X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=arch%2Fx86%2Fkvm%2Fpaging_tmpl.h;h=258e4591e1ca02246801fc380ae19472b2375b0a;hb=c2d0ee46e6e633a3c23ecbcb9b03ad731906cd79;hp=dc169e8148b12d404b6adae13e04dfa82e4d7904;hpb=a7052897b3bcd568a9f5bfaa558957039e7e7ec0;p=safe%2Fjmp%2Flinux-2.6

diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h
index dc169e8..258e459 100644
--- a/arch/x86/kvm/paging_tmpl.h
+++ b/arch/x86/kvm/paging_tmpl.h
@@ -25,7 +25,6 @@
 #if PTTYPE == 64
 	#define pt_element_t u64
 	#define guest_walker guest_walker64
-	#define shadow_walker shadow_walker64
 	#define FNAME(name) paging##64_##name
 	#define PT_BASE_ADDR_MASK PT64_BASE_ADDR_MASK
 	#define PT_DIR_BASE_ADDR_MASK PT64_DIR_BASE_ADDR_MASK
@@ -42,7 +41,6 @@
 #elif PTTYPE == 32
 	#define pt_element_t u32
 	#define guest_walker guest_walker32
-	#define shadow_walker shadow_walker32
 	#define FNAME(name) paging##32_##name
 	#define PT_BASE_ADDR_MASK PT32_BASE_ADDR_MASK
 	#define PT_DIR_BASE_ADDR_MASK PT32_DIR_BASE_ADDR_MASK
@@ -73,17 +71,6 @@ struct guest_walker {
 	u32 error_code;
 };
 
-struct shadow_walker {
-	struct kvm_shadow_walk walker;
-	struct guest_walker *guest_walker;
-	int user_fault;
-	int write_fault;
-	int largepage;
-	int *ptwrite;
-	pfn_t pfn;
-	u64 *sptep;
-};
-
 static gfn_t gpte_to_gfn(pt_element_t gpte)
 {
 	return (gpte & PT_BASE_ADDR_MASK) >> PAGE_SHIFT;
@@ -136,6 +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:
@@ -170,6 +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;
@@ -222,7 +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;
 	}
 
@@ -246,6 +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;
 }
 
@@ -274,96 +267,85 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page,
 		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 int FNAME(shadow_walk_entry)(struct kvm_shadow_walk *_sw,
-				    struct kvm_vcpu *vcpu, u64 addr,
-				    u64 *sptep, int level)
+static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr,
+			 struct guest_walker *gw,
+			 int user_fault, int write_fault, int largepage,
+			 int *ptwrite, pfn_t pfn)
 {
-	struct shadow_walker *sw =
-		container_of(_sw, struct shadow_walker, walker);
-	struct guest_walker *gw = sw->guest_walker;
 	unsigned access = gw->pt_access;
 	struct kvm_mmu_page *shadow_page;
-	u64 spte;
-	int metaphysical;
+	u64 spte, *sptep;
+	int direct;
 	gfn_t table_gfn;
 	int r;
+	int level;
 	pt_element_t curr_pte;
+	struct kvm_shadow_walk_iterator iterator;
 
-	if (level == PT_PAGE_TABLE_LEVEL
-	    || (sw->largepage && level == PT_DIRECTORY_LEVEL)) {
-		mmu_set_spte(vcpu, sptep, access, gw->pte_access & access,
-			     sw->user_fault, sw->write_fault,
-			     gw->ptes[gw->level-1] & PT_DIRTY_MASK,
-			     sw->ptwrite, sw->largepage, gw->gfn, sw->pfn,
-			     false);
-		sw->sptep = sptep;
-		return 1;
-	}
+	if (!is_present_pte(gw->ptes[gw->level - 1]))
+		return NULL;
 
-	if (is_shadow_present_pte(*sptep) && !is_large_pte(*sptep))
-		return 0;
+	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_large_pte(*sptep)) {
-		set_shadow_pte(sptep, shadow_trap_nonpresent_pte);
-		kvm_flush_remote_tlbs(vcpu->kvm);
-		rmap_remove(vcpu->kvm, sptep);
-	}
+		if (is_shadow_present_pte(*sptep) && !is_large_pte(*sptep))
+			continue;
 
-	if (level == PT_DIRECTORY_LEVEL && gw->level == PT_DIRECTORY_LEVEL) {
-		metaphysical = 1;
-		if (!is_dirty_pte(gw->ptes[level - 1]))
-			access &= ~ACC_WRITE_MASK;
-		table_gfn = gpte_to_gfn(gw->ptes[level - 1]);
-	} else {
-		metaphysical = 0;
-		table_gfn = gw->table_gfn[level - 2];
-	}
-	shadow_page = kvm_mmu_get_page(vcpu, table_gfn, (gva_t)addr, level-1,
-				       metaphysical, access, sptep);
-	if (!metaphysical) {
-		r = kvm_read_guest_atomic(vcpu->kvm, gw->pte_gpa[level - 2],
-					  &curr_pte, sizeof(curr_pte));
-		if (r || curr_pte != gw->ptes[level - 2]) {
-			kvm_release_pfn_clean(sw->pfn);
-			sw->sptep = NULL;
-			return 1;
+		if (is_large_pte(*sptep)) {
+			rmap_remove(vcpu->kvm, sptep);
+			set_shadow_pte(sptep, shadow_trap_nonpresent_pte);
+			kvm_flush_remote_tlbs(vcpu->kvm);
 		}
-	}
-
-	spte = __pa(shadow_page->spt) | PT_PRESENT_MASK | PT_ACCESSED_MASK
-		| PT_WRITABLE_MASK | PT_USER_MASK;
-	*sptep = spte;
-	return 0;
-}
 
-static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr,
-			 struct guest_walker *guest_walker,
-			 int user_fault, int write_fault, int largepage,
-			 int *ptwrite, pfn_t pfn)
-{
-	struct shadow_walker walker = {
-		.walker = { .entry = FNAME(shadow_walk_entry), },
-		.guest_walker = guest_walker,
-		.user_fault = user_fault,
-		.write_fault = write_fault,
-		.largepage = largepage,
-		.ptwrite = ptwrite,
-		.pfn = pfn,
-	};
-
-	if (!is_present_pte(guest_walker->ptes[guest_walker->level - 1]))
-		return NULL;
+		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(gw->ptes[level - 1]);
+		} else {
+			direct = 0;
+			table_gfn = gw->table_gfn[level - 2];
+		}
+		shadow_page = kvm_mmu_get_page(vcpu, table_gfn, addr, level-1,
+					       direct, access, sptep);
+		if (!direct) {
+			r = kvm_read_guest_atomic(vcpu->kvm,
+						  gw->pte_gpa[level - 2],
+						  &curr_pte, sizeof(curr_pte));
+			if (r || curr_pte != gw->ptes[level - 2]) {
+				kvm_mmu_put_page(shadow_page, sptep);
+				kvm_release_pfn_clean(pfn);
+				sptep = NULL;
+				break;
+			}
+		}
 
-	walk_shadow(&walker.walker, vcpu, addr);
+		spte = __pa(shadow_page->spt)
+			| PT_PRESENT_MASK | PT_ACCESSED_MASK
+			| PT_WRITABLE_MASK | PT_USER_MASK;
+		*sptep = spte;
+	}
 
-	return walker.sptep;
+	return sptep;
 }
 
 /*
@@ -402,7 +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);
@@ -461,29 +443,58 @@ out_unlock:
 	return 0;
 }
 
-static int FNAME(shadow_invlpg_entry)(struct kvm_shadow_walk *_sw,
-				      struct kvm_vcpu *vcpu, u64 addr,
-				      u64 *sptep, int level)
+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;
 
-	if (level == PT_PAGE_TABLE_LEVEL) {
-		if (is_shadow_present_pte(*sptep))
-			rmap_remove(vcpu->kvm, sptep);
-		set_shadow_pte(sptep, shadow_trap_nonpresent_pte);
-		return 1;
+	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 (!is_shadow_present_pte(*sptep))
-		return 1;
-	return 0;
-}
 
-static void FNAME(invlpg)(struct kvm_vcpu *vcpu, gva_t gva)
-{
-	struct shadow_walker walker = {
-		.walker = { .entry = FNAME(shadow_invlpg_entry), },
-	};
+	if (need_flush)
+		kvm_flush_remote_tlbs(vcpu->kvm);
+	spin_unlock(&vcpu->kvm->mmu_lock);
 
-	walk_shadow(&walker.walker, vcpu, gva);
+	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)
@@ -509,7 +520,7 @@ static void FNAME(prefetch_page)(struct kvm_vcpu *vcpu,
 	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;
@@ -580,7 +591,7 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 		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);
+			 spte_to_pfn(sp->spt[i]), true, false);
 	}
 
 	return !nr_present;
@@ -588,7 +599,6 @@ static int FNAME(sync_page)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *sp)
 
 #undef pt_element_t
 #undef guest_walker
-#undef shadow_walker
 #undef FNAME
 #undef PT_BASE_ADDR_MASK
 #undef PT_INDEX