From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ian Jackson Subject: Re: [PATCH 20/29] libxl: cancellation: Note that driver domain task cannot be usefully cancelled Date: Tue, 31 Mar 2015 19:37:38 +0100 Message-ID: <21786.59762.369959.680022@mariner.uk.xensource.com> References: <1423599016-32639-1-git-send-email-ian.jackson@eu.citrix.com> <1423599016-32639-21-git-send-email-ian.jackson@eu.citrix.com> <1427197873.21742.363.camel@citrix.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1427197873.21742.363.camel@citrix.com> List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Sender: xen-devel-bounces@lists.xen.org Errors-To: xen-devel-bounces@lists.xen.org To: Ian Campbell Cc: xen-devel@lists.xensource.com, Euan Harris , Roger Pau Monne List-Id: xen-devel@lists.xenproject.org Ian Campbell writes ("Re: [Xen-devel] [PATCH 20/29] libxl: cancellation: Note that driver domain task cannot be usefully cancelled"): > On Tue, 2015-02-10 at 20:10 +0000, Ian Jackson wrote: > > In practice, cancelling this task will cause all subsequent actual > > backend operations to fail, but will not actually cause the > > libxl_device_events_handler operation to complete. ... > Acked-by: Ian Campbell > > This is an unfortunate short-coming though, and I presume one which > could be fixed by updates to the toolstack<->driver domain protocol? It could be fixed by updates to the code inside libxl for the driver domain task. But it is not clear that the benefit would be large, and the memory management in the driver domain task means that it would be tricky to implement - there would have to be more tracking of what operations are ongoing. The behaviour of a driver domain service task which is in the process of being cancelled is not anything that you would want, so a cancel-and-restart isn't a nice operation anyway. Ian.