Re: alsa drivers; was Re: kernels for Lenny.
On Thursday 07 August 2008 11:46, peasthope@shaw.ca wrote:
> Nigel and anyone interested in alsa and USB development,
>
> nh> ... both snd-intel8x0, and snd-hda-intel are
> being loaded, and it seems like they are both trying to grab card0, which
> is resulting in neither getting card0, and snd-usb-audio is getting card0,
> because the other 2 are having a non resolvable argument as to who gets
> card0.
>
> There are two sound devices. Intel 82801BA/BAM
> AC'97 on the system board and C-media USB plugged in.
>
> >From my previous investigation.
>
> dalton:~# cat /proc/asound/cards
> 0 [default ]: USB-Audio - C-Media USB Headphone Set
> C-Media USB Headphone Set at usb-0000:00:1f.2-2,
> full speed
This is normal from a long time back, and in my case, when I was using Fedora
Core 1. I have a usb midi keyboard, and it was always being set as card0, as
the the usb starts early in the boot sequence, and consequently the actual
sound card was unable to use the card0 slot, resulting in no sounds.
The fix was to add the following lines to, in the case of
Fedora's, /etc/modprobe.conf
options snd-emu10k1 index=0 (that's the actual soundcard)
options snd-usb-audio index=1 (that's for the usb midi keyboard)
>
> If the C-Media is card 0, then the Intel device
> should be card 1. The Intel drivers should not
> involve a USB device ... unless it contains
> an Intel or Intel-like chip.
I havn't found that that works. Logically yes, but in reality no. Looking back
at my situation in Fedora 1 above (that is with no options set), the usb
starts first in the bootup sequence, and thinks that my usb midi keyboard is
a soundcard, and sets it as card0. the actual soundcard using the snd-emu10k1
driver is looking for setting itself as card0, but the slot for card0 has
already been taken by the usb midi keyboard, and consequently the actual
soundcard doesn't exist in /proc/asound/cards
Btw, if I switch off the usb midi keyboard before booting, then bootup, the
emu10k1 is set as card0, and if I then switch the usb midi keyboard on, that
is set as card1, which is much as you see below.
>
> nh> Try adding the 3 lines below to the bottom of
> the alsa-base file.
>
> There is a simple test before editing alsa-base.
> Unplug the C-Media device before starting the
> system and plug it after the on-board Intel chip
> is initialized. This is from my first trial.
>
> dalton:~# cat /proc/asound/cards
> 0 [I82801BAICH2 ]: ICH - Intel 82801BA-ICH2
> Intel 82801BA-ICH2 with AD1885 at irq 17
> 1 [default ]: USB-Audio - C-Media USB Headphone Set
> C-Media USB Headphone Set at usb-0000:00:1f.2-2,
> full speed
Ok, I see where your going here. Having plugged in the C-Media device after
booting up, the big question is, does the sound work? It should do, as
snd-intel8x0 is set as card0
>
> Which seems consistent with my hypothesis.
>
> Further hypothesizing: the Intel 82801 invokes
> snd_intel8x0 while the C-Media USB invokes snd_hda_intel.
>
> I'll try this unplug/replug procedure a few more times
> and report next week.
>
> USB and udev are appealing concepts but implementation
> is certainly non-trivial.
>
> Thanks for the help. Progress would be difficult
> without it.
>
> Regards, ... Peter E.
Did you try adding the 3 lines that I suggested to /etc/modprobe.d/alsa-base?
It would be nice to find out if the C-Media device is actually causing
snd-hda-intel to load, thus causing the conflict when booting with the usb
device plugged in.
Just trying to help.
Nigel.
>
>
>
>
>
>
>
>
>
>
>
> --
> http://members.shaw.ca/peasthope/
> http://carnot.yi.org/ = http://carnot.pathology.ubc.ca/
Reply to: