From: Trond Myklebust Subject: Re: [PATCH] xprtrdma: fix XDR tail buf marshalling for all ops Date: Thu, 06 Dec 2007 02:23:36 -0500 Message-ID: <1196925817.7663.1.camel@heimdal.trondhjem.org> References: <1196902607.17724.2.camel@heimdal.trondhjem.org> <1196916625.25528.43.camel@trinity.ogc.int> Mime-Version: 1.0 Content-Type: text/plain Cc: James Lentini , Thomas Talpey , linux-nfs@vger.kernel.org, nfs@lists.sourceforge.net To: Tom Tucker Return-path: Received: from pat.uio.no ([129.240.10.15]:41901 "EHLO pat.uio.no" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752329AbXLFHZl (ORCPT ); Thu, 6 Dec 2007 02:25:41 -0500 In-Reply-To: <1196916625.25528.43.camel-SMNkleLxa3ZimH42XvhXlA@public.gmane.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Wed, 2007-12-05 at 22:50 -0600, Tom Tucker wrote: > xdr->len is 0 Mind explaining where we're missing a call to xdr_adjust_iovec? > > Alternatively, just a single line fix should fix the bug AFAICS: > > - pos += len; > > + pos += xdrbuf->page_len > > xdr->len notwithstanding, for the purpose of this code, the pos > manipulation isn't necessary because the condition being checked (no > tail) can be trivially determined just by looking at tail[0].iov_len. I know that, but we're in an -rc4 release. Trond