public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757
@ 2021-10-28 18:30 seurer at gcc dot gnu.org
2021-10-29 6:49 ` [Bug target/102991] [12 regression] " rguenth at gcc dot gnu.org
` (9 more replies)
0 siblings, 10 replies; 11+ messages in thread
From: seurer at gcc dot gnu.org @ 2021-10-28 18:30 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
Bug ID: 102991
Summary: [12 regress] gcc.dg/vect/vect-simd-17.c fails after
r12-4757
Product: gcc
Version: 12.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: seurer at gcc dot gnu.org
Target Milestone: ---
g:9222481ffc69a6c0b73ec81e1bf04289fa3db0ed, r12-4757
make -k check-gcc RUNTESTFLAGS="vect.exp=gcc.dg/vect/vect-simd-17.c"
FAIL: gcc.dg/vect/vect-simd-17.c execution test
FAIL: gcc.dg/vect/vect-simd-17.c -flto -ffat-lto-objects execution test
# of expected passes 2
# of unexpected failures 2
I am only seeing this failure on power 10.
(gdb) run
Starting program: /home/seurer/gcc/git/build/gcc-test/vect-simd-17.exe
Missing separate debuginfos, use: yum debuginfo-install
glibc-2.28-151.el8.ppc64le
Program received signal SIGABRT, Aborted.
0x00002000002144d8 in raise () from /lib64/libc.so.6
(gdb) where
#0 0x00002000002144d8 in raise () from /lib64/libc.so.6
#1 0x00002000001f462c in abort () from /lib64/libc.so.6
#2 0x0000000010003578 in doit () at
/home/seurer/gcc/git/gcc-test/gcc/testsuite/gcc.dg/vect/vect-simd-17.c:37
#3 0x0000000010000614 in main () at
/home/seurer/gcc/git/gcc-test/gcc/testsuite/gcc.dg/vect/vect-simd-17.c:302
commit 9222481ffc69a6c0b73ec81e1bf04289fa3db0ed (HEAD, refs/bisect/bad)
Author: Xionghu Luo <luoxhu@linux.ibm.com>
Date: Wed Oct 27 21:21:20 2021 -0500
rs6000: Fix wrong code generation for vec_sel [PR94613]
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
@ 2021-10-29 6:49 ` rguenth at gcc dot gnu.org
2021-10-29 9:11 ` luoxhu at gcc dot gnu.org
` (8 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu.org @ 2021-10-29 6:49 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|[12 regress] |[12 regression]
|gcc.dg/vect/vect-simd-17.c |gcc.dg/vect/vect-simd-17.c
|fails after r12-4757 |fails after r12-4757
Target Milestone|--- |12.0
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
2021-10-29 6:49 ` [Bug target/102991] [12 regression] " rguenth at gcc dot gnu.org
@ 2021-10-29 9:11 ` luoxhu at gcc dot gnu.org
2021-10-29 9:55 ` linkw at gcc dot gnu.org
` (7 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: luoxhu at gcc dot gnu.org @ 2021-10-29 9:11 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
--- Comment #1 from luoxhu at gcc dot gnu.org ---
Couldn't reproduce on rain6p1 (P10):
Test run by luoxhu on Fri Oct 29 04:08:49 2021
Native configuration is powerpc64le-unknown-linux-gnu
=== gcc tests ===
Schedule of variations:
unix
Running target unix
Running /home/luoxhu/workspace/gcc/gcc/testsuite/gcc.dg/vect/vect.exp ...
PASS: gcc.dg/vect/vect-simd-17.c (test for excess errors)
PASS: gcc.dg/vect/vect-simd-17.c execution test
PASS: gcc.dg/vect/vect-simd-17.c -flto -ffat-lto-objects (test for excess
errors)
PASS: gcc.dg/vect/vect-simd-17.c -flto -ffat-lto-objects execution test
=== gcc Summary ===
# of expected passes 4
/home/luoxhu/workspace/build/gcc/xgcc version 12.0.0 20211029 (experimental)
(GCC)
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
2021-10-29 6:49 ` [Bug target/102991] [12 regression] " rguenth at gcc dot gnu.org
2021-10-29 9:11 ` luoxhu at gcc dot gnu.org
@ 2021-10-29 9:55 ` linkw at gcc dot gnu.org
2021-11-01 5:03 ` luoxhu at gcc dot gnu.org
` (6 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: linkw at gcc dot gnu.org @ 2021-10-29 9:55 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
Kewen Lin <linkw at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |linkw at gcc dot gnu.org
--- Comment #2 from Kewen Lin <linkw at gcc dot gnu.org> ---
(In reply to luoxhu from comment #1)
> Couldn't reproduce on rain6p1 (P10):
>
It's weird, I can reproduce this on rain6p1.
FAIL: gcc.dg/vect/vect-simd-17.c execution test
FAIL: gcc.dg/vect/vect-simd-17.c -flto -ffat-lto-objects execution test
>------->-------=== gcc Summary ===
# of expected passes>--->-------2
# of unexpected failures>-------2
Probably due to you still specified --with-cpu=power9 instead of
--with-cpu=power10 in gcc configuration?
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
` (2 preceding siblings ...)
2021-10-29 9:55 ` linkw at gcc dot gnu.org
@ 2021-11-01 5:03 ` luoxhu at gcc dot gnu.org
2021-11-02 8:18 ` luoxhu at gcc dot gnu.org
` (5 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: luoxhu at gcc dot gnu.org @ 2021-11-01 5:03 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
--- Comment #3 from luoxhu at gcc dot gnu.org ---
(In reply to Kewen Lin from comment #2)
> (In reply to luoxhu from comment #1)
> > Couldn't reproduce on rain6p1 (P10):
> >
>
> It's weird, I can reproduce this on rain6p1.
>
> FAIL: gcc.dg/vect/vect-simd-17.c execution test
> FAIL: gcc.dg/vect/vect-simd-17.c -flto -ffat-lto-objects execution test
>
> >------->-------=== gcc Summary ===
>
> # of expected passes>--->-------2
> # of unexpected failures>-------2
>
> Probably due to you still specified --with-cpu=power9 instead of
> --with-cpu=power10 in gcc configuration?
Thanks, confirmed. --with-cpu=power9 doesn't fail on both P9 and P10 with the
patch.
It aborts at vect-simd-17.c of line 274.
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
` (3 preceding siblings ...)
2021-11-01 5:03 ` luoxhu at gcc dot gnu.org
@ 2021-11-02 8:18 ` luoxhu at gcc dot gnu.org
2021-11-02 8:36 ` luoxhu at gcc dot gnu.org
` (4 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: luoxhu at gcc dot gnu.org @ 2021-11-02 8:18 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
--- Comment #4 from luoxhu at gcc dot gnu.org ---
vect-simd-17.p10.c.335r.final:
3379: %v1:V16QI=unspec[%v1:V16QI,%v1:V16QI,%v9:V16QI] 254
3372: {%v11:V4SI=~%v0:V4SI&%v13:V4SI|%v11:V4SI;clobber %r10:V4SI;} // wrong
code.
REG_DEAD %v0:V4SI
REG_UNUSED %r10:V4SI
3373: [%r1:DI+0x80]=%v11:V4SI
ASM:
.L149:
lxvx %vs32,%r9,%r8
vadduwm %v12,%v12,%v13
mfvsrd %r5,%vs42
mfvsrld %r4,%vs42
vadduwm %v10,%v10,%v8
stxv %vs44,112(%r1)
xxperm %vs32,%vs32,%vs41
vadduwm %v1,%v13,%v0
vcmpequw %v0,%v0,%v13
xxperm %vs33,%vs33,%vs41
vandc %r10,%v13,%v0 // wrong code
vor %v11,%r10,%v11 // wrong code
stxv %vs43,128(%r1)
stxvx %vs33,%r9,%r8
addi %r8,%r8,-16
bdnz .L149
But the binary is (/opt/binutils-power10/bin/objdump -d vect-simd-17.p10 |
less):
10002ea0: 19 42 09 7c lxvx vs32,r9,r8
10002ea4: 80 68 8c 11 vadduwm v12,v12,v13
10002ea8: 67 00 45 7d mfvrd r5,v10
10002eac: 67 02 44 7d mfvsrld r4,vs42
10002eb0: 80 40 4a 11 vadduwm v10,v10,v8
10002eb4: 7d 00 81 f5 stxv vs44,112(r1)
10002eb8: d7 48 00 f0 xxperm vs32,vs32,vs41
10002ebc: 80 00 2d 10 vadduwm v1,v13,v0
10002ec0: 86 68 00 10 vcmpequw v0,v0,v13
10002ec4: d7 48 21 f0 xxperm vs33,vs33,vs41
10002ec8: 44 04 4d 11 vandc v10,v13,v0 // wrong code
10002ecc: 84 5c 6a 11 vor v11,v10,v11 // wrong code
10002ed0: 8d 00 61 f5 stxv vs43,128(r1)
10002ed4: 19 43 29 7c stxvx vs33,r9,r8
10002ed8: f0 ff 08 39 addi r8,r8,-16
10002edc: c4 ff 00 42 bdnz 10002ea0 <doit+0x26f0>
%vs42 is a global constant data loaded from memory, it was modified at address
0x10002ec8, there r10 is changed to v10 from ASM to binary, which was supposed
to be never change in the loop.
(gdb)
0x0000000010002eb4 <doit+9988>: 7d 00 81 f5 stxv vs44,112(r1)
0x0000000010002eb8 <doit+9992>: d7 48 00 f0 xxperm vs32,vs32,vs41
0x0000000010002ebc <doit+9996>: 80 00 2d 10 vadduwm v1,v13,v0
0x0000000010002ec0 <doit+10000>: 86 68 00 10 vcmpequw v0,v0,v13
0x0000000010002ec4 <doit+10004>: d7 48 21 f0 xxperm vs33,vs33,vs41
=> 0x0000000010002ec8 <doit+10008>: 44 04 4d 11 vandc v10,v13,v0
0x0000000010002ecc <doit+10012>: 84 5c 6a 11 vor v11,v10,v11
0x0000000010002ed0 <doit+10016>: 8d 00 61 f5 stxv vs43,128(r1)
7: $vs42.v4_int32 = {-30, -29, -28, -27}
(gdb) si
0x0000000010002eb4 <doit+9988>: 7d 00 81 f5 stxv vs44,112(r1)
0x0000000010002eb8 <doit+9992>: d7 48 00 f0 xxperm vs32,vs32,vs41
0x0000000010002ebc <doit+9996>: 80 00 2d 10 vadduwm v1,v13,v0
0x0000000010002ec0 <doit+10000>: 86 68 00 10 vcmpequw v0,v0,v13
0x0000000010002ec4 <doit+10004>: d7 48 21 f0 xxperm vs33,vs33,vs41
0x0000000010002ec8 <doit+10008>: 44 04 4d 11 vandc v10,v13,v0
=> 0x0000000010002ecc <doit+10012>: 84 5c 6a 11 vor v11,v10,v11
0x0000000010002ed0 <doit+10016>: 8d 00 61 f5 stxv vs43,128(r1)
7: $vs42.v4_int32 = {0, 0, 0, 0}
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
` (4 preceding siblings ...)
2021-11-02 8:18 ` luoxhu at gcc dot gnu.org
@ 2021-11-02 8:36 ` luoxhu at gcc dot gnu.org
2021-11-05 5:39 ` cvs-commit at gcc dot gnu.org
` (3 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: luoxhu at gcc dot gnu.org @ 2021-11-02 8:36 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
--- Comment #5 from luoxhu at gcc dot gnu.org ---
P9:
.L149:
lxvx %vs32,%r8,%r10
vadduwm %v12,%v12,%v1
mfvsrd %r5,%vs43
mfvsrld %r4,%vs43
vadduwm %v11,%v11,%v9
stxv %vs44,112(%r1)
xxperm %vs32,%vs32,%vs42
vcmpequw %v13,%v0,%v1
vadduwm %v0,%v1,%v0
xxlandc %vs45,%vs33,%vs45 // here.
xxperm %vs32,%vs32,%vs42
xxlor %vs0,%vs0,%vs45
stxvx %vs32,%r8,%r10
stxv %vs0,128(%r1)
addi %r8,%r8,-16
bdnz .L149
$vs43 is not changed by xxlandc
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
` (5 preceding siblings ...)
2021-11-02 8:36 ` luoxhu at gcc dot gnu.org
@ 2021-11-05 5:39 ` cvs-commit at gcc dot gnu.org
2021-11-05 5:44 ` luoxhu at gcc dot gnu.org
` (2 subsequent siblings)
9 siblings, 0 replies; 11+ messages in thread
From: cvs-commit at gcc dot gnu.org @ 2021-11-05 5:39 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Xiong Hu Luo <luoxhu@gcc.gnu.org>:
https://gcc.gnu.org/g:614b39757b8b61f70ac1c666edb7a01a5fc19cd4
commit r12-4930-g614b39757b8b61f70ac1c666edb7a01a5fc19cd4
Author: Xionghu Luo <luoxhu@linux.ibm.com>
Date: Wed Nov 3 20:23:03 2021 -0500
rs6000: Fix incorrect fusion constraint [PR102991]
gcc/ChangeLog:
2021-11-05 Xionghu Luo <luoxhu@linux.ibm.com>
PR target/102991
* config/rs6000/fusion.md: Regenerate.
* config/rs6000/genfusion.pl: Fix incorrect clobber constraint.
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
` (6 preceding siblings ...)
2021-11-05 5:39 ` cvs-commit at gcc dot gnu.org
@ 2021-11-05 5:44 ` luoxhu at gcc dot gnu.org
2021-11-09 1:27 ` cvs-commit at gcc dot gnu.org
2021-11-09 1:28 ` luoxhu at gcc dot gnu.org
9 siblings, 0 replies; 11+ messages in thread
From: luoxhu at gcc dot gnu.org @ 2021-11-05 5:44 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
--- Comment #7 from luoxhu at gcc dot gnu.org ---
Fixed, will backport to gcc-11 in a week.
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
` (7 preceding siblings ...)
2021-11-05 5:44 ` luoxhu at gcc dot gnu.org
@ 2021-11-09 1:27 ` cvs-commit at gcc dot gnu.org
2021-11-09 1:28 ` luoxhu at gcc dot gnu.org
9 siblings, 0 replies; 11+ messages in thread
From: cvs-commit at gcc dot gnu.org @ 2021-11-09 1:27 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Xiong Hu Luo
<luoxhu@gcc.gnu.org>:
https://gcc.gnu.org/g:0c877d1c30ef1c5145817805ab944cddea61ad54
commit r11-9231-g0c877d1c30ef1c5145817805ab944cddea61ad54
Author: Xionghu Luo <luoxhu@linux.ibm.com>
Date: Wed Nov 3 20:23:03 2021 -0500
rs6000: Fix incorrect fusion constraint [PR102991]
gcc/ChangeLog:
2021-11-05 Xionghu Luo <luoxhu@linux.ibm.com>
PR target/102991
* config/rs6000/fusion.md: Regenerate.
* config/rs6000/genfusion.pl: Fix incorrect clobber constraint.
(cherry picked from commit 614b39757b8b61f70ac1c666edb7a01a5fc19cd4)
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug target/102991] [12 regression] gcc.dg/vect/vect-simd-17.c fails after r12-4757
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
` (8 preceding siblings ...)
2021-11-09 1:27 ` cvs-commit at gcc dot gnu.org
@ 2021-11-09 1:28 ` luoxhu at gcc dot gnu.org
9 siblings, 0 replies; 11+ messages in thread
From: luoxhu at gcc dot gnu.org @ 2021-11-09 1:28 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102991
luoxhu at gcc dot gnu.org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution|--- |FIXED
--- Comment #9 from luoxhu at gcc dot gnu.org ---
Fixed and backported to gcc-11.
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2021-11-09 1:28 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-10-28 18:30 [Bug target/102991] New: [12 regress] gcc.dg/vect/vect-simd-17.c fails after r12-4757 seurer at gcc dot gnu.org
2021-10-29 6:49 ` [Bug target/102991] [12 regression] " rguenth at gcc dot gnu.org
2021-10-29 9:11 ` luoxhu at gcc dot gnu.org
2021-10-29 9:55 ` linkw at gcc dot gnu.org
2021-11-01 5:03 ` luoxhu at gcc dot gnu.org
2021-11-02 8:18 ` luoxhu at gcc dot gnu.org
2021-11-02 8:36 ` luoxhu at gcc dot gnu.org
2021-11-05 5:39 ` cvs-commit at gcc dot gnu.org
2021-11-05 5:44 ` luoxhu at gcc dot gnu.org
2021-11-09 1:27 ` cvs-commit at gcc dot gnu.org
2021-11-09 1:28 ` luoxhu at gcc dot gnu.org
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).