KVM: Add coalesced MMIO support (common part)

This patch adds all needed structures to coalesce MMIOs. Until an architecture uses it, it is not compiled. Coalesced MMIO introduces two ioctl() to define where are the MMIO zones that can be coalesced: - KVM_REGISTER_COALESCED_MMIO registers a coalesced MMIO zone. It requests one parameter (struct kvm_coalesced_mmio_zone) which defines a memory area where MMIOs can be coalesced until the next switch to user space. The maximum number of MMIO zones is KVM_COALESCED_MMIO_ZONE_MAX. - KVM_UNREGISTER_COALESCED_MMIO cancels all registered zones inside the given bounds (bounds are also given by struct kvm_coalesced_mmio_zone). The userspace client can check kernel coalesced MMIO availability by asking ioctl(KVM_CHECK_EXTENSION) for the KVM_CAP_COALESCED_MMIO capability. The ioctl() call to KVM_CAP_COALESCED_MMIO will return 0 if not supported, or the page offset where will be stored the ring buffer. The page offset depends on the architecture. After an ioctl(KVM_RUN), the first page of the KVM memory mapped points to a kvm_run structure. The offset given by KVM_CAP_COALESCED_MMIO is an offset to the coalesced MMIO ring expressed in PAGE_SIZE relatively to the address of the start of th kvm_run structure. The MMIO ring buffer is defined by the structure kvm_coalesced_mmio_ring. [akio: fix oops during guest shutdown] Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net> Signed-off-by: Akio Takebe <takebe_akio@jp.fujitsu.com> Signed-off-by: Avi Kivity <avi@qumranet.com>

KVM: Add coalesced MMIO support (common part)
This patch adds all needed structures to coalesce MMIOs. Until an architecture uses it, it is not compiled. Coalesced MMIO introduces two ioctl() to define where are the MMIO zones that can be coalesced: - KVM_REGISTER_COALESCED_MMIO registers a coalesced MMIO zone. It requests one parameter (struct kvm_coalesced_mmio_zone) which defines a memory area where MMIOs can be coalesced until the next switch to user space. The maximum number of MMIO zones is KVM_COALESCED_MMIO_ZONE_MAX. - KVM_UNREGISTER_COALESCED_MMIO cancels all registered zones inside the given bounds (bounds are also given by struct kvm_coalesced_mmio_zone). The userspace client can check kernel coalesced MMIO availability by asking ioctl(KVM_CHECK_EXTENSION) for the KVM_CAP_COALESCED_MMIO capability. The ioctl() call to KVM_CAP_COALESCED_MMIO will return 0 if not supported, or the page offset where will be stored the ring buffer. The page offset depends on the architecture. After an ioctl(KVM_RUN), the first page of the KVM memory mapped points to a kvm_run structure. The offset given by KVM_CAP_COALESCED_MMIO is an offset to the coalesced MMIO ring expressed in PAGE_SIZE relatively to the address of the start of th kvm_run structure. The MMIO ring buffer is defined by the structure kvm_coalesced_mmio_ring. [akio: fix oops during guest shutdown] Signed-off-by: Laurent Vivier <Laurent.Vivier@bull.net> Signed-off-by: Akio Takebe <takebe_akio@jp.fujitsu.com> Signed-off-by: Avi Kivity <avi@qumranet.com>
Laurent Vivier · Avi Kivity
1 parent 92760499d0
Showing 5 changed files with 269 additions and 0 deletions Side-by-side Diff
include/linux/kvm.h
include/linux/kvm_host.h
virt/kvm/coalesced_mmio.c
virt/kvm/coalesced_mmio.h
virt/kvm/kvm_main.c
@@ -173,6 +173,30 @@
 	};
 };
  
+/* for KVM_REGISTER_COALESCED_MMIO / KVM_UNREGISTER_COALESCED_MMIO */
+
+struct kvm_coalesced_mmio_zone {
+	__u64 addr;
+	__u32 size;
+	__u32 pad;
+};
+
+struct kvm_coalesced_mmio {
+	__u64 phys_addr;
+	__u32 len;
+	__u32 pad;
+	__u8  data[8];
+};
+
+struct kvm_coalesced_mmio_ring {
+	__u32 first, last;
+	struct kvm_coalesced_mmio coalesced_mmio[0];
+};
+
+#define KVM_COALESCED_MMIO_MAX \
+	((PAGE_SIZE - sizeof(struct kvm_coalesced_mmio_ring)) / \
+	 sizeof(struct kvm_coalesced_mmio))
+
 /* for KVM_TRANSLATE */
 struct kvm_translation {
 	/* in */
@@ -346,6 +370,7 @@
 #define KVM_CAP_NOP_IO_DELAY 12
 #define KVM_CAP_PV_MMU 13
 #define KVM_CAP_MP_STATE 14
+#define KVM_CAP_COALESCED_MMIO 15
  
 /*
  * ioctls for VM fds
@@ -371,6 +396,10 @@
 #define KVM_CREATE_PIT		  _IO(KVMIO,  0x64)
 #define KVM_GET_PIT		  _IOWR(KVMIO, 0x65, struct kvm_pit_state)
 #define KVM_SET_PIT		  _IOR(KVMIO,  0x66, struct kvm_pit_state)
+#define KVM_REGISTER_COALESCED_MMIO \
+			_IOW(KVMIO,  0x67, struct kvm_coalesced_mmio_zone)
+#define KVM_UNREGISTER_COALESCED_MMIO \
+			_IOW(KVMIO,  0x68, struct kvm_coalesced_mmio_zone)
  
 /*
  * ioctls for vcpu fds
@@ -117,6 +117,10 @@
 	struct kvm_vm_stat stat;
 	struct kvm_arch arch;
 	atomic_t users_count;
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+	struct kvm_coalesced_mmio_dev *coalesced_mmio_dev;
+	struct kvm_coalesced_mmio_ring *coalesced_mmio_ring;
+#endif
 };
  
 /* The guest did something we don't support. */
+/*
+ * KVM coalesced MMIO
+ *
+ * Copyright (c) 2008 Bull S.A.S.
+ *
+ *  Author: Laurent Vivier <Laurent.Vivier@bull.net>
+ *
+ */
+
+#include "iodev.h"
+
+#include <linux/kvm_host.h>
+#include <linux/kvm.h>
+
+#include "coalesced_mmio.h"
+
+static int coalesced_mmio_in_range(struct kvm_io_device *this,
+				   gpa_t addr, int len, int is_write)
+{
+	struct kvm_coalesced_mmio_dev *dev =
+				(struct kvm_coalesced_mmio_dev*)this->private;
+	struct kvm_coalesced_mmio_zone *zone;
+	int next;
+	int i;
+
+	if (!is_write)
+		return 0;
+
+	/* kvm->lock is taken by the caller and must be not released before
+         * dev.read/write
+         */
+
+	/* Are we able to batch it ? */
+
+	/* last is the first free entry
+	 * check if we don't meet the first used entry
+	 * there is always one unused entry in the buffer
+	 */
+
+	next = (dev->kvm->coalesced_mmio_ring->last + 1) %
+							KVM_COALESCED_MMIO_MAX;
+	if (next == dev->kvm->coalesced_mmio_ring->first) {
+		/* full */
+		return 0;
+	}
+
+	/* is it in a batchable area ? */
+
+	for (i = 0; i < dev->nb_zones; i++) {
+		zone = &dev->zone[i];
+
+		/* (addr,len) is fully included in
+		 * (zone->addr, zone->size)
+		 */
+
+		if (zone->addr <= addr &&
+		    addr + len <= zone->addr + zone->size)
+			return 1;
+	}
+	return 0;
+}
+
+static void coalesced_mmio_write(struct kvm_io_device *this,
+				 gpa_t addr, int len, const void *val)
+{
+	struct kvm_coalesced_mmio_dev *dev =
+				(struct kvm_coalesced_mmio_dev*)this->private;
+	struct kvm_coalesced_mmio_ring *ring = dev->kvm->coalesced_mmio_ring;
+
+	/* kvm->lock must be taken by caller before call to in_range()*/
+
+	/* copy data in first free entry of the ring */
+
+	ring->coalesced_mmio[ring->last].phys_addr = addr;
+	ring->coalesced_mmio[ring->last].len = len;
+	memcpy(ring->coalesced_mmio[ring->last].data, val, len);
+	smp_wmb();
+	ring->last = (ring->last + 1) % KVM_COALESCED_MMIO_MAX;
+}
+
+static void coalesced_mmio_destructor(struct kvm_io_device *this)
+{
+	kfree(this);
+}
+
+int kvm_coalesced_mmio_init(struct kvm *kvm)
+{
+	struct kvm_coalesced_mmio_dev *dev;
+
+	dev = kzalloc(sizeof(struct kvm_coalesced_mmio_dev), GFP_KERNEL);
+	if (!dev)
+		return -ENOMEM;
+	dev->dev.write  = coalesced_mmio_write;
+	dev->dev.in_range  = coalesced_mmio_in_range;
+	dev->dev.destructor  = coalesced_mmio_destructor;
+	dev->dev.private  = dev;
+	dev->kvm = kvm;
+	kvm->coalesced_mmio_dev = dev;
+	kvm_io_bus_register_dev(&kvm->mmio_bus, &dev->dev);
+
+	return 0;
+}
+
+int kvm_vm_ioctl_register_coalesced_mmio(struct kvm *kvm,
+				         struct kvm_coalesced_mmio_zone *zone)
+{
+	struct kvm_coalesced_mmio_dev *dev = kvm->coalesced_mmio_dev;
+
+	if (dev == NULL)
+		return -EINVAL;
+
+	mutex_lock(&kvm->lock);
+	if (dev->nb_zones >= KVM_COALESCED_MMIO_ZONE_MAX) {
+		mutex_unlock(&kvm->lock);
+		return -ENOBUFS;
+	}
+
+	dev->zone[dev->nb_zones] = *zone;
+	dev->nb_zones++;
+
+	mutex_unlock(&kvm->lock);
+	return 0;
+}
+
+int kvm_vm_ioctl_unregister_coalesced_mmio(struct kvm *kvm,
+					   struct kvm_coalesced_mmio_zone *zone)
+{
+	int i;
+	struct kvm_coalesced_mmio_dev *dev = kvm->coalesced_mmio_dev;
+	struct kvm_coalesced_mmio_zone *z;
+
+	if (dev == NULL)
+		return -EINVAL;
+
+	mutex_lock(&kvm->lock);
+
+	i = dev->nb_zones;
+	while(i) {
+		z = &dev->zone[i - 1];
+
+		/* unregister all zones
+		 * included in (zone->addr, zone->size)
+		 */
+
+		if (zone->addr <= z->addr &&
+		    z->addr + z->size <= zone->addr + zone->size) {
+			dev->nb_zones--;
+			*z = dev->zone[dev->nb_zones];
+		}
+		i--;
+	}
+
+	mutex_unlock(&kvm->lock);
+
+	return 0;
+}
+/*
+ * KVM coalesced MMIO
+ *
+ * Copyright (c) 2008 Bull S.A.S.
+ *
+ *  Author: Laurent Vivier <Laurent.Vivier@bull.net>
+ *
+ */
+
+#define KVM_COALESCED_MMIO_ZONE_MAX 100
+
+struct kvm_coalesced_mmio_dev {
+	struct kvm_io_device dev;
+	struct kvm *kvm;
+	int nb_zones;
+	struct kvm_coalesced_mmio_zone zone[KVM_COALESCED_MMIO_ZONE_MAX];
+};
+
+int kvm_coalesced_mmio_init(struct kvm *kvm);
+int kvm_vm_ioctl_register_coalesced_mmio(struct kvm *kvm,
+                                       struct kvm_coalesced_mmio_zone *zone);
+int kvm_vm_ioctl_unregister_coalesced_mmio(struct kvm *kvm,
+                                         struct kvm_coalesced_mmio_zone *zone);
@@ -47,6 +47,10 @@
 #include <asm/uaccess.h>
 #include <asm/pgtable.h>
  
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+#include "coalesced_mmio.h"
+#endif
+
 MODULE_AUTHOR("Qumranet");
 MODULE_LICENSE("GPL");
  
  
@@ -185,10 +189,23 @@
 static struct kvm *kvm_create_vm(void)
 {
 	struct kvm *kvm = kvm_arch_create_vm();
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+	struct page *page;
+#endif
  
 	if (IS_ERR(kvm))
 		goto out;
  
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+	page = alloc_page(GFP_KERNEL | __GFP_ZERO);
+	if (!page) {
+		kfree(kvm);
+		return ERR_PTR(-ENOMEM);
+	}
+	kvm->coalesced_mmio_ring =
+			(struct kvm_coalesced_mmio_ring *)page_address(page);
+#endif
+
 	kvm->mm = current->mm;
 	atomic_inc(&kvm->mm->mm_count);
 	spin_lock_init(&kvm->mmu_lock);
@@ -200,6 +217,9 @@
 	spin_lock(&kvm_lock);
 	list_add(&kvm->vm_list, &vm_list);
 	spin_unlock(&kvm_lock);
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+	kvm_coalesced_mmio_init(kvm);
+#endif
 out:
 	return kvm;
 }
@@ -242,6 +262,10 @@
 	spin_unlock(&kvm_lock);
 	kvm_io_bus_destroy(&kvm->pio_bus);
 	kvm_io_bus_destroy(&kvm->mmio_bus);
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+	if (kvm->coalesced_mmio_ring != NULL)
+		free_page((unsigned long)kvm->coalesced_mmio_ring);
+#endif
 	kvm_arch_destroy_vm(kvm);
 	mmdrop(mm);
 }
@@ -826,6 +850,10 @@
 	else if (vmf->pgoff == KVM_PIO_PAGE_OFFSET)
 		page = virt_to_page(vcpu->arch.pio_data);
 #endif
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+	else if (vmf->pgoff == KVM_COALESCED_MMIO_PAGE_OFFSET)
+		page = virt_to_page(vcpu->kvm->coalesced_mmio_ring);
+#endif
 	else
 		return VM_FAULT_SIGBUS;
 	get_page(page);
@@ -1148,6 +1176,32 @@
 			goto out;
 		break;
 	}
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+	case KVM_REGISTER_COALESCED_MMIO: {
+		struct kvm_coalesced_mmio_zone zone;
+		r = -EFAULT;
+		if (copy_from_user(&zone, argp, sizeof zone))
+			goto out;
+		r = -ENXIO;
+		r = kvm_vm_ioctl_register_coalesced_mmio(kvm, &zone);
+		if (r)
+			goto out;
+		r = 0;
+		break;
+	}
+	case KVM_UNREGISTER_COALESCED_MMIO: {
+		struct kvm_coalesced_mmio_zone zone;
+		r = -EFAULT;
+		if (copy_from_user(&zone, argp, sizeof zone))
+			goto out;
+		r = -ENXIO;
+		r = kvm_vm_ioctl_unregister_coalesced_mmio(kvm, &zone);
+		if (r)
+			goto out;
+		r = 0;
+		break;
+	}
+#endif
 	default:
 		r = kvm_arch_vm_ioctl(filp, ioctl, arg);
 	}
@@ -1231,6 +1285,9 @@
 		r = PAGE_SIZE;     /* struct kvm_run */
 #ifdef CONFIG_X86
 		r += PAGE_SIZE;    /* pio data page */
+#endif
+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
+		r += PAGE_SIZE;    /* coalesced mmio ring page */
 #endif
 		break;
 	case KVM_TRACE_ENABLE:
...	...	@@ -173,6 +173,30 @@
173	173	};
174	174	};
175	175
	176	+/* for KVM_REGISTER_COALESCED_MMIO / KVM_UNREGISTER_COALESCED_MMIO */
	177	+
	178	+struct kvm_coalesced_mmio_zone {
	179	+ __u64 addr;
	180	+ __u32 size;
	181	+ __u32 pad;
	182	+};
	183	+
	184	+struct kvm_coalesced_mmio {
	185	+ __u64 phys_addr;
	186	+ __u32 len;
	187	+ __u32 pad;
	188	+ __u8 data[8];
	189	+};
	190	+
	191	+struct kvm_coalesced_mmio_ring {
	192	+ __u32 first, last;
	193	+ struct kvm_coalesced_mmio coalesced_mmio[0];
	194	+};
	195	+
	196	+#define KVM_COALESCED_MMIO_MAX \
	197	+ ((PAGE_SIZE - sizeof(struct kvm_coalesced_mmio_ring)) / \
	198	+ sizeof(struct kvm_coalesced_mmio))
	199	+
176	200	/* for KVM_TRANSLATE */
177	201	struct kvm_translation {
178	202	/* in */
...	...	@@ -346,6 +370,7 @@
346	370	#define KVM_CAP_NOP_IO_DELAY 12
347	371	#define KVM_CAP_PV_MMU 13
348	372	#define KVM_CAP_MP_STATE 14
	373	+#define KVM_CAP_COALESCED_MMIO 15
349	374
350	375	/*
351	376	* ioctls for VM fds
...	...	@@ -371,6 +396,10 @@
371	396	#define KVM_CREATE_PIT _IO(KVMIO, 0x64)
372	397	#define KVM_GET_PIT _IOWR(KVMIO, 0x65, struct kvm_pit_state)
373	398	#define KVM_SET_PIT _IOR(KVMIO, 0x66, struct kvm_pit_state)
	399	+#define KVM_REGISTER_COALESCED_MMIO \
	400	+ _IOW(KVMIO, 0x67, struct kvm_coalesced_mmio_zone)
	401	+#define KVM_UNREGISTER_COALESCED_MMIO \
	402	+ _IOW(KVMIO, 0x68, struct kvm_coalesced_mmio_zone)
374	403
375	404	/*
376	405	* ioctls for vcpu fds
...	...	@@ -117,6 +117,10 @@
117	117	struct kvm_vm_stat stat;
118	118	struct kvm_arch arch;
119	119	atomic_t users_count;
	120	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	121	+ struct kvm_coalesced_mmio_dev *coalesced_mmio_dev;
	122	+ struct kvm_coalesced_mmio_ring *coalesced_mmio_ring;
	123	+#endif
120	124	};
121	125
122	126	/* The guest did something we don't support. */
	1	+/*
	2	+ * KVM coalesced MMIO
	3	+ *
	4	+ * Copyright (c) 2008 Bull S.A.S.
	5	+ *
	6	+ * Author: Laurent Vivier <Laurent.Vivier@bull.net>
	7	+ *
	8	+ */
	9	+
	10	+#include "iodev.h"
	11	+
	12	+#include <linux/kvm_host.h>
	13	+#include <linux/kvm.h>
	14	+
	15	+#include "coalesced_mmio.h"
	16	+
	17	+static int coalesced_mmio_in_range(struct kvm_io_device *this,
	18	+ gpa_t addr, int len, int is_write)
	19	+{
	20	+ struct kvm_coalesced_mmio_dev *dev =
	21	+ (struct kvm_coalesced_mmio_dev*)this->private;
	22	+ struct kvm_coalesced_mmio_zone *zone;
	23	+ int next;
	24	+ int i;
	25	+
	26	+ if (!is_write)
	27	+ return 0;
	28	+
	29	+ /* kvm->lock is taken by the caller and must be not released before
	30	+ * dev.read/write
	31	+ */
	32	+
	33	+ /* Are we able to batch it ? */
	34	+
	35	+ /* last is the first free entry
	36	+ * check if we don't meet the first used entry
	37	+ * there is always one unused entry in the buffer
	38	+ */
	39	+
	40	+ next = (dev->kvm->coalesced_mmio_ring->last + 1) %
	41	+ KVM_COALESCED_MMIO_MAX;
	42	+ if (next == dev->kvm->coalesced_mmio_ring->first) {
	43	+ /* full */
	44	+ return 0;
	45	+ }
	46	+
	47	+ /* is it in a batchable area ? */
	48	+
	49	+ for (i = 0; i < dev->nb_zones; i++) {
	50	+ zone = &dev->zone[i];
	51	+
	52	+ /* (addr,len) is fully included in
	53	+ * (zone->addr, zone->size)
	54	+ */
	55	+
	56	+ if (zone->addr <= addr &&
	57	+ addr + len <= zone->addr + zone->size)
	58	+ return 1;
	59	+ }
	60	+ return 0;
	61	+}
	62	+
	63	+static void coalesced_mmio_write(struct kvm_io_device *this,
	64	+ gpa_t addr, int len, const void *val)
	65	+{
	66	+ struct kvm_coalesced_mmio_dev *dev =
	67	+ (struct kvm_coalesced_mmio_dev*)this->private;
	68	+ struct kvm_coalesced_mmio_ring *ring = dev->kvm->coalesced_mmio_ring;
	69	+
	70	+ /* kvm->lock must be taken by caller before call to in_range()*/
	71	+
	72	+ /* copy data in first free entry of the ring */
	73	+
	74	+ ring->coalesced_mmio[ring->last].phys_addr = addr;
	75	+ ring->coalesced_mmio[ring->last].len = len;
	76	+ memcpy(ring->coalesced_mmio[ring->last].data, val, len);
	77	+ smp_wmb();
	78	+ ring->last = (ring->last + 1) % KVM_COALESCED_MMIO_MAX;
	79	+}
	80	+
	81	+static void coalesced_mmio_destructor(struct kvm_io_device *this)
	82	+{
	83	+ kfree(this);
	84	+}
	85	+
	86	+int kvm_coalesced_mmio_init(struct kvm *kvm)
	87	+{
	88	+ struct kvm_coalesced_mmio_dev *dev;
	89	+
	90	+ dev = kzalloc(sizeof(struct kvm_coalesced_mmio_dev), GFP_KERNEL);
	91	+ if (!dev)
	92	+ return -ENOMEM;
	93	+ dev->dev.write = coalesced_mmio_write;
	94	+ dev->dev.in_range = coalesced_mmio_in_range;
	95	+ dev->dev.destructor = coalesced_mmio_destructor;
	96	+ dev->dev.private = dev;
	97	+ dev->kvm = kvm;
	98	+ kvm->coalesced_mmio_dev = dev;
	99	+ kvm_io_bus_register_dev(&kvm->mmio_bus, &dev->dev);
	100	+
	101	+ return 0;
	102	+}
	103	+
	104	+int kvm_vm_ioctl_register_coalesced_mmio(struct kvm *kvm,
	105	+ struct kvm_coalesced_mmio_zone *zone)
	106	+{
	107	+ struct kvm_coalesced_mmio_dev *dev = kvm->coalesced_mmio_dev;
	108	+
	109	+ if (dev == NULL)
	110	+ return -EINVAL;
	111	+
	112	+ mutex_lock(&kvm->lock);
	113	+ if (dev->nb_zones >= KVM_COALESCED_MMIO_ZONE_MAX) {
	114	+ mutex_unlock(&kvm->lock);
	115	+ return -ENOBUFS;
	116	+ }
	117	+
	118	+ dev->zone[dev->nb_zones] = *zone;
	119	+ dev->nb_zones++;
	120	+
	121	+ mutex_unlock(&kvm->lock);
	122	+ return 0;
	123	+}
	124	+
	125	+int kvm_vm_ioctl_unregister_coalesced_mmio(struct kvm *kvm,
	126	+ struct kvm_coalesced_mmio_zone *zone)
	127	+{
	128	+ int i;
	129	+ struct kvm_coalesced_mmio_dev *dev = kvm->coalesced_mmio_dev;
	130	+ struct kvm_coalesced_mmio_zone *z;
	131	+
	132	+ if (dev == NULL)
	133	+ return -EINVAL;
	134	+
	135	+ mutex_lock(&kvm->lock);
	136	+
	137	+ i = dev->nb_zones;
	138	+ while(i) {
	139	+ z = &dev->zone[i - 1];
	140	+
	141	+ /* unregister all zones
	142	+ * included in (zone->addr, zone->size)
	143	+ */
	144	+
	145	+ if (zone->addr <= z->addr &&
	146	+ z->addr + z->size <= zone->addr + zone->size) {
	147	+ dev->nb_zones--;
	148	+ *z = dev->zone[dev->nb_zones];
	149	+ }
	150	+ i--;
	151	+ }
	152	+
	153	+ mutex_unlock(&kvm->lock);
	154	+
	155	+ return 0;
	156	+}
...	...	@@ -47,6 +47,10 @@
47	47	#include <asm/uaccess.h>
48	48	#include <asm/pgtable.h>
49	49
	50	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	51	+#include "coalesced_mmio.h"
	52	+#endif
	53	+
50	54	MODULE_AUTHOR("Qumranet");
51	55	MODULE_LICENSE("GPL");
52	56
53	57
...	...	@@ -185,10 +189,23 @@
185	189	static struct kvm *kvm_create_vm(void)
186	190	{
187	191	struct kvm *kvm = kvm_arch_create_vm();
	192	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	193	+ struct page *page;
	194	+#endif
188	195
189	196	if (IS_ERR(kvm))
190	197	goto out;
191	198
	199	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	200	+ page = alloc_page(GFP_KERNEL \| __GFP_ZERO);
	201	+ if (!page) {
	202	+ kfree(kvm);
	203	+ return ERR_PTR(-ENOMEM);
	204	+ }
	205	+ kvm->coalesced_mmio_ring =
	206	+ (struct kvm_coalesced_mmio_ring *)page_address(page);
	207	+#endif
	208	+
192	209	kvm->mm = current->mm;
193	210	atomic_inc(&kvm->mm->mm_count);
194	211	spin_lock_init(&kvm->mmu_lock);
...	...	@@ -200,6 +217,9 @@
200	217	spin_lock(&kvm_lock);
201	218	list_add(&kvm->vm_list, &vm_list);
202	219	spin_unlock(&kvm_lock);
	220	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	221	+ kvm_coalesced_mmio_init(kvm);
	222	+#endif
203	223	out:
204	224	return kvm;
205	225	}
...	...	@@ -242,6 +262,10 @@
242	262	spin_unlock(&kvm_lock);
243	263	kvm_io_bus_destroy(&kvm->pio_bus);
244	264	kvm_io_bus_destroy(&kvm->mmio_bus);
	265	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	266	+ if (kvm->coalesced_mmio_ring != NULL)
	267	+ free_page((unsigned long)kvm->coalesced_mmio_ring);
	268	+#endif
245	269	kvm_arch_destroy_vm(kvm);
246	270	mmdrop(mm);
247	271	}
...	...	@@ -826,6 +850,10 @@
826	850	else if (vmf->pgoff == KVM_PIO_PAGE_OFFSET)
827	851	page = virt_to_page(vcpu->arch.pio_data);
828	852	#endif
	853	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	854	+ else if (vmf->pgoff == KVM_COALESCED_MMIO_PAGE_OFFSET)
	855	+ page = virt_to_page(vcpu->kvm->coalesced_mmio_ring);
	856	+#endif
829	857	else
830	858	return VM_FAULT_SIGBUS;
831	859	get_page(page);
...	...	@@ -1148,6 +1176,32 @@
1148	1176	goto out;
1149	1177	break;
1150	1178	}
	1179	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	1180	+ case KVM_REGISTER_COALESCED_MMIO: {
	1181	+ struct kvm_coalesced_mmio_zone zone;
	1182	+ r = -EFAULT;
	1183	+ if (copy_from_user(&zone, argp, sizeof zone))
	1184	+ goto out;
	1185	+ r = -ENXIO;
	1186	+ r = kvm_vm_ioctl_register_coalesced_mmio(kvm, &zone);
	1187	+ if (r)
	1188	+ goto out;
	1189	+ r = 0;
	1190	+ break;
	1191	+ }
	1192	+ case KVM_UNREGISTER_COALESCED_MMIO: {
	1193	+ struct kvm_coalesced_mmio_zone zone;
	1194	+ r = -EFAULT;
	1195	+ if (copy_from_user(&zone, argp, sizeof zone))
	1196	+ goto out;
	1197	+ r = -ENXIO;
	1198	+ r = kvm_vm_ioctl_unregister_coalesced_mmio(kvm, &zone);
	1199	+ if (r)
	1200	+ goto out;
	1201	+ r = 0;
	1202	+ break;
	1203	+ }
	1204	+#endif
1151	1205	default:
1152	1206	r = kvm_arch_vm_ioctl(filp, ioctl, arg);
1153	1207	}
...	...	@@ -1231,6 +1285,9 @@
1231	1285	r = PAGE_SIZE; /* struct kvm_run */
1232	1286	#ifdef CONFIG_X86
1233	1287	r += PAGE_SIZE; /* pio data page */
	1288	+#endif
	1289	+#ifdef KVM_COALESCED_MMIO_PAGE_OFFSET
	1290	+ r += PAGE_SIZE; /* coalesced mmio ring page */
1234	1291	#endif
1235	1292	break;
1236	1293	case KVM_TRACE_ENABLE: