[Bug middle-end/68134] New: vector-of-one-int64_t comparison ICE on aarch64-none-elf

[Bug middle-end/68134] New: vector-of-one-int64_t comparison ICE on aarch64-none-elf

[jgreenhalgh at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68134

            Bug ID: 68134
           Summary: vector-of-one-int64_t comparison ICE on
                    aarch64-none-elf
           Product: gcc
           Version: 6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: jgreenhalgh at gcc dot gnu.org
  Target Milestone: ---

Created attachment 36605
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36605&action=edit
Failing testcase

After r229128 I get an ICE at O0 on aarch64 targets when compiling the attached
testcase:

$ gcc random-test.c -O0

In file included from random-test.c:1:0:
..../build-aarch64-none-elf/install/lib/gcc/aarch64-none-elf/6.0.0/include/arm_neon.h:
In function 'foo':
..../build-aarch64-none-elf/install/lib/gcc/aarch64-none-elf/6.0.0/include/arm_neon.h:11941:28:
internal compiler error: in prepare_cmp_insn, at optabs.c:3820
   return (uint64x1_t) (__a == __b);
                            ^
0x996e40 prepare_cmp_insn
        ..../src/gcc/gcc/optabs.c:3820
0x997396 emit_cmp_and_jump_insns(rtx_def*, rtx_def*, rtx_code, rtx_def*,
machine_mode, int, rtx_def*, int)
        ..../src/gcc/gcc/optabs.c:3967
0x72e934 do_compare_rtx_and_jump(rtx_def*, rtx_def*, rtx_code, int,
machine_mode, rtx_def*, rtx_code_label*, rtx_code_label*, int)
        ..../src/gcc/gcc/dojump.c:1149
0x72f52a do_compare_and_jump
        ..../src/gcc/gcc/dojump.c:1228
0x73134f do_jump_1(tree_code, tree_node*, tree_node*, rtx_code_label*,
rtx_code_label*, int)
        ..../src/gcc/gcc/dojump.c:240
0x731c68 jumpifnot_1(tree_code, tree_node*, tree_node*, rtx_code_label*, int)
        ..../src/gcc/gcc/dojump.c:148
0x7c6e47 expand_expr_real_2(separate_ops*, rtx_def*, machine_mode,
expand_modifier)
        ..../src/gcc/gcc/expr.c:9030
0x6c5ad9 expand_gimple_stmt_1
        ..../src/gcc/gcc/cfgexpand.c:3617
0x6c5ad9 expand_gimple_stmt
        ..../src/gcc/gcc/cfgexpand.c:3677
0x6c8316 expand_gimple_basic_block
        ..../src/gcc/gcc/cfgexpand.c:5681
0x6cb8c4 execute
        ..../src/gcc/gcc/cfgexpand.c:6293
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <http://gcc.gnu.org/bugs.html> for instructions.

[Bug middle-end/68134] [6 Regression] vector-of-one-int64_t comparison ICE on aarch64-none-elf

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68134

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-10-28
                 CC|                            |ktkachov at gcc dot gnu.org
            Summary|vector-of-one-int64_t       |[6 Regression]
                   |comparison ICE on           |vector-of-one-int64_t
                   |aarch64-none-elf            |comparison ICE on
                   |                            |aarch64-none-elf
     Ever confirmed|0                           |1
      Known to fail|                            |6.0

--- Comment #1 from ktkachov at gcc dot gnu.org ---
Confirmed

[Bug middle-end/68134] [6 Regression] vector-of-one-int64_t comparison ICE on aarch64-none-elf

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68134

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |6.0

[Bug middle-end/68134] [6 Regression] float64x1_t comparison ICE on aarch64-none-elf

[jgreenhalgh at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68134

--- Comment #2 from James Greenhalgh <jgreenhalgh at gcc dot gnu.org> ---
Created attachment 36608
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36608&action=edit
testcasem, no arm_neon.h dependency

Second testcase stripping the arm_neon.h stuff and using only GCC vector
extensions. I couldn't get this to fail on s390 or i386 (two other targets
which V1DF mode stuff in their back-end).