Re: 2.6.38 page_test regression

From: Andrea Arcangeli
Date: Thu Apr 14 2011 - 19:39:42 EST

On Fri, Apr 15, 2011 at 12:16:48AM +0100, Mel Gorman wrote:
> Theories better than slightly-different-register-use are welcome.

__pte_alloc_one is now skipped 511 times every 512 page faults, so we
call alloc_pages 512 times less and we take page_table_lock 512 times
less, that explains it I think. We need it...
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at