From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753921AbdF0WeC (ORCPT ); Tue, 27 Jun 2017 18:34:02 -0400 Received: from mail-oi0-f68.google.com ([209.85.218.68]:34429 "EHLO mail-oi0-f68.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752856AbdF0Wd6 (ORCPT ); Tue, 27 Jun 2017 18:33:58 -0400 MIME-Version: 1.0 In-Reply-To: <61bbcc19-a818-6934-75e9-8aed28523aa0@redhat.com> References: <1498097214-6239-1-git-send-email-wanpeng.li@hotmail.com> <1498097214-6239-4-git-send-email-wanpeng.li@hotmail.com> <61bbcc19-a818-6934-75e9-8aed28523aa0@redhat.com> From: Wanpeng Li Date: Wed, 28 Jun 2017 06:33:56 +0800 Message-ID: Subject: Re: [PATCH v4 3/4] KVM: async_pf: Force a nested vmexit if the injected #PF is async_pf To: Paolo Bonzini Cc: "linux-kernel@vger.kernel.org" , kvm , =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= , Wanpeng Li Content-Type: text/plain; charset="UTF-8" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org 2017-06-27 21:40 GMT+08:00 Paolo Bonzini : > > > On 22/06/2017 04:06, Wanpeng Li wrote: >> From: Wanpeng Li >> >> Add an async_page_fault field to vcpu->arch.exception to identify an async >> page fault, and constructs the expected vm-exit information fields. Force >> a nested VM exit from nested_vmx_check_exception() if the injected #PF >> is async page fault. Extending the userspace interface KVM_GET_VCPU_EVENTS >> and KVM_SET_VCPU_EVENTS for live migration. > > I am not sure what would happen if new kernel (that can produce > async_page_fault=1) runs on top of old userspace (that can consume it). > > I think it would be safer to make the new field "nested_apf", and only > set it if in guest_mode, like > > vcpu->arch.exception.nested_apf = > is_guest_mode(vcpu) && fault->async_page_fault; > if (vcpu->arch.exception.nested_apf) > vcpu->arch.apf.nested_apf_token = fault->address; > else > vcpu->arch.cr2 = fault->address; I have already added the same logic in kvm_inject_page_fault in patch 3/4, in addition, there is a guarantee it is in guest mode when we set svm->vmcb->control.xxxx in nested_svm_check_exception, how about just as what we do in nested_vmx_check_exception? + if (svm->vcpu.arch.exception.async_page_fault) + svm->vmcb->control.exit_info_2 = svm->vcpu.arch.apf.nested_apf_token; + else + svm->vmcb->control.exit_info_2 = svm->vcpu.arch.cr2; Regards, Wanpeng Li