Description of problem: system-config-soundcord writes non-functional, and bogus info to /etc/modprobe.conf Version-Release number of selected component (if applicable): system-config-soundcard-2.0.6-11.fc8 How reproducible: Twice, log to follow Steps to Reproduce: 1. 2. 3. Actual results: Once it write /etc/modprobe.conf alias snd-card-0 snd-hda-intel options snd-card-0 index=0 options snd-hda-intel index=0 It also writes something like this: options snd cards_limit=8 alias snd-card-0 snd-hda-intel options snd-hda-intel index=0 alias snd-card-7 snd-usb-audio options snd-usb-audio index=7 Expected results: The working manual editing to /etc/modprobe.conf is: alias snd-card-0 snd-hda-intel options snd-card-0 index=0 options snd-hda-intel model=toshiba Additional info: System-config-soundcard also seems to have problem unloading the kernel module and reloading (conflict with pulse audio?)
Created attachment 294510 [details] log of first failure
Created attachment 294511 [details] log of 2nd failure
Created attachment 294512 [details] scsrun log
> options snd-hda-intel model=toshiba This specific configuration has to be fixed by upstream for all distros - file a bug at www.alsa-project.org. We can keep a specific config for all models. > System-config-soundcard also seems to have problem unloading > the kernel module and reloading (conflict with pulse audio?) Yes, it's possible. s-c-s can't unload a kernel module when it's used by some other app or the sound server.
fair enough - I just thought there are a few similiar bugs in redhat's bugzilla and it would be more useful if redhat people file one bug on behalf of the bunch upstream. I think the kernel module should try harder guessing what options to use (e.g. default to model=auto), but there are still a few criticism against sysmte-config-sound-card: 1) what it did insert, "model=0" is none of any of the documented options, and letting the default (no kernel module parameters, or "model=auto") taking over may work better in some cases (not in mine), (2) it was inserting bogus entries on my 2nd attempt. (3) it could try to shut down the sound server, or gives some hints on how to do so when it fails to unload the kernel module. anyway, my current configuration works, and future version of the sound module might just probe the hardware better, so this is probably not important... the information about identifying hardware, etc in the logs may still be useful if forwarded to the alsa people in a bundle together with some of the other bugs.
(In reply to comment #5) > fair enough - I just thought there are a few similiar bugs in redhat's bugzilla > and it would be more useful if redhat people file one bug on behalf of the > bunch upstream. I can't file the bug because i don't have your hardware available. There are usually many hw-specific question when you file any hw specific bug. > I think the kernel module should try harder guessing what options to use > (e.g. default to model=auto), but there are still a few criticism > against sysmte-config-sound-card: 1) what it did insert, "model=0" is > none of any of the documented options, and letting the default (no kernel > module parameters, or "model=auto") > taking over may work better in some cases (not in mine), (2) it was inserting > bogus entries on my 2nd attempt. (3) it could try to shut down the sound server, > or gives some hints on how to do so when it fails to unload the kernel module. Yes, the current fedora sound setup is quite old and doesn't reflect the latest changes. Unfortunately i don't have a time to work on it now. Feel free to create any patch that improves it and i can review/merge it then.