Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752789Ab0G3EzA (ORCPT ); Fri, 30 Jul 2010 00:55:00 -0400 Received: from fgwmail7.fujitsu.co.jp ([192.51.44.37]:56223 "EHLO fgwmail7.fujitsu.co.jp" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752195Ab0G3Ey6 (ORCPT ); Fri, 30 Jul 2010 00:54:58 -0400 X-SecurityPolicyCheck-FJ: OK by FujitsuOutboundMailChecker v1.3.1 From: KOSAKI Motohiro To: Mel Gorman Subject: Re: Why PAGEOUT_IO_SYNC stalls for a long time Cc: kosaki.motohiro@jp.fujitsu.com, Wu Fengguang , Andrew Morton , stable@kernel.org, Rik van Riel , Christoph Hellwig , "linux-kernel@vger.kernel.org" , "linux-fsdevel@vger.kernel.org" , "linux-mm@kvack.org" , Dave Chinner , Chris Mason , Nick Piggin , Johannes Weiner , KAMEZAWA Hiroyuki , Andrea Arcangeli , Minchan Kim , Andreas Mohr , Bill Davidsen , Ben Gamari In-Reply-To: <20100729142413.GB3571@csn.ul.ie> References: <20100729153719.4ABD.A69D9226@jp.fujitsu.com> <20100729142413.GB3571@csn.ul.ie> Message-Id: <20100730115222.4AD8.A69D9226@jp.fujitsu.com> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit X-Mailer: Becky! ver. 2.50.07 [ja] Date: Fri, 30 Jul 2010 13:54:53 +0900 (JST) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2844 Lines: 59 > > (1) and (8) might be solved > > by sleeping awhile, but it's unrelated on io-congestion. but might not be. It only works > > by lucky. So I don't like to depned on luck. > > In this case, waiting a while really in the right thing to do. It stalls > the caller, but it's a high-order allocation. The alternative is for it > to keep scanning which when under memory pressure could result in far > too many pages being evicted. How long to wait is a tricky one to answer > but I would recommend making this a low priority. For case (1), just lock_page() instead trylock is brilliant way than random sleep. Is there any good reason to give up synchrounous lumpy reclaim when trylock_page() failed? IOW, briefly lock_page() and wait_on_page_writeback() have the same latency. why should we only avoid former? side note: page lock contention is very common case. For case (8), I don't think sleeping is right way. get_page() is used in really various place of our kernel. so we can't assume it's only temporary reference count increasing. In the other hand, this contention is not so common because shrink_page_list() is excluded from IO activity by page-lock and wait_on_page_writeback(). so I think giving up this case don't makes too many pages eviction. If you disagree, can you please explain your expected bad scinario? > > > > 3. pageout() is intended anynchronous api. but doesn't works so. > > > > > > > > pageout() call ->writepage with wbc->nonblocking=1. because if the system have > > > > default vm.dirty_ratio (i.e. 20), we have 80% clean memory. so, getting stuck > > > > on one page is stupid, we should scan much pages as soon as possible. > > > > > > > > HOWEVER, block layer ignore this argument. if slow usb memory device connect > > > > to the system, ->writepage() will sleep long time. because submit_bio() call > > > > get_request_wait() unconditionally and it doesn't have any PF_MEMALLOC task > > > > bonus. > > > > > > Is this not a problem in the writeback layer rather than pageout() > > > specifically? > > > > Well, outside pageout(), probably only XFS makes PF_MEMALLOC + writeout. > > because PF_MEMALLOC is enabled only very limited situation. but I don't know > > XFS detail at all. I can't tell this area... > > > > All direct reclaimers have PF_MEMALLOC set so it's not that limited a > situation. See here Yes, all direct reclaimers have PF_MEMALLOC. but usually all direct reclaimers don't call any IO related function except pageout(). As far as I know, current shrink_icache() and shrink_dcache() doesn't make IO. Am I missing something? -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/