From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH 3/3] VMX: Enhance invalid guest state emulation Date: Mon, 31 Aug 2009 15:35:05 +0300 Message-ID: <4A9BC379.2060804@redhat.com> References: <1251470963-14542-1-git-send-email-m.gamal005@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: kvm@vger.kernel.org To: Mohammed Gamal Return-path: Received: from mx1.redhat.com ([209.132.183.28]:15166 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751417AbZHaMfO (ORCPT ); Mon, 31 Aug 2009 08:35:14 -0400 In-Reply-To: <1251470963-14542-1-git-send-email-m.gamal005@gmail.com> Sender: kvm-owner@vger.kernel.org List-ID: On 08/28/2009 05:49 PM, Mohammed Gamal wrote: > - Change returned handle_invalid_guest_state() to return relevant exit codes > - Move triggering the emulation from vmx_vcpu_run() to vmx_handle_exit() > - Return to userspace instead of repeatedly trying to emulate > instructions that have already failed > > Signed-off-by: Mohammed Gamal > --- > arch/x86/kvm/vmx.c | 24 ++++++++++++++---------- > 1 files changed, 14 insertions(+), 10 deletions(-) > > diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c > index 78101dd..e422470 100644 > --- a/arch/x86/kvm/vmx.c > +++ b/arch/x86/kvm/vmx.c > @@ -3318,10 +3318,11 @@ static int handle_nmi_window(struct kvm_vcpu *vcpu) > return 1; > } > > -static void handle_invalid_guest_state(struct kvm_vcpu *vcpu) > +static int handle_invalid_guest_state(struct kvm_vcpu *vcpu) > { > struct vcpu_vmx *vmx = to_vmx(vcpu); > enum emulation_result err = EMULATE_DONE; > + int ret = 1; > > local_irq_enable(); > preempt_enable(); > These are now wrong, since handle_invalid_exit() is called with interrupts and preemption enabled. > @@ -3329,11 +3330,16 @@ static void handle_invalid_guest_state(struct kvm_vcpu *vcpu) > while (!guest_state_valid(vcpu)) { > err = emulate_instruction(vcpu, 0, 0, 0); > > - if (err == EMULATE_DO_MMIO) > + if (err == EMULATE_DO_MMIO) { > + ret = 0; > break; > + } > > if (err != EMULATE_DONE) { > kvm_report_emulation_failure(vcpu, "emulation failure"); > + vcpu->run->exit_reason = KVM_EXIT_INTERNAL_ERROR; > + vcpu->run->internal.suberror = KVM_INTERNAL_ERROR_EMULATION; > + ret = 0; > break; > } > > @@ -3347,6 +3353,7 @@ static void handle_invalid_guest_state(struct kvm_vcpu *vcpu) > local_irq_disable(); > > vmx->invalid_state_emulation_result = err; > + return ret; > } > > /* > @@ -3405,9 +3412,12 @@ static int vmx_handle_exit(struct kvm_vcpu *vcpu) > /* If we need to emulate an MMIO from handle_invalid_guest_state > * we just return 0 */ > if (vmx->emulation_required&& emulate_invalid_guest_state) { > - if (guest_state_valid(vcpu)) > + if (guest_state_valid(vcpu)) { > vmx->emulation_required = 0; > - return vmx->invalid_state_emulation_result != EMULATE_DO_MMIO; > + return vmx->invalid_state_emulation_result != EMULATE_DO_MMIO; > This looks fishy. Can't say exactly why but vmx_handle_exit() should only depend on the current guest execution, not the previous guest execution. > + } else { > + return handle_invalid_guest_state(vcpu); > + } > } > > /* Access CR3 don't cause VMExit in paging mode, so we need > @@ -3603,12 +3613,6 @@ static void vmx_vcpu_run(struct kvm_vcpu *vcpu) > if (unlikely(!cpu_has_virtual_nmis()&& vmx->soft_vnmi_blocked)) > vmx->entry_time = ktime_get(); > > - /* Handle invalid guest state instead of entering VMX */ > - if (vmx->emulation_required&& emulate_invalid_guest_state) { > - handle_invalid_guest_state(vcpu); > - return; > - } > - > Don't we still need to return here? Otherwise we attempt guest entry needlessly. > if (test_bit(VCPU_REGS_RSP, (unsigned long *)&vcpu->arch.regs_dirty)) > vmcs_writel(GUEST_RSP, vcpu->arch.regs[VCPU_REGS_RSP]); > if (test_bit(VCPU_REGS_RIP, (unsigned long *)&vcpu->arch.regs_dirty)) > -- error compiling committee.c: too many arguments to function