[Bug target/107262] New: ICE: in as_a, at machmode.h:366 with __bf16 arithmetics at -Ofast

[Bug target/107262] New: ICE: in as_a, at machmode.h:366 with __bf16 arithmetics at -Ofast

[zsojka at seznam dot cz]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107262

            Bug ID: 107262
           Summary: ICE: in as_a, at machmode.h:366 with __bf16
                    arithmetics at -Ofast
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: zsojka at seznam dot cz
  Target Milestone: ---
              Host: x86_64-pc-linux-gnu
            Target: x86_64-pc-linux-gnu

Created attachment 53704
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53704&action=edit
reduced testcase

Compiler output:
$ x86_64-pc-linux-gnu-gcc -Ofast testcase.c 
during RTL pass: expand
testcase.c: In function 'foo':
testcase.c:8:10: internal compiler error: in as_a, at machmode.h:366
    8 |   return b;
      |          ^
0x6ddea3 scalar_int_mode as_a<scalar_int_mode>(machine_mode)
        /repo/gcc-trunk/gcc/machmode.h:366
0x6e4005 scalar_mode as_a<scalar_mode>(machine_mode)
        /repo/gcc-trunk/gcc/expr.cc:373
0x6e4005 convert_mode_scalar
        /repo/gcc-trunk/gcc/expr.cc:325
0xfe4442 convert_mode_scalar
        /repo/gcc-trunk/gcc/expr.cc:420
0xfe57f8 convert_modes(machine_mode, machine_mode, rtx_def*, int)
        /repo/gcc-trunk/gcc/expr.cc:924
0x1282633 expand_binop(machine_mode, optab_tag, rtx_def*, rtx_def*, rtx_def*,
int, optab_methods)
        /repo/gcc-trunk/gcc/optabs.cc:1668
0xfd8f7c expand_expr_real_2(separate_ops*, rtx_def*, machine_mode,
expand_modifier)
        /repo/gcc-trunk/gcc/expr.cc:10593
0xfe0295 expand_expr_real_1(tree_node*, rtx_def*, machine_mode,
expand_modifier, rtx_def**, bool)
        /repo/gcc-trunk/gcc/expr.cc:10795
0xeac81c expand_expr
        /repo/gcc-trunk/gcc/expr.h:310
0xeac81c expand_return
        /repo/gcc-trunk/gcc/cfgexpand.cc:3809
0xeac81c expand_gimple_stmt_1
        /repo/gcc-trunk/gcc/cfgexpand.cc:3918
0xeac81c expand_gimple_stmt
        /repo/gcc-trunk/gcc/cfgexpand.cc:4044
0xeb251e expand_gimple_basic_block
        /repo/gcc-trunk/gcc/cfgexpand.cc:6096
0xeb411f execute
        /repo/gcc-trunk/gcc/cfgexpand.cc:6822
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

$ x86_64-pc-linux-gnu-gcc -v
Using built-in specs.
COLLECT_GCC=/repo/gcc-trunk/binary-latest-amd64/bin/x86_64-pc-linux-gnu-gcc
COLLECT_LTO_WRAPPER=/repo/gcc-trunk/binary-trunk-r13-3298-20221014120235-ga75e9bee7c2-checking-yes-rtl-df-extra-nobootstrap-amd64/bin/../libexec/gcc/x86_64-pc-linux-gnu/13.0.0/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: /repo/gcc-trunk//configure --enable-languages=c,c++
--enable-valgrind-annotations --disable-nls --enable-checking=yes,rtl,df,extra
--disable-bootstrap --with-cloog --with-ppl --with-isl
--build=x86_64-pc-linux-gnu --host=x86_64-pc-linux-gnu
--target=x86_64-pc-linux-gnu --with-ld=/usr/bin/x86_64-pc-linux-gnu-ld
--with-as=/usr/bin/x86_64-pc-linux-gnu-as --disable-libstdcxx-pch
--prefix=/repo/gcc-trunk//binary-trunk-r13-3298-20221014120235-ga75e9bee7c2-checking-yes-rtl-df-extra-nobootstrap-amd64
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 13.0.0 20221014 (experimental) (GCC)

[Bug middle-end/107262] ICE: in as_a, at machmode.h:366 with __bf16 arithmetics at -Ofast

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107262

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2022-10-18
     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> ---
Created attachment 53721
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53721&action=edit
gcc13-pr107262.patch

Untested fix.  Thanks for the report.

[Bug middle-end/107262] ICE: in as_a, at machmode.h:366 with __bf16 arithmetics at -Ofast

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107262

--- Comment #2 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:65b98fc763d14d371fcf37a17e33519012ec2bac

commit r13-3374-g65b98fc763d14d371fcf37a17e33519012ec2bac
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Wed Oct 19 11:29:44 2022 +0200

    expr: Fix ICE on BFmode -> SFmode conversion of constant [PR107262]

    I forgot to handle the case where lowpart_subreg returns a VOIDmode
    CONST_INT, in that case convert_mode_scalar obviously doesn't work.

    The following patch fixes that.

    2022-10-19  Jakub Jelinek  <jakub@redhat.com>

            PR middle-end/107262
            * expr.cc (convert_mode_scalar): For BFmode -> SFmode conversions
            of constants, use simplify_unary_operation if fromi has VOIDmode
            instead of recursive convert_mode_scalar.

            * gcc.dg/pr107262.c: New test.

[Bug middle-end/107262] ICE: in as_a, at machmode.h:366 with __bf16 arithmetics at -Ofast

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107262

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED
   Target Milestone|---                         |13.0

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed now.