From mboxrd@z Thu Jan 1 00:00:00 1970 Message-Id: <200108280631.f7S6VWJ16822@maile.telia.com> Content-Type: text/plain; charset="iso-8859-1" From: Roger Larsson Subject: Re: kernel: __alloc_pages: 1-order allocation failed Date: Tue, 28 Aug 2001 08:27:04 +0200 References: In-Reply-To: MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: owner-linux-mm@kvack.org Return-Path: To: Marcelo Tosatti , Daniel Phillips Cc: Andrew Kay , linux-mm@kvack.org List-ID: On Tuesdayen den 28 August 2001 00:28, Marcelo Tosatti wrote: > On Tue, 28 Aug 2001, Daniel Phillips wrote: > > On August 27, 2001 10:14 pm, Andrew Kay wrote: > > > I am having some rather serious problems with the memory management (i > > > think) in the 2.4.x kernels. I am currently on the 2.4.9 and get lots > > > of these errors in /var/log/messages. > > > > > > Aug 24 15:08:04 dell63 kernel: __alloc_pages: 1-order allocation > > > failed. Aug 24 15:08:35 dell63 last message repeated 448 times > > > Aug 24 15:09:37 dell63 last message repeated 816 times > > > Aug 24 15:10:38 dell63 last message repeated 1147 times > > > > > > I am running a Redhat 7.1 distro w/2.4.9 kernel on a Dell poweredge > > > 6300 (4x500Mhz cpu, 4Gb ram). I get this error while running the > > > specmail 2001 benchmarking software against our email server, > > > Intrastore. The system is very idle from what I can see. The sar > > > output shows user cpu at around 1% and everything else rather low as > > > well. It seems to pop up randomly and requires a reboot to fix it. > > > > > > Is there any workarounds or something I can do to get a more useful > > > debug message than this? > > > > Please apply this patch: > > > > --- 2.4.9.clean/mm/page_alloc.c Thu Aug 16 12:43:02 2001 > > +++ 2.4.9/mm/page_alloc.c Mon Aug 20 22:05:40 2001 > > @@ -502,7 +502,8 @@ > > } > > > > /* No luck.. */ > > - printk(KERN_ERR "__alloc_pages: %lu-order allocation failed.\n", > > order); + printk(KERN_ERR "__alloc_pages: %lu-order allocation failed > > (gfp=0x%x/%i).\n", + order, gfp_mask, !!(current->flags & PF_MEMALLOC)); > > return NULL; > > } > > Daniel, > > Its probably the bounce buffering thingie. > > I'll send a patch to Linus soon. I have seen reports of this problem when running without HIGHMEM (from Stephan von Krawczynski ). But he is running with knfs(d) and raiserfs... In this configuration I really want to know who is responsible for the allocs. So, please add a show_trace(NULL); too... /RogerL -- Roger Larsson Skelleftea Sweden -- 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/