From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1161049AbWG0NMh (ORCPT ); Thu, 27 Jul 2006 09:12:37 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1161056AbWG0NMh (ORCPT ); Thu, 27 Jul 2006 09:12:37 -0400 Received: from omx2-ext.sgi.com ([192.48.171.19]:35241 "EHLO omx2.sgi.com") by vger.kernel.org with ESMTP id S1161049AbWG0NMh (ORCPT ); Thu, 27 Jul 2006 09:12:37 -0400 Date: Thu, 27 Jul 2006 06:12:32 -0700 From: Paul Jackson To: Andrew Morton Cc: linux-kernel@vger.kernel.org Subject: Should cpuset ABBA deadlock fix be in 2.6.18-rc2-mmx? Message-Id: <20060727061232.044fc927.pj@sgi.com> In-Reply-To: <20060727015639.9c89db57.akpm@osdl.org> References: <20060727015639.9c89db57.akpm@osdl.org> Organization: SGI X-Mailer: Sylpheed version 2.2.4 (GTK+ 2.8.3; i686-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I'd like to encourage including the patch: Cpuset: fix ABBA deadlock with cpu hotplug lock in 2.6.18-rc2-mmx. This patch was sent to lkml, copied to akpm, 14 July. It was sent again to lkml, in response to a request for a copy from Linus, on 23 July, copied again to akpm. Granted, the case for including it is not absolute. The fix is simple enough and fixes a definite deadlock, as verified by testing results on the system that was seeing this deadlock. However ... only one system, world-wide, has the magic workload to provoke this deadlock, so far. And it involves the cpu hotplug lock, which perhaps Linus would like to nuke entirely. -- I won't rest till it's the best ... Programmer, Linux Scalability Paul Jackson 1.925.600.0401