Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261526AbUKWVAM (ORCPT ); Tue, 23 Nov 2004 16:00:12 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261544AbUKWU7D (ORCPT ); Tue, 23 Nov 2004 15:59:03 -0500 Received: from facesaver.epoch.ncsc.mil ([144.51.25.10]:64958 "EHLO epoch.ncsc.mil") by vger.kernel.org with ESMTP id S261610AbUKWU6g (ORCPT ); Tue, 23 Nov 2004 15:58:36 -0500 Subject: Re: SELinux performance issue with large systems (32 cpus) From: Stephen Smalley To: keith Cc: devnull@us.ibm.com, djwong , Chris McDermott , "linux-kernel@vger.kernel.org" In-Reply-To: <1101237725.27848.301.camel@knk> References: <1101237725.27848.301.camel@knk> Content-Type: text/plain Organization: National Security Agency Message-Id: <1101243252.19785.316.camel@moss-spartans.epoch.ncsc.mil> Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.6 (1.4.6-2) Date: Tue, 23 Nov 2004 15:54:12 -0500 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 832 Lines: 19 On Tue, 2004-11-23 at 14:22, keith wrote: > After some lock profiling (keeping track of what locks were last used > and how many cycles were spent waiting) it became quite clean the the > avc_lock was to blame. The avc_lock is a SELinux lock. Thanks to work by Kaigai Kohei of NEC, the global avc spinlock has been replaced by an RCU-based scheme. Those changes are in the -mm patches (e.g. 2.6.10-rc2-mm3) and will hopefully go upstream after 2.6.10 is released. There is also ongoing work on baseline performance. -- Stephen Smalley National Security Agency - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/