24 Mar
2005
24 Mar
'05
2:49 p.m.
On Thu, 2005-03-24 at 06:17 -0600, Ryan Gallagher wrote:
On Thu, 2005-03-24 at 10:17 +0100, Clemens Ladisch
wrote:
It means JACK was delayed in waking up for more than half the period
time. For a PCI card this means the processing of the audio interrupt
was delayed for a long time.
With PCI cards this is a highly unusual and pathological situation. It
could be caused by a broken config like having the sound card IRQ
threaded but with a low RT priority. For a PCI card you will very
rately run into this situation.
I think the problem could be in the kernel or the ALSA driver. The
problem is related to the fact that it's a USB device, but it should be
possible to get this working (it works on other OS).
You are running at a pretty low latency (1.3ms or 64 frames at 48KHz).
But, since it looks like the wakeup delay is 16ms, as you probably know
you will have big problems running at any reasonable latency setting.
ISTR someone tried this config with Ingo's RT preempt kernel, and the
problem was still present? Maybe the latency tracing facility would
help.
Is the USB irq shared with anything?
Anyway, it's hard for me to diagnose the problem more without having the
hardware.
Lee
Steve Fosdick wrote:
Clemens (or any other informed reader),
Can you give me an idea of what these;
delay of 16535.000 usecs exceeds estimated spare time of 1307.000;
restart ...
Messages mean? Where do they originate? Are there any specific steps I
should take in getting rid of them? Is it somehow related to the fact
that this is a USB device? I am right in thinking the UA-25 used only
standard USB Audio
drivers?
Yes.
Do people find these keyboard lockups occur with
a variety of USB
sound cards but not with other USB devices?
USB Audio devices are somewhat different because they use isochronous
transfers. Video capture devices use those, too.