dsdreamer wrote:
>
> EDIT: I finally just copied the entire code of the static void
> *output_thread(void *arg) {} function from Revision: c0d549245689 of the
> output branch and used it to replace the equivalent code in Revision:
> de34f5cf5a5d, and recompiled. Now I have the benefit of resampling by
> exception and the ability to switch between sampling rates without fear
> of absurd speed-up effects (even if we are concealing errors rather than
> properly fixing them).
>
> 2nd EDIT: But this workaround only works for mmap=1. Just tried it with
> mmap=0 and it fails as might be expected. Of course I normally use
> mmap=1, which is the default so I don't have a practical problem but it
> is worth mentioning. Of course, if Clive is right and the F20 kernel
> fixes this properly, that would definitely be the best outcome.
To test the hypothesis of what is broken - what happens if you use the
new code, but duplicate the call to alsa_open so it is called two times
in a row. This will mean that the device is opened, closed and then
opened again whenever the sample rate changes. If it is getting a bad
feedback message then I think this would work around it for both mmap=0
and 1 cases.
------------------------------------------------------------------------
Triode's Profile: http://forums.slimdevices.com/member.php?userid=17
View this thread: http://forums.slimdevices.com/showthread.php?t=99395
_______________________________________________
unix mailing list
[email protected]
http://lists.slimdevices.com/mailman/listinfo/unix
