Re: More weird latency trace output (was Re: 2.6.17-rt1)

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

On Sun, 2006-06-25 at 00:12 +0200, Ingo Molnar wrote:
> * Lee Revell <rlrevell@joe-job.com> wrote:
> 
> The latency tracer uses get_cycles() for
> > > timestamping, which uses rdtsc, which is unusable for timing on dual
> > > core AMD64 machines

> does it get better if you boot with idle=poll? [that could work around 
> the rdtsc drifting problem] Calling gettimeofday() from within the 
> tracer is close to impossible - way too many opportunities for 
> recursion. It's also pretty slow that way.

That seemed to solve the drifting problem, but I still get some weird
behavior.  The max reported trace in dmesg still does not
match /proc/latency_trace:

( posix_cpu_timer-3    |#0): new 38 us maximum-latency wakeup.


preemption latency trace v1.1.5 on 2.6.17-rt1-smp-latency-trace
--------------------------------------------------------------------
 latency: 19 us, #74/74, CPU#0 | (M:rt VP:0, KP:0, SP:1 HP:1 #P:2)
    -----------------
    | task: posix_cpu_timer-3 (uid:0 nice:0 policy:1 rt_prio:99)
    -----------------

Lee