From: "J. Bruce Fields" Subject: Re: [PATCH] NLM: Revert parts of commit 24e36663 Date: Fri, 26 Sep 2008 14:45:54 -0400 Message-ID: <20080926184554.GA1492@fieldses.org> References: <20080924184732.4078.83334.stgit@tarkus.1015granger.net> <20080925213831.GF14349@fieldses.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: trond.myklebust@netapp.com, linux-nfs@vger.kernel.org To: Chuck Lever Return-path: Received: from mail.fieldses.org ([66.93.2.214]:42343 "EHLO fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751762AbYIZSp5 (ORCPT ); Fri, 26 Sep 2008 14:45:57 -0400 In-Reply-To: Sender: linux-nfs-owner@vger.kernel.org List-ID: On Fri, Sep 26, 2008 at 01:53:58PM -0400, Chuck Lever wrote: > On Sep 25, 2008, at Sep 25, 2008, 5:38 PM, J. Bruce Fields wrote: >> On Wed, Sep 24, 2008 at 02:50:36PM -0400, Chuck Lever wrote: >>> statd and some servers always need lockd to listen on UDP, so always >>> start both lockd listener sockets. >>> >>> This probably leaks an svc_serv if the TCP listener can't be created, >>> but it will do for now. >>> >>> Signed-off-by: Chuck Lever >>> --- >>> >>> I currently have this workaround lurking in my git repo. Is there >>> some kind >>> of fix for this issue planned for 2.6.28? >> >> Not that I know of. >> >>> This is only a reminder, not a suggested patch. >> >> Any chance you could generate a patch? > > I can, but no guarantee it will be in time for the 2.6.28 merge window. > > Would we consider this a regression fix? I suppose so. (Are there any bug reports? If so, could we get URL's for them into the commit message?) > If so, we could post it after > the merge window closes. It would give a little more time to do some > testing. That sound suspiciously like "since this is a high priority, it will be accepted later, so we can make it a lower priority." Help! But anyway, yes, we might have more time. --b.