http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
Jakub Jelinek changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #18 from Jakub Jelinek 2012-01-26
14:09:34 UTC ---
Author: jakub
Date: Thu Jan 26 14:09:29 2012
New Revision: 183560
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=183560
Log:
PR middle-end/51895
* expr.c (expand_expr_
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #17 from Richard Guenther 2012-01-23
11:59:57 UTC ---
Author: rguenth
Date: Mon Jan 23 11:59:53 2012
New Revision: 183429
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=183429
Log:
2012-01-23 Richard Guenther
PR tree-o
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #16 from Jakub Jelinek 2012-01-20
16:49:52 UTC ---
Created attachment 26396
--> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26396
l.ii
Note that the BLKmode MEM_REF with non-BLKmode base expansion can be triggered
also without SR
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #15 from Richard Guenther 2012-01-20
14:40:09 UTC ---
Testing
@@ -3891,6 +3853,8 @@ decide_one_param_reduction (struct acces
{
by_ref = false;
agg_size = cur_parm_size;
+ if (DECL_MODE (parm) != BLKmode)
+
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
Richard Guenther changed:
What|Removed |Added
Target|powerpc64-linux |powerpc64-linux, i?86-*-*
--- Comment
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
Richard Guenther changed:
What|Removed |Added
CC||rguenth at gcc dot gnu.org
--- Comment
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #12 from Eric Botcazou 2012-01-20
09:41:41 UTC ---
> Sure, but I believe just that eipa_sra is what is causing this pessimization
> by
> doing a bad decision.
Precisely, so why not change that?
> Anyway, if you prefer one of the ot
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #11 from Jakub Jelinek 2012-01-20
09:07:54 UTC ---
(In reply to comment #10)
> > I wonder why it does this, instead of just using type S, and if it really
> > has
> > to for some reason, why it can't at least make sure it has the sam
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #10 from Eric Botcazou 2012-01-19
16:13:55 UTC ---
> I wonder why it does this, instead of just using type S, and if it really has
> to for some reason, why it can't at least make sure it has the same TYPE_MODE.
> Changing a TImode ar
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #9 from Jakub Jelinek 2012-01-19
13:55:31 UTC ---
Peter's patch does as well. The thing is that we still do wrong expand_expr on
that kind of MEM_EXPR, and I don't see anything that would prevent such
MEM_EXPRs in all kinds of other
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #8 from Richard Guenther 2012-01-19
13:17:25 UTC ---
For example (untested, and probably completely bogus):
Index: expr.c
===
--- expr.c (revision 183296)
+++ expr.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
Jakub Jelinek changed:
What|Removed |Added
CC||ebotcazou at gcc dot
|
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
Richard Guenther changed:
What|Removed |Added
Priority|P3 |P1
--- Comment #6 from Richard Guenthe
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #5 from Jakub Jelinek 2012-01-19
12:33:55 UTC ---
Created attachment 26377
--> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26377
gcc47-pr51895.patch
Untested patch that attempts to fix BLKmode MEM_REF expansion with
non-DECL_ADDR
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #4 from Jakub Jelinek 2012-01-19
10:16:07 UTC ---
(In reply to comment #3)
> > bftype = TREE_TYPE (base);
> > if (TYPE_MODE (TREE_TYPE (exp)) != BLKmode)
> > bftype = TREE_TYPE (exp);
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
Richard Guenther changed:
What|Removed |Added
Status|UNCONFIRMED |NEW
Last reconfirmed|
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
Jakub Jelinek changed:
What|Removed |Added
CC||jamborm at gcc dot gnu.org
--- Comment #2
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51895
--- Comment #1 from Peter Bergner 2012-01-19
02:42:44 UTC ---
We ICE when we try and do a BLK move from a reg:TI. In GCC 4.6, we don't see
this, since the args[i].tree_value is different:
(gdb-gcc4.6) ptree args[i].tree_value
unit siz
19 matches
Mail list logo
