public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <bug-2480-1008@http.gcc.gnu.org/bugzilla/>
@ 2008-03-14 20:07 ` rguenth at gcc dot gnu dot org
2008-04-07 1:33 ` pinskia at gcc dot gnu dot org
` (3 subsequent siblings)
4 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2008-03-14 20:07 UTC (permalink / raw)
To: gcc-bugs
------- Comment #10 from rguenth at gcc dot gnu dot org 2008-03-14 20:06 -------
This is related to PR27799. It is also fixed with -fstrict-aliasing on the
tree level:
bar ()
{
struct example * ex1.0;
<bb 2>:
ex1.0 = ex1;
ex1.0->a = 1;
ex1.0->b = 2;
ex1.0->c = 3;
return;
}
Without -fstrict-aliasing we get (see PR27799):
bar ()
{
struct example * ex1.0;
<bb 2>:
# VUSE <ex1_4(D)>
ex1.0_1 = ex1;
# ex1_6 = VDEF <ex1_4(D)>
# SMT.5_7 = VDEF <SMT.5_5(D)>
ex1.0_1->a = 1;
# VUSE <ex1_6>
ex1.0_2 = ex1;
# ex1_8 = VDEF <ex1_6>
# SMT.5_9 = VDEF <SMT.5_7>
ex1.0_2->b = 2;
# VUSE <ex1_8>
ex1.0_3 = ex1;
# ex1_10 = VDEF <ex1_8>
# SMT.5_11 = VDEF <SMT.5_9>
ex1.0_3->c = 3;
return;
while with -fstrict-aliasing we have
bar ()
{
struct example * ex1.0;
<bb 2>:
# VUSE <ex1_4(D)>
ex1.0_1 = ex1;
# SMT.5_6 = VDEF <SMT.5_5(D)>
ex1.0_1->a = 1;
# VUSE <ex1_4(D)>
ex1.0_2 = ex1;
# SMT.5_7 = VDEF <SMT.5_6>
ex1.0_2->b = 2;
# VUSE <ex1_4(D)>
ex1.0_3 = ex1;
# SMT.5_8 = VDEF <SMT.5_7>
ex1.0_3->c = 3;
return;
}
If you disable all tree optimizations that do the optimization you still
have the missed optimization on the RTL level and get
bar:
pushl %ebp
movl %esp, %ebp
movl ex1, %eax
movb $1, (%eax)
movl ex1, %eax
movl $2, 4(%eax)
movb $3, 8(%eax)
popl %ebp
ret
without strict-aliasing it gets even
bar:
pushl %ebp
movl %esp, %ebp
movl ex1, %eax
movb $1, (%eax)
movl ex1, %eax
movl $2, 4(%eax)
movl ex1, %eax
movb $3, 8(%eax)
popl %ebp
ret
--
rguenth at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rguenth at gcc dot gnu dot
| |org
OtherBugsDependingO| |27799
nThis| |
Last reconfirmed|2006-02-16 21:27:01 |2008-03-14 20:06:36
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <bug-2480-1008@http.gcc.gnu.org/bugzilla/>
2008-03-14 20:07 ` [Bug tree-optimization/2480] aliasing problem with global structures rguenth at gcc dot gnu dot org
@ 2008-04-07 1:33 ` pinskia at gcc dot gnu dot org
2008-04-07 9:01 ` rguenth at gcc dot gnu dot org
` (2 subsequent siblings)
4 siblings, 0 replies; 11+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2008-04-07 1:33 UTC (permalink / raw)
To: gcc-bugs
------- Comment #11 from pinskia at gcc dot gnu dot org 2008-04-07 01:32 -------
Hmm, we do get something different on the trunk:
bar ()
{
struct example * ex1.0;
<bb 2>:
ex1.0 = ex1;
ex1.0->a = 1;
ex1.0->b = 2;
ex1.0->c = 3;
if (ex1.0->b != 2)
goto <bb 3>;
else
goto <bb 5>;
<bb 3>:
link_error ();
if (ex1->c != 3)
goto <bb 4>;
else
goto <bb 5>;
<bb 4>:
Invalid sum of incoming frequencies 5123, should be 6216
link_error () [tail call];
<bb 5>:
Invalid sum of incoming frequencies 11093, should be 10000
return;
}
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <bug-2480-1008@http.gcc.gnu.org/bugzilla/>
2008-03-14 20:07 ` [Bug tree-optimization/2480] aliasing problem with global structures rguenth at gcc dot gnu dot org
2008-04-07 1:33 ` pinskia at gcc dot gnu dot org
@ 2008-04-07 9:01 ` rguenth at gcc dot gnu dot org
2009-04-03 11:27 ` rguenth at gcc dot gnu dot org
2009-04-03 12:38 ` rguenth at gcc dot gnu dot org
4 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2008-04-07 9:01 UTC (permalink / raw)
To: gcc-bugs
------- Comment #12 from rguenth at gcc dot gnu dot org 2008-04-07 09:00 -------
This is because the oracle is confused by the link_error () call which clobbers
*ex2 and *ex1. Note that SCCVN does not consider control-dependence as it
considers all edges executable all the time.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <bug-2480-1008@http.gcc.gnu.org/bugzilla/>
` (2 preceding siblings ...)
2008-04-07 9:01 ` rguenth at gcc dot gnu dot org
@ 2009-04-03 11:27 ` rguenth at gcc dot gnu dot org
2009-04-03 12:38 ` rguenth at gcc dot gnu dot org
4 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2009-04-03 11:27 UTC (permalink / raw)
To: gcc-bugs
------- Comment #13 from rguenth at gcc dot gnu dot org 2009-04-03 11:27 -------
Fixed at -O2 since 4.4.
--
rguenth at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
Target Milestone|--- |4.4.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <bug-2480-1008@http.gcc.gnu.org/bugzilla/>
` (3 preceding siblings ...)
2009-04-03 11:27 ` rguenth at gcc dot gnu dot org
@ 2009-04-03 12:38 ` rguenth at gcc dot gnu dot org
4 siblings, 0 replies; 11+ messages in thread
From: rguenth at gcc dot gnu dot org @ 2009-04-03 12:38 UTC (permalink / raw)
To: gcc-bugs
------- Comment #14 from rguenth at gcc dot gnu dot org 2009-04-03 12:38 -------
Subject: Bug 2480
Author: rguenth
Date: Fri Apr 3 12:38:08 2009
New Revision: 145499
URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=145499
Log:
2009-04-03 Richard Guenther <rguenther@suse.de>
PR tree-optimization/2480
PR tree-optimization/23086
* gcc.dg/tree-ssa/pr2480.c: New testcase.
* gcc.dg/tree-ssa/pr23086.c: Likewise.
Added:
trunk/gcc/testsuite/gcc.dg/tree-ssa/pr23086.c
trunk/gcc/testsuite/gcc.dg/tree-ssa/pr2480.c
Modified:
trunk/gcc/testsuite/ChangeLog
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <20010403165559.2480.dann@godzilla.ics.uci.edu>
` (4 preceding siblings ...)
2005-01-31 6:57 ` pinskia at gcc dot gnu dot org
@ 2005-06-09 16:45 ` pinskia at gcc dot gnu dot org
5 siblings, 0 replies; 11+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2005-06-09 16:45 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2005-06-09 16:45 -------
Part of this has been fixed, there is only one loading of ex1 now on the mainline.
--
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed|2005-05-02 00:52:06 |2005-06-09 16:45:02
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <20010403165559.2480.dann@godzilla.ics.uci.edu>
` (3 preceding siblings ...)
2005-01-23 15:42 ` pinskia at gcc dot gnu dot org
@ 2005-01-31 6:57 ` pinskia at gcc dot gnu dot org
2005-06-09 16:45 ` pinskia at gcc dot gnu dot org
5 siblings, 0 replies; 11+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2005-01-31 6:57 UTC (permalink / raw)
To: gcc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|4.0.0 |---
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <20010403165559.2480.dann@godzilla.ics.uci.edu>
` (2 preceding siblings ...)
2005-01-23 15:23 ` steven at gcc dot gnu dot org
@ 2005-01-23 15:42 ` pinskia at gcc dot gnu dot org
2005-01-31 6:57 ` pinskia at gcc dot gnu dot org
2005-06-09 16:45 ` pinskia at gcc dot gnu dot org
5 siblings, 0 replies; 11+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2005-01-23 15:42 UTC (permalink / raw)
To: gcc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
Status|REOPENED |NEW
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <20010403165559.2480.dann@godzilla.ics.uci.edu>
2005-01-23 14:55 ` steven at gcc dot gnu dot org
2005-01-23 15:15 ` pinskia at gcc dot gnu dot org
@ 2005-01-23 15:23 ` steven at gcc dot gnu dot org
2005-01-23 15:42 ` pinskia at gcc dot gnu dot org
` (2 subsequent siblings)
5 siblings, 0 replies; 11+ messages in thread
From: steven at gcc dot gnu dot org @ 2005-01-23 15:23 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From steven at gcc dot gnu dot org 2005-01-23 15:23 -------
Whoops. I forgot to link.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <20010403165559.2480.dann@godzilla.ics.uci.edu>
2005-01-23 14:55 ` steven at gcc dot gnu dot org
@ 2005-01-23 15:15 ` pinskia at gcc dot gnu dot org
2005-01-23 15:23 ` steven at gcc dot gnu dot org
` (3 subsequent siblings)
5 siblings, 0 replies; 11+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2005-01-23 15:15 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2005-01-23 15:15 -------
This is not fixed, see example #2.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|FIXED |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
* [Bug tree-optimization/2480] aliasing problem with global structures
[not found] <20010403165559.2480.dann@godzilla.ics.uci.edu>
@ 2005-01-23 14:55 ` steven at gcc dot gnu dot org
2005-01-23 15:15 ` pinskia at gcc dot gnu dot org
` (4 subsequent siblings)
5 siblings, 0 replies; 11+ messages in thread
From: steven at gcc dot gnu dot org @ 2005-01-23 14:55 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From steven at gcc dot gnu dot org 2005-01-23 14:54 -------
CVS today:
.text
.p2align 4,,15
.globl bar
.type bar, @function
bar:
movl ex1, %eax
pushl %ebp
movl %esp, %ebp
movb $1, (%eax)
movl $2, 4(%eax)
movb $3, 8(%eax)
leave
ret
.size bar, .-bar
.p2align 4,,15
.globl foo
.type foo, @function
foo:
pushl %ebp
movl %esp, %ebp
movl 8(%ebp), %eax
movb $1, (%eax)
movl $2, 4(%eax)
movb $3, 8(%eax)
leave
ret
.size foo, .-foo
--
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
Target Milestone|--- |4.0.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=2480
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~2009-04-03 12:38 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
[not found] <bug-2480-1008@http.gcc.gnu.org/bugzilla/>
2008-03-14 20:07 ` [Bug tree-optimization/2480] aliasing problem with global structures rguenth at gcc dot gnu dot org
2008-04-07 1:33 ` pinskia at gcc dot gnu dot org
2008-04-07 9:01 ` rguenth at gcc dot gnu dot org
2009-04-03 11:27 ` rguenth at gcc dot gnu dot org
2009-04-03 12:38 ` rguenth at gcc dot gnu dot org
[not found] <20010403165559.2480.dann@godzilla.ics.uci.edu>
2005-01-23 14:55 ` steven at gcc dot gnu dot org
2005-01-23 15:15 ` pinskia at gcc dot gnu dot org
2005-01-23 15:23 ` steven at gcc dot gnu dot org
2005-01-23 15:42 ` pinskia at gcc dot gnu dot org
2005-01-31 6:57 ` pinskia at gcc dot gnu dot org
2005-06-09 16:45 ` pinskia at gcc dot gnu dot org
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).