[Bug c++/107029] New: ice during GIMPLE pass: reassoc

[Bug c++/107029] New: ice during GIMPLE pass: reassoc

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107029

            Bug ID: 107029
           Summary: ice during GIMPLE pass: reassoc
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
  Target Milestone: ---

This C++ code:

class ImplicitBool ImplicitBool::*Reset___trans_tmp_1;
int ImplicitBool::*Reset___trans_tmp_2;
struct Chips {
  struct Clocks {
    void Reset(bool, bool);
    int next;
  };
  void Reset();
  Clocks clocks;
};
void Chips::Clocks::Reset(bool mmc5Chip, bool fdsChip) {
  next = mmc5Chip || fdsChip;
}
void Chips::Reset() { clocks.Reset(Reset___trans_tmp_1, Reset___trans_tmp_2); }

when compiled as follows by recent g++

$ ../results/bin/g++ -c -w -O2 -std=c++98 bug849.cc

does this:

bug849.cc: In member function ‘void Chips::Reset()’:
bug849.cc:14:6: error: invalid types for ‘bit_and_expr’
   14 | void Chips::Reset() { clocks.Reset(Reset___trans_tmp_1,
Reset___trans_tmp_2); }
      |      ^~~~~
<<< Unknown tree: offset_type >>>
<<< Unknown tree: offset_type >>>
<<< Unknown tree: offset_type >>>
_11 = _10 & Reset___trans_tmp_1.1_3;
during GIMPLE pass: reassoc
bug849.cc:14:6: internal compiler error: verify_gimple failed

The bug first seems to occur sometime between git hash 7d7e2149cdb6b0c3
and 9a1bc4b463e61cd5, a day later. There are 24 commits in that range.

[Bug c++/107029] [13 Regression] ice during GIMPLE pass: reassoc

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107029

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2022-09-24
     Ever confirmed|0                           |1
            Version|12.0                        |13.0
            Summary|ice during GIMPLE pass:     |[13 Regression] ice during
                   |reassoc                     |GIMPLE pass: reassoc
   Target Milestone|---                         |13.0
                 CC|                            |jakub at gcc dot gnu.org
             Status|UNCONFIRMED                 |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org
           Priority|P3                          |P1

[Bug tree-optimization/107029] [13 Regression] ice during GIMPLE pass: reassoc

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107029

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 53624
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53624&action=edit
gcc13-pr107029.patch

Untested fix.

[Bug tree-optimization/107029] [13 Regression] ice during GIMPLE pass: reassoc

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107029

--- Comment #2 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:cb8f25c5dc9f6d5207c826c2dafe25f68458ceaf

commit r13-2891-gcb8f25c5dc9f6d5207c826c2dafe25f68458ceaf
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue Sep 27 08:26:18 2022 +0200

    reassoc: Handle OFFSET_TYPE like POINTER_TYPE in
optimize_range_tests_cmp_bitwise [PR107029[

    As the testcase shows, OFFSET_TYPE needs the same treatment as
    POINTER_TYPE/REFERENCE_TYPE, otherwise we fail the same during the
    newly added verification.  OFFSET_TYPE is signed though, so unlike
    POINTER_TYPE/REFERENCE_TYPE it can also trigger with the
    x < 0 && y < 0 && z < 0 to (x | y | z) < 0
    optimization.

    2022-09-27  Jakub Jelinek  <jakub@redhat.com>

            PR tree-optimization/107029
            * tree-ssa-reassoc.cc (optimize_range_tests_cmp_bitwise): Treat
            OFFSET_TYPE like POINTER_TYPE, except that OFFSET_TYPE may be
            signed and so can trigger even the (b % 4) == 3 case.

            * g++.dg/torture/pr107029.C: New test.

[Bug tree-optimization/107029] [13 Regression] ice during GIMPLE pass: reassoc

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107029

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

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

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Should be fixed now.