From be38d276b0189fa86231fc311428622a1981ad62 Mon Sep 17 00:00:00 2001 From: Avi Kivity Date: Sun, 6 Jun 2010 14:31:27 +0300 Subject: [PATCH] KVM: MMU: Introduce drop_spte() When we call rmap_remove(), we (almost) always immediately follow it by an __set_spte() to a nonpresent pte. Since we need to perform the two operations atomically, to avoid losing the dirty and accessed bits, introduce a helper drop_spte() and convert all call sites. The operation is still nonatomic at this point. Signed-off-by: Avi Kivity --- arch/x86/kvm/mmu.c | 30 +++++++++++++++++------------- arch/x86/kvm/paging_tmpl.h | 13 ++++++------- 2 files changed, 23 insertions(+), 20 deletions(-) diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index 70cdf6876b5f..1ad39cf70e18 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -658,6 +658,12 @@ static void rmap_remove(struct kvm *kvm, u64 *spte) } } +static void drop_spte(struct kvm *kvm, u64 *sptep, u64 new_spte) +{ + rmap_remove(kvm, sptep); + __set_spte(sptep, new_spte); +} + static u64 *rmap_next(struct kvm *kvm, unsigned long *rmapp, u64 *spte) { struct kvm_rmap_desc *desc; @@ -722,9 +728,9 @@ static int rmap_write_protect(struct kvm *kvm, u64 gfn) BUG_ON((*spte & (PT_PAGE_SIZE_MASK|PT_PRESENT_MASK)) != (PT_PAGE_SIZE_MASK|PT_PRESENT_MASK)); pgprintk("rmap_write_protect(large): spte %p %llx %lld\n", spte, *spte, gfn); if (is_writable_pte(*spte)) { - rmap_remove(kvm, spte); + drop_spte(kvm, spte, + shadow_trap_nonpresent_pte); --kvm->stat.lpages; - __set_spte(spte, shadow_trap_nonpresent_pte); spte = NULL; write_protected = 1; } @@ -744,8 +750,7 @@ static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp, while ((spte = rmap_next(kvm, rmapp, NULL))) { BUG_ON(!(*spte & PT_PRESENT_MASK)); rmap_printk("kvm_rmap_unmap_hva: spte %p %llx\n", spte, *spte); - rmap_remove(kvm, spte); - __set_spte(spte, shadow_trap_nonpresent_pte); + drop_spte(kvm, spte, shadow_trap_nonpresent_pte); need_tlb_flush = 1; } return need_tlb_flush; @@ -767,8 +772,7 @@ static int kvm_set_pte_rmapp(struct kvm *kvm, unsigned long *rmapp, rmap_printk("kvm_set_pte_rmapp: spte %p %llx\n", spte, *spte); need_flush = 1; if (pte_write(*ptep)) { - rmap_remove(kvm, spte); - __set_spte(spte, shadow_trap_nonpresent_pte); + drop_spte(kvm, spte, shadow_trap_nonpresent_pte); spte = rmap_next(kvm, rmapp, NULL); } else { new_spte = *spte &~ (PT64_BASE_ADDR_MASK); @@ -1464,7 +1468,8 @@ static void kvm_mmu_page_unlink_children(struct kvm *kvm, } else { if (is_large_pte(ent)) --kvm->stat.lpages; - rmap_remove(kvm, &pt[i]); + drop_spte(kvm, &pt[i], + shadow_trap_nonpresent_pte); } } pt[i] = shadow_trap_nonpresent_pte; @@ -1868,9 +1873,8 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep, if (level > PT_PAGE_TABLE_LEVEL && has_wrprotected_page(vcpu->kvm, gfn, level)) { ret = 1; - rmap_remove(vcpu->kvm, sptep); - spte = shadow_trap_nonpresent_pte; - goto set_pte; + drop_spte(vcpu->kvm, sptep, shadow_trap_nonpresent_pte); + goto done; } spte |= PT_WRITABLE_MASK; @@ -1902,6 +1906,7 @@ static int set_spte(struct kvm_vcpu *vcpu, u64 *sptep, set_pte: __set_spte(sptep, spte); +done: return ret; } @@ -1938,8 +1943,7 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep, } else if (pfn != spte_to_pfn(*sptep)) { pgprintk("hfn old %lx new %lx\n", spte_to_pfn(*sptep), pfn); - rmap_remove(vcpu->kvm, sptep); - __set_spte(sptep, shadow_trap_nonpresent_pte); + drop_spte(vcpu->kvm, sptep, shadow_trap_nonpresent_pte); kvm_flush_remote_tlbs(vcpu->kvm); } else was_rmapped = 1; @@ -2591,7 +2595,7 @@ static void mmu_pte_write_zap_pte(struct kvm_vcpu *vcpu, pte = *spte; if (is_shadow_present_pte(pte)) { if (is_last_spte(pte, sp->role.level)) - rmap_remove(vcpu->kvm, spte); + drop_spte(vcpu->kvm, spte, shadow_trap_nonpresent_pte); else { child = page_header(pte & PT64_BASE_ADDR_MASK); mmu_page_remove_parent_pte(child, spte); diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h index 59e750c1a269..796a325c7e59 100644 --- a/arch/x86/kvm/paging_tmpl.h +++ b/arch/x86/kvm/paging_tmpl.h @@ -353,8 +353,7 @@ static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr, } if (is_large_pte(*sptep)) { - rmap_remove(vcpu->kvm, sptep); - __set_spte(sptep, shadow_trap_nonpresent_pte); + drop_spte(vcpu->kvm, sptep, shadow_trap_nonpresent_pte); kvm_flush_remote_tlbs(vcpu->kvm); } @@ -516,12 +515,13 @@ static void FNAME(invlpg)(struct kvm_vcpu *vcpu, gva_t gva) pte_gpa += (sptep - sp->spt) * sizeof(pt_element_t); if (is_shadow_present_pte(*sptep)) { - rmap_remove(vcpu->kvm, sptep); if (is_large_pte(*sptep)) --vcpu->kvm->stat.lpages; + drop_spte(vcpu->kvm, sptep, + shadow_trap_nonpresent_pte); need_flush = 1; - } - __set_spte(sptep, shadow_trap_nonpresent_pte); + } else + __set_spte(sptep, shadow_trap_nonpresent_pte); break; } @@ -637,12 +637,11 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp, !is_present_gpte(gpte) || !(gpte & PT_ACCESSED_MASK)) { u64 nonpresent; - rmap_remove(vcpu->kvm, &sp->spt[i]); if (is_present_gpte(gpte) || !clear_unsync) nonpresent = shadow_trap_nonpresent_pte; else nonpresent = shadow_notrap_nonpresent_pte; - __set_spte(&sp->spt[i], nonpresent); + drop_spte(vcpu->kvm, &sp->spt[i], nonpresent); continue; }