From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: kernel 3.1.1 message: warn_alloc_failed Date: Fri, 9 Dec 2011 09:50:14 -0800 Message-ID: <20111209095014.14a647cd@nehalam.linuxnetplumber.net> References: <6.2.5.6.2.20111208210303.03a06228@flumedata.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 8BIT Cc: linux-kernel@vger.kernel.org, netdev To: starlight@binnacle.cx, Sony Chacko , Rajesh Borundia Return-path: In-Reply-To: <6.2.5.6.2.20111208210303.03a06228@flumedata.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Thu, 08 Dec 2011 21:10:02 -0500 starlight@binnacle.cx wrote: > Hello. > > FYI saw several of these when running > multiple 'rcp' commands copying multi-gigabyte > files to/from system. > > 'rcp' target files were all on an encrypted > mirrored USB3-attached pair of 3TB hard > drives. > > One outgoing 'rcp' was to a system with > a much slower disk (due to a USB2 attachment) > and that transfer was putting back-pressure > on the connection. > > Kernel is running on a CentOS 6.0 > distribution. > > Not subscribing to the lists, so please > CC me with any queries. Please note that > email originating from most non-US MTAs > will bounce, but I will check the list > archives if I see any. > > Can't spend a lot of time on this but am > willing to answer a few questions. Hopefully > the call stack will tell the story well > enough. You are seeing memory allocation failures because device is allocating a 16K (order 2) size socket buffer. You are using netxen device, and it looks like the problem. >>From reading the netxen driver source. The LRO buffers in this device are very large (8060+skb overhead). Until the driver is fixed to use fragmented page size memory, I recommend turning off LRO.