public inbox for [email protected]
help / color / mirror / Atom feedpgcrypto/des tests fail on riscv64 due to clang's code generation anomaly
4+ messages / 3 participants
[nested] [flat]
* pgcrypto/des tests fail on riscv64 due to clang's code generation anomaly
@ 2026-03-15 12:00 Alexander Lakhin <[email protected]>
0 siblings, 1 reply; 4+ messages in thread
From: Alexander Lakhin @ 2026-03-15 12:00 UTC (permalink / raw)
To: pgsql-hackers; +Cc: Greg Burd <[email protected]>
Hello hackers,
As recent failures from greenfly, e.g. [1], show:
--- /home/gburd/build/HEAD/pgsql/contrib/pgcrypto/expected/crypt-des.out 2026-03-13 18:06:10.669824704 +0000
+++ /home/gburd/build/HEAD/pgsql.build/testrun/pgcrypto/regress/results/crypt-des.out 2026-03-13 18:38:39.830105127 +0000
@@ -4,13 +4,13 @@
SELECT crypt('', 'NB');
crypt
---------------
- NBPx/38Y48kHg
+ NB3TsPvodZDgs
(1 row)
adding -march=rv64gcv to CFLAGS makes clang generate different (wrong?)
code for crypt-des.c. Namely, having reproduced the diff locally, with
this debugging addition:
--- a/contrib/pgcrypto/crypt-des.c
+++ b/contrib/pgcrypto/crypt-des.c
@@ -355,6 +355,11 @@ des_init(void)
for (i = 0; i < 32; i++)
un_pbox[pbox[i] - 1] = i;
+fprintf(stderr, "!!!des_init| un_pbox:");
+ for (i = 0; i < 32; i++)
+ fprintf(stderr, " %d", un_pbox[i]);
+fprintf(stderr, "\n");
+
for (b = 0; b < 4; b++)
I can see the following during a normal execution of:
CREATE EXTENSION pgcrypto; SELECT crypt('', 'NB');
crypt
---------------
NBPx/38Y48kHg
!!!des_init| un_pbox: 8 16 22 30 12 27 1 17 23 15 29 5 25 19 9 0 7 13 24 2 3 28 10 18 31 11 21 6 4 26 14 20
vs:
crypt
---------------
NB3TsPvodZDgs
!!!des_init| un_pbox: 15 6 19 20 28 11 27 16 0 14 22 25 4 17 30 9 1 7 23 13 31 26 2 8 18 12 29 5 21 10 3 24
when the code compiled with CC=clang-20 CPPFLAGS="-march=rv64gcv".
I don't observe this issue with gcc-14, nor with this modification (using
clang-20, -march=rv64gcv):
for (i = 0; i < 32; i++)
+{
un_pbox[pbox[i] - 1] = i;
+pg_compiler_barrier();
+}
[1] https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=greenfly&dt=2026-03-13%2018%3A06%3A09
Best regards,
Alexander
^ permalink raw reply [nested|flat] 4+ messages in thread
* Re: pgcrypto/des tests fail on riscv64 due to clang's code generation anomaly
@ 2026-03-16 01:08 Michael Paquier <[email protected]>
parent: Alexander Lakhin <[email protected]>
0 siblings, 1 reply; 4+ messages in thread
From: Michael Paquier @ 2026-03-16 01:08 UTC (permalink / raw)
To: Alexander Lakhin <[email protected]>; +Cc: pgsql-hackers; Greg Burd <[email protected]>
On Sun, Mar 15, 2026 at 02:00:00PM +0200, Alexander Lakhin wrote:
> adding -march=rv64gcv to CFLAGS makes clang generate different (wrong?)
> code for crypt-des.c. Namely, having reproduced the diff locally, with
> this debugging addition:
>
> when the code compiled with CC=clang-20 CPPFLAGS="-march=rv64gcv".
Thanks for the detailed investigation. I was puzzled by it myself for
a couple of hours, and did not see something else than a broken
compiler here..
--
Michael
Attachments:
[application/pgp-signature] signature.asc (833B, 2-signature.asc)
download
^ permalink raw reply [nested|flat] 4+ messages in thread
* Re: pgcrypto/des tests fail on riscv64 due to clang's code generation anomaly
@ 2026-03-16 12:13 Greg Burd <[email protected]>
parent: Michael Paquier <[email protected]>
0 siblings, 1 reply; 4+ messages in thread
From: Greg Burd @ 2026-03-16 12:13 UTC (permalink / raw)
To: Tom Lane <[email protected]>; Michael Paquier <[email protected]>; Alexander Lakhin <[email protected]>; +Cc: pgsql-hackers; Thomas Munro <[email protected]>
On Sun, Mar 15, 2026, at 9:40 PM, Tom Lane wrote:
> Michael Paquier <[email protected]> writes:
>> On Sun, Mar 15, 2026 at 02:00:00PM +0200, Alexander Lakhin wrote:
>>> adding -march=rv64gcv to CFLAGS makes clang generate different (wrong?)
>>> code for crypt-des.c. Namely, having reproduced the diff locally, with
>>> this debugging addition:
>>> when the code compiled with CC=clang-20 CPPFLAGS="-march=rv64gcv".
>
>> Thanks for the detailed investigation. I was puzzled by it myself for
>> a couple of hours, and did not see something else than a broken
>> compiler here..
>
> Seems like this might be worth a bug report to the clang people.
Thanks Alexander for digging and a solid report of the issue, and Michael for spending some time on this, and Tom for chiming in as well. I agree this looks like a compiler issue. I know that Thomas (added/CC'ed) you've posted bugs against clang before, any thoughts on this one?
>
> regards, tom lane
best.
-greg
^ permalink raw reply [nested|flat] 4+ messages in thread
* Re: pgcrypto/des tests fail on riscv64 due to clang's code generation anomaly
@ 2026-03-16 18:00 Alexander Lakhin <[email protected]>
parent: Greg Burd <[email protected]>
0 siblings, 0 replies; 4+ messages in thread
From: Alexander Lakhin @ 2026-03-16 18:00 UTC (permalink / raw)
To: Greg Burd <[email protected]>; Tom Lane <[email protected]>; Michael Paquier <[email protected]>; +Cc: pgsql-hackers; Thomas Munro <[email protected]>
Hello,
16.03.2026 14:13, Greg Burd wrote:
>> Seems like this might be worth a bug report to the clang people.
> Thanks Alexander for digging and a solid report of the issue, and Michael for spending some time on this, and Tom for chiming in as well. I agree this looks like a compiler issue. I know that Thomas (added/CC'ed) you've posted bugs against clang before, any thoughts on this one?
>
Please find attached the reduced test code, which works for me as follows:
$ clang-20 -O2 -march=rv64gcv crypt-des-test.c -o crypt-des-test && ./crypt-des-test
!!!des_init| un_pbox: 15 6 19 20 28 11 27 16 0 14 22 25 4 17 30 9 1 7 23 13 31 26 2 8 18 12 29 5 21 10 3 24
$ clang-20 -O2 -march=rv64gc crypt-des-test.c -o crypt-des-test && ./crypt-des-test
!!!des_init| un_pbox: 8 16 22 30 12 27 1 17 23 15 29 5 25 19 9 0 7 13 24 2 3 28 10 18 31 11 21 6 4 26 14 20
$ clang-20 -O1 -march=rv64gcv crypt-des-test.c -o crypt-des-test && ./crypt-des-test
!!!des_init| un_pbox: 8 16 22 30 12 27 1 17 23 15 29 5 25 19 9 0 7 13 24 2 3 28 10 18 31 11 21 6 4 26 14 20
$ clang-20 --version
Ubuntu clang version 20.1.2 (0ubuntu1~24.04.2)
Best regards,
Alexander
Attachments:
[text/x-csrc] crypt-des-test.c (741B, 2-crypt-des-test.c)
download | inline:
#include <stdio.h>
#define uint8 unsigned char
#define uint32 unsigned int
#define pg_compiler_barrier() __asm__ __volatile__("" ::: "memory")
static uint8 un_pbox[32];
static uint8 pbox[32] = {
16, 7, 20, 21, 29, 12, 28, 17, 1, 15, 23, 26, 5, 18, 31, 10,
2, 8, 24, 14, 32, 27, 3, 9, 19, 13, 30, 6, 22, 11, 4, 25
};
static void
des_init(void)
{
int i;
/*
* Invert the P-box permutation, and convert into OR-masks for handling
* the output of the S-box arrays setup above.
*/
for (i = 0; i < 32; i++)
{
un_pbox[pbox[i] - 1] = i;
//pg_compiler_barrier();
}
fprintf(stderr, "!!!des_init| un_pbox:");
for (i = 0; i < 32; i++)
fprintf(stderr, " %d", un_pbox[i]);
fprintf(stderr, "\n");
}
int main()
{
des_init();
}
^ permalink raw reply [nested|flat] 4+ messages in thread
end of thread, other threads:[~2026-03-16 18:00 UTC | newest]
Thread overview: 4+ messages (download: mbox mbox.gz follow: Atom feed)
-- links below jump to the message on this page --
2026-03-15 12:00 pgcrypto/des tests fail on riscv64 due to clang's code generation anomaly Alexander Lakhin <[email protected]>
2026-03-16 01:08 ` Michael Paquier <[email protected]>
2026-03-16 12:13 ` Greg Burd <[email protected]>
2026-03-16 18:00 ` Alexander Lakhin <[email protected]>
This inbox is served by agora; see mirroring instructions
for how to clone and mirror all data and code used for this inbox