From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266568AbUG0UK3 (ORCPT ); Tue, 27 Jul 2004 16:10:29 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266582AbUG0UK3 (ORCPT ); Tue, 27 Jul 2004 16:10:29 -0400 Received: from atrey.karlin.mff.cuni.cz ([195.113.31.123]:24785 "EHLO atrey.karlin.mff.cuni.cz") by vger.kernel.org with ESMTP id S266568AbUG0UKY (ORCPT ); Tue, 27 Jul 2004 16:10:24 -0400 Date: Mon, 26 Jul 2004 23:02:30 +0200 From: Pavel Machek To: Avi Kivity Cc: linux-kernel@vger.kernel.org Subject: Re: [PATCH] Deadlock during heavy write activity to userspace NFS server on local NFS mount Message-ID: <20040726210229.GC21889@openzaurus.ucw.cz> References: <41050300.90800@exanet.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <41050300.90800@exanet.com> User-Agent: Mutt/1.3.27i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi! > On heavy write activity, allocators wait synchronously for kswapd to > free some memory. But if kswapd is freeing memory via a userspace NFS > server, that server could be waiting for kswapd, and the system seizes > instantly. > > This patch (against RHEL 2.4.21-15EL, but should apply either > literally > or conceptually to other kernels) allows a process to declare itself > as > kswapd's little helper, and thus will not have to wait on kswapd. Ok, but what if its memory runs out, anyway? Pavel -- 64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms