Merge tag 'kvm-x86-mmu-6.4' of https://github.com/kvm-x86/linux into HEAD

KVM x86 MMU changes for 6.4:

 - Tweak FNAME(sync_spte) to avoid unnecessary writes+flushes when the
   guest is only adding new PTEs

 - Overhaul .sync_page() and .invlpg() to share the .sync_page()
   implementation, i.e. utilize .sync_page()'s optimizations when emulating
   invalidations

 - Clean up the range-based flushing APIs

 - Revamp the TDP MMU's reaping of Accessed/Dirty bits to clear a single
   A/D bit using a LOCK AND instead of XCHG, and skip all of the "handle
   changed SPTE" overhead associated with writing the entire entry

 - Track the number of "tail" entries in a pte_list_desc to avoid having
   to walk (potentially) all descriptors during insertion and deletion,
   which gets quite expensive if the guest is spamming fork()

 - Misc cleanups

Author: bonzini

arch/x86/include/asm/kvm-x86-ops.h

@@ -54,8 +54,8 @@ KVM_X86_OP(set_rflags)
 KVM_X86_OP(get_if_flag)
 KVM_X86_OP(flush_tlb_all)
 KVM_X86_OP(flush_tlb_current)
-KVM_X86_OP_OPTIONAL(tlb_remote_flush)
-KVM_X86_OP_OPTIONAL(tlb_remote_flush_with_range)
+KVM_X86_OP_OPTIONAL(flush_remote_tlbs)
+KVM_X86_OP_OPTIONAL(flush_remote_tlbs_range)
 KVM_X86_OP(flush_tlb_gva)
 KVM_X86_OP(flush_tlb_guest)
 KVM_X86_OP(vcpu_pre_run)

arch/x86/include/asm/kvm_host.h

@@ -420,6 +420,10 @@ struct kvm_mmu_root_info {
 
 #define KVM_MMU_NUM_PREV_ROOTS 3
 
+#define KVM_MMU_ROOT_CURRENT		BIT(0)
+#define KVM_MMU_ROOT_PREVIOUS(i)	BIT(1+i)
+#define KVM_MMU_ROOTS_ALL		(BIT(1 + KVM_MMU_NUM_PREV_ROOTS) - 1)
+
 #define KVM_HAVE_MMU_RWLOCK
 
 struct kvm_mmu_page;
@@ -439,9 +443,8 @@ struct kvm_mmu {
 	gpa_t (*gva_to_gpa)(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
 			    gpa_t gva_or_gpa, u64 access,
 			    struct x86_exception *exception);
-	int (*sync_page)(struct kvm_vcpu *vcpu,
-			 struct kvm_mmu_page *sp);
-	void (*invlpg)(struct kvm_vcpu *vcpu, gva_t gva, hpa_t root_hpa);
+	int (*sync_spte)(struct kvm_vcpu *vcpu,
+			 struct kvm_mmu_page *sp, int i);
 	struct kvm_mmu_root_info root;
 	union kvm_cpu_role cpu_role;
 	union kvm_mmu_page_role root_role;
@@ -479,11 +482,6 @@ struct kvm_mmu {
 	u64 pdptrs[4]; /* pae */
 };
 
-struct kvm_tlb_range {
-	u64 start_gfn;
-	u64 pages;
-};
-
 enum pmc_type {
 	KVM_PMC_GP = 0,
 	KVM_PMC_FIXED,
@@ -1585,9 +1583,9 @@ struct kvm_x86_ops {
 
 	void (*flush_tlb_all)(struct kvm_vcpu *vcpu);
 	void (*flush_tlb_current)(struct kvm_vcpu *vcpu);
-	int  (*tlb_remote_flush)(struct kvm *kvm);
-	int  (*tlb_remote_flush_with_range)(struct kvm *kvm,
-			struct kvm_tlb_range *range);
+	int  (*flush_remote_tlbs)(struct kvm *kvm);
+	int  (*flush_remote_tlbs_range)(struct kvm *kvm, gfn_t gfn,
+					gfn_t nr_pages);
 
 	/*
 	 * Flush any TLB entries associated with the given GVA.
@@ -1791,8 +1789,8 @@ void kvm_arch_free_vm(struct kvm *kvm);
 #define __KVM_HAVE_ARCH_FLUSH_REMOTE_TLB
 static inline int kvm_arch_flush_remote_tlb(struct kvm *kvm)
 {
-	if (kvm_x86_ops.tlb_remote_flush &&
-	    !static_call(kvm_x86_tlb_remote_flush)(kvm))
+	if (kvm_x86_ops.flush_remote_tlbs &&
+	    !static_call(kvm_x86_flush_remote_tlbs)(kvm))
 		return 0;
 	else
 		return -ENOTSUPP;
@@ -1997,10 +1995,6 @@ static inline int __kvm_irq_line_state(unsigned long *irq_state,
 	return !!(*irq_state);
 }
 
-#define KVM_MMU_ROOT_CURRENT		BIT(0)
-#define KVM_MMU_ROOT_PREVIOUS(i)	BIT(1+i)
-#define KVM_MMU_ROOTS_ALL		(~0UL)
-
 int kvm_pic_set_irq(struct kvm_pic *pic, int irq, int irq_source_id, int level);
 void kvm_pic_clear_all(struct kvm_pic *pic, int irq_source_id);
 
@@ -2044,8 +2038,8 @@ int kvm_emulate_hypercall(struct kvm_vcpu *vcpu);
 int kvm_mmu_page_fault(struct kvm_vcpu *vcpu, gpa_t cr2_or_gpa, u64 error_code,
 		       void *insn, int insn_len);
 void kvm_mmu_invlpg(struct kvm_vcpu *vcpu, gva_t gva);
-void kvm_mmu_invalidate_gva(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
-			    gva_t gva, hpa_t root_hpa);
+void kvm_mmu_invalidate_addr(struct kvm_vcpu *vcpu, struct kvm_mmu *mmu,
+			     u64 addr, unsigned long roots);
 void kvm_mmu_invpcid_gva(struct kvm_vcpu *vcpu, gva_t gva, unsigned long pcid);
 void kvm_mmu_new_pgd(struct kvm_vcpu *vcpu, gpa_t new_pgd);
 

arch/x86/kvm/kvm_onhyperv.c

@@ -10,17 +10,22 @@
 #include "hyperv.h"
 #include "kvm_onhyperv.h"
 
+struct kvm_hv_tlb_range {
+	u64 start_gfn;
+	u64 pages;
+};
+
 static int kvm_fill_hv_flush_list_func(struct hv_guest_mapping_flush_list *flush,
 		void *data)
 {
-	struct kvm_tlb_range *range = data;
+	struct kvm_hv_tlb_range *range = data;
 
 	return hyperv_fill_flush_guest_mapping_list(flush, range->start_gfn,
 			range->pages);
 }
 
 static inline int hv_remote_flush_root_tdp(hpa_t root_tdp,
-					   struct kvm_tlb_range *range)
+					   struct kvm_hv_tlb_range *range)
 {
 	if (range)
 		return hyperv_flush_guest_mapping_range(root_tdp,
@@ -29,8 +34,8 @@ static inline int hv_remote_flush_root_tdp(hpa_t root_tdp,
 		return hyperv_flush_guest_mapping(root_tdp);
 }
 
-int hv_remote_flush_tlb_with_range(struct kvm *kvm,
-		struct kvm_tlb_range *range)
+static int __hv_flush_remote_tlbs_range(struct kvm *kvm,
+					struct kvm_hv_tlb_range *range)
 {
 	struct kvm_arch *kvm_arch = &kvm->arch;
 	struct kvm_vcpu *vcpu;
@@ -86,19 +91,29 @@ int hv_remote_flush_tlb_with_range(struct kvm *kvm,
 	spin_unlock(&kvm_arch->hv_root_tdp_lock);
 	return ret;
 }
-EXPORT_SYMBOL_GPL(hv_remote_flush_tlb_with_range);
 
-int hv_remote_flush_tlb(struct kvm *kvm)
+int hv_flush_remote_tlbs_range(struct kvm *kvm, gfn_t start_gfn, gfn_t nr_pages)
+{
+	struct kvm_hv_tlb_range range = {
+		.start_gfn = start_gfn,
+		.pages = nr_pages,
+	};
+
+	return __hv_flush_remote_tlbs_range(kvm, &range);
+}
+EXPORT_SYMBOL_GPL(hv_flush_remote_tlbs_range);
+
+int hv_flush_remote_tlbs(struct kvm *kvm)
 {
-	return hv_remote_flush_tlb_with_range(kvm, NULL);
+	return __hv_flush_remote_tlbs_range(kvm, NULL);
 }
-EXPORT_SYMBOL_GPL(hv_remote_flush_tlb);
+EXPORT_SYMBOL_GPL(hv_flush_remote_tlbs);
 
 void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp)
 {
 	struct kvm_arch *kvm_arch = &vcpu->kvm->arch;
 
-	if (kvm_x86_ops.tlb_remote_flush == hv_remote_flush_tlb) {
+	if (kvm_x86_ops.flush_remote_tlbs == hv_flush_remote_tlbs) {
 		spin_lock(&kvm_arch->hv_root_tdp_lock);
 		vcpu->arch.hv_root_tdp = root_tdp;
 		if (root_tdp != kvm_arch->hv_root_tdp)

arch/x86/kvm/kvm_onhyperv.h

@@ -7,9 +7,8 @@
 #define __ARCH_X86_KVM_KVM_ONHYPERV_H__
 
 #if IS_ENABLED(CONFIG_HYPERV)
-int hv_remote_flush_tlb_with_range(struct kvm *kvm,
-		struct kvm_tlb_range *range);
-int hv_remote_flush_tlb(struct kvm *kvm);
+int hv_flush_remote_tlbs_range(struct kvm *kvm, gfn_t gfn, gfn_t nr_pages);
+int hv_flush_remote_tlbs(struct kvm *kvm);
 void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp);
 #else /* !CONFIG_HYPERV */
 static inline void hv_track_root_tdp(struct kvm_vcpu *vcpu, hpa_t root_tdp)