'Twas brillig, and David Henningsson at 10/01/10 18:35 did gyre and gimble: > David Henningsson wrote: >>> 'Twas brillig, and David Henningsson at 09/01/10 09:00 did gyre and >>> gimble: >>>> The pulse ALSA plugin has been known, for a while, to not work properly, >>>> causing underruns, hangs etc. I sat down yesterday trying to figure it >>>> out, and I'm pretty certain this patch improves the situation, but I >>>> don't mind getting some help testing it before it is committed upstream. >>>> >>>> Related bug: https://bugs.launchpad.net/bugs/485488 >>>> >>>> The patch is for the alsa-plugins tree at git.alsa-projects.org. >>> I applied this to 1.0.22 (cleanly) and sadly it's not helping my test >>> case using mpg123. >> >> I've tried to debug the mpg123 issue a little further. > > To sum up, I have two solutions/workarounds which both (independently) > fixes the mpg123 issue, but I could use some discussion regarding we > should apply either or both. And as said earlier, this is something > different from the patch posted earlier. > > First, we can avoid the initial underrun altogether by avoiding calling > pa_stream_cork() and pa_stream_trigger() after having written just a few > bytes. If we do that, playback will start when prebuf number of bytes > have been written instead. > > Second, we can ignore underruns if more buffers have been pushed since > the underrun occurred. Assuming the first fix is not applied, this will > cause an initial - to me inaudible - underrun, but since more buffers > are pushed in queue, the underrun will end automatically and after that > the stream will hopefully flow the way it should. > > Both fixes can, at least in theory, cause regressions. (And according to > both Murphy's law and own experience - if it breaks in theory, it > probably breaks in practice as well!) The first fix could cause a > regression for short event sounds where we never reach the prebuf limit. > The second could perhaps cause regressions for ALSA clients > automatically increasing buffer size if they find an underrun. Hiding > that underrun would then be a bad thing. Awesome work, thanks for looking into this in such depth. Lennart is the best person to comment on this as I'm nowhere near qualified. Hopefully he'll do a purge of his ML queue soon and provide some feedback :) Col -- Colin Guthrie gmane(at)colin.guthr.ie http://colin.guthr.ie/ Day Job: Tribalogic Limited [http://www.tribalogic.net/] Open Source: Mandriva Linux Contributor [http://www.mandriva.com/] PulseAudio Hacker [http://www.pulseaudio.org/] Trac Hacker [http://trac.edgewall.org/]
