[Bug debug/99808] New: [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[Bug debug/99808] New: [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

            Bug ID: 99808
           Summary: [8/9/10/11 Regression] ICE in as_a, at machmode.h:365
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: asolokha at gmx dot com
  Target Milestone: ---
            Target: aarch64-linux-gnu

gcc-11.0.1-alpha20210328 snapshot (g:499fa254ae8c9752d8c2cf3130b13ffddfd83546)
ICEs when compiling the following testcase, reduced from
gcc/testsuite/gcc.target/aarch64/advsimd-intrinsics/vcvt_high_1.c, w/ -Og -fweb
-fno-forward-propagate -g:

typedef __Float32x2_t float32x2_t;
typedef __Float32x4_t float32x4_t;
typedef __Float64x1_t float64x1_t;
typedef __Float64x2_t float64x2_t;

float32x4_t
foo (void)
{
  float64x2_t arg2 = __builtin_aarch64_combinedf (0, ((float64x1_t) 1UL) [0]);

  return __builtin_aarch64_float_truncate_hi_v4sf ((float32x2_t) 1UL, arg2);
}

% aarch64-linux-gnu-gcc-11.0.1 -Og -fweb -fno-forward-propagate -g -c
bhwx2dh3.c
during RTL pass: final
bhwx2dh3.c: In function 'foo':
bhwx2dh3.c:12:1: internal compiler error: in as_a, at machmode.h:365
   12 | }
      | ^
0x6c306f scalar_int_mode as_a<scalar_int_mode>(machine_mode)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/machmode.h:365
0x6c467f scalar_float_mode as_a<scalar_float_mode>(machine_mode)
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/tree.h:3467
0x6c467f insert_float
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:19697
0xacb35e add_const_value_attribute
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:19805
0xaef2e8 add_location_or_const_value_attribute
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:20367
0xaf0b52 gen_variable_die
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:24104
0xad3272 gen_decl_die
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:26516
0xaf497f decls_for_scope
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:25991
0xacf194 gen_subprogram_die
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:23522
0xad3360 gen_decl_die
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:26428
0xad4a22 dwarf2out_decl
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:27006
0xad539e dwarf2out_function_decl
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/dwarf2out.c:27021
0xb5ce43 rest_of_handle_final
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/final.c:4712
0xb5ce43 execute
       
/var/tmp/portage/cross-aarch64-linux-gnu/gcc-11.0.1_alpha20210328/work/gcc-11-20210328/gcc/final.c:4754

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2021-03-29
   Target Milestone|---                         |8.5
           Keywords|                            |wrong-code
          Component|debug                       |target
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
           Priority|P3                          |P2

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  We're having a

(note 32 30 31 (var_location arg2 (const_vector:V2DF [
        (const_double:DF 0.0 [0x0.0p+0])
        (const_int 0 [0])
    ])) NOTE_INSN_VAR_LOCATION)

note which obviously will fail a scalar_float_mode assessment.  This first
appears in CSE1 as

(insn 8 7 9 2 (set (reg/v:V2DF 92 [ arg2 ])
        (const_vector:V2DF [
                (const_double:DF 0.0 [0x0.0p+0])
                (const_int 0 [0])
            ])) "t.c":9:22 1167 {*aarch64_simd_movv2df}
     (expr_list:REG_DEAD (reg:DF 94)
        (nil)))

and is likely a backend pattern issue.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Started with r7-8376-g726e7a70b911f4676de4a97b19e042552ceedd17

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Sorry, meant r8-4318-g26895c21eb10cfd6c00285e13e6f13a75cccc1d9

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
The bug is clearly on the aarch64 backend side,
(insn 7 6 8 (set (reg/v:V2DF 73 [ arg2 ])
        (vec_concat:V2DF (reg:DF 75)
            (const_int 0 [0]))) "pr99808.c":9 -1
     (nil))

(insn 8 7 0 (set (reg/v:V2DF 73 [ arg2 ])
        (vec_concat:V2DF (vec_select:DF (reg/v:V2DF 73 [ arg2 ])
                (parallel:V2DF [
                        (const_int 0 [0])
                    ]))
            (reg:DF 76))) "pr99808.c":9 -1
     (nil))
is IMNSHO invalid RTL in both cases.  const0_rtx is not valid constant for
DFmode.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

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 #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 50482
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=50482&action=edit
gcc11-pr99808.patch

Untested fix.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
                 CC|                            |ktkachov at gcc dot gnu.org
         Resolution|---                         |DUPLICATE

--- Comment #6 from ktkachov at gcc dot gnu.org ---
This is a dup of PR 99037 then. I was planning on fixing it in GCC 12 but since
there is an ICE reproducer now I can push it in now

*** This bug has been marked as a duplicate of bug 99037 ***

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Kyrylo Tkachov <ktkachov@gcc.gnu.org>:

https://gcc.gnu.org/g:37d9074e12082132ae62c12fbe958c697f638c0a

commit r11-7888-g37d9074e12082132ae62c12fbe958c697f638c0a
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Mon Mar 29 11:52:24 2021 +0100

    aarch64: PR target/99037 Fix RTL represntation in move_lo_quad patterns

    This patch fixes the RTL representation of the move_lo_quad patterns to use
aarch64_simd_or_scalar_imm_zero
    for the zero part rather than a vec_duplicate of zero or a const_int 0.
    The expander that generates them is also adjusted so that we use and match
the correct const_vector forms throughout.

    Co-Authored-By: Jakub Jelinek <jakub@redhat.com>

    gcc/ChangeLog:

            PR target/99037
            * config/aarch64/aarch64-simd.md (move_lo_quad_internal_<mode>):
Use
            aarch64_simd_or_scalar_imm_zero to match zeroes.  Remove pattern
            matching const_int 0.
            (move_lo_quad_internal_be_<mode>): Likewise.
            (move_lo_quad_<mode>): Update for the above.
            * config/aarch64/iterators.md (VQ_2E): Delete.

    gcc/testsuite/ChangeLog:

            PR target/99808
            * gcc.target/aarch64/pr99808.c: New test.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Kyrylo Tkachov <ktkachov@gcc.gnu.org>:

https://gcc.gnu.org/g:41d57b2a97c44ae7b0a5b01ae703a8f0d0495238

commit r11-7907-g41d57b2a97c44ae7b0a5b01ae703a8f0d0495238
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Tue Mar 30 14:07:50 2021 +0100

    aarch64: Fix gcc.target/aarch64/pr99808.c for ILP32

    Fix test for -mabi=ilp32

    gcc/testsuite/ChangeLog:

            PR target/99808
            * gcc.target/aarch64/pr99808.c: Use ULL constant suffix.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Kyrylo Tkachov
<ktkachov@gcc.gnu.org>:

https://gcc.gnu.org/g:1a92899b08e61d503a2897f2f66b064eb84706bc

commit r10-9628-g1a92899b08e61d503a2897f2f66b064eb84706bc
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Mon Mar 29 11:52:24 2021 +0100

    aarch64: PR target/99037 Fix RTL represntation in move_lo_quad patterns

    This patch fixes the RTL representation of the move_lo_quad patterns to use
aarch64_simd_or_scalar_imm_zero
    for the zero part rather than a vec_duplicate of zero or a const_int 0.
    The expander that generates them is also adjusted so that we use and match
the correct const_vector forms throughout.

    Co-Authored-By: Jakub Jelinek <jakub@redhat.com>

    gcc/ChangeLog:

            PR target/99037
            * config/aarch64/aarch64-simd.md (move_lo_quad_internal_<mode>):
Use
            aarch64_simd_or_scalar_imm_zero to match zeroes.  Remove pattern
            matching const_int 0.
            (move_lo_quad_internal_be_<mode>): Likewise.
            (move_lo_quad_<mode>): Update for the above.
            * config/aarch64/iterators.md (VQ_2E): Delete.

    gcc/testsuite/ChangeLog:

            PR target/99808
            * gcc.target/aarch64/pr99808.c: New test.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #10 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Kyrylo Tkachov
<ktkachov@gcc.gnu.org>:

https://gcc.gnu.org/g:c611209a3422d2a2dc10bc804986db9bfb80a62e

commit r10-9629-gc611209a3422d2a2dc10bc804986db9bfb80a62e
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Tue Mar 30 14:07:50 2021 +0100

    aarch64: Fix gcc.target/aarch64/pr99808.c for ILP32

    Fix test for -mabi=ilp32

    gcc/testsuite/ChangeLog:

            PR target/99808
            * gcc.target/aarch64/pr99808.c: Use ULL constant suffix.

    (cherry picked from commit 41d57b2a97c44ae7b0a5b01ae703a8f0d0495238)

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #11 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Kyrylo Tkachov
<ktkachov@gcc.gnu.org>:

https://gcc.gnu.org/g:3a9504e1f03bc2980062dc62261212e8635bcf93

commit r9-9519-g3a9504e1f03bc2980062dc62261212e8635bcf93
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Fri May 7 11:35:29 2021 +0100

    aarch64: PR target/99037 Fix RTL represntation in move_lo_quad patterns

    This patch fixes the RTL representation of the move_lo_quad patterns to use
aarch64_simd_or_scalar_imm_zero
    for the zero part rather than a vec_duplicate of zero or a const_int 0.
    The expander that generates them is also adjusted so that we use and match
the correct const_vector forms throughout.

    Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
    gcc/ChangeLog:

            PR target/99037
            PR target/100441
            * config/aarch64/aarch64-simd.md (move_lo_quad_internal_<mode>):
Use
            aarch64_simd_or_scalar_imm_zero to match zeroes.  Remove pattern
            matching const_int 0.
            (move_lo_quad_internal_be_<mode>): Likewise.
            (move_lo_quad_<mode>): Update for the above.
            * config/aarch64/iterators.md (VQ_2E): Delete.

    gcc/testsuite/ChangeLog:

            PR target/99808
            * gcc.target/aarch64/pr99808.c: New test.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #12 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-8 branch has been updated by Kyrylo Tkachov
<ktkachov@gcc.gnu.org>:

https://gcc.gnu.org/g:64b6b1d1bc796a7fea189a4c4b0e290540f51ae3

commit r8-10956-g64b6b1d1bc796a7fea189a4c4b0e290540f51ae3
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Fri May 7 11:36:54 2021 +0100

    aarch64: PR target/99037 Fix RTL represntation in move_lo_quad patterns

    This patch fixes the RTL representation of the move_lo_quad patterns to use
aarch64_simd_or_scalar_imm_zero
    for the zero part rather than a vec_duplicate of zero or a const_int 0.
    The expander that generates them is also adjusted so that we use and match
the correct const_vector forms throughout.

    Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
    gcc/ChangeLog:

            PR target/99037
            PR target/100441
            * config/aarch64/aarch64-simd.md (move_lo_quad_internal_<mode>):
Use
            aarch64_simd_or_scalar_imm_zero to match zeroes.  Remove pattern
            matching const_int 0.
            (move_lo_quad_internal_be_<mode>): Likewise.
            (move_lo_quad_<mode>): Update for the above.
            * config/aarch64/iterators.md (VQ_2E): Delete.

    gcc/testsuite/ChangeLog:

            PR target/99808
            * gcc.target/aarch64/pr99808.c: New test.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #13 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Christophe Lyon
<clyon@gcc.gnu.org>:

https://gcc.gnu.org/g:c9c429cf986c885cf90259866186849de44e1e1f

commit r9-9520-gc9c429cf986c885cf90259866186849de44e1e1f
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Tue Mar 30 14:07:50 2021 +0100

    aarch64: Fix gcc.target/aarch64/pr99808.c for ILP32

    Fix test for -mabi=ilp32

    gcc/testsuite/ChangeLog:

            PR target/99808
            * gcc.target/aarch64/pr99808.c: Use ULL constant suffix.

[Bug target/99808] [8/9/10/11 Regression] ICE in as_a, at machmode.h:365

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99808

--- Comment #14 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-8 branch has been updated by Christophe Lyon
<clyon@gcc.gnu.org>:

https://gcc.gnu.org/g:bcc4f85667c88f9be098f2671b01831d4b8d9f7c

commit r8-10957-gbcc4f85667c88f9be098f2671b01831d4b8d9f7c
Author: Kyrylo Tkachov <kyrylo.tkachov@arm.com>
Date:   Tue Mar 30 14:07:50 2021 +0100

    aarch64: Fix gcc.target/aarch64/pr99808.c for ILP32

    Fix test for -mabi=ilp32

    gcc/testsuite/ChangeLog:

            PR target/99808
            * gcc.target/aarch64/pr99808.c: Use ULL constant suffix.