[LAU] alsa-jack plugin... some patches.

[torbenh]

On Fri, Oct 29, 2010 at 02:42:48PM +0200, Robin Gareus wrote:
> On 10/29/10 13:07, Paul Davis wrote:
> > On Thu, Oct 28, 2010 at 11:08 PM, Robin Gareus <robin at gareus.org> wrote:
> >> On 10/29/10 05:00, Paul Davis wrote:
> >>> On Thu, Oct 28, 2010 at 10:47 PM, Robin Gareus <robin at gareus.org> wrote:
> >>>
> >>>> ..but I could not get your .asoundrc working. It fails with
> >>>>
> >>>> [AO_ALSA] alsa-lib: pcm_hw.c:1293:(snd_pcm_hw_open) open
> >>>> '/dev/snd/pcmC0D0p' failed (-16): Device or resource busy
> >>>
> >>> you're not opening the JACK device, clearly - this message references
> >>> /dev/snd/pcmC0D0p
> >>
> >> I hazard a guess that it's just mplayer with a misleading error message.
> > 
> > i don't think so. i think its you not telling mplayer to use the
> > correct ALSA device.
> > 
> >> I prefixed torben's config with
> >>
> >> pcm.!default {
> >>  type plug
> >>  slave { pcm "jack" }
> >> }
> > 
> > which confirms my guess since this now makes the default ALSA device
> > be the JACK one, and thus apps which open "default" (which they should
> > if no other device name is given by the user) will get the JACK one.
> 
> Mmh I don't follow. You're say it should work, no?
> 
> The problem here is the parsing of the inline 'pcm { .. }' with
> @args and $NAME.. If i remove this part or simply do a separate
> definition of pcm.jackraw instead of using it inline in pcm.jack it works.
> 
> But then the jack-ports are just named "alsa-jack_P*" or "(null)_P*" and
> I don't see the client name(s).
> 
> From Torben's 'git log 'I see that he only changed jack/pcm_jack.c and
> it's 1.0.23 otherwise. Still, the issue could be related that I did not
> install all of alsa-plugins but only libasound_module_pcm_jack.so ?!

no. i just copied over the jack plugin. and i am on squeeze too.
this is not the issue.

please post your complete .asoundrc  
i think your default entry is just a bit wrong.


> 
> Anyway the new jack-plug works (tested with skype, firefox/flash &
> mplayer) for half a day.
> 
> For day-to-day usage I moved away from jack-plug a while ago and am
> using snd_aloop with alsa_in/out - it's a bit more heavy on CPU usage
> but much more robust.
> 
> One remaining issue with jack-plug that I don't have with snd_aloop is
> that some clients (fi. firefox) do not close the audio-device.
> If the client is still connected to alsa/jack-plug and I re-start jackd,
> sound becomes unavailable until the client (here: firefox) is also
> restarted.

i guess there is not much we can do.
what does pulse do under these circumstances ?

but since the goal is just to never stop jack, this issue will be
ignored ;D

> 
> ciao,
> robin
> 

-- 
torben Hohn