Return-Path: linux-nfs-owner@vger.kernel.org Received: from mail-qc0-f182.google.com ([209.85.216.182]:64183 "EHLO mail-qc0-f182.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753222AbaEELPI (ORCPT ); Mon, 5 May 2014 07:15:08 -0400 Received: by mail-qc0-f182.google.com with SMTP id e16so5058795qcx.13 for ; Mon, 05 May 2014 04:15:07 -0700 (PDT) Date: Mon, 5 May 2014 07:08:12 -0400 From: Jeff Layton To: David Rientjes Cc: Marc Dietrich , linux-kernel@vger.kernel.org, linux-nfs@vger.kernel.org, Trond Myklebust , Mel Gorman Subject: Re: nfs oom on 3.15 Message-ID: <20140505070812.470c66bd@tlielax.poochiereds.net> In-Reply-To: References: <2562595.ix17LO7uqe@fb07-iapwap2> MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-nfs-owner@vger.kernel.org List-ID: On Mon, 5 May 2014 02:42:35 -0700 (PDT) David Rientjes wrote: > On Mon, 5 May 2014, Marc Dietrich wrote: > > > Hi, > > > > this is spaming my log quite a bit. > > > > Any idea? > > > > I think rpc_malloc() really wants to be __GFP_NOWARN so that the warnings > are suppressed, doing order-1 allocations with GFP_NOWAIT is never > guaranteed to succeed and will have difficulty if memory is fragmented. Yes. We made a similar fix to xprt_alloc_slot a while back. The RPC engine has a mechanism to handle retrying these allocations, so this warning isn't particularly helpful. -- Jeff Layton