From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail191.messagelabs.com (mail191.messagelabs.com [216.82.242.19]) by kanga.kvack.org (Postfix) with SMTP id D47EB6B00A5 for ; Wed, 1 Dec 2010 14:29:39 -0500 (EST) Date: Wed, 1 Dec 2010 14:29:35 -0500 (EST) From: CAI Qian Message-ID: <1330724443.975931291231775834.JavaMail.root@zmail06.collab.prod.int.phx2.redhat.com> In-Reply-To: <919384632.877731291171499343.JavaMail.root@zmail06.collab.prod.int.phx2.redhat.com> Subject: Re: oom is broken in mmotm 2010-11-09-15-31 tree? MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 7bit Sender: owner-linux-mm@kvack.org To: linux-mm Cc: Michel Lespinasse , Rik van Riel , Linus Torvalds , Wu Fengguang , "H. Peter Anvin" List-ID: > Hi, just a head-up. When testing oom for this tree, my workstation is > immediately having no response to ssh, Desktop actions and so on apart > from ping. I am trying to bisect but looks like git public server is > having problem. > > # git pull > fatal: read error: Connection reset by peer > > # git clone git://zen-kernel.org/kernel/mmotm.git > Cloning into mmotm... > fatal: read error: Connection reset by peer This turned out that it was introduced by, d065bd810b6deb67d4897a14bfe21f8eb526ba99 mm: retry page fault when blocking on disk transfer It was reproduced by: 1) ssh to the test box. 2) try to trigger oom a few times using a malloc program there. Then, the test box will be unable to process any oom to kill the memory allocation program. If switch VCs for the test box and hit a few ENTER keys locally manually, it may process further oom. After roll-back this one commit, it had no problem to cope the above reproducers and always correctly killed the allocation programs. CAI Qian -- 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