Re: [Xen-devel] [PATCH v5 0/9] toolstack-based approach to pvhvm guest kexec

[George Dunlap <George.Dunlap@xxxxxxxxxxxxx>]

On Wed, Jan 7, 2015 at 11:03 AM, Olaf Hering <olaf@xxxxxxxxx> wrote:
> On Wed, Jan 07, Vitaly Kuznetsov wrote:
>
>> The thing is .. we don't have these pages when kexec is being performed,
>> they are already ballooned out and the hypervisor doesn't have the
>> knowledge of which GFNs should be re-populated. I think it is possible
>> to keep track of all pages the guest balloons out for this purpose, but
>> ..
>
> Have you tried to make the new guest a PoD guest? That way it may work
> out of the box already.

There's no difference between a PoD guest and a normal guest, except
that the PoD guest starts with some PoD entries.  At the moment, when
the balloon driver hands a page back to Xen, it puts a hole in the
p2m.  Once the balloon driver loads and inflates the balloon,
eventually all the pages will be touched, and there will be no more
PoD entries; at which point it will look exactly like any other guest.

An interesting idea to explore, however, would be to make it policy
rather than putting a hole in the p2m table when ballooning, to put a
p2m entry there instead.  Then when we kexec, it will look pretty much
like a normal PoD guest does booting, but just with some of the memory
pre-populated.  As long as there are enough zero pages lying around,
the PoD system should be able shuffle memory around enough to allow
the kexec'd guest to run until the balloon driver can come up and
inflate the balloon, similar to the way it does in the PoD case.

 -George

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel