On Thu, 2007-12-13 at 22:54 +0100, Peter Zijlstra wrote:
> On Thu, 2007-12-13 at 11:30 +0800, zhejiang wrote:
> > __percpu_counter_add() cache the result in percpu variable until it
> > exceeds the batch.
> > The prop_norm_percpu() use the percpu_counter_read(&pl->eve
__percpu_counter_add() cache the result in percpu variable until it
exceeds the batch.
The prop_norm_percpu() use the percpu_counter_read(&pl->events) to read
the counter ,and use percpu_counter_add(&pl->events, -half) to half the
counter.
There are potential problems:
1.The counter may be negativ
The patch 04fbfdc14e5f48463820d6b9807daa5e9c92c51f implemented bdi per
device dirty threshold. It works well.
However, the period for proportion calculation may be too large.
For 8G memory, the calc_period_shift() will return 19 as the shift.
When we switch writing operation between different disk
3 matches
Mail list logo