[Bug tree-optimization/103311] New: [12 Regression] ICE in complex_mul_pattern::build(vec_info*)

[Bug tree-optimization/103311] New: [12 Regression] ICE in complex_mul_pattern::build(vec_info*)

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

            Bug ID: 103311
           Summary: [12 Regression] ICE in
                    complex_mul_pattern::build(vec_info*)
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: asolokha at gmx dot com
  Target Milestone: ---
            Target: aarch64-linux-gnu

gcc-12.0.0-alpha20111114 snapshot (g:3057f1ab737582a9fb37a3fb967ed8bf3659f2f4)
ICEs when compiling the following testcase w/ -march=armv8.3-a -O1
-ftree-slp-vectorize:

_Complex double y;

void
cbknu (_Complex double f)
{
  const _Complex double cone = 1.0e0;

  f = f * cone;
  y = f * cone;
}

% aarch64-linux-gnu-gcc-12.0.0 -march=armv8.3-a -O1 -ftree-slp-vectorize -c
cbknu.c
during GIMPLE pass: slp
cbknu.c: In function 'cbknu':
cbknu.c:4:1: internal compiler error: in operator[], at vec.h:889
    4 | cbknu (_Complex double f)
      | ^~~~~
0x7e18be vec<tree_node*, va_heap, vl_embed>::operator[](unsigned int)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/vec.h:889
0x7e1a73 vec<_slp_tree*, va_heap, vl_embed>::operator[](unsigned int)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp-patterns.c:1119
0x7e1a73 vec<_slp_tree*, va_heap, vl_ptr>::operator[](unsigned int)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/vec.h:1495
0x7e1a73 complex_mul_pattern::build(vec_info*)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp-patterns.c:1083
0x11f8b00 vect_match_slp_patterns_2
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:2915
0x11f8aa6 vect_match_slp_patterns_2
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:2907
0x11f8aa6 vect_match_slp_patterns_2
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:2907
0x11fce1e vect_match_slp_patterns
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:2943
0x11fce1e vect_analyze_slp(vec_info*, unsigned int)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:3426
0x12030dc vect_slp_analyze_bb_1
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:5744
0x12030dc vect_slp_region
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:5846
0x1204b24 vect_slp_bbs
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:6038
0x1204f68 vect_slp_function(function*)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vect-slp.c:6126
0x120c28a execute
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-12.0.0_alpha20211114/work/gcc-12-20211114/gcc/tree-vectorizer.c:1459

[Bug tree-optimization/103311] [12 Regression] ICE in complex_mul_pattern::build(vec_info*)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tnfchris at gcc dot gnu.org
   Target Milestone|---                         |12.0

[Bug tree-optimization/103311] [12 Regression] ICE in complex_mul_pattern::build(vec_info*) since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-11-18
             Status|UNCONFIRMED                 |NEW
            Summary|[12 Regression] ICE in      |[12 Regression] ICE in
                   |complex_mul_pattern::build( |complex_mul_pattern::build(
                   |vec_info*)                  |vec_info*) since
                   |                            |r12-4785-ged3de62ac949c92ad
                   |                            |41ef6de7cc926fbb2a510ce
                 CC|                            |marxin at gcc dot gnu.org

--- Comment #1 from Martin Liška <marxin at gcc dot gnu.org> ---
Started with r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce.

[Bug tree-optimization/103311] [12 Regression] ICE in complex_mul_pattern::build(vec_info*) since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

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

--- Comment #2 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
Both sides of the VEC_PERM_EXPR need to be a multiply,

the check is currently

 if (!mul0 && !mul1)
    return IFN_LAST;

while it should be

 if (!mul0 || !mul1)
    return IFN_LAST;

testing a patch.

[Bug tree-optimization/103311] [12 Regression] ICE in complex_mul_pattern::build(vec_info*) since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

--- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Tamar Christina <tnfchris@gcc.gnu.org>:

https://gcc.gnu.org/g:4f0a2f5a3ddb1024b885c066a18caae4d733bb6c

commit r12-5378-g4f0a2f5a3ddb1024b885c066a18caae4d733bb6c
Author: Tamar Christina <tamar.christina@arm.com>
Date:   Thu Nov 18 17:10:36 2021 +0000

    middle-end: check that both sides of complex expression is a mul.

    Both sides of the VEC_PERM_EXPR need to be a MULT but the check
    was accidentally checking if both sides are a mul.

    The FMS case would be handled by the validate_multiplication but
    this makes the requirement more explicit and we exit earlier.

    gcc/ChangeLog:

            PR tree-optimization/103311
            * tree-vect-slp-patterns.c (complex_mul_pattern::matches,
            complex_fms_pattern::matches): Check for multiplications.

    gcc/testsuite/ChangeLog:

            PR tree-optimization/103311
            * gcc.target/aarch64/pr103311.c: New test.

[Bug tree-optimization/103311] [12 Regression] ICE in complex_mul_pattern::build(vec_info*) since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

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

--- Comment #4 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
Fixed on trunk.

[Bug tree-optimization/103311] [12 Regression] ICE in complex_mul_pattern::build(vec_info*) since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

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

--- Comment #5 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
MLA ended up getting not detected, for some reason the testcases where getting
skipped.

[Bug tree-optimization/103311] [12 Regression] ICE in complex_mul_pattern::build(vec_info*) since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Tamar Christina <tnfchris@gcc.gnu.org>:

https://gcc.gnu.org/g:487d604b6fa0f0a981eadc216d9e481d08ed7e7b

commit r12-5407-g487d604b6fa0f0a981eadc216d9e481d08ed7e7b
Author: Tamar Christina <tamar.christina@arm.com>
Date:   Fri Nov 19 15:12:38 2021 +0000

    middle-end: Handle FMA_CONJ correctly after SLP layout update.

    Apologies, I got dinged by the i386 regressions bot for a test I didn't
have in
    my tree at the time I made the previous patch.  The bot was telling me that
FMA
    stopped working after I strengthened the FMA check in the previous patch.

    The reason is that the check is slightly early.  The first check can indeed
only
    exit early when either node isn't a mult.  However we need to delay till we
know
    if the node is a MUL or FMA before enforcing that both nodes must be a MULT
    since the node to inspect is different if the operation is a MUL or FMA.

    Also with the update patch for GCC 11 tree layout update to the new GCC 12
one
    I had missed that the difference in which node is conjucated is not
symmetrical.

    So the test for it can just be testing the inverse order.  It was Currently
    no detecting when the first node was conjucated instead of the second one.

    This also made me wonder why my own test didn't detect this.  It turns out
that
    the tests, being copied from the _Float16 ones were incorrectly marked as
    xfail.  The _Float16 ones are marked as xfail since C doesn't have a conj
    operation for _Float16, which means you get extra type-casts in between.

    While you could use the GCC _Complex extension here I opted to mark them
xfail
    since I wanted to include detection over the widenings next year.

    Secondly the double tests were being skipped because Adv. SIMD was missing
from
    targets supporting Complex Double vectorization.

    With these changes all other tests run and pass and only XFAIL ones are
    correctly the _Float16 ones.  Sorry for missing this before, testing should
now
    cover all cases.

    gcc/ChangeLog:

            PR tree-optimization/103311
            PR target/103330
            * tree-vect-slp-patterns.c (vect_validate_multiplication): Fix CONJ
            test to new codegen.
            (complex_mul_pattern::matches): Move check downwards.

    gcc/testsuite/ChangeLog:

            PR tree-optimization/103311
            PR target/103330
            * gcc.dg/vect/complex/fast-math-bb-slp-complex-mla-double.c: Fix
it.
            * gcc.dg/vect/complex/fast-math-bb-slp-complex-mla-float.c:
Likewise.
            * gcc.dg/vect/complex/fast-math-bb-slp-complex-mls-double.c:
Likewise.
            * gcc.dg/vect/complex/fast-math-bb-slp-complex-mls-float.c:
Likewise.
            * gcc.dg/vect/complex/fast-math-bb-slp-complex-mul-double.c:
Likewise.
            * gcc.dg/vect/complex/fast-math-bb-slp-complex-mul-float.c:
Likewise.
            * lib/target-supports.exp
            (check_effective_target_vect_complex_add_double): Add Adv. SIMD.

[Bug tree-optimization/103311] [12 Regression] ICE in complex_mul_pattern::build(vec_info*) since r12-4785-ged3de62ac949c92ad41ef6de7cc926fbb2a510ce

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103311

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

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

--- Comment #7 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
Really fixed now.