From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S268013AbUIKBn1 (ORCPT ); Fri, 10 Sep 2004 21:43:27 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S268021AbUIKBn1 (ORCPT ); Fri, 10 Sep 2004 21:43:27 -0400 Received: from smtp204.mail.sc5.yahoo.com ([216.136.130.127]:22136 "HELO smtp204.mail.sc5.yahoo.com") by vger.kernel.org with SMTP id S268013AbUIKBnZ (ORCPT ); Fri, 10 Sep 2004 21:43:25 -0400 Message-ID: <41424E71.3050107@yahoo.com.au> Date: Sat, 11 Sep 2004 11:01:37 +1000 From: Nick Piggin User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.2) Gecko/20040810 Debian/1.7.2-2 X-Accept-Language: en MIME-Version: 1.0 To: Hugh Dickins CC: Nikita Danilov , Linux Kernel Mailing List Subject: Re: 2.6.9-rc1: page_referenced_one() CPU consumption References: In-Reply-To: Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hugh Dickins wrote: > On Fri, 10 Sep 2004, Hugh Dickins wrote: > >>I'm quite content to go back to a trylock in page_referenced_one - and >>in try_to_unmap_one? But yours is the first report of an issue there, >>so I'm inclined to wait for more reports (which should come flooding in >>now you mention it!), and input from those with a better grasp than I >>of how vmscan pans out in practice (Andrew, Nick, Con spring to mind). > > > Just want to add, that there'd be little point in changing that back > to a trylock, if vmscan ends up cycling hopelessly around a larger > loop - though if the larger loop is more preemptible, that's a plus. > Yeah - I'm not sure why a trylock would perform better. If it is just one big address space, and memory needs to be freed, presumably the scanner will just choose a different page, and try the lock again. Feel like doing a few more quick tests Nikita? ;)