[Bug libgomp/106045] New: Incorrect testcase in libgomp.c/target-31.c at -O0

[Bug libgomp/106045] New: Incorrect testcase in libgomp.c/target-31.c at -O0

[piannetta at kalrayinc dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106045

            Bug ID: 106045
           Summary: Incorrect testcase in libgomp.c/target-31.c at -O0
           Product: gcc
           Version: 10.3.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libgomp
          Assignee: unassigned at gcc dot gnu.org
          Reporter: piannetta at kalrayinc dot com
                CC: jakub at gcc dot gnu.org
  Target Milestone: ---

Created attachment 53183
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53183&action=edit
patch which add the private clause on i

When investigating a test failure in the openmp testsuite, we found that the
test target-31.c exhibit a wrong behavior at -O0.

This is due to a missing private clause on the i variable.  At higher
optimization levels, the i variable is optimized out which makes the wrong
behavior disappear.

In attachment, a patch adding the missing private clause.

[Bug libgomp/106045] Incorrect testcase in libgomp.c/target-31.c at -O0

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106045

--- Comment #1 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:85d613da341b76308edea48359a5dbc7061937c4

commit r13-1190-g85d613da341b76308edea48359a5dbc7061937c4
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue Jun 21 17:51:08 2022 +0200

    libgomp: Fix up target-31.c test [PR106045]

    The i variable is used inside of the parallel in:
          #pragma omp simd safelen(32) private (v)
          for (i = 0; i < 64; i++)
            {
              v = 3 * i;
              ll[i] = u1 + v * u2[0] + u2[1] + x + y[0] + y[1] + v + h[0] +
u3[i];
            }
    where i is predetermined linear (so while inside of the body
    it is safe, private per SIMD lane var) the final value is written to
    the shared variable, and in:
          for (i = 0; i < 64; i++)
            if (ll[i] != u1 + 3 * i * u2[0] + u2[1] + x + y[0] + y[1] + 3 * i +
13 + 14 + i)
              #pragma omp atomic write
                err = 1;
    which is a normal loop and so it isn't in any way privatized there.
    So we have a data race, fixed by adding private (i) clause to the
    parallel.

    2022-06-21  Jakub Jelinek  <jakub@redhat.com>
                Paul Iannetta  <piannetta@kalrayinc.com>

            PR libgomp/106045
            * testsuite/libgomp.c/target-31.c: Add private (i) clause.

[Bug libgomp/106045] Incorrect testcase in libgomp.c/target-31.c at -O0

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106045

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

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

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Thanks, fixed for 13 now.

[Bug libgomp/106045] Incorrect testcase in libgomp.c/target-31.c at -O0

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106045

--- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:d67dfc5f80e537460ebf809758a0e673028ebad7

commit r12-8521-gd67dfc5f80e537460ebf809758a0e673028ebad7
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue Jun 21 17:51:08 2022 +0200

    libgomp: Fix up target-31.c test [PR106045]

    The i variable is used inside of the parallel in:
          #pragma omp simd safelen(32) private (v)
          for (i = 0; i < 64; i++)
            {
              v = 3 * i;
              ll[i] = u1 + v * u2[0] + u2[1] + x + y[0] + y[1] + v + h[0] +
u3[i];
            }
    where i is predetermined linear (so while inside of the body
    it is safe, private per SIMD lane var) the final value is written to
    the shared variable, and in:
          for (i = 0; i < 64; i++)
            if (ll[i] != u1 + 3 * i * u2[0] + u2[1] + x + y[0] + y[1] + 3 * i +
13 + 14 + i)
              #pragma omp atomic write
                err = 1;
    which is a normal loop and so it isn't in any way privatized there.
    So we have a data race, fixed by adding private (i) clause to the
    parallel.

    2022-06-21  Jakub Jelinek  <jakub@redhat.com>
                Paul Iannetta  <piannetta@kalrayinc.com>

            PR libgomp/106045
            * testsuite/libgomp.c/target-31.c: Add private (i) clause.

    (cherry picked from commit 85d613da341b76308edea48359a5dbc7061937c4)

[Bug libgomp/106045] Incorrect testcase in libgomp.c/target-31.c at -O0

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106045

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:732e2315e6782bfc93241027f01cd3dfbdb5269c

commit r11-10352-g732e2315e6782bfc93241027f01cd3dfbdb5269c
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue Jun 21 17:51:08 2022 +0200

    libgomp: Fix up target-31.c test [PR106045]

    The i variable is used inside of the parallel in:
          #pragma omp simd safelen(32) private (v)
          for (i = 0; i < 64; i++)
            {
              v = 3 * i;
              ll[i] = u1 + v * u2[0] + u2[1] + x + y[0] + y[1] + v + h[0] +
u3[i];
            }
    where i is predetermined linear (so while inside of the body
    it is safe, private per SIMD lane var) the final value is written to
    the shared variable, and in:
          for (i = 0; i < 64; i++)
            if (ll[i] != u1 + 3 * i * u2[0] + u2[1] + x + y[0] + y[1] + 3 * i +
13 + 14 + i)
              #pragma omp atomic write
                err = 1;
    which is a normal loop and so it isn't in any way privatized there.
    So we have a data race, fixed by adding private (i) clause to the
    parallel.

    2022-06-21  Jakub Jelinek  <jakub@redhat.com>
                Paul Iannetta  <piannetta@kalrayinc.com>

            PR libgomp/106045
            * testsuite/libgomp.c/target-31.c: Add private (i) clause.

    (cherry picked from commit 85d613da341b76308edea48359a5dbc7061937c4)

[Bug libgomp/106045] Incorrect testcase in libgomp.c/target-31.c at -O0

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106045

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |11.4
      Known to work|                            |11.3.1, 12.2.0
      Known to fail|                            |11.3.0, 12.1.0

[Bug libgomp/106045] Incorrect testcase in libgomp.c/target-31.c at -O0

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106045

--- 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:58c396a2ce4f8f66e6c379d74e834e3c8d93f36b

commit r10-11331-g58c396a2ce4f8f66e6c379d74e834e3c8d93f36b
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue Jun 21 17:51:08 2022 +0200

    libgomp: Fix up target-31.c test [PR106045]

    The i variable is used inside of the parallel in:
          #pragma omp simd safelen(32) private (v)
          for (i = 0; i < 64; i++)
            {
              v = 3 * i;
              ll[i] = u1 + v * u2[0] + u2[1] + x + y[0] + y[1] + v + h[0] +
u3[i];
            }
    where i is predetermined linear (so while inside of the body
    it is safe, private per SIMD lane var) the final value is written to
    the shared variable, and in:
          for (i = 0; i < 64; i++)
            if (ll[i] != u1 + 3 * i * u2[0] + u2[1] + x + y[0] + y[1] + 3 * i +
13 + 14 + i)
              #pragma omp atomic write
                err = 1;
    which is a normal loop and so it isn't in any way privatized there.
    So we have a data race, fixed by adding private (i) clause to the
    parallel.

    2022-06-21  Jakub Jelinek  <jakub@redhat.com>
                Paul Iannetta  <piannetta@kalrayinc.com>

            PR libgomp/106045
            * testsuite/libgomp.c/target-31.c: Add private (i) clause.

    (cherry picked from commit 85d613da341b76308edea48359a5dbc7061937c4)