[LAD] alsa card order

Simon Fielding s.fielding at wmcce.org
Wed Apr 29 11:45:08 UTC 2009

On Tue, 2009-04-28 at 21:22 -0400, drew Roberts wrote:
> >
> > The naming system exists for a reason, i.e. to allow you to easily
> > identify devices without remembering any cryptic numbers. It also
> > allows you to change the name during runtime via sysfs in case you
> > want to fix them up.
> >
> > > 1. I *want* stable indices, and
> >
> > Why?
> Well, I don't care how I would do it so long as I can finally find a way to do 
> it.
> I want to load up 1 box with 4 sound cards (preferably el cheapo USB cards) 
> and hoot the outputs of 4 radios to the inputs of the cards. (I prefer 
> identical cards if possible.) I want to tune each radio to a different 
> station and record each 24x7. now I want the stations to record consistently.
> So far I have not run across a solution that didn't assume different cards. 
> Any ideas anyone/
> all the best,

As far as I am aware, the only way to do anything like this is with
udev. I don't have any spare sound cards in the office so I've done some
experimenting with 2 identical mice. Running the following command

udevadm info -a -p $(udevadm info -q path -n /dev/input/mouse0)

for each mouse (soundcard in your case) replacing mouse0 with mouse1,
mouse2 etc and comparing the results shows that the only way to uniquely
identify them is by the usb port number eg :-


Providing that your radio/soundcard/usb port triplets remain the same,
either by never unplugging them, or by labelling them so that the same
radio connects to the same soundcard connects to the same usb port every
time, you could use the above ATTRS to write udev rules
in /etc/udev/rules.d/ which will allow you to name your devices whatever
you wish eg :-
/dev/radios/luxembourg /dev/radios/texasbigbeat etc.

Caveat: I have never written a udev rule in my life so you may need to
do some research and testing of that yourself.

Hope this is of at least some slight help and/or interest.


More information about the Linux-audio-dev mailing list