[Bug target/100302] New: ICE in abs_hwi, at hwint.h:324

[Bug target/100302] New: ICE in abs_hwi, at hwint.h:324

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

            Bug ID: 100302
           Summary: ICE in abs_hwi, at hwint.h:324
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: asolokha at gmx dot com
  Target Milestone: ---
            Target: aarch64-linux-gnu

gcc-11.0.1-alpha20210426 snapshot (g:d3212299e2cfc3c16dd23bab26ec6c49024105f8)
ICEs when compiling the following testcase w/ -mcpu=zeus -O1
-ftree-loop-vectorize -fno-tree-scev-cprop --param vect-partial-vector-usage=0:

long int ct;

void
pt (void)
{
  for (ct = 0; ct >= 0; ++ct)
    ;
}

% aarch64-linux-gnu-gcc-11.0.1 -mcpu=zeus -O1 -ftree-loop-vectorize
-fno-tree-scev-cprop --param vect-partial-vector-usage=0 -c ppzrgnc5.c
during RTL pass: expand
ppzrgnc5.c: In function 'pt':
ppzrgnc5.c:4:1: internal compiler error: in abs_hwi, at hwint.h:324
    4 | pt (void)
      | ^~
0x7f23c9 abs_hwi(long)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/hwint.h:324
0x7f23c9 abs_hwi(long)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/hwint.h:322
0x7f23c9 aarch64_add_offset_1_temporaries
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/config/aarch64/aarch64.c:4739
0x7f23c9 aarch64_offset_temporaries
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/config/aarch64/aarch64.c:4854
0x12a1c6c aarch64_legitimate_constant_p
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/config/aarch64/aarch64.c:18023
0xe455bb general_operand(rtx_def*, machine_mode)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/recog.c:1434
0xb23e7a copy_to_mode_reg(machine_mode, rtx_def*)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/explow.c:648
0xdbc022 prepare_cmp_insn
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/optabs.c:4397
0xdbc8e6 emit_cmp_and_jump_insns(rtx_def*, rtx_def*, rtx_code, rtx_def*,
machine_mode, int, rtx_def*, profile_probability)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/optabs.c:4583
0xa91020 do_compare_rtx_and_jump(rtx_def*, rtx_def*, rtx_code, int,
machine_mode, rtx_def*, rtx_code_label*, rtx_code_label*, profile_probability)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/dojump.c:1220
0xa92574 do_compare_and_jump
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/dojump.c:1294
0xa93c83 do_jump_1
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/dojump.c:242
0xa230cd expand_gimple_cond
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/cfgexpand.c:2646
0xa2396c expand_gimple_basic_block
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/cfgexpand.c:5904
0xa2539f execute
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210426/work/gcc-11-20210426/gcc/cfgexpand.c:6729

[Bug target/100302] ICE in abs_hwi, at hwint.h:324

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
                 CC|                            |jakub at gcc dot gnu.org
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2021-04-28

[Bug target/100302] ICE in abs_hwi, at hwint.h:324

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Related to PR94121, just another abs_hwi needs to change to absu_hwi.

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |11.2
            Summary|ICE in abs_hwi, at          |[11/12 Regression] ICE in
                   |hwint.h:324                 |abs_hwi, at hwint.h:324
                   |                            |since r11-7861
           Priority|P3                          |P2

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r11-7861-ge4180ab2fea0d3e8010f23b5e73095ac13cedafa

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
With -march=armv8.2-a+sve -O1 -ftree-loop-vectorize -fno-tree-scev-cprop
--param vect-partial-vector-usage=0 -fvect-cost-model=unlimited
it started to ICE already with
r11-2221-g9fb832ce382d649b7687426e6bc4e5d3715cb78a
when vect-parial-vector-usage param has been introduced.  Anyway, latent before
that.

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 50698
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=50698&action=edit
gcc12-pr100302.patch

Untested fix.

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

--- Comment #5 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:1bb3e2c0ce6ed363c72caf814a6ba6d7b17c3e0a

commit r12-282-g1bb3e2c0ce6ed363c72caf814a6ba6d7b17c3e0a
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu Apr 29 11:34:50 2021 +0200

    aarch64: Fix ICE in aarch64_add_offset_1_temporaries [PR100302]

    In PR94121 I've changed aarch64_add_offset_1 to use absu_hwi instead of
    abs_hwi because offset can be HOST_WIDE_INT_MIN.  As can be seen with
    the testcase below, aarch64_add_offset_1_temporaries suffers from the same
    problem and should be in sync with aarch64_add_offset_1, i.e. for
    HOST_WIDE_INT_MIN it needs a temporary.

    2021-04-29  Jakub Jelinek  <jakub@redhat.com>

            PR target/100302
            * config/aarch64/aarch64.c (aarch64_add_offset_1_temporaries): Use
            absu_hwi instead of abs_hwi.

            * gcc.target/aarch64/sve/pr100302.c: New test.

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

--- Comment #6 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:5ac1313f32c5cd875ad047f6575dd4608e1949cf

commit r11-8326-g5ac1313f32c5cd875ad047f6575dd4608e1949cf
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu Apr 29 11:34:50 2021 +0200

    aarch64: Fix ICE in aarch64_add_offset_1_temporaries [PR100302]

    In PR94121 I've changed aarch64_add_offset_1 to use absu_hwi instead of
    abs_hwi because offset can be HOST_WIDE_INT_MIN.  As can be seen with
    the testcase below, aarch64_add_offset_1_temporaries suffers from the same
    problem and should be in sync with aarch64_add_offset_1, i.e. for
    HOST_WIDE_INT_MIN it needs a temporary.

    2021-04-29  Jakub Jelinek  <jakub@redhat.com>

            PR target/100302
            * config/aarch64/aarch64.c (aarch64_add_offset_1_temporaries): Use
            absu_hwi instead of abs_hwi.

            * gcc.target/aarch64/sve/pr100302.c: New test.

    (cherry picked from commit 1bb3e2c0ce6ed363c72caf814a6ba6d7b17c3e0a)

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

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

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

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 11.2 and on the trunk.

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:42ba897862fbdc0d6b27fae95a10fcc1a6436b1d

commit r10-9794-g42ba897862fbdc0d6b27fae95a10fcc1a6436b1d
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu Apr 29 11:34:50 2021 +0200

    aarch64: Fix ICE in aarch64_add_offset_1_temporaries [PR100302]

    In PR94121 I've changed aarch64_add_offset_1 to use absu_hwi instead of
    abs_hwi because offset can be HOST_WIDE_INT_MIN.  As can be seen with
    the testcase below, aarch64_add_offset_1_temporaries suffers from the same
    problem and should be in sync with aarch64_add_offset_1, i.e. for
    HOST_WIDE_INT_MIN it needs a temporary.

    2021-04-29  Jakub Jelinek  <jakub@redhat.com>

            PR target/100302
            * config/aarch64/aarch64.c (aarch64_add_offset_1_temporaries): Use
            absu_hwi instead of abs_hwi.

    (cherry picked from commit 1bb3e2c0ce6ed363c72caf814a6ba6d7b17c3e0a)

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:a322c55e05a4f74cb5cc7513826ec4900eb50d14

commit r9-9506-ga322c55e05a4f74cb5cc7513826ec4900eb50d14
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu Apr 29 11:34:50 2021 +0200

    aarch64: Fix ICE in aarch64_add_offset_1_temporaries [PR100302]

    In PR94121 I've changed aarch64_add_offset_1 to use absu_hwi instead of
    abs_hwi because offset can be HOST_WIDE_INT_MIN.  As can be seen with
    the testcase below, aarch64_add_offset_1_temporaries suffers from the same
    problem and should be in sync with aarch64_add_offset_1, i.e. for
    HOST_WIDE_INT_MIN it needs a temporary.

    2021-04-29  Jakub Jelinek  <jakub@redhat.com>

            PR target/100302
            * config/aarch64/aarch64.c (aarch64_add_offset_1_temporaries): Use
            absu_hwi instead of abs_hwi.

    (cherry picked from commit 1bb3e2c0ce6ed363c72caf814a6ba6d7b17c3e0a)

[Bug target/100302] [11/12 Regression] ICE in abs_hwi, at hwint.h:324 since r11-7861

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100302

--- Comment #10 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-8 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:e704ac6615e1e4badeb048ac3a0551f02a94dfc4

commit r8-10947-ge704ac6615e1e4badeb048ac3a0551f02a94dfc4
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Thu Apr 29 11:34:50 2021 +0200

    aarch64: Fix ICE in aarch64_add_offset_1_temporaries [PR100302]

    In PR94121 I've changed aarch64_add_offset_1 to use absu_hwi instead of
    abs_hwi because offset can be HOST_WIDE_INT_MIN.  As can be seen with
    the testcase below, aarch64_add_offset_1_temporaries suffers from the same
    problem and should be in sync with aarch64_add_offset_1, i.e. for
    HOST_WIDE_INT_MIN it needs a temporary.

    2021-04-29  Jakub Jelinek  <jakub@redhat.com>

            PR target/100302
            * config/aarch64/aarch64.c (aarch64_add_offset_1_temporaries): Use
            absu_hwi instead of abs_hwi.

    (cherry picked from commit 1bb3e2c0ce6ed363c72caf814a6ba6d7b17c3e0a)