[Bug c/106016] New: [PowerPC] crash with attempt to initialize array of MMA accumulators

[Bug c/106016] New: [PowerPC] crash with attempt to initialize array of MMA accumulators

[nemanja.i.ibm at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

            Bug ID: 106016
           Summary: [PowerPC] crash with attempt to initialize array of
                    MMA accumulators
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: nemanja.i.ibm at gmail dot com
  Target Milestone: ---

Code:
$ cat aa.c 
void array_crash(__vector_quad *a, __vector_quad *b) {
  __vector_quad arr[2] = {*a, *b}; // Should not crash
}

Compile:
$ /opt/gcc-nightly/12/bin/gcc -mcpu=power10 -O3 aa.c -S
aa.c: In function 'array_crash':
aa.c:2:17: internal compiler error: in count_type_elements, at expr.cc:6407
    2 |   __vector_quad arr[2] = {*a, *b}; // Should not crash
      |                 ^~~
0x105583df count_type_elements
        /home/gccbuild/gcc_12_git/gcc/gcc/expr.cc:6407
0x1055cd3b categorize_ctor_elements_1
        /home/gccbuild/gcc_12_git/gcc/gcc/expr.cc:6519
0x10646293 gimplify_init_constructor
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:5179
0x10647643 gimplify_modify_expr
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:6040
0x10639d93 gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:15098
0x10643fd3 gimplify_stmt(tree_node**, gimple**)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:7151
0x10643fd3 gimplify_and_add(tree_node*, gimple**)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:496
0x10643fd3 gimplify_decl_expr
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:1936
0x1063ad0b gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:15295
0x10642b6f gimplify_stmt(tree_node**, gimple**)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:7151
0x10642b6f gimplify_bind_expr
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:1428
0x1063a79b gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:15299
0x1063e3d7 gimplify_stmt(tree_node**, gimple**)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:7151
0x1063e3d7 gimplify_body(tree_node*, bool)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:16355
0x1063e83b gimplify_function_tree(tree_node*)
        /home/gccbuild/gcc_12_git/gcc/gcc/gimplify.cc:16509
0x1044cab7 cgraph_node::analyze()
        /home/gccbuild/gcc_12_git/gcc/gcc/cgraphunit.cc:676
0x1044fbf7 analyze_functions
        /home/gccbuild/gcc_12_git/gcc/gcc/cgraphunit.cc:1241
0x10450893 symbol_table::finalize_compilation_unit()
        /home/gccbuild/gcc_12_git/gcc/gcc/cgraphunit.cc:2501
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.

Version:
$ /opt/gcc-nightly/12/bin/gcc --version
gcc (GCC) 12.1.1 20220524 [releases/gcc-12 r12-8410-gf0a0aeec44]
Copyright (C) 2022 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

[Bug target/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

Segher Boessenkool <segher at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2022-06-17

--- Comment #1 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Confirmed.  Thanks for reporting these bugs!

[Bug target/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #2 from Peter Bergner <bergner at gcc dot gnu.org> ---
Given where this is ICEing, I'm guessing this is non-target specific issue in
handling opaque types.

[Bug target/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #3 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Yeah.  It should just return 1 like the other scalar types?

[Bug target/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #4 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Segher Boessenkool from comment #3)
> Yeah.  It should just return 1 like the other scalar types?

So the code did look for OPAQUE_TYPE and expected never to see it, so it was on
an error path.  I agree with your comment above and I'm guessing we want
something like:

diff --git a/gcc/expr.cc b/gcc/expr.cc
index 78c839ab425..1675198a146 100644
--- a/gcc/expr.cc
+++ b/gcc/expr.cc
@@ -6423,13 +6423,13 @@ count_type_elements (const_tree type, bool for_ctor_p)
     case OFFSET_TYPE:
     case REFERENCE_TYPE:
     case NULLPTR_TYPE:
+    case OPAQUE_TYPE:
       return 1;

     case ERROR_MARK:
       return 0;

     case VOID_TYPE:
-    case OPAQUE_TYPE:
     case METHOD_TYPE:
     case FUNCTION_TYPE:
     case LANG_TYPE:

[Bug target/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #5 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Peter Bergner from comment #4)
> diff --git a/gcc/expr.cc b/gcc/expr.cc
> index 78c839ab425..1675198a146 100644
> --- a/gcc/expr.cc
> +++ b/gcc/expr.cc
> @@ -6423,13 +6423,13 @@ count_type_elements (const_tree type, bool
> for_ctor_p)
>      case OFFSET_TYPE:
>      case REFERENCE_TYPE:
>      case NULLPTR_TYPE:
> +    case OPAQUE_TYPE:
>        return 1;
>  
>      case ERROR_MARK:
>        return 0;
>  
>      case VOID_TYPE:
> -    case OPAQUE_TYPE:
>      case METHOD_TYPE:
>      case FUNCTION_TYPE:
>      case LANG_TYPE:

It does fix the ICE.

[Bug target/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #6 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Like that yes.  Pre-approved if it survives regcheck, too.  Thanks!

Please add the testcase as well of course :-)

[Bug target/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

Peter Bergner <bergner at gcc dot gnu.org> changed:

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

--- Comment #7 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Segher Boessenkool from comment #6)
> Like that yes.  Pre-approved if it survives regcheck, too.  Thanks!
> 
> Please add the testcase as well of course :-)

Ok, bootstrap and regtests are running.

[Bug target/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #8 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Peter Bergner from comment #7)
> (In reply to Segher Boessenkool from comment #6)
> > Like that yes.  Pre-approved if it survives regcheck, too.  Thanks!
> > 
> > Please add the testcase as well of course :-)
> 
> Ok, bootstrap and regtests are running.

Bootstrap and regression testing were clean.

[Bug c/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|target                      |c

--- Comment #9 from Peter Bergner <bergner at gcc dot gnu.org> ---
This is not a target issue, so changing the Component to C.

[Bug middle-end/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

Segher Boessenkool <segher at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|target                      |middle-end

--- Comment #10 from Segher Boessenkool <segher at gcc dot gnu.org> ---
No, this is *not* a target issue.  Let's try middle-end, then.

[Bug middle-end/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #11 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Peter Bergner <bergner@gcc.gnu.org>:

https://gcc.gnu.org/g:975658b782f36dcf6eb190966d5b705977bfd5eb

commit r13-1845-g975658b782f36dcf6eb190966d5b705977bfd5eb
Author: Peter Bergner <bergner@linux.ibm.com>
Date:   Fri Jun 17 23:43:23 2022 -0500

    c: Handle initializations of opaque types [PR106016]

    The initial commit that added opaque types thought that there couldn't
    be any valid initializations for variables of these types, but the test
    case in the bug report shows that isn't true.  The solution is to handle
    OPAQUE_TYPE initializations like the other scalar types.

    2022-06-17  Peter Bergner  <bergner@linux.ibm.com>

    gcc/
            PR c/106016
            * expr.cc (count_type_elements): Handle OPAQUE_TYPE.

    gcc/testsuite/
            PR c/106016
            * gcc.target/powerpc/pr106016.c: New test.

[Bug middle-end/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #12 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Peter Bergner
<bergner@gcc.gnu.org>:

https://gcc.gnu.org/g:6aaaf20ee4ad9c85f3099ef425720547644fb08d

commit r12-8682-g6aaaf20ee4ad9c85f3099ef425720547644fb08d
Author: Peter Bergner <bergner@linux.ibm.com>
Date:   Fri Jun 17 23:43:23 2022 -0500

    c: Handle initializations of opaque types [PR106016]

    The initial commit that added opaque types thought that there couldn't
    be any valid initializations for variables of these types, but the test
    case in the bug report shows that isn't true.  The solution is to handle
    OPAQUE_TYPE initializations like the other scalar types.

    2022-06-17  Peter Bergner  <bergner@linux.ibm.com>

    gcc/
            PR c/106016
            * expr.cc (count_type_elements): Handle OPAQUE_TYPE.

    gcc/testsuite/
            PR c/106016
            * gcc.target/powerpc/pr106016.c: New test.

    (cherry picked from commit 975658b782f36dcf6eb190966d5b705977bfd5eb)

[Bug middle-end/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

--- Comment #13 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Peter Bergner
<bergner@gcc.gnu.org>:

https://gcc.gnu.org/g:df8567535a1f3eaa20b700be62374b2fb4f09204

commit r11-10201-gdf8567535a1f3eaa20b700be62374b2fb4f09204
Author: Peter Bergner <bergner@linux.ibm.com>
Date:   Fri Jun 17 23:43:23 2022 -0500

    c: Handle initializations of opaque types [PR106016]

    The initial commit that added opaque types thought that there couldn't
    be any valid initializations for variables of these types, but the test
    case in the bug report shows that isn't true.  The solution is to handle
    OPAQUE_TYPE initializations like the other scalar types.

    2022-06-17  Peter Bergner  <bergner@linux.ibm.com>

    gcc/
            PR c/106016
            * expr.c (count_type_elements): Handle OPAQUE_TYPE.

    gcc/testsuite/
            PR c/106016
            * gcc.target/powerpc/pr106016.c: New test.

    (cherry picked from commit 975658b782f36dcf6eb190966d5b705977bfd5eb)

[Bug middle-end/106016] [PowerPC] crash with attempt to initialize array of MMA accumulators

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106016

Peter Bergner <bergner at gcc dot gnu.org> changed:

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

--- Comment #14 from Peter Bergner <bergner at gcc dot gnu.org> ---
Fixed everywhere.