X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=arch%2Fx86%2Fkvm%2Fpaging_tmpl.h;h=258e4591e1ca02246801fc380ae19472b2375b0a;hb=c2d0ee46e6e633a3c23ecbcb9b03ad731906cd79;hp=156fe10288ae330c6d9ac2d3e0b857a9d6c853b5;hpb=35149e2129fe34fc8cb5917e1ecf5156b0fa3415;p=safe%2Fjmp%2Flinux-2.6

diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index 156fe10..258e459 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -29,7 +29,6 @@
 	#define PT_BASE_ADDR_MASK PT64_BASE_ADDR_MASK
 	#define PT_DIR_BASE_ADDR_MASK PT64_DIR_BASE_ADDR_MASK
 	#define PT_INDEX(addr, level) PT64_INDEX(addr, level)
-	#define SHADOW_PT_INDEX(addr, level) PT64_INDEX(addr, level)
 	#define PT_LEVEL_MASK(level) PT64_LEVEL_MASK(level)
 	#define PT_LEVEL_BITS PT64_LEVEL_BITS
 	#ifdef CONFIG_X86_64
@@ -46,7 +45,6 @@
 	#define PT_BASE_ADDR_MASK PT32_BASE_ADDR_MASK
 	#define PT_DIR_BASE_ADDR_MASK PT32_DIR_BASE_ADDR_MASK
 	#define PT_INDEX(addr, level) PT32_INDEX(addr, level)
-	#define SHADOW_PT_INDEX(addr, level) PT64_INDEX(addr, level)
 	#define PT_LEVEL_MASK(level) PT32_LEVEL_MASK(level)
 	#define PT_LEVEL_BITS PT32_LEVEL_BITS
 	#define PT_MAX_FULL_LEVELS 2
@@ -91,14 +89,10 @@ static bool FNAME(cmpxchg_gpte)(struct kvm *kvm,
 	pt_element_t *table;
 	struct page *page;
 
-	down_read(&current->mm->mmap_sem);
 	page = gfn_to_page(kvm, table_gfn);
-	up_read(&current->mm->mmap_sem);
 
 	table = kmap_atomic(page, KM_USER0);
-
 	ret = CMPXCHG(&table[index], orig_pte, new_pte);
-
 	kunmap_atomic(table, KM_USER0);
 
 	kvm_release_page_dirty(page);
@@ -129,6 +123,7 @@ static int FNAME(walk_addr)(struct guest_walker *walker,
 	gfn_t table_gfn;
 	unsigned index, pt_access, pte_access;
 	gpa_t pte_gpa;
+	int rsvd_fault = 0;
 
 	pgprintk("%s: addr %lx\n", __func__, addr);
 walk:
@@ -163,6 +158,10 @@ walk:
 		if (!is_present_pte(pte))
 			goto not_present;
 
+		rsvd_fault = is_rsvd_bits_set(vcpu, pte, walker->level);
+		if (rsvd_fault)
+			goto access_error;
+
 		if (write_fault && !is_writeble_pte(pte))
 			if (user_fault || is_write_protection(vcpu))
 				goto access_error;
@@ -215,7 +214,6 @@ walk:
 		if (ret)
 			goto walk;
 		pte |= PT_DIRTY_MASK;
-		kvm_mmu_pte_write(vcpu, pte_gpa, (u8 *)&pte, sizeof(pte));
 		walker->ptes[walker->level - 1] = pte;
 	}
 
@@ -239,6 +237,8 @@ err:
 		walker->error_code |= PFERR_USER_MASK;
 	if (fetch_fault)
 		walker->error_code |= PFERR_FETCH_MASK;
+	if (rsvd_fault)
+		walker->error_code |= PFERR_RSVD_MASK;
 	return 0;
 }
 
@@ -263,95 +263,89 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page,
 	pfn = vcpu->arch.update_pte.pfn;
 	if (is_error_pfn(pfn))
 		return;
+	if (mmu_notifier_retry(vcpu, vcpu->arch.update_pte.mmu_seq))
+		return;
 	kvm_get_pfn(pfn);
 	mmu_set_spte(vcpu, spte, page->role.access, pte_access, 0, 0,
-		     gpte & PT_DIRTY_MASK, NULL, largepage, gpte_to_gfn(gpte),
-		     pfn, true);
+		     gpte & PT_DIRTY_MASK, NULL, largepage,
+		     gpte_to_gfn(gpte), pfn, true);
 }
 
 /*
  * Fetch a shadow pte for a specific level in the paging hierarchy.
  */
 static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr,
-			 struct guest_walker *walker,
+			 struct guest_walker *gw,
 			 int user_fault, int write_fault, int largepage,
 			 int *ptwrite, pfn_t pfn)
 {
-	hpa_t shadow_addr;
+	unsigned access = gw->pt_access;
+	struct kvm_mmu_page *shadow_page;
+	u64 spte, *sptep;
+	int direct;
+	gfn_t table_gfn;
+	int r;
 	int level;
-	u64 *shadow_ent;
-	unsigned access = walker->pt_access;
+	pt_element_t curr_pte;
+	struct kvm_shadow_walk_iterator iterator;
 
-	if (!is_present_pte(walker->ptes[walker->level - 1]))
+	if (!is_present_pte(gw->ptes[gw->level - 1]))
 		return NULL;
 
-	shadow_addr = vcpu->arch.mmu.root_hpa;
-	level = vcpu->arch.mmu.shadow_root_level;
-	if (level == PT32E_ROOT_LEVEL) {
-		shadow_addr = vcpu->arch.mmu.pae_root[(addr >> 30) & 3];
-		shadow_addr &= PT64_BASE_ADDR_MASK;
-		--level;
-	}
-
-	for (; ; level--) {
-		u32 index = SHADOW_PT_INDEX(addr, level);
-		struct kvm_mmu_page *shadow_page;
-		u64 shadow_pte;
-		int metaphysical;
-		gfn_t table_gfn;
-
-		shadow_ent = ((u64 *)__va(shadow_addr)) + index;
-		if (level == PT_PAGE_TABLE_LEVEL)
-			break;
-
-		if (largepage && level == PT_DIRECTORY_LEVEL)
+	for_each_shadow_entry(vcpu, addr, iterator) {
+		level = iterator.level;
+		sptep = iterator.sptep;
+		if (level == PT_PAGE_TABLE_LEVEL
+		    || (largepage && level == PT_DIRECTORY_LEVEL)) {
+			mmu_set_spte(vcpu, sptep, access,
+				     gw->pte_access & access,
+				     user_fault, write_fault,
+				     gw->ptes[gw->level-1] & PT_DIRTY_MASK,
+				     ptwrite, largepage,
+				     gw->gfn, pfn, false);
 			break;
+		}
 
-		if (is_shadow_present_pte(*shadow_ent)
-		    && !is_large_pte(*shadow_ent)) {
-			shadow_addr = *shadow_ent & PT64_BASE_ADDR_MASK;
+		if (is_shadow_present_pte(*sptep) && !is_large_pte(*sptep))
 			continue;
-		}
 
-		if (is_large_pte(*shadow_ent))
-			rmap_remove(vcpu->kvm, shadow_ent);
+		if (is_large_pte(*sptep)) {
+			rmap_remove(vcpu->kvm, sptep);
+			set_shadow_pte(sptep, shadow_trap_nonpresent_pte);
+			kvm_flush_remote_tlbs(vcpu->kvm);
+		}
 
-		if (level - 1 == PT_PAGE_TABLE_LEVEL
-		    && walker->level == PT_DIRECTORY_LEVEL) {
-			metaphysical = 1;
-			if (!is_dirty_pte(walker->ptes[level - 1]))
+		if (level == PT_DIRECTORY_LEVEL
+		    && gw->level == PT_DIRECTORY_LEVEL) {
+			direct = 1;
+			if (!is_dirty_pte(gw->ptes[level - 1]))
 				access &= ~ACC_WRITE_MASK;
-			table_gfn = gpte_to_gfn(walker->ptes[level - 1]);
+			table_gfn = gpte_to_gfn(gw->ptes[level - 1]);
 		} else {
-			metaphysical = 0;
-			table_gfn = walker->table_gfn[level - 2];
+			direct = 0;
+			table_gfn = gw->table_gfn[level - 2];
 		}
 		shadow_page = kvm_mmu_get_page(vcpu, table_gfn, addr, level-1,
-					       metaphysical, access,
-					       shadow_ent);
-		if (!metaphysical) {
-			int r;
-			pt_element_t curr_pte;
+					       direct, access, sptep);
+		if (!direct) {
 			r = kvm_read_guest_atomic(vcpu->kvm,
-						  walker->pte_gpa[level - 2],
+						  gw->pte_gpa[level - 2],
 						  &curr_pte, sizeof(curr_pte));
-			if (r || curr_pte != walker->ptes[level - 2]) {
+			if (r || curr_pte != gw->ptes[level - 2]) {
+				kvm_mmu_put_page(shadow_page, sptep);
 				kvm_release_pfn_clean(pfn);
-				return NULL;
+				sptep = NULL;
+				break;
 			}
 		}
-		shadow_addr = __pa(shadow_page->spt);
-		shadow_pte = shadow_addr | PT_PRESENT_MASK | PT_ACCESSED_MASK
+
+		spte = __pa(shadow_page->spt)
+			| PT_PRESENT_MASK | PT_ACCESSED_MASK
 			| PT_WRITABLE_MASK | PT_USER_MASK;
-		*shadow_ent = shadow_pte;
+		*sptep = spte;
 	}
 
-	mmu_set_spte(vcpu, shadow_ent, access, walker->pte_access & access,
-		     user_fault, write_fault,
-		     walker->ptes[walker->level-1] & PT_DIRTY_MASK,
-		     ptwrite, largepage, walker->gfn, pfn, false);
-
-	return shadow_ent;
+	return sptep;
 }
 
 /*
@@ -380,6 +374,7 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, gva_t addr,
 	int r;
 	pfn_t pfn;
 	int largepage = 0;
+	unsigned long mmu_seq;
 
 	pgprintk("%s: addr %lx err %x\n", __func__, addr, error_code);
 	kvm_mmu_audit(vcpu, "pre page fault");
@@ -389,7 +384,7 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, gva_t addr,
 		return r;
 
 	/*
-	 * Look up the shadow pte for the faulting address.
+	 * Look up the guest pte for the faulting address.
 	 */
 	r = FNAME(walk_addr)(&walker, vcpu, addr, write_fault, user_fault,
 			     fetch_fault);
@@ -404,7 +399,6 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, gva_t addr,
 		return 0;
 	}
 
-	down_read(&current->mm->mmap_sem);
 	if (walker.level == PT_DIRECTORY_LEVEL) {
 		gfn_t large_gfn;
 		large_gfn = walker.gfn & ~(KVM_PAGES_PER_HPAGE-1);
@@ -413,17 +407,20 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, gva_t addr,
 			largepage = 1;
 		}
 	}
+	mmu_seq = vcpu->kvm->mmu_notifier_seq;
+	smp_rmb();
 	pfn = gfn_to_pfn(vcpu->kvm, walker.gfn);
-	up_read(&current->mm->mmap_sem);
 
 	/* mmio */
 	if (is_error_pfn(pfn)) {
-		pgprintk("gfn %x is mmio\n", walker.gfn);
+		pgprintk("gfn %lx is mmio\n", walker.gfn);
 		kvm_release_pfn_clean(pfn);
 		return 1;
 	}
 
 	spin_lock(&vcpu->kvm->mmu_lock);
+	if (mmu_notifier_retry(vcpu, mmu_seq))
+		goto out_unlock;
 	kvm_mmu_free_some_pages(vcpu);
 	shadow_pte = FNAME(fetch)(vcpu, addr, &walker, user_fault, write_fault,
 				  largepage, &write_pt, pfn);
@@ -439,6 +436,65 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, gva_t addr,
 	spin_unlock(&vcpu->kvm->mmu_lock);
 
 	return write_pt;
+
+out_unlock:
+	spin_unlock(&vcpu->kvm->mmu_lock);
+	kvm_release_pfn_clean(pfn);
+	return 0;
+}
+
+static void FNAME(invlpg)(struct kvm_vcpu *vcpu, gva_t gva)
+{
+	struct kvm_shadow_walk_iterator iterator;
+	pt_element_t gpte;
+	gpa_t pte_gpa = -1;
+	int level;
+	u64 *sptep;
+	int need_flush = 0;
+
+	spin_lock(&vcpu->kvm->mmu_lock);
+
+	for_each_shadow_entry(vcpu, gva, iterator) {
+		level = iterator.level;
+		sptep = iterator.sptep;
+
+		/* FIXME: properly handle invlpg on large guest pages */
+		if (level == PT_PAGE_TABLE_LEVEL ||
+		    ((level == PT_DIRECTORY_LEVEL) && is_large_pte(*sptep))) {
+			struct kvm_mmu_page *sp = page_header(__pa(sptep));
+
+			pte_gpa = (sp->gfn << PAGE_SHIFT);
+			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;
+				need_flush = 1;
+			}
+			set_shadow_pte(sptep, shadow_trap_nonpresent_pte);
+			break;
+		}
+
+		if (!is_shadow_present_pte(*sptep))
+			break;
+	}
+
+	if (need_flush)
+		kvm_flush_remote_tlbs(vcpu->kvm);
+	spin_unlock(&vcpu->kvm->mmu_lock);
+
+	if (pte_gpa == -1)
+		return;
+	if (kvm_read_guest_atomic(vcpu->kvm, pte_gpa, &gpte,
+				  sizeof(pt_element_t)))
+		return;
+	if (is_present_pte(gpte) && (gpte & PT_ACCESSED_MASK)) {
+		if (mmu_topup_memory_caches(vcpu))
+			return;
+		kvm_mmu_pte_write(vcpu, pte_gpa, (const u8 *)&gpte,
+				  sizeof(pt_element_t), 0);
+	}
 }
 
 static gpa_t FNAME(gva_to_gpa)(struct kvm_vcpu *vcpu, gva_t vaddr)
@@ -460,29 +516,85 @@ static gpa_t FNAME(gva_to_gpa)(struct kvm_vcpu *vcpu, gva_t vaddr)
 static void FNAME(prefetch_page)(struct kvm_vcpu *vcpu,
 				 struct kvm_mmu_page *sp)
 {
-	int i, offset = 0, r = 0;
-	pt_element_t pt;
+	int i, j, offset, r;
+	pt_element_t pt[256 / sizeof(pt_element_t)];
+	gpa_t pte_gpa;
 
-	if (sp->role.metaphysical
+	if (sp->role.direct
 	    || (PTTYPE == 32 && sp->role.level > PT_PAGE_TABLE_LEVEL)) {
 		nonpaging_prefetch_page(vcpu, sp);
 		return;
 	}
 
+	pte_gpa = gfn_to_gpa(sp->gfn);
+	if (PTTYPE == 32) {
+		offset = sp->role.quadrant << PT64_LEVEL_BITS;
+		pte_gpa += offset * sizeof(pt_element_t);
+	}
+
+	for (i = 0; i < PT64_ENT_PER_PAGE; i += ARRAY_SIZE(pt)) {
+		r = kvm_read_guest_atomic(vcpu->kvm, pte_gpa, pt, sizeof pt);
+		pte_gpa += ARRAY_SIZE(pt) * sizeof(pt_element_t);
+		for (j = 0; j < ARRAY_SIZE(pt); ++j)
+			if (r || is_present_pte(pt[j]))
+				sp->spt[i+j] = shadow_trap_nonpresent_pte;
+			else
+				sp->spt[i+j] = shadow_notrap_nonpresent_pte;
+	}
+}
+
+/*
+ * Using the cached information from sp->gfns is safe because:
+ * - The spte has a reference to the struct page, so the pfn for a given gfn
+ *   can't change unless all sptes pointing to it are nuked first.
+ * - Alias changes zap the entire shadow cache.
+ */
+static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
+{
+	int i, offset, nr_present;
+
+	offset = nr_present = 0;
+
 	if (PTTYPE == 32)
 		offset = sp->role.quadrant << PT64_LEVEL_BITS;
 
-	for (i = 0; i < PT64_ENT_PER_PAGE; ++i) {
-		gpa_t pte_gpa = gfn_to_gpa(sp->gfn);
+	for (i = 0; i < PT64_ENT_PER_PAGE; i++) {
+		unsigned pte_access;
+		pt_element_t gpte;
+		gpa_t pte_gpa;
+		gfn_t gfn = sp->gfns[i];
+
+		if (!is_shadow_present_pte(sp->spt[i]))
+			continue;
+
+		pte_gpa = gfn_to_gpa(sp->gfn);
 		pte_gpa += (i+offset) * sizeof(pt_element_t);
 
-		r = kvm_read_guest_atomic(vcpu->kvm, pte_gpa, &pt,
-					  sizeof(pt_element_t));
-		if (r || is_present_pte(pt))
-			sp->spt[i] = shadow_trap_nonpresent_pte;
-		else
-			sp->spt[i] = shadow_notrap_nonpresent_pte;
+		if (kvm_read_guest_atomic(vcpu->kvm, pte_gpa, &gpte,
+					  sizeof(pt_element_t)))
+			return -EINVAL;
+
+		if (gpte_to_gfn(gpte) != gfn || !is_present_pte(gpte) ||
+		    !(gpte & PT_ACCESSED_MASK)) {
+			u64 nonpresent;
+
+			rmap_remove(vcpu->kvm, &sp->spt[i]);
+			if (is_present_pte(gpte))
+				nonpresent = shadow_trap_nonpresent_pte;
+			else
+				nonpresent = shadow_notrap_nonpresent_pte;
+			set_shadow_pte(&sp->spt[i], nonpresent);
+			continue;
+		}
+
+		nr_present++;
+		pte_access = sp->role.access & FNAME(gpte_access)(vcpu, gpte);
+		set_spte(vcpu, &sp->spt[i], pte_access, 0, 0,
+			 is_dirty_pte(gpte), 0, gfn,
+			 spte_to_pfn(sp->spt[i]), true, false);
 	}
+
+	return !nr_present;
 }
 
 #undef pt_element_t
@@ -490,7 +602,6 @@ static void FNAME(prefetch_page)(struct kvm_vcpu *vcpu,
 #undef FNAME
 #undef PT_BASE_ADDR_MASK
 #undef PT_INDEX
-#undef SHADOW_PT_INDEX
 #undef PT_LEVEL_MASK
 #undef PT_DIR_BASE_ADDR_MASK
 #undef PT_LEVEL_BITS