[Bug target/115028] New: [15 regression] gcc.target/i386/pr101950-2.c FAILs

[Bug target/115028] New: [15 regression] gcc.target/i386/pr101950-2.c FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

            Bug ID: 115028
           Summary: [15 regression] gcc.target/i386/pr101950-2.c FAILs
           Product: gcc
           Version: 15.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ro at gcc dot gnu.org
  Target Milestone: ---
            Target: i386-pc-solaris2.11

Between 20240506 (80c03ac8041340b29325f86ed58ea8bd40a55b99) and 20240507
(bf10f0db20db1598157505b373098bc93c66b915),
the 64-bit gcc.target/i386/pr101950-2.c test regressed on Solaris/x86:

FAIL: gcc.target/i386/pr101950-2.c scan-assembler-times \\txor[ql]\\t 2

There are 3 instances of xor[ql] now instead of the expected two:

        xorq    %rax, %rdi
        xorl    %eax, %edi
        xorl    %eax, %eax

where the gcc-14 branch has

        xorq    %rdi, %rax
        xorl    %edi, %eax

The additional one is from

@@ -20,8 +20,9 @@
 .LFB1:
        movl    %edi, %eax
        sarl    $31, %eax
-       xorl    %edi, %eax
-       lzcntl  %eax, %eax
+       xorl    %eax, %edi
+       xorl    %eax, %eax
+       lzcntl  %edi, %eax
        subl    $1, %eax
        ret
 .LFE1:

I'm attaching both gcc-14 and trunk assembler output.

[Bug target/115028] [15 regression] gcc.target/i386/pr101950-2.c FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

--- Comment #1 from Rainer Orth <ro at gcc dot gnu.org> ---
Created attachment 58170
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=58170&action=edit
trunk assembler input

[Bug target/115028] [15 regression] gcc.target/i386/pr101950-2.c FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

--- Comment #2 from Rainer Orth <ro at gcc dot gnu.org> ---
Created attachment 58171
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=58171&action=edit
gcc-14 assembler input

[Bug target/115028] [15 regression] gcc.target/i386/pr101950-2.c FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

Rainer Orth <ro at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|i386-pc-solaris2.11         |i386-pc-solaris2.11,
                   |                            |x86_64-pc-linux-gnu

--- Comment #3 from Rainer Orth <ro at gcc dot gnu.org> ---
I'm seeing this on Linux/x86_64, too.

[Bug target/115028] [15 regression] gcc.target/i386/pr101950-2.c FAILs

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |testsuite-fail
   Target Milestone|---                         |15.0

[Bug target/115028] [15 regression] gcc.target/i386/pr101950-2.c FAILs

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

Rainer Orth <ro at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rguenth at gcc dot gnu.org

--- Comment #4 from Rainer Orth <ro at gcc dot gnu.org> ---
A reghunt identified

commit 9dbff9c05520a74e6cd337578f27b56c941f64f3
Author: Richard Biener <rguenther@suse.de>
Date:   Tue May 7 10:14:19 2024 +0200

    Revert "Revert "combine: Don't combine if I2 does not change""

as the culprit.

[Bug target/115028] [15 regression] gcc.target/i386/pr101950-2.c FAILs

[haochen.jiang at intel dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

Haochen Jiang <haochen.jiang at intel dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |haochen.jiang at intel dot com

--- Comment #5 from Haochen Jiang <haochen.jiang at intel dot com> ---
I suppose Richard should have known the issue.
See: https://gcc.gnu.org/pipermail/gcc-regression/2024-May/079828.html

[Bug target/115028] [15 regression] gcc.target/i386/pr101950-2.c FAILs

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2024-05-15
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Haochen Jiang from comment #5)
> I suppose Richard should have known the issue.
> See: https://gcc.gnu.org/pipermail/gcc-regression/2024-May/079828.html

Yep, this and others were the reason the change was reverted against the
authors wish for GCC 14 but I promised to re-instantiate in stage1.  That
means it's Segher who's intending to fix this and the rest of the fallout
in a more appropriate way that wasn't suitable for late stage4.

But confirmed, I also see this.

[Bug target/115028] [15 regression] gcc.target/i386/pr101950-2.c FAILs

[lingling.kong7 at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115028

kong lingling <lingling.kong7 at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |lingling.kong7 at gmail dot com

--- Comment #7 from kong lingling <lingling.kong7 at gmail dot com> ---
Just mention and tag commit 9dbff9c05520a74e6cd337578f27b56c941f64f3 also
causes another case FAIL.
FAIL: gcc.target/i386/apx-ndd-tls-1b.c scan-assembler-times addq[
\t]+%r[a-z0-9]+, a@gottpoff\\(%rip\\), %r[a-z0-9]+ 1