[Bug middle-end/46388] New: [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[Bug middle-end/46388] New: [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46388

           Summary: [4.6 Regression] ICE in int_mode_for_mode, at
                    stor-layout.c:493
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: rguenth@gcc.gnu.org
            Target: ia64-linux


Created attachment 22342
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22342
preprocessed source

openmpi fails to build on ia64-linux.

./cc1 -fpreprocessed file.i -quiet -dumpbase file.c -auxbase-strip
file/.libs/file.o -O -version -fmessage-length=0 -o file.s
In file included from ../opal/class/opal_list.h:67:0,
                 from ../ompi/file/file.h:23,
                 from file/file.c:21:
../opal/class/opal_object.h: In function 'ompi_file_init':
../opal/class/opal_object.h:456:33: internal compiler error: in
int_mode_for_mode, at stor-layout.c:493
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://bugs.opensuse.org/> for instructions.

[Bug middle-end/46388] [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46388

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.0

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-09 11:47:52 UTC ---
Reduced testcase:

typedef struct opal_object_t opal_object_t;
typedef struct opal_class_t opal_class_t;
struct opal_object_t {
    opal_class_t *obj_class;
};
typedef struct ompi_pointer_array_t ompi_pointer_array_t;
extern opal_class_t ompi_pointer_array_t_class;
extern ompi_pointer_array_t ompi_file_f_to_c_table;
void
ompi_file_init(void)
{
  ((opal_object_t *) ((&ompi_file_f_to_c_table)))->obj_class
      = ((&(ompi_pointer_array_t_class)));
}


#2  0x00000000006aaeb3 in store_bit_field_1 (str_rtx=0x7ffff7fb8ca8, 
    bitsize=64, bitnum=0, fieldmode=DImode, value=0x7ffff7fbb840, 
    fallback_p=1 '\001') at /space/rguenther/src/svn/trunk/gcc/expmed.c:480
480         enum machine_mode imode = int_mode_for_mode (GET_MODE (op0));
(gdb) p op0
$1 = (rtx) 0x7ffff7fb8ca8
(gdb) call debug_rtx (op0)
(mem/s/f/j/c (reg/f:DI 340) [0 MEM[(struct opal_object_t
*)&ompi_file_f_to_c_table].obj_class+0 S8 A8])


we have a VOIDmode MEM.

[Bug middle-end/46388] [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46388

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2010.11.10 10:00:41
                 CC|                            |jakub at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-10 10:00:41 UTC ---
struct S;
struct T
{
  struct S *t;
};
extern struct S s, u;

void
foo (void)
{
  ((struct T *) &u)->t = &s;
}

[Bug middle-end/46388] [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46388

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

[Bug middle-end/46388] [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46388

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

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

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-10 10:52:53 UTC ---
Created attachment 22364
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=22364
gcc46-pr46388.patch

Untested fix.

I think it is a bad idea to keep around VOIDmode MEMs, it can break too much.
I think this should work (alternatively we could use BLKmode in that case
instead).

[Bug middle-end/46388] [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46388

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-11 15:48:43 UTC ---
Author: jakub
Date: Thu Nov 11 15:48:39 2010
New Revision: 166603

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=166603
Log:
    PR middle-end/46388
    * expr.c (expand_assignment): If to_rtx is a VOIDmode MEM, use
    BLKmode mode for it.
    (expand_expr_real_1): Similarly for op0.

    * gcc.c-torture/compile/pr46388.c: New test.

Added:
    trunk/gcc/testsuite/gcc.c-torture/compile/pr46388.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/expr.c
    trunk/gcc/testsuite/ChangeLog

[Bug middle-end/46388] [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46388

--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-11 17:14:47 UTC ---
Fixed.

[Bug middle-end/46388] [4.6 Regression] ICE in int_mode_for_mode, at stor-layout.c:493

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46388

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

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

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> 2010-11-11 23:09:02 UTC ---
.