Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751099AbdHQGF7 (ORCPT ); Thu, 17 Aug 2017 02:05:59 -0400 Received: from mail-oi0-f66.google.com ([209.85.218.66]:34699 "EHLO mail-oi0-f66.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750800AbdHQGF5 (ORCPT ); Thu, 17 Aug 2017 02:05:57 -0400 MIME-Version: 1.0 In-Reply-To: References: <1502848704-7474-1-git-send-email-tianyu.lan@intel.com> <1cf50aea-0c1c-f2f1-952e-9bcdd76722e8@redhat.com> From: Wanpeng Li Date: Thu, 17 Aug 2017 14:05:56 +0800 Message-ID: Subject: Re: [PATCH] KVM/VMX: Avoid CR3 VMEXIT during guest real mode when "unrestricted guest" is supported. To: Paolo Bonzini Cc: Lan Tianyu , Radim Krcmar , Thomas Gleixner , Ingo Molnar , "H. Peter Anvin" , "the arch/x86 maintainers" , kvm , "linux-kernel@vger.kernel.org" , Nadav Amit Content-Type: text/plain; charset="UTF-8" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3351 Lines: 75 Cc Nadav, 2017-08-17 13:58 GMT+08:00 Wanpeng Li : > 2017-08-16 17:25 GMT+08:00 Paolo Bonzini : >> On 16/08/2017 03:58, Lan Tianyu wrote: >>> These CR3 VMEXITs was introduced for platform without "unrestricted guest" >>> support. This is to set ept identity table to guest CR3 in guest real >>> mode because these platforms don't support ept real mode(CR0.PE and CR0.PG >>> must be set to 1). But these VMEXITs is redundant for platforms with >>> "unrestricted guest" support. >> >> This is true, but I'm not sure it's a good idea to make things more >> complex... It is working code and is not a bottleneck anywhere. > > Sorry, I have another question. There is no A20 emulation in kvm > currently, then whether or not the vm8086 which is emulated by kvm can > guarantee real mode guest which access occurs at 1 MByte will incur > address-wraparound? > > Regards, > Wanpeng Li > >> >> Paolo >> >>> Signed-off-by: Lan Tianyu >>> --- >>> arch/x86/kvm/vmx.c | 22 +++++++++++++--------- >>> 1 file changed, 13 insertions(+), 9 deletions(-) >>> >>> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c >>> index 9b21b12..46dcf50 100644 >>> --- a/arch/x86/kvm/vmx.c >>> +++ b/arch/x86/kvm/vmx.c >>> @@ -4221,18 +4221,20 @@ static void ept_update_paging_mode_cr0(unsigned long *hw_cr0, >>> vmx_decache_cr3(vcpu); >>> if (!(cr0 & X86_CR0_PG)) { >>> /* From paging/starting to nonpaging */ >>> - vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, >>> - vmcs_read32(CPU_BASED_VM_EXEC_CONTROL) | >>> - (CPU_BASED_CR3_LOAD_EXITING | >>> - CPU_BASED_CR3_STORE_EXITING)); >>> + if (!enable_unrestricted_guest) >>> + vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, >>> + vmcs_read32(CPU_BASED_VM_EXEC_CONTROL) | >>> + (CPU_BASED_CR3_LOAD_EXITING | >>> + CPU_BASED_CR3_STORE_EXITING)); >>> vcpu->arch.cr0 = cr0; >>> vmx_set_cr4(vcpu, kvm_read_cr4(vcpu)); >>> } else if (!is_paging(vcpu)) { >>> /* From nonpaging to paging */ >>> - vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, >>> - vmcs_read32(CPU_BASED_VM_EXEC_CONTROL) & >>> - ~(CPU_BASED_CR3_LOAD_EXITING | >>> - CPU_BASED_CR3_STORE_EXITING)); >>> + if (!enable_unrestricted_guest) >>> + vmcs_write32(CPU_BASED_VM_EXEC_CONTROL, >>> + vmcs_read32(CPU_BASED_VM_EXEC_CONTROL) & >>> + ~(CPU_BASED_CR3_LOAD_EXITING | >>> + CPU_BASED_CR3_STORE_EXITING)); >>> vcpu->arch.cr0 = cr0; >>> vmx_set_cr4(vcpu, kvm_read_cr4(vcpu)); >>> } >>> @@ -4311,7 +4313,9 @@ static void vmx_set_cr3(struct kvm_vcpu *vcpu, unsigned long cr3) >>> } >>> >>> vmx_flush_tlb(vcpu); >>> - vmcs_writel(GUEST_CR3, guest_cr3); >>> + >>> + if (!enable_unrestricted_guest || !enable_ept) >>> + vmcs_writel(GUEST_CR3, guest_cr3); >>> } >>> >>> static int vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4) >>> >>