[Bug target/112088] New: [14 Regression] GCN target testing broken by "amdgcn: add -march=gfx1030 EXPERIMENTAL"

[Bug target/112088] New: [14 Regression] GCN target testing broken by "amdgcn: add -march=gfx1030 EXPERIMENTAL"

[tschwinge at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112088

            Bug ID: 112088
           Summary: [14 Regression] GCN target testing broken by "amdgcn:
                    add -march=gfx1030 EXPERIMENTAL"
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: major
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: tschwinge at gcc dot gnu.org
                CC: ams at gcc dot gnu.org, jules at gcc dot gnu.org
  Target Milestone: ---
            Target: GCN

Assuming that I did my bisecting correctly, it's the recent commit
r14-4787-gc7ec7bd1c6590cf4eed267feab490288e0b8d691 "amdgcn: add -march=gfx1030
EXPERIMENTAL" that completely distorts GCN target testing: execution test FAILs
all over the place due to 'gcn-run' failing with:

    Possible kernel exit value corruption, 2 most significant bytes aren't
0xffff, 0xcafe, or 0: 0x9d4c0038

(The last value different each time.)

I've tested '-march=gfx90a' and '-march=fiji'.

[Bug target/112088] [14 Regression] GCN target testing broken by "amdgcn: add -march=gfx1030 EXPERIMENTAL"

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112088

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |14.0

[Bug target/112088] [14 Regression] GCN target testing broken by "amdgcn: add -march=gfx1030 EXPERIMENTAL"

[ams at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112088

Andrew Stubbs <ams at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2023-10-27
           Assignee|unassigned at gcc dot gnu.org      |ams at gcc dot gnu.org

--- Comment #1 from Andrew Stubbs <ams at gcc dot gnu.org> ---
I'm testing a fix for this.

[Bug target/112088] [14 Regression] GCN target testing broken by "amdgcn: add -march=gfx1030 EXPERIMENTAL"

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112088

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

https://gcc.gnu.org/g:9f3c4c673d769bd3f1d504d85b743ba250e68d7b

commit r14-4982-g9f3c4c673d769bd3f1d504d85b743ba250e68d7b
Author: Andrew Stubbs <ams@codesourcery.com>
Date:   Fri Oct 27 17:53:10 2023 +0100

    amdgcn: Fix bug in gfx1030 support patch

    The previous patch to add gfx1030 support introduced an issue with passing
    exit codes from kernels run under gcn-run (offload kernels were
unaffected).

    gcc/ChangeLog:

            PR target/112088
            * config/gcn/gcn.cc (gcn_expand_epilogue): Fix kernel epilogue
register
            conflict.

[Bug target/112088] [14 Regression] GCN target testing broken by "amdgcn: add -march=gfx1030 EXPERIMENTAL"

[ams at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112088

Andrew Stubbs <ams at gcc dot gnu.org> changed:

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

--- Comment #3 from Andrew Stubbs <ams at gcc dot gnu.org> ---
The patch should fix the bug.