From mboxrd@z Thu Jan 1 00:00:00 1970 From: Konrad Rzeszutek Wilk Subject: Re: Re: [PATCH] xen/blkback: add xen-blkback support Date: Mon, 18 Apr 2011 16:14:43 -0400 Message-ID: <20110418201443.GC12562@dumpdata.com> References: <1303149464-875-1-git-send-email-konrad.wilk@oracle.com> <1303149464-875-2-git-send-email-konrad.wilk@oracle.com> <20110418181211.GA1149@dumpdata.com> <1303151503.9965.65.camel@agari.van.xensource.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1303151503.9965.65.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: Daniel Stodden Cc: "xen-devel@lists.xensource.com" List-Id: xen-devel@lists.xenproject.org > > Hmm, let me remove those get/puts. They are actually not necessary as we > > already hold a reference count when we opened the device. > > Ack. I never noticed these weren't exported. > > But bdev_get_queue is not an addref, just an accessor macro, as you're > probably aware of. Yes. > > But get/put is normally managed entirely in block/ and implied by the > request and bdev interfaces. And I found the reason for all of that. Way back in the past (2.6.18) we would be checking the request_queue->flag to make sure it did not have the QUEUE_FLAG_DEAD flag set. Over the years the block API started doing that in the submit_bio path and .. well, we don't need this check anymore.