From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Mike Snitzer" Subject: Re: [PATCH 1/6] Convert bonding timers to workqueues Date: Mon, 22 Oct 2007 10:55:17 -0400 Message-ID: <170fa0d20710220755w744e8ec8w1d85cae4f4be822e@mail.gmail.com> References: <11920760494096-git-send-email-fubar@us.ibm.com> <11920760502756-git-send-email-fubar@us.ibm.com> <4713B44E.7030902@pobox.com> <20038.1192475339@death> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: "Jay Vosburgh" , netdev@vger.kernel.org, andy@greyhouse.net To: "Jeff Garzik" Return-path: Received: from nz-out-0506.google.com ([64.233.162.238]:2141 "EHLO nz-out-0506.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752365AbXJVOzS (ORCPT ); Mon, 22 Oct 2007 10:55:18 -0400 Received: by nz-out-0506.google.com with SMTP id s18so442999nze for ; Mon, 22 Oct 2007 07:55:18 -0700 (PDT) In-Reply-To: <20038.1192475339@death> Content-Disposition: inline Sender: netdev-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On 10/15/07, Jay Vosburgh wrote: > Jeff Garzik wrote: > > >Jay Vosburgh wrote: > >> Convert bonding timers to workqueues. This converts the various > >> monitor functions to run in periodic work queues instead of timers. This > >> patch introduces the framework and convers the calls, but does not resolve > >> various locking issues, and does not stand alone. > >> > >> Signed-off-by: Andy Gospodarek > >> Signed-off-by: Jay Vosburgh > > > >"does not stand alone" == it is not bisectable? That's a problem. > > The patch will compile fine (this is true for any point in the > series), but it's possible for bonding to deadlock or misbehave when > doing certain operations. Those problems are addressed in the later > patches. > > I didn't want to just vomit out one big mega-patch that has the > whole set of changes, because that's harder to understand, and the later > patches generally address discrete issues. > > Does that make you more or less nervous about its bisectability? > I can repost the whole thing as a big blob if that's what you'd prefer. > > >Anyway, we have more fun fish to fry: after applying the IPoIB bonding > >patchset, this no longer applies... > > Yah, I'll rebase it (once I know your blob vs. not-blob > preference). I wasn't sure how long the IPoIB mystery bug hunt was > going to take. Jeff, Given that Jay rebased and posted updated bonding locking patches on Oct 17 will you be pushing them for inclusion in 2.6.24-rc1? or 2.6.24-rcX? I don't see them queued in your netdev-2.6 tree. These bonding locking issues have been a problem for some time and it would be very good to get them fixed upstream sooner rather than later. thanks, Mike