>It's really discarded as spurious? All voices has separate spurious
>interrupt checking based on the global frame counter (stimer). Please, can
>you verify it? I still think, that it would be better to call
>snd_trident_capture_pointer() and see the actual ptr. My patch should
>fix the problem, if ptr is less than period_size.
thats not the central issue, it turns out.
what seems to happen very often is something like:
@frame N: interrupt for playback+capture
@frame N+n: interrupt for capture
where `n' is a small value. sometimes, we should be paying attention
to the interrupt a time N, sometimes the one at N+n.
changing the ESO to be 2 frames larger definitely helped, as did
keeping separate playback and capture stimers for the voice, but i
still don't have it working properly yet. the code i have works better
that the code in CVS, but there are still many xruns when running in
full duplex mode at 44.1kHz, 64 frame period, caused by the driver
failing to wake up a task at the correct time.
work continues. the more i look at this stuff, the more i understand
how clean and elegant the RME Hammerfall design really is :)
--p
_______________________________________________
Alsa-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/alsa-devel
