[Bug c++/110444] New: ice in real_can_shorten_arithmetic, at real.cc:1398

[Bug c++/110444] New: ice in real_can_shorten_arithmetic, at real.cc:1398

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

            Bug ID: 110444
           Summary: ice in real_can_shorten_arithmetic, at real.cc:1398
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
  Target Milestone: ---

Created attachment 55408
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=55408&action=edit
gzipped C++ source code

The attached C++ code does this with recent gcc trunk:

$ ~/gcc/results/bin/gcc -c -w -O2 -std=c++11 bug934.cc 
during GIMPLE pass: fre
In file included from
/home/dcb38/rpmbuild/BUILD/vigra-1.11.1/vigranumpy/src/cor
e/non_local_mean.cxx:41:
/home/dcb38/rpmbuild/BUILD/vigra-1.11.1/include/vigra/non_local_mean.hxx: In
fun
ction ‘void vigra::gaussianMeanAndVariance(const MultiArrayView<DIM,
PIXEL_TYPE_
IN>&, double, MultiArrayView<DIM, PIXEL_TYPE_OUT>&, MultiArrayView<DIM,
PIXEL_TY
PE_OUT>&, MultiArrayView<DIM, PIXEL_TYPE_OUT>&) [with int DIM = 2;
PIXEL_TYPE_IN
 = TinyVector<float, 3>; SMOOTH_POLICY = TinyVector<float, 3>; PIXEL_TYPE_OUT = 
TinyVector<float, 3>]’:
/home/dcb38/rpmbuild/BUILD/vigra-1.11.1/include/vigra/non_local_mean.hxx:799:13:
 internal compiler error: in real_can_shorten_arithmetic, at real.cc:1398
  799 | inline void gaussianMeanAndVariance(
      |             ^~~~~~~~~~~~~~~~~~~~~~~
0x107bebe real_can_shorten_arithmetic(machine_mode, machine_mode)
        ../../trunk.year/gcc/real.cc:1398

[Bug middle-end/110444] ice in real_can_shorten_arithmetic, at real.cc:1398

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c++                         |middle-end
           Keywords|                            |ice-on-valid-code

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
0x8df39f real_can_shorten_arithmetic(machine_mode, machine_mode)
        /home/apinski/src/upstream-gcc-git/gcc/gcc/real.cc:1398
0x216236b gimple_simplify_123(gimple_match_op*, gimple**, tree_node*
(*)(tree_node*), tree_node*, tree_node**, tree_code)
       
/home/apinski/src/upstream-gcc-git/gcc/objdir/gcc/gimple-match-2.cc:1071
0x215a036 gimple_simplify_CONVERT_EXPR(gimple_match_op*, gimple**, tree_node*
(*)(tree_node*), code_helper, tree_node*, tree_node*)
       
/home/apinski/src/upstream-gcc-git/gcc/objdir/gcc/gimple-match-1.cc:4814
0x186c971 gimple_resimplify1
        /home/apinski/src/upstream-gcc-git/gcc/gcc/gimple-match-exports.cc:890
0x186d427 gimple_simplify(gimple*, gimple_match_op*, gimple**, tree_node*
(*)(tree_node*), tree_node* (*)(tree_node*))
        /home/apinski/src/upstream-gcc-git/gcc/gcc/gimple-match-exports.cc:834
0xf67c6e gimple_fold_stmt_to_constant_1(gimple*, tree_node* (*)(tree_node*),
tree_node* (*)(tree_node*))
        /home/apinski/src/upstream-gcc-git/gcc/gcc/gimple-fold.cc:7472
0x149bbb4 try_to_simplify
        /home/apinski/src/upstream-gcc-git/gcc/gcc/tree-ssa-sccvn.cc:6096
0x149bbb4 visit_stmt
        /home/apinski/src/upstream-gcc-git/gcc/gcc/tree-ssa-sccvn.cc:6139
0x149c6b5 process_bb
        /home/apinski/src/upstream-gcc-git/gcc/gcc/tree-ssa-sccvn.cc:7945
0x149e1c9 do_rpo_vn_1
        /home/apinski/src/upstream-gcc-git/gcc/gcc/tree-ssa-sccvn.cc:8544
0x149fb38 execute
        /home/apinski/src/upstream-gcc-git/gcc/gcc/tree-ssa-sccvn.cc:8702
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

[Bug middle-end/110444] ice in real_can_shorten_arithmetic, at real.cc:1398

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Reducing ...

[Bug middle-end/110444] ice in real_can_shorten_arithmetic, at real.cc:1398

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
r14-2041

[Bug middle-end/110444] ice in real_can_shorten_arithmetic, at real.cc:1398

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

--- Comment #4 from David Binderman <dcb314 at hotmail dot com> ---
The bug seems to appear sometime between g:f10a4ce044a62c92
and g:319e6fe96027f537, which is 173 commits.

[Bug middle-end/110444] [14 Regression] ice in real_can_shorten_arithmetic, at real.cc:1398

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |14.0
             Target|                            |x86_64-linux-gnu
            Summary|ice in                      |[14 Regression] ice in
                   |real_can_shorten_arithmetic |real_can_shorten_arithmetic
                   |, at real.cc:1398           |, at real.cc:1398
            Version|unknown                     |14.0

[Bug middle-end/110444] [14 Regression] ice in real_can_shorten_arithmetic, at real.cc:1398

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Note this might be already fixed by:
https://gcc.gnu.org/pipermail/gcc-patches/2023-June/622984.html
https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=d915762ea9043da858d388b60b2d8093ff77eeab

Once I reduce the testcase, I will retest it to see if it has been fixed or
not.

[Bug middle-end/110444] [14 Regression] ice in real_can_shorten_arithmetic, at real.cc:1398

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2023-06-28

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Reduced testcase:
```
void f(float *a, float *b, float *c, int size)
{
  float t[2];
  t[0] = b[0] - (float)__builtin_pow(c[0], 2);
  t[1] = b[1] - (float)__builtin_pow(c[1], 2);
  a[0] = t[0];
  a[1] = t[1];
}
```

[Bug middle-end/110444] [14 Regression] ice in real_can_shorten_arithmetic, at real.cc:1398

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

--- Comment #7 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #5)
> Note this might be already fixed by:
> https://gcc.gnu.org/pipermail/gcc-patches/2023-June/622984.html
> https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;
> h=d915762ea9043da858d388b60b2d8093ff77eeab
> 
> Once I reduce the testcase, I will retest it to see if it has been fixed or
> not.

Yes it was, going to add a testcase and close the bug as fixed.

[Bug middle-end/110444] [14 Regression] ice in real_can_shorten_arithmetic, at real.cc:1398

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Andrew Pinski <pinskia@gcc.gnu.org>:

https://gcc.gnu.org/g:857e1f93ff8e3b93a7a3dcca9e50fe32a4c93950

commit r14-2151-g857e1f93ff8e3b93a7a3dcca9e50fe32a4c93950
Author: Andrew Pinski <apinski@marvell.com>
Date:   Wed Jun 28 00:21:08 2023 -0700

    Add testcase for PR 110444

    This testcase was fixed after r14-2135-gd915762ea9043da85 and
    there was no testcase for it before so adding one is a good thing.

    Committed as obvious after testing the testcase to make sure it works.

    gcc/testsuite/ChangeLog:

            PR tree-optimization/110444
            * gcc.c-torture/compile/pr110444-1.c: New test.

[Bug middle-end/110444] [14 Regression] ice in real_can_shorten_arithmetic, at real.cc:1398

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110444

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #9 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Fixed.