Re: Overruns are killing my recordings.

From: Mark Knecht
Date: Fri Oct 28 2005 - 13:55:33 EST


On 10/28/05, Lee Revell <rlrevell@xxxxxxxxxxx> wrote:
> On Fri, 2005-10-28 at 07:52 -0700, Mark Knecht wrote:
> > On 10/27/05, Lee Revell <rlrevell@xxxxxxxxxxx> wrote:
> > > On Thu, 2005-10-27 at 17:00 -0700, Avuton Olrich wrote:
> > > > aggh. Sorry for all the noise,
> > > >
> > > > I have all my drives on a linear raid and I had hdparm set to put my
> > > > IDE drives to sleep after a while, I didn't put it together because it
> > > > was happening in the middle of recording.
> > >
> > > Hey, I think it's a testament to the progress that has been made in the
> > > past year and a half that people now consider audio dropouts in a "known
> > > good" app like ecasound to be a kernel bug. For the longest time the
> > > answer was "linux isn't an RTOS, deal with it".
> > >
> > > Lee
> >
> > Lee, et. all,
> > Could this possibly be part of what is causing my xrun problems? I
> > had a huge rash of xruns yesterday. I seem to run into issues after
> > longer times of inactivity. I hadn't considered this possibility
> > before.
>
> I really doubt it. It's more likely that the xruns are caused by a bug
> in the new ktimers system. I am seeing "xruns" here too with -rt1, but
> the latency tracer does not report anything over ~120 usecs. Previous
> to all the ktimers/HRT stuff going in, I was xrun free for months.
>
> The reason I think it's a ktimers bug is because sometimes JACK reports
> an xrun of negative length which I'd NEVER seen before.
>
> I suspect this might all be fixed in the latest -rt patch but I have not
> had time to build it.
>
> Lee

OK then I'll just hang tinght. I've not seen any response on the email
I sent about 2.6.14-rc5-rt7. I cannot build it. It fails like this:

CC arch/x86_64/kernel/sys_x86_64.o
CC arch/x86_64/kernel/x8664_ksyms.o
CC arch/x86_64/kernel/i387.o
CC arch/x86_64/kernel/syscall.o
CC arch/x86_64/kernel/vsyscall.o
arch/x86_64/kernel/vsyscall.c:57: error: `SEQLOCK_UNLOCKED' undeclared
here (not in a function)
make[1]: *** [arch/x86_64/kernel/vsyscall.o] Error 1
make: *** [arch/x86_64/kernel] Error 2
lightning linux #

This is a new failure here since -rc5-rt3.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/