--- Comment #4 from eric dot weddington at atmel dot com 2008-04-09 22:09
---
Confirmed.
Andy's fwprop patch from bug #35542 did not solve this.
--
eric dot weddington at atmel dot com changed:
What|Removed |Added
gcc can place code into the .rodata section. I found this while debugging a
crash in Ingo Molnar's sched-devel kernel. In the kernel code
arch/x86/kernel/alternative.c the static inline function costant_test_bit was
placed into the rodata. Later on in the kernel boot up sequence, the rodata
--- Comment #1 from rostedt at goodmis dot org 2008-04-09 22:20 ---
Created an attachment (id=15461)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15461action=view)
alternative.c file compiled with -E and with some stipping
This is a gzipped version that shows the problem.
Do the
--- Comment #2 from pinskia at gcc dot gnu dot org 2008-04-09 22:34 ---
The toplevel asm has:
asm(\t.section .rodata, \a\\nintelnops:
.byte 0x90\n .byte 0x89,0xf6\n .byte 0x8d,0x76,0x00\n .byte
0x8d,0x74,0x26,0x00\n .byte 0x90\n .byte 0x8d,0x74,0x26,0x00\n .byte
Four tests in libgomp.fortran (omp_parse3.f90, pr25162.f, and
threadprivate[23].f90 started failing with this patch:
http://gcc.gnu.org/viewcvs?view=revrev=133801
r133801 | george | 2008-04-01 17:23:36 -0400 (Tue, 01 Apr 2008)
This small testcase fails in the same way when compiled with -m64
--- Comment #1 from janis at gcc dot gnu dot org 2008-04-09 22:46 ---
Created an attachment (id=15462)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15462action=view)
generated code before patch
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35896
--- Comment #2 from janis at gcc dot gnu dot org 2008-04-09 22:47 ---
Created an attachment (id=15463)
-- (http://gcc.gnu.org/bugzilla/attachment.cgi?id=15463action=view)
generated code after patch
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35896
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
Keywords||openmp
Target Milestone|--- |4.4.0
--- Comment #3 from pinskia at gcc dot gnu dot org 2008-04-09 22:50 ---
I think we are emitting the debug info too early for _testcom before we marked
it as a TLS.
--
pinskia at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #6 from hutchinsonandy at gcc dot gnu dot org 2008-04-09 22:51
---
Subject: Bug 35519
Author: hutchinsonandy
Date: Wed Apr 9 22:50:42 2008
New Revision: 134152
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=134152
Log:
2008-04-09 Andy Hutchinson [EMAIL PROTECTED]
--- Comment #12 from hutchinsonandy at gcc dot gnu dot org 2008-04-09
22:51 ---
Subject: Bug 34916
Author: hutchinsonandy
Date: Wed Apr 9 22:50:42 2008
New Revision: 134152
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=134152
Log:
2008-04-09 Andy Hutchinson [EMAIL PROTECTED]
--- Comment #4 from pinskia at gcc dot gnu dot org 2008-04-09 22:51 ---
Before:
.quad [EMAIL PROTECTED]
After:
.8byte testcom_
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35896
--- Comment #5 from pinskia at gcc dot gnu dot org 2008-04-09 22:54 ---
Actually we just don't take into account TLSness of the decl/RTL.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35896
--- Comment #6 from jvdelisle at gcc dot gnu dot org 2008-04-09 23:47
---
After a clean bootstrap (empty build directory) I get a pass. I suspect we
have a dependency somewhere not getting taken car of. Hans-Peter can you try
this and see whta happends.
--
--- Comment #7 from hp at gcc dot gnu dot org 2008-04-09 23:49 ---
For cris-elf, this graduated to a runtime error equivalent to SEGV (no
valgrind); worked with 134139, failed from 134147.
--
hp at gcc dot gnu dot org changed:
What|Removed |Added
--- Comment #8 from hp at gcc dot gnu dot org 2008-04-09 23:56 ---
Try what? I always start with a clean slate; an empty build directory!
Perhaps you mean something else?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34955
--- Comment #9 from hp at gcc dot gnu dot org 2008-04-09 23:57 ---
(In reply to comment #8)
Try what? I always start with a clean slate; an empty build directory!
Perhaps you mean something else?
Oops, replied in wrong PR... please ignore.
--
--- Comment #7 from hp at gcc dot gnu dot org 2008-04-09 23:57 ---
Try what? I always start with a clean slate; an empty build directory!
Perhaps you mean something else?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35864
--- Comment #2 from hutchinsonandy at gcc dot gnu dot org 2008-04-09 23:59
---
Subject: Bug 33782
Author: hutchinsonandy
Date: Wed Apr 9 23:58:39 2008
New Revision: 134153
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=134153
Log:
2008-04-09 Andy Hutchinson [EMAIL PROTECTED]
--- Comment #2 from hutchinsonandy at gcc dot gnu dot org 2008-04-09 23:59
---
Subject: Bug 34894
Author: hutchinsonandy
Date: Wed Apr 9 23:58:39 2008
New Revision: 134153
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=134153
Log:
2008-04-09 Andy Hutchinson [EMAIL PROTECTED]
--- Comment #8 from hp at gcc dot gnu dot org 2008-04-10 00:00 ---
I haven't tried it myself, but since we're in voodoo-debug mode, try
--enable-checking=yes,valgrind.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35864
--- Comment #7 from hp at gcc dot gnu dot org 2008-04-10 00:03 ---
(In reply to comment #6)
I am discussing with Zdenek the proper fix.
Any conclusion in sight? If not, I'll xfail this as per protocol.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35629
--- Comment #3 from joseph at codesourcery dot com 2008-04-10 00:15 ---
Subject: Re: Vector load/store from a packed struct does
not work (without -mstrict-align)
On Tue, 8 Apr 2008, pinskia at gcc dot gnu dot org wrote:
With -mstrict-align, we get the correct (but suboptimal)
dse.c has
struct store_info
{
...
/* An bitmask as wide as the number of bytes in the word that
contains a 1 if the byte may be needed. The store is unused if
all of the bits are 0. */
long positions_needed;
};
...
record_store ()
{
...
gcc_assert ((unsigned) width sizeof
--- Comment #1 from pinskia at gcc dot gnu dot org 2008-04-10 00:41 ---
Well first it should be made HOST_WIDE_INT and x86 should move over to 64bit
HWI.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35897
--- Comment #2 from hjl at gcc dot gnu dot org 2008-04-10 00:53 ---
Subject: Bug 35897
Author: hjl
Date: Thu Apr 10 00:53:04 2008
New Revision: 134160
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=134160
Log:
2008-04-09 H.J. Lu [EMAIL PROTECTED]
PR middle-end/35897
Using built-in specs.
Target: x86_64-linux-gnu
Configured with: ../src/configure -v
--enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr
--enable-shared --with-system-zlib --libexecdir=/usr/lib
--without-included-gettext --enable-threads=posix --enable-nls
--- Comment #1 from vania at liama dot ia dot ac dot cn 2008-04-10 03:34
---
I have a simplified version of this program that works fine however !!!
When the exception thrown is an int instead of a class instance, everything
works fine!
--
vania at liama dot ia dot ac dot cn
--- Comment #9 from irar at gcc dot gnu dot org 2008-04-10 05:52 ---
Subject: Bug 35821
Author: irar
Date: Thu Apr 10 05:51:59 2008
New Revision: 134162
URL: http://gcc.gnu.org/viewcvs?root=gccview=revrev=134162
Log:
PR tree-optimization/35821
* tree-vect-transform.c
101 - 129 of 129 matches
Mail list logo