On Mon, Apr 20, 2020 at 07:46:11PM -0700, Dexuan Cui wrote:
> 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. not
> 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 used by hv_apic_eoi_write(), and is also
> used in the case of nested virtualization (running KVM atop Hyper-V).
>
> For hv_apic_eoi_write(), when the page is not properly re-enabled,
> hvp->apic_assist is always 0, so the HV_X64_MSR_EOI MSR is always written.
> This is not ideal with respect to performance, but Hyper-V can still
> correctly handle this according to the Hyper-V spec; nevertheless, Linux
> still must update the Hyper-V hypervisor with the correct VP assist page
> to prevent Hyper-V from writing to the stale page, which causes guest
> memory corruption and consequently may have caused the hangs and triple
> faults seen during non-boot CPUs resume.
>
> Fix the issue by calling hv_cpu_die()/hv_cpu_init() in the syscore ops.
> Without the fix, hibernation can fail at a rate of 1/300 ~ 1/500.
> With the fix, hibernation can pass a long-haul test of 2000 runs.
>
> In the case of nested virtualization, disabling/reenabling the assist
> page upon hibernation may be unsafe if there are active L2 guests.
> It looks KVM should be enhanced to abort the hibernation request if
> there is any active L2 guest.
>
> Fixes: 05bd330a7fd8 ("x86/hyperv: Suspend/resume the hypercall page for hibernation")
> Cc: stable@xxxxxxxxxxxxxxx
> Signed-off-by: Dexuan Cui <decui@xxxxxxxxxxxxx>
Applied to hyperv-fixes. Thanks.
