From: James Pearson Subject: Support for larger than 32Kb [rw]size in older 2.6 kernels? Date: Wed, 22 Apr 2009 17:36:39 +0100 Message-ID: <49EF4797.2070402@moving-picture.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed To: linux-nfs@vger.kernel.org Return-path: Received: from mpc-26.sohonet.co.uk ([193.203.82.251]:53227 "EHLO moving-picture.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752317AbZDVRG4 (ORCPT ); Wed, 22 Apr 2009 13:06:56 -0400 Received: from cassini.mpc.local ([172.16.15.37] helo=moving-picture.com) by moving-picture.com with esmtp (Exim 4.43) id 1LwfR9-0002gp-Bz for linux-nfs@vger.kernel.org; Wed, 22 Apr 2009 17:36:39 +0100 Sender: linux-nfs-owner@vger.kernel.org List-ID: We have a number of machines that use RHEL4 kernels that run 3rd party software/hardware that we can't upgrade to run on newer kernels. I've managed (I think!) to backport the large [rw]size patch that went into 2.6.16 to the RHEL4 2.6.9 based kernels - taken from: Doing some simple tests over NFS using this patch seems to suggest that it works fine. Although the NFS code between the two kernel is significantly different, it was fairly straight forward to match up where the changes should go - however, I don't pretend to fully understand the code, so is there anything that I should be aware of that might cause a problem? i.e. does larger [rw]size depend on anything else that that would have been added to the kernel since 2.6.9? One thing that isn't in the RHEL4 is 'rpc_max_payload' - which is used as a check on the [rw]size - however, this appears not to be important as the maximum [rw]size I can get is 1Mb, which matches the max [rw]size with more recent kernels. Thanks James Pearson