Re: [PATCH v3 2/5] KVM: MMU: check guest CR3 reserved bits based on its physical address width.

From: Yu Zhang
Date: Thu Aug 24 2017 - 12:43:49 EST




On 8/25/2017 12:27 AM, Paolo Bonzini wrote:
On 24/08/2017 17:38, Yu Zhang wrote:

In practice, MAXPHYADDR will never be 59 even because the PKRU bits are
at bits 59..62.
Thanks, Paolo.
I see. I had made an assumption that MAXPHYADDR shall not exceed the
physical one,
which is 52 I believe. But I'm not sure there's any place to check this.
Maybe we should make sure the vcpu->arch.maxphyaddr will not be greater
than the value of the host?
That's a separate change anyway. In any case, since currently the
MAXPHYADDR is not validated, your change to rsvd_bits makes sense.

Thanks, Paolo.
As to this patch series, any change I need do?

BTW, I have written a patch for kvm-unit-test access test, but the test failed.
Not sure if my patch is erroneous or due to a simulator error. I'll send out the
test patch after it works.:-)

Yu
Paolo