| View previous topic :: View next topic |
| Author |
Message |
Jim6
Tux's lil' helper
Joined: 08 Apr 2005
Posts: 102
|
 Posted: Fri Jul 24, 2009 12:11 am Post subject: [solved-ish] hal segfaults in libglib |
 |
|
hal segfaults every time it is started. The instruction pointer is always the same.
| Code: | hald[17893]: segfault at 8a38 ip b7e54e15 sp bfbb7ba8 error 4 in libglib-2.0.so.0.2000.4[b7dfa000+d0000]
hald[27080]: segfault at 8a38 ip b7e83e15 sp bfaf8288 error 4 in libglib-2.0.so.0.2000.4[b7e29000+d0000]
hald[19667]: segfault at 9c58 ip b7e58e15 sp bff27718 error 4 in libglib-2.0.so.0.2000.4[b7dfe000+d0000]
hald[19160]: segfault at 9c58 ip b7e93e15 sp bf80a868 error 4 in libglib-2.0.so.0.2000.4[b7e39000+d0000]
hald[17517]: segfault at 9c58 ip b7e3ee15 sp bf987ef8 error 4 in libglib-2.0.so.0.2000.4[b7de4000+d0000]
hald[17541]: segfault at 9c58 ip b7e80e15 sp bfcf0e98 error 4 in libglib-2.0.so.0.2000.4[b7e26000+d0000]
hald[21838]: segfault at 9c58 ip b7e2ce15 sp bfa724d8 error 4 in libglib-2.0.so.0.2000.4[b7dd2000+d0000]
hald[1314]: segfault at 9c58 ip b7ecfe15 sp bfe12078 error 4 in libglib-2.0.so.0.2000.4[b7e75000+d0000]
hald[1463]: segfault at 9c58 ip b7e09e15 sp bfaec2a8 error 4 in libglib-2.0.so.0.2000.4[b7daf000+d0000]
hald[1619]: segfault at 9c58 ip b7e70e15 sp bfe463d8 error 4 in libglib-2.0.so.0.2000.4[b7e16000+d0000] |
I have tried the following versions of hal and glib
hal:
0.5.9.1-r3
0.5.11-r8
0.5.12_rc1-r7
glib:
2.18.4-r1
2.20.4
I have run revdep-rebuild many many times at this point 
I believe that this thread may be related, but maybe not.
make.conf
dmesg
hald --daemon=no --verbose=yes &> hald.txt
strace hald &> hald-strace.txt
hald --daemon=no --verbose=yes &> hald-strace2.txt
Does anyone have any ideas?
Last edited by Jim6 on Fri Jul 24, 2009 6:01 pm; edited 1 time in total |
|
| Back to top |
|
 |
Jim6
Tux's lil' helper
Joined: 08 Apr 2005
Posts: 102
|
| Posted: Fri Jul 24, 2009 1:49 pm Post subject: |
|
|
I discovered that I was using gcc-4.1.2, so I'm going to try to switch to 4.3.2 and emerge -e world.
I have already tried re-building hal and glib using gcc-4.3.2, so this is a long shot
| Code: | sofabox ~ # gcc-config -l
[1] i686-pc-linux-gnu-4.1.2
[2] i686-pc-linux-gnu-4.3.2 *
|
|
|
| Back to top |
|
|
Jim6
Tux's lil' helper
Joined: 08 Apr 2005
Posts: 102
|
| Posted: Fri Jul 24, 2009 5:57 pm Post subject: |
|
|
Well, I have it working, but only with no optimisations:
For posterity:
| Code: | ~ # gcc-config -l
[1] i686-pc-linux-gnu-4.1.2
[2] i686-pc-linux-gnu-4.3.2
[3] i686-pc-linux-gnu-4.4.1 *
|
| Code: | | CFLAGS="-O0 -march=native -pipe -fomit-frame-pointer" |
then an emerge -av hal glib
Barely a fix, but it's working and that's good enough for the moment. |
|
| Back to top |
|
|
Jim6
Tux's lil' helper
Joined: 08 Apr 2005
Posts: 102
|
| Posted: Tue Jan 05, 2010 3:26 am Post subject: |
|
|
I just ran into this problem again upgrading hal.
You don't have to re-build glib with -O0. (only hal)
Check out the bug for details |
|
| Back to top |
|
|
|
Kernel & Hardware
