbuf->markPos);
+}
+
+/* ----------------
+ * index_batch_restore_pos - restore scan to a previously marked position
+ *
+ * Restores the scan to a position previously saved by index_batch_mark_pos.
+ * The marked batch is restored as the current batch, allowing the scan to
+ * resume from the marked position. Also notifies the index AM via a call to
+ * its amposreset routine, which allows it to invalidate any private state
+ * that independently tracks scan progress (such as array key state)
+ *
+ * Function currently just discards most batch ring buffer state. It might
+ * make sense to teach it to hold on to other nearby batches (still-held
+ * batches that are likely to be needed once the scan finishes returning
+ * matching items from the restored batch) as an optimization. Such a scheme
+ * would have the benefit of avoiding repeat calls to amgetbatch/repeatedly
+ * reading the same index pages.
+ * ----------------
+ */
+void
+index_batch_restore_pos(IndexScanDesc scan)
+{
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+ BatchRingItemPos *markPos = &ringbuf->markPos;
+ BatchRingItemPos *scanPos = &ringbuf->scanPos ;
+ BatchIndexScan markBatch = ringbuf->markBatch;
+
+ if (scanPos->batch == markPos->batch &&
+ scanPos->batch == ringbuf->headBatch)
+ {
+ /*
+ * We don't have to discard the scan's state after all, since the
+ * current headBatch is also the batch that we're restoring to
+ */
+ scanPos->item = markPos->item;
+ return;
+ }
+
+ /*
+ * Call amposreset to let index AM know to invalidate any private state
+ * that independently tracks the scan's progress
+ */
+ scan->indexRelation->rd_indam->amposreset(scan, markBatch);
+
+ /*
+ * Reset the batching state, except for the marked batch, and make it look
+ * like we have a single batch -- the marked one.
+ */
+ index_batch_reset(scan, false);
+
+ ringbuf->scanPos = *markPos;
+ ringbuf->nextBatch = ringbuf->headBatch = markPos->batch;
+
+ INDEX_SCAN_BATCH_APPEND(scan, markBatch);
+}
+
+/*
+ * batch_free
+ * Release resources associated with a batch returned by the index AM.
+ *
+ * Called by table AM's ordered index scan implementation when it is finished
+ * with a batch and wishes to release its resources.
+ *
+ * This calls the index AM's amfreebatch callback to release AM-specific
+ * resources, and to set LP_DEAD bits on the batch's index page. It isn't
+ * safe for table AMs to fetch table tuples using TIDs saved from a batch that
+ * was already freed: 'dropPin' scans need the index AM to retain a pin on the
+ * TID's index page, as an interlock against concurrent TID recycling.
+ */
+void
+batch_free(IndexScanDesc scan, BatchIndexScan batch)
+{
+ batch_assert_batch_valid(scan, batch);
+
+ /* don't free the batch that is marked */
+ if (batch == scan->ringbuf->markBatch)
+ return;
+
+ /*
+ * killedItems[] is now in whatever order the scan returned items in.
+ * Scrollable cursor scans might have even saved the same item/TID twice.
+ *
+ * Sort and unique-ify killedItems[]. That way the index AM can safely
+ * assume that items will always be in their original index page order.
+ */
+ if (batch->numKilled > 1)
+ {
+ qsort(batch->killedItems, batch->numKilled, sizeof(int),
+ batch_compare_int);
+ batch->numKilled = qunique(batch->killedItems, batch->numKilled,
+ sizeof(int), batch_compare_int);
+ }
+
+ scan->indexRelation->rd_indam->amfreebatch(scan, batch);
+}
+
+/* ----------------
+ * index_batch_kill_item - record item for deferred LP_DEAD marking
+ *
+ * Records the item index of the currently-read tuple in scanBatch's
+ * killedItems array. The items' index tuples will later be marked LP_DEAD
+ * when current scanBatch is freed by amfreebatch routine (see batch_free).
+ * ----------------
+ */
+void
+index_batch_kill_item(IndexScanDesc scan)
+{
+ BatchRingItemPos *scanPos = &scan->ringbuf->scanPos;
+ BatchIndexScan scanBatch = INDEX_SCAN_BATCH(scan, scanPos->batch);
+
+ batch_assert_pos_valid(scan, scanPos);
+
+ if (scanBatch->killedItems == NULL)
+ scanBatch->killedItems = palloc_array(int, scan->maxitemsbatch);
+ if (scanBatch->numKilled < scan->maxitemsbatch)
+ scanBatch->killedItems[scanBatch->numKilled++] = scanPos->item;
+}
+
+/* ----------------
+ * index_batch_end - end a batch scan and free all resources
+ *
+ * Called when an index scan is being ended, right before the owning scan
+ * descriptor goes away. Cleans up all batch related resources.
+ * ----------------
+ */
+void
+index_batch_end(IndexScanDesc scan)
+{
+ index_batch_reset(scan, true);
+
+ /* bail out without batching */
+ if (!scan->ringbuf)
+ return;
+
+ for (int i = 0; i < INDEX_SCAN_CACHE_BATCHES; i++)
+ {
+ BatchIndexScan cached = scan->ringbuf->cache[i];
+
+ if (cached == NULL)
+ continue;
+
+ if (cached->killedItems)
+ pfree(cached->killedItems);
+ if (cached->currTuples)
+ pfree(cached->currTuples);
+ pfree(cached);
+ }
+
+ pfree(scan->ringbuf);
+}
+
+/* ----------------------------------------------------------------
+ * utility functions called by amgetbatch index AMs
+ *
+ * These functions manage batch allocation, unlock/pin management, and batch
+ * resource recycling. Index AMs implementing amgetbatch should use these
+ * rather than managing buffers directly.
+ * ----------------------------------------------------------------
+ */
+
+/*
+ * indexam_util_batch_unlock - Drop lock and conditionally drop pin on batch page
+ *
+ * Unlocks caller's batch->buf in preparation for amgetbatch returning items
+ * saved in that batch. Manages the details of dropping the lock and possibly
+ * the pin for index AM caller (dropping the pin prevents VACUUM from blocking
+ * on acquiring a cleanup lock, but isn't always safe).
+ *
+ * Only call here when a batch has one or more matching items to return using
+ * amgetbatch (or for amgetbitmap to load into its bitmap of matching TIDs).
+ * When an index page has no matches, it's always safe for index AMs to drop
+ * both the lock and the pin for themselves.
+ *
+ * Note: It is convenient for index AMs that implement both amgetbitmap and
+ * amgetbitmap to consistently use the same batch management approach, since
+ * that avoids introducing special cases to lower-level code. We always drop
+ * both the lock and the pin on batch's page on behalf of amgetbitmap callers.
+ * Such amgetbitmap callers must be careful to free all batches with matching
+ * items once they're done saving the matching TIDs (there will never be any
+ * calls to amfreebatch, so amgetbitmap must call indexam_util_batch_release
+ * directly, in lieu of a deferred call to amfreebatch from core code).
+ */
+void
+indexam_util_batch_unlock(IndexScanDesc scan, BatchIndexScan batch)
+{
+ Relation rel = scan->indexRelation;
+ bool dropPin = scan->dropPin;
+
+ /* batch must have one or more matching items returned by index AM */
+ Assert(batch->firstItem >= 0 && batch->firstItem <= batch->lastItem);
+
+ if (!dropPin)
+ {
+ if (!RelationUsesLocalBuffers(rel))
+ VALGRIND_MAKE_MEM_NOACCESS(BufferGetPage(batch->buf), BLCKSZ);
+
+ /* Just drop the lock (not the pin) */
+ LockBuffer(batch->buf, BUFFER_LOCK_UNLOCK);
+ return;
+ }
+
+ if (scan->ringbuf)
+ {
+ /* amgetbatch (not amgetbitmap) caller */
+ Assert(scan->heapRelation != NULL);
+
+ /*
+ * Have to set batch->lsn so that amfreebatch has a way to detect when
+ * concurrent heap TID recycling by VACUUM might have taken place.
+ * It'll only be safe to set any index tuple LP_DEAD bits when the
+ * page LSN hasn't advanced.
+ */
+ Assert(RelationNeedsWAL(rel));
+ Assert(!scan->xs_want_itup);
+ batch->lsn = BufferGetLSNAtomic(batch->buf);
+ }
+
+ /* Drop both the lock and the pin */
+ LockBuffer(batch->buf, BUFFER_LOCK_UNLOCK);
+ if (!RelationUsesLocalBuffers(rel))
+ VALGRIND_MAKE_MEM_NOACCESS(BufferGetPage(batch->buf), BLCKSZ);
+ ReleaseBuffer(batch->buf);
+ batch->buf = InvalidBuffer;
+}
+
+/*
+ * indexam_util_batch_alloc
+ * Allocate a batch during a amgetbatch (or amgetbitmap) index scan.
+ *
+ * Returns BatchIndexScan with space to fit scan->maxitemsbatch-many
+ * BatchMatchingItem entries. This will either be a newly allocated batch, or
+ * a batch recycled from the cache managed by indexam_util_batch_release. See
+ * comments above indexam_util_batch_release.
+ *
+ * Index AMs that use batches should call this from either their amgetbatch or
+ * amgetbitmap routines only. Note in particular that it cannot safely be
+ * called from a amfreebatch routine.
+ */
+BatchIndexScan
+indexam_util_batch_alloc(IndexScanDesc scan)
+{
+ BatchIndexScan batch = NULL;
+
+ /* First look for an existing batch from ring buffer */
+ if (scan->ringbuf != NULL)
+ {
+ for (int i = 0; i < INDEX_SCAN_CACHE_BATCHES; i++)
+ {
+ if (scan->ringbuf->cache[i] != NULL)
+ {
+ /* Return cached unreferenced batch */
+ batch = scan->ringbuf->cache[i];
+ scan->ringbuf->cache[i] = NULL;
+ break;
+ }
+ }
+ }
+
+ if (!batch)
+ {
+ batch = palloc(offsetof(BatchIndexScanData, items) +
+ sizeof(BatchMatchingItem) * scan->maxitemsbatch);
+
+ /*
+ * If we are doing an index-only scan, we need a tuple storage
+ * workspace. We allocate BLCKSZ for this, which should always give
+ * the index AM enough space to fit a full page's worth of tuples.
+ */
+ batch->currTuples = NULL;
+ if (scan->xs_want_itup)
+ batch->currTuples = palloc(BLCKSZ);
+
+ /*
+ * Batches allocate killedItems lazily (though note that cached
+ * batches keep their killedItems allocation when recycled)
+ */
+ batch->killedItems = NULL;
+ }
+
+ /* xs_want_itup scans must get a currTuples space */
+ Assert(!(scan->xs_want_itup && (batch->currTuples == NULL)));
+
+ /* shared initialization */
+ batch->buf = InvalidBuffer;
+ batch->firstItem = -1;
+ batch->lastItem = -1;
+ batch->numKilled = 0;
+
+ return batch;
+}
+
+/*
+ * indexam_util_batch_release
+ * Either stash the batch in a small cache for reuse, or free it.
+ *
+ * This function is called by index AMs to release a batch allocated by
+ * indexam_util_batch_alloc. Batches are cached here for reuse (when scan
+ * hasn't already finished) to reduce palloc/pfree overhead.
+ *
+ * It's safe to release a batch immediately when it was used to read a page
+ * that returned no matches to the scan. Batches actually returned by index
+ * AM's amgetbatch routine (i.e. batches for pages with one or more matches)
+ * must be released by calling here at the end of their amfreebatch routine.
+ * Index AMs that uses batches should call here to release a batch from any of
+ * their amgetbatch, amgetbitmap, and amfreebatch routines.
+ */
+void
+indexam_util_batch_release(IndexScanDesc scan, BatchIndexScan batch)
+{
+ Assert(batch->buf == InvalidBuffer);
+
+ if (scan->ringbuf)
+ {
+ /* amgetbatch scan caller */
+ Assert(scan->heapRelation != NULL);
+
+ if (scan->finished)
+ {
+ /* Don't bother using cache when scan is ending */
+ }
+ else
+ {
+ /*
+ * Use cache. This is generally only beneficial when there are
+ * many small rescans of an index.
+ */
+ for (int i = 0; i < INDEX_SCAN_CACHE_BATCHES; i++)
+ {
+ if (scan->ringbuf->cache[i] == NULL)
+ {
+ /* found empty slot, we're done */
+ scan->ringbuf->cache[i] = batch;
+ return;
+ }
+ }
+ }
+
+ /*
+ * Failed to find a free slot for this batch. We'll just free it
+ * ourselves. This isn't really expected; it's just defensive.
+ */
+ if (batch->killedItems)
+ pfree(batch->killedItems);
+ if (batch->currTuples)
+ pfree(batch->currTuples);
+ }
+ else
+ {
+ /* amgetbitmap scan caller */
+ Assert(scan->heapRelation == NULL);
+ Assert(batch->killedItems == NULL);
+ Assert(batch->currTuples == NULL);
+ }
+
+ /* no free slot to save this batch (expected with amgetbitmap callers) */
+ pfree(batch);
+}
+
+/*
+ * qsort comparison function for int arrays
+ */
+static int
+batch_compare_int(const void *va, const void *vb)
+{
+ int a = *((const int *) va);
+ int b = *((const int *) vb);
+
+ return pg_cmp_s32(a, b);
+}
+
+static void
+batch_debug_print_batches(const char *label, IndexScanDesc scan)
+{
+#ifdef INDEXAM_DEBUG
+ BatchRingBuffer *ringbuf = scan->ringbuf;
+
+ if (!ringbuf)
+ return;
+
+ if (!AmRegularBackendProcess())
+ return;
+ if (IsCatalogRelation(scan->indexRelation))
+ return;
+
+ DEBUG_LOG("%s: batches headBatch %d nextBatch %d",
+ label,
+ ringbuf->headBatch, ringbuf->nextBatch);
+
+ for (int i = ringbuf->headBatch; i < ringbuf->nextBatch; i++)
+ {
+ BatchIndexScan batch = INDEX_SCAN_BATCH(scan, i);
+
+ DEBUG_LOG(" batch %d currPage %u %p firstItem %d lastItem %d killed %d",
+ i, batch->currPage, batch, batch->firstItem,
+ batch->lastItem, batch->numKilled);
+ }
+#endif
+}
diff --git a/src/backend/access/index/meson.build b/src/backend/access/index/meson.build
index da64cb595..83dfa3f2b 100644
--- a/src/backend/access/index/meson.build
+++ b/src/backend/access/index/meson.build
@@ -5,4 +5,5 @@ backend_sources += files(
'amvalidate.c',
'genam.c',
'indexam.c',
+ 'indexbatch.c',
)
diff --git a/src/backend/access/nbtree/README b/src/backend/access/nbtree/README
index 53d4a61dc..231da20e4 100644
--- a/src/backend/access/nbtree/README
+++ b/src/backend/access/nbtree/README
@@ -471,7 +471,7 @@ proper. A plain index scan will even recognize LP_UNUSED items in the
heap (items that could be recycled but haven't been just yet) as "not
visible" -- even when the heap page is generally considered all-visible.
-LP_DEAD setting of index tuples by the kill_prior_tuple optimization
+Opportunistic LP_DEAD setting of known-dead index tuples during index scans
(described in full in simple deletion, below) is also more complicated for
index scans that drop their leaf page pins. We must be careful to avoid
LP_DEAD-marking any new index tuple that looks like a known-dead index
@@ -481,7 +481,7 @@ new, unrelated index tuple, on the same leaf page, which has the same
original TID. It would be totally wrong to LP_DEAD-set this new,
unrelated index tuple.
-We handle this kill_prior_tuple race condition by having affected index
+We handle this LP_DEAD setting race condition by having affected index
scans conservatively assume that any change to the leaf page at all
implies that it was reached by btbulkdelete in the interim period when no
buffer pin was held. This is implemented by not setting any LP_DEAD bits
@@ -735,7 +735,7 @@ of readers could still move right to recover if we didn't couple
same-level locks), but we prefer to be conservative here.
During recovery all index scans start with ignore_killed_tuples = false
-and we never set kill_prior_tuple. We do this because the oldest xmin
+and we never LP_DEAD-mark tuples. We do this because the oldest xmin
on the standby server can be older than the oldest xmin on the primary
server, which means tuples can be marked LP_DEAD even when they are
still visible on the standby. We don't WAL log tuple LP_DEAD bits, but
diff --git a/src/backend/access/nbtree/nbtpage.c b/src/backend/access/nbtree/nbtpage.c
index 4125c185e..21bb80168 100644
--- a/src/backend/access/nbtree/nbtpage.c
+++ b/src/backend/access/nbtree/nbtpage.c
@@ -1033,6 +1033,9 @@ _bt_relbuf(Relation rel, Buffer buf)
* Lock is acquired without acquiring another pin. This is like a raw
* LockBuffer() call, but performs extra steps needed by Valgrind.
*
+ * Note: indexam_util_batch_unlock has similar Valgrind buffer lock
+ * instrumentation, which we rely on here.
+ *
* Note: Caller may need to call _bt_checkpage() with buf when pin on buf
* wasn't originally acquired in _bt_getbuf() or _bt_relandgetbuf().
*/
diff --git a/src/backend/access/nbtree/nbtreadpage.c b/src/backend/access/nbtree/nbtreadpage.c
index 2ba1ca660..69cbe4784 100644
--- a/src/backend/access/nbtree/nbtreadpage.c
+++ b/src/backend/access/nbtree/nbtreadpage.c
@@ -32,6 +32,7 @@ typedef struct BTReadPageState
{
/* Input parameters, set by _bt_readpage for _bt_checkkeys */
ScanDirection dir; /* current scan direction */
+ BlockNumber currpage; /* current page being read */
OffsetNumber minoff; /* Lowest non-pivot tuple's offset */
OffsetNumber maxoff; /* Highest non-pivot tuple's offset */
IndexTuple finaltup; /* Needed by scans with array keys */
@@ -63,14 +64,13 @@ static bool _bt_scanbehind_checkkeys(IndexScanDesc scan, ScanDirection dir,
IndexTuple finaltup);
static bool _bt_oppodir_checkkeys(IndexScanDesc scan, ScanDirection dir,
IndexTuple finaltup);
-static void _bt_saveitem(BTScanOpaque so, int itemIndex,
- OffsetNumber offnum, IndexTuple itup);
-static int _bt_setuppostingitems(BTScanOpaque so, int itemIndex,
+static void _bt_saveitem(BatchIndexScan newbatch, int itemIndex, OffsetNumber offnum,
+ IndexTuple itup, int *tupleOffset);
+static int _bt_setuppostingitems(BatchIndexScan newbatch, int itemIndex,
OffsetNumber offnum, const ItemPointerData *heapTid,
- IndexTuple itup);
-static inline void _bt_savepostingitem(BTScanOpaque so, int itemIndex,
- OffsetNumber offnum,
- ItemPointer heapTid, int tupleOffset);
+ IndexTuple itup, int *tupleOffset);
+static inline void _bt_savepostingitem(BatchIndexScan newbatch, int itemIndex, OffsetNumber offnum,
+ ItemPointer heapTid, int baseOffset);
static bool _bt_checkkeys(IndexScanDesc scan, BTReadPageState *pstate, bool arrayKeys,
IndexTuple tuple, int tupnatts);
static bool _bt_check_compare(IndexScanDesc scan, ScanDirection dir,
@@ -111,15 +111,15 @@ static bool _bt_verify_keys_with_arraykeys(IndexScanDesc scan);
/*
- * _bt_readpage() -- Load data from current index page into so->currPos
+ * _bt_readpage() -- Load data from current index page into newbatch.
*
- * Caller must have pinned and read-locked so->currPos.buf; the buffer's state
- * is not changed here. Also, currPos.moreLeft and moreRight must be valid;
- * they are updated as appropriate. All other fields of so->currPos are
+ * Caller must have pinned and read-locked newbatch.buf; the buffer's state is
+ * not changed here. Also, newbatch's moreLeft and moreRight must be valid;
+ * they are updated as appropriate. All other fields of newbatch are
* initialized from scratch here.
*
* We scan the current page starting at offnum and moving in the indicated
- * direction. All items matching the scan keys are loaded into currPos.items.
+ * direction. All items matching the scan keys are saved in newbatch.items.
* moreLeft or moreRight (as appropriate) is cleared if _bt_checkkeys reports
* that there can be no more matching tuples in the current scan direction
* (could just be for the current primitive index scan when scan has arrays).
@@ -131,8 +131,8 @@ static bool _bt_verify_keys_with_arraykeys(IndexScanDesc scan);
* Returns true if any matching items found on the page, false if none.
*/
bool
-_bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
- bool firstpage)
+_bt_readpage(IndexScanDesc scan, BatchIndexScan newbatch, ScanDirection dir,
+ OffsetNumber offnum, bool firstpage)
{
Relation rel = scan->indexRelation;
BTScanOpaque so = (BTScanOpaque) scan->opaque;
@@ -144,23 +144,20 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
bool arrayKeys,
ignore_killed_tuples = scan->ignore_killed_tuples;
int itemIndex,
+ tupleOffset = 0,
indnatts;
/* save the page/buffer block number, along with its sibling links */
- page = BufferGetPage(so->currPos.buf);
+ page = BufferGetPage(newbatch->buf);
opaque = BTPageGetOpaque(page);
- so->currPos.currPage = BufferGetBlockNumber(so->currPos.buf);
- so->currPos.prevPage = opaque->btpo_prev;
- so->currPos.nextPage = opaque->btpo_next;
- /* delay setting so->currPos.lsn until _bt_drop_lock_and_maybe_pin */
- pstate.dir = so->currPos.dir = dir;
- so->currPos.nextTupleOffset = 0;
+ pstate.currpage = newbatch->currPage = BufferGetBlockNumber(newbatch->buf);
+ newbatch->prevPage = opaque->btpo_prev;
+ newbatch->nextPage = opaque->btpo_next;
+ pstate.dir = newbatch->dir = dir;
/* either moreRight or moreLeft should be set now (may be unset later) */
- Assert(ScanDirectionIsForward(dir) ? so->currPos.moreRight :
- so->currPos.moreLeft);
+ Assert(ScanDirectionIsForward(dir) ? newbatch->moreRight : newbatch->moreLeft);
Assert(!P_IGNORE(opaque));
- Assert(BTScanPosIsPinned(so->currPos));
Assert(!so->needPrimScan);
/* initialize local variables */
@@ -188,14 +185,12 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
{
/* allow next/prev page to be read by other worker without delay */
if (ScanDirectionIsForward(dir))
- _bt_parallel_release(scan, so->currPos.nextPage,
- so->currPos.currPage);
+ _bt_parallel_release(scan, newbatch->nextPage, newbatch->currPage);
else
- _bt_parallel_release(scan, so->currPos.prevPage,
- so->currPos.currPage);
+ _bt_parallel_release(scan, newbatch->prevPage, newbatch->currPage);
}
- PredicateLockPage(rel, so->currPos.currPage, scan->xs_snapshot);
+ PredicateLockPage(rel, pstate.currpage, scan->xs_snapshot);
if (ScanDirectionIsForward(dir))
{
@@ -212,11 +207,10 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
!_bt_scanbehind_checkkeys(scan, dir, pstate.finaltup))
{
/* Schedule another primitive index scan after all */
- so->currPos.moreRight = false;
+ newbatch->moreRight = false;
so->needPrimScan = true;
if (scan->parallel_scan)
- _bt_parallel_primscan_schedule(scan,
- so->currPos.currPage);
+ _bt_parallel_primscan_schedule(scan, newbatch->currPage);
return false;
}
}
@@ -280,26 +274,26 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
if (!BTreeTupleIsPosting(itup))
{
/* Remember it */
- _bt_saveitem(so, itemIndex, offnum, itup);
+ _bt_saveitem(newbatch, itemIndex, offnum, itup, &tupleOffset);
itemIndex++;
}
else
{
- int tupleOffset;
+ int baseOffset;
/* Set up posting list state (and remember first TID) */
- tupleOffset =
- _bt_setuppostingitems(so, itemIndex, offnum,
+ baseOffset =
+ _bt_setuppostingitems(newbatch, itemIndex, offnum,
BTreeTupleGetPostingN(itup, 0),
- itup);
+ itup, &tupleOffset);
itemIndex++;
/* Remember all later TIDs (must be at least one) */
for (int i = 1; i < BTreeTupleGetNPosting(itup); i++)
{
- _bt_savepostingitem(so, itemIndex, offnum,
+ _bt_savepostingitem(newbatch, itemIndex, offnum,
BTreeTupleGetPostingN(itup, i),
- tupleOffset);
+ baseOffset);
itemIndex++;
}
}
@@ -339,12 +333,11 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
}
if (!pstate.continuescan)
- so->currPos.moreRight = false;
+ newbatch->moreRight = false;
Assert(itemIndex <= MaxTIDsPerBTreePage);
- so->currPos.firstItem = 0;
- so->currPos.lastItem = itemIndex - 1;
- so->currPos.itemIndex = 0;
+ newbatch->firstItem = 0;
+ newbatch->lastItem = itemIndex - 1;
}
else
{
@@ -361,11 +354,10 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
!_bt_scanbehind_checkkeys(scan, dir, pstate.finaltup))
{
/* Schedule another primitive index scan after all */
- so->currPos.moreLeft = false;
+ newbatch->moreLeft = false;
so->needPrimScan = true;
if (scan->parallel_scan)
- _bt_parallel_primscan_schedule(scan,
- so->currPos.currPage);
+ _bt_parallel_primscan_schedule(scan, newbatch->currPage);
return false;
}
}
@@ -466,27 +458,27 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
{
/* Remember it */
itemIndex--;
- _bt_saveitem(so, itemIndex, offnum, itup);
+ _bt_saveitem(newbatch, itemIndex, offnum, itup, &tupleOffset);
}
else
{
uint16 nitems = BTreeTupleGetNPosting(itup);
- int tupleOffset;
+ int baseOffset;
/* Set up posting list state (and remember last TID) */
itemIndex--;
- tupleOffset =
- _bt_setuppostingitems(so, itemIndex, offnum,
+ baseOffset =
+ _bt_setuppostingitems(newbatch, itemIndex, offnum,
BTreeTupleGetPostingN(itup, nitems - 1),
- itup);
+ itup, &tupleOffset);
/* Remember all prior TIDs (must be at least one) */
for (int i = nitems - 2; i >= 0; i--)
{
itemIndex--;
- _bt_savepostingitem(so, itemIndex, offnum,
+ _bt_savepostingitem(newbatch, itemIndex, offnum,
BTreeTupleGetPostingN(itup, i),
- tupleOffset);
+ baseOffset);
}
}
}
@@ -502,12 +494,11 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
* be found there
*/
if (!pstate.continuescan)
- so->currPos.moreLeft = false;
+ newbatch->moreLeft = false;
Assert(itemIndex >= 0);
- so->currPos.firstItem = itemIndex;
- so->currPos.lastItem = MaxTIDsPerBTreePage - 1;
- so->currPos.itemIndex = MaxTIDsPerBTreePage - 1;
+ newbatch->firstItem = itemIndex;
+ newbatch->lastItem = MaxTIDsPerBTreePage - 1;
}
/*
@@ -524,7 +515,7 @@ _bt_readpage(IndexScanDesc scan, ScanDirection dir, OffsetNumber offnum,
*/
Assert(!pstate.forcenonrequired);
- return (so->currPos.firstItem <= so->currPos.lastItem);
+ return (newbatch->firstItem <= newbatch->lastItem);
}
/*
@@ -1027,90 +1018,96 @@ _bt_oppodir_checkkeys(IndexScanDesc scan, ScanDirection dir,
return true;
}
-/* Save an index item into so->currPos.items[itemIndex] */
+/* Save an index item into newbatch.items[itemIndex] */
static void
-_bt_saveitem(BTScanOpaque so, int itemIndex,
- OffsetNumber offnum, IndexTuple itup)
+_bt_saveitem(BatchIndexScan newbatch, int itemIndex, OffsetNumber offnum,
+ IndexTuple itup, int *tupleOffset)
{
- BTScanPosItem *currItem = &so->currPos.items[itemIndex];
-
Assert(!BTreeTupleIsPivot(itup) && !BTreeTupleIsPosting(itup));
- currItem->heapTid = itup->t_tid;
- currItem->indexOffset = offnum;
- if (so->currTuples)
+ /* copy the populated part of the items array */
+ newbatch->items[itemIndex].heapTid = itup->t_tid;
+ newbatch->items[itemIndex].indexOffset = offnum;
+ newbatch->items[itemIndex].allVisible = false;
+
+ if (newbatch->currTuples)
{
Size itupsz = IndexTupleSize(itup);
- currItem->tupleOffset = so->currPos.nextTupleOffset;
- memcpy(so->currTuples + so->currPos.nextTupleOffset, itup, itupsz);
- so->currPos.nextTupleOffset += MAXALIGN(itupsz);
+ newbatch->items[itemIndex].tupleOffset = *tupleOffset;
+ memcpy(newbatch->currTuples + *tupleOffset, itup, itupsz);
+ *tupleOffset += MAXALIGN(itupsz);
}
}
/*
* Setup state to save TIDs/items from a single posting list tuple.
*
- * Saves an index item into so->currPos.items[itemIndex] for TID that is
- * returned to scan first. Second or subsequent TIDs for posting list should
- * be saved by calling _bt_savepostingitem().
+ * Saves an index item into newbatch.items[itemIndex] for TID that is returned
+ * to scan first. Second or subsequent TIDs for posting list should be saved
+ * by calling _bt_savepostingitem().
*
- * Returns an offset into tuple storage space that main tuple is stored at if
- * needed.
+ * Returns baseOffset, an offset into tuple storage space that main tuple is
+ * stored at if needed.
*/
static int
-_bt_setuppostingitems(BTScanOpaque so, int itemIndex, OffsetNumber offnum,
- const ItemPointerData *heapTid, IndexTuple itup)
+_bt_setuppostingitems(BatchIndexScan newbatch, int itemIndex,
+ OffsetNumber offnum, const ItemPointerData *heapTid,
+ IndexTuple itup, int *tupleOffset)
{
- BTScanPosItem *currItem = &so->currPos.items[itemIndex];
+ BatchMatchingItem *item = &newbatch->items[itemIndex];
Assert(BTreeTupleIsPosting(itup));
- currItem->heapTid = *heapTid;
- currItem->indexOffset = offnum;
- if (so->currTuples)
+ /* copy the populated part of the items array */
+ item->heapTid = *heapTid;
+ item->indexOffset = offnum;
+ item->allVisible = false;
+
+ if (newbatch->currTuples)
{
/* Save base IndexTuple (truncate posting list) */
IndexTuple base;
Size itupsz = BTreeTupleGetPostingOffset(itup);
itupsz = MAXALIGN(itupsz);
- currItem->tupleOffset = so->currPos.nextTupleOffset;
- base = (IndexTuple) (so->currTuples + so->currPos.nextTupleOffset);
+ item->tupleOffset = *tupleOffset;
+ base = (IndexTuple) (newbatch->currTuples + *tupleOffset);
memcpy(base, itup, itupsz);
/* Defensively reduce work area index tuple header size */
base->t_info &= ~INDEX_SIZE_MASK;
base->t_info |= itupsz;
- so->currPos.nextTupleOffset += itupsz;
+ *tupleOffset += itupsz;
- return currItem->tupleOffset;
+ return item->tupleOffset;
}
return 0;
}
/*
- * Save an index item into so->currPos.items[itemIndex] for current posting
+ * Save an index item into newbatch.items[itemIndex] for current posting
* tuple.
*
* Assumes that _bt_setuppostingitems() has already been called for current
- * posting list tuple. Caller passes its return value as tupleOffset.
+ * posting list tuple. Caller passes its return value as baseOffset.
*/
static inline void
-_bt_savepostingitem(BTScanOpaque so, int itemIndex, OffsetNumber offnum,
- ItemPointer heapTid, int tupleOffset)
+_bt_savepostingitem(BatchIndexScan newbatch, int itemIndex, OffsetNumber offnum,
+ ItemPointer heapTid, int baseOffset)
{
- BTScanPosItem *currItem = &so->currPos.items[itemIndex];
+ BatchMatchingItem *item = &newbatch->items[itemIndex];
- currItem->heapTid = *heapTid;
- currItem->indexOffset = offnum;
+ item->heapTid = *heapTid;
+ item->indexOffset = offnum;
+ item->allVisible = false;
/*
* Have index-only scans return the same base IndexTuple for every TID
* that originates from the same posting list
*/
- if (so->currTuples)
- currItem->tupleOffset = tupleOffset;
+ if (newbatch->currTuples)
+ item->tupleOffset = baseOffset;
}
#define LOOK_AHEAD_REQUIRED_RECHECKS 3
@@ -2822,13 +2819,13 @@ new_prim_scan:
* Note: We make a soft assumption that the current scan direction will
* also be used within _bt_next, when it is asked to step off this page.
* It is up to _bt_next to cancel this scheduled primitive index scan
- * whenever it steps to a page in the direction opposite currPos.dir.
+ * whenever it steps to a page in the direction opposite pstate->dir.
*/
pstate->continuescan = false; /* Tell _bt_readpage we're done... */
so->needPrimScan = true; /* ...but call _bt_first again */
if (scan->parallel_scan)
- _bt_parallel_primscan_schedule(scan, so->currPos.currPage);
+ _bt_parallel_primscan_schedule(scan, pstate->currpage);
/* Caller's tuple doesn't match the new qual */
return false;
@@ -2913,14 +2910,6 @@ _bt_advance_array_keys_increment(IndexScanDesc scan, ScanDirection dir,
* Restore the array keys to the state they were in immediately before we
* were called. This ensures that the arrays only ever ratchet in the
* current scan direction.
- *
- * Without this, scans could overlook matching tuples when the scan
- * direction gets reversed just before btgettuple runs out of items to
- * return, but just after _bt_readpage prepares all the items from the
- * scan's final page in so->currPos. When we're on the final page it is
- * typical for so->currPos to get invalidated once btgettuple finally
- * returns false, which'll effectively invalidate the scan's array keys.
- * That hasn't happened yet, though -- and in general it may never happen.
*/
_bt_start_array_keys(scan, -dir);
diff --git a/src/backend/access/nbtree/nbtree.c b/src/backend/access/nbtree/nbtree.c
index 3dec1ee65..4877ccb48 100644
--- a/src/backend/access/nbtree/nbtree.c
+++ b/src/backend/access/nbtree/nbtree.c
@@ -159,11 +159,12 @@ bthandler(PG_FUNCTION_ARGS)
.amadjustmembers = btadjustmembers,
.ambeginscan = btbeginscan,
.amrescan = btrescan,
- .amgettuple = btgettuple,
+ .amgettuple = NULL,
+ .amgetbatch = btgetbatch,
+ .amfreebatch = btfreebatch,
.amgetbitmap = btgetbitmap,
.amendscan = btendscan,
- .ammarkpos = btmarkpos,
- .amrestrpos = btrestrpos,
+ .amposreset = btposreset,
.amestimateparallelscan = btestimateparallelscan,
.aminitparallelscan = btinitparallelscan,
.amparallelrescan = btparallelrescan,
@@ -222,13 +223,13 @@ btinsert(Relation rel, Datum *values, bool *isnull,
}
/*
- * btgettuple() -- Get the next tuple in the scan.
+ * btgetbatch() -- Get the first or next batch of tuples in the scan
*/
-bool
-btgettuple(IndexScanDesc scan, ScanDirection dir)
+BatchIndexScan
+btgetbatch(IndexScanDesc scan, BatchIndexScan priorbatch, ScanDirection dir)
{
BTScanOpaque so = (BTScanOpaque) scan->opaque;
- bool res;
+ BatchIndexScan batch = priorbatch;
Assert(scan->heapRelation != NULL);
@@ -241,45 +242,20 @@ btgettuple(IndexScanDesc scan, ScanDirection dir)
/*
* If we've already initialized this scan, we can just advance it in
* the appropriate direction. If we haven't done so yet, we call
- * _bt_first() to get the first item in the scan.
+ * _bt_first() to get the first batch in the scan.
*/
- if (!BTScanPosIsValid(so->currPos))
- res = _bt_first(scan, dir);
+ if (batch == NULL)
+ batch = _bt_first(scan, dir);
else
- {
- /*
- * Check to see if we should kill the previously-fetched tuple.
- */
- if (scan->kill_prior_tuple)
- {
- /*
- * Yes, remember it for later. (We'll deal with all such
- * tuples at once right before leaving the index page.) The
- * test for numKilled overrun is not just paranoia: if the
- * caller reverses direction in the indexscan then the same
- * item might get entered multiple times. It's not worth
- * trying to optimize that, so we don't detect it, but instead
- * just forget any excess entries.
- */
- if (so->killedItems == NULL)
- so->killedItems = palloc_array(int, MaxTIDsPerBTreePage);
- if (so->numKilled < MaxTIDsPerBTreePage)
- so->killedItems[so->numKilled++] = so->currPos.itemIndex;
- }
+ batch = _bt_next(scan, dir, batch);
- /*
- * Now continue the scan.
- */
- res = _bt_next(scan, dir);
- }
-
- /* If we have a tuple, return it ... */
- if (res)
+ /* If we have a batch, return it ... */
+ if (batch)
break;
/* ... otherwise see if we need another primitive index scan */
} while (so->numArrayKeys && _bt_start_prim_scan(scan));
- return res;
+ return batch;
}
/*
@@ -289,6 +265,7 @@ int64
btgetbitmap(IndexScanDesc scan, TIDBitmap *tbm)
{
BTScanOpaque so = (BTScanOpaque) scan->opaque;
+ BatchIndexScan batch;
int64 ntids = 0;
ItemPointer heapTid;
@@ -297,29 +274,29 @@ btgetbitmap(IndexScanDesc scan, TIDBitmap *tbm)
/* Each loop iteration performs another primitive index scan */
do
{
- /* Fetch the first page & tuple */
- if (_bt_first(scan, ForwardScanDirection))
+ /* Fetch the first batch */
+ if ((batch = _bt_first(scan, ForwardScanDirection)))
{
- /* Save tuple ID, and continue scanning */
- heapTid = &scan->xs_heaptid;
+ int itemIndex = 0;
+
+ /* Save first tuple's TID */
+ heapTid = &batch->items[itemIndex].heapTid;
tbm_add_tuples(tbm, heapTid, 1, false);
ntids++;
for (;;)
{
- /*
- * Advance to next tuple within page. This is the same as the
- * easy case in _bt_next().
- */
- if (++so->currPos.itemIndex > so->currPos.lastItem)
+ /* Advance to next TID within page-sized batch */
+ if (++itemIndex > batch->lastItem)
{
/* let _bt_next do the heavy lifting */
- if (!_bt_next(scan, ForwardScanDirection))
+ itemIndex = 0;
+ batch = _bt_next(scan, ForwardScanDirection, batch);
+ if (!batch)
break;
}
- /* Save tuple ID, and continue scanning */
- heapTid = &so->currPos.items[so->currPos.itemIndex].heapTid;
+ heapTid = &batch->items[itemIndex].heapTid;
tbm_add_tuples(tbm, heapTid, 1, false);
ntids++;
}
@@ -347,8 +324,6 @@ btbeginscan(Relation rel, int nkeys, int norderbys)
/* allocate private workspace */
so = palloc_object(BTScanOpaqueData);
- BTScanPosInvalidate(so->currPos);
- BTScanPosInvalidate(so->markPos);
if (scan->numberOfKeys > 0)
so->keyData = (ScanKey) palloc(scan->numberOfKeys * sizeof(ScanKeyData));
else
@@ -362,19 +337,9 @@ btbeginscan(Relation rel, int nkeys, int norderbys)
so->orderProcs = NULL;
so->arrayContext = NULL;
- so->killedItems = NULL; /* until needed */
- so->numKilled = 0;
-
- /*
- * We don't know yet whether the scan will be index-only, so we do not
- * allocate the tuple workspace arrays until btrescan. However, we set up
- * scan->xs_itupdesc whether we'll need it or not, since that's so cheap.
- */
- so->currTuples = so->markTuples = NULL;
-
- scan->xs_itupdesc = RelationGetDescr(rel);
-
scan->opaque = so;
+ scan->xs_itupdesc = RelationGetDescr(rel);
+ scan->maxitemsbatch = MaxTIDsPerBTreePage;
return scan;
}
@@ -388,72 +353,37 @@ btrescan(IndexScanDesc scan, ScanKey scankey, int nscankeys,
{
BTScanOpaque so = (BTScanOpaque) scan->opaque;
- /* we aren't holding any read locks, but gotta drop the pins */
- if (BTScanPosIsValid(so->currPos))
- {
- /* Before leaving current page, deal with any killed items */
- if (so->numKilled > 0)
- _bt_killitems(scan);
- BTScanPosUnpinIfPinned(so->currPos);
- BTScanPosInvalidate(so->currPos);
- }
-
- /*
- * We prefer to eagerly drop leaf page pins before btgettuple returns.
- * This avoids making VACUUM wait to acquire a cleanup lock on the page.
- *
- * We cannot safely drop leaf page pins during index-only scans due to a
- * race condition involving VACUUM setting pages all-visible in the VM.
- * It's also unsafe for plain index scans that use a non-MVCC snapshot.
- *
- * When we drop pins eagerly, the mechanism that marks so->killedItems[]
- * index tuples LP_DEAD has to deal with concurrent TID recycling races.
- * The scheme used to detect unsafe TID recycling won't work when scanning
- * unlogged relations (since it involves saving an affected page's LSN).
- * Opt out of eager pin dropping during unlogged relation scans for now
- * (this is preferable to opting out of kill_prior_tuple LP_DEAD setting).
- *
- * Also opt out of dropping leaf page pins eagerly during bitmap scans.
- * Pins cannot be held for more than an instant during bitmap scans either
- * way, so we might as well avoid wasting cycles on acquiring page LSNs.
- *
- * See nbtree/README section on making concurrent TID recycling safe.
- *
- * Note: so->dropPin should never change across rescans.
- */
- so->dropPin = (!scan->xs_want_itup &&
- IsMVCCSnapshot(scan->xs_snapshot) &&
- RelationNeedsWAL(scan->indexRelation) &&
- scan->heapRelation != NULL);
-
- so->markItemIndex = -1;
- so->needPrimScan = false;
- so->scanBehind = false;
- so->oppositeDirCheck = false;
- BTScanPosUnpinIfPinned(so->markPos);
- BTScanPosInvalidate(so->markPos);
-
- /*
- * Allocate tuple workspace arrays, if needed for an index-only scan and
- * not already done in a previous rescan call. To save on palloc
- * overhead, both workspaces are allocated as one palloc block; only this
- * function and btendscan know that.
- */
- if (scan->xs_want_itup && so->currTuples == NULL)
- {
- so->currTuples = (char *) palloc(BLCKSZ * 2);
- so->markTuples = so->currTuples + BLCKSZ;
- }
-
/*
* Reset the scan keys
*/
if (scankey && scan->numberOfKeys > 0)
memcpy(scan->keyData, scankey, scan->numberOfKeys * sizeof(ScanKeyData));
+ so->needPrimScan = false;
+ so->scanBehind = false;
+ so->oppositeDirCheck = false;
so->numberOfKeys = 0; /* until _bt_preprocess_keys sets it */
so->numArrayKeys = 0; /* ditto */
}
+/*
+ * btfreebatch() -- Free batch resources, including its buffer pin
+ */
+void
+btfreebatch(IndexScanDesc scan, BatchIndexScan batch)
+{
+ if (batch->numKilled > 0)
+ _bt_killitems(scan, batch);
+
+ if (!scan->dropPin)
+ {
+ /* indexam_util_batch_unlock didn't unpin page earlier, do it now */
+ ReleaseBuffer(batch->buf);
+ batch->buf = InvalidBuffer;
+ }
+
+ indexam_util_batch_release(scan, batch);
+}
+
/*
* btendscan() -- close down a scan
*/
@@ -462,116 +392,48 @@ btendscan(IndexScanDesc scan)
{
BTScanOpaque so = (BTScanOpaque) scan->opaque;
- /* we aren't holding any read locks, but gotta drop the pins */
- if (BTScanPosIsValid(so->currPos))
- {
- /* Before leaving current page, deal with any killed items */
- if (so->numKilled > 0)
- _bt_killitems(scan);
- BTScanPosUnpinIfPinned(so->currPos);
- }
-
- so->markItemIndex = -1;
- BTScanPosUnpinIfPinned(so->markPos);
-
- /* No need to invalidate positions, the RAM is about to be freed. */
-
/* Release storage */
if (so->keyData != NULL)
pfree(so->keyData);
/* so->arrayKeys and so->orderProcs are in arrayContext */
if (so->arrayContext != NULL)
MemoryContextDelete(so->arrayContext);
- if (so->killedItems != NULL)
- pfree(so->killedItems);
- if (so->currTuples != NULL)
- pfree(so->currTuples);
- /* so->markTuples should not be pfree'd, see btrescan */
pfree(so);
}
/*
- * btmarkpos() -- save current scan position
+ * btposreset() -- invalidate scan's array keys
*/
void
-btmarkpos(IndexScanDesc scan)
+btposreset(IndexScanDesc scan, BatchIndexScan markbatch)
{
BTScanOpaque so = (BTScanOpaque) scan->opaque;
- /* There may be an old mark with a pin (but no lock). */
- BTScanPosUnpinIfPinned(so->markPos);
+ if (!so->numArrayKeys)
+ return;
/*
- * Just record the current itemIndex. If we later step to next page
- * before releasing the marked position, _bt_steppage makes a full copy of
- * the currPos struct in markPos. If (as often happens) the mark is moved
- * before we leave the page, we don't have to do that work.
+ * Core system is about to restore a mark associated with a previously
+ * returned batch. Reset the scan's arrays to make all this safe.
*/
- if (BTScanPosIsValid(so->currPos))
- so->markItemIndex = so->currPos.itemIndex;
+ _bt_start_array_keys(scan, markbatch->dir);
+
+ /*
+ * Core system will invalidate all other batches.
+ *
+ * Deal with this by unsetting needPrimScan as well as moreRight (or as
+ * well as moreLeft, when scanning backwards). That way, the next time
+ * _bt_next is called it will step to the right (or to the left). At that
+ * point _bt_readpage will restore the scan's arrays to elements that
+ * correctly track the next page's position in the index's key space.
+ */
+ if (ScanDirectionIsForward(markbatch->dir))
+ markbatch->moreRight = true;
else
- {
- BTScanPosInvalidate(so->markPos);
- so->markItemIndex = -1;
- }
-}
-
-/*
- * btrestrpos() -- restore scan to last saved position
- */
-void
-btrestrpos(IndexScanDesc scan)
-{
- BTScanOpaque so = (BTScanOpaque) scan->opaque;
-
- if (so->markItemIndex >= 0)
- {
- /*
- * The scan has never moved to a new page since the last mark. Just
- * restore the itemIndex.
- *
- * NB: In this case we can't count on anything in so->markPos to be
- * accurate.
- */
- so->currPos.itemIndex = so->markItemIndex;
- }
- else
- {
- /*
- * The scan moved to a new page after last mark or restore, and we are
- * now restoring to the marked page. We aren't holding any read
- * locks, but if we're still holding the pin for the current position,
- * we must drop it.
- */
- if (BTScanPosIsValid(so->currPos))
- {
- /* Before leaving current page, deal with any killed items */
- if (so->numKilled > 0)
- _bt_killitems(scan);
- BTScanPosUnpinIfPinned(so->currPos);
- }
-
- if (BTScanPosIsValid(so->markPos))
- {
- /* bump pin on mark buffer for assignment to current buffer */
- if (BTScanPosIsPinned(so->markPos))
- IncrBufferRefCount(so->markPos.buf);
- memcpy(&so->currPos, &so->markPos,
- offsetof(BTScanPosData, items[1]) +
- so->markPos.lastItem * sizeof(BTScanPosItem));
- if (so->currTuples)
- memcpy(so->currTuples, so->markTuples,
- so->markPos.nextTupleOffset);
- /* Reset the scan's array keys (see _bt_steppage for why) */
- if (so->numArrayKeys)
- {
- _bt_start_array_keys(scan, so->currPos.dir);
- so->needPrimScan = false;
- }
- }
- else
- BTScanPosInvalidate(so->currPos);
- }
+ markbatch->moreLeft = true;
+ so->needPrimScan = false;
+ so->scanBehind = false;
+ so->oppositeDirCheck = false;
}
/*
@@ -887,15 +749,6 @@ _bt_parallel_seize(IndexScanDesc scan, BlockNumber *next_scan_page,
*next_scan_page = InvalidBlockNumber;
*last_curr_page = InvalidBlockNumber;
- /*
- * Reset so->currPos, and initialize moreLeft/moreRight such that the next
- * call to _bt_readnextpage treats this backend similarly to a serial
- * backend that steps from *last_curr_page to *next_scan_page (unless this
- * backend's so->currPos is initialized by _bt_readfirstpage before then).
- */
- BTScanPosInvalidate(so->currPos);
- so->currPos.moreLeft = so->currPos.moreRight = true;
-
if (first)
{
/*
@@ -1045,8 +898,6 @@ _bt_parallel_done(IndexScanDesc scan)
BTParallelScanDesc btscan;
bool status_changed = false;
- Assert(!BTScanPosIsValid(so->currPos));
-
/* Do nothing, for non-parallel scans */
if (parallel_scan == NULL)
return;
diff --git a/src/backend/access/nbtree/nbtsearch.c b/src/backend/access/nbtree/nbtsearch.c
index 32ae0bda8..ee1c96861 100644
--- a/src/backend/access/nbtree/nbtsearch.c
+++ b/src/backend/access/nbtree/nbtsearch.c
@@ -26,53 +26,23 @@
#include "utils/rel.h"
-static inline void _bt_drop_lock_and_maybe_pin(Relation rel, BTScanOpaque so);
static Buffer _bt_moveright(Relation rel, Relation heaprel, BTScanInsert key,
Buffer buf, bool forupdate, BTStack stack,
int access);
static OffsetNumber _bt_binsrch(Relation rel, BTScanInsert key, Buffer buf);
static int _bt_binsrch_posting(BTScanInsert key, Page page,
OffsetNumber offnum);
-static inline void _bt_returnitem(IndexScanDesc scan, BTScanOpaque so);
-static bool _bt_steppage(IndexScanDesc scan, ScanDirection dir);
-static bool _bt_readfirstpage(IndexScanDesc scan, OffsetNumber offnum,
- ScanDirection dir);
-static bool _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno,
- BlockNumber lastcurrblkno, ScanDirection dir,
- bool seized);
+static BatchIndexScan _bt_readfirstpage(IndexScanDesc scan, BatchIndexScan firstbatch,
+ OffsetNumber offnum, ScanDirection dir);
+static BatchIndexScan _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno,
+ BlockNumber lastcurrblkno,
+ ScanDirection dir, bool firstpage);
static Buffer _bt_lock_and_validate_left(Relation rel, BlockNumber *blkno,
BlockNumber lastcurrblkno);
-static bool _bt_endpoint(IndexScanDesc scan, ScanDirection dir);
+static BatchIndexScan _bt_endpoint(IndexScanDesc scan, ScanDirection dir,
+ BatchIndexScan firstbatch);
-/*
- * _bt_drop_lock_and_maybe_pin()
- *
- * Unlock so->currPos.buf. If scan is so->dropPin, drop the pin, too.
- * Dropping the pin prevents VACUUM from blocking on acquiring a cleanup lock.
- */
-static inline void
-_bt_drop_lock_and_maybe_pin(Relation rel, BTScanOpaque so)
-{
- if (!so->dropPin)
- {
- /* Just drop the lock (not the pin) */
- _bt_unlockbuf(rel, so->currPos.buf);
- return;
- }
-
- /*
- * Drop both the lock and the pin.
- *
- * Have to set so->currPos.lsn so that _bt_killitems has a way to detect
- * when concurrent heap TID recycling by VACUUM might have taken place.
- */
- Assert(RelationNeedsWAL(rel));
- so->currPos.lsn = BufferGetLSNAtomic(so->currPos.buf);
- _bt_relbuf(rel, so->currPos.buf);
- so->currPos.buf = InvalidBuffer;
-}
-
/*
* _bt_search() -- Search the tree for a particular scankey,
* or more precisely for the first leaf page it could be on.
@@ -861,20 +831,16 @@ _bt_compare(Relation rel,
* conditions, and the tree ordering. We find the first item (or,
* if backwards scan, the last item) in the tree that satisfies the
* qualifications in the scan key. On success exit, data about the
- * matching tuple(s) on the page has been loaded into so->currPos. We'll
- * drop all locks and hold onto a pin on page's buffer, except during
- * so->dropPin scans, when we drop both the lock and the pin.
- * _bt_returnitem sets the next item to return to scan on success exit.
+ * matching tuple(s) on the page has been loaded into the returned batch.
*
- * If there are no matching items in the index, we return false, with no
- * pins or locks held. so->currPos will remain invalid.
+ * If there are no matching items in the index, we just return NULL.
*
* Note that scan->keyData[], and the so->keyData[] scankey built from it,
* are both search-type scankeys (see nbtree/README for more about this).
* Within this routine, we build a temporary insertion-type scankey to use
* in locating the scan start position.
*/
-bool
+BatchIndexScan
_bt_first(IndexScanDesc scan, ScanDirection dir)
{
Relation rel = scan->indexRelation;
@@ -888,8 +854,10 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
StrategyNumber strat_total = InvalidStrategy;
BlockNumber blkno = InvalidBlockNumber,
lastcurrblkno;
+ BatchIndexScan firstbatch;
- Assert(!BTScanPosIsValid(so->currPos));
+ /* Allocate space for first batch */
+ firstbatch = indexam_util_batch_alloc(scan);
/*
* Examine the scan keys and eliminate any redundant keys; also mark the
@@ -905,6 +873,7 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
{
Assert(!so->needPrimScan);
_bt_parallel_done(scan);
+ indexam_util_batch_release(scan, firstbatch);
return false;
}
@@ -914,7 +883,10 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
*/
if (scan->parallel_scan != NULL &&
!_bt_parallel_seize(scan, &blkno, &lastcurrblkno, true))
- return false;
+ {
+ indexam_util_batch_release(scan, firstbatch);
+ return false; /* definitely done (so->needPrimScan is unset) */
+ }
/*
* Initialize the scan's arrays (if any) for the current scan direction
@@ -931,14 +903,10 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
* _bt_readnextpage releases the scan for us (not _bt_readfirstpage).
*/
Assert(scan->parallel_scan != NULL);
- Assert(!so->needPrimScan);
- Assert(blkno != P_NONE);
- if (!_bt_readnextpage(scan, blkno, lastcurrblkno, dir, true))
- return false;
+ indexam_util_batch_release(scan, firstbatch);
- _bt_returnitem(scan, so);
- return true;
+ return _bt_readnextpage(scan, blkno, lastcurrblkno, dir, true);
}
/*
@@ -1238,7 +1206,7 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
* Note: calls _bt_readfirstpage for us, which releases the parallel scan.
*/
if (keysz == 0)
- return _bt_endpoint(scan, dir);
+ return _bt_endpoint(scan, dir, firstbatch);
/*
* We want to start the scan somewhere within the index. Set up an
@@ -1506,12 +1474,12 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
* position ourselves on the target leaf page.
*/
Assert(ScanDirectionIsBackward(dir) == inskey.backward);
- stack = _bt_search(rel, NULL, &inskey, &so->currPos.buf, BT_READ);
+ stack = _bt_search(rel, NULL, &inskey, &firstbatch->buf, BT_READ);
/* don't need to keep the stack around... */
_bt_freestack(stack);
- if (!BufferIsValid(so->currPos.buf))
+ if (unlikely(!BufferIsValid(firstbatch->buf)))
{
Assert(!so->needPrimScan);
@@ -1527,23 +1495,24 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
if (IsolationIsSerializable())
{
PredicateLockRelation(rel, scan->xs_snapshot);
- stack = _bt_search(rel, NULL, &inskey, &so->currPos.buf, BT_READ);
+ stack = _bt_search(rel, NULL, &inskey, &firstbatch->buf, BT_READ);
_bt_freestack(stack);
}
- if (!BufferIsValid(so->currPos.buf))
+ if (!BufferIsValid(firstbatch->buf))
{
_bt_parallel_done(scan);
+ indexam_util_batch_release(scan, firstbatch);
return false;
}
}
/* position to the precise item on the page */
- offnum = _bt_binsrch(rel, &inskey, so->currPos.buf);
+ offnum = _bt_binsrch(rel, &inskey, firstbatch->buf);
/*
* Now load data from the first page of the scan (usually the page
- * currently in so->currPos.buf).
+ * currently in firstbatch.buf).
*
* If inskey.nextkey = false and inskey.backward = false, offnum is
* positioned at the first non-pivot tuple >= inskey.scankeys.
@@ -1561,168 +1530,69 @@ _bt_first(IndexScanDesc scan, ScanDirection dir)
* for the page. For example, when inskey is both < the leaf page's high
* key and > all of its non-pivot tuples, offnum will be "maxoff + 1".
*/
- if (!_bt_readfirstpage(scan, offnum, dir))
- return false;
-
- _bt_returnitem(scan, so);
- return true;
+ return _bt_readfirstpage(scan, firstbatch, offnum, dir);
}
/*
* _bt_next() -- Get the next item in a scan.
*
- * On entry, so->currPos describes the current page, which may be pinned
- * but is not locked, and so->currPos.itemIndex identifies which item was
- * previously returned.
+ * On entry, priorbatch describes the batch that was last returned by
+ * btgetbatch. We'll use the prior batch's positioning information to
+ * decide which page to read next.
*
- * On success exit, so->currPos is updated as needed, and _bt_returnitem
- * sets the next item to return to the scan. so->currPos remains valid.
+ * On success exit, returns the next batch. There must be at least one
+ * matching tuple on any returned batch (else we'd just return NULL).
*
- * On failure exit (no more tuples), we invalidate so->currPos. It'll
- * still be possible for the scan to return tuples by changing direction,
- * though we'll need to call _bt_first anew in that other direction.
+ * On failure exit (no more tuples), we return NULL. It'll still be
+ * possible for the scan to return tuples by changing direction, though
+ * we'll need to call _bt_first anew in that other direction.
*/
-bool
-_bt_next(IndexScanDesc scan, ScanDirection dir)
-{
- BTScanOpaque so = (BTScanOpaque) scan->opaque;
-
- Assert(BTScanPosIsValid(so->currPos));
-
- /*
- * Advance to next tuple on current page; or if there's no more, try to
- * step to the next page with data.
- */
- if (ScanDirectionIsForward(dir))
- {
- if (++so->currPos.itemIndex > so->currPos.lastItem)
- {
- if (!_bt_steppage(scan, dir))
- return false;
- }
- }
- else
- {
- if (--so->currPos.itemIndex < so->currPos.firstItem)
- {
- if (!_bt_steppage(scan, dir))
- return false;
- }
- }
-
- _bt_returnitem(scan, so);
- return true;
-}
-
-/*
- * Return the index item from so->currPos.items[so->currPos.itemIndex] to the
- * index scan by setting the relevant fields in caller's index scan descriptor
- */
-static inline void
-_bt_returnitem(IndexScanDesc scan, BTScanOpaque so)
-{
- BTScanPosItem *currItem = &so->currPos.items[so->currPos.itemIndex];
-
- /* Most recent _bt_readpage must have succeeded */
- Assert(BTScanPosIsValid(so->currPos));
- Assert(so->currPos.itemIndex >= so->currPos.firstItem);
- Assert(so->currPos.itemIndex <= so->currPos.lastItem);
-
- /* Return next item, per amgettuple contract */
- scan->xs_heaptid = currItem->heapTid;
- if (so->currTuples)
- scan->xs_itup = (IndexTuple) (so->currTuples + currItem->tupleOffset);
-}
-
-/*
- * _bt_steppage() -- Step to next page containing valid data for scan
- *
- * Wrapper on _bt_readnextpage that performs final steps for the current page.
- *
- * On entry, so->currPos must be valid. Its buffer will be pinned, though
- * never locked. (Actually, when so->dropPin there won't even be a pin held,
- * though so->currPos.currPage must still be set to a valid block number.)
- */
-static bool
-_bt_steppage(IndexScanDesc scan, ScanDirection dir)
+BatchIndexScan
+_bt_next(IndexScanDesc scan, ScanDirection dir, BatchIndexScan priorbatch)
{
BTScanOpaque so = (BTScanOpaque) scan->opaque;
BlockNumber blkno,
lastcurrblkno;
- Assert(BTScanPosIsValid(so->currPos));
-
- /* Before leaving current page, deal with any killed items */
- if (so->numKilled > 0)
- _bt_killitems(scan);
-
- /*
- * Before we modify currPos, make a copy of the page data if there was a
- * mark position that needs it.
- */
- if (so->markItemIndex >= 0)
- {
- /* bump pin on current buffer for assignment to mark buffer */
- if (BTScanPosIsPinned(so->currPos))
- IncrBufferRefCount(so->currPos.buf);
- memcpy(&so->markPos, &so->currPos,
- offsetof(BTScanPosData, items[1]) +
- so->currPos.lastItem * sizeof(BTScanPosItem));
- if (so->markTuples)
- memcpy(so->markTuples, so->currTuples,
- so->currPos.nextTupleOffset);
- so->markPos.itemIndex = so->markItemIndex;
- so->markItemIndex = -1;
-
- /*
- * If we're just about to start the next primitive index scan
- * (possible with a scan that has arrays keys, and needs to skip to
- * continue in the current scan direction), moreLeft/moreRight only
- * indicate the end of the current primitive index scan. They must
- * never be taken to indicate that the top-level index scan has ended
- * (that would be wrong).
- *
- * We could handle this case by treating the current array keys as
- * markPos state. But depending on the current array state like this
- * would add complexity. Instead, we just unset markPos's copy of
- * moreRight or moreLeft (whichever might be affected), while making
- * btrestrpos reset the scan's arrays to their initial scan positions.
- * In effect, btrestrpos leaves advancing the arrays up to the first
- * _bt_readpage call (that takes place after it has restored markPos).
- */
- if (so->needPrimScan)
- {
- if (ScanDirectionIsForward(so->currPos.dir))
- so->markPos.moreRight = true;
- else
- so->markPos.moreLeft = true;
- }
-
- /* mark/restore not supported by parallel scans */
- Assert(!scan->parallel_scan);
- }
-
- BTScanPosUnpinIfPinned(so->currPos);
+ Assert(BlockNumberIsValid(priorbatch->currPage));
/* Walk to the next page with data */
if (ScanDirectionIsForward(dir))
- blkno = so->currPos.nextPage;
+ blkno = priorbatch->nextPage;
else
- blkno = so->currPos.prevPage;
- lastcurrblkno = so->currPos.currPage;
+ blkno = priorbatch->prevPage;
+ lastcurrblkno = priorbatch->currPage;
/*
- * Cancel primitive index scans that were scheduled when the call to
- * _bt_readpage for currPos happened to use the opposite direction to the
- * one that we're stepping in now. (It's okay to leave the scan's array
- * keys as-is, since the next _bt_readpage will advance them.)
+ * Cancel primitive index scans that were scheduled when priorbatch's call
+ * to _bt_readpage happened to use the opposite direction to the one that
+ * we're stepping in now. (It's okay to leave the scan's array keys
+ * as-is, since the next _bt_readpage will advance them.)
*/
- if (so->currPos.dir != dir)
+ if (priorbatch->dir != dir)
so->needPrimScan = false;
+ if (blkno == P_NONE ||
+ (ScanDirectionIsForward(dir) ?
+ !priorbatch->moreRight : !priorbatch->moreLeft))
+ {
+ /*
+ * priorbatch _bt_readpage call ended scan in this direction (though
+ * if so->needPrimScan was set the scan will continue in _bt_first)
+ */
+ _bt_parallel_done(scan);
+ return NULL;
+ }
+
+ /* parallel scan must seize the scan to get next blkno */
+ if (scan->parallel_scan != NULL &&
+ !_bt_parallel_seize(scan, &blkno, &lastcurrblkno, false))
+ return NULL; /* done iff so->needPrimScan wasn't set */
+
return _bt_readnextpage(scan, blkno, lastcurrblkno, dir, false);
}
+
/*
* _bt_readfirstpage() -- Read first page containing valid data for _bt_first
*
@@ -1732,73 +1602,90 @@ _bt_steppage(IndexScanDesc scan, ScanDirection dir)
* to stop the scan on this page by calling _bt_checkkeys against the high
* key. See _bt_readpage for full details.
*
- * On entry, so->currPos must be pinned and locked (so offnum stays valid).
+ * On entry, firstbatch must be pinned and locked (so offnum stays valid).
* Parallel scan callers must have seized the scan before calling here.
*
- * On exit, we'll have updated so->currPos and retained locks and pins
+ * On exit, we'll have updated firstbatch and retained locks and pins
* according to the same rules as those laid out for _bt_readnextpage exit.
- * Like _bt_readnextpage, our return value indicates if there are any matching
- * records in the given direction.
*
* We always release the scan for a parallel scan caller, regardless of
* success or failure; we'll call _bt_parallel_release as soon as possible.
*/
-static bool
-_bt_readfirstpage(IndexScanDesc scan, OffsetNumber offnum, ScanDirection dir)
+static BatchIndexScan
+_bt_readfirstpage(IndexScanDesc scan, BatchIndexScan firstbatch,
+ OffsetNumber offnum, ScanDirection dir)
{
BTScanOpaque so = (BTScanOpaque) scan->opaque;
+ BlockNumber blkno,
+ lastcurrblkno;
- so->numKilled = 0; /* just paranoia */
- so->markItemIndex = -1; /* ditto */
-
- /* Initialize so->currPos for the first page (page in so->currPos.buf) */
+ /* Initialize firstbatch's position for the first page */
if (so->needPrimScan)
{
Assert(so->numArrayKeys);
- so->currPos.moreLeft = true;
- so->currPos.moreRight = true;
+ firstbatch->moreLeft = true;
+ firstbatch->moreRight = true;
so->needPrimScan = false;
}
else if (ScanDirectionIsForward(dir))
{
- so->currPos.moreLeft = false;
- so->currPos.moreRight = true;
+ firstbatch->moreLeft = false;
+ firstbatch->moreRight = true;
}
else
{
- so->currPos.moreLeft = true;
- so->currPos.moreRight = false;
+ firstbatch->moreLeft = true;
+ firstbatch->moreRight = false;
}
/*
* Attempt to load matching tuples from the first page.
*
- * Note that _bt_readpage will finish initializing the so->currPos fields.
+ * Note that _bt_readpage will finish initializing the firstbatch fields.
* _bt_readpage also releases parallel scan (even when it returns false).
*/
- if (_bt_readpage(scan, dir, offnum, true))
+ if (_bt_readpage(scan, firstbatch, dir, offnum, true))
{
- Relation rel = scan->indexRelation;
-
- /*
- * _bt_readpage succeeded. Drop the lock (and maybe the pin) on
- * so->currPos.buf in preparation for btgettuple returning tuples.
- */
- Assert(BTScanPosIsPinned(so->currPos));
- _bt_drop_lock_and_maybe_pin(rel, so);
- return true;
+ /* _bt_readpage saved one or more matches in firstbatch.items[] */
+ indexam_util_batch_unlock(scan, firstbatch);
+ return firstbatch;
}
- /* There's no actually-matching data on the page in so->currPos.buf */
- _bt_unlockbuf(scan->indexRelation, so->currPos.buf);
+ /* There's no actually-matching data on the page */
+ _bt_relbuf(scan->indexRelation, firstbatch->buf);
+ firstbatch->buf = InvalidBuffer;
- /* Call _bt_readnextpage using its _bt_steppage wrapper function */
- if (!_bt_steppage(scan, dir))
- return false;
+ /* Walk to the next page with data */
+ if (ScanDirectionIsForward(dir))
+ blkno = firstbatch->nextPage;
+ else
+ blkno = firstbatch->prevPage;
+ lastcurrblkno = firstbatch->currPage;
- /* _bt_readpage for a later page (now in so->currPos) succeeded */
- return true;
+ Assert(firstbatch->dir == dir);
+
+ if (blkno == P_NONE ||
+ (ScanDirectionIsForward(dir) ?
+ !firstbatch->moreRight : !firstbatch->moreLeft))
+ {
+ /*
+ * firstbatch _bt_readpage call ended scan in this direction (though
+ * if so->needPrimScan was set the scan will continue in _bt_first)
+ */
+ indexam_util_batch_release(scan, firstbatch);
+ _bt_parallel_done(scan);
+ return NULL;
+ }
+
+ indexam_util_batch_release(scan, firstbatch);
+
+ /* parallel scan must seize the scan to get next blkno */
+ if (scan->parallel_scan != NULL &&
+ !_bt_parallel_seize(scan, &blkno, &lastcurrblkno, false))
+ return NULL; /* done iff so->needPrimScan wasn't set */
+
+ return _bt_readnextpage(scan, blkno, lastcurrblkno, dir, false);
}
/*
@@ -1808,102 +1695,65 @@ _bt_readfirstpage(IndexScanDesc scan, OffsetNumber offnum, ScanDirection dir)
* previously-saved right link or left link. lastcurrblkno is the page that
* was current at the point where the blkno link was saved, which we use to
* reason about concurrent page splits/page deletions during backwards scans.
- * In the common case where seized=false, blkno is either so->currPos.nextPage
- * or so->currPos.prevPage, and lastcurrblkno is so->currPos.currPage.
+ * blkno is the prior batch's nextPage or prevPage (depending on the current
+ * scan direction), and lastcurrblkno is the prior batch's currPage.
*
- * On entry, so->currPos shouldn't be locked by caller. so->currPos.buf must
- * be InvalidBuffer/unpinned as needed by caller (note that lastcurrblkno
- * won't need to be read again in almost all cases). Parallel scan callers
- * that seized the scan before calling here should pass seized=true; such a
- * caller's blkno and lastcurrblkno arguments come from the seized scan.
- * seized=false callers just pass us the blkno/lastcurrblkno taken from their
- * so->currPos, which (along with so->currPos itself) can be used to end the
- * scan. A seized=false caller's blkno can never be assumed to be the page
- * that must be read next during a parallel scan, though. We must figure that
- * part out for ourselves by seizing the scan (the correct page to read might
- * already be beyond the seized=false caller's blkno during a parallel scan,
- * unless blkno/so->currPos.nextPage/so->currPos.prevPage is already P_NONE,
- * or unless so->currPos.moreRight/so->currPos.moreLeft is already unset).
+ * On entry, no page should be locked by caller.
*
- * On success exit, so->currPos is updated to contain data from the next
- * interesting page, and we return true. We hold a pin on the buffer on
- * success exit (except during so->dropPin index scans, when we drop the pin
- * eagerly to avoid blocking VACUUM).
+ * On success exit, returns batch containing data from the next page that has
+ * at least one matching item. If there are no more matching items in the
+ * given scan direction, we just return NULL.
*
- * If there are no more matching records in the given direction, we invalidate
- * so->currPos (while ensuring it retains no locks or pins), and return false.
- *
- * We always release the scan for a parallel scan caller, regardless of
- * success or failure; we'll call _bt_parallel_release as soon as possible.
+ * Parallel scan callers must seize the scan before calling here. blkno and
+ * lastcurrblkno should come from the seized scan. We'll release the scan as
+ * soon as possible.
*/
-static bool
+static BatchIndexScan
_bt_readnextpage(IndexScanDesc scan, BlockNumber blkno,
- BlockNumber lastcurrblkno, ScanDirection dir, bool seized)
+ BlockNumber lastcurrblkno, ScanDirection dir, bool firstpage)
{
Relation rel = scan->indexRelation;
- BTScanOpaque so = (BTScanOpaque) scan->opaque;
+ BatchIndexScan newbatch;
- Assert(so->currPos.currPage == lastcurrblkno || seized);
- Assert(!(blkno == P_NONE && seized));
- Assert(!BTScanPosIsPinned(so->currPos));
+ /* Allocate space for next batch */
+ newbatch = indexam_util_batch_alloc(scan);
/*
- * Remember that the scan already read lastcurrblkno, a page to the left
- * of blkno (or remember reading a page to the right, for backwards scans)
+ * newbatch will be the batch for lastcurrblkno, a page to the left of
+ * blkno (or to the right, when the scan is moving backwards)
*/
- if (ScanDirectionIsForward(dir))
- so->currPos.moreLeft = true;
- else
- so->currPos.moreRight = true;
+ newbatch->moreLeft = true;
+ newbatch->moreRight = true;
for (;;)
{
Page page;
BTPageOpaque opaque;
- if (blkno == P_NONE ||
- (ScanDirectionIsForward(dir) ?
- !so->currPos.moreRight : !so->currPos.moreLeft))
- {
- /* most recent _bt_readpage call (for lastcurrblkno) ended scan */
- Assert(so->currPos.currPage == lastcurrblkno && !seized);
- BTScanPosInvalidate(so->currPos);
- _bt_parallel_done(scan); /* iff !so->needPrimScan */
- return false;
- }
-
- Assert(!so->needPrimScan);
-
- /* parallel scan must never actually visit so->currPos blkno */
- if (!seized && scan->parallel_scan != NULL &&
- !_bt_parallel_seize(scan, &blkno, &lastcurrblkno, false))
- {
- /* whole scan is now done (or another primitive scan required) */
- BTScanPosInvalidate(so->currPos);
- return false;
- }
+ Assert(!((BTScanOpaque) scan->opaque)->needPrimScan);
+ Assert(blkno != P_NONE && lastcurrblkno != P_NONE);
if (ScanDirectionIsForward(dir))
{
/* read blkno, but check for interrupts first */
CHECK_FOR_INTERRUPTS();
- so->currPos.buf = _bt_getbuf(rel, blkno, BT_READ);
+ newbatch->buf = _bt_getbuf(rel, blkno, BT_READ);
}
else
{
/* read blkno, avoiding race (also checks for interrupts) */
- so->currPos.buf = _bt_lock_and_validate_left(rel, &blkno,
- lastcurrblkno);
- if (so->currPos.buf == InvalidBuffer)
+ newbatch->buf = _bt_lock_and_validate_left(rel, &blkno,
+ lastcurrblkno);
+ if (newbatch->buf == InvalidBuffer)
{
/* must have been a concurrent deletion of leftmost page */
- BTScanPosInvalidate(so->currPos);
_bt_parallel_done(scan);
- return false;
+ indexam_util_batch_release(scan, newbatch);
+ return NULL;
}
}
- page = BufferGetPage(so->currPos.buf);
+ page = BufferGetPage(newbatch->buf);
opaque = BTPageGetOpaque(page);
lastcurrblkno = blkno;
if (likely(!P_IGNORE(opaque)))
@@ -1911,17 +1761,17 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno,
/* see if there are any matches on this page */
if (ScanDirectionIsForward(dir))
{
- /* note that this will clear moreRight if we can stop */
- if (_bt_readpage(scan, dir, P_FIRSTDATAKEY(opaque), seized))
+ if (_bt_readpage(scan, newbatch, dir,
+ P_FIRSTDATAKEY(opaque), firstpage))
break;
- blkno = so->currPos.nextPage;
+ blkno = newbatch->nextPage;
}
else
{
- /* note that this will clear moreLeft if we can stop */
- if (_bt_readpage(scan, dir, PageGetMaxOffsetNumber(page), seized))
+ if (_bt_readpage(scan, newbatch, dir,
+ PageGetMaxOffsetNumber(page), firstpage))
break;
- blkno = so->currPos.prevPage;
+ blkno = newbatch->prevPage;
}
}
else
@@ -1936,19 +1786,39 @@ _bt_readnextpage(IndexScanDesc scan, BlockNumber blkno,
}
/* no matching tuples on this page */
- _bt_relbuf(rel, so->currPos.buf);
- seized = false; /* released by _bt_readpage (or by us) */
+ _bt_relbuf(rel, newbatch->buf);
+ newbatch->buf = InvalidBuffer;
+
+ /* Continue the scan in this direction? */
+ if (blkno == P_NONE ||
+ (ScanDirectionIsForward(dir) ?
+ !newbatch->moreRight : !newbatch->moreLeft))
+ {
+ /*
+ * blkno _bt_readpage call ended scan in this direction (though if
+ * so->needPrimScan was set the scan will continue in _bt_first)
+ */
+ _bt_parallel_done(scan);
+ indexam_util_batch_release(scan, newbatch);
+ return NULL;
+ }
+
+ /* parallel scan must seize the scan to get next blkno */
+ if (scan->parallel_scan != NULL &&
+ !_bt_parallel_seize(scan, &blkno, &lastcurrblkno, false))
+ {
+ indexam_util_batch_release(scan, newbatch);
+ return NULL; /* done iff so->needPrimScan wasn't set */
+ }
+
+ firstpage = false; /* next page cannot be first */
}
- /*
- * _bt_readpage succeeded. Drop the lock (and maybe the pin) on
- * so->currPos.buf in preparation for btgettuple returning tuples.
- */
- Assert(so->currPos.currPage == blkno);
- Assert(BTScanPosIsPinned(so->currPos));
- _bt_drop_lock_and_maybe_pin(rel, so);
+ /* _bt_readpage saved one or more matches in newbatch.items[] */
+ Assert(newbatch->currPage == blkno);
+ indexam_util_batch_unlock(scan, newbatch);
- return true;
+ return newbatch;
}
/*
@@ -2174,25 +2044,23 @@ _bt_get_endpoint(Relation rel, uint32 level, bool rightmost)
* Parallel scan callers must have seized the scan before calling here.
* Exit conditions are the same as for _bt_first().
*/
-static bool
-_bt_endpoint(IndexScanDesc scan, ScanDirection dir)
+static BatchIndexScan
+_bt_endpoint(IndexScanDesc scan, ScanDirection dir, BatchIndexScan firstbatch)
{
Relation rel = scan->indexRelation;
- BTScanOpaque so = (BTScanOpaque) scan->opaque;
Page page;
BTPageOpaque opaque;
OffsetNumber start;
- Assert(!BTScanPosIsValid(so->currPos));
- Assert(!so->needPrimScan);
+ Assert(!((BTScanOpaque) scan->opaque)->needPrimScan);
/*
* Scan down to the leftmost or rightmost leaf page. This is a simplified
* version of _bt_search().
*/
- so->currPos.buf = _bt_get_endpoint(rel, 0, ScanDirectionIsBackward(dir));
+ firstbatch->buf = _bt_get_endpoint(rel, 0, ScanDirectionIsBackward(dir));
- if (!BufferIsValid(so->currPos.buf))
+ if (!BufferIsValid(firstbatch->buf))
{
/*
* Empty index. Lock the whole relation, as nothing finer to lock
@@ -2203,7 +2071,7 @@ _bt_endpoint(IndexScanDesc scan, ScanDirection dir)
return false;
}
- page = BufferGetPage(so->currPos.buf);
+ page = BufferGetPage(firstbatch->buf);
opaque = BTPageGetOpaque(page);
Assert(P_ISLEAF(opaque));
@@ -2229,9 +2097,5 @@ _bt_endpoint(IndexScanDesc scan, ScanDirection dir)
/*
* Now load data from the first page of the scan.
*/
- if (!_bt_readfirstpage(scan, start, dir))
- return false;
-
- _bt_returnitem(scan, so);
- return true;
+ return _bt_readfirstpage(scan, firstbatch, start, dir);
}
diff --git a/src/backend/access/nbtree/nbtutils.c b/src/backend/access/nbtree/nbtutils.c
index 5c50f0dd1..b505d7876 100644
--- a/src/backend/access/nbtree/nbtutils.c
+++ b/src/backend/access/nbtree/nbtutils.c
@@ -21,15 +21,12 @@
#include "access/reloptions.h"
#include "access/relscan.h"
#include "commands/progress.h"
-#include "common/int.h"
-#include "lib/qunique.h"
#include "miscadmin.h"
#include "utils/datum.h"
#include "utils/lsyscache.h"
#include "utils/rel.h"
-static int _bt_compare_int(const void *va, const void *vb);
static int _bt_keep_natts(Relation rel, IndexTuple lastleft,
IndexTuple firstright, BTScanInsert itup_key);
@@ -160,104 +157,69 @@ _bt_freestack(BTStack stack)
}
}
-/*
- * qsort comparison function for int arrays
- */
-static int
-_bt_compare_int(const void *va, const void *vb)
-{
- int a = *((const int *) va);
- int b = *((const int *) vb);
-
- return pg_cmp_s32(a, b);
-}
-
/*
* _bt_killitems - set LP_DEAD state for items an indexscan caller has
* told us were killed
*
- * scan->opaque, referenced locally through so, contains information about the
- * current page and killed tuples thereon (generally, this should only be
- * called if so->numKilled > 0).
+ * The batch parameter contains information about the current page and killed
+ * tuples thereon (this should only be called if batch->numKilled > 0).
*
- * Caller should not have a lock on the so->currPos page, but must hold a
- * buffer pin when !so->dropPin. When we return, it still won't be locked.
- * It'll continue to hold whatever pins were held before calling here.
+ * Caller should not have a lock on the batch position's page, but must hold a
+ * buffer pin when !dropPin. When we return, it still won't be locked. It'll
+ * continue to hold whatever pins were held before calling here.
*
* We match items by heap TID before assuming they are the right ones to set
* LP_DEAD. If the scan is one that holds a buffer pin on the target page
* continuously from initially reading the items until applying this function
- * (if it is a !so->dropPin scan), VACUUM cannot have deleted any items on the
+ * (if it is a !dropPin scan), VACUUM cannot have deleted any items on the
* page, so the page's TIDs can't have been recycled by now. There's no risk
* that we'll confuse a new index tuple that happens to use a recycled TID
* with a now-removed tuple with the same TID (that used to be on this same
* page). We can't rely on that during scans that drop buffer pins eagerly
- * (so->dropPin scans), though, so we must condition setting LP_DEAD bits on
+ * (i.e. dropPin scans), though, so we must condition setting LP_DEAD bits on
* the page LSN having not changed since back when _bt_readpage saw the page.
* We totally give up on setting LP_DEAD bits when the page LSN changed.
*
- * We give up much less often during !so->dropPin scans, but it still happens.
+ * We tend to give up less often during !dropPin scans, but it still happens.
* We cope with cases where items have moved right due to insertions. If an
* item has moved off the current page due to a split, we'll fail to find it
* and just give up on it.
*/
void
-_bt_killitems(IndexScanDesc scan)
+_bt_killitems(IndexScanDesc scan, BatchIndexScan batch)
{
Relation rel = scan->indexRelation;
- BTScanOpaque so = (BTScanOpaque) scan->opaque;
Page page;
BTPageOpaque opaque;
OffsetNumber minoff;
OffsetNumber maxoff;
- int numKilled = so->numKilled;
bool killedsomething = false;
Buffer buf;
- Assert(numKilled > 0);
- Assert(BTScanPosIsValid(so->currPos));
+ Assert(batch->numKilled > 0);
+ Assert(BlockNumberIsValid(batch->currPage));
Assert(scan->heapRelation != NULL); /* can't be a bitmap index scan */
- /* Always invalidate so->killedItems[] before leaving so->currPos */
- so->numKilled = 0;
-
- /*
- * We need to iterate through so->killedItems[] in leaf page order; the
- * loop below expects this (when marking posting list tuples, at least).
- * so->killedItems[] is now in whatever order the scan returned items in.
- * Scrollable cursor scans might have even saved the same item/TID twice.
- *
- * Sort and unique-ify so->killedItems[] to deal with all this.
- */
- if (numKilled > 1)
- {
- qsort(so->killedItems, numKilled, sizeof(int), _bt_compare_int);
- numKilled = qunique(so->killedItems, numKilled, sizeof(int),
- _bt_compare_int);
- }
-
- if (!so->dropPin)
+ if (!scan->dropPin)
{
/*
* We have held the pin on this page since we read the index tuples,
* so all we need to do is lock it. The pin will have prevented
* concurrent VACUUMs from recycling any of the TIDs on the page.
*/
- Assert(BTScanPosIsPinned(so->currPos));
- buf = so->currPos.buf;
+ buf = batch->buf;
_bt_lockbuf(rel, buf, BT_READ);
}
else
{
XLogRecPtr latestlsn;
- Assert(!BTScanPosIsPinned(so->currPos));
Assert(RelationNeedsWAL(rel));
- buf = _bt_getbuf(rel, so->currPos.currPage, BT_READ);
+ buf = _bt_getbuf(rel, batch->currPage, BT_READ);
latestlsn = BufferGetLSNAtomic(buf);
- Assert(so->currPos.lsn <= latestlsn);
- if (so->currPos.lsn != latestlsn)
+ Assert(batch->lsn <= latestlsn);
+ if (batch->lsn != latestlsn)
{
/* Modified, give up on hinting */
_bt_relbuf(rel, buf);
@@ -272,17 +234,16 @@ _bt_killitems(IndexScanDesc scan)
minoff = P_FIRSTDATAKEY(opaque);
maxoff = PageGetMaxOffsetNumber(page);
- /* Iterate through so->killedItems[] in leaf page order */
- for (int i = 0; i < numKilled; i++)
+ /* Iterate through batch->killedItems[] in leaf page order */
+ for (int i = 0; i < batch->numKilled; i++)
{
- int itemIndex = so->killedItems[i];
- BTScanPosItem *kitem = &so->currPos.items[itemIndex];
+ int itemIndex = batch->killedItems[i];
+ BatchMatchingItem *kitem = &batch->items[itemIndex];
OffsetNumber offnum = kitem->indexOffset;
- Assert(itemIndex >= so->currPos.firstItem &&
- itemIndex <= so->currPos.lastItem);
+ Assert(itemIndex >= batch->firstItem && itemIndex <= batch->lastItem);
Assert(i == 0 ||
- offnum >= so->currPos.items[so->killedItems[i - 1]].indexOffset);
+ offnum >= batch->items[batch->killedItems[i - 1]].indexOffset);
if (offnum < minoff)
continue; /* pure paranoia */
@@ -300,7 +261,7 @@ _bt_killitems(IndexScanDesc scan)
/*
* Note that the page may have been modified in almost any way
- * since we first read it (in the !so->dropPin case), so it's
+ * since we first read it (in the !dropPin case), so it's
* possible that this posting list tuple wasn't a posting list
* tuple when we first encountered its heap TIDs.
*/
@@ -316,7 +277,7 @@ _bt_killitems(IndexScanDesc scan)
* though only in the common case where the page can't
* have been concurrently modified
*/
- Assert(kitem->indexOffset == offnum || !so->dropPin);
+ Assert(kitem->indexOffset == offnum || !scan->dropPin);
/*
* Read-ahead to later kitems here.
@@ -332,8 +293,8 @@ _bt_killitems(IndexScanDesc scan)
* kitem is also the last heap TID in the last index tuple
* correctly -- posting tuple still gets killed).
*/
- if (pi < numKilled)
- kitem = &so->currPos.items[so->killedItems[pi++]];
+ if (pi < batch->numKilled)
+ kitem = &batch->items[batch->killedItems[pi++]];
}
/*
@@ -383,7 +344,7 @@ _bt_killitems(IndexScanDesc scan)
MarkBufferDirtyHint(buf, true);
}
- if (!so->dropPin)
+ if (!scan->dropPin)
_bt_unlockbuf(rel, buf);
else
_bt_relbuf(rel, buf);
diff --git a/src/backend/access/spgist/spgutils.c b/src/backend/access/spgist/spgutils.c
index 9f5379b87..a18a2fa9e 100644
--- a/src/backend/access/spgist/spgutils.c
+++ b/src/backend/access/spgist/spgutils.c
@@ -88,10 +88,11 @@ spghandler(PG_FUNCTION_ARGS)
.ambeginscan = spgbeginscan,
.amrescan = spgrescan,
.amgettuple = spggettuple,
+ .amgetbatch = NULL,
+ .amfreebatch = NULL,
.amgetbitmap = spggetbitmap,
.amendscan = spgendscan,
- .ammarkpos = NULL,
- .amrestrpos = NULL,
+ .amposreset = NULL,
.amestimateparallelscan = NULL,
.aminitparallelscan = NULL,
.amparallelrescan = NULL,
diff --git a/src/backend/commands/explain.c b/src/backend/commands/explain.c
index b7bb11168..ea0add1c9 100644
--- a/src/backend/commands/explain.c
+++ b/src/backend/commands/explain.c
@@ -135,7 +135,7 @@ static void show_recursive_union_info(RecursiveUnionState *rstate,
static void show_memoize_info(MemoizeState *mstate, List *ancestors,
ExplainState *es);
static void show_hashagg_info(AggState *aggstate, ExplainState *es);
-static void show_indexsearches_info(PlanState *planstate, ExplainState *es);
+static void show_indexscan_info(PlanState *planstate, ExplainState *es);
static void show_tidbitmap_info(BitmapHeapScanState *planstate,
ExplainState *es);
static void show_instrumentation_count(const char *qlabel, int which,
@@ -1972,7 +1972,7 @@ ExplainNode(PlanState *planstate, List *ancestors,
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1,
planstate, es);
- show_indexsearches_info(planstate, es);
+ show_indexscan_info(planstate, es);
break;
case T_IndexOnlyScan:
show_scan_qual(((IndexOnlyScan *) plan)->indexqual,
@@ -1986,15 +1986,12 @@ ExplainNode(PlanState *planstate, List *ancestors,
if (plan->qual)
show_instrumentation_count("Rows Removed by Filter", 1,
planstate, es);
- if (es->analyze)
- ExplainPropertyFloat("Heap Fetches", NULL,
- planstate->instrument->ntuples2, 0, es);
- show_indexsearches_info(planstate, es);
+ show_indexscan_info(planstate, es);
break;
case T_BitmapIndexScan:
show_scan_qual(((BitmapIndexScan *) plan)->indexqualorig,
"Index Cond", planstate, ancestors, es);
- show_indexsearches_info(planstate, es);
+ show_indexscan_info(planstate, es);
break;
case T_BitmapHeapScan:
show_scan_qual(((BitmapHeapScan *) plan)->bitmapqualorig,
@@ -3858,15 +3855,16 @@ show_hashagg_info(AggState *aggstate, ExplainState *es)
}
/*
- * Show the total number of index searches for a
+ * Show index scan related executor instrumentation for a
* IndexScan/IndexOnlyScan/BitmapIndexScan node
*/
static void
-show_indexsearches_info(PlanState *planstate, ExplainState *es)
+show_indexscan_info(PlanState *planstate, ExplainState *es)
{
Plan *plan = planstate->plan;
SharedIndexScanInstrumentation *SharedInfo = NULL;
- uint64 nsearches = 0;
+ uint64 nsearches = 0,
+ nheapfetches = 0;
if (!es->analyze)
return;
@@ -3887,6 +3885,7 @@ show_indexsearches_info(PlanState *planstate, ExplainState *es)
IndexOnlyScanState *indexstate = ((IndexOnlyScanState *) planstate);
nsearches = indexstate->ioss_Instrument.nsearches;
+ nheapfetches = indexstate->ioss_Instrument.nheapfetches;
SharedInfo = indexstate->ioss_SharedInfo;
break;
}
@@ -3910,9 +3909,13 @@ show_indexsearches_info(PlanState *planstate, ExplainState *es)
IndexScanInstrumentation *winstrument = &SharedInfo->winstrument[i];
nsearches += winstrument->nsearches;
+ nheapfetches += winstrument->nheapfetches;
}
}
+ if (nodeTag(plan) == T_IndexOnlyScan)
+ ExplainPropertyUInteger("Heap Fetches", NULL, nheapfetches, es);
+
ExplainPropertyUInteger("Index Searches", NULL, nsearches, es);
}
diff --git a/src/backend/commands/indexcmds.c b/src/backend/commands/indexcmds.c
index 635679cc1..54c2403da 100644
--- a/src/backend/commands/indexcmds.c
+++ b/src/backend/commands/indexcmds.c
@@ -884,7 +884,7 @@ DefineIndex(ParseState *pstate,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("access method \"%s\" does not support multicolumn indexes",
accessMethodName)));
- if (exclusion && amRoutine->amgettuple == NULL)
+ if (exclusion && amRoutine->amgettuple == NULL && amRoutine->amgetbatch == NULL)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("access method \"%s\" does not support exclusion constraints",
diff --git a/src/backend/executor/execAmi.c b/src/backend/executor/execAmi.c
index 90a68c0d1..0dfb01337 100644
--- a/src/backend/executor/execAmi.c
+++ b/src/backend/executor/execAmi.c
@@ -428,7 +428,7 @@ ExecSupportsMarkRestore(Path *pathnode)
case T_IndexOnlyScan:
/*
- * Not all index types support mark/restore.
+ * Not all index types support restoring a mark
*/
return castNode(IndexPath, pathnode)->indexinfo->amcanmarkpos;
diff --git a/src/backend/executor/execIndexing.c b/src/backend/executor/execIndexing.c
index 6ae0f9595..3041d5c72 100644
--- a/src/backend/executor/execIndexing.c
+++ b/src/backend/executor/execIndexing.c
@@ -816,10 +816,12 @@ check_exclusion_or_unique_constraint(Relation heap, Relation index,
retry:
conflict = false;
found_self = false;
- index_scan = index_beginscan(heap, index, &DirtySnapshot, NULL, indnkeyatts, 0);
+ index_scan = index_beginscan(heap, index, false, &DirtySnapshot, NULL,
+ indnkeyatts, 0);
index_rescan(index_scan, scankeys, indnkeyatts, NULL, 0);
- while (index_getnext_slot(index_scan, ForwardScanDirection, existing_slot))
+ while (table_index_getnext_slot(index_scan, ForwardScanDirection,
+ existing_slot))
{
TransactionId xwait;
XLTW_Oper reason_wait;
diff --git a/src/backend/executor/execReplication.c b/src/backend/executor/execReplication.c
index 72f2bff77..ab487b9e6 100644
--- a/src/backend/executor/execReplication.c
+++ b/src/backend/executor/execReplication.c
@@ -205,7 +205,7 @@ RelationFindReplTupleByIndex(Relation rel, Oid idxoid,
skey_attoff = build_replindex_scan_key(skey, rel, idxrel, searchslot);
/* Start an index scan. */
- scan = index_beginscan(rel, idxrel, &snap, NULL, skey_attoff, 0);
+ scan = index_beginscan(rel, idxrel, false, &snap, NULL, skey_attoff, 0);
retry:
found = false;
@@ -213,7 +213,7 @@ retry:
index_rescan(scan, skey, skey_attoff, NULL, 0);
/* Try to find the tuple */
- while (index_getnext_slot(scan, ForwardScanDirection, outslot))
+ while (table_index_getnext_slot(scan, ForwardScanDirection, outslot))
{
/*
* Avoid expensive equality check if the index is primary key or
@@ -666,12 +666,12 @@ RelationFindDeletedTupleInfoByIndex(Relation rel, Oid idxoid,
* not yet committed or those just committed prior to the scan are
* excluded in update_most_recent_deletion_info().
*/
- scan = index_beginscan(rel, idxrel, SnapshotAny, NULL, skey_attoff, 0);
+ scan = index_beginscan(rel, idxrel, false, SnapshotAny, NULL, skey_attoff, 0);
index_rescan(scan, skey, skey_attoff, NULL, 0);
/* Try to find the tuple */
- while (index_getnext_slot(scan, ForwardScanDirection, scanslot))
+ while (table_index_getnext_slot(scan, ForwardScanDirection, scanslot))
{
/*
* Avoid expensive equality check if the index is primary key or
diff --git a/src/backend/executor/nodeBitmapIndexscan.c b/src/backend/executor/nodeBitmapIndexscan.c
index 058a59ef5..2580a0139 100644
--- a/src/backend/executor/nodeBitmapIndexscan.c
+++ b/src/backend/executor/nodeBitmapIndexscan.c
@@ -202,6 +202,7 @@ ExecEndBitmapIndexScan(BitmapIndexScanState *node)
* which will have a new BitmapIndexScanState and zeroed stats.
*/
winstrument->nsearches += node->biss_Instrument.nsearches;
+ Assert(node->biss_Instrument.nheapfetches == 0);
}
/*
diff --git a/src/backend/executor/nodeIndexonlyscan.c b/src/backend/executor/nodeIndexonlyscan.c
index c2d093745..ff3e8f302 100644
--- a/src/backend/executor/nodeIndexonlyscan.c
+++ b/src/backend/executor/nodeIndexonlyscan.c
@@ -34,14 +34,12 @@
#include "access/relscan.h"
#include "access/tableam.h"
#include "access/tupdesc.h"
-#include "access/visibilitymap.h"
#include "catalog/pg_type.h"
#include "executor/executor.h"
#include "executor/nodeIndexonlyscan.h"
#include "executor/nodeIndexscan.h"
#include "miscadmin.h"
#include "storage/bufmgr.h"
-#include "storage/predicate.h"
#include "utils/builtins.h"
#include "utils/rel.h"
@@ -65,7 +63,6 @@ IndexOnlyNext(IndexOnlyScanState *node)
ScanDirection direction;
IndexScanDesc scandesc;
TupleTableSlot *slot;
- ItemPointer tid;
/*
* extract necessary information from index scan node
@@ -90,18 +87,14 @@ IndexOnlyNext(IndexOnlyScanState *node)
* parallel.
*/
scandesc = index_beginscan(node->ss.ss_currentRelation,
- node->ioss_RelationDesc,
+ node->ioss_RelationDesc, true,
estate->es_snapshot,
&node->ioss_Instrument,
node->ioss_NumScanKeys,
node->ioss_NumOrderByKeys);
node->ioss_ScanDesc = scandesc;
-
-
- /* Set it up for index-only scan */
- node->ioss_ScanDesc->xs_want_itup = true;
- node->ioss_VMBuffer = InvalidBuffer;
+ Assert(node->ioss_ScanDesc->xs_want_itup);
/*
* If no run-time keys to calculate or they are ready, go ahead and
@@ -118,78 +111,10 @@ IndexOnlyNext(IndexOnlyScanState *node)
/*
* OK, now that we have what we need, fetch the next tuple.
*/
- while ((tid = index_getnext_tid(scandesc, direction)) != NULL)
+ while (table_index_getnext_slot(scandesc, direction, node->ioss_TableSlot))
{
- bool tuple_from_heap = false;
-
CHECK_FOR_INTERRUPTS();
- /*
- * We can skip the heap fetch if the TID references a heap page on
- * which all tuples are known visible to everybody. In any case,
- * we'll use the index tuple not the heap tuple as the data source.
- *
- * Note on Memory Ordering Effects: visibilitymap_get_status does not
- * lock the visibility map buffer, and therefore the result we read
- * here could be slightly stale. However, it can't be stale enough to
- * matter.
- *
- * We need to detect clearing a VM bit due to an insert right away,
- * because the tuple is present in the index page but not visible. The
- * reading of the TID by this scan (using a shared lock on the index
- * buffer) is serialized with the insert of the TID into the index
- * (using an exclusive lock on the index buffer). Because the VM bit
- * is cleared before updating the index, and locking/unlocking of the
- * index page acts as a full memory barrier, we are sure to see the
- * cleared bit if we see a recently-inserted TID.
- *
- * Deletes do not update the index page (only VACUUM will clear out
- * the TID), so the clearing of the VM bit by a delete is not
- * serialized with this test below, and we may see a value that is
- * significantly stale. However, we don't care about the delete right
- * away, because the tuple is still visible until the deleting
- * transaction commits or the statement ends (if it's our
- * transaction). In either case, the lock on the VM buffer will have
- * been released (acting as a write barrier) after clearing the bit.
- * And for us to have a snapshot that includes the deleting
- * transaction (making the tuple invisible), we must have acquired
- * ProcArrayLock after that time, acting as a read barrier.
- *
- * It's worth going through this complexity to avoid needing to lock
- * the VM buffer, which could cause significant contention.
- */
- if (!VM_ALL_VISIBLE(scandesc->heapRelation,
- ItemPointerGetBlockNumber(tid),
- &node->ioss_VMBuffer))
- {
- /*
- * Rats, we have to visit the heap to check visibility.
- */
- InstrCountTuples2(node, 1);
- if (!index_fetch_heap(scandesc, node->ioss_TableSlot))
- continue; /* no visible tuple, try next index entry */
-
- ExecClearTuple(node->ioss_TableSlot);
-
- /*
- * Only MVCC snapshots are supported here, so there should be no
- * need to keep following the HOT chain once a visible entry has
- * been found. If we did want to allow that, we'd need to keep
- * more state to remember not to call index_getnext_tid next time.
- */
- if (scandesc->xs_heap_continue)
- elog(ERROR, "non-MVCC snapshots are not supported in index-only scans");
-
- /*
- * Note: at this point we are holding a pin on the heap page, as
- * recorded in scandesc->xs_cbuf. We could release that pin now,
- * but it's not clear whether it's a win to do so. The next index
- * entry might require a visit to the same heap page.
- */
-
- tuple_from_heap = true;
- }
-
/*
* Fill the scan tuple slot with data from the index. This might be
* provided in either HeapTuple or IndexTuple format. Conceivably an
@@ -238,16 +163,6 @@ IndexOnlyNext(IndexOnlyScanState *node)
ereport(ERROR,
(errcode(ERRCODE_FEATURE_NOT_SUPPORTED),
errmsg("lossy distance functions are not supported in index-only scans")));
-
- /*
- * If we didn't access the heap, then we'll need to take a predicate
- * lock explicitly, as if we had. For now we do that at page level.
- */
- if (!tuple_from_heap)
- PredicateLockPage(scandesc->heapRelation,
- ItemPointerGetBlockNumber(tid),
- estate->es_snapshot);
-
return slot;
}
@@ -407,13 +322,6 @@ ExecEndIndexOnlyScan(IndexOnlyScanState *node)
indexRelationDesc = node->ioss_RelationDesc;
indexScanDesc = node->ioss_ScanDesc;
- /* Release VM buffer pin, if any. */
- if (node->ioss_VMBuffer != InvalidBuffer)
- {
- ReleaseBuffer(node->ioss_VMBuffer);
- node->ioss_VMBuffer = InvalidBuffer;
- }
-
/*
* When ending a parallel worker, copy the statistics gathered by the
* worker back into shared memory so that it can be picked up by the main
@@ -433,6 +341,7 @@ ExecEndIndexOnlyScan(IndexOnlyScanState *node)
* which will have a new IndexOnlyScanState and zeroed stats.
*/
winstrument->nsearches += node->ioss_Instrument.nsearches;
+ winstrument->nheapfetches += node->ioss_Instrument.nheapfetches;
}
/*
@@ -784,13 +693,12 @@ ExecIndexOnlyScanInitializeDSM(IndexOnlyScanState *node,
node->ioss_ScanDesc =
index_beginscan_parallel(node->ss.ss_currentRelation,
- node->ioss_RelationDesc,
+ node->ioss_RelationDesc, true,
&node->ioss_Instrument,
node->ioss_NumScanKeys,
node->ioss_NumOrderByKeys,
piscan);
- node->ioss_ScanDesc->xs_want_itup = true;
- node->ioss_VMBuffer = InvalidBuffer;
+ Assert(node->ioss_ScanDesc->xs_want_itup);
/*
* If no run-time keys to calculate or they are ready, go ahead and pass
@@ -850,12 +758,12 @@ ExecIndexOnlyScanInitializeWorker(IndexOnlyScanState *node,
node->ioss_ScanDesc =
index_beginscan_parallel(node->ss.ss_currentRelation,
- node->ioss_RelationDesc,
+ node->ioss_RelationDesc, true,
&node->ioss_Instrument,
node->ioss_NumScanKeys,
node->ioss_NumOrderByKeys,
piscan);
- node->ioss_ScanDesc->xs_want_itup = true;
+ Assert(node->ioss_ScanDesc->xs_want_itup);
/*
* If no run-time keys to calculate or they are ready, go ahead and pass
diff --git a/src/backend/executor/nodeIndexscan.c b/src/backend/executor/nodeIndexscan.c
index 84823f0b6..c34a13a87 100644
--- a/src/backend/executor/nodeIndexscan.c
+++ b/src/backend/executor/nodeIndexscan.c
@@ -107,7 +107,7 @@ IndexNext(IndexScanState *node)
* serially executing an index scan that was planned to be parallel.
*/
scandesc = index_beginscan(node->ss.ss_currentRelation,
- node->iss_RelationDesc,
+ node->iss_RelationDesc, false,
estate->es_snapshot,
&node->iss_Instrument,
node->iss_NumScanKeys,
@@ -128,7 +128,7 @@ IndexNext(IndexScanState *node)
/*
* ok, now that we have what we need, fetch the next tuple.
*/
- while (index_getnext_slot(scandesc, direction, slot))
+ while (table_index_getnext_slot(scandesc, direction, slot))
{
CHECK_FOR_INTERRUPTS();
@@ -203,7 +203,7 @@ IndexNextWithReorder(IndexScanState *node)
* serially executing an index scan that was planned to be parallel.
*/
scandesc = index_beginscan(node->ss.ss_currentRelation,
- node->iss_RelationDesc,
+ node->iss_RelationDesc, false,
estate->es_snapshot,
&node->iss_Instrument,
node->iss_NumScanKeys,
@@ -260,7 +260,7 @@ IndexNextWithReorder(IndexScanState *node)
* Fetch next tuple from the index.
*/
next_indextuple:
- if (!index_getnext_slot(scandesc, ForwardScanDirection, slot))
+ if (!table_index_getnext_slot(scandesc, ForwardScanDirection, slot))
{
/*
* No more tuples from the index. But we still need to drain any
@@ -812,6 +812,7 @@ ExecEndIndexScan(IndexScanState *node)
* which will have a new IndexOnlyScanState and zeroed stats.
*/
winstrument->nsearches += node->iss_Instrument.nsearches;
+ Assert(node->iss_Instrument.nheapfetches == 0);
}
/*
@@ -1719,7 +1720,7 @@ ExecIndexScanInitializeDSM(IndexScanState *node,
node->iss_ScanDesc =
index_beginscan_parallel(node->ss.ss_currentRelation,
- node->iss_RelationDesc,
+ node->iss_RelationDesc, false,
&node->iss_Instrument,
node->iss_NumScanKeys,
node->iss_NumOrderByKeys,
@@ -1783,7 +1784,7 @@ ExecIndexScanInitializeWorker(IndexScanState *node,
node->iss_ScanDesc =
index_beginscan_parallel(node->ss.ss_currentRelation,
- node->iss_RelationDesc,
+ node->iss_RelationDesc, false,
&node->iss_Instrument,
node->iss_NumScanKeys,
node->iss_NumOrderByKeys,
diff --git a/src/backend/optimizer/path/indxpath.c b/src/backend/optimizer/path/indxpath.c
index 51b9d6677..03598b50d 100644
--- a/src/backend/optimizer/path/indxpath.c
+++ b/src/backend/optimizer/path/indxpath.c
@@ -45,7 +45,7 @@
/* Whether we are looking for plain indexscan, bitmap scan, or either */
typedef enum
{
- ST_INDEXSCAN, /* must support amgettuple */
+ ST_INDEXSCAN, /* must support amgettuple or amgetbatch */
ST_BITMAPSCAN, /* must support amgetbitmap */
ST_ANYSCAN, /* either is okay */
} ScanTypeControl;
diff --git a/src/backend/optimizer/util/plancat.c b/src/backend/optimizer/util/plancat.c
index a90e1c9ee..39a5ea299 100644
--- a/src/backend/optimizer/util/plancat.c
+++ b/src/backend/optimizer/util/plancat.c
@@ -313,11 +313,11 @@ get_relation_info(PlannerInfo *root, Oid relationObjectId, bool inhparent,
info->amsearcharray = amroutine->amsearcharray;
info->amsearchnulls = amroutine->amsearchnulls;
info->amcanparallel = amroutine->amcanparallel;
- info->amhasgettuple = (amroutine->amgettuple != NULL);
+ info->amhasgettuple = (amroutine->amgettuple != NULL ||
+ amroutine->amgetbatch != NULL);
info->amhasgetbitmap = amroutine->amgetbitmap != NULL &&
relation->rd_tableam->scan_bitmap_next_tuple != NULL;
- info->amcanmarkpos = (amroutine->ammarkpos != NULL &&
- amroutine->amrestrpos != NULL);
+ info->amcanmarkpos = amroutine->amposreset != NULL;
info->amcostestimate = amroutine->amcostestimate;
Assert(info->amcostestimate != NULL);
diff --git a/src/backend/replication/logical/relation.c b/src/backend/replication/logical/relation.c
index 0b1d80b5b..76b0d035f 100644
--- a/src/backend/replication/logical/relation.c
+++ b/src/backend/replication/logical/relation.c
@@ -890,7 +890,8 @@ IsIndexUsableForReplicaIdentityFull(Relation idxrel, AttrMap *attrmap)
* The given index access method must implement "amgettuple", which will
* be used later to fetch the tuples. See RelationFindReplTupleByIndex().
*/
- if (GetIndexAmRoutineByAmId(idxrel->rd_rel->relam, false)->amgettuple == NULL)
+ if (GetIndexAmRoutineByAmId(idxrel->rd_rel->relam, false)->amgettuple == NULL &&
+ GetIndexAmRoutineByAmId(idxrel->rd_rel->relam, false)->amgetbatch == NULL)
return false;
return true;
diff --git a/src/backend/utils/adt/amutils.c b/src/backend/utils/adt/amutils.c
index c81fb61a0..ddfd1b55c 100644
--- a/src/backend/utils/adt/amutils.c
+++ b/src/backend/utils/adt/amutils.c
@@ -363,10 +363,11 @@ indexam_property(FunctionCallInfo fcinfo,
PG_RETURN_BOOL(routine->amclusterable);
case AMPROP_INDEX_SCAN:
- PG_RETURN_BOOL(routine->amgettuple ? true : false);
+ PG_RETURN_BOOL(routine->amgettuple != NULL ||
+ routine->amgetbatch != NULL);
case AMPROP_BITMAP_SCAN:
- PG_RETURN_BOOL(routine->amgetbitmap ? true : false);
+ PG_RETURN_BOOL(routine->amgetbitmap != NULL);
case AMPROP_BACKWARD_SCAN:
PG_RETURN_BOOL(routine->amcanbackward);
@@ -392,7 +393,8 @@ indexam_property(FunctionCallInfo fcinfo,
PG_RETURN_BOOL(routine->amcanmulticol);
case AMPROP_CAN_EXCLUDE:
- PG_RETURN_BOOL(routine->amgettuple ? true : false);
+ PG_RETURN_BOOL(routine->amgettuple != NULL ||
+ routine->amgetbatch != NULL);
case AMPROP_CAN_INCLUDE:
PG_RETURN_BOOL(routine->amcaninclude);
diff --git a/src/backend/utils/adt/selfuncs.c b/src/backend/utils/adt/selfuncs.c
index 29fec6555..a7213654a 100644
--- a/src/backend/utils/adt/selfuncs.c
+++ b/src/backend/utils/adt/selfuncs.c
@@ -102,7 +102,6 @@
#include "access/gin.h"
#include "access/table.h"
#include "access/tableam.h"
-#include "access/visibilitymap.h"
#include "catalog/pg_collation.h"
#include "catalog/pg_operator.h"
#include "catalog/pg_statistic.h"
@@ -7124,13 +7123,10 @@ get_actual_variable_endpoint(Relation heapRel,
bool have_data = false;
SnapshotData SnapshotNonVacuumable;
IndexScanDesc index_scan;
- Buffer vmbuffer = InvalidBuffer;
- BlockNumber last_heap_block = InvalidBlockNumber;
- int n_visited_heap_pages = 0;
- ItemPointer tid;
Datum values[INDEX_MAX_KEYS];
bool isnull[INDEX_MAX_KEYS];
MemoryContext oldcontext;
+ IndexScanInstrumentation instrument;
/*
* We use the index-only-scan machinery for this. With mostly-static
@@ -7179,56 +7175,32 @@ get_actual_variable_endpoint(Relation heapRel,
InitNonVacuumableSnapshot(SnapshotNonVacuumable,
GlobalVisTestFor(heapRel));
- index_scan = index_beginscan(heapRel, indexRel,
- &SnapshotNonVacuumable, NULL,
+ /*
+ * Set it up for instrumented index-only scan. We need the
+ * instrumentation to monitor the number of heap fetches.
+ */
+ memset(&instrument, 0, sizeof(instrument));
+ index_scan = index_beginscan(heapRel, indexRel, true,
+ &SnapshotNonVacuumable, &instrument,
1, 0);
- /* Set it up for index-only scan */
- index_scan->xs_want_itup = true;
+ Assert(index_scan->xs_want_itup);
index_rescan(index_scan, scankeys, 1, NULL, 0);
/* Fetch first/next tuple in specified direction */
- while ((tid = index_getnext_tid(index_scan, indexscandir)) != NULL)
+ while (table_index_getnext_slot(index_scan, indexscandir, tableslot))
{
- BlockNumber block = ItemPointerGetBlockNumber(tid);
+ /* We don't actually need the heap tuple for anything */
+ ExecClearTuple(tableslot);
- if (!VM_ALL_VISIBLE(heapRel,
- block,
- &vmbuffer))
- {
- /* Rats, we have to visit the heap to check visibility */
- if (!index_fetch_heap(index_scan, tableslot))
- {
- /*
- * No visible tuple for this index entry, so we need to
- * advance to the next entry. Before doing so, count heap
- * page fetches and give up if we've done too many.
- *
- * We don't charge a page fetch if this is the same heap page
- * as the previous tuple. This is on the conservative side,
- * since other recently-accessed pages are probably still in
- * buffers too; but it's good enough for this heuristic.
- */
+ /*
+ * No visible tuple for this index entry, so we need to advance to the
+ * next entry. Before doing so, count heap page fetches and give up
+ * if we've done too many.
+ */
#define VISITED_PAGES_LIMIT 100
- if (block != last_heap_block)
- {
- last_heap_block = block;
- n_visited_heap_pages++;
- if (n_visited_heap_pages > VISITED_PAGES_LIMIT)
- break;
- }
-
- continue; /* no visible tuple, try next index entry */
- }
-
- /* We don't actually need the heap tuple for anything */
- ExecClearTuple(tableslot);
-
- /*
- * We don't care whether there's more than one visible tuple in
- * the HOT chain; if any are visible, that's good enough.
- */
- }
+ if (instrument.nheapfetches > VISITED_PAGES_LIMIT)
+ break;
/*
* We expect that the index will return data in IndexTuple not
@@ -7261,8 +7233,6 @@ get_actual_variable_endpoint(Relation heapRel,
break;
}
- if (vmbuffer != InvalidBuffer)
- ReleaseBuffer(vmbuffer);
index_endscan(index_scan);
return have_data;
diff --git a/contrib/bloom/blutils.c b/contrib/bloom/blutils.c
index 5111cdc6d..7fd98dba6 100644
--- a/contrib/bloom/blutils.c
+++ b/contrib/bloom/blutils.c
@@ -146,10 +146,11 @@ blhandler(PG_FUNCTION_ARGS)
.ambeginscan = blbeginscan,
.amrescan = blrescan,
.amgettuple = NULL,
+ .amgetbatch = NULL,
+ .amfreebatch = NULL,
.amgetbitmap = blgetbitmap,
.amendscan = blendscan,
- .ammarkpos = NULL,
- .amrestrpos = NULL,
+ .amposreset = NULL,
.amestimateparallelscan = NULL,
.aminitparallelscan = NULL,
.amparallelrescan = NULL,
diff --git a/doc/src/sgml/indexam.sgml b/doc/src/sgml/indexam.sgml
index f48da3185..8df91bc38 100644
--- a/doc/src/sgml/indexam.sgml
+++ b/doc/src/sgml/indexam.sgml
@@ -167,10 +167,11 @@ typedef struct IndexAmRoutine
ambeginscan_function ambeginscan;
amrescan_function amrescan;
amgettuple_function amgettuple; /* can be NULL */
+ amgetbatch_function amgetbatch; /* can be NULL */
+ amfreebatch_function amfreebatch; /* can be NULL */
amgetbitmap_function amgetbitmap; /* can be NULL */
amendscan_function amendscan;
- ammarkpos_function ammarkpos; /* can be NULL */
- amrestrpos_function amrestrpos; /* can be NULL */
+ amposreset_function amposreset; /* can be NULL */
/* interface functions to support parallel index scans */
amestimateparallelscan_function amestimateparallelscan; /* can be NULL */
@@ -749,6 +750,138 @@ amgettuple (IndexScanDesc scan,
<structfield>amgettuple</structfield> field in its <structname>IndexAmRoutine</structname>
struct must be set to NULL.
</para>
+ <note>
+ <para>
+ As of <productname>PostgreSQL</productname> version 19, position marking
+ and restoration of scans is no longer supported for the
+ <function>amgettuple</function> interface; only the
+ <function>amgetbatch</function> interface supports this feature through
+ the <function>amposreset</function> callback.
+ </para>
+ </note>
+
+ <para>
+<programlisting>
+BatchIndexScan
+amgetbatch (IndexScanDesc scan,
+ BatchIndexScan priorbatch,
+ ScanDirection direction);
+</programlisting>
+ Return the next batch of index tuples in the given scan, moving in the
+ given direction (forward or backward in the index). Returns an instance of
+ <type>BatchIndexScan</type> with index tuples loaded, or
+ <literal>NULL</literal> if there are no more index tuples.
+ </para>
+
+ <para>
+ The <literal>priorbatch</literal> parameter passes the batch previously
+ returned by an earlier <function>amgetbatch</function> call (or
+ <literal>NULL</literal> on the first call). The index AM uses
+ <literal>priorbatch</literal> to determine which index page to read next,
+ typically by following page links found in <literal>priorbatch</literal>.
+ The returned batch contains matching items immediately adjacent to those
+ from <literal>priorbatch</literal> in the common case where
+ <literal>priorbatch</literal> is the batch that was returned by the most
+ recent call to <function>amgetbatch</function> call (though not when the
+ most recent call used the opposite scan direction to this call, and not
+ when a mark has been restored).
+ </para>
+
+ <para>
+ A batch returned by <function>amgetbatch</function> is guaranteed to be
+ associated with an index page containing at least one matching tuple.
+ The index page associated with the batch may be retained in a buffer with
+ its pin held as an interlock against concurrent TID recycling by
+ <command>VACUUM</command>. See <xref linkend="index-locking"/> for details
+ on buffer pin management during <quote>plain</quote> index scans.
+ </para>
+
+ <para>
+ The <function>amgetbatch</function> interface does not support index-only
+ scans that return data via the <literal>xs_hitup</literal> mechanism.
+ Index-only scans are supported through the <literal>xs_itup</literal>
+ mechanism only.
+ </para>
+
+ <para>
+ The <function>amgetbatch</function> function need only be provided if the
+ access method supports <quote>plain</quote> index scans. If it doesn't,
+ the <function>amgetbatch</function> field in its
+ <structname>IndexAmRoutine</structname> struct must be set to NULL.
+ </para>
+
+ <para>
+ A <type>BatchIndexScan</type> that is returned by
+ <function>amgetbatch</function> is no longer managed by the access method.
+ It is up to the table AM caller to decide when it should be freed by
+ passing it to <function>amfreebatch</function>. Note also that
+ <function>amgetbatch</function> functions must never modify the
+ <structfield>priorbatch</structfield> parameter.
+ </para>
+
+ <para>
+ The access method may provide only one of <function>amgettuple</function>
+ and <function>amgetbatch</function> callbacks, not both (XXX uncertain).
+ When the access method provides <function>amgetbatch</function>, it must
+ also provide <function>amfreebatch</function>.
+ </para>
+
+ <para>
+ The same caveats described for <function>amgettuple</function> apply here
+ too: an entry in the returned batch means only that the index contains
+ an entry that matches the scan keys, not that the tuple necessarily still
+ exists in the heap or will pass the caller's snapshot test.
+ </para>
+
+ <para>
+<programlisting>
+void
+amfreebatch (IndexScanDesc scan,
+ BatchIndexScan batch);
+</programlisting>
+ Releases a batch returned by the <function>amgetbatch</function> callback.
+ This function is called exclusively by table access methods to indicate
+ that processing of the batch is complete; it should never be called within
+ the index access method itself.
+ </para>
+
+ <para>
+ <function>amfreebatch</function> frees buffer pins held on the batch's
+ associated index page and releases related memory and resources. These
+ buffer pins serve as an interlock against concurrent TID recycling by
+ <command>VACUUM</command>, protecting the table access method from confusion
+ about which TID corresponds to which logical row. See <xref
+ linkend="index-locking"/> for detailed discussion of buffer pin management.
+ </para>
+
+ <para>
+ The index AM may choose to retain its own buffer pins across multiple
+ <function>amfreebatch</function> calls when this serves an internal purpose
+ (for example, maintaining a descent stack of pinned index pages for reuse
+ across <function>amgetbatch</function> calls). However, any scheme that
+ retains buffer pins must keep the number of retained pins fixed and small,
+ to avoid exhausting the backend's buffer pin limit.
+ </para>
+
+ <para>
+ The index AM has the option of setting <literal>LP_DEAD</literal> bits in
+ the index page to mark dead tuples before releasing the buffer pin. When
+ <literal>IndexScanDescData.dropPin</literal> is true and the buffer pin is being
+ dropped eagerly, the index AM must check <literal>BatchIndexScan.lsn</literal>
+ to verify that the page LSN has not advanced since the batch was originally
+ read before setting <literal>LP_DEAD</literal> bits, to avoid concurrent
+ TID recycling hazards. When <literal>IndexScanDescData.dropPin</literal>
+ is false (requiring that a buffer pin be held throughout first reading the
+ index leaf page and calling <function>amfreebatch</function>),
+ <literal>LP_DEAD</literal> bits can always be set safely without an LSN
+ check.
+ </para>
+
+ <para>
+ The <function>amfreebatch</function> function need only be provided if the
+ access method provides <function>amgetbatch</function>. Otherwise it has to
+ remain set to <literal>NULL</literal>.
+ </para>
<para>
<programlisting>
@@ -768,8 +901,8 @@ amgetbitmap (IndexScanDesc scan,
itself, and therefore callers recheck both the scan conditions and the
partial index predicate (if any) for recheckable tuples. That might not
always be true, however.
- <function>amgetbitmap</function> and
- <function>amgettuple</function> cannot be used in the same index scan; there
+ Only one of <function>amgetbitmap</function>, <function>amgettuple</function>,
+ or <function>amgetbatch</function> can be used in any given index scan; there
are other restrictions too when using <function>amgetbitmap</function>, as explained
in <xref linkend="index-scanning"/>.
</para>
@@ -795,32 +928,25 @@ amendscan (IndexScanDesc scan);
<para>
<programlisting>
void
-ammarkpos (IndexScanDesc scan);
+amposreset (IndexScanDesc scan);
</programlisting>
- Mark current scan position. The access method need only support one
- remembered scan position per scan.
+ Notify index AM that core code will change the scan's position to an item
+ returned as part of an earlier batch. The index AM must therefore
+ invalidate any state that independently tracks the scan's progress
+ (e.g., array keys used with a ScalarArrayOpExpr qual). Called by the core
+ system when it is about to restore a mark.
</para>
<para>
- The <function>ammarkpos</function> function need only be provided if the access
- method supports ordered scans. If it doesn't,
- the <structfield>ammarkpos</structfield> field in its <structname>IndexAmRoutine</structname>
- struct may be set to NULL.
- </para>
-
- <para>
-<programlisting>
-void
-amrestrpos (IndexScanDesc scan);
-</programlisting>
- Restore the scan to the most recently marked position.
- </para>
-
- <para>
- The <function>amrestrpos</function> function need only be provided if the access
- method supports ordered scans. If it doesn't,
- the <structfield>amrestrpos</structfield> field in its <structname>IndexAmRoutine</structname>
- struct may be set to NULL.
+ The <function>amposreset</function> function can only be provided if the
+ access method supports ordered scans through the <function>amgetbatch</function>
+ interface. If it doesn't, the <structfield>amposreset</structfield> field
+ in its <structname>IndexAmRoutine</structname> struct should be set to
+ NULL. Index AMs that don't have any private state that might need to be
+ invalidated might still find it useful to provide an empty
+ <structfield>amposreset</structfield> function; if <function>amposreset</function>
+ is set to NULL, the core system will assume that it is unsafe to restore a
+ marked position.
</para>
<para>
@@ -994,30 +1120,47 @@ amtranslatecmptype (CompareType cmptype, Oid opfamily, Oid opcintype);
</para>
<para>
- The <function>amgettuple</function> function has a <literal>direction</literal> argument,
+ The <function>amgettuple</function> and <function>amgetbatch</function>
+ functions have a <literal>direction</literal> argument,
which can be either <literal>ForwardScanDirection</literal> (the normal case)
or <literal>BackwardScanDirection</literal>. If the first call after
<function>amrescan</function> specifies <literal>BackwardScanDirection</literal>, then the
set of matching index entries is to be scanned back-to-front rather than in
- the normal front-to-back direction, so <function>amgettuple</function> must return
- the last matching tuple in the index, rather than the first one as it
- normally would. (This will only occur for access
- methods that set <structfield>amcanorder</structfield> to true.) After the
- first call, <function>amgettuple</function> must be prepared to advance the scan in
+ the normal front-to-back direction. In this case,
+ <function>amgettuple</function> must return the last matching tuple in the
+ index, rather than the first one as it normally would. Similarly,
+ <function>amgetbatch</function> must return the last matching batch of items
+ when either the first call after <function>amrescan</function> specifies
+ <literal>BackwardScanDirection</literal>, or a subsequent call has
+ <literal>NULL</literal> as its <structfield>priorbatch</structfield> argument
+ (indicating a backward scan restart). (This backward-scan behavior will
+ only occur for access methods that set <structfield>amcanorder</structfield>
+ to true.) After the first call, both <function>amgettuple</function> and
+ <function>amgetbatch</function> must be prepared to advance the scan in
either direction from the most recently returned entry. (But if
<structfield>amcanbackward</structfield> is false, all subsequent
calls will have the same direction as the first one.)
</para>
<para>
- Access methods that support ordered scans must support <quote>marking</quote> a
- position in a scan and later returning to the marked position. The same
- position might be restored multiple times. However, only one position need
- be remembered per scan; a new <function>ammarkpos</function> call overrides the
- previously marked position. An access method that does not support ordered
- scans need not provide <function>ammarkpos</function> and <function>amrestrpos</function>
- functions in <structname>IndexAmRoutine</structname>; set those pointers to NULL
- instead.
+ Access methods using the <function>amgetbatch</function> interface may
+ support <quote>marking</quote> a position in a scan and later returning to
+ the marked position, though this is optional. If the same marked position
+ might be restored multiple times, the core system manages marking and
+ restoration through the <function>index_batch_mark_pos</function> and
+ <function>index_batch_restore_pos</function> internal functions. When a
+ marked position is restored, the index AM is notified via the
+ <function>amposreset</function> callback so it can invalidate any private
+ state that independently tracks the scan's progress (such as array key
+ state).
+ </para>
+
+ <para>
+ The <function>amposreset</function> function in <structname>IndexAmRoutine</structname>
+ should be set to NULL for access methods that do not support mark/restore.
+ For access methods that do support this feature, <function>amposreset</function>
+ must be provided (though it can be a no-op function if the AM has no private
+ state to invalidate).
</para>
<para>
@@ -1186,6 +1329,94 @@ amtranslatecmptype (CompareType cmptype, Oid opfamily, Oid opcintype);
reduce the frequency of such transaction cancellations.
</para>
+ <sect2 id="index-locking-batches">
+ <title>Batch Scanning and Buffer Pin Management</title>
+
+ <para>
+ Index access methods that implement the <function>amgetbatch</function>
+ interface must cooperate with the core system to manage buffer pins in a
+ way that prevents concurrent <command>VACUUM</command> from creating
+ TID recycling hazards. Unlike <function>amgettuple</function> scans,
+ which keep the index access method in control of scan progression,
+ <function>amgetbatch</function> scans give control to the table access
+ method, which may fetch table tuples in a different order than the index
+ entries were returned. This creates the need for explicit buffer pin
+ management to ensure the table access method does not confuse a recycled
+ TID with the original row it meant to reference.
+ </para>
+
+ <para>
+ When <function>amgetbatch</function> returns a batch, the batch's
+ associated index page may be retained in a buffer with a pin held on it.
+ This pin serves as an interlock: <command>VACUUM</command> cannot recycle
+ TIDs on a pinned page. The buffer pin protects only the table access
+ method's ability to map TIDs to rows correctly; it does not protect the
+ index structure itself. Index access methods may use pins for other
+ purposes (for example, maintaining a descent stack of pinned pages), but
+ those uses are internal to the access method and independent of the
+ table-AM synchronization described here.
+ </para>
+
+ <para>
+ Whether a pin should be held when returning a batch is controlled by the
+ <structfield>dropPin</structfield> flag in the <type>BatchRingBuffer</type>
+ structure. When <literal>dropPin</literal> is true, the index access method
+ drops the pin before returning the batch, which avoids blocking
+ <command>VACUUM</command>. When <literal>dropPin</literal> is false, the
+ index access method must hold the pin until the batch is freed via
+ <function>amfreebatch</function>. The core system sets the
+ <literal>dropPin</literal> flag based on scan type: it is true for
+ MVCC-compliant snapshots on logged relations (unless index-only scans are
+ in use), and false otherwise.
+ </para>
+
+ <para>
+ When <literal>dropPin</literal> is true and the index access method is
+ eager about dropping pins, it must save the page's LSN in the batch before
+ returning. Later, when <function>amfreebatch</function> is called and the
+ access method wishes to set <literal>LP_DEAD</literal> bits to mark dead
+ tuples, it must verify that the page's LSN has not changed since the batch
+ was read. If the LSN has changed, the page may have been modified by
+ concurrent activity and it is unsafe to set <literal>LP_DEAD</literal> bits.
+ This LSN-based validation scheme protects against TID recycling races when
+ pins have been dropped. When <literal>dropPin</literal> is false, the pin
+ prevents unsafe concurrent removal of table TID references by
+ <command>VACUUM</command>, so no LSN check is necessary.
+ </para>
+
+ <para>
+ The core system provides three utility functions for managing batch
+ resources:
+ <function>indexam_util_batch_alloc</function> allocates a new batch or
+ reuses a cached one,
+ <function>indexam_util_batch_unlock</function> drops the lock and
+ conditionally drops the pin on a batch's index page (based on the
+ <literal>dropPin</literal> setting), and
+ <function>indexam_util_batch_release</function> frees or caches a batch.
+ Index access methods should use these utilities rather than managing
+ buffers directly. The <filename>src/backend/access/nbtree/</filename>
+ implementation provides a reference example of correct usage.
+ </para>
+
+ <para>
+ Note that <function>amfreebatch</function> is called only by the core code
+ and table access method, never by the index access method itself. The
+ index AM must not assume that a call to <function>amfreebatch</function>
+ will take place before another call to <function>amgetbatch</function>
+ (for the same index scan) takes place.
+ </para>
+
+ <para>
+ The index AM must also avoid relying on the core code calling
+ <function>amfreebatch</function> with batches that are in any particular
+ order. For example, it is not okay for an index AM to assume that calls
+ to <function>amfreebatch</function> will take place in the same order as
+ the <function>amgetbatch</function> calls that initially
+ allocated/populated/returned each batch.
+ </para>
+
+ </sect2>
+
</sect1>
<sect1 id="index-unique-checks">
diff --git a/doc/src/sgml/ref/create_table.sgml b/doc/src/sgml/ref/create_table.sgml
index 77c5a763d..55b7222e9 100644
--- a/doc/src/sgml/ref/create_table.sgml
+++ b/doc/src/sgml/ref/create_table.sgml
@@ -1152,12 +1152,13 @@ WITH ( MODULUS <replaceable class="parameter">numeric_literal</replaceable>, REM
</para>
<para>
- The access method must support <literal>amgettuple</literal> (see <xref
- linkend="indexam"/>); at present this means <acronym>GIN</acronym>
- cannot be used. Although it's allowed, there is little point in using
- B-tree or hash indexes with an exclusion constraint, because this
- does nothing that an ordinary unique constraint doesn't do better.
- So in practice the access method will always be <acronym>GiST</acronym> or
+ The access method must support either <literal>amgettuple</literal>
+ or <literal>amgetbatch</literal> (see <xref linkend="indexam"/>); at
+ present this means <acronym>GIN</acronym> cannot be used. Although
+ it's allowed, there is little point in using B-tree or hash indexes
+ with an exclusion constraint, because this does nothing that an
+ ordinary unique constraint doesn't do better. So in practice the
+ access method will always be <acronym>GiST</acronym> or
<acronym>SP-GiST</acronym>.
</para>
diff --git a/src/test/modules/dummy_index_am/dummy_index_am.c b/src/test/modules/dummy_index_am/dummy_index_am.c
index 9eb8f0a6c..9d4fddec4 100644
--- a/src/test/modules/dummy_index_am/dummy_index_am.c
+++ b/src/test/modules/dummy_index_am/dummy_index_am.c
@@ -317,10 +317,11 @@ dihandler(PG_FUNCTION_ARGS)
.ambeginscan = dibeginscan,
.amrescan = direscan,
.amgettuple = NULL,
+ .amgetbatch = NULL,
+ .amfreebatch = NULL,
.amgetbitmap = NULL,
.amendscan = diendscan,
- .ammarkpos = NULL,
- .amrestrpos = NULL,
+ .amposreset = NULL,
.amestimateparallelscan = NULL,
.aminitparallelscan = NULL,
.amparallelrescan = NULL,
diff --git a/src/tools/pgindent/typedefs.list b/src/tools/pgindent/typedefs.list
index 14dec2d49..d25ba3412 100644
--- a/src/tools/pgindent/typedefs.list
+++ b/src/tools/pgindent/typedefs.list
@@ -226,8 +226,6 @@ BTScanInsertData
BTScanKeyPreproc
BTScanOpaque
BTScanOpaqueData
-BTScanPosData
-BTScanPosItem
BTShared
BTSortArrayContext
BTSpool
@@ -3471,12 +3469,10 @@ amgettuple_function
aminitparallelscan_function
aminsert_function
aminsertcleanup_function
-ammarkpos_function
amoptions_function
amparallelrescan_function
amproperty_function
amrescan_function
-amrestrpos_function
amtranslate_cmptype_function
amtranslate_strategy_function
amvacuumcleanup_function
--
2.51.0
view thread (2+ messages) latest in thread
reply
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Reply to all the recipients using the --to and --cc options:
reply via email
To: [email protected]
Cc: [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected], [email protected]
Subject: Re: index prefetching
In-Reply-To: <CAH2-Wz=6a7fGz2rALDX+xiFDuEaGQWpZ49xEaBUDKiPH8gcL+Q@mail.gmail.com>
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox