http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
--- Comment #12 from Andrey Belevantsev abel at gcc dot gnu.org 2011-04-07
06:53:48 UTC ---
Author: abel
Date: Thu Apr 7 06:53:44 2011
New Revision: 172080
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=172080
Log:
Backport from
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
--- Comment #10 from Alexander Monakov amonakov at gcc dot gnu.org 2010-11-22
10:35:10 UTC ---
Author: amonakov
Date: Mon Nov 22 10:35:06 2010
New Revision: 167025
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=167025
Log:
PR
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
Alexander Monakov amonakov at gcc dot gnu.org changed:
What|Removed |Added
Status|ASSIGNED|RESOLVED
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
Eric Botcazou ebotcazou at gcc dot gnu.org changed:
What|Removed |Added
CC||ebotcazou at
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
--- Comment #6 from Richard Guenther rguenth at gcc dot gnu.org 2010-11-17
14:08:01 UTC ---
(In reply to comment #5)
P1 sounds a little overrated, it's a specialized option and an artificial
test.
Only IA-64 really uses this thing, right?
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
--- Comment #7 from Richard Guenther rguenth at gcc dot gnu.org 2010-11-17
14:08:35 UTC ---
Btw, the scheduler guys have a fix in the works already.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
--- Comment #8 from Eric Botcazou ebotcazou at gcc dot gnu.org 2010-11-17
15:31:36 UTC ---
Btw, the scheduler guys have a fix in the works already.
Then they should assign the PR to someone to avoid duplicating efforts.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
Alexander Monakov amonakov at gcc dot gnu.org changed:
What|Removed |Added
Status|NEW |ASSIGNED
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
--- Comment #3 from Richard Guenther rguenth at gcc dot gnu.org 2010-11-04
16:35:55 UTC ---
It looks like scheduling itself adds the instruction that references
(mem/s:DI (reg:DI 4 si) [0 MEM[(struct S * {ref-all})addr.0_2]+0 S8 A64])
making use
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
--- Comment #4 from Richard Guenther rguenth at gcc dot gnu.org 2010-11-04
16:48:49 UTC ---
And indeed --param selsched-insns-to-rename=0 fixes the testcase.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
Richard Guenther rguenth at gcc dot gnu.org changed:
What|Removed |Added
Priority|P3 |P1
--- Comment #2 from amonakov at gcc dot gnu dot org 2010-09-13 16:53
---
Confirmed. Not related to PR43949 since selective scheduling does not use
cselib. The miscompilation seems to come from RTL aliasing: sel-sched lifts a
load that references stack via a general-purpose register
--- Comment #1 from zsojka at seznam dot cz 2010-09-12 15:44 ---
Created an attachment (id=21781)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=21781action=view)
reduced testcase
$ /mnt/svn/gcc-trunk/binary-163636-lto-fortran-checking-yes-rtl-df/bin/gcc -O2
-fselective-scheduling2
--
rguenth at gcc dot gnu dot org changed:
What|Removed |Added
Target Milestone|--- |4.6.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=45652
14 matches
Mail list logo