KVM: introduce kvm->srcu and convert kvm_set_memory_region to SRCU update

Use two steps for memslot deletion: mark the slot invalid (which stops
instantiation of new shadow pages for that slot, but allows destruction),
then instantiate the new empty slot.

Also simplifies kvm_handle_hva locking.

Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

3 files changed, 22 insertions, 16 deletions

diff --git a/arch/ia64/kvm/kvm-ia64.c b/arch/ia64/kvm/kvm-ia64.c
index 0757c7027986..b2e4d16dd39e 100644
--- a/arch/ia64/kvm/kvm-ia64.c
+++ b/arch/ia64/kvm/kvm-ia64.c
@@ -1382,7 +1382,7 @@ static void kvm_release_vm_pages(struct kvm *kvm)
 	int i, j;
 	unsigned long base_gfn;
 
-	slots = kvm->memslots;
+	slots = rcu_dereference(kvm->memslots);
 	for (i = 0; i < slots->nmemslots; i++) {
 		memslot = &slots->memslots[i];
 		base_gfn = memslot->base_gfn;
@@ -1837,6 +1837,7 @@ int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
 	struct kvm_memory_slot *memslot;
 	int is_dirty = 0;
 
+	down_write(&kvm->slots_lock);
 	spin_lock(&kvm->arch.dirty_log_lock);
 
 	r = kvm_ia64_sync_dirty_log(kvm, log);
@@ -1856,6 +1857,7 @@ int kvm_vm_ioctl_get_dirty_log(struct kvm *kvm,
 	}
 	r = 0;
 out:
+	up_write(&kvm->slots_lock);
 	spin_unlock(&kvm->arch.dirty_log_lock);
 	return r;
 }
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 81f84d326a84..f8bf42a25995 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -29,6 +29,7 @@
 #include <linux/swap.h>
 #include <linux/hugetlb.h>
 #include <linux/compiler.h>
+#include <linux/srcu.h>
 
 #include <asm/page.h>
 #include <asm/cmpxchg.h>
@@ -807,21 +808,15 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 {
 	int i, j;
 	int retval = 0;
-	struct kvm_memslots *slots = kvm->memslots;
+	struct kvm_memslots *slots;
+
+	slots = rcu_dereference(kvm->memslots);
 
-	/*
-	 * If mmap_sem isn't taken, we can look the memslots with only
-	 * the mmu_lock by skipping over the slots with userspace_addr == 0.
-	 */
 	for (i = 0; i < slots->nmemslots; i++) {
 		struct kvm_memory_slot *memslot = &slots->memslots[i];
 		unsigned long start = memslot->userspace_addr;
 		unsigned long end;
 
-		/* mmu_lock protects userspace_addr */
-		if (!start)
-			continue;
-
 		end = start + (memslot->npages << PAGE_SHIFT);
 		if (hva >= start && hva < end) {
 			gfn_t gfn_offset = (hva - start) >> PAGE_SHIFT;
@@ -1617,7 +1612,7 @@ static void mmu_unshadow(struct kvm *kvm, gfn_t gfn)
 
 static void page_header_update_slot(struct kvm *kvm, void *pte, gfn_t gfn)
 {
-	int slot = memslot_id(kvm, gfn_to_memslot(kvm, gfn));
+	int slot = memslot_id(kvm, gfn);
 	struct kvm_mmu_page *sp = page_header(__pa(pte));
 
 	__set_bit(slot, sp->slot_bitmap);
@@ -3021,9 +3016,11 @@ unsigned int kvm_mmu_calculate_mmu_pages(struct kvm *kvm)
 	int i;
 	unsigned int nr_mmu_pages;
 	unsigned int  nr_pages = 0;
+	struct kvm_memslots *slots;
 
-	for (i = 0; i < kvm->memslots->nmemslots; i++)
-		nr_pages += kvm->memslots->memslots[i].npages;
+	slots = rcu_dereference(kvm->memslots);
+	for (i = 0; i < slots->nmemslots; i++)
+		nr_pages += slots->memslots[i].npages;
 
 	nr_mmu_pages = nr_pages * KVM_PERMILLE_MMU_PAGES / 1000;
 	nr_mmu_pages = max(nr_mmu_pages,
@@ -3293,10 +3290,12 @@ static void audit_mappings(struct kvm_vcpu *vcpu)
 static int count_rmaps(struct kvm_vcpu *vcpu)
 {
 	int nmaps = 0;
-	int i, j, k;
+	int i, j, k, idx;
 
+	idx = srcu_read_lock(&kvm->srcu);
+	slots = rcu_dereference(kvm->memslots);
 	for (i = 0; i < KVM_MEMORY_SLOTS; ++i) {
-		struct kvm_memory_slot *m = &vcpu->kvm->memslots->memslots[i];
+		struct kvm_memory_slot *m = &slots->memslots[i];
 		struct kvm_rmap_desc *d;
 
 		for (j = 0; j < m->npages; ++j) {
@@ -3319,6 +3318,7 @@ static int count_rmaps(struct kvm_vcpu *vcpu)
 			}
 		}
 	}
+	srcu_read_unlock(&kvm->srcu, idx);
 	return nmaps;
 }
 
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index 18698799e365..f1cae7d6113d 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -1503,7 +1503,11 @@ static void enter_pmode(struct kvm_vcpu *vcpu)
 static gva_t rmode_tss_base(struct kvm *kvm)
 {
 	if (!kvm->arch.tss_addr) {
-		gfn_t base_gfn = kvm->memslots->memslots[0].base_gfn +
+		struct kvm_memslots *slots;
+		gfn_t base_gfn;
+
+		slots = rcu_dereference(kvm->memslots);
+		base_gfn = kvm->memslots->memslots[0].base_gfn +
 				 kvm->memslots->memslots[0].npages - 3;
 		return base_gfn << PAGE_SHIFT;
 	}