From: Mike Snitzer Subject: Re: dm snapshot: rework writing to snapshot origin Date: Tue, 17 Nov 2009 12:49:03 -0500 Message-ID: <20091117174902.GB5223@redhat.com> References: <1258472477-5795-1-git-send-email-snitzer@redhat.com> Reply-To: device-mapper development Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1258472477-5795-1-git-send-email-snitzer@redhat.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com To: Alasdair G Kergon Cc: dm-devel@redhat.com, Mikulas Patocka List-Id: dm-devel.ids Alasdair, I am not aware of any "cons" associated with this patch. Mikulas are you? Alasdair would like both "pros" and "cons" documented in the patch header. Here is a revised patch header that now documents why this patch is a prerequisite for snapshot-merge (specifically this patch: http://patchwork.kernel.org/patch/60708/): From: Mikulas Patocka The previous code selected one exception as "primary_pe", linked all other exceptions on it and used reference counting to wait until all exceptions are reallocated. All the complexity with exceptions linking and reference counting has been removed. Now, a bio is linked to one exception and when that exception is reallocated, the bio is retried to possibly wait for other exceptions. The new __origin_write() interface affords the snapshot-merge support, snapshot_merge_process() specifically, the ability to trigger exceptions in other snapshots without having a need for an associated bio (which snapshot-merge does not generate). As such this patch is a prerequisite for snapshot-merge. Signed-off-by: Mikulas Patocka Signed-off-by: Mike Snitzer