From mboxrd@z Thu Jan  1 00:00:00 1970
From: Ian Campbell <ian.campbell@citrix.com>
Subject: Re: [PATCH V13 4/7] xen/arm: Data abort exception (R/W)
 mem_events.
Date: Thu, 12 Mar 2015 15:35:54 +0000
Message-ID: <1426174554.32572.16.camel@citrix.com>
References: <1425677073-13729-1-git-send-email-tklengyel@sec.in.tum.de>
	<1425677073-13729-5-git-send-email-tklengyel@sec.in.tum.de>
	<1426167305.21353.423.camel@citrix.com>
	<CABfawhmPXzMn_=SqkHUjgWzBW+h5uWm78-w-NZNTQDzPJU2n5g@mail.gmail.com>
	<CABfawhnqna-uZMQvwM=a8ia1ESJJbBHeXCwHF5YqQAJNGVgYMQ@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Return-path: <xen-devel-bounces@lists.xen.org>
In-Reply-To: <CABfawhnqna-uZMQvwM=a8ia1ESJJbBHeXCwHF5YqQAJNGVgYMQ@mail.gmail.com>
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: Tamas K Lengyel <tklengyel@sec.in.tum.de>
Cc: wei.liu2@citrix.com, Stefano Stabellini <stefano.stabellini@eu.citrix.com>, Tim Deegan <tim@xen.org>, Julien Grall <julien.grall@linaro.org>, Ian Jackson <ian.jackson@eu.citrix.com>, xen-devel@lists.xen.org, stefano.stabellini@citrix.com, Jan Beulich <jbeulich@suse.com>, Keir Fraser <keir@xen.org>
List-Id: xen-devel@lists.xenproject.org

On Thu, 2015-03-12 at 16:19 +0100, Tamas K Lengyel wrote:
> 
> >  out:
>                 > +    if ( flush )
>                 > +    {
>                 > +        flush_tlb_domain(d);
>                 > +        iommu_iotlb_flush(d, sgfn, egfn - sgfn);
>                 > +    }
>                 
>                 Is moving the flush out of the loop an independent bug
>                 fix? If so please
>                 do in a separate commit with a rationale in the commit
>                 log. If it is
>                 somehow related to the changes here then please
>                 mention it in this
>                 commit log, since it's a bit subtle.
>         
>         
>         Right, it's not a bugfix and not required to be outside the
>         loop, I think I just moved it because it made sense to me to
>         flush it only once instead at every iteration. I'll place it
>         back.

> 
> Sorry, the flush wasn't actually part of the loop to begin with. I
> just moved it under the label out so that the TLB gets flushed when
> the memaccess setting hypercall gets preempted. I will just set a
> separate label for it before out so that the existing behavior is
> preserved but the tlb is still flushed when memaccess is preempted.

I wonder why this isn't needed for the other uses of goto out, i.e. on
the relinquish check.

I'm not convinced this isn't just a straight up bug. Anyone remember any
reasoning why we don't flush on exit if any work has been done?

Ian.