Return-Path: Received: from mx2.netapp.com ([216.240.18.37]:33429 "EHLO mx2.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752415Ab0KDPWu (ORCPT ); Thu, 4 Nov 2010 11:22:50 -0400 Received: from localhost.localdomain (dnsindia.hq.netapp.com [10.58.52.94] (may be forged)) by smtp1.corp.netapp.com (8.13.1/8.13.1/NTAP-1.6) with ESMTP id oA4FMYMM029583 for ; Thu, 4 Nov 2010 08:22:35 -0700 (PDT) From: Fred Isaman To: linux-nfs@vger.kernel.org Subject: [PATCH 00/18] rewrite of CB_LAYOUTRECALL and layoutstate code Date: Thu, 4 Nov 2010 11:22:13 -0400 Message-Id: <1288884151-11128-1-git-send-email-iisaman@netapp.com> Sender: linux-nfs-owner@vger.kernel.org List-ID: Content-Type: text/plain MIME-Version: 1.0 This code rewrites the layout state handling and the CB_LAYOUTRECALL paths. It applies to Benny's current pnfs-submit branch (commit eb04948a). Patch 01 - A session's callback patch that passes the CB_SEQUENCE info (the session in particular) up to subsequent operations. Andy is looking at issues Trond pointed out with this patch, but something very much like it will come. 02-12 - These are what smallish changes I could pull out of the rewrite. Patch 03 in particular introduces changes to refcounting that are at the heart of io drain notification. 13 - Sorry this patch is so huge, but I am simultaneously changing the stateid and RPC processing for LAYOUTGET, LAYOUTRETURN, and CB_LAYOUTRECALL, and I did not see an easy way to split them bisectably. 14 - More CB_LAYOUTRECALL changes that could be more easily split out 15-18 Andy's code that puts LAYOUTCOMMIT and LAYOUTRETURN in the same compound as CLOSE Fred