Return-Path: linux-nfs-owner@vger.kernel.org Received: from mail-ob0-f174.google.com ([209.85.214.174]:54405 "EHLO mail-ob0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753671Ab2HCMJK (ORCPT ); Fri, 3 Aug 2012 08:09:10 -0400 Received: by obbuo13 with SMTP id uo13so998719obb.19 for ; Fri, 03 Aug 2012 05:09:09 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <866.1343987269@warthog.procyon.org.uk> References: <866.1343987269@warthog.procyon.org.uk> From: William Dauchy Date: Fri, 3 Aug 2012 14:08:49 +0200 Message-ID: Subject: Re: nfs_idmap_legacy_upcall To: David Howells Cc: Trond Myklebust , linux-nfs@vger.kernel.org, "Schumaker, Bryan" Content-Type: text/plain; charset=ISO-8859-1 Sender: linux-nfs-owner@vger.kernel.org List-ID: Hi David, Thank you for your reply. On Fri, Aug 3, 2012 at 11:47 AM, David Howells wrote: > I'm afraid I'm not going to be able to do a lot with this until the 13th as > I'm about to go on holiday. If it is still a problem when I get back, I'll > look at it then. Thanks. I started to dig it and went through the huge backlog between 3.4 and 3.5. It seems like a race during mount since it's only happening sometimes and can't reproduce it easily. I first thought there was still an issue in this fix: b102743: NFS: Force the legacy idmapper to be single threaded but then found: 4697bd5:NFSv4: Fix a race in the net namespace mount notification I guess this last one might be related to my issues since idmaper is using struct nfs_client. I started new tests with it. Regards, -- William