From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from eggs.gnu.org ([2001:4830:134:3::10]:52327)
	by lists.gnu.org with esmtp (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1WR0M2-0004w9-Ee
	for qemu-devel@nongnu.org; Fri, 21 Mar 2014 10:23:32 -0400
Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)
	(envelope-from <pbonzini@redhat.com>) id 1WR0Ls-0000zk-BP
	for qemu-devel@nongnu.org; Fri, 21 Mar 2014 10:23:26 -0400
Message-ID: <532C4B4A.6030807@redhat.com>
Date: Fri, 21 Mar 2014 15:23:06 +0100
From: Paolo Bonzini <pbonzini@redhat.com>
MIME-Version: 1.0
References: <1394603550-11556-1-git-send-email-aik@ozlabs.ru>		
	<1394603550-11556-8-git-send-email-aik@ozlabs.ru>	
	<1395259061.8201.242.camel@ul30vt.home>
	<532BF15C.7030802@ozlabs.ru> <1395411438.632.52.camel@ul30vt.home>
In-Reply-To: <1395411438.632.52.camel@ul30vt.home>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Subject: Re: [Qemu-devel] [PATCH v5 07/11] vfio: Add guest side IOMMU support
List-Id: <qemu-devel.nongnu.org>
List-Unsubscribe: <https://lists.nongnu.org/mailman/options/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=unsubscribe>
List-Archive: <http://lists.nongnu.org/archive/html/qemu-devel>
List-Post: <mailto:qemu-devel@nongnu.org>
List-Help: <mailto:qemu-devel-request@nongnu.org?subject=help>
List-Subscribe: <https://lists.nongnu.org/mailman/listinfo/qemu-devel>,
	<mailto:qemu-devel-request@nongnu.org?subject=subscribe>
To: Alex Williamson <alex.williamson@redhat.com>, Alexey Kardashevskiy <aik@ozlabs.ru>
Cc: Alexander Graf <agraf@suse.de>, qemu-ppc@nongnu.org, qemu-devel@nongnu.org, David Gibson <david@gibson.dropbear.id.au>

Il 21/03/2014 15:17, Alex Williamson ha scritto:
>> > > Is this possible?  Assuming len is initially a power-of-2, would the
>> > > translate function change it?  Maybe worth a comment to explain.
>>
>>
>> Oh. address_space_translate() actually changes @len to min(len,
>> TARGET_PAGE_SIZE) and TARGET_PAGE_SIZE is hardcoded to 4K. So far it was ok
>> but lately I have been implementing a huge DMA window (plus one
>> sPAPRTCETable and one VFIOGuestIOMMU objects) which currently operates with
>> 16MB pages (can do 64K pages too) and now this "granularity incompatible"
>> is happening.
>>
>> I disabled that check but I need to think of better fix...
>>
>> Adding Paolo to cc, may be he picks the context and gives good piece of
>> advise :)

I think that "if" in address_space_translate should be restricted to 
Xen, since that's what it was added for.

Paolo