[LAT] 2.6.29-rt1

Robin Gareus robin at gareus.org
Thu Mar 26 16:30:40 EDT 2009

Hash: SHA1

hollunder at gmx.at wrote:
> On Thu, 26 Mar 2009 19:52:40 +0100
> Robin Gareus <robin at gareus.org> wrote:
>> Hash: SHA1
>> We've got a winner! 2.6.29-rt1
>> I've just succeeded with a few tests: jackd realtime (64fpp with a USB
>> UA25) recording/playing Audio and MIDI with qtractor & qsynth and I
>> played a 32 channel ardour session - no xrun in sight!
>> I know this is old news since 64studio already ships a 2.6.29
>> pre-release - However, I run a debian squeeze/sid and
>> 2.6.29-release-candidates were causing an OOPS on this system.
>> As it turned out it was caused by "irqbalance"! It took me quite a
>> while to track down that the rt kernel is no longer compatible with
>> irqbalance (it works fine with 2.6.24-rt).
>> Here's a debian package built for core2:
>> http://rg42.org/_media/wiki/kernel/linux-image-2.6.29-rt1_2.6.29-rt1-10.00.custom_i386.deb
>> and the corresponding .config file:
>> http://rg42.org/_media/wiki/kernel/config-2.6.29-rt1.txt
>> kudos to the rt-linux devs.
>> have fun,
>> robin
> I compiled 2.6.29-rc8-rt4 yesterday and 2.6.29-rt1 today (my first self
> compiled kernels). 

Congrats. welcome to the club ;)

> I happen to have the same interface and it seems to
> work well here so far, everything seems to work and misbehaving
> applications still produce xruns.

How do you define "misbehaving" ?!

I managed to crash ardour3 and that got me an xrun - but I've got no
complaints there ;)

> Thanks for the warning about irqbalance, I was about to try it..
> (Is this the script by fernando lopez leczano or something else?)

It is not Rui's rtirq script. You should definitely get that:

"irqbalance" is a different beast. By default all IRQs are handled by
the first CPU only. If you have a SMP system irqbalance can distribute
interrupt handling to all available CPUs.

Version: GnuPG v1.4.9 (GNU/Linux)


More information about the Linux-audio-tuning mailing list