From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S263151AbUB0WY2 (ORCPT ); Fri, 27 Feb 2004 17:24:28 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S263172AbUB0WY2 (ORCPT ); Fri, 27 Feb 2004 17:24:28 -0500 Received: from 64-186-161-006.cyclades.com ([64.186.161.6]:11193 "EHLO intra.cyclades.com") by vger.kernel.org with ESMTP id S263151AbUB0WYU (ORCPT ); Fri, 27 Feb 2004 17:24:20 -0500 Date: Fri, 27 Feb 2004 20:18:29 -0300 (BRT) From: Marcelo Tosatti X-X-Sender: marcelo@logos.cnet To: Andrew Morton Cc: Rik van Riel , andrea@suse.de, linux-kernel@vger.kernel.org Subject: Re: 2.4.23aa2 (bugfixes and important VM improvements for the high end) In-Reply-To: <20040227122936.4c1be1fd.akpm@osdl.org> Message-ID: References: <20040227173250.GC8834@dualathlon.random> <20040227122936.4c1be1fd.akpm@osdl.org> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Cyclades-MailScanner-Information: Please contact the ISP for more information X-Cyclades-MailScanner: Found to be clean Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 27 Feb 2004, Andrew Morton wrote: > Oh, and can we please have testcases? It's all very well to assert "it > sucks doing X and I fixed it" but it's a lot more useful if one can > distrubute testcases as well so others can evaluate the fix and can explore > alternative solutions. > > Andrea, this shmem problem is a case in point, please. > > > > in small machines the current 2.4 stock algo works just fine too, it's > > > only when the lru has the million pages queued that without my new vm > > > algo you'll do million swapouts before freeing the memleak^Wcache. > > > > Same for Arjan's O(1) VM. For machines in the single and low > > double digit number of gigabytes of memory either would work > > similarly well ... > > Case in point. We went round the O(1) page reclaim loop a year ago and I > was never able to obtain a testcase which demonstrated the problem on 2.4, > let alone on 2.6. > > I had previously found some workloads in which the 2.4 VM collapsed for > similar reasons and those were fixed with the rotate_reclaimable_page() > logic. Without testcases we will not be able to verify that anything else > needs doing. Btw, Andrew, are your testcases online somewhere? I heard once someone was going to collect VM tests to make a "official testing package", but that has never happened AFAIK.