Guenter Geiger discourseth:

There are a few guesses here. Possibly the input is not correctly started, and the program hangs in a read in kernel code. You could put debug before and after the read and write, and look if it really happens there.

Unfortunately it hangs so hard that even an strace doesn't work.

...

I doubt that's the problem in my case. Coincidentally, I've done a lot of work with buffer configuration for my audio driver. :)

Ahh, very good, still I suspect an audio driver problem.

I suspect you're right! The -dac option works for me. I hacked around the soundcard initialization code a bit to see if the problem was there (*many* unclean reboots :) but with no luck. I think my driver just has a full duplex problem.

...

It's a PCI128 soundcard. My kernel is the latest update from redhat: 2.0.36-3. I *did* modify the sounddriver to fix a bug in the mixer map.

Yes ? I always wanted to do that correctly, there is a separate mixer ioctl for turning on and of the "monitor feature". Did you implement that ? My version only uses a "fast hack" to get things done the way I want. We may come up with a clean version and submit the patch to the author.

If you're talking about the separate output channel, a ioctl() for that is in the latest version of the driver. :)

Mhmm, I never tried the PCI128 with 2.0.36, only with recent kernels. You might try it with 2.2.x.

Indeed..the latest drivers seem to require 2.2.x. I expect I'll make the move to that soon.

Here a short quote from a previous thread about PCI128 (just to give a little bit of hope)

[...]

(you probably have read all of this already)

Uh..well, I should have. :) This certainly encourages me to try the latest kernels and drivers.

Many Thanks,

Eric