Most recent kernel where this bug did not occur: Bug occurs in all kernels using the alsa drivers. Distribution: Slackware 10 Problem also occurs in Slackware 9 Hardware Environment: Celeron 400mhz -- Non Copermine processor. Asus P2-99 Motherboard (please no comments on how much my system sucks I already know but I need this software to work for now.) 300MB PC133 SDRAM 20GB hard Drive 46% full. Sound Blaster AWE (64) ISA card. (pnp) (P2-99 has two ISA slots sound card uses one. The other is empty) Software Environment: Xwindows Systmem. KDE 3.2 (Arts server is turned off) KMixer is used to adjust sound card settings. Skype (Version 1.2.0.11) Latest available Linux version. Problem Description: Entire Alsa subsystem crashes when using the Skype software. This driver crashed on the 2.6.11.8 Kernel the first time after performing a voice chat for 2 hours. Entire sound system goes down and I lose sound until I restart. Attempting to restart the driver using modprobe doesn't work. I am getting no visible errors from the kernel at this time. Search of syslog files shows no visible errors. Once the initial crash happens it takes considerably less time for the driver to crash again the next time. So something on the system is getting changed to induce the problem. This issue remains in place even after reboots. Reinstallation of the skype software appears to have no effect on the amount of time between crashes after the first crash appeared during the first installation. So I don't believe its a problem with skype but with the ALSA driver. Or at least with the OSS emulator as I believe skype still uses the OSS subsystem. After crash occurs turning the arts server back on produces no sound. Problem occurs in versions 2.4 and 2.6 kernels. (ALL of them) Steps to reproduce: Running a voice chat in skype for 2 hours first time. And no more then 10 minutes after the first crash has occured. (I've placed a bug report also to the Alsa Project but I haven't had any response. They either never got around to it or are ignoring it. Currently the bug report is unassigned.)
If you have any questions about this bug. Please feel free to email me at philip@digitalinfinity.biz. thank you.
Sorry, we can't debug problems with proprietary software. If the bug is still present with the latest versions of skype and ALSA, please try to reproduce it with an open source VoIP app like kiax.
Please reopen this bug if: - it is still present in kernel 2.6.16.5 and - you can reproduce it with an open source application.
This bug as far as I can see only happens with skype. I haven't been able to reproduce it with Linphone or gnome meeting. However, I'm no longer using that computer due to the problem. The SB Audigy 2 card doesn't have the problem. It only happens with the ISA driver with the SB awe 32 and 64. It seems to be present prodominately in slackware. However I haven't tried other distros due to the fact that the computer is too slow for me to use another distro. It may have also been fixed already in the latest version. But it was a problem that prevented me from sometimes using the program for more then an hour. This was meant as a bug for the SB 16, awe 32 and awe 64 sound driver. As they are all the same driver. I've also filed a bug report on the alsa website. They were trouble shooting the problem with me however no further information was given to me since.
One more thing the Alsa people seemed to think it wasn't a problem with the skype program but a problem with the sound driver itself.
What is the alsa bugreport number, Philip? Has this been solved? It's a kernel-side problem? (are there oopses in dmesg or something?) If it is, I don't see why it's neccesary to reproduce with a open source app, if a propietary userspace app breaks the kernel using valid kernel APIs then it's a valid kernel bug.
Bug Identification number for this problem on the www.alsa-project.org is ID: 0001337 The bug still appears to be open. This bug has all the dmesg information that I was able to aquire for the fix. I honestly couldn't tell you if its been fixed. I gave up and decided to buy a new computer with a new sound blaster card. As far as I can see its not a problem with my new card. It appears to be a driver issue.
I can't find the dmesg there in the alsa bug for some reason... But in any case, it's been a while. There were fixes done to sound system lately. Philip, can you please confirm this is still a problem with the latest kernel? Thanks.
I suppose we can close the bug, since there is no recent activity (and no complaints).