From 9594a4986192f99c01a7c0a1779b5ac0eff8e208 Mon Sep 17 00:00:00 2001
From: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Date: Mon, 2 Jul 2012 17:53:25 +0900
Subject: KVM: MMU: Use __gfn_to_rmap() to clean up kvm_handle_hva()

We can treat every level uniformly.

Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 28c8fbc..2beb95b 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1281,14 +1281,14 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 			gfn_t gfn_offset = (hva - start) >> PAGE_SHIFT;
 			gfn_t gfn = memslot->base_gfn + gfn_offset;
 
-			ret = handler(kvm, &memslot->rmap[gfn_offset], data);
+			ret = 0;
 
-			for (j = 0; j < KVM_NR_PAGE_SIZES - 1; ++j) {
-				struct kvm_lpage_info *linfo;
+			for (j = PT_PAGE_TABLE_LEVEL;
+			     j < PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES; ++j) {
+				unsigned long *rmapp;
 
-				linfo = lpage_info_slot(gfn, memslot,
-							PT_DIRECTORY_LEVEL + j);
-				ret |= handler(kvm, &linfo->rmap_pde, data);
+				rmapp = __gfn_to_rmap(gfn, j, memslot);
+				ret |= handler(kvm, rmapp, data);
 			}
 			trace_kvm_age_page(hva, memslot, ret);
 			retval |= ret;
-- 
cgit v0.10.2


From d19a748b1c42b133e9263e9023c1d162efa6f4ad Mon Sep 17 00:00:00 2001
From: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Date: Mon, 2 Jul 2012 17:54:30 +0900
Subject: KVM: Introduce hva_to_gfn_memslot() for kvm_handle_hva()

This restricts hva handling in mmu code and makes it easier to extend
kvm_handle_hva() so that it can treat a range of addresses later in this
patch series.

Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Cc: Alexander Graf <agraf@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/powerpc/kvm/book3s_64_mmu_hv.c b/arch/powerpc/kvm/book3s_64_mmu_hv.c
index d03eb6f..3703755 100644
--- a/arch/powerpc/kvm/book3s_64_mmu_hv.c
+++ b/arch/powerpc/kvm/book3s_64_mmu_hv.c
@@ -772,10 +772,10 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 
 		end = start + (memslot->npages << PAGE_SHIFT);
 		if (hva >= start && hva < end) {
-			gfn_t gfn_offset = (hva - start) >> PAGE_SHIFT;
+			gfn_t gfn = hva_to_gfn_memslot(hva, memslot);
+			gfn_t gfn_offset = gfn - memslot->base_gfn;
 
-			ret = handler(kvm, &memslot->rmap[gfn_offset],
-				      memslot->base_gfn + gfn_offset);
+			ret = handler(kvm, &memslot->rmap[gfn_offset], gfn);
 			retval |= ret;
 		}
 	}
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 2beb95b..170a632 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1278,8 +1278,7 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 
 		end = start + (memslot->npages << PAGE_SHIFT);
 		if (hva >= start && hva < end) {
-			gfn_t gfn_offset = (hva - start) >> PAGE_SHIFT;
-			gfn_t gfn = memslot->base_gfn + gfn_offset;
+			gfn_t gfn = hva_to_gfn_memslot(hva, memslot);
 
 			ret = 0;
 
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index e3c86f8..6f6c18a 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -740,6 +740,14 @@ static inline gfn_t gfn_to_index(gfn_t gfn, gfn_t base_gfn, int level)
 		(base_gfn >> KVM_HPAGE_GFN_SHIFT(level));
 }
 
+static inline gfn_t
+hva_to_gfn_memslot(unsigned long hva, struct kvm_memory_slot *slot)
+{
+	gfn_t gfn_offset = (hva - slot->userspace_addr) >> PAGE_SHIFT;
+
+	return slot->base_gfn + gfn_offset;
+}
+
 static inline unsigned long gfn_to_hva_memslot(struct kvm_memory_slot *slot,
 					       gfn_t gfn)
 {
-- 
cgit v0.10.2


From 84504ef38673fa021b3d8f3da2b79cf878b33315 Mon Sep 17 00:00:00 2001
From: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Date: Mon, 2 Jul 2012 17:55:48 +0900
Subject: KVM: MMU: Make kvm_handle_hva() handle range of addresses

When guest's memory is backed by THP pages, MMU notifier needs to call
kvm_unmap_hva(), which in turn leads to kvm_handle_hva(), in a loop to
invalidate a range of pages which constitute one huge page:

  for each page
    for each memslot
      if page is in memslot
        unmap using rmap

This means although every page in that range is expected to be found in
the same memslot, we are forced to check unrelated memslots many times.
If the guest has more memslots, the situation will become worse.

Furthermore, if the range does not include any pages in the guest's
memory, the loop over the pages will just consume extra time.

This patch, together with the following patches, solves this problem by
introducing kvm_handle_hva_range() which makes the loop look like this:

  for each memslot
    for each page in memslot
      unmap using rmap

In this new processing, the actual work is converted to a loop over rmap
which is much more cache friendly than before.

Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Cc: Alexander Graf <agraf@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/powerpc/kvm/book3s_64_mmu_hv.c b/arch/powerpc/kvm/book3s_64_mmu_hv.c
index 3703755..1a470bc 100644
--- a/arch/powerpc/kvm/book3s_64_mmu_hv.c
+++ b/arch/powerpc/kvm/book3s_64_mmu_hv.c
@@ -756,9 +756,12 @@ int kvmppc_book3s_hv_page_fault(struct kvm_run *run, struct kvm_vcpu *vcpu,
 	goto out_put;
 }
 
-static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
-			  int (*handler)(struct kvm *kvm, unsigned long *rmapp,
-					 unsigned long gfn))
+static int kvm_handle_hva_range(struct kvm *kvm,
+				unsigned long start,
+				unsigned long end,
+				int (*handler)(struct kvm *kvm,
+					       unsigned long *rmapp,
+					       unsigned long gfn))
 {
 	int ret;
 	int retval = 0;
@@ -767,12 +770,22 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 
 	slots = kvm_memslots(kvm);
 	kvm_for_each_memslot(memslot, slots) {
-		unsigned long start = memslot->userspace_addr;
-		unsigned long end;
+		unsigned long hva_start, hva_end;
+		gfn_t gfn, gfn_end;
+
+		hva_start = max(start, memslot->userspace_addr);
+		hva_end = min(end, memslot->userspace_addr +
+					(memslot->npages << PAGE_SHIFT));
+		if (hva_start >= hva_end)
+			continue;
+		/*
+		 * {gfn(page) | page intersects with [hva_start, hva_end)} =
+		 * {gfn, gfn+1, ..., gfn_end-1}.
+		 */
+		gfn = hva_to_gfn_memslot(hva_start, memslot);
+		gfn_end = hva_to_gfn_memslot(hva_end + PAGE_SIZE - 1, memslot);
 
-		end = start + (memslot->npages << PAGE_SHIFT);
-		if (hva >= start && hva < end) {
-			gfn_t gfn = hva_to_gfn_memslot(hva, memslot);
+		for (; gfn < gfn_end; ++gfn) {
 			gfn_t gfn_offset = gfn - memslot->base_gfn;
 
 			ret = handler(kvm, &memslot->rmap[gfn_offset], gfn);
@@ -783,6 +796,13 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 	return retval;
 }
 
+static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
+			  int (*handler)(struct kvm *kvm, unsigned long *rmapp,
+					 unsigned long gfn))
+{
+	return kvm_handle_hva_range(kvm, hva, hva + 1, handler);
+}
+
 static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp,
 			   unsigned long gfn)
 {
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 170a632..7235b0c 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1259,10 +1259,13 @@ static int kvm_set_pte_rmapp(struct kvm *kvm, unsigned long *rmapp,
 	return 0;
 }
 
-static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
-			  unsigned long data,
-			  int (*handler)(struct kvm *kvm, unsigned long *rmapp,
-					 unsigned long data))
+static int kvm_handle_hva_range(struct kvm *kvm,
+				unsigned long start,
+				unsigned long end,
+				unsigned long data,
+				int (*handler)(struct kvm *kvm,
+					       unsigned long *rmapp,
+					       unsigned long data))
 {
 	int j;
 	int ret;
@@ -1273,13 +1276,22 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 	slots = kvm_memslots(kvm);
 
 	kvm_for_each_memslot(memslot, slots) {
-		unsigned long start = memslot->userspace_addr;
-		unsigned long end;
+		unsigned long hva_start, hva_end;
+		gfn_t gfn, gfn_end;
 
-		end = start + (memslot->npages << PAGE_SHIFT);
-		if (hva >= start && hva < end) {
-			gfn_t gfn = hva_to_gfn_memslot(hva, memslot);
+		hva_start = max(start, memslot->userspace_addr);
+		hva_end = min(end, memslot->userspace_addr +
+					(memslot->npages << PAGE_SHIFT));
+		if (hva_start >= hva_end)
+			continue;
+		/*
+		 * {gfn(page) | page intersects with [hva_start, hva_end)} =
+		 * {gfn, gfn+1, ..., gfn_end-1}.
+		 */
+		gfn = hva_to_gfn_memslot(hva_start, memslot);
+		gfn_end = hva_to_gfn_memslot(hva_end + PAGE_SIZE - 1, memslot);
 
+		for (; gfn < gfn_end; ++gfn) {
 			ret = 0;
 
 			for (j = PT_PAGE_TABLE_LEVEL;
@@ -1289,7 +1301,9 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 				rmapp = __gfn_to_rmap(gfn, j, memslot);
 				ret |= handler(kvm, rmapp, data);
 			}
-			trace_kvm_age_page(hva, memslot, ret);
+			trace_kvm_age_page(memslot->userspace_addr +
+					(gfn - memslot->base_gfn) * PAGE_SIZE,
+					memslot, ret);
 			retval |= ret;
 		}
 	}
@@ -1297,6 +1311,14 @@ static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 	return retval;
 }
 
+static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
+			  unsigned long data,
+			  int (*handler)(struct kvm *kvm, unsigned long *rmapp,
+					 unsigned long data))
+{
+	return kvm_handle_hva_range(kvm, hva, hva + 1, data, handler);
+}
+
 int kvm_unmap_hva(struct kvm *kvm, unsigned long hva)
 {
 	return kvm_handle_hva(kvm, hva, 0, kvm_unmap_rmapp);
-- 
cgit v0.10.2


From b3ae2096974b12c3af2ad1a4e7716b084949867f Mon Sep 17 00:00:00 2001
From: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Date: Mon, 2 Jul 2012 17:56:33 +0900
Subject: KVM: Introduce kvm_unmap_hva_range() for
 kvm_mmu_notifier_invalidate_range_start()

When we tested KVM under memory pressure, with THP enabled on the host,
we noticed that MMU notifier took a long time to invalidate huge pages.

Since the invalidation was done with mmu_lock held, it not only wasted
the CPU but also made the host harder to respond.

This patch mitigates this by using kvm_handle_hva_range().

Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Cc: Alexander Graf <agraf@suse.de>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/powerpc/include/asm/kvm_host.h b/arch/powerpc/include/asm/kvm_host.h
index 50ea12f..572ad01 100644
--- a/arch/powerpc/include/asm/kvm_host.h
+++ b/arch/powerpc/include/asm/kvm_host.h
@@ -52,6 +52,8 @@
 
 struct kvm;
 extern int kvm_unmap_hva(struct kvm *kvm, unsigned long hva);
+extern int kvm_unmap_hva_range(struct kvm *kvm,
+			       unsigned long start, unsigned long end);
 extern int kvm_age_hva(struct kvm *kvm, unsigned long hva);
 extern int kvm_test_age_hva(struct kvm *kvm, unsigned long hva);
 extern void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte);
diff --git a/arch/powerpc/kvm/book3s_64_mmu_hv.c b/arch/powerpc/kvm/book3s_64_mmu_hv.c
index 1a470bc..3c635c0 100644
--- a/arch/powerpc/kvm/book3s_64_mmu_hv.c
+++ b/arch/powerpc/kvm/book3s_64_mmu_hv.c
@@ -870,6 +870,13 @@ int kvm_unmap_hva(struct kvm *kvm, unsigned long hva)
 	return 0;
 }
 
+int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end)
+{
+	if (kvm->arch.using_mmu_notifiers)
+		kvm_handle_hva_range(kvm, start, end, kvm_unmap_rmapp);
+	return 0;
+}
+
 static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 			 unsigned long gfn)
 {
diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index a3e9409..d4aab86 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -944,6 +944,7 @@ extern bool kvm_rebooting;
 
 #define KVM_ARCH_WANT_MMU_NOTIFIER
 int kvm_unmap_hva(struct kvm *kvm, unsigned long hva);
+int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end);
 int kvm_age_hva(struct kvm *kvm, unsigned long hva);
 int kvm_test_age_hva(struct kvm *kvm, unsigned long hva);
 void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte);
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 7235b0c..d2855f8 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1324,6 +1324,11 @@ int kvm_unmap_hva(struct kvm *kvm, unsigned long hva)
 	return kvm_handle_hva(kvm, hva, 0, kvm_unmap_rmapp);
 }
 
+int kvm_unmap_hva_range(struct kvm *kvm, unsigned long start, unsigned long end)
+{
+	return kvm_handle_hva_range(kvm, start, end, 0, kvm_unmap_rmapp);
+}
+
 void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte)
 {
 	kvm_handle_hva(kvm, hva, (unsigned long)&pte, kvm_set_pte_rmapp);
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index b3ce91c..e2b1a15 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -332,8 +332,7 @@ static void kvm_mmu_notifier_invalidate_range_start(struct mmu_notifier *mn,
 	 * count is also read inside the mmu_lock critical section.
 	 */
 	kvm->mmu_notifier_count++;
-	for (; start < end; start += PAGE_SIZE)
-		need_tlb_flush |= kvm_unmap_hva(kvm, start);
+	need_tlb_flush = kvm_unmap_hva_range(kvm, start, end);
 	need_tlb_flush |= kvm->tlbs_dirty;
 	/* we've to flush the tlb before the pages can be freed */
 	if (need_tlb_flush)
-- 
cgit v0.10.2


From 77d11309b3a10e1ce112058ec2c9b7b979bcf311 Mon Sep 17 00:00:00 2001
From: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Date: Mon, 2 Jul 2012 17:57:17 +0900
Subject: KVM: Separate rmap_pde from kvm_lpage_info->write_count

This makes it possible to loop over rmap_pde arrays in the same way as
we do over rmap so that we can optimize kvm_handle_hva_range() easily in
the following patch.

Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
index d4aab86..4f98da9 100644
--- a/arch/x86/include/asm/kvm_host.h
+++ b/arch/x86/include/asm/kvm_host.h
@@ -500,11 +500,11 @@ struct kvm_vcpu_arch {
 };
 
 struct kvm_lpage_info {
-	unsigned long rmap_pde;
 	int write_count;
 };
 
 struct kvm_arch_memory_slot {
+	unsigned long *rmap_pde[KVM_NR_PAGE_SIZES - 1];
 	struct kvm_lpage_info *lpage_info[KVM_NR_PAGE_SIZES - 1];
 };
 
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index d2855f8..3b3f5ae 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -960,13 +960,13 @@ static void pte_list_walk(unsigned long *pte_list, pte_list_walk_fn fn)
 static unsigned long *__gfn_to_rmap(gfn_t gfn, int level,
 				    struct kvm_memory_slot *slot)
 {
-	struct kvm_lpage_info *linfo;
+	unsigned long idx;
 
 	if (likely(level == PT_PAGE_TABLE_LEVEL))
 		return &slot->rmap[gfn - slot->base_gfn];
 
-	linfo = lpage_info_slot(gfn, slot, level);
-	return &linfo->rmap_pde;
+	idx = gfn_to_index(gfn, slot->base_gfn, level);
+	return &slot->arch.rmap_pde[level - PT_DIRECTORY_LEVEL][idx];
 }
 
 /*
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 59b5950..829b4e9 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -6314,6 +6314,10 @@ void kvm_arch_free_memslot(struct kvm_memory_slot *free,
 	int i;
 
 	for (i = 0; i < KVM_NR_PAGE_SIZES - 1; ++i) {
+		if (!dont || free->arch.rmap_pde[i] != dont->arch.rmap_pde[i]) {
+			kvm_kvfree(free->arch.rmap_pde[i]);
+			free->arch.rmap_pde[i] = NULL;
+		}
 		if (!dont || free->arch.lpage_info[i] != dont->arch.lpage_info[i]) {
 			kvm_kvfree(free->arch.lpage_info[i]);
 			free->arch.lpage_info[i] = NULL;
@@ -6333,6 +6337,11 @@ int kvm_arch_create_memslot(struct kvm_memory_slot *slot, unsigned long npages)
 		lpages = gfn_to_index(slot->base_gfn + npages - 1,
 				      slot->base_gfn, level) + 1;
 
+		slot->arch.rmap_pde[i] =
+			kvm_kvzalloc(lpages * sizeof(*slot->arch.rmap_pde[i]));
+		if (!slot->arch.rmap_pde[i])
+			goto out_free;
+
 		slot->arch.lpage_info[i] =
 			kvm_kvzalloc(lpages * sizeof(*slot->arch.lpage_info[i]));
 		if (!slot->arch.lpage_info[i])
@@ -6361,7 +6370,9 @@ int kvm_arch_create_memslot(struct kvm_memory_slot *slot, unsigned long npages)
 
 out_free:
 	for (i = 0; i < KVM_NR_PAGE_SIZES - 1; ++i) {
+		kvm_kvfree(slot->arch.rmap_pde[i]);
 		kvm_kvfree(slot->arch.lpage_info[i]);
+		slot->arch.rmap_pde[i] = NULL;
 		slot->arch.lpage_info[i] = NULL;
 	}
 	return -ENOMEM;
-- 
cgit v0.10.2


From 048212d0bc0b1769a4bbecd7ace8c8d237577d1b Mon Sep 17 00:00:00 2001
From: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Date: Mon, 2 Jul 2012 17:57:59 +0900
Subject: KVM: MMU: Add memslot parameter to hva handlers

This is needed to push trace_kvm_age_page() into kvm_age_rmapp() in the
following patch.

Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 3b3f5ae..dfd7a9a 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1200,7 +1200,7 @@ static bool rmap_write_protect(struct kvm *kvm, u64 gfn)
 }
 
 static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp,
-			   unsigned long data)
+			   struct kvm_memory_slot *slot, unsigned long data)
 {
 	u64 *sptep;
 	struct rmap_iterator iter;
@@ -1218,7 +1218,7 @@ static int kvm_unmap_rmapp(struct kvm *kvm, unsigned long *rmapp,
 }
 
 static int kvm_set_pte_rmapp(struct kvm *kvm, unsigned long *rmapp,
-			     unsigned long data)
+			     struct kvm_memory_slot *slot, unsigned long data)
 {
 	u64 *sptep;
 	struct rmap_iterator iter;
@@ -1265,6 +1265,7 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 				unsigned long data,
 				int (*handler)(struct kvm *kvm,
 					       unsigned long *rmapp,
+					       struct kvm_memory_slot *slot,
 					       unsigned long data))
 {
 	int j;
@@ -1299,7 +1300,7 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 				unsigned long *rmapp;
 
 				rmapp = __gfn_to_rmap(gfn, j, memslot);
-				ret |= handler(kvm, rmapp, data);
+				ret |= handler(kvm, rmapp, memslot, data);
 			}
 			trace_kvm_age_page(memslot->userspace_addr +
 					(gfn - memslot->base_gfn) * PAGE_SIZE,
@@ -1314,6 +1315,7 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
 			  unsigned long data,
 			  int (*handler)(struct kvm *kvm, unsigned long *rmapp,
+					 struct kvm_memory_slot *slot,
 					 unsigned long data))
 {
 	return kvm_handle_hva_range(kvm, hva, hva + 1, data, handler);
@@ -1335,7 +1337,7 @@ void kvm_set_spte_hva(struct kvm *kvm, unsigned long hva, pte_t pte)
 }
 
 static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
-			 unsigned long data)
+			 struct kvm_memory_slot *slot, unsigned long data)
 {
 	u64 *sptep;
 	struct rmap_iterator uninitialized_var(iter);
@@ -1350,7 +1352,7 @@ static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 	 * out actively used pages or breaking up actively used hugepages.
 	 */
 	if (!shadow_accessed_mask)
-		return kvm_unmap_rmapp(kvm, rmapp, data);
+		return kvm_unmap_rmapp(kvm, rmapp, slot, data);
 
 	for (sptep = rmap_get_first(*rmapp, &iter); sptep;
 	     sptep = rmap_get_next(&iter)) {
@@ -1367,7 +1369,7 @@ static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 }
 
 static int kvm_test_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
-			      unsigned long data)
+			      struct kvm_memory_slot *slot, unsigned long data)
 {
 	u64 *sptep;
 	struct rmap_iterator iter;
@@ -1405,7 +1407,7 @@ static void rmap_recycle(struct kvm_vcpu *vcpu, u64 *spte, gfn_t gfn)
 
 	rmapp = gfn_to_rmap(vcpu->kvm, gfn, sp->role.level);
 
-	kvm_unmap_rmapp(vcpu->kvm, rmapp, 0);
+	kvm_unmap_rmapp(vcpu->kvm, rmapp, NULL, 0);
 	kvm_flush_remote_tlbs(vcpu->kvm);
 }
 
-- 
cgit v0.10.2


From f395302e09ef783b8f82d1160510a95aa8c66dbc Mon Sep 17 00:00:00 2001
From: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Date: Mon, 2 Jul 2012 17:58:48 +0900
Subject: KVM: MMU: Push trace_kvm_age_page() into kvm_age_rmapp()

This restricts the tracing to page aging and makes it possible to
optimize kvm_handle_hva_range() further in the following patch.

Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index dfd7a9a..58adec3 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1269,8 +1269,7 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 					       unsigned long data))
 {
 	int j;
-	int ret;
-	int retval = 0;
+	int ret = 0;
 	struct kvm_memslots *slots;
 	struct kvm_memory_slot *memslot;
 
@@ -1293,8 +1292,6 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 		gfn_end = hva_to_gfn_memslot(hva_end + PAGE_SIZE - 1, memslot);
 
 		for (; gfn < gfn_end; ++gfn) {
-			ret = 0;
-
 			for (j = PT_PAGE_TABLE_LEVEL;
 			     j < PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES; ++j) {
 				unsigned long *rmapp;
@@ -1302,14 +1299,10 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 				rmapp = __gfn_to_rmap(gfn, j, memslot);
 				ret |= handler(kvm, rmapp, memslot, data);
 			}
-			trace_kvm_age_page(memslot->userspace_addr +
-					(gfn - memslot->base_gfn) * PAGE_SIZE,
-					memslot, ret);
-			retval |= ret;
 		}
 	}
 
-	return retval;
+	return ret;
 }
 
 static int kvm_handle_hva(struct kvm *kvm, unsigned long hva,
@@ -1351,8 +1344,10 @@ static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 	 * This has some overhead, but not as much as the cost of swapping
 	 * out actively used pages or breaking up actively used hugepages.
 	 */
-	if (!shadow_accessed_mask)
-		return kvm_unmap_rmapp(kvm, rmapp, slot, data);
+	if (!shadow_accessed_mask) {
+		young = kvm_unmap_rmapp(kvm, rmapp, slot, data);
+		goto out;
+	}
 
 	for (sptep = rmap_get_first(*rmapp, &iter); sptep;
 	     sptep = rmap_get_next(&iter)) {
@@ -1364,7 +1359,9 @@ static int kvm_age_rmapp(struct kvm *kvm, unsigned long *rmapp,
 				 (unsigned long *)sptep);
 		}
 	}
-
+out:
+	/* @data has hva passed to kvm_age_hva(). */
+	trace_kvm_age_page(data, slot, young);
 	return young;
 }
 
@@ -1413,7 +1410,7 @@ static void rmap_recycle(struct kvm_vcpu *vcpu, u64 *spte, gfn_t gfn)
 
 int kvm_age_hva(struct kvm *kvm, unsigned long hva)
 {
-	return kvm_handle_hva(kvm, hva, 0, kvm_age_rmapp);
+	return kvm_handle_hva(kvm, hva, hva, kvm_age_rmapp);
 }
 
 int kvm_test_age_hva(struct kvm *kvm, unsigned long hva)
-- 
cgit v0.10.2


From bcd3ef58283a471d6b65855b83f78bd39eb55391 Mon Sep 17 00:00:00 2001
From: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Date: Mon, 2 Jul 2012 17:59:33 +0900
Subject: KVM: MMU: Avoid handling same rmap_pde in kvm_handle_hva_range()

When we invalidate a THP page, we call the handler with the same
rmap_pde argument 512 times in the following loop:

  for each guest page in the range
    for each level
      unmap using rmap

This patch avoids these extra handler calls by changing the loop order
like this:

  for each level
    for each rmap in the range
      unmap using rmap

With the preceding patches in the patch series, this made THP page
invalidation more than 5 times faster on our x86 host: the host became
more responsive during swapping the guest's memory as a result.

Signed-off-by: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 58adec3..a5d6ef7 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -1277,7 +1277,7 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 
 	kvm_for_each_memslot(memslot, slots) {
 		unsigned long hva_start, hva_end;
-		gfn_t gfn, gfn_end;
+		gfn_t gfn_start, gfn_end;
 
 		hva_start = max(start, memslot->userspace_addr);
 		hva_end = min(end, memslot->userspace_addr +
@@ -1286,19 +1286,27 @@ static int kvm_handle_hva_range(struct kvm *kvm,
 			continue;
 		/*
 		 * {gfn(page) | page intersects with [hva_start, hva_end)} =
-		 * {gfn, gfn+1, ..., gfn_end-1}.
+		 * {gfn_start, gfn_start+1, ..., gfn_end-1}.
 		 */
-		gfn = hva_to_gfn_memslot(hva_start, memslot);
+		gfn_start = hva_to_gfn_memslot(hva_start, memslot);
 		gfn_end = hva_to_gfn_memslot(hva_end + PAGE_SIZE - 1, memslot);
 
-		for (; gfn < gfn_end; ++gfn) {
-			for (j = PT_PAGE_TABLE_LEVEL;
-			     j < PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES; ++j) {
-				unsigned long *rmapp;
+		for (j = PT_PAGE_TABLE_LEVEL;
+		     j < PT_PAGE_TABLE_LEVEL + KVM_NR_PAGE_SIZES; ++j) {
+			unsigned long idx, idx_end;
+			unsigned long *rmapp;
 
-				rmapp = __gfn_to_rmap(gfn, j, memslot);
-				ret |= handler(kvm, rmapp, memslot, data);
-			}
+			/*
+			 * {idx(page_j) | page_j intersects with
+			 *  [hva_start, hva_end)} = {idx, idx+1, ..., idx_end}.
+			 */
+			idx = gfn_to_index(gfn_start, memslot->base_gfn, j);
+			idx_end = gfn_to_index(gfn_end - 1, memslot->base_gfn, j);
+
+			rmapp = __gfn_to_rmap(gfn_start, j, memslot);
+
+			for (; idx <= idx_end; ++idx)
+				ret |= handler(kvm, rmapp++, memslot, data);
 		}
 	}
 
-- 
cgit v0.10.2


From 9d3c92af47d853d4e31ee971dba7bc086275b7b3 Mon Sep 17 00:00:00 2001
From: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Date: Tue, 17 Jul 2012 21:50:48 +0800
Subject: KVM: x86: remove unnecessary mark_page_dirty

fix:
[  132.474633] 3.5.0-rc1+ #50 Not tainted
[  132.474634] -------------------------------
[  132.474635] include/linux/kvm_host.h:369 suspicious rcu_dereference_check() usage!
[  132.474636]
[  132.474636] other info that might help us debug this:
[  132.474636]
[  132.474638]
[  132.474638] rcu_scheduler_active = 1, debug_locks = 1
[  132.474640] 1 lock held by qemu-kvm/2832:
[  132.474657]  #0:  (&vcpu->mutex){+.+.+.}, at: [<ffffffffa01e1636>] vcpu_load+0x1e/0x91 [kvm]
[  132.474658]
[  132.474658] stack backtrace:
[  132.474660] Pid: 2832, comm: qemu-kvm Not tainted 3.5.0-rc1+ #50
[  132.474661] Call Trace:
[  132.474665]  [<ffffffff81092f40>] lockdep_rcu_suspicious+0xfc/0x105
[  132.474675]  [<ffffffffa01e0c85>] kvm_memslots+0x6d/0x75 [kvm]
[  132.474683]  [<ffffffffa01e0ca1>] gfn_to_memslot+0x14/0x4c [kvm]
[  132.474693]  [<ffffffffa01e3575>] mark_page_dirty+0x17/0x2a [kvm]
[  132.474706]  [<ffffffffa01f21ea>] kvm_arch_vcpu_ioctl+0xbcf/0xc07 [kvm]

Actually, we do not write vcpu->arch.time at this time, mark_page_dirty
should be removed.

Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 829b4e9..ecc71dd 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -2632,7 +2632,6 @@ static int kvm_set_guest_paused(struct kvm_vcpu *vcpu)
 	if (!vcpu->arch.time_page)
 		return -EINVAL;
 	src->flags |= PVCLOCK_GUEST_STOPPED;
-	mark_page_dirty(vcpu->kvm, vcpu->arch.time >> PAGE_SHIFT);
 	kvm_make_request(KVM_REQ_CLOCK_UPDATE, vcpu);
 	return 0;
 }
-- 
cgit v0.10.2


From 86fde74cf5b829627b37ca86322acfdd99b524b8 Mon Sep 17 00:00:00 2001
From: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Date: Tue, 17 Jul 2012 21:52:52 +0800
Subject: KVM: MMU: track the refcount when unmap the page

It will trigger a WARN_ON if the page has been freed but it is still
used in mmu, it can help us to detect mm bug early

Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index a5d6ef7..685a485 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -556,6 +556,14 @@ static int mmu_spte_clear_track_bits(u64 *sptep)
 		return 0;
 
 	pfn = spte_to_pfn(old_spte);
+
+	/*
+	 * KVM does not hold the refcount of the page used by
+	 * kvm mmu, before reclaiming the page, we should
+	 * unmap it from mmu first.
+	 */
+	WARN_ON(!kvm_is_mmio_pfn(pfn) && !page_count(pfn_to_page(pfn)));
+
 	if (!shadow_accessed_mask || old_spte & shadow_accessed_mask)
 		kvm_set_pfn_accessed(pfn);
 	if (!shadow_dirty_mask || (old_spte & shadow_dirty_mask))
-- 
cgit v0.10.2


From 903816fa4d016e20ec71a1a97700cfcdda115580 Mon Sep 17 00:00:00 2001
From: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Date: Tue, 17 Jul 2012 21:54:11 +0800
Subject: KVM: using get_fault_pfn to get the fault pfn

Using get_fault_pfn to cleanup the code

Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index 685a485..f85cc21 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -2513,10 +2513,8 @@ static pfn_t pte_prefetch_gfn_to_pfn(struct kvm_vcpu *vcpu, gfn_t gfn,
 	unsigned long hva;
 
 	slot = gfn_to_memslot_dirty_bitmap(vcpu, gfn, no_dirty_log);
-	if (!slot) {
-		get_page(fault_page);
-		return page_to_pfn(fault_page);
-	}
+	if (!slot)
+		return get_fault_pfn();
 
 	hva = gfn_to_hva_memslot(slot, gfn);
 
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 6f6c18a..1a7f838 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -383,15 +383,11 @@ id_to_memslot(struct kvm_memslots *slots, int id)
 static inline int is_error_hpa(hpa_t hpa) { return hpa >> HPA_MSB; }
 
 extern struct page *bad_page;
-extern struct page *fault_page;
-
 extern pfn_t bad_pfn;
-extern pfn_t fault_pfn;
 
 int is_error_page(struct page *page);
 int is_error_pfn(pfn_t pfn);
 int is_hwpoison_pfn(pfn_t pfn);
-int is_fault_pfn(pfn_t pfn);
 int is_noslot_pfn(pfn_t pfn);
 int is_invalid_pfn(pfn_t pfn);
 int kvm_is_error_hva(unsigned long addr);
@@ -441,6 +437,7 @@ void kvm_release_pfn_clean(pfn_t pfn);
 void kvm_set_pfn_dirty(pfn_t pfn);
 void kvm_set_pfn_accessed(pfn_t pfn);
 void kvm_get_pfn(pfn_t pfn);
+pfn_t get_fault_pfn(void);
 
 int kvm_read_guest_page(struct kvm *kvm, gfn_t gfn, void *data, int offset,
 			int len);
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index e2b1a15..0fbbf2d 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -103,8 +103,8 @@ static bool largepages_enabled = true;
 static struct page *hwpoison_page;
 static pfn_t hwpoison_pfn;
 
-struct page *fault_page;
-pfn_t fault_pfn;
+static struct page *fault_page;
+static pfn_t fault_pfn;
 
 inline int kvm_is_mmio_pfn(pfn_t pfn)
 {
@@ -949,12 +949,6 @@ int is_hwpoison_pfn(pfn_t pfn)
 }
 EXPORT_SYMBOL_GPL(is_hwpoison_pfn);
 
-int is_fault_pfn(pfn_t pfn)
-{
-	return pfn == fault_pfn;
-}
-EXPORT_SYMBOL_GPL(is_fault_pfn);
-
 int is_noslot_pfn(pfn_t pfn)
 {
 	return pfn == bad_pfn;
@@ -1038,11 +1032,12 @@ unsigned long gfn_to_hva(struct kvm *kvm, gfn_t gfn)
 }
 EXPORT_SYMBOL_GPL(gfn_to_hva);
 
-static pfn_t get_fault_pfn(void)
+pfn_t get_fault_pfn(void)
 {
 	get_page(fault_page);
 	return fault_pfn;
 }
+EXPORT_SYMBOL_GPL(get_fault_pfn);
 
 int get_user_page_nowait(struct task_struct *tsk, struct mm_struct *mm,
 	unsigned long start, int write, struct page **page)
-- 
cgit v0.10.2


From ca0565f5736e67af3172d188577b57e303dd082a Mon Sep 17 00:00:00 2001
From: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Date: Tue, 17 Jul 2012 21:54:52 +0800
Subject: KVM: make bad_pfn static to kvm_main.c

bad_pfn is not used out of kvm_main.c, so mark it static, also move it near
hwpoison_pfn and fault_pfn

Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 1a7f838..5f956cd 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -383,7 +383,6 @@ id_to_memslot(struct kvm_memslots *slots, int id)
 static inline int is_error_hpa(hpa_t hpa) { return hpa >> HPA_MSB; }
 
 extern struct page *bad_page;
-extern pfn_t bad_pfn;
 
 int is_error_page(struct page *page);
 int is_error_pfn(pfn_t pfn);
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 0fbbf2d..f955eee 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -100,6 +100,9 @@ EXPORT_SYMBOL_GPL(kvm_rebooting);
 
 static bool largepages_enabled = true;
 
+struct page *bad_page;
+static pfn_t bad_pfn;
+
 static struct page *hwpoison_page;
 static pfn_t hwpoison_pfn;
 
@@ -2691,9 +2694,6 @@ static struct syscore_ops kvm_syscore_ops = {
 	.resume = kvm_resume,
 };
 
-struct page *bad_page;
-pfn_t bad_pfn;
-
 static inline
 struct kvm_vcpu *preempt_notifier_to_vcpu(struct preempt_notifier *pn)
 {
-- 
cgit v0.10.2


From f340a51b7e41abbe92ae3a327c0020974a059f95 Mon Sep 17 00:00:00 2001
From: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Date: Tue, 17 Jul 2012 21:55:34 +0800
Subject: KVM: remove is_error_hpa

Remove them since they are not used anymore

Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 5f956cd..e8d13a0 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -378,10 +378,6 @@ id_to_memslot(struct kvm_memslots *slots, int id)
 	return slot;
 }
 
-#define HPA_MSB ((sizeof(hpa_t) * 8) - 1)
-#define HPA_ERR_MASK ((hpa_t)1 << HPA_MSB)
-static inline int is_error_hpa(hpa_t hpa) { return hpa >> HPA_MSB; }
-
 extern struct page *bad_page;
 
 int is_error_page(struct page *page);
-- 
cgit v0.10.2


From d566104853361cc377c61f70e41c1ad3d44b86c6 Mon Sep 17 00:00:00 2001
From: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Date: Tue, 17 Jul 2012 21:56:16 +0800
Subject: KVM: remove the unused parameter of gfn_to_pfn_memslot

The parameter, 'kvm', is not used in gfn_to_pfn_memslot, we can happily remove
it

Signed-off-by: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/powerpc/kvm/e500_tlb.c b/arch/powerpc/kvm/e500_tlb.c
index c510fc9..c8f6c58 100644
--- a/arch/powerpc/kvm/e500_tlb.c
+++ b/arch/powerpc/kvm/e500_tlb.c
@@ -520,7 +520,7 @@ static inline void kvmppc_e500_shadow_map(struct kvmppc_vcpu_e500 *vcpu_e500,
 
 	if (likely(!pfnmap)) {
 		unsigned long tsize_pages = 1 << (tsize + 10 - PAGE_SHIFT);
-		pfn = gfn_to_pfn_memslot(vcpu_e500->vcpu.kvm, slot, gfn);
+		pfn = gfn_to_pfn_memslot(slot, gfn);
 		if (is_error_pfn(pfn)) {
 			printk(KERN_ERR "Couldn't get real page for gfn %lx!\n",
 					(long)gfn);
diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c
index f85cc21..4f77f7a 100644
--- a/arch/x86/kvm/mmu.c
+++ b/arch/x86/kvm/mmu.c
@@ -2518,7 +2518,7 @@ static pfn_t pte_prefetch_gfn_to_pfn(struct kvm_vcpu *vcpu, gfn_t gfn,
 
 	hva = gfn_to_hva_memslot(slot, gfn);
 
-	return hva_to_pfn_atomic(vcpu->kvm, hva);
+	return hva_to_pfn_atomic(hva);
 }
 
 static int direct_pte_prefetch_many(struct kvm_vcpu *vcpu,
diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index e8d13a0..db9aa917 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -418,15 +418,14 @@ void kvm_release_page_dirty(struct page *page);
 void kvm_set_page_dirty(struct page *page);
 void kvm_set_page_accessed(struct page *page);
 
-pfn_t hva_to_pfn_atomic(struct kvm *kvm, unsigned long addr);
+pfn_t hva_to_pfn_atomic(unsigned long addr);
 pfn_t gfn_to_pfn_atomic(struct kvm *kvm, gfn_t gfn);
 pfn_t gfn_to_pfn_async(struct kvm *kvm, gfn_t gfn, bool *async,
 		       bool write_fault, bool *writable);
 pfn_t gfn_to_pfn(struct kvm *kvm, gfn_t gfn);
 pfn_t gfn_to_pfn_prot(struct kvm *kvm, gfn_t gfn, bool write_fault,
 		      bool *writable);
-pfn_t gfn_to_pfn_memslot(struct kvm *kvm,
-			 struct kvm_memory_slot *slot, gfn_t gfn);
+pfn_t gfn_to_pfn_memslot(struct kvm_memory_slot *slot, gfn_t gfn);
 void kvm_release_pfn_dirty(pfn_t);
 void kvm_release_pfn_clean(pfn_t pfn);
 void kvm_set_pfn_dirty(pfn_t pfn);
diff --git a/virt/kvm/iommu.c b/virt/kvm/iommu.c
index e9fff98..c03f1fb 100644
--- a/virt/kvm/iommu.c
+++ b/virt/kvm/iommu.c
@@ -42,13 +42,13 @@ static int kvm_iommu_unmap_memslots(struct kvm *kvm);
 static void kvm_iommu_put_pages(struct kvm *kvm,
 				gfn_t base_gfn, unsigned long npages);
 
-static pfn_t kvm_pin_pages(struct kvm *kvm, struct kvm_memory_slot *slot,
-			   gfn_t gfn, unsigned long size)
+static pfn_t kvm_pin_pages(struct kvm_memory_slot *slot, gfn_t gfn,
+			   unsigned long size)
 {
 	gfn_t end_gfn;
 	pfn_t pfn;
 
-	pfn     = gfn_to_pfn_memslot(kvm, slot, gfn);
+	pfn     = gfn_to_pfn_memslot(slot, gfn);
 	end_gfn = gfn + (size >> PAGE_SHIFT);
 	gfn    += 1;
 
@@ -56,7 +56,7 @@ static pfn_t kvm_pin_pages(struct kvm *kvm, struct kvm_memory_slot *slot,
 		return pfn;
 
 	while (gfn < end_gfn)
-		gfn_to_pfn_memslot(kvm, slot, gfn++);
+		gfn_to_pfn_memslot(slot, gfn++);
 
 	return pfn;
 }
@@ -105,7 +105,7 @@ int kvm_iommu_map_pages(struct kvm *kvm, struct kvm_memory_slot *slot)
 		 * Pin all pages we are about to map in memory. This is
 		 * important because we unmap and unpin in 4kb steps later.
 		 */
-		pfn = kvm_pin_pages(kvm, slot, gfn, page_size);
+		pfn = kvm_pin_pages(slot, gfn, page_size);
 		if (is_error_pfn(pfn)) {
 			gfn += 1;
 			continue;
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index f955eee..68dda51 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -1062,8 +1062,8 @@ static inline int check_user_page_hwpoison(unsigned long addr)
 	return rc == -EHWPOISON;
 }
 
-static pfn_t hva_to_pfn(struct kvm *kvm, unsigned long addr, bool atomic,
-			bool *async, bool write_fault, bool *writable)
+static pfn_t hva_to_pfn(unsigned long addr, bool atomic, bool *async,
+			bool write_fault, bool *writable)
 {
 	struct page *page[1];
 	int npages = 0;
@@ -1143,9 +1143,9 @@ static pfn_t hva_to_pfn(struct kvm *kvm, unsigned long addr, bool atomic,
 	return pfn;
 }
 
-pfn_t hva_to_pfn_atomic(struct kvm *kvm, unsigned long addr)
+pfn_t hva_to_pfn_atomic(unsigned long addr)
 {
-	return hva_to_pfn(kvm, addr, true, NULL, true, NULL);
+	return hva_to_pfn(addr, true, NULL, true, NULL);
 }
 EXPORT_SYMBOL_GPL(hva_to_pfn_atomic);
 
@@ -1163,7 +1163,7 @@ static pfn_t __gfn_to_pfn(struct kvm *kvm, gfn_t gfn, bool atomic, bool *async,
 		return page_to_pfn(bad_page);
 	}
 
-	return hva_to_pfn(kvm, addr, atomic, async, write_fault, writable);
+	return hva_to_pfn(addr, atomic, async, write_fault, writable);
 }
 
 pfn_t gfn_to_pfn_atomic(struct kvm *kvm, gfn_t gfn)
@@ -1192,11 +1192,10 @@ pfn_t gfn_to_pfn_prot(struct kvm *kvm, gfn_t gfn, bool write_fault,
 }
 EXPORT_SYMBOL_GPL(gfn_to_pfn_prot);
 
-pfn_t gfn_to_pfn_memslot(struct kvm *kvm,
-			 struct kvm_memory_slot *slot, gfn_t gfn)
+pfn_t gfn_to_pfn_memslot(struct kvm_memory_slot *slot, gfn_t gfn)
 {
 	unsigned long addr = gfn_to_hva_memslot(slot, gfn);
-	return hva_to_pfn(kvm, addr, false, NULL, true, NULL);
+	return hva_to_pfn(addr, false, NULL, true, NULL);
 }
 
 int gfn_to_page_many_atomic(struct kvm *kvm, gfn_t gfn, struct page **pages,
-- 
cgit v0.10.2


From 0fa060714753ef65aef294b3462efb0212520933 Mon Sep 17 00:00:00 2001
From: Guo Chao <yan@linux.vnet.ibm.com>
Date: Thu, 28 Jun 2012 15:16:19 +0800
Subject: KVM: VMX: Fix typos

Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c
index c39b607..2300e53 100644
--- a/arch/x86/kvm/vmx.c
+++ b/arch/x86/kvm/vmx.c
@@ -1343,7 +1343,7 @@ static bool update_transition_efer(struct vcpu_vmx *vmx, int efer_offset)
 	guest_efer = vmx->vcpu.arch.efer;
 
 	/*
-	 * NX is emulated; LMA and LME handled by hardware; SCE meaninless
+	 * NX is emulated; LMA and LME handled by hardware; SCE meaningless
 	 * outside long mode
 	 */
 	ignore_bits = EFER_NX | EFER_SCE;
@@ -3261,7 +3261,7 @@ static void vmx_set_segment(struct kvm_vcpu *vcpu,
 	 * qemu binaries.
 	 *   IA32 arch specifies that at the time of processor reset the
 	 * "Accessed" bit in the AR field of segment registers is 1. And qemu
-	 * is setting it to 0 in the usedland code. This causes invalid guest
+	 * is setting it to 0 in the userland code. This causes invalid guest
 	 * state vmexit when "unrestricted guest" mode is turned on.
 	 *    Fix for this setup issue in cpu_reset is being pushed in the qemu
 	 * tree. Newer qemu binaries with that qemu fix would not need this
@@ -4446,7 +4446,7 @@ vmx_patch_hypercall(struct kvm_vcpu *vcpu, unsigned char *hypercall)
 	hypercall[2] = 0xc1;
 }
 
-/* called to set cr0 as approriate for a mov-to-cr0 exit. */
+/* called to set cr0 as appropriate for a mov-to-cr0 exit. */
 static int handle_set_cr0(struct kvm_vcpu *vcpu, unsigned long val)
 {
 	if (to_vmx(vcpu)->nested.vmxon &&
-- 
cgit v0.10.2


From c5ec2e56d0a654797ee43a31001738ccd05eef0b Mon Sep 17 00:00:00 2001
From: Guo Chao <yan@linux.vnet.ibm.com>
Date: Thu, 28 Jun 2012 15:16:43 +0800
Subject: KVM: SVM: Fix typos

Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c
index baead95..687d0c3 100644
--- a/arch/x86/kvm/svm.c
+++ b/arch/x86/kvm/svm.c
@@ -2063,7 +2063,7 @@ static inline bool nested_svm_intr(struct vcpu_svm *svm)
 	if (svm->nested.intercept & 1ULL) {
 		/*
 		 * The #vmexit can't be emulated here directly because this
-		 * code path runs with irqs and preemtion disabled. A
+		 * code path runs with irqs and preemption disabled. A
 		 * #vmexit emulation might sleep. Only signal request for
 		 * the #vmexit here.
 		 */
@@ -2409,7 +2409,7 @@ static bool nested_svm_vmrun_msrpm(struct vcpu_svm *svm)
 {
 	/*
 	 * This function merges the msr permission bitmaps of kvm and the
-	 * nested vmcb. It is omptimized in that it only merges the parts where
+	 * nested vmcb. It is optimized in that it only merges the parts where
 	 * the kvm msr permission bitmap may contain zero bits
 	 */
 	int i;
-- 
cgit v0.10.2


From 4a9699807c491740c4dfe7b6a06703e1d262e802 Mon Sep 17 00:00:00 2001
From: Guo Chao <yan@linux.vnet.ibm.com>
Date: Thu, 28 Jun 2012 15:17:27 +0800
Subject: KVM: x86: Fix typos in x86.c

Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index ecc71dd..3d9d08ed 100644
--- a/arch/x86/kvm/x86.c
+++ b/arch/x86/kvm/x86.c
@@ -1093,7 +1093,7 @@ void kvm_write_tsc(struct kvm_vcpu *vcpu, u64 data)
 		 * For each generation, we track the original measured
 		 * nanosecond time, offset, and write, so if TSCs are in
 		 * sync, we can match exact offset, and if not, we can match
-		 * exact software computaion in compute_guest_tsc()
+		 * exact software computation in compute_guest_tsc()
 		 *
 		 * These values are tracked in kvm->arch.cur_xxx variables.
 		 */
@@ -1500,7 +1500,7 @@ static int kvm_pv_enable_async_pf(struct kvm_vcpu *vcpu, u64 data)
 {
 	gpa_t gpa = data & ~0x3f;
 
-	/* Bits 2:5 are resrved, Should be zero */
+	/* Bits 2:5 are reserved, Should be zero */
 	if (data & 0x3c)
 		return 1;
 
@@ -1723,7 +1723,7 @@ int kvm_set_msr_common(struct kvm_vcpu *vcpu, u32 msr, u64 data)
 		 * Ignore all writes to this no longer documented MSR.
 		 * Writes are only relevant for old K7 processors,
 		 * all pre-dating SVM, but a recommended workaround from
-		 * AMD for these chips. It is possible to speicify the
+		 * AMD for these chips. It is possible to specify the
 		 * affected processor models on the command line, hence
 		 * the need to ignore the workaround.
 		 */
@@ -4491,7 +4491,7 @@ static bool reexecute_instruction(struct kvm_vcpu *vcpu, gva_t gva)
 
 	/*
 	 * if emulation was due to access to shadowed page table
-	 * and it failed try to unshadow page and re-entetr the
+	 * and it failed try to unshadow page and re-enter the
 	 * guest to let CPU execute the instruction.
 	 */
 	if (kvm_mmu_unprotect_page_virt(vcpu, gva))
@@ -5587,7 +5587,7 @@ int kvm_arch_vcpu_ioctl_get_regs(struct kvm_vcpu *vcpu, struct kvm_regs *regs)
 		/*
 		 * We are here if userspace calls get_regs() in the middle of
 		 * instruction emulation. Registers state needs to be copied
-		 * back from emulation context to vcpu. Usrapace shouldn't do
+		 * back from emulation context to vcpu. Userspace shouldn't do
 		 * that usually, but some bad designed PV devices (vmware
 		 * backdoor interface) need this to work
 		 */
@@ -6116,7 +6116,7 @@ int kvm_arch_hardware_enable(void *garbage)
 	 * as we reset last_host_tsc on all VCPUs to stop this from being
 	 * called multiple times (one for each physical CPU bringup).
 	 *
-	 * Platforms with unnreliable TSCs don't have to deal with this, they
+	 * Platforms with unreliable TSCs don't have to deal with this, they
 	 * will be compensated by the logic in vcpu_load, which sets the TSC to
 	 * catchup mode.  This will catchup all VCPUs to real time, but cannot
 	 * guarantee that they stay in perfect synchronization.
@@ -6391,7 +6391,7 @@ int kvm_arch_prepare_memory_region(struct kvm *kvm,
 		map_flags = MAP_SHARED | MAP_ANONYMOUS;
 
 	/*To keep backward compatibility with older userspace,
-	 *x86 needs to hanlde !user_alloc case.
+	 *x86 needs to handle !user_alloc case.
 	 */
 	if (!user_alloc) {
 		if (npages && !old.rmap) {
-- 
cgit v0.10.2


From fc0586807dc4e307da6d3ba4ed5c927b6d27276c Mon Sep 17 00:00:00 2001
From: Guo Chao <yan@linux.vnet.ibm.com>
Date: Thu, 28 Jun 2012 15:19:51 +0800
Subject: KVM: x86: Fix typos in emulate.c

Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index 97d9a99..85b611e 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -642,7 +642,7 @@ static int __linearize(struct x86_emulate_ctxt *ctxt,
 			if (addr.ea > lim || (u32)(addr.ea + size - 1) > lim)
 				goto bad;
 		} else {
-			/* exapand-down segment */
+			/* expand-down segment */
 			if (addr.ea <= lim || (u32)(addr.ea + size - 1) <= lim)
 				goto bad;
 			lim = desc.d ? 0xffffffff : 0xffff;
@@ -1383,7 +1383,7 @@ static int load_segment_descriptor(struct x86_emulate_ctxt *ctxt,
 	err_code = selector & 0xfffc;
 	err_vec = GP_VECTOR;
 
-	/* can't load system descriptor into segment selecor */
+	/* can't load system descriptor into segment selector */
 	if (seg <= VCPU_SREG_GS && !seg_desc.s)
 		goto exception;
 
@@ -2398,7 +2398,7 @@ static int load_state_from_tss16(struct x86_emulate_ctxt *ctxt,
 	set_segment_selector(ctxt, tss->ds, VCPU_SREG_DS);
 
 	/*
-	 * Now load segment descriptors. If fault happenes at this stage
+	 * Now load segment descriptors. If fault happens at this stage
 	 * it is handled in a context of new task
 	 */
 	ret = load_segment_descriptor(ctxt, tss->ldt, VCPU_SREG_LDTR);
@@ -2640,7 +2640,7 @@ static int emulator_do_task_switch(struct x86_emulate_ctxt *ctxt,
 	 *
 	 * 1. jmp/call/int to task gate: Check against DPL of the task gate
 	 * 2. Exception/IRQ/iret: No check is performed
-	 * 3. jmp/call to TSS: Check agains DPL of the TSS
+	 * 3. jmp/call to TSS: Check against DPL of the TSS
 	 */
 	if (reason == TASK_SWITCH_GATE) {
 		if (idt_index != -1) {
@@ -2681,7 +2681,7 @@ static int emulator_do_task_switch(struct x86_emulate_ctxt *ctxt,
 		ctxt->eflags = ctxt->eflags & ~X86_EFLAGS_NT;
 
 	/* set back link to prev task only if NT bit is set in eflags
-	   note that old_tss_sel is not used afetr this point */
+	   note that old_tss_sel is not used after this point */
 	if (reason != TASK_SWITCH_CALL && reason != TASK_SWITCH_GATE)
 		old_tss_sel = 0xffff;
 
-- 
cgit v0.10.2


From bbbda79510b6e5d399fae76bdb9b999286eb1b59 Mon Sep 17 00:00:00 2001
From: Guo Chao <yan@linux.vnet.ibm.com>
Date: Thu, 28 Jun 2012 15:20:58 +0800
Subject: KVM: x86: Fix typos in cpuid.c

Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c
index 0595f13..b496da6 100644
--- a/arch/x86/kvm/cpuid.c
+++ b/arch/x86/kvm/cpuid.c
@@ -316,7 +316,7 @@ static int do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function,
 	}
 	case 7: {
 		entry->flags |= KVM_CPUID_FLAG_SIGNIFCANT_INDEX;
-		/* Mask ebx against host capbability word 9 */
+		/* Mask ebx against host capability word 9 */
 		if (index == 0) {
 			entry->ebx &= kvm_supported_word9_x86_features;
 			cpuid_mask(&entry->ebx, 9);
-- 
cgit v0.10.2


From d5b0b5b196b474ef26f46f2036c8fbaa8cca2cf5 Mon Sep 17 00:00:00 2001
From: Guo Chao <yan@linux.vnet.ibm.com>
Date: Thu, 28 Jun 2012 15:22:57 +0800
Subject: KVM: x86: Fix typos in lapic.c

Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c
index ce87878..fff7173 100644
--- a/arch/x86/kvm/lapic.c
+++ b/arch/x86/kvm/lapic.c
@@ -719,7 +719,7 @@ static int apic_reg_read(struct kvm_lapic *apic, u32 offset, int len,
 {
 	unsigned char alignment = offset & 0xf;
 	u32 result;
-	/* this bitmask has a bit cleared for each reserver register */
+	/* this bitmask has a bit cleared for each reserved register */
 	static const u64 rmask = 0x43ff01ffffffe70cULL;
 
 	if ((alignment + len) > 4) {
@@ -792,7 +792,7 @@ static void start_apic_timer(struct kvm_lapic *apic)
 	atomic_set(&apic->lapic_timer.pending, 0);
 
 	if (apic_lvtt_period(apic) || apic_lvtt_oneshot(apic)) {
-		/* lapic timer in oneshot or peroidic mode */
+		/* lapic timer in oneshot or periodic mode */
 		now = apic->lapic_timer.timer.base->get_time();
 		apic->lapic_timer.period = (u64)apic_get_reg(apic, APIC_TMICT)
 			    * APIC_BUS_CYCLE_NS * apic->divide_count;
-- 
cgit v0.10.2


From c7a7062fa00db7dc66280a72cd9dad0f3595bc66 Mon Sep 17 00:00:00 2001
From: Guo Chao <yan@linux.vnet.ibm.com>
Date: Thu, 28 Jun 2012 15:23:08 +0800
Subject: KVM: x86: Fix typos in pmu.c

Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/pmu.c b/arch/x86/kvm/pmu.c
index 2e88438..db206a4 100644
--- a/arch/x86/kvm/pmu.c
+++ b/arch/x86/kvm/pmu.c
@@ -1,5 +1,5 @@
 /*
- * Kernel-based Virtual Machine -- Performane Monitoring Unit support
+ * Kernel-based Virtual Machine -- Performance Monitoring Unit support
  *
  * Copyright 2011 Red Hat, Inc. and/or its affiliates.
  *
-- 
cgit v0.10.2


From 93b6547e2219784b2df790353e083e0bdbebd2c2 Mon Sep 17 00:00:00 2001
From: "Michael S. Tsirkin" <mst@redhat.com>
Date: Thu, 19 Jul 2012 13:55:53 +0300
Subject: KVM: switch to symbolic name for irq_states size

Use PIC_NUM_PINS instead of hard-coded 16 for pic pins.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

diff --git a/arch/x86/kvm/irq.h b/arch/x86/kvm/irq.h
index 2086f2b..2d03568 100644
--- a/arch/x86/kvm/irq.h
+++ b/arch/x86/kvm/irq.h
@@ -70,7 +70,7 @@ struct kvm_pic {
 	struct kvm_io_device dev_slave;
 	struct kvm_io_device dev_eclr;
 	void (*ack_notifier)(void *opaque, int irq);
-	unsigned long irq_states[16];
+	unsigned long irq_states[PIC_NUM_PINS];
 };
 
 struct kvm_pic *kvm_create_pic(struct kvm *kvm);
diff --git a/virt/kvm/irq_comm.c b/virt/kvm/irq_comm.c
index a6a0365..22aae8f 100644
--- a/virt/kvm/irq_comm.c
+++ b/virt/kvm/irq_comm.c
@@ -343,11 +343,11 @@ static int setup_routing_entry(struct kvm_irq_routing_table *rt,
 		switch (ue->u.irqchip.irqchip) {
 		case KVM_IRQCHIP_PIC_MASTER:
 			e->set = kvm_set_pic_irq;
-			max_pin = 16;
+			max_pin = PIC_NUM_PINS;
 			break;
 		case KVM_IRQCHIP_PIC_SLAVE:
 			e->set = kvm_set_pic_irq;
-			max_pin = 16;
+			max_pin = PIC_NUM_PINS;
 			delta = 8;
 			break;
 		case KVM_IRQCHIP_IOAPIC:
-- 
cgit v0.10.2


From f2a743473194a1ad44a85f8b63aeef9d63e5bf47 Mon Sep 17 00:00:00 2001
From: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Date: Wed, 18 Jul 2012 19:07:32 +0530
Subject: KVM: Add config to support ple or cpu relax optimzation

Suggested-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Reviewed-by: Marcelo Tosatti <mtosatti@redhat.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> # on s390x
Signed-off-by: Avi Kivity <avi@redhat.com>

diff --git a/arch/s390/kvm/Kconfig b/arch/s390/kvm/Kconfig
index 78eb984..a6e2677 100644
--- a/arch/s390/kvm/Kconfig
+++ b/arch/s390/kvm/Kconfig
@@ -21,6 +21,7 @@ config KVM
 	depends on HAVE_KVM && EXPERIMENTAL
 	select PREEMPT_NOTIFIERS
 	select ANON_INODES
+	select HAVE_KVM_CPU_RELAX_INTERCEPT
 	---help---
 	  Support hosting paravirtualized guest machines using the SIE
 	  virtualization capability on the mainframe. This should work
diff --git a/arch/x86/kvm/Kconfig b/arch/x86/kvm/Kconfig
index a28f338..45c044f 100644
--- a/arch/x86/kvm/Kconfig
+++ b/arch/x86/kvm/Kconfig
@@ -37,6 +37,7 @@ config KVM
 	select TASK_DELAY_ACCT
 	select PERF_EVENTS
 	select HAVE_KVM_MSI
+	select HAVE_KVM_CPU_RELAX_INTERCEPT
 	---help---
 	  Support hosting fully virtualized guest machines using hardware
 	  virtualization extensions.  You will need a fairly recent
diff --git a/virt/kvm/Kconfig b/virt/kvm/Kconfig
index 28694f4..d01b24b 100644
--- a/virt/kvm/Kconfig
+++ b/virt/kvm/Kconfig
@@ -21,3 +21,6 @@ config KVM_ASYNC_PF
 
 config HAVE_KVM_MSI
        bool
+
+config HAVE_KVM_CPU_RELAX_INTERCEPT
+       bool
-- 
cgit v0.10.2


From 4c088493c8d07e4e27bad53a99dcfdc14cdf45f8 Mon Sep 17 00:00:00 2001
From: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Date: Wed, 18 Jul 2012 19:07:46 +0530
Subject: KVM: Note down when cpu relax intercepted or pause loop exited

Noting pause loop exited vcpu or cpu relax intercepted helps in
filtering right candidate to yield. Wrong selection of vcpu;
i.e., a vcpu that just did a pl-exit or cpu relax intercepted may
contribute to performance degradation.

Signed-off-by: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Reviewed-by: Marcelo Tosatti <mtosatti@redhat.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> # on s390x
Signed-off-by: Avi Kivity <avi@redhat.com>

diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index db9aa917..361b36f 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -183,6 +183,18 @@ struct kvm_vcpu {
 	} async_pf;
 #endif
 
+#ifdef CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT
+	/*
+	 * Cpu relax intercept or pause loop exit optimization
+	 * in_spin_loop: set when a vcpu does a pause loop exit
+	 *  or cpu relax intercepted.
+	 * dy_eligible: indicates whether vcpu is eligible for directed yield.
+	 */
+	struct {
+		bool in_spin_loop;
+		bool dy_eligible;
+	} spin_loop;
+#endif
 	struct kvm_vcpu_arch arch;
 };
 
@@ -898,5 +910,27 @@ static inline bool kvm_check_request(int req, struct kvm_vcpu *vcpu)
 	}
 }
 
+#ifdef CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT
+
+static inline void kvm_vcpu_set_in_spin_loop(struct kvm_vcpu *vcpu, bool val)
+{
+	vcpu->spin_loop.in_spin_loop = val;
+}
+static inline void kvm_vcpu_set_dy_eligible(struct kvm_vcpu *vcpu, bool val)
+{
+	vcpu->spin_loop.dy_eligible = val;
+}
+
+#else /* !CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT */
+
+static inline void kvm_vcpu_set_in_spin_loop(struct kvm_vcpu *vcpu, bool val)
+{
+}
+
+static inline void kvm_vcpu_set_dy_eligible(struct kvm_vcpu *vcpu, bool val)
+{
+}
+
+#endif /* CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT */
 #endif
 
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 68dda51..0892b75 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -239,6 +239,9 @@ int kvm_vcpu_init(struct kvm_vcpu *vcpu, struct kvm *kvm, unsigned id)
 	}
 	vcpu->run = page_address(page);
 
+	kvm_vcpu_set_in_spin_loop(vcpu, false);
+	kvm_vcpu_set_dy_eligible(vcpu, false);
+
 	r = kvm_arch_vcpu_init(vcpu);
 	if (r < 0)
 		goto fail_free_run;
@@ -1585,6 +1588,7 @@ void kvm_vcpu_on_spin(struct kvm_vcpu *me)
 	int pass;
 	int i;
 
+	kvm_vcpu_set_in_spin_loop(me, true);
 	/*
 	 * We boost the priority of a VCPU that is runnable but not
 	 * currently running, because it got preempted by something
@@ -1610,6 +1614,7 @@ void kvm_vcpu_on_spin(struct kvm_vcpu *me)
 			}
 		}
 	}
+	kvm_vcpu_set_in_spin_loop(me, false);
 }
 EXPORT_SYMBOL_GPL(kvm_vcpu_on_spin);
 
-- 
cgit v0.10.2


From 06e48c510aa37f6e791602e6420422ea7071fe94 Mon Sep 17 00:00:00 2001
From: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Date: Thu, 19 Jul 2012 15:17:52 +0530
Subject: KVM: Choose better candidate for directed yield

Currently, on a large vcpu guests, there is a high probability of
yielding to the same vcpu who had recently done a pause-loop exit or
cpu relax intercepted. Such a yield can lead to the vcpu spinning
again and hence degrade the performance.

The patchset keeps track of the pause loop exit/cpu relax interception
and gives chance to a vcpu which:
 (a) Has not done pause loop exit or cpu relax intercepted at all
     (probably he is preempted lock-holder)
 (b) Was skipped in last iteration because it did pause loop exit or
     cpu relax intercepted, and probably has become eligible now
     (next eligible lock holder)

Signed-off-by: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com>
Reviewed-by: Marcelo Tosatti <mtosatti@redhat.com>
Reviewed-by: Rik van Riel <riel@redhat.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com> # on s390x
Signed-off-by: Avi Kivity <avi@redhat.com>

diff --git a/include/linux/kvm_host.h b/include/linux/kvm_host.h
index 361b36f..74a78d0 100644
--- a/include/linux/kvm_host.h
+++ b/include/linux/kvm_host.h
@@ -931,6 +931,11 @@ static inline void kvm_vcpu_set_dy_eligible(struct kvm_vcpu *vcpu, bool val)
 {
 }
 
+static inline bool kvm_vcpu_eligible_for_directed_yield(struct kvm_vcpu *vcpu)
+{
+	return true;
+}
+
 #endif /* CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT */
 #endif
 
diff --git a/virt/kvm/kvm_main.c b/virt/kvm/kvm_main.c
index 0892b75..1e10ebe 100644
--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -1579,6 +1579,43 @@ bool kvm_vcpu_yield_to(struct kvm_vcpu *target)
 }
 EXPORT_SYMBOL_GPL(kvm_vcpu_yield_to);
 
+#ifdef CONFIG_HAVE_KVM_CPU_RELAX_INTERCEPT
+/*
+ * Helper that checks whether a VCPU is eligible for directed yield.
+ * Most eligible candidate to yield is decided by following heuristics:
+ *
+ *  (a) VCPU which has not done pl-exit or cpu relax intercepted recently
+ *  (preempted lock holder), indicated by @in_spin_loop.
+ *  Set at the beiginning and cleared at the end of interception/PLE handler.
+ *
+ *  (b) VCPU which has done pl-exit/ cpu relax intercepted but did not get
+ *  chance last time (mostly it has become eligible now since we have probably
+ *  yielded to lockholder in last iteration. This is done by toggling
+ *  @dy_eligible each time a VCPU checked for eligibility.)
+ *
+ *  Yielding to a recently pl-exited/cpu relax intercepted VCPU before yielding
+ *  to preempted lock-holder could result in wrong VCPU selection and CPU
+ *  burning. Giving priority for a potential lock-holder increases lock
+ *  progress.
+ *
+ *  Since algorithm is based on heuristics, accessing another VCPU data without
+ *  locking does not harm. It may result in trying to yield to  same VCPU, fail
+ *  and continue with next VCPU and so on.
+ */
+bool kvm_vcpu_eligible_for_directed_yield(struct kvm_vcpu *vcpu)
+{
+	bool eligible;
+
+	eligible = !vcpu->spin_loop.in_spin_loop ||
+			(vcpu->spin_loop.in_spin_loop &&
+			 vcpu->spin_loop.dy_eligible);
+
+	if (vcpu->spin_loop.in_spin_loop)
+		kvm_vcpu_set_dy_eligible(vcpu, !vcpu->spin_loop.dy_eligible);
+
+	return eligible;
+}
+#endif
 void kvm_vcpu_on_spin(struct kvm_vcpu *me)
 {
 	struct kvm *kvm = me->kvm;
@@ -1607,6 +1644,8 @@ void kvm_vcpu_on_spin(struct kvm_vcpu *me)
 				continue;
 			if (waitqueue_active(&vcpu->wq))
 				continue;
+			if (!kvm_vcpu_eligible_for_directed_yield(vcpu))
+				continue;
 			if (kvm_vcpu_yield_to(vcpu)) {
 				kvm->last_boosted_vcpu = i;
 				yielded = 1;
@@ -1615,6 +1654,9 @@ void kvm_vcpu_on_spin(struct kvm_vcpu *me)
 		}
 	}
 	kvm_vcpu_set_in_spin_loop(me, false);
+
+	/* Ensure vcpu is not eligible during next spinloop */
+	kvm_vcpu_set_dy_eligible(me, false);
 }
 EXPORT_SYMBOL_GPL(kvm_vcpu_on_spin);
 
-- 
cgit v0.10.2