[Bug target/52079] ICE in aggregate_value_p while building libgo on arm-linux-gnueabihf

["jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52079

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2012-02-01
         AssignedTo|unassigned at gcc dot       |jakub at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-02-01 13:43:03 UTC ---
Created attachment 26542
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26542
gcc47-pr52079.patch

Untested fix.  While the c-family/fortran/lto/ada
lang_hooks.types.type_for_mode returns non-NULL for TImode (it creates a type
for 2 * HOST_BITS_PER_WIDE_INT bitsize), go (and assuming java too) just return
NULL.
This is an attempt to try harder in this case, another alternative would be for
go/java type_for_mode and type_for_size langhooks to return non-NULL for TImode
resp. 128-bit size (using build_nonstandard_integer_type), at least when
HOST_BITS_PER_WIDE_INT >= 64 (like other FEs do).