Re: 2.6.14-rc4-rt1

[Mark Knecht <markknecht@gmail.com>]

On 10/12/05, Lee Revell <rlrevell@joe-job.com> wrote:
> On Wed, 2005-10-12 at 09:37 -0700, Mark Knecht wrote:
> > On 10/11/05, Lee Revell <rlrevell@joe-job.com> wrote:
> > > On Tue, 2005-10-11 at 14:13 -0700, Mark Knecht wrote:
> > > > The machine had been essentially 'User space idle' for the previous
> > > > two hours. The screen saver had kicked in. Audio was running and the
> > > > machine was busy. I woke it up, gave xscreensaver my password, read
> > > > email, sent the previous mail, then picked up the telephone to make a
> > > > call. Not 2 seconds later the xruns occurred!
> > >
> > > So what does /proc/latency_trace report?
> > >
> > > Lee
> >
> > Well, unfortunately it doesn't appear to report anythign helpful. The
> > maximum latency report did not change when the xrun occurred. This was
> > the last one reported and it happened long before the xrun:
>
> Sounds like an application bug (some JACK client doing something not RT
> safe).  Can you reproduce the xruns if you just run jackd with no
> clients?

I don't know. These xruns take hours to generate. I'd probably have to
dedicate a whole day of doing nothing on the machine to try, and then
if I didn't produce anything I'm not sure what it proves. If I do get
one then we get to see if there's data.

I'd really like to do the IRQ-off tests that you do before I go that
direction but unfortunately it's broken very badly since yesterday's
-rt1 release.

Maybe a better path to take would just be pushing the machine harder.
Do some real work in Ardour. Build up a big session and let it rip for
a while and see what that produces?

I don't know really.

- Mark
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/