At Sun, 18 Oct 2009 18:37:28 +0200, Lubomir Rintel wrote:
On Mon, 2009-10-12 at 10:08 +0200, Takashi Iwai wrote:
At Mon, 12 Oct 2009 10:03:53 +0200, Lubomir Rintel wrote:
On Mon, 2009-10-12 at 09:18 +0200, Takashi Iwai wrote:
At Mon, 12 Oct 2009 08:46:26 +0200, Clemens Ladisch wrote:
Lubomir Rintel wrote:
Entry for JMTek LLC., SSS USB Headphone Set in the quirk table.
Please add an explaination why this entry is needed. At first glance, this entry seems to describe a class-compliant device that should not need a quirk.
Ah right. It'd really helpful if Lubomir can give more details...
A 2.6.31.1-based kernel on my Fedora 12 workstation at hone seemed to require that, only the input driver attached to the device when plugged in, the snd-usb-audio didn't seem to load and did not care about the device when loaded manually. I did not have an idea why, since "alias: usb:v*p*d*dc*dsc*dp*ic01isc01ip*" really seemed to match my device. Nevertheless, after finding out that adding an entry to the quirk table solves my problem I concluded that my understanding (or lack of thereof) was wrong and that alias is really not meant to match my device.
Now I plugged the adapter into my work lappy with 2.6.30.8-based Fedora 11 and the audio interfaces on my adapter got instantly recognized and claimed by snd-usb-audio (and the input interface by the input subsystem), without modifying anything.
Any clues what could have gone wrong then?
Hm, I don't see any affecting changes between 2.6.30 and 31. Just as a test, could you copy sound/usb/*.[ch] from 2.6.30 tree to 2.6.31 or 32-rc and check whether the same problem appears?
I did not try that, but updated to a newer snapshot and it works as it used to again, so I don't thing there's much point in finding out what went wrong now. I'm very sorry for the noise.
OK, thanks for confirmation.
Takashi