[Bug c/107969] New: ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[Bug c/107969] New: ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107969

            Bug ID: 107969
           Summary: ICE in extract_insn, at recog.cc:2791 since
                    r13-3292-gc2565a31c1622ab0
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: marxin at gcc dot gnu.org
                CC: jakub at gcc dot gnu.org
  Target Milestone: ---

The follow crashes:

$ cat prec.c
int i;
__bf16 f;

void
bar() { i *= 0 <= f; }

$ gcc prec.c -Ofast -fexcess-precision=16 -m32 -msoft-float -c
prec.c: In function ‘bar’:
prec.c:5:22: error: unrecognizable insn:
    5 | bar() { i *= 0 <= f; }
      |                      ^
(insn 10 9 12 2 (set (reg:CCFP 17 flags)
        (compare:CCFP (reg:SF 89)
            (reg:SF 90))) "prec.c":5:11 -1
     (nil))
during RTL pass: vregs
prec.c:5:22: internal compiler error: in extract_insn, at recog.cc:2791
0x75acd4 _fatal_insn(char const*, rtx_def const*, char const*, int, char
const*)
        /home/marxin/Programming/gcc/gcc/rtl-error.cc:108
0x75acf0 _fatal_insn_not_found(rtx_def const*, char const*, int, char const*)
        /home/marxin/Programming/gcc/gcc/rtl-error.cc:116
0x7592b5 extract_insn(rtx_insn*)
        /home/marxin/Programming/gcc/gcc/recog.cc:2791
0xb88100 instantiate_virtual_regs_in_insn
        /home/marxin/Programming/gcc/gcc/function.cc:1611
0xb88100 instantiate_virtual_regs
        /home/marxin/Programming/gcc/gcc/function.cc:1985
0xb88100 execute
        /home/marxin/Programming/gcc/gcc/function.cc:2034
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 c/107969] ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107969

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |13.0
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2022-12-05

[Bug target/107969] ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107969

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Almost nobody uses msoft-float on x86_64 or even i386 these days.

[Bug target/107969] ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107969

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
Sure, it's a result of option fuzzing, I admit.

[Bug target/107969] ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107969

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

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

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 54014
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54014&action=edit
gcc13-pr107969.patch

Indeed, because libgcc doesn't define the functions that soft-float uses.
Anyway, this patch should fix that.

[Bug target/107969] ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107969

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
I mean should fix the ICE, nothing else.

[Bug target/107969] ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107969

--- 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:bcf89f05f7b0b1732553e224548f1de55321e437

commit r13-4505-gbcf89f05f7b0b1732553e224548f1de55321e437
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Tue Dec 6 12:16:37 2022 +0100

    i386: Fix up expander conditions on cbranchbf4 and cstorebf4 [PR107969]

    With -msoft-float we ICE on __bf16 comparisons, because the
    insns we want to use under the hood (cbranchsf4 and cstoresf4)
    after performing the fast extensions aren't available.

    The following patch copies the conditions from the c*sf4 expanders
    to the corresponding c*bf4 expanders.

    2022-12-06  Jakub Jelinek  <jakub@redhat.com>

            PR target/107969
            * config/i386/i386.md (cbranchbf4, cstorebf4): Guard expanders
            with the same condition as cbranchsf4 or cstoresf4 expanders.

            * gcc.target/i386/pr107969.c: New test.

[Bug target/107969] [13 Regression] ICE in extract_insn, at recog.cc:2791 since r13-3292-gc2565a31c1622ab0

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107969

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED
            Summary|ICE in extract_insn, at     |[13 Regression] ICE in
                   |recog.cc:2791 since         |extract_insn, at
                   |r13-3292-gc2565a31c1622ab0  |recog.cc:2791 since
                   |                            |r13-3292-gc2565a31c1622ab0

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed.