[linux-audio-dev] Re: [Jackit-devel] Maybe I'm going at this wrong...?

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: [linux-audio-dev] Re: [Jackit-devel] Maybe I'm going at this wrong...?
From: Lee Revell (rlrevell_AT_joe-job.com)
Date: Thu Sep 09 2004 - 07:19:32 EEST


On Wed, 2004-09-08 at 23:53, Eric Dantan Rzewnicki wrote:
> Lee Revell wrote:
> > On Wed, 2004-09-08 at 17:38, Lee Revell wrote:
> >>If you want to try the new 2.6 low latency kernel, here is a QuickStart:
> >>Start with:
> >> http://kernel.org/pub/linux/kernel/v2.6/linux-2.6.8.tar.bz2
> >>Then apply these patches in this order:
> >> http://kernel.org/pub/linux/kernel/v2.6/testing/patch-2.6.9-rc1.bz2
> >> http://redhat.com/~mingo/voluntary-preempt/patch-2.6.9-rc1-bk12.bz2
> >> http://redhat.com/~mingo/voluntary-preempt/voluntary-preempt-2.6.9-rc1-bk12-R9
> >>This will give you the irq threading settings in /proc that other posters referred to.
> >>Please keep in mind that this is not completely stable yet, I would not
> >>recommend it for production use, but it works very well for many
> >>people. You are likely to have better results on a regular 32 bit
> >>single processor (that means no hyperthreading) x86 desktop system,
> >>because this has been well tested.
> > Ugh, typo, the last patch should be 'R6' and not 'R9'.
>
> Well, actually there is an R9 out today ...
>

Yeah, I know, but R9 had a bug that caused it to not compile for me; R6
was the last version that worked. Maybe you are just not hitting that
bug...

> I have it patched, compiled and installed.
>
> audiobox:~# uname -r
> 2.6.9-rc1-bk12-vp-r9
>
> (had to lowercase the extraversion stuff to comply with debian policy
> when making .deb with make-kpkg)
>

OK, great! I have been meaning to make .debs of these but have not had
time.

> I've set all the /proc and /sys stuff I could find in the wiki (console
> output below). Still, I get xruns reported by my ecasound 8 analogueOsc
> thingy. Interestingly, though, they now only happen when I start the
> second ecasound process to record from the jack outputs of the first.
> I'll double check that with previous kernel versions when I get a
> chance, but I'm fairly sure this is new and an improvement.
>
> Subjectively, the sound of the analogueOscs is smoother and the zipper
> noise from controller movements seems much less, but still the first
> ecasound reports xruns when it exits.
>
> As I said earlier today, I'm not complaining. I'm just trying to help
> out by reporting. I'm eager and willing to help with debugging. Is there
> anything I may have missed?
>
> What other info should I provide?
>

All your settings look good. I think you are in good shape. xruns are
normal when connecting and disconnecting a jack client, this does not
seem to hurt anything, and they should not happen once you are rolling.

If you do get any xruns, here is how to debug them. Reset the
preemption latency counter like this:

echo 0 > /proc/sys/kernel/preempt_max_latency

Then start jackd. When you see an xrun, copy the contents of
/proc/latency_trace to a file, then post them here.

Also, are you running debian unstable? What output do you get when you
run 'getconf GNU_LIBPTHREAD_VERSION'? There is a known bug affecting
the version of glibc that debian unstable comes with.

Lee


New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Thu Sep 09 2004 - 07:24:52 EEST