1 files changed, 102 insertions, 82 deletions

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 7a69cf053711..f551962ac294 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -42,6 +42,7 @@
 #include <linux/kern_levels.h>
 
 #include <asm/page.h>
+#include <asm/pat.h>
 #include <asm/cmpxchg.h>
 #include <asm/io.h>
 #include <asm/vmx.h>
@@ -150,6 +151,20 @@ module_param(dbg, bool, 0644);
 /* make pte_list_desc fit well in cache line */
 #define PTE_LIST_EXT 3
 
+/*
+ * Return values of handle_mmio_page_fault and mmu.page_fault:
+ * RET_PF_RETRY: let CPU fault again on the address.
+ * RET_PF_EMULATE: mmio page fault, emulate the instruction directly.
+ *
+ * For handle_mmio_page_fault only:
+ * RET_PF_INVALID: the spte is invalid, let the real page fault path update it.
+ */
+enum {
+	RET_PF_RETRY = 0,
+	RET_PF_EMULATE = 1,
+	RET_PF_INVALID = 2,
+};
+
 struct pte_list_desc {
 	u64 *sptes[PTE_LIST_EXT];
 	struct pte_list_desc *more;
@@ -367,7 +382,7 @@ void kvm_mmu_set_mask_ptes(u64 user_mask, u64 accessed_mask,
 }
 EXPORT_SYMBOL_GPL(kvm_mmu_set_mask_ptes);
 
-void kvm_mmu_clear_all_pte_masks(void)
+static void kvm_mmu_clear_all_pte_masks(void)
 {
 	shadow_user_mask = 0;
 	shadow_accessed_mask = 0;
@@ -443,7 +458,7 @@ static u64 __update_clear_spte_slow(u64 *sptep, u64 spte)
 
 static u64 __get_spte_lockless(u64 *sptep)
 {
-	return ACCESS_ONCE(*sptep);
+	return READ_ONCE(*sptep);
 }
 #else
 union split_spte {
@@ -2424,7 +2439,7 @@ static void __shadow_walk_next(struct kvm_shadow_walk_iterator *iterator,
 
 static void shadow_walk_next(struct kvm_shadow_walk_iterator *iterator)
 {
-	return __shadow_walk_next(iterator, *iterator->sptep);
+	__shadow_walk_next(iterator, *iterator->sptep);
 }
 
 static void link_shadow_page(struct kvm_vcpu *vcpu, u64 *sptep,
@@ -2694,7 +2709,18 @@ static bool mmu_need_write_protect(struct kvm_vcpu *vcpu, gfn_t gfn,
 static bool kvm_is_mmio_pfn(kvm_pfn_t pfn)
 {
 	if (pfn_valid(pfn))
-		return !is_zero_pfn(pfn) && PageReserved(pfn_to_page(pfn));
+		return !is_zero_pfn(pfn) && PageReserved(pfn_to_page(pfn)) &&
+			/*
+			 * Some reserved pages, such as those from NVDIMM
+			 * DAX devices, are not for MMIO, and can be mapped
+			 * with cached memory type for better performance.
+			 * However, the above check misconceives those pages
+			 * as MMIO, and results in KVM mapping them with UC
+			 * memory type, which would hurt the performance.
+			 * Therefore, we check the host memory type in addition
+			 * and only treat UC/UC-/WC pages as MMIO.
+			 */
+			(!pat_enabled() || pat_pfn_immune_to_uc_mtrr(pfn));
 
 	return true;
 }
@@ -2794,13 +2820,13 @@ done:
 	return ret;
 }
 
-static bool mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep, unsigned pte_access,
-			 int write_fault, int level, gfn_t gfn, kvm_pfn_t pfn,
-			 bool speculative, bool host_writable)
+static int mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep, unsigned pte_access,
+			int write_fault, int level, gfn_t gfn, kvm_pfn_t pfn,
+		       	bool speculative, bool host_writable)
 {
 	int was_rmapped = 0;
 	int rmap_count;
-	bool emulate = false;
+	int ret = RET_PF_RETRY;
 
 	pgprintk("%s: spte %llx write_fault %d gfn %llx\n", __func__,
 		 *sptep, write_fault, gfn);
@@ -2830,12 +2856,12 @@ static bool mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep, unsigned pte_access,
 	if (set_spte(vcpu, sptep, pte_access, level, gfn, pfn, speculative,
 	      true, host_writable)) {
 		if (write_fault)
-			emulate = true;
+			ret = RET_PF_EMULATE;
 		kvm_make_request(KVM_REQ_TLB_FLUSH, vcpu);
 	}
 
 	if (unlikely(is_mmio_spte(*sptep)))
-		emulate = true;
+		ret = RET_PF_EMULATE;
 
 	pgprintk("%s: setting spte %llx\n", __func__, *sptep);
 	pgprintk("instantiating %s PTE (%s) at %llx (%llx) addr %p\n",
@@ -2855,7 +2881,7 @@ static bool mmu_set_spte(struct kvm_vcpu *vcpu, u64 *sptep, unsigned pte_access,
 
 	kvm_release_pfn_clean(pfn);
 
-	return emulate;
+	return ret;
 }
 
 static kvm_pfn_t pte_prefetch_gfn_to_pfn(struct kvm_vcpu *vcpu, gfn_t gfn,
@@ -2994,17 +3020,16 @@ static int kvm_handle_bad_page(struct kvm_vcpu *vcpu, gfn_t gfn, kvm_pfn_t pfn)
 	 * Do not cache the mmio info caused by writing the readonly gfn
 	 * into the spte otherwise read access on readonly gfn also can
 	 * caused mmio page fault and treat it as mmio access.
-	 * Return 1 to tell kvm to emulate it.
 	 */
 	if (pfn == KVM_PFN_ERR_RO_FAULT)
-		return 1;
+		return RET_PF_EMULATE;
 
 	if (pfn == KVM_PFN_ERR_HWPOISON) {
 		kvm_send_hwpoison_signal(kvm_vcpu_gfn_to_hva(vcpu, gfn), current);
-		return 0;
+		return RET_PF_RETRY;
 	}
 
-	return -EFAULT;
+	return RET_PF_EMULATE;
 }
 
 static void transparent_hugepage_adjust(struct kvm_vcpu *vcpu,
@@ -3286,13 +3311,13 @@ static int nonpaging_map(struct kvm_vcpu *vcpu, gva_t v, u32 error_code,
 	}
 
 	if (fast_page_fault(vcpu, v, level, error_code))
-		return 0;
+		return RET_PF_RETRY;
 
 	mmu_seq = vcpu->kvm->mmu_notifier_seq;
 	smp_rmb();
 
 	if (try_async_pf(vcpu, prefault, gfn, v, &pfn, write, &map_writable))
-		return 0;
+		return RET_PF_RETRY;
 
 	if (handle_abnormal_pfn(vcpu, v, gfn, pfn, ACC_ALL, &r))
 		return r;
@@ -3312,7 +3337,7 @@ static int nonpaging_map(struct kvm_vcpu *vcpu, gva_t v, u32 error_code,
 out_unlock:
 	spin_unlock(&vcpu->kvm->mmu_lock);
 	kvm_release_pfn_clean(pfn);
-	return 0;
+	return RET_PF_RETRY;
 }
 
 
@@ -3382,7 +3407,7 @@ static int mmu_alloc_direct_roots(struct kvm_vcpu *vcpu)
 		spin_lock(&vcpu->kvm->mmu_lock);
 		if(make_mmu_pages_available(vcpu) < 0) {
 			spin_unlock(&vcpu->kvm->mmu_lock);
-			return 1;
+			return -ENOSPC;
 		}
 		sp = kvm_mmu_get_page(vcpu, 0, 0,
 				vcpu->arch.mmu.shadow_root_level, 1, ACC_ALL);
@@ -3397,7 +3422,7 @@ static int mmu_alloc_direct_roots(struct kvm_vcpu *vcpu)
 			spin_lock(&vcpu->kvm->mmu_lock);
 			if (make_mmu_pages_available(vcpu) < 0) {
 				spin_unlock(&vcpu->kvm->mmu_lock);
-				return 1;
+				return -ENOSPC;
 			}
 			sp = kvm_mmu_get_page(vcpu, i << (30 - PAGE_SHIFT),
 					i << 30, PT32_ROOT_LEVEL, 1, ACC_ALL);
@@ -3437,7 +3462,7 @@ static int mmu_alloc_shadow_roots(struct kvm_vcpu *vcpu)
 		spin_lock(&vcpu->kvm->mmu_lock);
 		if (make_mmu_pages_available(vcpu) < 0) {
 			spin_unlock(&vcpu->kvm->mmu_lock);
-			return 1;
+			return -ENOSPC;
 		}
 		sp = kvm_mmu_get_page(vcpu, root_gfn, 0,
 				vcpu->arch.mmu.shadow_root_level, 0, ACC_ALL);
@@ -3474,7 +3499,7 @@ static int mmu_alloc_shadow_roots(struct kvm_vcpu *vcpu)
 		spin_lock(&vcpu->kvm->mmu_lock);
 		if (make_mmu_pages_available(vcpu) < 0) {
 			spin_unlock(&vcpu->kvm->mmu_lock);
-			return 1;
+			return -ENOSPC;
 		}
 		sp = kvm_mmu_get_page(vcpu, root_gfn, i << 30, PT32_ROOT_LEVEL,
 				      0, ACC_ALL);
@@ -3659,54 +3684,38 @@ exit:
 	return reserved;
 }
 
-/*
- * Return values of handle_mmio_page_fault:
- * RET_MMIO_PF_EMULATE: it is a real mmio page fault, emulate the instruction
- *			directly.
- * RET_MMIO_PF_INVALID: invalid spte is detected then let the real page
- *			fault path update the mmio spte.
- * RET_MMIO_PF_RETRY: let CPU fault again on the address.
- * RET_MMIO_PF_BUG: a bug was detected (and a WARN was printed).
- */
-enum {
-	RET_MMIO_PF_EMULATE = 1,
-	RET_MMIO_PF_INVALID = 2,
-	RET_MMIO_PF_RETRY = 0,
-	RET_MMIO_PF_BUG = -1
-};
-
 static int handle_mmio_page_fault(struct kvm_vcpu *vcpu, u64 addr, bool direct)
 {
 	u64 spte;
 	bool reserved;
 
 	if (mmio_info_in_cache(vcpu, addr, direct))
-		return RET_MMIO_PF_EMULATE;
+		return RET_PF_EMULATE;
 
 	reserved = walk_shadow_page_get_mmio_spte(vcpu, addr, &spte);
 	if (WARN_ON(reserved))
-		return RET_MMIO_PF_BUG;
+		return -EINVAL;
 
 	if (is_mmio_spte(spte)) {
 		gfn_t gfn = get_mmio_spte_gfn(spte);
 		unsigned access = get_mmio_spte_access(spte);
 
 		if (!check_mmio_spte(vcpu, spte))
-			return RET_MMIO_PF_INVALID;
+			return RET_PF_INVALID;
 
 		if (direct)
 			addr = 0;
 
 		trace_handle_mmio_page_fault(addr, gfn, access);
 		vcpu_cache_mmio_info(vcpu, addr, gfn, access);
-		return RET_MMIO_PF_EMULATE;
+		return RET_PF_EMULATE;
 	}
 
 	/*
 	 * If the page table is zapped by other cpus, let CPU fault again on
 	 * the address.
 	 */
-	return RET_MMIO_PF_RETRY;
+	return RET_PF_RETRY;
 }
 EXPORT_SYMBOL_GPL(handle_mmio_page_fault);
 
@@ -3756,7 +3765,7 @@ static int nonpaging_page_fault(struct kvm_vcpu *vcpu, gva_t gva,
 	pgprintk("%s: gva %lx error %x\n", __func__, gva, error_code);
 
 	if (page_fault_handle_page_track(vcpu, error_code, gfn))
-		return 1;
+		return RET_PF_EMULATE;
 
 	r = mmu_topup_memory_caches(vcpu);
 	if (r)
@@ -3784,7 +3793,8 @@ static int kvm_arch_setup_async_pf(struct kvm_vcpu *vcpu, gva_t gva, gfn_t gfn)
 bool kvm_can_do_async_pf(struct kvm_vcpu *vcpu)
 {
 	if (unlikely(!lapic_in_kernel(vcpu) ||
-		     kvm_event_needs_reinjection(vcpu)))
+		     kvm_event_needs_reinjection(vcpu) ||
+		     vcpu->arch.exception.pending))
 		return false;
 
 	if (!vcpu->arch.apf.delivery_as_pf_vmexit && is_guest_mode(vcpu))
@@ -3820,8 +3830,7 @@ static bool try_async_pf(struct kvm_vcpu *vcpu, bool prefault, gfn_t gfn,
 }
 
 int kvm_handle_page_fault(struct kvm_vcpu *vcpu, u64 error_code,
-				u64 fault_address, char *insn, int insn_len,
-				bool need_unprotect)
+				u64 fault_address, char *insn, int insn_len)
 {
 	int r = 1;
 
@@ -3829,7 +3838,7 @@ int kvm_handle_page_fault(struct kvm_vcpu *vcpu, u64 error_code,
 	default:
 		trace_kvm_page_fault(fault_address, error_code);
 
-		if (need_unprotect && kvm_event_needs_reinjection(vcpu))
+		if (kvm_event_needs_reinjection(vcpu))
 			kvm_mmu_unprotect_page_virt(vcpu, fault_address);
 		r = kvm_mmu_page_fault(vcpu, fault_address, error_code, insn,
 				insn_len);
@@ -3876,7 +3885,7 @@ static int tdp_page_fault(struct kvm_vcpu *vcpu, gva_t gpa, u32 error_code,
 	MMU_WARN_ON(!VALID_PAGE(vcpu->arch.mmu.root_hpa));
 
 	if (page_fault_handle_page_track(vcpu, error_code, gfn))
-		return 1;
+		return RET_PF_EMULATE;
 
 	r = mmu_topup_memory_caches(vcpu);
 	if (r)
@@ -3893,13 +3902,13 @@ static int tdp_page_fault(struct kvm_vcpu *vcpu, gva_t gpa, u32 error_code,
 	}
 
 	if (fast_page_fault(vcpu, gpa, level, error_code))
-		return 0;
+		return RET_PF_RETRY;
 
 	mmu_seq = vcpu->kvm->mmu_notifier_seq;
 	smp_rmb();
 
 	if (try_async_pf(vcpu, prefault, gfn, gpa, &pfn, write, &map_writable))
-		return 0;
+		return RET_PF_RETRY;
 
 	if (handle_abnormal_pfn(vcpu, 0, gfn, pfn, ACC_ALL, &r))
 		return r;
@@ -3919,7 +3928,7 @@ static int tdp_page_fault(struct kvm_vcpu *vcpu, gva_t gpa, u32 error_code,
 out_unlock:
 	spin_unlock(&vcpu->kvm->mmu_lock);
 	kvm_release_pfn_clean(pfn);
-	return 0;
+	return RET_PF_RETRY;
 }
 
 static void nonpaging_init_context(struct kvm_vcpu *vcpu,
@@ -4819,7 +4828,7 @@ static void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa,
 	 * If we don't have indirect shadow pages, it means no page is
 	 * write-protected, so we can exit simply.
 	 */
-	if (!ACCESS_ONCE(vcpu->kvm->arch.indirect_shadow_pages))
+	if (!READ_ONCE(vcpu->kvm->arch.indirect_shadow_pages))
 		return;
 
 	remote_flush = local_flush = false;
@@ -4918,25 +4927,25 @@ int kvm_mmu_page_fault(struct kvm_vcpu *vcpu, gva_t cr2, u64 error_code,
 		vcpu->arch.gpa_val = cr2;
 	}
 
+	r = RET_PF_INVALID;
 	if (unlikely(error_code & PFERR_RSVD_MASK)) {
 		r = handle_mmio_page_fault(vcpu, cr2, direct);
-		if (r == RET_MMIO_PF_EMULATE) {
+		if (r == RET_PF_EMULATE) {
 			emulation_type = 0;
 			goto emulate;
 		}
-		if (r == RET_MMIO_PF_RETRY)
-			return 1;
-		if (r < 0)
-			return r;
-		/* Must be RET_MMIO_PF_INVALID.  */
 	}
 
-	r = vcpu->arch.mmu.page_fault(vcpu, cr2, lower_32_bits(error_code),
-				      false);
+	if (r == RET_PF_INVALID) {
+		r = vcpu->arch.mmu.page_fault(vcpu, cr2, lower_32_bits(error_code),
+					      false);
+		WARN_ON(r == RET_PF_INVALID);
+	}
+
+	if (r == RET_PF_RETRY)
+		return 1;
 	if (r < 0)
 		return r;
-	if (!r)
-		return 1;
 
 	/*
 	 * Before emulating the instruction, check if the error code
@@ -4954,6 +4963,16 @@ int kvm_mmu_page_fault(struct kvm_vcpu *vcpu, gva_t cr2, u64 error_code,
 	if (mmio_info_in_cache(vcpu, cr2, direct))
 		emulation_type = 0;
 emulate:
+	/*
+	 * On AMD platforms, under certain conditions insn_len may be zero on #NPF.
+	 * This can happen if a guest gets a page-fault on data access but the HW
+	 * table walker is not able to read the instruction page (e.g instruction
+	 * page is not present in memory). In those cases we simply restart the
+	 * guest.
+	 */
+	if (unlikely(insn && !insn_len))
+		return 1;
+
 	er = x86_emulate_instruction(vcpu, cr2, emulation_type, insn, insn_len);
 
 	switch (er) {
@@ -4993,8 +5012,7 @@ EXPORT_SYMBOL_GPL(kvm_disable_tdp);
 static void free_mmu_pages(struct kvm_vcpu *vcpu)
 {
 	free_page((unsigned long)vcpu->arch.mmu.pae_root);
-	if (vcpu->arch.mmu.lm_root != NULL)
-		free_page((unsigned long)vcpu->arch.mmu.lm_root);
+	free_page((unsigned long)vcpu->arch.mmu.lm_root);
 }
 
 static int alloc_mmu_pages(struct kvm_vcpu *vcpu)
@@ -5062,7 +5080,7 @@ void kvm_mmu_uninit_vm(struct kvm *kvm)
 typedef bool (*slot_level_handler) (struct kvm *kvm, struct kvm_rmap_head *rmap_head);
 
 /* The caller should hold mmu-lock before calling this function. */
-static bool
+static __always_inline bool
 slot_handle_level_range(struct kvm *kvm, struct kvm_memory_slot *memslot,
 			slot_level_handler fn, int start_level, int end_level,
 			gfn_t start_gfn, gfn_t end_gfn, bool lock_flush_tlb)
@@ -5092,7 +5110,7 @@ slot_handle_level_range(struct kvm *kvm, struct kvm_memory_slot *memslot,
 	return flush;
 }
 
-static bool
+static __always_inline bool
 slot_handle_level(struct kvm *kvm, struct kvm_memory_slot *memslot,
 		  slot_level_handler fn, int start_level, int end_level,
 		  bool lock_flush_tlb)
@@ -5103,7 +5121,7 @@ slot_handle_level(struct kvm *kvm, struct kvm_memory_slot *memslot,
 			lock_flush_tlb);
 }
 
-static bool
+static __always_inline bool
 slot_handle_all_level(struct kvm *kvm, struct kvm_memory_slot *memslot,
 		      slot_level_handler fn, bool lock_flush_tlb)
 {
@@ -5111,7 +5129,7 @@ slot_handle_all_level(struct kvm *kvm, struct kvm_memory_slot *memslot,
 				 PT_MAX_HUGEPAGE_LEVEL, lock_flush_tlb);
 }
 
-static bool
+static __always_inline bool
 slot_handle_large_level(struct kvm *kvm, struct kvm_memory_slot *memslot,
 			slot_level_handler fn, bool lock_flush_tlb)
 {
@@ -5119,7 +5137,7 @@ slot_handle_large_level(struct kvm *kvm, struct kvm_memory_slot *memslot,
 				 PT_MAX_HUGEPAGE_LEVEL, lock_flush_tlb);
 }
 
-static bool
+static __always_inline bool
 slot_handle_leaf(struct kvm *kvm, struct kvm_memory_slot *memslot,
 		 slot_level_handler fn, bool lock_flush_tlb)
 {
@@ -5464,38 +5482,40 @@ static struct shrinker mmu_shrinker = {
 
 static void mmu_destroy_caches(void)
 {
-	if (pte_list_desc_cache)
-		kmem_cache_destroy(pte_list_desc_cache);
-	if (mmu_page_header_cache)
-		kmem_cache_destroy(mmu_page_header_cache);
+	kmem_cache_destroy(pte_list_desc_cache);
+	kmem_cache_destroy(mmu_page_header_cache);
 }
 
 int kvm_mmu_module_init(void)
 {
+	int ret = -ENOMEM;
+
 	kvm_mmu_clear_all_pte_masks();
 
 	pte_list_desc_cache = kmem_cache_create("pte_list_desc",
 					    sizeof(struct pte_list_desc),
-					    0, 0, NULL);
+					    0, SLAB_ACCOUNT, NULL);
 	if (!pte_list_desc_cache)
-		goto nomem;
+		goto out;
 
 	mmu_page_header_cache = kmem_cache_create("kvm_mmu_page_header",
 						  sizeof(struct kvm_mmu_page),
-						  0, 0, NULL);
+						  0, SLAB_ACCOUNT, NULL);
 	if (!mmu_page_header_cache)
-		goto nomem;
+		goto out;
 
 	if (percpu_counter_init(&kvm_total_used_mmu_pages, 0, GFP_KERNEL))
-		goto nomem;
+		goto out;
 
-	register_shrinker(&mmu_shrinker);
+	ret = register_shrinker(&mmu_shrinker);
+	if (ret)
+		goto out;
 
 	return 0;
 
-nomem:
+out:
 	mmu_destroy_caches();
-	return -ENOMEM;
+	return ret;
 }
 
 /*