From mboxrd@z Thu Jan 1 00:00:00 1970 Date: Sun, 16 Nov 2008 21:28:08 +0000 (GMT) From: Hugh Dickins Subject: Re: [PATCH] mm: evict streaming IO cache first In-Reply-To: <49208E9A.5080801@redhat.com> Message-ID: References: <20081115181748.3410.KOSAKI.MOTOHIRO@jp.fujitsu.com> <20081115210039.537f59f5.akpm@linux-foundation.org> <49208E9A.5080801@redhat.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org Return-Path: To: Rik van Riel Cc: Linus Torvalds , Andrew Morton , KOSAKI Motohiro , LKML , linux-mm , Gene Heskett List-ID: On Sun, 16 Nov 2008, Rik van Riel wrote: > > I wonder if the "do not do mark_page_accessed at page fault time" > patch is triggering the current troublesome behaviour in the VM, > because actively used file pages are not moved out of the way of > the VM - which leads get_scan_ratio to believe that we are already > hitting the working set on the file side and should also start > scanning the anon LRUs. That patch is only in the -mm tree, not in 2.6.28-rc. Hugh -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: email@kvack.org