This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
x86-64 bootstrap broken (was Re: [PATCH] tree level if-conversion for vectorizer)
- From: Jan Hubicka <hubicka at ucw dot cz>
- To: Devang Patel <dpatel at apple dot com>
- Cc: Richard Henderson <rth at redhat dot com>,Diego Novillo <dnovillo at redhat dot com>,GCC Patches <gcc-patches at gcc dot gnu dot org>,Dorit Naishlos <DORIT at il dot ibm dot com>, Ayal Zaks <ZAKS at il dot ibm dot com>,gcc-bugs at gcc dot gnu dot org, stevenb at suse dot de
- Date: Sat, 4 Sep 2004 14:55:11 +0200
- Subject: x86-64 bootstrap broken (was Re: [PATCH] tree level if-conversion for vectorizer)
- References: <20040826085655.GB3418@redhat.com> <8A2B21DE-F785-11D8-8706-000393A91CAA@apple.com> <20040826175940.GD5543@redhat.com> <704A4F05-F893-11D8-A1AD-000393A91CAA@apple.com> <20040830225141.GC3700@redhat.com> <80448B92-FB00-11D8-A1E0-000393A91CAA@apple.com> <20040831044334.GA4561@redhat.com> <3C950169-FB10-11D8-A1E0-000393A91CAA@apple.com> <20040831063608.GA5008@redhat.com> <473D0A88-FC52-11D8-9C9F-000393A91CAA@apple.com>
Devang,
it looks like your patch is responsible for misscompilation of stage2 at
x86-64 compiler (so we die building libgcc with memory corruption). I
checked it by disabling the pass from tree-optimize and we got past it.
Would be possible to look into it soon? I've got similarly looking
failure on PPC-linux too:
/abuild/gcc-test/gcc/gcc/libgcc2.c:359: internal compiler error: in
pool_free, at alloc-pool.c:287
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
/abuild/gcc-test/gcc/gcc/libgcc2.c: In function `__ashldi3':
/abuild/gcc-test/gcc/gcc/libgcc2.c:387: internal compiler error:
Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
/abuild/gcc-test/gcc/gcc/libgcc2.c: In function `__ashrdi3':
/abuild/gcc-test/gcc/gcc/libgcc2.c:416: internal compiler error: in
pool_free, at alloc-pool.c:287
Please submit a full bug report,
with preprocessed source if appropriate.
Honza
> I have addressed following feedback in this iteration:
>
> o Remove tree-gimple.c change
> o Update comments regarding vectorizer needs
> o Create one temp., for condition, per block
> o Allow only var decls as LHS for candidate assignments and expose
> assignment. This may transform 'uninitialized use' into
> 'un-intentionally initialized use', but that is fine.
> o Remove expr.c change from previous patch to allow (void)0 as one of
> the COND_EXPR operand.
>
> Bootstrapped on powerpc-darwin. DejaGNU test run is in progress. OK if
> it succeeds ?
>
> Thanks,
> -
> Devang
>
> 2004-09-01 Devang Patel <dpatel@apple.com>
>
> * Makefile.in (OBJS-common): Add tree-if-conv.o
> (tree-if-conv.o): New rule.
> * cfgloop.c (flow_loop_exit_edges_find): Set EDGE_LOOP_EXIT
> flag.
> (get_loop_body_in_bfs_order): New.
> * cfgloop.h (get_loop_body_in_bfs_order): New.
> * tree-flow.h (enum move_pos): Move here from ..
> * tree-ssa-loop-im.c (enum move_pos): here.
> (movement_possibility): Make externally visible.
> * tree-optimize.c (init_tree_optimization_passes): New entry for
> if conversion pass.
> * tree-pass.h (pass_if_conversion): New.
> * tree-ssa-operands.c (get_expr_operands): Handle COND_EXPR.
> * tree-if-conv.c: New file.
> * doc/passes.texi: Document tree if-conversion pass.
> * doc/tree-ssa.texi: Same.
>
> testsuite:
>
> * gcc.dg/tree-ssa/ifc-20040816-1.c: New test.
> * gcc.dg/tree-ssa/ifc-20040816-2.c: New test.
>
>
>