Bug 6329

Summary: OOPS in snd_pcm_oss
Product: Drivers Reporter: Konrad Rzepecki (hannibal)
Component: Sound(ALSA)Assignee: Jaroslav Kysela (perex)
Status: CLOSED CODE_FIX    
Severity: normal CC: akpm, tiwai
Priority: P2    
Hardware: i386   
OS: Linux   
Kernel Version: 2.6.17-rc1 Subsystem:
Regression: --- Bisected commit-id:
Attachments: OOPS from syslog
config
OOPS from syslog - without NVIDIA
Other oops in snd_pcm_oss
Fix PCM OSS bug for 2.6.17rc1
Better fix for PCM-OSS Oops on 2.6.17rc1

Description Konrad Rzepecki 2006-04-04 08:13:17 UTC
Most recent kernel where this bug did not occur: ?
Distribution: Slackware-current
Hardware Environment: GIGABYTE 8I848E(ICH4) P4 1.8 512 MB RAM HDD 120GB 
SEAGATE 
Software Environment: KDE 3.5.2, XOrg 6.9, glibc 2.3.6
Problem Description: OOPS in snd_pcm_oss on exit fish fillets ng 
(http://fillets.sf.net/)

Steps to reproduce: Exit fish fillets v 0.7.0
Comment 1 Konrad Rzepecki 2006-04-04 08:16:44 UTC
Created attachment 7759 [details]
OOPS from syslog
Comment 2 Konrad Rzepecki 2006-04-04 08:17:53 UTC
Created attachment 7760 [details]
config
Comment 3 Andrew Morton 2006-04-04 13:18:08 UTC
This is repeatable, yes?

Could you please check that it happens in a kernel which hasn't
been tainted by the nvidia driver?

Did earlier kernels have this problem? 2.6.16?
Comment 4 Konrad Rzepecki 2006-04-04 16:09:13 UTC
Yes, it is repeatable.

2.6.16 works OK. Nvidia module absence has no influence to it.

I discovered that filles during crash on this oops returns (translations):
Audio write: Deskryptor pliku w złym stanie (File descriptor in bad shape)
Naruszenie ochrony pamięci (Segmentation fault)
Comment 5 Konrad Rzepecki 2006-04-04 16:27:34 UTC
Created attachment 7763 [details]
OOPS from syslog - without NVIDIA
Comment 6 Konrad Rzepecki 2006-04-05 08:50:26 UTC
Created attachment 7777 [details]
Other oops in snd_pcm_oss

I found out that other oss-using apps also can cause oopses (flash plugin). It
occurs on closing browser using flash plugin to play some sounds or music.
Comment 7 Takashi Iwai 2006-04-05 09:01:58 UTC
Created attachment 7778 [details]
Fix PCM OSS bug for 2.6.17rc1
Comment 8 Takashi Iwai 2006-04-05 09:02:30 UTC
Could you try the patch above?
Comment 9 Konrad Rzepecki 2006-04-05 09:35:44 UTC
It works OK. Both type of OOPS-es didn't occur again. Thanks for the patch.
Comment 10 Takashi Iwai 2006-04-05 12:05:39 UTC
Created attachment 7779 [details]
Better fix for PCM-OSS Oops on 2.6.17rc1
Comment 11 Konrad Rzepecki 2006-04-05 15:04:04 UTC
The second patch also works.
Comment 12 Martin Mokrejs 2006-04-13 00:01:26 UTC
The "Better fix for PCM-OSS Oops on 2.6.17rc1" patch fixed my Ooops reported here:
http://marc.theaimsgroup.com/?t=114479997800001&r=1&w=2
Comment 13 Konrad Rzepecki 2006-06-27 09:02:55 UTC
I'm closing this bug, because in 2.6.17 it no longer exists. Patch was applied  
long time ago (AFAIR before 2.6.17-rc2) into mainline.


BTW. What is the bug closing policy? I think, that bug owner should closing it 
(or subsystem mantainer). Submiter not always have enough knowlege about 
problem (this including myself ;-), and shouldn't do it. But also bug cannot 
be "new" forever, especially when it is in fact resolved...