Skip to contentSkip to conversation

Kernel 2.6.17.1 bug/oops in snd_usb_audio subsystem

DISCUSSION Martin van Es 3 participants 5 messages Jul 14, 2006

lore
Share & tools
lore
#0 lore raw

Hi,

I already sent this mail to [email protected] a couple of days ago since I thought that was the closest match in the MAINTAINERS list for this oops. Since I didn't receive any reply I am not sure if the information in this mail is known to the developing community so I'll send a one time copy to the list. Please forgive me for not being a member of the list, I just happen to follow (stable) kernel development and stumbled upon this oops...

If there's anything I forgot, or should test/send to this list or person, please CC me on the reply.

Regards,
Martin van Es

-- original message --

Device:
The mentioned usb SOUND device is part of a philips webcam, dmesg output:
pwc: Philips PCVC675K (Vesta) USB webcam detected.

and registers as /dev/dsp1

I was able to reproduce the bug every time I start up ffmpeg (using /dev/dsp1) in 2.6.17.1, but for 1 time. I tried the same with 2.6.16 (only) once which worked fine.

Maybe helpful:
I deselected support for old ALSA API in 2.6.17.1

I didn't find a reference to the snd_usb_audio subsystem in recent 2.6.17 releases (> .1) changelog so didn't try to compile a new version of the 2.6.17.x kernel.

Grtz.
Martin van Es

dmesg output:

BUG: unable to handle kernel NULL pointer dereference at virtual address 

000001b8
 printing eip:
deae2514
*pde = 00000000

Oops: 0002 [#1]
Modules linked in: snd_usb_audio pwc snd_usb_lib videodev v4l2_common 

snd_rawmidi snd_hwdep capability commoncap kqemu ipw2200
CPU:    0
EIP:    0060:[<deae2514>]    Tainted: P      VLI
EFLAGS: 00210202   (2.6.17.1 #5)
EIP is at snd_usb_pcm_open+0x34/0x4d0 [snd_usb_audio]
eax: db67e780   ebx: 00000000   ecx: 0000000e   edx: 000001a8
esi: deaf0d20   edi: c5be1e50   ebp: c56379c0   esp: c5be1d9c
ds: 007b   es: 007b   ss: 0068
Process ffmpeg (pid: 19504, threadinfo=c5be0000 task=c678aa70)

Stack: 00000020 c5be1dd8 00000000 d1d49000 00000010 c56379c0 c030f7f4 d1d49000
       00000000 00000010 d1d49000 000001a8 00000011 0000000b ffffffff cfcdf600
       00000000 0000000d c5be1e50 c56379c0 c030fb17 db67e780 00000001 c61a4c80
Call Trace:
 <c030f7f4> snd_pcm_hw_constraints_init+0x704/0x790  <c030fb17> 
snd_pcm_open_substream+0x57/0xb0
 <c031df41> snd_pcm_oss_open+0x221/0x4a0  <c0163800> 
generic_permission+0x110/0x120
 <c02126c7> kobject_get+0x17/0x20  <c01185c0> default_wake_function+0x0/0x20
 <c030482d> soundcore_open+0x8d/0x1c0  <c015e826> chrdev_open+0x76/0x160
 <c015e7b0> chrdev_open+0x0/0x160  <c0153aeb> __dentry_open+0xbb/0x200
 <c0153d4c> do_filp_open+0x5c/0x70  <c01539cb> get_unused_fd+0x5b/0xc0
 <c0153db3> do_sys_open+0x53/0x100  <c0153eb7> sys_open+0x27/0x30
 <c0102f47> syscall_call+0x7/0xb
Code: 8b 48 5c 8d 14 52 fc 89 4c 24 28 89 d1 c1 e1 04 01 ca b9 0e 00 00 00 8d 

14 d5 10 00 00 00 89 54 24 2c 8b 58 08 01 da 89 54 24 2c <c7> 42 10 ff ff ff ff c7 42 24 00 00 00 00 8b 7c 24 28 81 c7 d0
EIP: [<deae2514>] snd_usb_pcm_open+0x34/0x4d0 [snd_usb_audio] SS:ESP 0068:c5be1d9c

#1 lore raw
Quoted 3 lines · Martin van Es

I already sent this mail to [email protected] a couple of days ago since I thought that was the closest match in the MAINTAINERS list for this oops. Since I didn't receive any reply

Your kernel is tainted. Please complain to kqemu developers. Don't waste Jaroslav's time.

Quoted 10 lines
BUG: unable to handle kernel NULL pointer dereference at virtual address

000001b8
 printing eip:
deae2514
*pde = 00000000

Oops: 0002 [#1]
Modules linked in: snd_usb_audio pwc snd_usb_lib videodev v4l2_common

snd_rawmidi snd_hwdep capability commoncap kqemu ipw2200

EIP:    0060:[<deae2514>]    Tainted: P      VLI

#2 lore raw
Quoted 7 lines · Alexey Dobriyan
Quoted 3 lines · Martin van Es

I already sent this mail to [email protected] a couple of days ago since I thought that was the closest match in the MAINTAINERS list for this oops. Since I didn't receive any reply

Your kernel is tainted. Please complain to kqemu developers. Don't waste Jaroslav's time.

Thank you Alexey for pointing out I reported this oops while running a tainted kernel (kqemu was indeed tainting it). I interpreted the reply as a request for reproducing the oops on a untainted version of the Kernel of which I'll post the dmesg output below.

Hope this helps.
Is there anything else I can do to contribute?

BUG: unable to handle kernel NULL pointer dereference at virtual address 

000001b8
printing eip:
dea9a514
*pde = 00000000

Oops: 0002 [#1]
Modules linked in: snd_usb_audio snd_usb_lib pwc snd_rawmidi snd_hwdep 
videodev v4l2_common capability commoncap ipw2200
CPU:    0
EIP:    0060:[<dea9a514>]    Not tainted VLI
EFLAGS: 00210202   (2.6.17.1 #5)

EIP is at snd_usb_pcm_open+0x34/0x4d0 [snd_usb_audio]
eax: d681e9c0 ebx: 00000000 ecx: 0000000e edx: 000001a8
esi: deaa8d20 edi: ca6a7e50 ebp: c8f5afc0 esp: ca6a7d9c
ds: 007b es: 007b ss: 0068
Process ffmpeg (pid: 4793, threadinfo=ca6a6000 task=caf275b0)

Stack: 00000020 ca6a7dd8 00000000 cedf1c00 00000010 c8f5afc0 c030f7f4 cedf1c00
       00000000 00000010 cedf1c00 000001a8 00000011 0000000b ffffffff d3eaba00
       00000000 0000000d ca6a7e50 c8f5afc0 c030fb17 d681e9c0 00000001 c8edeb00
Call Trace:
 <c030f7f4> snd_pcm_hw_constraints_init+0x704/0x790  <c030fb17> 
snd_pcm_open_substream+0x57/0xb0
 <c031df41> snd_pcm_oss_open+0x221/0x4a0  <c0163800> 
generic_permission+0x110/0x120
 <c02126c7> kobject_get+0x17/0x20  <c01185c0> default_wake_function+0x0/0x20
 <c030482d> soundcore_open+0x8d/0x1c0  <c015e826> chrdev_open+0x76/0x160
 <c015e7b0> chrdev_open+0x0/0x160  <c0153aeb> __dentry_open+0xbb/0x200
 <c0153d4c> do_filp_open+0x5c/0x70  <c01539cb> get_unused_fd+0x5b/0xc0
 <c0153db3> do_sys_open+0x53/0x100  <c0153eb7> sys_open+0x27/0x30
 <c0102f47> syscall_call+0x7/0xb
Code: 8b 48 5c 8d 14 52 fc 89 4c 24 28 89 d1 c1 e1 04 01 ca b9 0e 00 00 00 8d 

14 d5 10 00 00 00 89 54 24 2c 8b 58 08 01 da 89 54 24 2c <c7> 42 10 ff ff ff ff c7 42 24 00 00 00 00 8b 7c 24 28 81 c7 d0
EIP: [<dea9a514>] snd_usb_pcm_open+0x34/0x4d0 [snd_usb_audio] SS:ESP 0068:ca6a7d9c

#3 lore raw

At Fri, 14 Jul 2006 13:28:40 +0200,

Quoted 10 lines · Martin van Es
Quoted 7 lines · Alexey Dobriyan
Quoted 3 lines · Martin van Es

I already sent this mail to [email protected] a couple of days ago since I thought that was the closest match in the MAINTAINERS list for this oops. Since I didn't receive any reply

Your kernel is tainted. Please complain to kqemu developers. Don't waste Jaroslav's time.

Quoted 7 lines

Thank you Alexey for pointing out I reported this oops while running a tainted kernel (kqemu was indeed tainting it). I interpreted the reply as a request for reproducing the oops on a untainted version of the Kernel of which I'll post the dmesg output below.

Hope this helps.
Is there anything else I can do to contribute?

Try to compile with CONFIG_SND_DEBUG=y. Also, you can try to build modules from alsa-driver-1.0.12rc1 whether the bug still exists in the latest version.

The problem is on OSS emulation, so check other OSS apps whether the similar problem happens.

Takashi

#4 lore raw
Quoted 17 lines · Takashi Iwai

At Fri, 14 Jul 2006 13:28:40 +0200,

Quoted 14 lines · Martin van Es
Quoted 7 lines · Alexey Dobriyan
Quoted 3 lines · Martin van Es

I already sent this mail to [email protected] a couple of days ago since I thought that was the closest match in the MAINTAINERS list for this oops. Since I didn't receive any reply

Your kernel is tainted. Please complain to kqemu developers. Don't waste Jaroslav's time.

Thank you Alexey for pointing out I reported this oops while running a tainted kernel (kqemu was indeed tainting it). I interpreted the reply as a request for reproducing the oops on a untainted version of the Kernel of which I'll post the dmesg output below.

Quoted 6 lines
Quoted 2 lines

Hope this helps.
Is there anything else I can do to contribute?

Try to compile with CONFIG_SND_DEBUG=y. Also, you can try to build modules from alsa-driver-1.0.12rc1 whether the bug still exists in the latest version.

Thx!

This is dmesg output with CONFIG_SND_DEBUG=y:

BUG: unable to handle kernel NULL pointer dereference at virtual address 

000001b8
printing eip:
dea9a514
*pde = 00000000

Oops: 0002 [#1]
Modules linked in: snd_usb_audio snd_usb_lib pwc snd_rawmidi snd_hwdep 
videodev v4l2_common capability commoncap ipw2200
CPU:    0
EIP:    0060:[<dea9a514>]    Not tainted VLI
EFLAGS: 00210202   (2.6.17.1 #5)

EIP is at snd_usb_pcm_open+0x34/0x4d0 [snd_usb_audio]
eax: d6f115c0 ebx: 00000000 ecx: 0000000e edx: 000001a8
esi: deaa8d20 edi: c8ed7e50 ebp: d5631ee0 esp: c8ed7d9c
ds: 007b es: 007b ss: 0068
Process ffmpeg (pid: 4950, threadinfo=c8ed6000 task=ca731a30)

Stack: 00000020 c8ed7dd8 00000000 c8ec7000 00000010 d5631ee0 c030f7f4 c8ec7000
       00000000 00000010 c8ec7000 000001a8 00000011 0000000b ffffffff cd1a0400
       00000000 0000000d c8ed7e50 d5631ee0 c030fb17 d6f115c0 00000001 c921c900
Call Trace:
 <c030f7f4> snd_pcm_hw_constraints_init+0x704/0x790  <c030fb17> 
snd_pcm_open_substream+0x57/0xb0
 <c031df41> snd_pcm_oss_open+0x221/0x4a0  <c0163800> 
generic_permission+0x110/0x120
 <c02126c7> kobject_get+0x17/0x20  <c01185c0> default_wake_function+0x0/0x20
 <c030482d> soundcore_open+0x8d/0x1c0  <c015e826> chrdev_open+0x76/0x160
 <c015e7b0> chrdev_open+0x0/0x160  <c0153aeb> __dentry_open+0xbb/0x200
 <c0153d4c> do_filp_open+0x5c/0x70  <c01539cb> get_unused_fd+0x5b/0xc0
 <c0153db3> do_sys_open+0x53/0x100  <c0153eb7> sys_open+0x27/0x30
 <c0102f47> syscall_call+0x7/0xb
Code: 8b 48 5c 8d 14 52 fc 89 4c 24 28 89 d1 c1 e1 04 01 ca b9 0e 00 00 00 8d 

14 d5 10 00 00 00 89 54 24 2c 8b 58 08 01 da 89 54 24 2c <c7> 42 10 ff ff ff ff c7 42 24 00 00 00 00 8b 7c 24 28 81 c7 d0
EIP: [<dea9a514>] snd_usb_pcm_open+0x34/0x4d0 [snd_usb_audio] SS:ESP 0068:c8ed7d9c

Regards,
Martin