* [Xenomai-help] Kernel oops on posix threads
@ 2005-10-12 15:33 Luotao Fu
2005-10-12 15:52 ` Gilles Chanteperdrix
2005-10-12 15:59 ` Gilles Chanteperdrix
0 siblings, 2 replies; 6+ messages in thread
From: Luotao Fu @ 2005-10-12 15:33 UTC (permalink / raw)
To: xenomai
[-- Attachment #1: Type: text/plain, Size: 1200 bytes --]
Hi folks,
I found out that parallel threads using the Posix skin cause kernel oops
somewhere between the finishing of the threads and the main thread. I
have two threads, say thread A and thread B. Both created as joinable
threads and they have both the same priority 99. The main routine hold a
semaphore semA to synchronize the threads. After the creation thread A
waits on the Semaphore semA, while the thread B runs and releases semA,
so that semA can finish. On exiting the thread B I always get a Kernel
ooops. The kernerl says:
unable to handle kernel paging request at virtual address 6b6b6b77
etc.......
I scratched some simple code which causes this Kernel Ooops. Packed a
tar ball together with the makefile and attached the tar ball it to this
mail so that you might have a look at it yourself.
I'm using
Kernel: Vanilla 2.6.12 patched with adeos-linux-2.6.12-i386-r12.patch
Distr. Debian 3.1
Testhost: vmware 5.0
compiler gcc 3.3.5
I also wrote some code doing the same tasks in RTAI native context and
it ran without problems. Hence I suppose the problem to be in the posix
skin.
Am I making any mistakes in the code or have I evtl. found a bug in the
posix skin?
Cheers
Luotao Fu
[-- Attachment #2: test_posix.tar.gz --]
[-- Type: application/gzip, Size: 1198 bytes --]
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Xenomai-help] Kernel oops on posix threads
2005-10-12 15:33 [Xenomai-help] Kernel oops on posix threads Luotao Fu
@ 2005-10-12 15:52 ` Gilles Chanteperdrix
2005-10-12 15:59 ` Gilles Chanteperdrix
1 sibling, 0 replies; 6+ messages in thread
From: Gilles Chanteperdrix @ 2005-10-12 15:52 UTC (permalink / raw)
To: Luotao Fu; +Cc: xenomai
Luotao Fu wrote:
> Hi folks,
>
> I found out that parallel threads using the Posix skin cause kernel oops
> somewhere between the finishing of the threads and the main thread. I
> have two threads, say thread A and thread B. Both created as joinable
> threads and they have both the same priority 99. The main routine hold a
> semaphore semA to synchronize the threads. After the creation thread A
> waits on the Semaphore semA, while the thread B runs and releases semA,
> so that semA can finish. On exiting the thread B I always get a Kernel
> ooops. The kernerl says:
>
> unable to handle kernel paging request at virtual address 6b6b6b77
> etc.......
Thanks a lot for your report, it is almost perfect. However, it would be
really helpful to have the "etc...." in order to have a stack dump and
have a quick look at the code before trying to reproduce the situation
here.
>
> I scratched some simple code which causes this Kernel Ooops. Packed a
> tar ball together with the makefile and attached the tar ball it to this
> mail so that you might have a look at it yourself.
> I'm using
> Kernel: Vanilla 2.6.12 patched with adeos-linux-2.6.12-i386-r12.patch
> Distr. Debian 3.1
> Testhost: vmware 5.0
> compiler gcc 3.3.5
>
> I also wrote some code doing the same tasks in RTAI native context and
> it ran without problems. Hence I suppose the problem to be in the posix
> skin.
> Am I making any mistakes in the code or have I evtl. found a bug in the
> posix skin?
A kernel oops is almost always certainly a bug... Could you tell me
whether this happens even the first time after insertion of the
rtai_posix module ?
Also, why do you link your program with the rtai library, rtai-config
--posix-ldflags should be enough ?
--
Gilles Chanteperdrix.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Xenomai-help] Kernel oops on posix threads
2005-10-12 15:33 [Xenomai-help] Kernel oops on posix threads Luotao Fu
2005-10-12 15:52 ` Gilles Chanteperdrix
@ 2005-10-12 15:59 ` Gilles Chanteperdrix
2005-10-12 16:17 ` Luotao Fu
1 sibling, 1 reply; 6+ messages in thread
From: Gilles Chanteperdrix @ 2005-10-12 15:59 UTC (permalink / raw)
To: Luotao Fu; +Cc: xenomai
Luotao Fu wrote:
> I scratched some simple code which causes this Kernel Ooops. Packed a
> tar ball together with the makefile and attached the tar ball it to this
> mail so that you might have a look at it yourself.
> I'm using
> Kernel: Vanilla 2.6.12 patched with adeos-linux-2.6.12-i386-r12.patch
> Distr. Debian 3.1
> Testhost: vmware 5.0
> compiler gcc 3.3.5
Sorry for the second answer: what version of RTAI/Fusion ?
If it is easy for you, could you try with the current version of
Xenomai ?
--
Gilles Chanteperdrix.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [Xenomai-help] Kernel oops on posix threads
2005-10-12 15:59 ` Gilles Chanteperdrix
@ 2005-10-12 16:17 ` Luotao Fu
2005-10-12 16:26 ` Gilles Chanteperdrix
0 siblings, 1 reply; 6+ messages in thread
From: Luotao Fu @ 2005-10-12 16:17 UTC (permalink / raw)
To: Gilles Chanteperdrix, xenomai
[-- Attachment #1: Type: text/plain, Size: 903 bytes --]
Hi,
Gilles Chanteperdrix schrieb:
> Luotao Fu wrote:
> > I scratched some simple code which causes this Kernel Ooops. Packed a
> > tar ball together with the makefile and attached the tar ball it to this
> > mail so that you might have a look at it yourself.
> > I'm using
> > Kernel: Vanilla 2.6.12 patched with adeos-linux-2.6.12-i386-r12.patch
> > Distr. Debian 3.1
> > Testhost: vmware 5.0
> > compiler gcc 3.3.5
>
> Sorry for the second answer: what version of RTAI/Fusion ?
> If it is easy for you, could you try with the current version of
> Xenomai ?
>
Thanx for you answer.
Sorry for the insufficiency, I forgot to mention it. I use fusion 0.9.1,
compiled it just two weeks ago. To the questions in you last mail:
1.Yes I get the oops also the first time after I inserted the posix module.
2. The complete Oops message: see attachment
Hope it helps.
Thanx again
Cheers
Luotao Fu
[-- Attachment #2: oops.txt --]
[-- Type: text/plain, Size: 1511 bytes --]
Unable to handle kernel paging request at virtual address 6b6b6b77
printing eip:
c4aaef89
*pde = 00000000
Oops: 0000 [#1]
PREEMPT
Modules linked in: nfs lockd sunrpc rtai_posix rtai_nucleus rtai_hal md5 ipv6 af_packet piix ide_generic intel_agp evdev sd_mod ide_cd cdrom ide_disk ide_core floppy vga16fb vgastate mptscsih mptbase scsi_mod unix
CPU: 0
EIP: 0060:[<c4aaef89>] Not tainted VLI
EFLAGS: 00010082 (2.6.12.6-adeos-patched)
EIP is at xnpod_schedule+0x754/0x87b [rtai_nucleus]
eax: 6b6b6b6b ebx: c4abf110 ecx: 8005003b edx: c2d5b020
esi: 00000000 edi: c4b58330 ebp: c2d5b540 esp: c2f2bedc
ds: 007b es: 007b ss: 0068
Process test (pid: 1658, threadinfo=c2f2a000 task=c2d5b540)
Stack: 00000000 00200000 00000001 c4abede0 c4abf110 c4b50033 c0100000 00000000
8005003b c4abf460 c4b58330 00000000 00000000 c4aadcf9 00000000 c4abede0
00000100 00000000 c4b58330 c4a5bfa8 c0353868 c4ab1de1 c4b58330 00000100
Call Trace:
[<c4aadcf9>] xnpod_suspend_thread+0x178/0x1d1 [rtai_nucleus]
[<c4ab1de1>] xnshadow_relax+0xa6/0xf4 [rtai_nucleus]
[<c4ab27cd>] hisyscall_event+0x261/0x2af [rtai_nucleus]
[<c0129c76>] __adeos_handle_event+0x70/0x113
[<c010d239>] __adeos_enter_syscall+0x16/0x83
[<c0102851>] system_call+0x25/0x46
Code: 00 74 05 0f ae 08 eb 02 dd 20 8b 54 24 10 8b 44 24 0c 89 90 20 03 00 00 eb 5a 8b 5c 24 10 8b 93 10 02 00 00 85 d2 74 09 8b 42 04 <f6> 40 0c 01 74 43 0f 06 a1 8c d2 2a c0 a9 00 00 00 01 75 35 85
<6>note: test[1658] exited with preempt_count 2
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [Xenomai-help] Kernel oops on posix threads
2005-10-12 16:17 ` Luotao Fu
@ 2005-10-12 16:26 ` Gilles Chanteperdrix
2005-10-12 16:30 ` Gilles Chanteperdrix
0 siblings, 1 reply; 6+ messages in thread
From: Gilles Chanteperdrix @ 2005-10-12 16:26 UTC (permalink / raw)
To: Luotao Fu; +Cc: xenomai
Luotao Fu wrote:
> Hi,
>
> Gilles Chanteperdrix schrieb:
> > Luotao Fu wrote:
> > > I scratched some simple code which causes this Kernel Ooops. Packed a
> > > tar ball together with the makefile and attached the tar ball it to this
> > > mail so that you might have a look at it yourself.
> > > I'm using
> > > Kernel: Vanilla 2.6.12 patched with adeos-linux-2.6.12-i386-r12.patch
> > > Distr. Debian 3.1
> > > Testhost: vmware 5.0
> > > compiler gcc 3.3.5
> >
> > Sorry for the second answer: what version of RTAI/Fusion ?
> > If it is easy for you, could you try with the current version of
> > Xenomai ?
> >
> Thanx for you answer.
>
> Sorry for the insufficiency, I forgot to mention it. I use fusion 0.9.1,
> compiled it just two weeks ago. To the questions in you last mail:
> 1.Yes I get the oops also the first time after I inserted the posix module.
> 2. The complete Oops message: see attachment
>
>
> Hope it helps.
>
> Thanx again
> Cheers
> Luotao Fu
> Unable to handle kernel paging request at virtual address 6b6b6b77
> printing eip:
> c4aaef89
> *pde = 00000000
> Oops: 0000 [#1]
> PREEMPT
> Modules linked in: nfs lockd sunrpc rtai_posix rtai_nucleus rtai_hal md5 ipv6 af_packet piix ide_generic intel_agp evdev sd_mod ide_cd cdrom ide_disk ide_core floppy vga16fb vgastate mptscsih mptbase scsi_mod unix
> CPU: 0
> EIP: 0060:[<c4aaef89>] Not tainted VLI
> EFLAGS: 00010082 (2.6.12.6-adeos-patched)
> EIP is at xnpod_schedule+0x754/0x87b [rtai_nucleus]
> eax: 6b6b6b6b ebx: c4abf110 ecx: 8005003b edx: c2d5b020
> esi: 00000000 edi: c4b58330 ebp: c2d5b540 esp: c2f2bedc
> ds: 007b es: 007b ss: 0068
> Process test (pid: 1658, threadinfo=c2f2a000 task=c2d5b540)
> Stack: 00000000 00200000 00000001 c4abede0 c4abf110 c4b50033 c0100000 00000000
> 8005003b c4abf460 c4b58330 00000000 00000000 c4aadcf9 00000000 c4abede0
> 00000100 00000000 c4b58330 c4a5bfa8 c0353868 c4ab1de1 c4b58330 00000100
> Call Trace:
> [<c4aadcf9>] xnpod_suspend_thread+0x178/0x1d1 [rtai_nucleus]
> [<c4ab1de1>] xnshadow_relax+0xa6/0xf4 [rtai_nucleus]
> [<c4ab27cd>] hisyscall_event+0x261/0x2af [rtai_nucleus]
> [<c0129c76>] __adeos_handle_event+0x70/0x113
> [<c010d239>] __adeos_enter_syscall+0x16/0x83
> [<c0102851>] system_call+0x25/0x46
> Code: 00 74 05 0f ae 08 eb 02 dd 20 8b 54 24 10 8b 44 24 0c 89 90 20 03 00 00 eb 5a 8b 5c 24 10 8b 93 10 02 00 00 85 d2 74 09 8b 42 04 <f6> 40 0c 01 74 43 0f 06 a1 8c d2 2a c0 a9 00 00 00 01 75 35 85
> <6>note: test[1658] exited with preempt_count 2
>
Ok, you really ought to try xenomai, because this strangely looks like a
bug that was fixed. It was not as reproducible as that in my test cases
though...
--
Gilles Chanteperdrix.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2005-10-12 16:30 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-10-12 15:33 [Xenomai-help] Kernel oops on posix threads Luotao Fu
2005-10-12 15:52 ` Gilles Chanteperdrix
2005-10-12 15:59 ` Gilles Chanteperdrix
2005-10-12 16:17 ` Luotao Fu
2005-10-12 16:26 ` Gilles Chanteperdrix
2005-10-12 16:30 ` Gilles Chanteperdrix
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.