Re: [Qemu-devel] [RFC PATCH RDMA support v1: 10/13] introduce new command migrate_check_for_zero

["Michael S. Tsirkin" <mst@redhat.com>]

On Thu, Apr 11, 2013 at 09:19:43AM -0400, Michael R. Hines wrote:
> On 04/11/2013 07:13 AM, Michael S. Tsirkin wrote:
> >On Thu, Apr 11, 2013 at 11:18:38AM +0200, Paolo Bonzini wrote:
> >>Il 11/04/2013 09:38, Michael S. Tsirkin ha scritto:
> >>>On Wed, Apr 10, 2013 at 06:28:18PM -0400, mrhines@linux.vnet.ibm.com wrote:
> >>>>From: "Michael R. Hines" <mrhines@us.ibm.com>
> >>>>
> >>>>This allows the user to disable zero page checking during migration
> >>>>
> >>>>Signed-off-by: Michael R. Hines <mrhines@us.ibm.com>
> >>>IMO this knob is too low level to expose to management.
> >>>Why not disable this automatically when migrating with rdma?
> >>Thinking more about it, I'm not sure why it is important to disable it.
> >This just illustrates the point. There's no place for such low level
> >knobs in the management interface.
> >
> 
> I disagree with that: We already have precedent for this in the
> XBZRLE capability.

My understanding is the issue is protocol compatibility,
not optimization. E.g. you can migrate to file, for each
new feature you need a way to disable it to stay compatible.

> Zero page checking is no "more" low-level than this
> capability already is and the community has already agreed to expose
> this capability to management.
> 
> Since zero page scanning does in fact affect performance, we not give
> the user the option?
> 
> Why would the community agree to expose one low-level feature and
> not expose another?
> 
> - Michael