From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965241AbVHJR7O (ORCPT ); Wed, 10 Aug 2005 13:59:14 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S965242AbVHJR7O (ORCPT ); Wed, 10 Aug 2005 13:59:14 -0400 Received: from holly.csn.ul.ie ([136.201.105.4]:34210 "EHLO holly.csn.ul.ie") by vger.kernel.org with ESMTP id S965241AbVHJR7N (ORCPT ); Wed, 10 Aug 2005 13:59:13 -0400 Date: Wed, 10 Aug 2005 18:59:12 +0100 (IST) From: Mel Gorman X-X-Sender: mel@skynet To: Dave Hansen Cc: Andrew Morton , Linux Kernel Mailing List Subject: Re: How to reclaim inode pages on demand In-Reply-To: <1123696225.15970.10.camel@localhost> Message-ID: References: <20050808160844.04d1f7ac.akpm@osdl.org> <20050810101714.147e1333.akpm@osdl.org> <1123696225.15970.10.camel@localhost> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 10 Aug 2005, Dave Hansen wrote: > On Wed, 2005-08-10 at 18:27 +0100, Mel Gorman wrote: > > I later linearly scan the mem_map looking for pages that can be freed up > > (usually LRU pages). I was expecting any page with PG_inode set to have a > > page->mapping but not all of them do. It is the pages without a ->mapping > > that are confusing the hell out of me. > > How about putting a check for PG_inode and a periodic dump_stack() > wherever page->mapping is cleared? That should at least let you catch > the culprit who is clearing them. __remove_from_page_cache() is the only > real place I see this being done. > Will do, it might shed some light on whats happening. > Are you remembering to clean PG_inode when a page is freed? Yes, I was using the allocator to count how many inode-related pages there were in the system. The answer was lots, another reason why I'm thinking that I picked the worst possible benchmark to test this stuff with. > Perhaps > it's allocated for the page cache, reclaimed, returned to the allocator, > and reallocated as anonymous memory where you're seeing the null > ->mapping. Might want to check for PG_inode in free_pages_check(). > I am 99.99% certain I make that check but I can't double check the code until late in the weekend. -- Mel Gorman Part-time Phd Student Java Applications Developer University of Limerick IBM Dublin Software Lab