[Bug c++/99007] New: [11 Regression] ICE in dominated_by_p, at dominance.c:1124

[Bug c++/99007] New: [11 Regression] ICE in dominated_by_p, at dominance.c:1124

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

            Bug ID: 99007
           Summary: [11 Regression] ICE in dominated_by_p, at
                    dominance.c:1124
           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: ---

Started between 20201018 and 20201108 :


$ cat z1.cc
template <typename T> void bar (T *)
{
  T s[0/0];
  #pragma omp teams distribute parallel for reduction(+:s) allocate(s)
  for (int i = 0; i < 8; i++);
}
void foo ()
{
  long *a;
  bar (a);
}


$ g++-11-20210207 -c z1.cc -fopenmp
z1.cc: In function 'void bar(T*)':
z1.cc:3:8: warning: division by zero [-Wdiv-by-zero]
    3 |   T s[0/0];
      |       ~^~
z1.cc: In instantiation of 'void bar(T*) [with T = long int]':
z1.cc:10:9:   required from here
z1.cc:3:8: warning: division by zero [-Wdiv-by-zero]
during GIMPLE pass: ssa
z1.cc: In function '_Z3barIlEvPT_._omp_fn.0':
z1.cc:11:1: internal compiler error: Segmentation fault
   11 | }
      | ^
0xfe6cdf crash_signal
        ../../gcc/toplev.c:327
0xb6e764 dominated_by_p(cdi_direction, basic_block_def const*, basic_block_def
const*)
        ../../gcc/dominance.c:1124
0x1259775 verify_use
        ../../gcc/tree-ssa.c:892
0x125e462 verify_ssa(bool, bool)
        ../../gcc/tree-ssa.c:1167
0xef8a87 execute_function_todo
        ../../gcc/passes.c:2049
0xef97f2 execute_todo
        ../../gcc/passes.c:2096

[Bug c++/99007] [11 Regression] ICE in dominated_by_p, at dominance.c:1124

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org
   Target Milestone|---                         |11.0
   Last reconfirmed|                            |2021-02-08
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with my r11-4710-g9649031577028310e853025672c71dfb500494f0 change when
allocate clause has been implemented for this.

[Bug c++/99007] [11 Regression] ICE in dominated_by_p, at dominance.c:1124

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Simplified testcase, no need for templates, and no need for UB in the code and
no need for C++:
void
bar (int n)
{
  long s[n];
  #pragma omp teams distribute parallel for reduction(+:s) allocate(s)
  for (int i = 0; i < 8; i++)
    ;
}

[Bug middle-end/99007] [11 Regression] ICE in dominated_by_p, at dominance.c:1124

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
And allocate clause isn't needed either, this ICEs since
r9-3941-g28567c40e2c7c88e424283e8a1a6ff8cb7ba440c
when host teams has been introduced:
int
bar (int n)
{
  int s[n];
  int i, j;
  for (i = 0; i < n; i++)
    s[i] = 0;
  #pragma omp teams distribute parallel for reduction(+:s) private (j)
  for (i = 0; i < 8; i++)
    for (j = 0; j < n; j++)
      s[j] += i;
  return s[0] + s[n - 1];
}
and
int
bar (int n)
{
  int s[n];
  int i, j;
  for (i = 0; i < n; i++)
    s[i] = 0;
  #pragma omp parallel reduction(+:s) num_threads(2)
  #pragma omp parallel for reduction(+:s) private (j)
  for (i = 0; i < 8; i++)
    for (j = 0; j < n; j++)
      s[j] += i;
  return s[0] + s[n - 1];
}
ICEs since r7-536-g381cdae49785fc4b1941a6b37bad3e8c6f3f428d

[Bug middle-end/99007] [11 Regression] ICE in dominated_by_p, at dominance.c:1124

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 50152
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=50152&action=edit
gcc11-pr99007.patch

Untested fix.

[Bug middle-end/99007] [11 Regression] ICE in dominated_by_p, at dominance.c:1124

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

--- Comment #5 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:bd0e37f68a3aed944df4eb739a0734bb87153749

commit r11-7167-gbd0e37f68a3aed944df4eb739a0734bb87153749
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Feb 10 10:34:58 2021 +0100

    openmp: Temporarily disable into_ssa when gimplifying OpenMP reduction
clauses [PR99007]

    gimplify_scan_omp_clauses was already calling gimplify_expr with false as
    last argument to make sure it is not an SSA_NAME, but as the testcases
show,
    that is not enough, SSA_NAME temporaries created during that gimplification
    can be reused too and we can't allow SSA_NAMEs to be used across OpenMP
    region boundaries, as we can only firstprivatize decls.

    Fixed by temporarily disabling into_ssa.

    2021-02-10  Jakub Jelinek  <jakub@redhat.com>

            PR middle-end/99007
            * gimplify.c (gimplify_scan_omp_clauses): For MEM_REF on
reductions,
            temporarily disable gimplify_ctxp->into_ssa around gimplify_expr
            calls.

            * g++.dg/gomp/pr99007.C: New test.
            * gcc.dg/gomp/pr99007-1.c: New test.
            * gcc.dg/gomp/pr99007-2.c: New test.
            * gcc.dg/gomp/pr99007-3.c: New test.

[Bug middle-end/99007] [8/9/10 Regression] ICE in dominated_by_p, at dominance.c:1124

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|11.0                        |8.5
            Summary|[11 Regression] ICE in      |[8/9/10 Regression] ICE in
                   |dominated_by_p, at          |dominated_by_p, at
                   |dominance.c:1124            |dominance.c:1124

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed on the trunk so far.

[Bug middle-end/99007] [8/9/10 Regression] ICE in dominated_by_p, at dominance.c:1124

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

[Bug middle-end/99007] [8/9/10 Regression] ICE in dominated_by_p, at dominance.c:1124

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

--- Comment #7 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:deba6b20a3889aa23f0e4b3a5248de4172a0167d

commit r10-9465-gdeba6b20a3889aa23f0e4b3a5248de4172a0167d
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Feb 10 10:34:58 2021 +0100

    openmp: Temporarily disable into_ssa when gimplifying OpenMP reduction
clauses [PR99007]

    gimplify_scan_omp_clauses was already calling gimplify_expr with false as
    last argument to make sure it is not an SSA_NAME, but as the testcases
show,
    that is not enough, SSA_NAME temporaries created during that gimplification
    can be reused too and we can't allow SSA_NAMEs to be used across OpenMP
    region boundaries, as we can only firstprivatize decls.

    Fixed by temporarily disabling into_ssa.

    2021-02-10  Jakub Jelinek  <jakub@redhat.com>

            PR middle-end/99007
            * gimplify.c (gimplify_scan_omp_clauses): For MEM_REF on
reductions,
            temporarily disable gimplify_ctxp->into_ssa around gimplify_expr
            calls.

            * g++.dg/gomp/pr99007.C: New test.
            * gcc.dg/gomp/pr99007-1.c: New test.
            * gcc.dg/gomp/pr99007-2.c: New test.
            * gcc.dg/gomp/pr99007-3.c: New test.

    (cherry picked from commit bd0e37f68a3aed944df4eb739a0734bb87153749)

[Bug middle-end/99007] [8/9 Regression] ICE in dominated_by_p, at dominance.c:1124

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[8/9/10 Regression] ICE in  |[8/9 Regression] ICE in
                   |dominated_by_p, at          |dominated_by_p, at
                   |dominance.c:1124            |dominance.c:1124

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 10.3 too.

[Bug middle-end/99007] [8/9 Regression] ICE in dominated_by_p, at dominance.c:1124

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

--- Comment #9 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:024f7908cbdd82e4e23a06e6df47f3823e8bdf59

commit r9-9417-g024f7908cbdd82e4e23a06e6df47f3823e8bdf59
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Feb 10 10:34:58 2021 +0100

    openmp: Temporarily disable into_ssa when gimplifying OpenMP reduction
clauses [PR99007]

    gimplify_scan_omp_clauses was already calling gimplify_expr with false as
    last argument to make sure it is not an SSA_NAME, but as the testcases
show,
    that is not enough, SSA_NAME temporaries created during that gimplification
    can be reused too and we can't allow SSA_NAMEs to be used across OpenMP
    region boundaries, as we can only firstprivatize decls.

    Fixed by temporarily disabling into_ssa.

    2021-02-10  Jakub Jelinek  <jakub@redhat.com>

            PR middle-end/99007
            * gimplify.c (gimplify_scan_omp_clauses): For MEM_REF on
reductions,
            temporarily disable gimplify_ctxp->into_ssa around gimplify_expr
            calls.

            * g++.dg/gomp/pr99007.C: New test.
            * gcc.dg/gomp/pr99007-1.c: New test.
            * gcc.dg/gomp/pr99007-2.c: New test.
            * gcc.dg/gomp/pr99007-3.c: New test.

    (cherry picked from commit deba6b20a3889aa23f0e4b3a5248de4172a0167d)

[Bug middle-end/99007] [8/9 Regression] ICE in dominated_by_p, at dominance.c:1124

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

--- Comment #10 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-8 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:5dfeb10fd2d21eebf329272cce687ad34c2ae2a9

commit r8-10882-g5dfeb10fd2d21eebf329272cce687ad34c2ae2a9
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Feb 10 10:34:58 2021 +0100

    openmp: Temporarily disable into_ssa when gimplifying OpenMP reduction
clauses [PR99007]

    gimplify_scan_omp_clauses was already calling gimplify_expr with false as
    last argument to make sure it is not an SSA_NAME, but as the testcases
show,
    that is not enough, SSA_NAME temporaries created during that gimplification
    can be reused too and we can't allow SSA_NAMEs to be used across OpenMP
    region boundaries, as we can only firstprivatize decls.

    Fixed by temporarily disabling into_ssa.

    2021-02-10  Jakub Jelinek  <jakub@redhat.com>

            PR middle-end/99007
            * gimplify.c (gimplify_scan_omp_clauses): For MEM_REF on
reductions,
            temporarily disable gimplify_ctxp->into_ssa around gimplify_expr
            calls.

            * g++.dg/gomp/pr99007.C: New test.
            * gcc.dg/gomp/pr99007-1.c: New test.
            * gcc.dg/gomp/pr99007-2.c: New test.
            * gcc.dg/gomp/pr99007-3.c: New test.

    (cherry picked from commit deba6b20a3889aa23f0e4b3a5248de4172a0167d)

[Bug middle-end/99007] [8/9 Regression] ICE in dominated_by_p, at dominance.c:1124

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99007

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

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

--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.