From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail144.messagelabs.com (mail144.messagelabs.com [216.82.254.51]) by kanga.kvack.org (Postfix) with ESMTP id D6AEE6B0087 for ; Thu, 2 Dec 2010 10:42:56 -0500 (EST) Date: Thu, 2 Dec 2010 15:42:35 +0000 From: Mel Gorman Subject: Re: [patch]vmscan: make kswapd use a correct order Message-ID: <20101202154235.GY13268@csn.ul.ie> References: <1291172911.12777.58.camel@sli10-conroe> <20101201132730.ABC2.A69D9226@jp.fujitsu.com> <20101201155854.GA3372@barrios-desktop> <20101202101234.GR13268@csn.ul.ie> <20101202153526.GB1735@barrios-desktop> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-15 Content-Disposition: inline In-Reply-To: <20101202153526.GB1735@barrios-desktop> Sender: owner-linux-mm@kvack.org To: Minchan Kim Cc: KOSAKI Motohiro , Shaohua Li , linux-mm , Andrew Morton List-ID: On Fri, Dec 03, 2010 at 12:35:26AM +0900, Minchan Kim wrote: > > > @@ -2550,8 +2558,13 @@ static int kswapd(void *p) > > > */ > > > order = new_order; > > > } else { > > > - kswapd_try_to_sleep(pgdat, order); > > > - order = pgdat->kswapd_max_order; > > > + /* > > > + * If we wake up after enough sleeping, let's > > > + * start new order. Otherwise, it was a premature > > > + * sleep so we keep going on. > > > + */ > > > + if (kswapd_try_to_sleep(pgdat, order)) > > > + order = pgdat->kswapd_max_order; > > > > Ok, we lose the old order if we slept enough. That is fine because if we > > slept enough it implies that reclaiming at that order was no longer > > necessary. > > > > This needs a repost with a full changelog explaining why order has to be > > preserved if kswapd fails to go to sleep. There will be merge difficulties > > with the series aimed at fixing Simon's problem but it's unavoidable. > > Rebasing on top of my series isn't an option as I'm still patching > > against mainline until that issue is resolved. > > So what's your point? Only point was to comment "I think this part of the patch is fine". > Do you want me to send this patch alone > regardless of your series for Simon's problem? > Yes, because I do not believe the problems are directly related. When/if I get something working with Simon, I'll backport your patch on top of it for testing by him just in case but I don't think it'll affect him. -- Mel Gorman Part-time Phd Student Linux Technology Center University of Limerick IBM Dublin Software Lab -- 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/ . Fight unfair telecom policy in Canada: sign http://dissolvethecrtc.ca/ Don't email: email@kvack.org