From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bandan Das Subject: Re: [PATCH v4 3/3] KVM: nVMX: Emulate EPTP switching for the L1 hypervisor Date: Thu, 13 Jul 2017 13:08:49 -0400 Message-ID: References: <20170710204936.4001-1-bsd@redhat.com> <20170710204936.4001-4-bsd@redhat.com> <2d50ebc4-9328-ce08-b55b-6a331ee13cc3@redhat.com> <8c58f57d-091f-103d-5d8c-fc49b9d00e13@redhat.com> Mime-Version: 1.0 Content-Type: text/plain Cc: kvm@vger.kernel.org, pbonzini@redhat.com, linux-kernel@vger.kernel.org, Radim =?utf-8?B?S3LEjW3DocWZ?= To: David Hildenbrand Return-path: In-Reply-To: <8c58f57d-091f-103d-5d8c-fc49b9d00e13@redhat.com> (David Hildenbrand's message of "Thu, 13 Jul 2017 17:39:48 +0200") Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org David Hildenbrand writes: >>>> + /* >>>> + * If the (L2) guest does a vmfunc to the currently >>>> + * active ept pointer, we don't have to do anything else >>>> + */ >>>> + if (vmcs12->ept_pointer != address) { >>>> + if (address >> cpuid_maxphyaddr(vcpu) || >>>> + !IS_ALIGNED(address, 4096)) >>> >>> Couldn't the pfn still be invalid and make kvm_mmu_reload() fail? >>> (triggering a KVM_REQ_TRIPLE_FAULT) >> >> If there's a triple fault, I think it's a good idea to inject it >> back. Basically, there's no need to take care of damage control >> that L1 is intentionally doing. > > I quickly rushed over the massive amount of comments. Sounds like you'll > be preparing a v5. Would be great if you could add some comments that > were the result of this discussion (for parts that are not that obvious > - triple faults) - thanks! Will do. Basically, we agreed that we don't need to do anything with mmu_reload() faillures because the invalid eptp that mmu_unload will write to root_hpa will result in an ept violation. Bandan