From mboxrd@z Thu Jan 1 00:00:00 1970 From: Xiao Guangrong Subject: [PATCH v3 2/6] KVM: MMU: don't drop spte if overwrite it from W to RO Date: Fri, 19 Nov 2010 17:02:35 +0800 Message-ID: <4CE63D2B.7090909@cn.fujitsu.com> References: <4CE63CF4.80502@cn.fujitsu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , KVM , LKML To: Avi Kivity Return-path: Received: from cn.fujitsu.com ([222.73.24.84]:55653 "EHLO song.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751370Ab0KSI6V (ORCPT ); Fri, 19 Nov 2010 03:58:21 -0500 In-Reply-To: <4CE63CF4.80502@cn.fujitsu.com> Sender: kvm-owner@vger.kernel.org List-ID: We just need flush tlb if overwrite a writable spte with a read-only one. And we should move this operation to set_spte() for sync_page path Signed-off-by: Xiao Guangrong --- arch/x86/kvm/mmu.c | 20 +++++++++----------- 1 files changed, 9 insertions(+), 11 deletions(-) diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index bdb9fa9..59b5bd2 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -1960,7 +1960,7 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep, gfn_t gfn, pfn_t pfn, bool speculative, bool can_unsync, bool reset_host_protection) { - u64 spte; + u64 spte, entry = *sptep; int ret = 0; /* @@ -2031,6 +2031,14 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep, set_pte: update_spte(sptep, spte); + /* + * If we overwrite a writable spte with a read-only one we + * should flush remote TLBs. Otherwise rmap_write_protect + * will find a read-only spte, even though the writable spte + * might be cached on a CPU's TLB. + */ + if (is_writable_pte(entry) && !is_writable_pte(*sptep)) + kvm_flush_remote_tlbs(vcpu->kvm); done: return ret; } @@ -2069,16 +2077,6 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep, spte_to_pfn(*sptep), pfn); drop_spte(vcpu->kvm, sptep, shadow_trap_nonpresent_pte); kvm_flush_remote_tlbs(vcpu->kvm); - /* - * If we overwrite a writable spte with a read-only one, - * drop it and flush remote TLBs. Otherwise rmap_write_protect - * will find a read-only spte, even though the writable spte - * might be cached on a CPU's TLB. - */ - } else if (is_writable_pte(*sptep) && - (!(pte_access & ACC_WRITE_MASK) || !dirty)) { - drop_spte(vcpu->kvm, sptep, shadow_trap_nonpresent_pte); - kvm_flush_remote_tlbs(vcpu->kvm); } else was_rmapped = 1; } -- 1.7.0.4