[Bug rtl-optimization/98229] New: ICE at -O1 in 32-bit mode on x86_64-pc-linux-gnu in decompose, at rtl.h:2282

[Bug rtl-optimization/98229] New: ICE at -O1 in 32-bit mode on x86_64-pc-linux-gnu in decompose, at rtl.h:2282

[zhendong.su at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98229

            Bug ID: 98229
           Summary: ICE at -O1 in 32-bit mode on x86_64-pc-linux-gnu in
                    decompose, at rtl.h:2282
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: zhendong.su at inf dot ethz.ch
  Target Milestone: ---

[555] % gcctk -v
Using built-in specs.
COLLECT_GCC=gcctk
COLLECT_LTO_WRAPPER=/local/suz-local/software/local/gcc-trunk/libexec/gcc/x86_64-pc-linux-gnu/11.0.0/lto-wrapper
Target: x86_64-pc-linux-gnu
Configured with: ../gcc-trunk/configure --disable-bootstrap
--prefix=/local/suz-local/software/local/gcc-trunk --enable-languages=c,c++
--disable-werror --enable-multilib --with-system-zlib
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 11.0.0 20201210 (experimental) [master revision
b46dd03fe94:30c63a5c82a:afc14c8d0a9e7af13698a7eec84226a3cc4b0e67] (GCC) 
[556] % 
[556] % gcctk -m32 -O0 -c small.c
[557] % 
[557] % gcctk -m32 -O1 -c small.c
during RTL pass: expand
small.c: In function ‘f’:
small.c:2:14: internal compiler error: in decompose, at rtl.h:2282
    2 | void f() { a = a % ~0UL; }
      |            ~~^~~~~~~~~~
0x927acb wi::int_traits<std::pair<rtx_def*, machine_mode> >::decompose(long*,
unsigned int, std::pair<rtx_def*, machine_mode> const&)
        ../../gcc-trunk/gcc/rtl.h:2280
0x927acb wide_int_ref_storage<false,
false>::wide_int_ref_storage<std::pair<rtx_def*, machine_mode>
>(std::pair<rtx_def*, machine_mode> const&)
        ../../gcc-trunk/gcc/wide-int.h:1025
0x927acb generic_wide_int<wide_int_ref_storage<false, false>
>::generic_wide_int<std::pair<rtx_def*, machine_mode> >(std::pair<rtx_def*,
machine_mode> const&)
        ../../gcc-trunk/gcc/wide-int.h:782
0x927acb wide_int_storage::wide_int_storage<std::pair<rtx_def*, machine_mode>
>(std::pair<rtx_def*, machine_mode> const&)
        ../../gcc-trunk/gcc/wide-int.h:1115
0x927acb
generic_wide_int<wide_int_storage>::generic_wide_int<std::pair<rtx_def*,
machine_mode> >(std::pair<rtx_def*, machine_mode> const&)
        ../../gcc-trunk/gcc/wide-int.h:782
0x927acb expand_divmod(int, tree_code, machine_mode, rtx_def*, rtx_def*,
rtx_def*, int, optab_methods)
        ../../gcc-trunk/gcc/expmed.c:4214
0xbec854 expand_doubleword_mod
        ../../gcc-trunk/gcc/optabs.c:1085
0xbecca1 expand_doubleword_divmod(machine_mode, rtx_def*, rtx_def*, rtx_def**,
bool)
        ../../gcc-trunk/gcc/optabs.c:1159
0xbea858 expand_binop(machine_mode, optab_tag, rtx_def*, rtx_def*, rtx_def*,
int, optab_methods)
        ../../gcc-trunk/gcc/optabs.c:2119
0xbedf03 sign_expand_binop(machine_mode, optab_tag, optab_tag, rtx_def*,
rtx_def*, rtx_def*, int, optab_methods)
        ../../gcc-trunk/gcc/optabs.c:2282
0x928606 expand_divmod(int, tree_code, machine_mode, rtx_def*, rtx_def*,
rtx_def*, int, optab_methods)
        ../../gcc-trunk/gcc/expmed.c:5216
0x938553 expand_expr_real_2(separate_ops*, rtx_def*, machine_mode,
expand_modifier)
        ../../gcc-trunk/gcc/expr.c:9243
0x93f5e0 expand_expr_real_1(tree_node*, rtx_def*, machine_mode,
expand_modifier, rtx_def**, bool)
        ../../gcc-trunk/gcc/expr.c:10177
0x94b9d1 store_expr(tree_node*, rtx_def*, int, bool, bool)
        ../../gcc-trunk/gcc/expr.c:5859
0x94d84f expand_assignment(tree_node*, tree_node*, bool)
        ../../gcc-trunk/gcc/expr.c:5595
0x7f483a expand_gimple_stmt_1
        ../../gcc-trunk/gcc/cfgexpand.c:3901
0x7f483a expand_gimple_stmt
        ../../gcc-trunk/gcc/cfgexpand.c:3999
0x7fc087 expand_gimple_basic_block
        ../../gcc-trunk/gcc/cfgexpand.c:6036
0x7fe9d6 execute
        ../../gcc-trunk/gcc/cfgexpand.c:6720
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.
[558] % 
[558] % cat small.c
unsigned long long a;
void f() { a = a % ~0UL; }

[Bug rtl-optimization/98229] [11 Regression] ICE at -O1 in 32-bit mode on x86_64-pc-linux-gnu in decompose, at rtl.h:2282

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98229

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2020-12-10
            Summary|ICE at -O1 in 32-bit mode   |[11 Regression] ICE at -O1
                   |on x86_64-pc-linux-gnu in   |in 32-bit mode on
                   |decompose, at rtl.h:2282    |x86_64-pc-linux-gnu in
                   |                            |decompose, at rtl.h:2282
   Target Milestone|---                         |11.0
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org
           Priority|P3                          |P1
                 CC|                            |jakub at gcc dot gnu.org

[Bug rtl-optimization/98229] [11 Regression] ICE at -O1 in 32-bit mode on x86_64-pc-linux-gnu in decompose, at rtl.h:2282

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98229

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 49731
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=49731&action=edit
gcc11-pr98229.patch

Untested fix.

[Bug rtl-optimization/98229] [11 Regression] ICE at -O1 in 32-bit mode on x86_64-pc-linux-gnu in decompose, at rtl.h:2282

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98229

--- 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:1423318fa778649365f5a73c2c4657733e191820

commit r11-5929-g1423318fa778649365f5a73c2c4657733e191820
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Dec 11 12:47:52 2020 +0100

    expand: Fix up expand_doubleword_mod on 32-bit targets [PR98229]

    As the testcase shows, for 32-bit word size we can end up with op1
    up to 0xffffffff (0x100000000 % 0xffffffff == 1 and so we use bit == 32
    for that), but the CONST_INT we got from caller is for DImode in that case
    and not valid for SImode operations.

    The following patch canonicalizes the two spots where the constant needs
    canonicalization.

    2020-12-10  Jakub Jelinek  <jakub@redhat.com>

            PR rtl-optimization/98229
            * optabs.c (expand_doubleword_mod): Canonicalize op1 and
            1 - INTVAL (op1) as word_mode constants when used in
            word_mode arithmetics.

            * gcc.c-torture/compile/pr98229.c: New test.

[Bug rtl-optimization/98229] [11 Regression] ICE at -O1 in 32-bit mode on x86_64-pc-linux-gnu in decompose, at rtl.h:2282

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98229

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

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

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