From mboxrd@z Thu Jan  1 00:00:00 1970
From: Ian Campbell <ian.campbell@citrix.com>
Subject: Re: [PATCH v4 26/33] xen/passthrough: Extend
 XEN_DOMCTL_*assign_device to support DT device
Date: Tue, 31 Mar 2015 14:16:41 +0100
Message-ID: <1427807801.2115.123.camel@citrix.com>
References: <1426793399-6283-1-git-send-email-julien.grall@linaro.org>
	<1426793399-6283-27-git-send-email-julien.grall@linaro.org>
	<1427801079.2115.84.camel@citrix.com> <551A934C.6090908@linaro.org>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Return-path: <xen-devel-bounces@lists.xen.org>
Received: from mail6.bemta5.messagelabs.com ([195.245.231.135])
	by lists.xen.org with esmtp (Exim 4.72)
	(envelope-from <Ian.Campbell@citrix.com>) id 1Ycw2A-0003Tq-D6
	for xen-devel@lists.xenproject.org; Tue, 31 Mar 2015 13:16:46 +0000
In-Reply-To: <551A934C.6090908@linaro.org>
List-Unsubscribe: <http://lists.xen.org/cgi-bin/mailman/options/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xen.org>
List-Help: <mailto:xen-devel-request@lists.xen.org?subject=help>
List-Subscribe: <http://lists.xen.org/cgi-bin/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=subscribe>
Sender: xen-devel-bounces@lists.xen.org
Errors-To: xen-devel-bounces@lists.xen.org
To: Julien Grall <julien.grall@linaro.org>
Cc: Wei Liu <wei.liu2@citrix.com>, Ian Jackson <ian.jackson@eu.citrix.com>, tim@xen.org, stefano.stabellini@citrix.com, Jan Beulich <jbeulich@suse.com>, xen-devel@lists.xenproject.org
List-Id: xen-devel@lists.xenproject.org

On Tue, 2015-03-31 at 13:30 +0100, Julien Grall wrote:
> >> +    /* TODO: Do we need to check is_dying? Mostly to protect against
> >> +     * hypercall trying to passthrough a device while we are
> >> +     * dying.
> > 
> > iommu_do_pci_domctl does in specific casses (i.e. assign device). I
> > guess you should follow that lead.
> 
> I'm not sure to fully understand when is_dying should be used or not.
> 
> Looking to the PCI code, the is_dying has been added when we add code to
> deal with page.
> 
> I would be inclined to say it's only necessary when deadling with page.
> Can someone confirm me?
> 
> Otherwise, I don't why is_dying should be check here and not in other call.

it makes little sense to add a device to a domain which is dying, but
it's just about plausible you might want to remove it (e.g. to unwedge
something which has gone wrong).

Overall I think being consistent with other similar cases (e.g. PCI) is
the way to go.