[Bug c/95108] New: [10/11 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[Bug c/95108] New: [10/11 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

            Bug ID: 95108
           Summary: [10/11 Regression] ICE in tree_fits_uhwi_p, at
                    tree.c:7292
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Changed between 20190811 and 20190818, with -fno-tree-ccp :
(test case derived from pr88588.c)


$ cat z1.c
int *v;

#pragma omp declare simd
void
foo (int x)
{
  int *a = &x + 1;

  for (;;)
    {
      *v = *a;
      a = v;
    }
}


$ gcc-10-20190811 -c z1.c -O2 -fopenmp -fno-tree-ccp
$
$ gcc-11-20200510 -c z1.c -O2 -fopenmp -fno-tree-ccp
during GIMPLE pass: alias
z1.c: In function 'foo.simdclone.0':
z1.c:5:1: internal compiler error: Segmentation fault
    5 | foo (int x)
      | ^~~
0xb0ed8f crash_signal
        ../../gcc/toplev.c:328
0xd62e47 tree_fits_uhwi_p(tree_node const*)
        ../../gcc/tree.c:7292
0xcb74ad create_variable_info_for_1
        ../../gcc/tree-ssa-structalias.c:6068
0xcb7c43 create_variable_info_for
        ../../gcc/tree-ssa-structalias.c:6251
0xcb7c43 get_vi_for_tree
        ../../gcc/tree-ssa-structalias.c:2939
0xcb839f get_constraint_for_ssa_var
        ../../gcc/tree-ssa-structalias.c:3031
0xcb8902 get_constraint_for_1
        ../../gcc/tree-ssa-structalias.c:3611
0xcb97c4 get_constraint_for_ptr_offset
        ../../gcc/tree-ssa-structalias.c:3166
0xcb8939 get_constraint_for_1
        ../../gcc/tree-ssa-structalias.c:3544
0xcb9172 get_constraint_for_address_of
        ../../gcc/tree-ssa-structalias.c:3462
0xcb8b6b get_constraint_for_1
        ../../gcc/tree-ssa-structalias.c:3530
0xcbce95 find_func_aliases
        ../../gcc/tree-ssa-structalias.c:5017
0xcbdafe compute_points_to_sets
        ../../gcc/tree-ssa-structalias.c:7380
0xcbdafe compute_may_aliases()
        ../../gcc/tree-ssa-structalias.c:7861
0xa4524e execute_function_todo
        ../../gcc/passes.c:1957
0xa46192 execute_todo
        ../../gcc/passes.c:2039


A test version (configured with --enable-checking=yes)
does not need that extra option :

$ gcc-11-20200510-chk -c z1.c -O2 -fopenmp
z1.c: In function 'foo.simdclone.7':
z1.c:14:1: error: definition in block 6 follows the use
   14 | }
      | ^
for SSA_NAME: _8 in statement:
_9 = &MEM <int> [(void *)_8 + 4B];
during IPA pass: simdclone
z1.c:14:1: internal compiler error: verify_ssa failed
0xf3faab verify_ssa(bool, bool)
        ../../gcc/tree-ssa.c:1208
0xbe8827 execute_function_todo
        ../../gcc/passes.c:1992
0xbe952d do_per_function
        ../../gcc/passes.c:1647
0xbe9592 execute_todo
        ../../gcc/passes.c:2039

[Bug c/95108] [9/10/11 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2020-05-13
                 CC|                            |jakub at gcc dot gnu.org
     Ever confirmed|0                           |1
   Target Milestone|---                         |9.4
             Status|UNCONFIRMED                 |NEW
            Summary|[10/11 Regression] ICE in   |[9/10/11 Regression] ICE in
                   |tree_fits_uhwi_p, at        |tree_fits_uhwi_p, at
                   |tree.c:7292                 |tree.c:7292

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r9-6508-g33813f1d703c95d4fc87d16a17f6c834135ab209

[Bug middle-end/95108] [9/10/11 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org
             Status|NEW                         |ASSIGNED

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 48523
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=48523&action=edit
gcc11-pr95108.patch

Untested fix.

[Bug middle-end/95108] [9/10/11 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

[Bug middle-end/95108] [9/10/11 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

--- Comment #3 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:d0fb9ffc1b8f3b86bbdf0e915cec2136141b329b

commit r11-381-gd0fb9ffc1b8f3b86bbdf0e915cec2136141b329b
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu May 14 09:51:05 2020 +0200

    openmp: Fix placement of 2nd+ preparation statement for PHIs in simd clone
lowering [PR95108]

    For normal stmts, preparation statements are inserted before the stmt, so
if we need multiple,
    they are in the correct order, but for PHIs we emit them after labels in
the entry successor
    bb, and we used to emit them in the reverse order that way.

    2020-05-14  Jakub Jelinek  <jakub@redhat.com>

            PR middle-end/95108
            * omp-simd-clone.c (struct modify_stmt_info): Add after_stmt
member.
            (ipa_simd_modify_stmt_ops): For PHIs, only add before first stmt in
            entry block if info->after_stmt is NULL, otherwise add after that
stmt
            and update it after adding each stmt.
            (ipa_simd_modify_function_body): Initialize info.after_stmt.

            * gcc.dg/gomp/pr95108.c: New test.

[Bug middle-end/95108] [9/10 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10/11 Regression] ICE in |[9/10 Regression] ICE in
                   |tree_fits_uhwi_p, at        |tree_fits_uhwi_p, at
                   |tree.c:7292                 |tree.c:7292

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed on the trunk so far.

[Bug middle-end/95108] [9/10 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:e09d626551fef9b5719073427170d11e97038c44

commit r10-8293-ge09d626551fef9b5719073427170d11e97038c44
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu May 14 09:51:05 2020 +0200

    openmp: Fix placement of 2nd+ preparation statement for PHIs in simd clone
lowering [PR95108]

    For normal stmts, preparation statements are inserted before the stmt, so
if we need multiple,
    they are in the correct order, but for PHIs we emit them after labels in
the entry successor
    bb, and we used to emit them in the reverse order that way.

    2020-05-14  Jakub Jelinek  <jakub@redhat.com>

            PR middle-end/95108
            * omp-simd-clone.c (struct modify_stmt_info): Add after_stmt
member.
            (ipa_simd_modify_stmt_ops): For PHIs, only add before first stmt in
            entry block if info->after_stmt is NULL, otherwise add after that
stmt
            and update it after adding each stmt.
            (ipa_simd_modify_function_body): Initialize info.after_stmt.

            * gcc.dg/gomp/pr95108.c: New test.

    (cherry picked from commit d0fb9ffc1b8f3b86bbdf0e915cec2136141b329b)

[Bug middle-end/95108] [9 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10 Regression] ICE in    |[9 Regression] ICE in
                   |tree_fits_uhwi_p, at        |tree_fits_uhwi_p, at
                   |tree.c:7292                 |tree.c:7292

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 10.2+ too.

[Bug middle-end/95108] [9 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:b78ea59edbd354fed914d0f8eb78109db7316ce8

commit r9-8899-gb78ea59edbd354fed914d0f8eb78109db7316ce8
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu May 14 09:51:05 2020 +0200

    openmp: Fix placement of 2nd+ preparation statement for PHIs in simd clone
lowering [PR95108]

    For normal stmts, preparation statements are inserted before the stmt, so
if we need multiple,
    they are in the correct order, but for PHIs we emit them after labels in
the entry successor
    bb, and we used to emit them in the reverse order that way.

    2020-05-14  Jakub Jelinek  <jakub@redhat.com>

            PR middle-end/95108
            * omp-simd-clone.c (struct modify_stmt_info): Add after_stmt
member.
            (ipa_simd_modify_stmt_ops): For PHIs, only add before first stmt in
            entry block if info->after_stmt is NULL, otherwise add after that
stmt
            and update it after adding each stmt.
            (ipa_simd_modify_function_body): Initialize info.after_stmt.

            * gcc.dg/gomp/pr95108.c: New test.

    (cherry picked from commit d0fb9ffc1b8f3b86bbdf0e915cec2136141b329b)

[Bug middle-end/95108] [9 Regression] ICE in tree_fits_uhwi_p, at tree.c:7292

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95108

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

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

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 8.5 in r8-10505-g38484dac88b157151db984790be0582bf66adcfb and by the
above commit for 9.4+ too.