Re: 2.6.14-rc3-rt9 - a few xruns misses

[Mark Knecht <markknecht@gmail.com>]

On 10/6/05, Mark Knecht <markknecht@gmail.com> wrote:
> On 10/6/05, Lee Revell <rlrevell@joe-job.com> wrote:
> > On Thu, 2005-10-06 at 09:00 -0700, Mark Knecht wrote:
> > > Even with Jack running I don't see the jackd process getting any
> > > special priority. Is this correct, or is that part that gets higher
> > > prioity just not listed here.
> >
> > ps does not show all threads of multithreaded processes by default.
> > Use:
> >
> > ps -Leo pid,pri,rtprio,cmd
> >
> > and you should see that 2 JACK threads get RT priority.
> >
> > Lee
>
> Thanks Lee. That's what I thought might be happening.
>
> 8398  24      - hdspmixer
>  8400  24      - qjackctl
>  8400  49      9 qjackctl
>  8402  20      - /usr/bin/jackd -R -dalsa -dhw:1 -r44100 -p128 -n2
>  8402  20      - /usr/bin/jackd -R -dalsa -dhw:1 -r44100 -p128 -n2
>  8402  23      - /usr/bin/jackd -R -dalsa -dhw:1 -r44100 -p128 -n2
>  8402  60     20 /usr/bin/jackd -R -dalsa -dhw:1 -r44100 -p128 -n2
>  8402  50     10 /usr/bin/jackd -R -dalsa -dhw:1 -r44100 -p128 -n2
>
> Cheers,
> Mark
>

Hi Ingo,
   OK, after about an hour or moderately heavy usage while running
Jack - I've built a new new kernel, done an emerge sync and am in the
process of an emerge world right now (it's building glibc) along with
playing music and browsing the web - I just had my first xrun:

configuring for 44100Hz, period = 128 frames, buffer = 2 periods
nperiods = 2 for capture
nperiods = 2 for playback
08:51:42.919 Server configuration saved to "/home/mark/.jackdrc".
08:51:42.919 Statistics reset.
08:51:43.037 Client activated.
08:51:43.038 Audio connection change.
08:51:43.042 Audio connection graph change.
09:11:52.063 Audio connection graph change.
09:11:52.122 Audio connection change.
09:11:54.217 Audio connection graph change.
subgraph starting at qjackctl-8400 timed out (subgraph_wait_fd=17,
status = 0, state = Finished)
09:54:17.656 XRUN callback (1).
**** alsa_pcm: xrun of at least 4.351 msecs

So....how do I begin to find out what caused that, or will cause
similar xruns in the future?

Is it that rt priorities are not set up correctly? Or is it something else?

I must say that the kernel has really improved for me over the last
couple of weeks. This is great performance from where I was before I
joined the LKML. I think you guys are making good progress, most
especially for my AMD64 platform.

Thanks!

Cheers,
Mark