Re: Locking in disk(9) api

[Masao Uebayashi <uebayasi%tombi.co.jp@localhost>]

> I think this is almost the canonical example of data we should not lock.
> The cost of using locking *or* atomic operations for statistics update
> will be extreme on busy MP systems, and there's almost no real benefit.

My understanding is that frequent & important counters should be collected
by per-cpu basis as sys/net* statistics does.  Rare & important ones should
fit better for atomic ops.  I don't think less important ones need atomicity.

Masao

-- 
Masao Uebayashi / Tombi Inc. / Tel: +81-90-9141-4635