From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933279Ab2FGWU2 (ORCPT ); Thu, 7 Jun 2012 18:20:28 -0400 Received: from avon.wwwdotorg.org ([70.85.31.133]:35645 "EHLO avon.wwwdotorg.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S933184Ab2FGWU1 (ORCPT ); Thu, 7 Jun 2012 18:20:27 -0400 Message-ID: <4FD12926.8050200@wwwdotorg.org> Date: Thu, 07 Jun 2012 16:20:22 -0600 From: Stephen Warren User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:12.0) Gecko/20120430 Thunderbird/12.0.1 MIME-Version: 1.0 To: Andrew Morton CC: Stephen Rothwell , linux-next@vger.kernel.org, linux-kernel@vger.kernel.org, Konstantin Khlebnikov , Oleg Nesterov Subject: Re: linux-next: boot problem with next-20120607 References: <20120607135320.7a1298385346bc238a0f92b0@canb.auug.org.au> <20120607150549.63f3ab81.akpm@linux-foundation.org> In-Reply-To: <20120607150549.63f3ab81.akpm@linux-foundation.org> X-Enigmail-Version: 1.5pre Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 06/07/2012 04:05 PM, Andrew Morton wrote: > On Thu, 7 Jun 2012 13:53:20 +1000 > Stephen Rothwell wrote: > >> Hi Andrew, >> >> My boot of next-20120607 in PowerPC produces lots of these warnings: >> >> BUG: Bad rss-counter state mm:c000000278508000 idx:0 val:-27 >> BUG: Bad rss-counter state mm:c000000278284000 idx:0 val:-27 >> BUG: Bad rss-counter state mm:c000000279e13000 idx:0 val:-27 >> BUG: Bad rss-counter state mm:c000000278285400 idx:0 val:-27 >> >> And so on (with different mm, idx and val numbers). Searching showed me >> commit "mm: correctly synchronize rss-counters at exit/exec" which seems >> to be there to fix this exact problem? > > Well that's a worry. Are you really sure > mm-correctly-synchronize-rss-counters-at-exit-exec.patch was applied at > the time? > > There's been an extended amount of inconclusive waffling on that patch, > but I thought it concerned the best way to fix it, rather than > *whether* it fixes it. I see those same errors on my ARM system. I found that reverting that patch (105d42c mm: correctly synchronize rss-counters at exit/exec) made them go away, although I didn't investigate whether that just removed some debug print or actually solved a problem.