4 messages, 3 authors, 2012-08-15 · open the first message on its own page
From: Rafael Aquini <hidden>
Date: 2012-08-14 20:11:33
On Tue, Aug 14, 2012 at 10:51:39PM +0300, Michael S. Tsirkin wrote:
What I think you should do is use rcu for access. And here sync rcu before freeing. Maybe an overkill but at least a documented synchronization primitive, and it is very light weight.
I liked your suggestion on barriers, as well. Rik, Mel ?
From: "Michael S. Tsirkin" <mst@redhat.com>
Date: 2012-08-14 20:31:23
On Tue, Aug 14, 2012 at 05:11:13PM -0300, Rafael Aquini wrote:
On Tue, Aug 14, 2012 at 10:51:39PM +0300, Michael S. Tsirkin wrote:quoted
What I think you should do is use rcu for access. And here sync rcu before freeing. Maybe an overkill but at least a documented synchronization primitive, and it is very light weight.I liked your suggestion on barriers, as well. Rik, Mel ?
Further instead of simple assignment I would add an api in mm to set/clear the balloon mapping, with proper locking. This could fail if already set, and thus fix crash with many ballons.
From: Mel Gorman <hidden>
Date: 2012-08-15 09:05:34
On Tue, Aug 14, 2012 at 05:11:13PM -0300, Rafael Aquini wrote:
On Tue, Aug 14, 2012 at 10:51:39PM +0300, Michael S. Tsirkin wrote:quoted
What I think you should do is use rcu for access. And here sync rcu before freeing. Maybe an overkill but at least a documented synchronization primitive, and it is very light weight.I liked your suggestion on barriers, as well.
I have not thought about this as deeply as I shouold but is simply rechecking the mapping under the pages_lock to make sure the page is still a balloon page an option? i.e. use pages_lock to stabilise page->mapping. -- Mel Gorman SUSE Labs
From: "Michael S. Tsirkin" <mst@redhat.com>
Date: 2012-08-15 09:25:28
On Wed, Aug 15, 2012 at 10:05:28AM +0100, Mel Gorman wrote:
On Tue, Aug 14, 2012 at 05:11:13PM -0300, Rafael Aquini wrote:quoted
On Tue, Aug 14, 2012 at 10:51:39PM +0300, Michael S. Tsirkin wrote:quoted
What I think you should do is use rcu for access. And here sync rcu before freeing. Maybe an overkill but at least a documented synchronization primitive, and it is very light weight.I liked your suggestion on barriers, as well.I have not thought about this as deeply as I shouold but is simply rechecking the mapping under the pages_lock to make sure the page is still a balloon page an option? i.e. use pages_lock to stabilise page->mapping.
To clarify, are you concerned about cost of rcu_read_lock for non balloon pages?
-- Mel Gorman SUSE Labs