[PlanetCCRMA] Configuring 2.6.11-0.3.rdt.rhfc2.ccrma for 0 xruns...

Rui Nuno Capela rncbc@rncbc.org
Thu Apr 14 01:16:01 2005

>> As I say, I think it was Ingo who
>> told me to try it back when I was having trouble with ACPI and USB
>> mice in October, 04.
>> Anyway, perfect results at 128/2.
> Good, maybe this should be something that rtirq should do by default.
> Rui: are you reading this thread? Maybe rtirq should set the snd
> interrupts to not threaded if it finds the proc controls for that.

Damn. I'm running PREEMPT_RT (full realtime) for so long that I forgot
that one could unthread the IRQ handlers with some other realtime-preempt
kernel configurations. With PREEMPT_RT all IRQs are threaded, period.

And, believe me, PREEMPT_RT its the top performer configuration I ever saw
for audio work. That's why I never thought of going to other lesser
realtime-preempt configurations, once I got that stable on my laptop (UP)
and desktop machine (SMP/HT).

Correct me if I'm wrong, the planet-edge kernels are configured as

OK. I'll check if I can remember how to unthread the snd related IRQs. For
example, on my PREEMPT_RT machines I have nothing found as
"/proc/irq/*/*/threaded" . May I assume that if "/proc/irq/*/*/threaded"
is found we are under a PREEMPT_DESKTOP kernel, and then we can do the
  for x in "/proc/irq/$IRQ/*/threaded"; do echo 0 > "$x"; done

>> I'll be trying 64/2 in a few minutes.
>> This can be set in rc.local, etc. and works every time I boot up on my
>> laptop.
> The threaded stuff? Yes, I guess it could go there. It would be best to
> incorporate it into rtirq, of course.

I think I'll blind check this and introduce a new rtirq configuration
switch as an option whether "snd" IRQs should be threaded or not. Default
should be threaded, I guess.

rncbc aka Rui Nuno Capela