* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
@ 2004-05-29 0:06 ` reichelt at gcc dot gnu dot org
2004-05-29 0:12 ` reichelt at gcc dot gnu dot org
` (9 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: reichelt at gcc dot gnu dot org @ 2004-05-29 0:06 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From reichelt at gcc dot gnu dot org 2004-05-28 16:07 -------
Confirmed.
BTW, I also had to set CXXFLAGS to the value mentioned above.
It's actually the program "pic" located in "groff-1.19.1/src/preproc/pic"
that crashes. I'll add a minimal input file for the miscompiled "pic" that
triggers the segfault.
--
What |Removed |Added
----------------------------------------------------------------------------
CC| |reichelt at gcc dot gnu dot
| |org
Component|c |target
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
2004-05-29 0:06 ` [Bug target/15617] " reichelt at gcc dot gnu dot org
@ 2004-05-29 0:12 ` reichelt at gcc dot gnu dot org
2004-06-02 1:47 ` reichelt at gcc dot gnu dot org
` (8 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: reichelt at gcc dot gnu dot org @ 2004-05-29 0:12 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From reichelt at gcc dot gnu dot org 2004-05-28 16:10 -------
Created an attachment (id=6418)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=6418&action=view)
trigger for bug in miscompiled "pic"
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
2004-05-29 0:06 ` [Bug target/15617] " reichelt at gcc dot gnu dot org
2004-05-29 0:12 ` reichelt at gcc dot gnu dot org
@ 2004-06-02 1:47 ` reichelt at gcc dot gnu dot org
2004-06-02 1:58 ` pinskia at gcc dot gnu dot org
` (7 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: reichelt at gcc dot gnu dot org @ 2004-06-02 1:47 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From reichelt at gcc dot gnu dot org 2004-06-02 01:47 -------
The culprit is in the function box_object::print() in object.cpp,
where the program crashes. Removing the following 4 lines from the
assembler code makes things work again (the lines are missing if
one compiles the code with -march=pentium3):
========================================================
--- object.s 2004-06-02 03:02:24.000000000 +0200
+++ object.s 2004-06-02 03:01:24.000000000 +0200
@@ -5217,11 +5217,9 @@ _ZN10box_object5printEv:
subl $8, %esp
leal -124(%ebp), %esi
fchs
- movsd .LC125, %xmm0
leal -108(%ebp), %ebx
fstpl (%esp)
pushl -24(%ebp)
- movaps %xmm0, -144(%ebp)
pushl -28(%ebp)
pushl %esi
.LCFI482:
@@ -5331,9 +5329,7 @@ _ZN10box_object5printEv:
subl $16, %esp
leal -124(%ebp), %ebx
fldl 76(%eax)
- movsd .LC126, %xmm0
fabs
- movaps %xmm0, -160(%ebp)
fstpl 8(%esp)
fldl 68(%eax)
fabs
========================================================
I'll attach a short self-contained testcase to reproduce the function
box_object::print(). Just compile with "g++ -Os -march=pentium4".
(The assembler labels are of course different, but there are only
four lines that contain %xmm0.)
To actually run the code one has to build groff, though.
So this testcase is not completely self-contained,
but maybe this is enough to identify the bug.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Ever Confirmed| |1
Last reconfirmed|0000-00-00 00:00:00 |2004-06-02 01:47:24
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
` (2 preceding siblings ...)
2004-06-02 1:47 ` reichelt at gcc dot gnu dot org
@ 2004-06-02 1:58 ` pinskia at gcc dot gnu dot org
2004-07-04 0:36 ` pinskia at gcc dot gnu dot org
` (6 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-06-02 1:58 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-06-02 01:58 -------
This looks like a case where the stack pointer is misaligned, movaps requires 16byte (128bit) aligned
addresses So this caused by -Os implying -mpreferred-stack-boundary=32 which is wrong for SSE.
Maybe the fix would be to imply -mpreferred-stack-boundary=128 if -msse (or is it -msse2) is turned
on with -Os, this also causes var_arg25.c (I think) to fail at -Os.
Jan I added you as CC to find your opinion on my suggestion. I think a huge amount of people will run
into this if they used -Os without knowning what is going on.
--
What |Removed |Added
----------------------------------------------------------------------------
CC| |hubicka at gcc dot gnu dot
| |org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
` (3 preceding siblings ...)
2004-06-02 1:58 ` pinskia at gcc dot gnu dot org
@ 2004-07-04 0:36 ` pinskia at gcc dot gnu dot org
2004-07-10 2:06 ` pinskia at gcc dot gnu dot org
` (5 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-07-04 0:36 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-07-04 00:36 -------
*** Bug 16354 has been marked as a duplicate of this bug. ***
--
What |Removed |Added
----------------------------------------------------------------------------
CC| |jbuck at gcc dot gnu dot org
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
` (4 preceding siblings ...)
2004-07-04 0:36 ` pinskia at gcc dot gnu dot org
@ 2004-07-10 2:06 ` pinskia at gcc dot gnu dot org
2004-07-10 2:07 ` pinskia at gcc dot gnu dot org
` (4 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-07-10 2:06 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-07-10 02:06 -------
Fixed by:
> * config/i386/i386.c (init_cumulative_args): Set mmx/sse registers
> available for use only when TARGET_MMX/_SSE is set, and check for
> variable argument function if any register count is non-zero.
> (function_arg): Correctly suppress repeated warnings for passing SSE
> vectors are arguments without SSE enabled.
> (ix86_function_arg_boundary): Without SSE enabled, the use of what
> would be SSE register modes does no longer influence the alignment.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
` (5 preceding siblings ...)
2004-07-10 2:06 ` pinskia at gcc dot gnu dot org
@ 2004-07-10 2:07 ` pinskia at gcc dot gnu dot org
2004-07-12 12:29 ` reichelt at gcc dot gnu dot org
` (3 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-07-10 2:07 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From pinskia at gcc dot gnu dot org 2004-07-10 02:07 -------
Fixed by:
> * config/i386/i386.c (init_cumulative_args): Set mmx/sse registers
> available for use only when TARGET_MMX/_SSE is set, and check for
> variable argument function if any register count is non-zero.
> (function_arg): Correctly suppress repeated warnings for passing SSE
> vectors are arguments without SSE enabled.
> (ix86_function_arg_boundary): Without SSE enabled, the use of what
> would be SSE register modes does no longer influence the alignment.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |RESOLVED
Resolution| |FIXED
Target Milestone|--- |3.5.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
` (6 preceding siblings ...)
2004-07-10 2:07 ` pinskia at gcc dot gnu dot org
@ 2004-07-12 12:29 ` reichelt at gcc dot gnu dot org
2004-07-17 19:29 ` pinskia at gcc dot gnu dot org
` (2 subsequent siblings)
10 siblings, 0 replies; 12+ messages in thread
From: reichelt at gcc dot gnu dot org @ 2004-07-12 12:29 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From reichelt at gcc dot gnu dot org 2004-07-12 12:29 -------
Not fixed.
We still crash at the same position.
--
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Resolution|FIXED |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
` (7 preceding siblings ...)
2004-07-12 12:29 ` reichelt at gcc dot gnu dot org
@ 2004-07-17 19:29 ` pinskia at gcc dot gnu dot org
2005-02-18 15:55 ` uros at kss-loka dot si
2005-08-15 0:46 ` mueller at kde dot org
10 siblings, 0 replies; 12+ messages in thread
From: pinskia at gcc dot gnu dot org @ 2004-07-17 19:29 UTC (permalink / raw)
To: gcc-bugs
--
What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|3.5.0 |---
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
` (8 preceding siblings ...)
2004-07-17 19:29 ` pinskia at gcc dot gnu dot org
@ 2005-02-18 15:55 ` uros at kss-loka dot si
2005-08-15 0:46 ` mueller at kde dot org
10 siblings, 0 replies; 12+ messages in thread
From: uros at kss-loka dot si @ 2005-02-18 15:55 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From uros at kss-loka dot si 2005-02-18 06:45 -------
FYI: gcc 4.0 doesn't generate any SSE instructions for testcase.cc:
gcc -Os -march=pentium4 -S testcase.cc
grep xmm testcase.s | wc -l
0
--
What |Removed |Added
----------------------------------------------------------------------------
Known to work| |4.0.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread* [Bug target/15617] building groff-1.19.1 with "-Os -march=pentium4" causes sig 11
2004-05-24 1:29 [Bug c/15617] New: " kpfleming at backtobasicsmgmt dot com
` (9 preceding siblings ...)
2005-02-18 15:55 ` uros at kss-loka dot si
@ 2005-08-15 0:46 ` mueller at kde dot org
10 siblings, 0 replies; 12+ messages in thread
From: mueller at kde dot org @ 2005-08-15 0:46 UTC (permalink / raw)
To: gcc-bugs
------- Additional Comments From mueller at kde dot org 2005-08-15 00:46 -------
duplicate of #13685 ?
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15617
^ permalink raw reply [flat|nested] 12+ messages in thread