From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: linux-nfs-owner@vger.kernel.org Received: from shards.monkeyblade.net ([198.137.202.13]:59014 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757861Ab2ENU2E (ORCPT ); Mon, 14 May 2012 16:28:04 -0400 Date: Mon, 14 May 2012 16:26:34 -0400 (EDT) Message-Id: <20120514.162634.1094732813264319951.davem@davemloft.net> To: mgorman@suse.de Cc: akpm@linux-foundation.org, linux-mm@kvack.org, netdev@vger.kernel.org, linux-nfs@vger.kernel.org, linux-kernel@vger.kernel.org, Trond.Myklebust@netapp.com, neilb@suse.de, hch@infradead.org, a.p.zijlstra@chello.nl, michaelc@cs.wisc.edu, emunson@mgebm.net Subject: Re: [PATCH 01/12] netvm: Prevent a stream-specific deadlock From: David Miller In-Reply-To: <20120514105604.GB29102@suse.de> References: <1336658065-24851-2-git-send-email-mgorman@suse.de> <20120511.011034.557833140906762226.davem@davemloft.net> <20120514105604.GB29102@suse.de> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Sender: linux-nfs-owner@vger.kernel.org List-ID: From: Mel Gorman Date: Mon, 14 May 2012 11:56:04 +0100 > On Fri, May 11, 2012 at 01:10:34AM -0400, David Miller wrote: >> From: Mel Gorman >> Date: Thu, 10 May 2012 14:54:14 +0100 >> >> > It could happen that all !SOCK_MEMALLOC sockets have buffered so >> > much data that we're over the global rmem limit. This will prevent >> > SOCK_MEMALLOC buffers from receiving data, which will prevent userspace >> > from running, which is needed to reduce the buffered data. >> > >> > Fix this by exempting the SOCK_MEMALLOC sockets from the rmem limit. >> > >> > Signed-off-by: Peter Zijlstra >> > Signed-off-by: Mel Gorman >> >> This introduces an invariant which I am not so sure is enforced. >> >> With this change it is absolutely required that once a socket >> becomes SOCK_MEMALLOC it must never _ever_ lose that attribute. >> > > This is effectively true. In the NFS case, the flag is cleared on > swapoff after all the entries have been paged in. In the NBD case, > SOCK_MEMALLOC is left set until the socket is destroyed. I'll update the > changelog. Bugs happen, you need to find a way to assert that nobody every does this. Because if a bug is introduced which makes this happen, it will otherwise be very difficult to debug.