[Bug target/106748] New: ICE in ix86_avx256_split_vector_move_misalign, at config/i386/i386-expand.cc:743 since r13-2111-g6910cad55ffc330d

[Bug target/106748] New: ICE in ix86_avx256_split_vector_move_misalign, at config/i386/i386-expand.cc:743 since r13-2111-g6910cad55ffc330d

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106748

            Bug ID: 106748
           Summary: ICE in ix86_avx256_split_vector_move_misalign, at
                    config/i386/i386-expand.cc:743 since
                    r13-2111-g6910cad55ffc330d
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: marxin at gcc dot gnu.org
                CC: lingling.kong7 at gmail dot com
  Target Milestone: ---
              Host: x86_64-linux-gnu

The following crashes:

$ cat test_passing_unions.i
typedef __bf16 __m256bf16 __attribute__((__vector_size__(32)));
typedef struct {
  __m256bf16 _m256bf16[1];
} YMM_T;

struct {
  YMM_T ymm0;
} fregs;

__m256bf16 do_test_u3b_0_0;
int do_test_i;

void
do_test() { (&fregs.ymm0)[do_test_i]._m256bf16[0] = do_test_u3b_0_0; }

$ gcc test_passing_unions.i -fpack-struct -mavx256-split-unaligned-store
-mavx512vp2intersect
during RTL pass: expand
test_passing_unions.i: In function ‘do_test’:
test_passing_unions.i:14:51: internal compiler error: in
ix86_avx256_split_vector_move_misalign, at config/i386/i386-expand.cc:743
   14 | do_test() { (&fregs.ymm0)[do_test_i]._m256bf16[0] = do_test_u3b_0_0; }
      |             ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~
0x82f27b ix86_avx256_split_vector_move_misalign
        /home/marxin/Programming/gcc/gcc/config/i386/i386-expand.cc:743
0x82f27b ix86_expand_vector_move_misalign(machine_mode, rtx_def**)
        /home/marxin/Programming/gcc/gcc/config/i386/i386-expand.cc:855
0x13e1a91 ix86_expand_vector_move(machine_mode, rtx_def**)
        /home/marxin/Programming/gcc/gcc/config/i386/i386-expand.cc:654
0x19002ea gen_movv16bf(rtx_def*, rtx_def*)
        /home/marxin/Programming/gcc/gcc/config/i386/sse.md:1314
0xb95ee2 rtx_insn* insn_gen_fn::operator()<rtx_def*, rtx_def*>(rtx_def*,
rtx_def*) const
        /home/marxin/Programming/gcc/gcc/recog.h:407
0xb95ee2 emit_move_ccmode
        /home/marxin/Programming/gcc/gcc/expr.cc:3871
0xb95ee2 emit_move_insn_1(rtx_def*, rtx_def*)
        /home/marxin/Programming/gcc/gcc/expr.cc:4016
0xb9630f emit_move_insn(rtx_def*, rtx_def*)
        /home/marxin/Programming/gcc/gcc/expr.cc:4167
0xb9f206 store_expr(tree_node*, rtx_def*, int, bool, bool)
        /home/marxin/Programming/gcc/gcc/expr.cc:6342
0xba1177 expand_assignment(tree_node*, tree_node*, bool)
        /home/marxin/Programming/gcc/gcc/expr.cc:5736
0xa60c80 expand_gimple_stmt_1
        /home/marxin/Programming/gcc/gcc/cfgexpand.cc:3946
0xa60c80 expand_gimple_stmt
        /home/marxin/Programming/gcc/gcc/cfgexpand.cc:4044
0xa67944 expand_gimple_basic_block
        /home/marxin/Programming/gcc/gcc/cfgexpand.cc:6096
0xa695c7 execute
        /home/marxin/Programming/gcc/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.

[Bug target/106748] ICE in ix86_avx256_split_vector_move_misalign, at config/i386/i386-expand.cc:743 since r13-2111-g6910cad55ffc330d

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106748

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2022-08-26
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
                 CC|                            |hjl at gcc dot gnu.org

[Bug target/106748] ICE in ix86_avx256_split_vector_move_misalign, at config/i386/i386-expand.cc:743 since r13-2111-g6910cad55ffc330d

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106748

H.J. Lu <hjl.tools at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |13.0

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> ---
A patch is posted at

https://gcc.gnu.org/pipermail/gcc-patches/2022-August/600429.html

[Bug target/106748] [13 Regression] ICE in ix86_avx256_split_vector_move_misalign, at config/i386/i386-expand.cc:743 since r13-2111-g6910cad55ffc330d

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106748

--- Comment #2 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by H.J. Lu <hjl@gcc.gnu.org>:

https://gcc.gnu.org/g:60d1d296b42b22b08d4eaa38bea02100c07261ac

commit r13-2253-g60d1d296b42b22b08d4eaa38bea02100c07261ac
Author: H.J. Lu <hjl.tools@gmail.com>
Date:   Fri Aug 26 10:26:06 2022 -0700

    x86: Handle V16BF in ix86_avx256_split_vector_move_misalign

    Handle E_V16BFmode in ix86_avx256_split_vector_move_misalign and add
    V16BF to V_256H iterator.

    gcc/

            PR target/106748
            * config/i386/i386-expand.cc
            (ix86_avx256_split_vector_move_misalign): Handle E_V16BFmode.
            * config/i386/sse.md (V_256H): Add V16BF.

    gcc/testsuite/

            PR target/106748
            * gcc.target/i386/pr106748.c: New test.

[Bug target/106748] [13 Regression] ICE in ix86_avx256_split_vector_move_misalign, at config/i386/i386-expand.cc:743 since r13-2111-g6910cad55ffc330d

[hjl.tools at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106748

H.J. Lu <hjl.tools at gmail dot com> changed:

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

--- Comment #3 from H.J. Lu <hjl.tools at gmail dot com> ---
Fixed.