From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266687AbUJAWwg (ORCPT ); Fri, 1 Oct 2004 18:52:36 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266810AbUJAWwc (ORCPT ); Fri, 1 Oct 2004 18:52:32 -0400 Received: from zcars04f.nortelnetworks.com ([47.129.242.57]:56822 "EHLO zcars04f.nortelnetworks.com") by vger.kernel.org with ESMTP id S266687AbUJAWts (ORCPT ); Fri, 1 Oct 2004 18:49:48 -0400 Message-ID: <415DDF09.8030703@nortelnetworks.com> Date: Fri, 01 Oct 2004 16:49:45 -0600 X-Sybari-Space: 00000000 00000000 00000000 00000000 From: Chris Friesen User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.6) Gecko/20040113 X-Accept-Language: en-us, en MIME-Version: 1.0 To: Linux kernel Subject: bug?: strange behaviour with memory hogs and 2.6.9-rc3 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I'm running 2.6.9-rc3 with 2GB of ram. I've got a test app that loops allocating chunks of memory 60 pages at a time, scribbles on each page, and mlock()s it, for a total of about a gig and a half. It then sleeps for 100 seconds, and exits. I ran one instance fine. I ran the second instance, and userspace locked up. The machine still responds to pings, but no userspace stuff works. It's been about 10 minutes, and userspace still hasn't come back. Obviously I'm out of memory, but I would have expected that the oom killer would wake up or something. Anyone have any ideas? Chris