Re: [patch] Real-Time Preemption, -RT-2.6.12-rc6-V0.7.48-00

[Ingo Molnar]

* William Weston <weston@xxxxxxxxx> wrote:

> I also regularly see high (>200us) wakeup latencies on the Xeon/HT, 
> which I don't see on the Athlon or non-HT Xeon systems.  Disabling IRQ 
> balancing doesn't seem to help.  It's been a while since the Xeon/HT 
> box has seen a non-debug kernel, but in the past, that hasn't helped 
> latencies by more than a few usec.

>   <idle>-0     0Dnh2    4us : find_next_bit (__schedule)
>   <idle>-0     0Dnh2    4us : _raw_spin_lock (__schedule)
>   <idle>-0     0Dnh3    4us!: find_next_bit (__schedule)
>   <idle>-0     0Dnh3  244us : find_next_bit (__schedule)
>   <idle>-0     0Dnh3  245us : _raw_spin_unlock (__schedule)

this does seem to be similar to other reports where the cause of such 
latencies was some sort of hardware-level latency. (DMA related delays 
or other, bus-arbitration related delays)

another possibility is that something interesting happened on another 
CPU while this latency occured - to debug this please enable all-CPUs 
tracing:

	echo 1 > /proc/sys/kernel/trace_all_cpus

and send me a new trace.

	Ingo
-
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/