Re: [Xen-devel] [PATCH RFC v2 8/9] xen/arm: Implement hypercall for dirty page tracing (shadow op)

[Eugene Fedotov <e.fedotov@xxxxxxxxxxx>]

1) Maybe it reasonable use a new HVM parameter for guest RAM base PFN 
instead of new DOMCTL and setup it from libxc through existing 
xc_set_hvm_param hypercall?
2) Since a static GUEST_RAM_BASE is also used in libxc zImage loader 
(xc_dom_armzimageloader.c) should we parse guest's DTB in function 
xc_dom_parse_zimage_kernel or simply provide dom->rambase_pfn to the 
hypervisor?
Best regards,
Evgeny

04.07.2013 14:36, Ian Campbell
> On Thu, 2013-07-04 at 11:29 +0100, Stefano Stabellini wrote:
> > On Thu, 4 Jul 2013, Jaeyong Yoo wrote:
> > > > > Yes, you are right. Since there is no way to figure out this value at
> > > > > the moment, we use this static value here. I think we need to parse
> > > > > dtb of domu guest and get this value from there. Do you have any
> > > > suggestion?
> > > >
> > > > The hypervisor should never be parsing (or even seeing) the domU DTB.
> > > > Instead the toolstack should poke this sort of information down via a
> > > > (presumably new) DOMCTL.
> > > Got it.
> > I wonder whether we could find the start of memory by looking at the p2m
> > and then storing the value in a per-domain variable to avoid a pagetable
> > walk every time.
> I think the toolstack should just tell the hypervisor.
>
> Ian.
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxx
> http://lists.xen.org/xen-devel

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