Re: [patch] latency tracer, 2.6.15-rc7

[Lee Revell <rlrevell@joe-job.com>]

On Fri, 2005-12-30 at 16:59 -0800, Mark Knecht wrote:
> On 12/30/05, Lee Revell <rlrevell@joe-job.com> wrote:
> > On Fri, 2005-12-30 at 09:00 +0100, Ingo Molnar wrote:
> > > * Lee Revell <rlrevell@joe-job.com> wrote:
> > >
> > > > It seems that debug_smp_processor_id is being called a lot, even
> > > > though I have a UP config, which I didn't see with the -rt kernel:
> > >
> > > do you have CONFIG_DEBUG_PREEMPT enabled? (if yes then disable it)
> > >
> > > > Was this optimized out on UP before?
> > >
> > > no, because smp_processor_id() debugging is useful on UP too: it checks
> > > whether smp_processor_id() is every called with preemption enabled, and
> > > reports such bugs.
> >
> > It seems that the networking code's use of RCU can cause 10ms+
> > latencies:
> >
> 
> 
> Hi,
>    I've noted for awhile that on my AMD64 machine that has xrun issues
> that at least annecdotally it has always seemed that the network
> interface was somehow involved. I wonder if this may turn out to be
> true?
> 
> Q: Did the latency trace stuff ever get fixed for AMD64? I'm running
> 2.6.15-rc7-rt1 as of today. I've had xruns all afternoon while working
> inside of MythTV. (Start, stop, rewind, pause all over the place...)

Yes, it should have.  What happens when you try it?

Lee