From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH] mlx4: allocate just enough pages instead of always 4 pages Date: Wed, 04 Apr 2012 20:34:48 -0400 (EDT) Message-ID: <20120404.203448.1118889674233934663.davem@davemloft.net> References: <1333568440-14494-1-git-send-email-cascardo@linux.vnet.ibm.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, roland@kernel.org, ogerlitz@mellanox.com, yevgenyp@mellanox.co.il, alexg@mellanox.co.il, klebers@linux.vnet.ibm.com To: cascardo@linux.vnet.ibm.com Return-path: Received: from shards.monkeyblade.net ([198.137.202.13]:51398 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754062Ab2DEAfF (ORCPT ); Wed, 4 Apr 2012 20:35:05 -0400 In-Reply-To: <1333568440-14494-1-git-send-email-cascardo@linux.vnet.ibm.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Thadeu Lima de Souza Cascardo Date: Wed, 4 Apr 2012 16:40:40 -0300 > The driver uses a 2-order allocation, which is too much on architectures > like ppc64, which has a 64KiB page. This particular allocation is used > for large packet fragments that may have a size of 512, 1024, 4096 or > fill the whole allocation. So, a minimum size of 16384 is good enough > and will be the same size that is used in architectures of 4KiB sized > pages. > > This will avoid allocation failures that we see when the system is under > stress, but still has plenty of memory, like the one below. > > This will also allow us to set the interface MTU to higher values like > 9000, which was not possible on ppc64 without this patch. ... > Signed-off-by: Thadeu Lima de Souza Cascardo > Signed-off-by: Kleber Sacilotto de Souza > Tested-by: Kleber Sacilotto de Souza Applied, thanks a lot.