Re: [PATCH] Scalable page cache

Benjamin LaHaise (bcrl@redhat.com)
Mon, 26 Nov 2001 13:34:50 -0500


On Mon, Nov 26, 2001 at 09:29:39PM +0100, Ingo Molnar wrote:
> this is a misunderstanding of the problem. The reason why the
> pagecache_lock is a performance problem is *not* contention, the reason is
> *not* the length of chains, or any other reason you listed. The problem is
> SMP cacheline invalidation costs, due to using the same cacheline from
> multiple CPUs. Thus the spreading out of locking gives good SMP cacheline
> usage properties.

Please reply to the rest of the message. Perhaps that item is a
misunderstanding, but the rest of it is applicable across the board.

-ben
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/