From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from psmtp.com (na3sys010amx104.postini.com [74.125.245.104]) by kanga.kvack.org (Postfix) with SMTP id 459C46B002C for ; Wed, 8 Feb 2012 14:27:22 -0500 (EST) Received: by bkty12 with SMTP id y12so1112077bkt.14 for ; Wed, 08 Feb 2012 11:27:20 -0800 (PST) MIME-Version: 1.0 In-Reply-To: References: <1328271538-14502-1-git-send-email-m.szyprowski@samsung.com> <1328271538-14502-12-git-send-email-m.szyprowski@samsung.com> <20120203140428.GG5796@csn.ul.ie> From: sandeep patil Date: Wed, 8 Feb 2012 11:26:40 -0800 Message-ID: Subject: Re: [Linaro-mm-sig] [PATCH 11/15] mm: trigger page reclaim in alloc_contig_range() to stabilize watermarks Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Sender: owner-linux-mm@kvack.org List-ID: To: Michal Nazarewicz Cc: Mel Gorman , Marek Szyprowski , Ohad Ben-Cohen , Daniel Walker , Russell King , Arnd Bergmann , Jesse Barker , Jonathan Corbet , linux-kernel@vger.kernel.org, Dave Hansen , linaro-mm-sig@lists.linaro.org, linux-mm@kvack.org, Kyungmin Park , KAMEZAWA Hiroyuki , Andrew Morton , Rob Clark , linux-arm-kernel@lists.infradead.org, linux-media@vger.kernel.org 2012/2/8 Michal Nazarewicz : > On Wed, 08 Feb 2012 03:04:18 +0100, sandeep patil > wrote: >> >> There's another problem I am facing with zone watermarks and CMA. >> >> Test details: >> Memory =A0: 480 MB of total memory, 128 MB CMA region >> Test case : around 600 MB of file transfer over USB RNDIS onto target >> System Load : ftpd with console running on target. >> No one is doing CMA allocations except for the DMA allocations done by t= he >> drivers. >> >> Result : After about 300MB transfer, I start getting GFP_ATOMIC >> allocation failures. =A0This only happens if CMA region is reserved. >> Total memory available is way above the zone watermarks. So, we ended >> up starving >> UNMOVABLE/RECLAIMABLE atomic allocations that cannot fallback on CMA >> region. > > > This looks like something Mel warned me about. =A0I don't really have a g= ood > solution for that yet. ;/ What if we have NR_FREE_CMA_PAGES in vmstat and use them to calculate __zone_watermark_ok()? However, it still doesn't solve the problem when we DON'T want to use NR_FREE_CMA_PAGES in case of MOVABLE allocations. Sandeep -- 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 internet charges in Canada: sign http://stopthemeter.ca/ Don't email: email@kvack.org