[Bug target/103625] New: ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[Bug target/103625] New: ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

            Bug ID: 103625
           Summary: ICE: in extract_insn, at recog.c:2769 (error:
                    unrecognizable insn)
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: asolokha at gmx dot com
  Target Milestone: ---
            Target: powerpc-e300c3-linux-gnu

gcc 12.0.0 20211205 snapshot (g:c9419faef0bfaf31e6a6f744baa064892e0d105c) ICEs
when compiling the following testcase, reduced from
gcc/testsuite/gcc.target/powerpc/pr70963.c, w/ -mvsx:

#define vector __attribute__((altivec(vector__)))

static void test()
{
  vector double x = { 81.0, 76.0 };
  vector long long y = { 81, 76 };
  vector long long a = __builtin_vec_cts (x, 0);

  if (!__builtin_vec_vcmpeq_p (2, a, y))
    __builtin_abort ();
}

% powerpc-e300c3-linux-gnu-gcc-12.0.0 -mvsx -c m3jai565.c
m3jai565.c: In function 'test':
m3jai565.c:11:1: error: unrecognizable insn:
   11 | }
      | ^
(insn 16 15 17 2 (parallel [
            (set (reg:CC 106 6)
                (unspec:CC [
                        (eq:CC (reg:V2DI 124)
                            (reg:V2DI 125))
                    ] UNSPEC_PREDICATE))
            (set (reg:V2DI 126)
                (eq:V2DI (reg:V2DI 124)
                    (reg:V2DI 125)))
        ]) "m3jai565.c":9:7 -1
     (nil))
during RTL pass: vregs
m3jai565.c:11:1: internal compiler error: in extract_insn, at recog.c:2769
0x6ac6f5 _fatal_insn(char const*, rtx_def const*, char const*, int, char
const*)
       
/var/tmp/portage/cross-powerpc-e300c3-linux-gnu/gcc-12.0.0_p20211205/work/gcc-12-20211205/gcc/rtl-error.c:108
0x6ac715 _fatal_insn_not_found(rtx_def const*, char const*, int, char const*)
       
/var/tmp/portage/cross-powerpc-e300c3-linux-gnu/gcc-12.0.0_p20211205/work/gcc-12-20211205/gcc/rtl-error.c:116
0x6aad23 extract_insn(rtx_insn*)
       
/var/tmp/portage/cross-powerpc-e300c3-linux-gnu/gcc-12.0.0_p20211205/work/gcc-12-20211205/gcc/recog.c:2769
0xb24c4c instantiate_virtual_regs_in_insn
       
/var/tmp/portage/cross-powerpc-e300c3-linux-gnu/gcc-12.0.0_p20211205/work/gcc-12-20211205/gcc/function.c:1660
0xb24c4c instantiate_virtual_regs
       
/var/tmp/portage/cross-powerpc-e300c3-linux-gnu/gcc-12.0.0_p20211205/work/gcc-12-20211205/gcc/function.c:1985
0xb24c4c execute
       
/var/tmp/portage/cross-powerpc-e300c3-linux-gnu/gcc-12.0.0_p20211205/work/gcc-12-20211205/gcc/function.c:2034

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[linkw at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

Kewen Lin <linkw at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |linkw at gcc dot gnu.org,
                   |                            |wschmidt at gcc dot gnu.org

--- Comment #1 from Kewen Lin <linkw at gcc dot gnu.org> ---
Confirmed, started to fail from r12-5752.

The __builtin_vec_vcmpeq_p for doubleword overload requires power8-vector. Its
current stanza is vsx, but should be power8-vector. The related bifs are:

  const vsll __builtin_altivec_vcmpequd (vull, vull);
  const int __builtin_altivec_vcmpequd_p (int, vsll, vsll);

  const vsll __builtin_altivec_vcmpgtsd (vsll, vsll);
  const int __builtin_altivec_vcmpgtsd_p (int, vsll, vsll);

  const vsll __builtin_altivec_vcmpgtud (vull, vull);
  const int __builtin_altivec_vcmpgtud_p (int, vsll, vsll);

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

Bill Schmidt <wschmidt at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |wschmidt at gcc dot gnu.org
   Target Milestone|---                         |12.0

--- Comment #2 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Mine.

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

--- Comment #3 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
BTW, Arseny, please CC me on any other built-in issues you see.

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

--- Comment #4 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Kewen, how did you confirm this?  My cross doesn't accept -mvsx as valid.

$ /home/wschmidt/gcc/build/gcc-e300/gcc/xgcc -c -O2 -mvsx pr103625.c
-B/home/wschmidt/gcc/build/gcc-e300/gcc 
/home/wschmidt/gcc/build/gcc-e300/gcc/as: line 114: exec: -m: invalid option
exec: usage: exec [-cl] [-a name] [command [arguments ...]] [redirection ...]

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

Bill Schmidt <wschmidt at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-12-12
             Status|UNCONFIRMED                 |NEW

--- Comment #5 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Confirmed by Kewen, so setting status.

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[linkw at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

--- Comment #6 from Kewen Lin <linkw at gcc dot gnu.org> ---
(In reply to Bill Schmidt from comment #4)
> Kewen, how did you confirm this?  My cross doesn't accept -mvsx as valid.
> 
> $ /home/wschmidt/gcc/build/gcc-e300/gcc/xgcc -c -O2 -mvsx pr103625.c
> -B/home/wschmidt/gcc/build/gcc-e300/gcc 
> /home/wschmidt/gcc/build/gcc-e300/gcc/as: line 114: exec: -m: invalid option
> exec: usage: exec [-cl] [-a name] [command [arguments ...]] [redirection ...]

I used the original command which doesn't specify -O2. I just confirmed that
it's reproducible with that, just noticed that with O2 all seems optimized out,
then it's going to invoke assembler (-S can make it silent).

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

--- Comment #7 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Oh, duh, sorry.  Yes, I can reproduce as well.

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

--- Comment #8 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
And Kewen's analysis is spot-on, will fix that.

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

--- Comment #9 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Patch posted here:

https://gcc.gnu.org/pipermail/gcc-patches/2021-December/586713.html

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

--- Comment #10 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by William Schmidt <wschmidt@gcc.gnu.org>:

https://gcc.gnu.org/g:936051f9241ee2eafae8f5b8a4ad99fd7ed693bc

commit r12-5964-g936051f9241ee2eafae8f5b8a4ad99fd7ed693bc
Author: Bill Schmidt <wschmidt@linux.ibm.com>
Date:   Mon Dec 13 09:30:18 2021 -0600

    rs6000: Builtins for doubleword compare should be in [power8-vector]
(PR103625)

    2021-12-13  Bill Schmidt  <wschmidt@linux.ibm.com>

    gcc/
            PR target/103625
            * config/rs6000/rs6000-builtin-new.def
(__builtin_altivec_vcmpequd):
            Move to power8-vector stanza.
            (__builtin_altivec_vcmpequd_p): Likewise.
            (__builtin_altivec_vcmpgtsd): Likewise.
            (__builtin_altivec_vcmpgtsd_p): Likewise.
            (__builtin_altivec_vcmpgtud): Likewise.
            (__builtin_altivec_vcmpgtud_p): Likewise.

[Bug target/103625] ICE: in extract_insn, at recog.c:2769 (error: unrecognizable insn)

[wschmidt at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103625

Bill Schmidt <wschmidt at gcc dot gnu.org> changed:

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

--- Comment #11 from Bill Schmidt <wschmidt at gcc dot gnu.org> ---
Fixed.