From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Stodden Subject: Re: [PATCH 00 of 10] blkfront pvops updates Date: Fri, 30 Apr 2010 04:34:30 -0700 Message-ID: <1272627270.11404.112.camel@ramone.somacoma.net> References: <4BD9E2EE.5060402@goop.org> <1272571912.25447.11.camel@agari.van.xensource.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1272571912.25447.11.camel@agari.van.xensource.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xensource.com Errors-To: xen-devel-bounces@lists.xensource.com To: Jeremy Fitzhardinge Cc: ksrinivasan , Xen , Jan Beulich List-Id: xen-devel@lists.xenproject.org On Thu, 2010-04-29 at 16:11 -0400, Daniel Stodden wrote: > On Thu, 2010-04-29 at 15:50 -0400, Jeremy Fitzhardinge wrote: > > On 04/28/2010 07:19 PM, Daniel Stodden wrote: > > > Hi. > > > > > > Here are couple of bugfixes for blkfront. > > > > > > 1. Bdev removal was slightly buggy. [Is nobody out there using > > > block-detach??] Also it seems like we'be been leaking the gendisk > > > structs for a long time. > > > > > > 2. Fixes most common occasions where disk open/close races against > > > backend state switches. This one didn't come for free: > > > * xenbus_switch_state growing transactions again. > > > * New mutex in blkfront_info > > > * Bdev .open/.close rewritten accordingly. > > > > > > > Thanks for this. However, when I went to apply it, it looks like I've > > made a mess of the various blkfront branches. > > > > xen/frontend has some changes from Jan and Ky which are merged into > > xen/next and the various stable branches. > > > > I also have xen/blkfront, which has your last set of changes, which I > > seem to have forgotten to merge into anything else, and now conflicts > > non-trivially with xen/blkfront. I'd like to add this new set of > > patches into xen/blkfront, but it doesn't apply cleanly. > > > > Could you look at this and advise me what I should do? > > Oh, NP. I'll try and see if I can gather those and what it takes to > merge. I moved on to xen/next. That still seems to apply okay. But as usual when looking at that kind of thing a day later, I promptly found 2 more loopholes I didn't see before. %} These should be closed now too, but I'll rather rerun this stuff tomorrow again before resubmitting. Thanks, Daniel