[Bug fortran/99840] New: [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[Bug fortran/99840] New: [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

            Bug ID: 99840
           Summary: [9/10/11 Regression] ICE in gfc_simplify_matmul, at
                    fortran/simplify.c:4777
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Started with r8, compiles with r7 :


$ cat z1.f90
program p
   integer, parameter :: x(0,0) = 0
   integer :: y(0,0)
   y = matmul(x, transpose(x))
end


$ gfortran-7 -c z1.f90
$
$ gfortran-11-20210328 -c z1.f90
 f951: internal compiler error: Segmentation fault
0xc09d4f crash_signal
        ../../gcc/toplev.c:327
0x716f87 gfc_simplify_matmul(gfc_expr*, gfc_expr*)
        ../../gcc/fortran/simplify.c:4777
0x69b9d3 do_simplify
        ../../gcc/fortran/intrinsic.c:4664
0x6a63fa gfc_intrinsic_func_interface(gfc_expr*, int)
        ../../gcc/fortran/intrinsic.c:5050
0x6f5699 resolve_unknown_f
        ../../gcc/fortran/resolve.c:2926
0x6f5699 resolve_function
        ../../gcc/fortran/resolve.c:3270
0x6f5699 gfc_resolve_expr(gfc_expr*)
        ../../gcc/fortran/resolve.c:7105
0x6fbb24 gfc_resolve_expr(gfc_expr*)
        ../../gcc/fortran/resolve.c:11728
0x6fbb24 gfc_resolve_code(gfc_code*, gfc_namespace*)
        ../../gcc/fortran/resolve.c:11824
0x6fd0b7 resolve_codes
        ../../gcc/fortran/resolve.c:17395
0x6fd17e gfc_resolve(gfc_namespace*)
        ../../gcc/fortran/resolve.c:17430
0x6e56e4 resolve_all_program_units
        ../../gcc/fortran/parse.c:6290
0x6e56e4 gfc_parse_file()
        ../../gcc/fortran/parse.c:6542
0x7320ef gfc_be_parse_file
        ../../gcc/fortran/f95-lang.c:212

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-03-30
      Known to fail|                            |10.2.1, 11.0, 8.4.1, 9.3.1
           Priority|P3                          |P4
                 CC|                            |anlauf at gcc dot gnu.org

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

This one is funny.  The TRANSPOSE seems to screw up the simplification.
Works without TRANSPOSE.

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[kargl at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

kargl at gcc dot gnu.org changed:

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

--- Comment #2 from kargl at gcc dot gnu.org ---
Patch.  Check if a or b is zero-sized.

diff --git a/gcc/fortran/simplify.c b/gcc/fortran/simplify.c
index 388aca7c38c..6c5259c648d 100644
--- a/gcc/fortran/simplify.c
+++ b/gcc/fortran/simplify.c
@@ -4728,7 +4728,9 @@ gfc_simplify_matmul (gfc_expr *matrix_a, gfc_expr
*matrix_b)
   int stride_a, offset_a, stride_b, offset_b;

   if (!is_constant_array_expr (matrix_a)
-      || !is_constant_array_expr (matrix_b))
+      || gfc_is_size_zero_array (matrix_a)
+      || !is_constant_array_expr (matrix_b)
+      || gfc_is_size_zero_array (matrix_b))
     return NULL;

   /* MATMUL should do mixed-mode arithmetic.  Set the result type.  */

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |9.4

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

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

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

--- Comment #3 from Martin Liška <marxin at gcc dot gnu.org> ---
Started with r8-4151-g6c6bde30706c29ff.

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

--- Comment #4 from anlauf at gcc dot gnu.org ---
For reasons I do not understand,

Breakpoint 1, gfc_simplify_matmul (matrix_a=0x292bbf0, matrix_b=0x292c550)
    at ../../gcc-trunk/gcc/fortran/simplify.c:4777
4777          result_columns = mpz_get_si (matrix_b->shape[1]);
(gdb) p matrix_b->shape[1]                         
$64 = {{_mp_alloc = 0, _mp_size = 0, _mp_d = 0x0}}
(gdb) p matrix_b->shape[0]
$65 = {{_mp_alloc = 0, _mp_size = 0, _mp_d = 0x0}}

This is a result from the mpz_set in gfc_simplify_transpose.
Isn't this a representation of zero?

Alternative patch, which passes the relevant regtesting:

diff --git a/gcc/fortran/simplify.c b/gcc/fortran/simplify.c
index 388aca7c38c..b884a81fce0 100644
--- a/gcc/fortran/simplify.c
+++ b/gcc/fortran/simplify.c
@@ -8123,16 +8123,16 @@ gfc_simplify_transpose (gfc_expr *matrix)
                               &matrix->where);
   result->rank = 2;
   result->shape = gfc_get_shape (result->rank);
-  mpz_set (result->shape[0], matrix->shape[1]);
-  mpz_set (result->shape[1], matrix->shape[0]);
+  matrix_rows = mpz_get_si (matrix->shape[0]);
+  matrix_cols = mpz_get_si (matrix->shape[1]);
+  mpz_init_set_si (result->shape[0], matrix_cols);
+  mpz_init_set_si (result->shape[1], matrix_rows);

   if (matrix->ts.type == BT_CHARACTER)
     result->ts.u.cl = matrix->ts.u.cl;
   else if (matrix->ts.type == BT_DERIVED)
     result->ts.u.derived = matrix->ts.u.derived;

-  matrix_rows = mpz_get_si (matrix->shape[0]);
-  matrix_cols = mpz_get_si (matrix->shape[1]);
   for (row = 0; row < matrix_rows; ++row)
     for (col = 0; col < matrix_cols; ++col)
       {

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

--- Comment #5 from anlauf at gcc dot gnu.org ---
OK, now I see it.  gfc_get_shape does not init the resulting shape.
The following simpler patch does the job:

diff --git a/gcc/fortran/simplify.c b/gcc/fortran/simplify.c
index 388aca7c38c..c27b47aa98f 100644
--- a/gcc/fortran/simplify.c
+++ b/gcc/fortran/simplify.c
@@ -8123,8 +8123,8 @@ gfc_simplify_transpose (gfc_expr *matrix)
                               &matrix->where);
   result->rank = 2;
   result->shape = gfc_get_shape (result->rank);
-  mpz_set (result->shape[0], matrix->shape[1]);
-  mpz_set (result->shape[1], matrix->shape[0]);
+  mpz_init_set (result->shape[0], matrix->shape[1]);
+  mpz_init_set (result->shape[1], matrix->shape[0]);

   if (matrix->ts.type == BT_CHARACTER)
     result->ts.u.cl = matrix->ts.u.cl;

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[sgk at troutmask dot apl.washington.edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

--- Comment #6 from Steve Kargl <sgk at troutmask dot apl.washington.edu> ---
On Wed, Mar 31, 2021 at 08:51:57PM +0000, anlauf at gcc dot gnu.org wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840
> 
> --- Comment #5 from anlauf at gcc dot gnu.org ---
> OK, now I see it.  gfc_get_shape does not init the resulting shape.
> The following simpler patch does the job:
> 
> diff --git a/gcc/fortran/simplify.c b/gcc/fortran/simplify.c
> index 388aca7c38c..c27b47aa98f 100644
> --- a/gcc/fortran/simplify.c
> +++ b/gcc/fortran/simplify.c
> @@ -8123,8 +8123,8 @@ gfc_simplify_transpose (gfc_expr *matrix)
>                                &matrix->where);
>    result->rank = 2;
>    result->shape = gfc_get_shape (result->rank);
> -  mpz_set (result->shape[0], matrix->shape[1]);
> -  mpz_set (result->shape[1], matrix->shape[0]);
> +  mpz_init_set (result->shape[0], matrix->shape[1]);
> +  mpz_init_set (result->shape[1], matrix->shape[0]);
> 
>    if (matrix->ts.type == BT_CHARACTER)
>      result->ts.u.cl = matrix->ts.u.cl;
> 

The simple patch in comment #2 also works.

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[anlauf at gmx dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

--- Comment #7 from Harald Anlauf <anlauf at gmx dot de> ---
> The simple patch in comment #2 also works.

I know.  But it only covers the issue in gfc_simplify_transpose.

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

--- Comment #8 from anlauf at gcc dot gnu.org ---
Patch: https://gcc.gnu.org/pipermail/fortran/2021-March/055897.html

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Harald Anlauf <anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:d7cef070bf43bfb3f3d77bac42eadea06c4b0281

commit r11-7943-gd7cef070bf43bfb3f3d77bac42eadea06c4b0281
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Apr 1 07:49:32 2021 +0200

    PR fortran/99840 - ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

    The simplification of the transposition of a constant array shall properly
    initialize and set the shape of the result.

    gcc/fortran/ChangeLog:

            PR fortran/99840
            * simplify.c (gfc_simplify_transpose): Properly initialize
            resulting shape.

    gcc/testsuite/ChangeLog:

            PR fortran/99840
            * gfortran.dg/transpose_5.f90: New test.

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

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

https://gcc.gnu.org/g:591b5c903f8d3fc73caa403038cbd680b367f23b

commit r10-9648-g591b5c903f8d3fc73caa403038cbd680b367f23b
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Apr 1 07:49:32 2021 +0200

    PR fortran/99840 - ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

    The simplification of the transposition of a constant array shall properly
    initialize and set the shape of the result.

    gcc/fortran/ChangeLog:

            PR fortran/99840
            * simplify.c (gfc_simplify_transpose): Properly initialize
            resulting shape.

    gcc/testsuite/ChangeLog:

            PR fortran/99840
            * gfortran.dg/transpose_5.f90: New test.

    (cherry picked from commit d7cef070bf43bfb3f3d77bac42eadea06c4b0281)

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

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

https://gcc.gnu.org/g:0c7d11dbd216df4c1ae2f47138321b399e7349f0

commit r9-9319-g0c7d11dbd216df4c1ae2f47138321b399e7349f0
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Apr 1 07:49:32 2021 +0200

    PR fortran/99840 - ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

    The simplification of the transposition of a constant array shall properly
    initialize and set the shape of the result.

    gcc/fortran/ChangeLog:

            PR fortran/99840
            * simplify.c (gfc_simplify_transpose): Properly initialize
            resulting shape.

    gcc/testsuite/ChangeLog:

            PR fortran/99840
            * gfortran.dg/transpose_5.f90: New test.

    (cherry picked from commit d7cef070bf43bfb3f3d77bac42eadea06c4b0281)

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

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

https://gcc.gnu.org/g:28ba017a7ef5d0d0ebe4a37e52db22f667c1c0e2

commit r8-10828-g28ba017a7ef5d0d0ebe4a37e52db22f667c1c0e2
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Apr 1 07:49:32 2021 +0200

    PR fortran/99840 - ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

    The simplification of the transposition of a constant array shall properly
    initialize and set the shape of the result.

    gcc/fortran/ChangeLog:

            PR fortran/99840
            * simplify.c (gfc_simplify_transpose): Properly initialize
            resulting shape.

    gcc/testsuite/ChangeLog:

            PR fortran/99840
            * gfortran.dg/transpose_5.f90: New test.

    (cherry picked from commit d7cef070bf43bfb3f3d77bac42eadea06c4b0281)

[Bug fortran/99840] [9/10/11 Regression] ICE in gfc_simplify_matmul, at fortran/simplify.c:4777

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99840

anlauf at gcc dot gnu.org changed:

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

--- Comment #13 from anlauf at gcc dot gnu.org ---
Fixed on mainline and all open branches.  Closing.

Thanks for the report!