------- Comment #2 from ro at gcc dot gnu dot org 2010-07-07 15:54 -------
A reghunt identified this patch as the culprit:
2010-07-05 Richard Guenther <rguent...@suse.de>
* tree.c (reference_alias_ptr_type): New function.
* tree.h (reference_alias_ptr_type): Declare.
* tree-ssa-loop-ivopts.c (copy_ref_info): Restructure to
allow non-TARGET_MEM_REF new refs.
(rewrite_use_address): Pass old alias pointer type to
create_mem_ref.
* tree-ssa-address.c (create_mem_ref_raw): Get alias pointer type.
Build a MEM_REF instead of a TARGET_MEM_REF if possible.
(create_mem_ref): Get alias pointer type. Adjust calls to
create_mem_ref_raw.
(maybe_fold_tmr): Likewise.
* tree-flow.h (create_mem_ref): Adjust prototype.
I'm attaching the preprocessed format.i (from libgfortran/io/format.c).
The patch changes format.s as follows:
--- format.s 2010-07-07 16:26:07.933989030 +0200
+++
/var/gcc/gcc-4.6.0-20100705/10-gcc/sparc-sun-solaris2.10/libgfortran/format.
s 2010-07-07 16:23:08.775373256 +0200
@@ -2267,7 +2267,7 @@ _gfortrani_free_format_hash_table:
ld [%i0+248], %g1
orcc %g1, 0, %o0
be,a,pn %icc, .LL256
- st %g0, [%i0+240]
+ stx %g0, [%i0+240]
.LLSM424:
call _gfortrani_free_format_data, 0
nop
I.e. a store to [%i0+240] is done with stx, although the destination isn't
8-byte aligned.
I'm compiling the .i file like this:
./cc1 -fpreprocessed format.i -mcpu=v9 -g -O2 -std=gnu99 -fcx-fortran-rules
-ffunction-sections -fdata-sections -fPIC -o format.s
$
--
ro at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rguenther at suse dot de
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44843
