Messages in this thread | | | Date | Wed, 02 Dec 2009 00:08:18 -0500 | From | Rik van Riel <> | Subject | Re: [PATCH 2/9] ksm: let shared pages be swappable |
| |
On 12/01/2009 04:59 AM, Andrea Arcangeli wrote: > On Tue, Dec 01, 2009 at 06:46:06PM +0900, KOSAKI Motohiro wrote: > >> Ah, well. please wait a bit. I'm under reviewing Larry's patch. I don't >> dislike your idea. last mail only pointed out implementation thing. >> > Yep thanks for pointing it out. It's an implementation thing I don't > like. The VM should not ever touch ptes when there's light VM pressure > and plenty of unmapped clean cache available, but I'm ok if others > disagree and want to keep it that way. > The VM needs to touch a few (but only a few) PTEs in that situation, to make sure that anonymous pages get moved to the inactive anon list and get to a real chance at being referenced before we try to evict anonymous pages.
Without a small amount of pre-aging, we would end up essentially doing FIFO replacement of anonymous memory, which has been known to be disastrous to performance for over 40 years now.
A two-handed clock mechanism needs to put some distance between the front and the back hands of the clock.
Having said that - it may be beneficial to keep very heavily shared pages on the active list, without ever trying to scan the ptes associated with them.
| |