Re: [PATCH MANUALSEL 5.14 4/9] KVM: x86: reset pdptrs_from_userspace when exiting smm

[Paolo Bonzini]

On 06/10/21 15:30, Sasha Levin wrote:
> From: Maxim Levitsky <mlevitsk@xxxxxxxxxx>
>
> [ Upstream commit 37687c403a641f251cb2ef2e7830b88aa0647ba9 ]
>
> When exiting SMM, pdpts are loaded again from the guest memory.
>
> This fixes a theoretical bug, when exit from SMM triggers entry to the
> nested guest which re-uses some of the migration
> code which uses this flag as a workaround for a legacy userspace.
>
> Signed-off-by: Maxim Levitsky <mlevitsk@xxxxxxxxxx>
> Message-Id: <20210913140954.165665-4-mlevitsk@xxxxxxxxxx>
> Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
> ---
>   arch/x86/kvm/x86.c | 7 +++++++
>   1 file changed, 7 insertions(+)
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index b3f855d48f72..1e7d629bbf36 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -7659,6 +7659,13 @@ static void kvm_smm_changed(struct kvm_vcpu *vcpu, bool entering_smm)
>   
>   		/* Process a latched INIT or SMI, if any.  */
>   		kvm_make_request(KVM_REQ_EVENT, vcpu);
> +
> +		/*
> +		 * Even if KVM_SET_SREGS2 loaded PDPTRs out of band,
> +		 * on SMM exit we still need to reload them from
> +		 * guest memory
> +		 */
> +		vcpu->arch.pdptrs_from_userspace = false;
>   	}
>   
>   	kvm_mmu_reset_context(vcpu);

Acked-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>