[Bug c/110640] New: Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[Bug c/110640] New: Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[shaohua.li at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

            Bug ID: 110640
           Summary: Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: shaohua.li at inf dot ethz.ch
  Target Milestone: ---

This looks like a regression since GCC-13. gcc produces wrong code at -O2/3.

Compiler explorer: https://godbolt.org/z/5MWe6eK8M

$ cat a.c
unsigned short a = 65535;
int b, f, g;
int *c = &b;
long d;
short e;
static int *h(int);
void i() { h(a); }
int *h(int j) {
  unsigned char k;
  for (; e != 8; e = e + 4)
    k = 0;
  for (; (unsigned char)(j-181249535) + k <= 1; k++) {
    *c = d;
    for (; f; f++)
      ;
  }
  return &g;
}
int main() { i(); }
$
$ gcc-tk -O0 a.c && ./a.out
$
$ gcc-tk -O2 a.c
$ ./a.out
Killed
$ gcc-tk -O3 a.c
$ ./a.out
Killed
$
$ gcc-tk -v
Using built-in specs.
COLLECT_GCC=gcc-tk
COLLECT_LTO_WRAPPER=/zdata/shaoli/compilers/ccbuilder-compilers/gcc-322d17ae51ea0137167424e0018d7fa355948f9f/libexec/gcc/x86_64-pc-linux-gnu/14.0.0/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: ../configure --disable-multilib --disable-bootstrap
--enable-languages=c,c++
--prefix=/zdata/shaoli/compilers/ccbuilder-compilers/gcc-322d17ae51ea0137167424e0018d7fa355948f9f
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 14.0.0 20230711 (experimental) (GCC)
$

[Bug c/110640] [13/14 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |13.2
            Summary|Wrong code at -O2/3 on      |[13/14 Regression] Wrong
                   |x86_64-linux-gnu since      |code at -O2/3 on
                   |GCC-13                      |x86_64-linux-gnu since
                   |                            |GCC-13
           Keywords|                            |needs-bisection, wrong-code

[Bug tree-optimization/110640] [13/14 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rguenth at gcc dot gnu.org
   Last reconfirmed|                            |2023-07-12
          Component|c                           |tree-optimization
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  Adding -fno-tree-vrp avoids the issue.

[Bug tree-optimization/110640] [13/14 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #1)
> Confirmed.  Adding -fno-tree-vrp avoids the issue.

Using -fdisable-tree-vrp1 avoids the issue ...

[Bug tree-optimization/110640] [13/14 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
We end up with

  <bb 10> [local count: 78745003]:

  <bb 11> [local count: 78745003]:
  *c.5_6 = _7;
  f.7_32 = f;
  if (f.7_32 != 0)
    goto <bb 12>; [89.00%]
  else
    goto <bb 10>; [11.00%]

  <bb 12> [local count: 70083053]:
  f = 0;
  goto <bb 10>; [100.00%]

disabling loop header copying (-fdisable-tree-ch2) also avoids the issue.

[Bug tree-optimization/110640] [13/14 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|13.2                        |13.3

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 13.2 is being released, retargeting bugs to GCC 13.3.

[Bug tree-optimization/110640] [13/14 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[shaohua.li at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

--- Comment #5 from Shaohua Li <shaohua.li at inf dot ethz.ch> ---
Bisect to r12-1653-gcb448ade74d

[Bug tree-optimization/110640] [13/14 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|needs-bisection             |
           Priority|P3                          |P2
                 CC|                            |aldyh at gcc dot gnu.org,
                   |                            |amacleod at redhat dot com

[Bug tree-optimization/110640] [13/14 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
This doesn't reproduce on the trunk since
r14-4790-g9692309ed6b625f0fb358c0e230404b5603f69a6 anymore.  Was that a fix for
this or just made it latent?

[Bug tree-optimization/110640] [13 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13 since r12-1653-gcb448ade74d

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |14.0
            Summary|[13/14 Regression] Wrong    |[13 Regression] Wrong code
                   |code at -O2/3 on            |at -O2/3 on
                   |x86_64-linux-gnu since      |x86_64-linux-gnu since
                   |GCC-13 since                |GCC-13 since
                   |r12-1653-gcb448ade74d       |r12-1653-gcb448ade74d
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org
         Depends on|                            |111445
             Status|NEW                         |ASSIGNED

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #6)
> This doesn't reproduce on the trunk since
> r14-4790-g9692309ed6b625f0fb358c0e230404b5603f69a6 anymore.  Was that a fix
> for this or just made it latent?

It does look similar enough:

   <bb 7> [local count: 118111600]:
   k_11 = k_38 + 1;
-  _13 = (unsigned char) j_12(D);
-  _14 = _13 + 1;
-  _15 = (int) _14;
+  _13 = _26;
+  _14 = _28;
+  _15 = _29;
   _16 = (int) k_11;
-  _17 = _15 + _16;
+  _17 = _29 + 1;

I'll add this testcase as well.


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111445
[Bug 111445] [12 Regression] Wrong code at -Os on x86_64-linux-gnu since
r12-1077-g57bf3751511

[Bug tree-optimization/110640] [13 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13 since r12-1653-gcb448ade74d

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
      Known to work|                            |13.2.1
         Resolution|---                         |FIXED

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
Verified fixed on the branch (patch was backported).

[Bug tree-optimization/110640] [13 Regression] Wrong code at -O2/3 on x86_64-linux-gnu since GCC-13 since r12-1653-gcb448ade74d

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110640

--- Comment #9 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Richard Biener <rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:9e20779a7d8e14e77df6efaeadc6bf3a001336ef

commit r14-6532-g9e20779a7d8e14e77df6efaeadc6bf3a001336ef
Author: Richard Biener <rguenther@suse.de>
Date:   Thu Dec 14 08:40:00 2023 +0100

    tree-optimization/110640 - testcase for fixed bug

            PR tree-optimization/110640
            * gcc.dg/torture/pr110640.c: New testcase.