[Bug rtl-optimization/105528] New: [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f

[Bug rtl-optimization/105528] New: [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f

[zsojka at seznam dot cz]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

            Bug ID: 105528
           Summary: [11/12/13 Regression] ICE: in expand_expr_real_2, at
                    expr.cc:10290 with -O2 -mavx512f
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          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 52940
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52940&action=edit
reduced testcase

Compiler output:
$ x86_64-pc-linux-gnu-gcc -O2 -mavx512f testcase.c 
during RTL pass: expand
In function 'bar',
    inlined from 'foo' at testcase.c:17:12:
testcase.c:11:12: internal compiler error: in expand_expr_real_2, at
expr.cc:10290
   11 |   return v + w;
      |          ~~^~~
0x6cce24 expand_expr_real_2(separate_ops*, rtx_def*, machine_mode,
expand_modifier)
        /repo/gcc-trunk/gcc/expr.cc:10290
0xfade5c expand_expr_real_1(tree_node*, rtx_def*, machine_mode,
expand_modifier, rtx_def**, bool)
        /repo/gcc-trunk/gcc/expr.cc:10504
0xfb0064 expand_expr
        /repo/gcc-trunk/gcc/expr.h:301
0xfb0064 expand_operands(tree_node*, tree_node*, rtx_def*, rtx_def**,
rtx_def**, expand_modifier)
        /repo/gcc-trunk/gcc/expr.cc:8314
0xfa720f expand_expr_real_2(separate_ops*, rtx_def*, machine_mode,
expand_modifier)
        /repo/gcc-trunk/gcc/expr.cc:9281
0xe72418 expand_gimple_stmt_1
        /repo/gcc-trunk/gcc/cfgexpand.cc:3972
0xe72418 expand_gimple_stmt
        /repo/gcc-trunk/gcc/cfgexpand.cc:4033
0xe784ee expand_gimple_basic_block
        /repo/gcc-trunk/gcc/cfgexpand.cc:6081
0xe7a0ef execute
        /repo/gcc-trunk/gcc/cfgexpand.cc:6807
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-r13-168-20220508001620-ga1947c92f7c-checking-yes-rtl-df-extra-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
--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-r13-168-20220508001620-ga1947c92f7c-checking-yes-rtl-df-extra-amd64
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 13.0.0 20220508 (experimental) (GCC)

[Bug target/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
This is rather interesting adding -g allows it to compile.

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Target Milestone|---                         |11.4
          Component|target                      |tree-optimization
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2022-05-09

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Without -g:
  _9 = _8 >= { 4294967295, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
  _10 = BIT_FIELD_REF <w, 32, 160>;
  _11 = VIEW_CONVERT_EXPR<unsigned int[16]>(w)[_10];
  _12 = _11 + 1;
  _18 = w;
  _19 = .VEC_SET (_18, _12, _10);
  w = _19;
  v_13 = VEC_COND_EXPR <_9, { 4294967264, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0 }, { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }>;

With -g:
  v_13 = .VCONDU (_8, { 4294967295, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0
}, { 4294967264, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, { 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, 111);


  vector(16) unsigned int _8;
  vector(16) <signed-boolean:1> _9;

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |compare-debug-failure

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Even though this is not directly a compare debug failure, it could be if expand
didn't crash.

Note GCC 10.x had:
  _9 = VEC_COND_EXPR <_8 >= { 4294967295, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0 }, { 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5 }, { 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }>;

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Most likely introduced by r11-1445-g502d63b6d61415.

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[11/12/13 Regression] ICE:  |[11/12/13 Regression] ICE:
                   |in expand_expr_real_2, at   |in expand_expr_real_2, at
                   |expr.cc:10290 with -O2      |expr.cc:10290 with -O2
                   |-mavx512f                   |-mavx512f since
                   |                            |r11-8059-g5240c5ca2e2b1f36
                 CC|                            |marxin at gcc dot gnu.org,
                   |                            |rguenth at gcc dot gnu.org,
                   |                            |rsandifo at gcc dot gnu.org

--- Comment #5 from Martin Liška <marxin at gcc dot gnu.org> ---
Started with r11-8059-g5240c5ca2e2b1f36.

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org
             Status|NEW                         |ASSIGNED

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
I'll have a look.

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 52941
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52941&action=edit
gcc12-pr105528.patch

Untested fix.  gsi_remove moves iterator to the next stmt after the removed
one, and the caller uses gsi_next, so if gsi_remove happened, the caller skips
processing one stmt.

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #7)
> Created attachment 52941 [details]
> gcc12-pr105528.patch
> 
> Untested fix.  gsi_remove moves iterator to the next stmt after the removed
> one, and the caller uses gsi_next, so if gsi_remove happened, the caller
> skips processing one stmt.

That explains the -g difference even.

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

--- Comment #9 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:ddd46293e2b508b260eefe1adbad6720df7a5dd2

commit r13-275-gddd46293e2b508b260eefe1adbad6720df7a5dd2
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue May 10 18:34:08 2022 +0200

    isel: Fix up gimple_expand_vec_set_expr [PR105528]

    The following testcase ICEs (and only without -g), because we don't replace
    one VEC_COND_EXPR with .VCOND* call.
    We don't do that because gimple_expand_vec_set_expr adds some stmts before
    *gsi and then uses gsi_remove to remove it.  gsi_remove moves the iterator
    to the next stmt and in the caller we then do gsi_next before looking at
    another stmt, which means we can skip processing of one stmt, which in this
    case happened to be a VEC_COND_EXPR but with -g is some debug stmt in
    between.  As we always emit some stmts before it, it is easy to update the
    iterator to the last stmt emitted there, so that caller continues really
    with the next stmt.

    2022-05-10  Jakub Jelinek  <jakub@redhat.com>

            PR tree-optimization/105528
            * gimple-isel.cc (gimple_expand_vec_set_expr): After gsi_remove
            set *gsi to gsi_for_stmt (ass_stmt).  Fix up function comment.

            * gcc.dg/pr105528.c: New test.

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

--- Comment #10 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:9468cc35daca9fd6a0ec479217bb40a77616fb25

commit r12-8362-g9468cc35daca9fd6a0ec479217bb40a77616fb25
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue May 10 18:34:08 2022 +0200

    isel: Fix up gimple_expand_vec_set_expr [PR105528]

    The following testcase ICEs (and only without -g), because we don't replace
    one VEC_COND_EXPR with .VCOND* call.
    We don't do that because gimple_expand_vec_set_expr adds some stmts before
    *gsi and then uses gsi_remove to remove it.  gsi_remove moves the iterator
    to the next stmt and in the caller we then do gsi_next before looking at
    another stmt, which means we can skip processing of one stmt, which in this
    case happened to be a VEC_COND_EXPR but with -g is some debug stmt in
    between.  As we always emit some stmts before it, it is easy to update the
    iterator to the last stmt emitted there, so that caller continues really
    with the next stmt.

    2022-05-10  Jakub Jelinek  <jakub@redhat.com>

            PR tree-optimization/105528
            * gimple-isel.cc (gimple_expand_vec_set_expr): After gsi_remove
            set *gsi to gsi_for_stmt (ass_stmt).  Fix up function comment.

            * gcc.dg/pr105528.c: New test.

    (cherry picked from commit ddd46293e2b508b260eefe1adbad6720df7a5dd2)

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

--- Comment #11 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:03f73056bf688da31b649fd04e70c9d7a0dcc001

commit r11-9975-g03f73056bf688da31b649fd04e70c9d7a0dcc001
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue May 10 18:34:08 2022 +0200

    isel: Fix up gimple_expand_vec_set_expr [PR105528]

    The following testcase ICEs (and only without -g), because we don't replace
    one VEC_COND_EXPR with .VCOND* call.
    We don't do that because gimple_expand_vec_set_expr adds some stmts before
    *gsi and then uses gsi_remove to remove it.  gsi_remove moves the iterator
    to the next stmt and in the caller we then do gsi_next before looking at
    another stmt, which means we can skip processing of one stmt, which in this
    case happened to be a VEC_COND_EXPR but with -g is some debug stmt in
    between.  As we always emit some stmts before it, it is easy to update the
    iterator to the last stmt emitted there, so that caller continues really
    with the next stmt.

    2022-05-10  Jakub Jelinek  <jakub@redhat.com>

            PR tree-optimization/105528
            * gimple-isel.cc (gimple_expand_vec_set_expr): After gsi_remove
            set *gsi to gsi_for_stmt (ass_stmt).  Fix up function comment.

            * gcc.dg/pr105528.c: New test.

    (cherry picked from commit ddd46293e2b508b260eefe1adbad6720df7a5dd2)

[Bug tree-optimization/105528] [11/12/13 Regression] ICE: in expand_expr_real_2, at expr.cc:10290 with -O2 -mavx512f since r11-8059-g5240c5ca2e2b1f36

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105528

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

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

--- Comment #12 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.