Re: [patch] voluntary-preempt-2.6.9-rc1-bk4-R1

[Ingo Molnar]

* Mark_H_Johnson@xxxxxxxxxxxx <Mark_H_Johnson@xxxxxxxxxxxx> wrote:

> 00000001 0.000ms (+0.000ms): spin_lock (get_swap_page)
> 00000001 0.000ms (+0.000ms): spin_lock (<00000000>)
> 00000001 0.000ms (+0.000ms): spin_lock (get_swap_page)
> 00000002 0.000ms (+0.113ms): spin_lock (<00000000>)
> 00010002 0.113ms (+0.000ms): do_nmi (get_swap_page)

> Are you SURE the spin lock counter works properly on SMP systems?
> I did a quick check of yesterday's results:
>   # grep -ir '<.*>' latencytest0.42-png/lt040909  | wc -l
>   6978
>   # grep -ir '<.*>' latencytest0.42-png/lt040909  | grep -v '<00000000>' |
> less -im
>   ...

> No entries that are non zero and lock related.

it works fine here. To double-check i've created a contention testcase:

 00000001 0.000ms (+0.000ms): spin_lock (sys_gettimeofday)
 00000001 0.000ms (+0.000ms): spin_lock (<000012ce>)

this spin_lock() spun 4814 times before it got the lock.

Linux locking is pretty uncontended on 2-way boxes.

	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/