From mboxrd@z Thu Jan 1 00:00:00 1970 From: Avi Kivity Subject: Re: [PATCH v4 3/3] KVM: MMU: retry #PF for softmmu Date: Mon, 06 Dec 2010 11:48:00 +0200 Message-ID: <4CFCB150.40202@redhat.com> References: <4CF76A8B.2040102@cn.fujitsu.com> <4CF76B06.6050708@cn.fujitsu.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Cc: Marcelo Tosatti , Gleb Natapov , LKML , KVM To: Xiao Guangrong Return-path: In-Reply-To: <4CF76B06.6050708@cn.fujitsu.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: kvm.vger.kernel.org On 12/02/2010 11:46 AM, Xiao Guangrong wrote: > Retry #PF for softmmu only when the current vcpu has the same cr3 as the time > when #PF occurs > > Changelog: > Just compare cr3 value since It's harmless to instantiate an spte for an > unused translation from Marcelo's comment > It's not harmless. We could be in a different process, so we have to re-fetch the gpte. Or we could have switched from one nested guest to another. > > diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c > index c6bb449..3f0d9a0 100644 > --- a/arch/x86/kvm/mmu.c > +++ b/arch/x86/kvm/mmu.c > @@ -2607,9 +2607,11 @@ static int nonpaging_page_fault(struct kvm_vcpu *vcpu, gva_t gva, > static int kvm_arch_setup_async_pf(struct kvm_vcpu *vcpu, gva_t gva, gfn_t gfn) > { > struct kvm_arch_async_pf arch; > + > arch.token = (vcpu->arch.apf.id++<< 12) | vcpu->vcpu_id; > arch.gfn = gfn; > arch.direct_map = vcpu->arch.mmu.direct_map; > + arch.cr3 = vcpu->arch.mmu.get_cr3(vcpu); > > return kvm_setup_async_pf(vcpu, gva, gfn,&arch); > } > diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h > index 23275d0..437e11a 100644 > --- a/arch/x86/kvm/paging_tmpl.h > +++ b/arch/x86/kvm/paging_tmpl.h > @@ -116,7 +116,7 @@ static unsigned FNAME(gpte_access)(struct kvm_vcpu *vcpu, pt_element_t gpte) > */ > static int FNAME(walk_addr_generic)(struct guest_walker *walker, > struct kvm_vcpu *vcpu, struct kvm_mmu *mmu, > - gva_t addr, u32 access) > + gva_t addr, u32 access, bool prefault) > { > pt_element_t pte; > gfn_t table_gfn; > @@ -194,6 +194,13 @@ walk: > #endif > > if (!eperm&& !rsvd_fault&& !(pte& PT_ACCESSED_MASK)) { > + /* > + * Don't set gpte accessed bit if it's on > + * speculative path. > + */ > + if (prefault) > + goto error; It's actually legal to set the accessed bit on speculative access. -- error compiling committee.c: too many arguments to function