From mboxrd@z Thu Jan  1 00:00:00 1970
From: Olaf Hering <olaf@aepfle.de>
Subject: paged granttable entries
Date: Thu, 26 Aug 2010 14:13:55 +0200
Message-ID: <20100826121355.GA6254@aepfle.de>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Return-path: <xen-devel-bounces@lists.xensource.com>
Content-Disposition: inline
List-Unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xensource.com?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xensource.com>
List-Help: <mailto:xen-devel-request@lists.xensource.com?subject=help>
List-Subscribe: <http://lists.xensource.com/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xensource.com?subject=subscribe>
Sender: xen-devel-bounces@lists.xensource.com
Errors-To: xen-devel-bounces@lists.xensource.com
To: Patrick Colp <pjcolp@cs.ubc.ca>
Cc: xen-devel@lists.xensource.com
List-Id: xen-devel@lists.xenproject.org


Patrick,

after adding some debug to __gnttab_map_grant_ref(), I found the gfn
referenced by sha1->frame has p2m_type_t == p2m_ram_paged. As a result,
the following code gets into the 'iomem_access_permitted(rd, frame, frame)' branch.

Some naive approach to call p2m_mem_paging_populate() and return
GNTST_eagain lead to a deadlock in the guest. The guest reponds to keypresses,
xm destroy does not work.


How are paged out granttable entries supposed to come back when the
guest did not do the GNTTABOP_map_grant_ref hypercall yet? I tried to
skip gfn listed in the guests granttable in p2m_mem_paging_nominate, but
that doesnt appear to work because new pages can be added to the
granttable at anytime by the drivers.


Olaf