When booting kernel-xen-2.6.16-1.2289_FC6 my soundcard (Audiophile 24/96) no longer gets initialized properly. The snd_ice1712 shows up in lsmod but /proc/asound/cards is empty. After booting back into a regular kernel audio works fine again. scsound.log attached
Created attachment 131205 [details] logfile of soundcard detection on xen enabled kernel
I switched to a new box but kept the same soundcard and this still doesn't work. The kernel (2.6.19-1.2898.2.3.fc7xen) does seem to detect the soundcard but misses out on the pcm bits. Here is what scsrun.log says: ALSA lib pcm_hw.c:1357:(_snd_pcm_hw_open) Invalid value for card aplay: main:550: audio open error: No such device cat: /etc/asound.conf: No such file or directory amixer: Mixer attach default error: No such device Running from command-line...
Can you please include kernel logs of the soundcard detection? All I can see attached so far is the user-space alsa/alsalib logs, but we really need to see what the kernel is reporting at boot time to debug this. Thanks!
Created attachment 150698 [details] s-c-s logfile This is a current logfile generated by s-c-s. I checked both the dmesg output and /var/log/messages but neither seem to contain any messages related to the soundcard whatsoever. How do I get to the information you need?
"dmesg" shortly after boot should have it. From your quote "The kernel (2.6.19-1.2898.2.3.fc7xen) does seem to detect the soundcard" can we assume that the kernel is giving at least _some_ indication of having found it?
The dmesg output is present in the logfiles above but I've never seen any output by the sound-subsystem in the kernel messages either using xen or the regular kernel. I assumed that the modules found something by looking at /proc/asound/devices which contained this: 2: : timer 3: : sequencer but thinking about it now these are probably some generic elements that would appear even without a soundcard present. So it looks like the kernel loads the appropriate modules without indicating an error but alsa doesn't quite seem to pick it up from there. Is there any way to get more verbose output when loading the modules or initializing the audio subsystem?
change QA contact
This report targets FC6, which is now end-of-life. Please re-test against Fedora 7 or later, and if the issue persists, open a new bug. Thanks