From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jens Axboe Subject: Re: block: Deprecate QUEUE_FLAG_CLUSTER and use queue_limits instead Date: Wed, 1 Dec 2010 19:43:36 +0100 Message-ID: <4CF69758.40900@fusionio.com> References: <85ABFDC4641C4847BA6088F2982EFD8F049D4EB8@nonamew.ptu.promise.com> <4CEEAE35.5060805@fusionio.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-scsi-owner@vger.kernel.org To: "Martin K. Petersen" Cc: device-mapper development , "linux-scsi@vger.kernel.org" , Promise_Linux List-Id: dm-devel.ids On 2010-11-26 01:37, Martin K. Petersen wrote: >>>>>> "Jens" == Jens Axboe writes: > > Jens> Great, the two different values and needing to sync them was > Jens> horrible. What kind of testing did you do? Have to be a little > Jens> extra careful at this point. > > Yeah, we should probably let it soak a bit in -next just to make sure. > > There really aren't many devices from this millennium that don't support > clustering. Which I guess is why we haven't seen any problems. > > I ended up disabling clustering in one of the FC drivers to test with a > real workload. Threw in a BUG_ON(nsegs > queue_max_segments(q)) for good > measure. > > I also tested mixing and matching clustered and non-clustered bottom > devices while stacking with DM. > > New version below, fixing the things you and Matthew pointed out... Thanks applied. Small plea for the future - please rewrite the subject line to indicate a v2 (or whatever) of the patch. That will make them harder to miss. -- Jens Axboe