> From: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
> Sent: Friday, April 17, 2020 5:04 AM
> To: Wei Liu <wei.liu@xxxxxxxxxx>
>
> Wei Liu <wei.liu@xxxxxxxxxx> writes:
>
> > On Fri, Apr 17, 2020 at 12:03:18PM +0200, Vitaly Kuznetsov wrote:
> >> Dexuan Cui <decui@xxxxxxxxxxxxx> writes:
> >>
> >> > Unlike the other CPUs, CPU0 is never offlined during hibernation. So in the
> >> > resume path, the "new" kernel's VP assist page is not suspended (i.e.
> >> > disabled), and later when we jump to the "old" kernel, the page is not
> >> > properly re-enabled for CPU0 with the allocated page from the old kernel.
> >> >
> >> > So far, the VP assist page is only used by hv_apic_eoi_write().
> >>
> >> No, not only for that ('git grep hv_get_vp_assist_page')
Sorry, I unintentionally ignored that, as I have few knowledge about the
optimization for nested virtualization. :-)
> >> KVM on Hyper-V also needs VP assist page to use Enlightened VMCS. In
> >> particular, Enlightened VMPTR is written there.
> >>
> >> This makes me wonder: how does hibernation work with KVM in case we
> use
> >> Enlightened VMCS and we have VMs running? We need to make sure VP
> Assist
> >> page content is preserved.
> >
> > The page itself is preserved, isn't it?
> >
>
> Right, unlike hyperv_pcpu_input_arg is is not freed.
>
> > hv_cpu_die never frees the vp_assit page. It merely disables it.
> > hv_cpu_init only allocates a new page if necessary.
>
> I'm not really sure that Hyper-V will like us when we disable VP Assist
> page and have an active L2 guest using Enlightened VMCS, who knows what
> it caches and when. I'll try to at least test if/how it works.
>
> This all is not really related to Dexuan's patch)
> --
> Vitaly
It looks you imply that: if there is no active L2 guests, it should be safe to
disable/reenable the assist page upon hibernation?
Can you please write a patch for KVM (when KVM runs on Hyper-V) to abort
the hibernation request if there is any active L2 guest? The pm_notifier can
be used for this.
Thanks,
-- Dexuan
