[Bug target/112313] New: [14 Regression] GCN target 'gcc.dg/pr111082.c' ICE, 'during RTL pass: vregs': 'error: unrecognizable insn'

[Bug target/112313] New: [14 Regression] GCN target 'gcc.dg/pr111082.c' ICE, 'during RTL pass: vregs': 'error: unrecognizable insn'

[tschwinge at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112313

            Bug ID: 112313
           Summary: [14 Regression] GCN target 'gcc.dg/pr111082.c' ICE,
                    'during RTL pass: vregs': 'error: unrecognizable insn'
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          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 my tacking is to be believed) something in Git commit
r14-2060-gad5ab848cc487b3f7fd82c7cb3c408747bd10422..r14-3575-g7f2ed06ddc825e8a4e0edfd1d66b5156e6dc1d34
triggers a new GCN target ICE (tested '-march=gfx90a'):

    [...]/gcc/testsuite/gcc.dg/pr111082.c: In function 'minarray2':
    [...]/gcc/testsuite/gcc.dg/pr111082.c:10:1: error: unrecognizable insn:
    (insn 10 9 11 2 (set (reg:V2DI 433)
            (smin:V2DI (reg:V2DI 434)
                (reg:V2DI 430))) -1
         (nil))
    during RTL pass: vregs
    [...]/gcc/testsuite/gcc.dg/pr111082.c:10:1: internal compiler error: in
extract_insn, at recog.cc:2791
    0x74b91f _fatal_insn(char const*, rtx_def const*, char const*, int, char
const*)
            [...]/gcc/rtl-error.cc:108
    0x74b9a0 _fatal_insn_not_found(rtx_def const*, char const*, int, char
const*)
            [...]/gcc/rtl-error.cc:116
    0xed182e extract_insn(rtx_insn*)
            [...]/gcc/recog.cc:2791
    0xb42cfc instantiate_virtual_regs_in_insn
            [...]/gcc/function.cc:1610
    0xb42cfc instantiate_virtual_regs
            [...]/gcc/function.cc:1983
    0xb42cfc execute
            [...]/gcc/function.cc:2030

[Bug target/112313] [14 Regression] GCN target 'gcc.dg/pr111082.c' ICE, 'during RTL pass: vregs': 'error: unrecognizable insn'

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112313

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

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

[Bug target/112313] [14 Regression] GCN target 'gcc.dg/pr111082.c' ICE, 'during RTL pass: vregs': 'error: unrecognizable insn'

[burnus at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112313

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |burnus at gcc dot gnu.org

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Fix written and committed by Andrew as r14-5316-g8da8b922576212

"The problem was a missed case in the vector reduction expand code."

See also https://gcc.gnu.org/pipermail/gcc-patches/2023-November/635982.html


commit 8da8b9225762126ca41b12243d6531cd41a831b3
Author: Andrew Stubbs <ams@codesourcery.com>
Date:   Fri Nov 10 09:43:21 2023 +0000

    amdgcn: Fix vector min/max ICE

    The DImode min/max instructions need a clobber that SImode does not, so
    add the special case to the reduction expand code.

    gcc/ChangeLog:

            * config/gcn/gcn.cc (gcn_expand_reduc_scalar): Add clobber to
DImode
            min/max instructions.

[Bug target/112313] [14 Regression] GCN target 'gcc.dg/pr111082.c' ICE, 'during RTL pass: vregs': 'error: unrecognizable insn'

[ams at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112313

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|UNCONFIRMED                 |RESOLVED
           Assignee|unassigned at gcc dot gnu.org      |ams at gcc dot gnu.org

--- Comment #2 from Andrew Stubbs <ams at gcc dot gnu.org> ---
This is now fixed.