[net-next-2.6.git] / drivers / virtio / virtio_ring.c

/* Virtio ring implementation.
 *
 *  Copyright 2007 Rusty Russell IBM Corporation
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License as published by
 *  the Free Software Foundation; either version 2 of the License, or
 *  (at your option) any later version.
 *
 *  This program is distributed in the hope that it will be useful,
 *  but WITHOUT ANY WARRANTY; without even the implied warranty of
 *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *  GNU General Public License for more details.
 *
 *  You should have received a copy of the GNU General Public License
 *  along with this program; if not, write to the Free Software
 *  Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 */
#include <linux/virtio.h>
#include <linux/virtio_ring.h>
#include <linux/virtio_config.h>
#include <linux/device.h>

#ifdef DEBUG
/* For development, we want to crash whenever the ring is screwed. */
#define BAD_RING(_vq, fmt, args...)				\
	do {							\
		dev_err(&(_vq)->vq.vdev->dev,			\
			"%s:"fmt, (_vq)->vq.name, ##args);	\
		BUG();						\
	} while (0)
/* Caller is supposed to guarantee no reentry. */
#define START_USE(_vq)						\
	do {							\
		if ((_vq)->in_use)				\
			panic("%s:in_use = %i\n",		\
			      (_vq)->vq.name, (_vq)->in_use);	\
		(_vq)->in_use = __LINE__;			\
		mb();						\
	} while (0)
#define END_USE(_vq) \
	do { BUG_ON(!(_vq)->in_use); (_vq)->in_use = 0; mb(); } while(0)
#else
#define BAD_RING(_vq, fmt, args...)				\
	do {							\
		dev_err(&_vq->vq.vdev->dev,			\
			"%s:"fmt, (_vq)->vq.name, ##args);	\
		(_vq)->broken = true;				\
	} while (0)
#define START_USE(vq)
#define END_USE(vq)
#endif

struct vring_virtqueue
{
	struct virtqueue vq;

	/* Actual memory layout for this queue */
	struct vring vring;

	/* Other side has made a mess, don't try any more. */
	bool broken;

	/* Host supports indirect buffers */
	bool indirect;

	/* Number of free buffers */
	unsigned int num_free;
	/* Head of free buffer list. */
	unsigned int free_head;
	/* Number we've added since last sync. */
	unsigned int num_added;

	/* Last used index we've seen. */
	u16 last_used_idx;

	/* How to notify other side. FIXME: commonalize hcalls! */
	void (*notify)(struct virtqueue *vq);

#ifdef DEBUG
	/* They're supposed to lock for us. */
	unsigned int in_use;
#endif

	/* Tokens for callbacks. */
	void *data[];
};

#define to_vvq(_vq) container_of(_vq, struct vring_virtqueue, vq)

/* Set up an indirect table of descriptors and add it to the queue. */
static int vring_add_indirect(struct vring_virtqueue *vq,
			      struct scatterlist sg[],
			      unsigned int out,
			      unsigned int in)
{
	struct vring_desc *desc;
	unsigned head;
	int i;

	desc = kmalloc((out + in) * sizeof(struct vring_desc), GFP_ATOMIC);
	if (!desc)
		return vq->vring.num;

	/* Transfer entries from the sg list into the indirect page */
	for (i = 0; i < out; i++) {
		desc[i].flags = VRING_DESC_F_NEXT;
		desc[i].addr = sg_phys(sg);
		desc[i].len = sg->length;
		desc[i].next = i+1;
		sg++;
	}
	for (; i < (out + in); i++) {
		desc[i].flags = VRING_DESC_F_NEXT|VRING_DESC_F_WRITE;
		desc[i].addr = sg_phys(sg);
		desc[i].len = sg->length;
		desc[i].next = i+1;
		sg++;
	}

	/* Last one doesn't continue. */
	desc[i-1].flags &= ~VRING_DESC_F_NEXT;
	desc[i-1].next = 0;

	/* We're about to use a buffer */
	vq->num_free--;

	/* Use a single buffer which doesn't continue */
	head = vq->free_head;
	vq->vring.desc[head].flags = VRING_DESC_F_INDIRECT;
	vq->vring.desc[head].addr = virt_to_phys(desc);
	vq->vring.desc[head].len = i * sizeof(struct vring_desc);

	/* Update free pointer */
	vq->free_head = vq->vring.desc[head].next;

	return head;
}

static int vring_add_buf(struct virtqueue *_vq,
			 struct scatterlist sg[],
			 unsigned int out,
			 unsigned int in,
			 void *data)
{
	struct vring_virtqueue *vq = to_vvq(_vq);
	unsigned int i, avail, head, uninitialized_var(prev);

	START_USE(vq);

	BUG_ON(data == NULL);

	/* If the host supports indirect descriptor tables, and we have multiple
	 * buffers, then go indirect. FIXME: tune this threshold */
	if (vq->indirect && (out + in) > 1 && vq->num_free) {
		head = vring_add_indirect(vq, sg, out, in);
		if (head != vq->vring.num)
			goto add_head;
	}

	BUG_ON(out + in > vq->vring.num);
	BUG_ON(out + in == 0);

	if (vq->num_free < out + in) {
		pr_debug("Can't add buf len %i - avail = %i\n",
			 out + in, vq->num_free);
		/* FIXME: for historical reasons, we force a notify here if
		 * there are outgoing parts to the buffer.  Presumably the
		 * host should service the ring ASAP. */
		if (out)
			vq->notify(&vq->vq);
		END_USE(vq);
		return -ENOSPC;
	}

	/* We're about to use some buffers from the free list. */
	vq->num_free -= out + in;

	head = vq->free_head;
	for (i = vq->free_head; out; i = vq->vring.desc[i].next, out--) {
		vq->vring.desc[i].flags = VRING_DESC_F_NEXT;
		vq->vring.desc[i].addr = sg_phys(sg);
		vq->vring.desc[i].len = sg->length;
		prev = i;
		sg++;
	}
	for (; in; i = vq->vring.desc[i].next, in--) {
		vq->vring.desc[i].flags = VRING_DESC_F_NEXT|VRING_DESC_F_WRITE;
		vq->vring.desc[i].addr = sg_phys(sg);
		vq->vring.desc[i].len = sg->length;
		prev = i;
		sg++;
	}
	/* Last one doesn't continue. */
	vq->vring.desc[prev].flags &= ~VRING_DESC_F_NEXT;

	/* Update free pointer */
	vq->free_head = i;

add_head:
	/* Set token. */
	vq->data[head] = data;

	/* Put entry in available array (but don't update avail->idx until they
	 * do sync).  FIXME: avoid modulus here? */
	avail = (vq->vring.avail->idx + vq->num_added++) % vq->vring.num;
	vq->vring.avail->ring[avail] = head;

	pr_debug("Added buffer head %i to %p\n", head, vq);
	END_USE(vq);

	/* If we're indirect, we can fit many (assuming not OOM). */
	if (vq->indirect)
		return vq->num_free ? vq->vring.num : 0;
	return vq->num_free;
}

static void vring_kick(struct virtqueue *_vq)
{
	struct vring_virtqueue *vq = to_vvq(_vq);
	START_USE(vq);
	/* Descriptors and available array need to be set before we expose the
	 * new available array entries. */
	wmb();

	vq->vring.avail->idx += vq->num_added;
	vq->num_added = 0;

	/* Need to update avail index before checking if we should notify */
	mb();

	if (!(vq->vring.used->flags & VRING_USED_F_NO_NOTIFY))
		/* Prod other side to tell it about changes. */
		vq->notify(&vq->vq);

	END_USE(vq);
}

static void detach_buf(struct vring_virtqueue *vq, unsigned int head)
{
	unsigned int i;

	/* Clear data ptr. */
	vq->data[head] = NULL;

	/* Put back on free list: find end */
	i = head;

	/* Free the indirect table */
	if (vq->vring.desc[i].flags & VRING_DESC_F_INDIRECT)
		kfree(phys_to_virt(vq->vring.desc[i].addr));

	while (vq->vring.desc[i].flags & VRING_DESC_F_NEXT) {
		i = vq->vring.desc[i].next;
		vq->num_free++;
	}

	vq->vring.desc[i].next = vq->free_head;
	vq->free_head = head;
	/* Plus final descriptor */
	vq->num_free++;
}

static inline bool more_used(const struct vring_virtqueue *vq)
{
	return vq->last_used_idx != vq->vring.used->idx;
}

static void *vring_get_buf(struct virtqueue *_vq, unsigned int *len)
{
	struct vring_virtqueue *vq = to_vvq(_vq);
	void *ret;
	unsigned int i;

	START_USE(vq);

	if (unlikely(vq->broken)) {
		END_USE(vq);
		return NULL;
	}

	if (!more_used(vq)) {
		pr_debug("No more buffers in queue\n");
		END_USE(vq);
		return NULL;
	}

	/* Only get used array entries after they have been exposed by host. */
	rmb();

	i = vq->vring.used->ring[vq->last_used_idx%vq->vring.num].id;
	*len = vq->vring.used->ring[vq->last_used_idx%vq->vring.num].len;

	if (unlikely(i >= vq->vring.num)) {
		BAD_RING(vq, "id %u out of range\n", i);
		return NULL;
	}
	if (unlikely(!vq->data[i])) {
		BAD_RING(vq, "id %u is not a head!\n", i);
		return NULL;
	}

	/* detach_buf clears data, so grab it now. */
	ret = vq->data[i];
	detach_buf(vq, i);
	vq->last_used_idx++;
	END_USE(vq);
	return ret;
}

static void vring_disable_cb(struct virtqueue *_vq)
{
	struct vring_virtqueue *vq = to_vvq(_vq);

	vq->vring.avail->flags |= VRING_AVAIL_F_NO_INTERRUPT;
}

static bool vring_enable_cb(struct virtqueue *_vq)
{
	struct vring_virtqueue *vq = to_vvq(_vq);

	START_USE(vq);

	/* We optimistically turn back on interrupts, then check if there was
	 * more to do. */
	vq->vring.avail->flags &= ~VRING_AVAIL_F_NO_INTERRUPT;
	mb();
	if (unlikely(more_used(vq))) {
		END_USE(vq);
		return false;
	}

	END_USE(vq);
	return true;
}

irqreturn_t vring_interrupt(int irq, void *_vq)
{
	struct vring_virtqueue *vq = to_vvq(_vq);

	if (!more_used(vq)) {
		pr_debug("virtqueue interrupt with no work for %p\n", vq);
		return IRQ_NONE;
	}

	if (unlikely(vq->broken))
		return IRQ_HANDLED;

	pr_debug("virtqueue callback for %p (%p)\n", vq, vq->vq.callback);
	if (vq->vq.callback)
		vq->vq.callback(&vq->vq);

	return IRQ_HANDLED;
}
EXPORT_SYMBOL_GPL(vring_interrupt);

static struct virtqueue_ops vring_vq_ops = {
	.add_buf = vring_add_buf,
	.get_buf = vring_get_buf,
	.kick = vring_kick,
	.disable_cb = vring_disable_cb,
	.enable_cb = vring_enable_cb,
};

struct virtqueue *vring_new_virtqueue(unsigned int num,
				      unsigned int vring_align,
				      struct virtio_device *vdev,
				      void *pages,
				      void (*notify)(struct virtqueue *),
				      void (*callback)(struct virtqueue *),
				      const char *name)
{
	struct vring_virtqueue *vq;
	unsigned int i;

	/* We assume num is a power of 2. */
	if (num & (num - 1)) {
		dev_warn(&vdev->dev, "Bad virtqueue length %u\n", num);
		return NULL;
	}

	vq = kmalloc(sizeof(*vq) + sizeof(void *)*num, GFP_KERNEL);
	if (!vq)
		return NULL;

	vring_init(&vq->vring, num, pages, vring_align);
	vq->vq.callback = callback;
	vq->vq.vdev = vdev;
	vq->vq.vq_ops = &vring_vq_ops;
	vq->vq.name = name;
	vq->notify = notify;
	vq->broken = false;
	vq->last_used_idx = 0;
	vq->num_added = 0;
	list_add_tail(&vq->vq.list, &vdev->vqs);
#ifdef DEBUG
	vq->in_use = false;
#endif

	vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC);

	/* No callback?  Tell other side not to bother us. */
	if (!callback)
		vq->vring.avail->flags |= VRING_AVAIL_F_NO_INTERRUPT;

	/* Put everything in free lists. */
	vq->num_free = num;
	vq->free_head = 0;
	for (i = 0; i < num-1; i++)
		vq->vring.desc[i].next = i+1;

	return &vq->vq;
}
EXPORT_SYMBOL_GPL(vring_new_virtqueue);

void vring_del_virtqueue(struct virtqueue *vq)
{
	list_del(&vq->list);
	kfree(to_vvq(vq));
}
EXPORT_SYMBOL_GPL(vring_del_virtqueue);

/* Manipulates transport-specific feature bits. */
void vring_transport_features(struct virtio_device *vdev)
{
	unsigned int i;

	for (i = VIRTIO_TRANSPORT_F_START; i < VIRTIO_TRANSPORT_F_END; i++) {
		switch (i) {
		case VIRTIO_RING_F_INDIRECT_DESC:
			break;
		default:
			/* We don't understand this bit. */
			clear_bit(i, vdev->features);
		}
	}
}
EXPORT_SYMBOL_GPL(vring_transport_features);

MODULE_LICENSE("GPL");
Commit	Line	Data
0a8a69dd RR	1	/* Virtio ring implementation.
	2	*
	3	* Copyright 2007 Rusty Russell IBM Corporation
	4	*
	5	* This program is free software; you can redistribute it and/or modify
	6	* it under the terms of the GNU General Public License as published by
	7	* the Free Software Foundation; either version 2 of the License, or
	8	* (at your option) any later version.
	9	*
	10	* This program is distributed in the hope that it will be useful,
	11	* but WITHOUT ANY WARRANTY; without even the implied warranty of
	12	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
	13	* GNU General Public License for more details.
	14	*
	15	* You should have received a copy of the GNU General Public License
	16	* along with this program; if not, write to the Free Software
	17	* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
	18	*/
	19	#include <linux/virtio.h>
	20	#include <linux/virtio_ring.h>
e34f8725	21	#include <linux/virtio_config.h>
0a8a69dd RR	22	#include <linux/device.h>
	23
	24	#ifdef DEBUG
	25	/* For development, we want to crash whenever the ring is screwed. */
9499f5e7 RR	26	#define BAD_RING(_vq, fmt, args...) \
	27	do { \
	28	dev_err(&(_vq)->vq.vdev->dev, \
	29	"%s:"fmt, (_vq)->vq.name, ##args); \
	30	BUG(); \
	31	} while (0)
c5f841f1 RR	32	/* Caller is supposed to guarantee no reentry. */
	33	#define START_USE(_vq) \
	34	do { \
	35	if ((_vq)->in_use) \
9499f5e7 RR	36	panic("%s:in_use = %i\n", \
9499f5e7 RR	37	(_vq)->vq.name, (_vq)->in_use); \
c5f841f1 RR	38	(_vq)->in_use = __LINE__; \
c5f841f1 RR	39	mb(); \
9499f5e7	40	} while (0)
3a35ce7d RK	41	#define END_USE(_vq) \
3a35ce7d RK	42	do { BUG_ON(!(_vq)->in_use); (_vq)->in_use = 0; mb(); } while(0)
0a8a69dd	43	#else
9499f5e7 RR	44	#define BAD_RING(_vq, fmt, args...) \
	45	do { \
	46	dev_err(&_vq->vq.vdev->dev, \
	47	"%s:"fmt, (_vq)->vq.name, ##args); \
	48	(_vq)->broken = true; \
	49	} while (0)
0a8a69dd RR	50	#define START_USE(vq)
	51	#define END_USE(vq)
	52	#endif
	53
	54	struct vring_virtqueue
	55	{
	56	struct virtqueue vq;
	57
	58	/* Actual memory layout for this queue */
	59	struct vring vring;
	60
	61	/* Other side has made a mess, don't try any more. */
	62	bool broken;
	63
9fa29b9d MM	64	/* Host supports indirect buffers */
	65	bool indirect;
	66
0a8a69dd RR	67	/* Number of free buffers */
	68	unsigned int num_free;
	69	/* Head of free buffer list. */
	70	unsigned int free_head;
	71	/* Number we've added since last sync. */
	72	unsigned int num_added;
	73
	74	/* Last used index we've seen. */
1bc4953e	75	u16 last_used_idx;
0a8a69dd RR	76
	77	/* How to notify other side. FIXME: commonalize hcalls! */
	78	void (notify)(struct virtqueue vq);
	79
	80	#ifdef DEBUG
	81	/* They're supposed to lock for us. */
	82	unsigned int in_use;
	83	#endif
	84
	85	/* Tokens for callbacks. */
	86	void *data[];
	87	};
	88
	89	#define to_vvq(_vq) container_of(_vq, struct vring_virtqueue, vq)
	90
9fa29b9d MM	91	/* Set up an indirect table of descriptors and add it to the queue. */
	92	static int vring_add_indirect(struct vring_virtqueue *vq,
	93	struct scatterlist sg[],
	94	unsigned int out,
	95	unsigned int in)
	96	{
	97	struct vring_desc *desc;
	98	unsigned head;
	99	int i;
	100
	101	desc = kmalloc((out + in) * sizeof(struct vring_desc), GFP_ATOMIC);
	102	if (!desc)
	103	return vq->vring.num;
	104
	105	/* Transfer entries from the sg list into the indirect page */
	106	for (i = 0; i < out; i++) {
	107	desc[i].flags = VRING_DESC_F_NEXT;
	108	desc[i].addr = sg_phys(sg);
	109	desc[i].len = sg->length;
	110	desc[i].next = i+1;
	111	sg++;
	112	}
	113	for (; i < (out + in); i++) {
	114	desc[i].flags = VRING_DESC_F_NEXT\|VRING_DESC_F_WRITE;
	115	desc[i].addr = sg_phys(sg);
	116	desc[i].len = sg->length;
	117	desc[i].next = i+1;
	118	sg++;
	119	}
	120
	121	/* Last one doesn't continue. */
	122	desc[i-1].flags &= ~VRING_DESC_F_NEXT;
	123	desc[i-1].next = 0;
	124
	125	/* We're about to use a buffer */
	126	vq->num_free--;
	127
	128	/* Use a single buffer which doesn't continue */
	129	head = vq->free_head;
	130	vq->vring.desc[head].flags = VRING_DESC_F_INDIRECT;
	131	vq->vring.desc[head].addr = virt_to_phys(desc);
	132	vq->vring.desc[head].len = i * sizeof(struct vring_desc);
	133
	134	/* Update free pointer */
	135	vq->free_head = vq->vring.desc[head].next;
	136
	137	return head;
	138	}
	139
0a8a69dd RR	140	static int vring_add_buf(struct virtqueue *_vq,
	141	struct scatterlist sg[],
	142	unsigned int out,
	143	unsigned int in,
	144	void *data)
	145	{
	146	struct vring_virtqueue *vq = to_vvq(_vq);
	147	unsigned int i, avail, head, uninitialized_var(prev);
	148
9fa29b9d MM	149	START_USE(vq);
9fa29b9d MM	150
0a8a69dd	151	BUG_ON(data == NULL);
9fa29b9d MM	152
	153	/* If the host supports indirect descriptor tables, and we have multiple
	154	* buffers, then go indirect. FIXME: tune this threshold */
	155	if (vq->indirect && (out + in) > 1 && vq->num_free) {
	156	head = vring_add_indirect(vq, sg, out, in);
	157	if (head != vq->vring.num)
	158	goto add_head;
	159	}
	160
0a8a69dd RR	161	BUG_ON(out + in > vq->vring.num);
	162	BUG_ON(out + in == 0);
	163
0a8a69dd RR	164	if (vq->num_free < out + in) {
	165	pr_debug("Can't add buf len %i - avail = %i\n",
	166	out + in, vq->num_free);
44653eae RR	167	/* FIXME: for historical reasons, we force a notify here if
	168	* there are outgoing parts to the buffer. Presumably the
	169	* host should service the ring ASAP. */
	170	if (out)
	171	vq->notify(&vq->vq);
0a8a69dd RR	172	END_USE(vq);
	173	return -ENOSPC;
	174	}
	175
	176	/* We're about to use some buffers from the free list. */
	177	vq->num_free -= out + in;
	178
	179	head = vq->free_head;
	180	for (i = vq->free_head; out; i = vq->vring.desc[i].next, out--) {
	181	vq->vring.desc[i].flags = VRING_DESC_F_NEXT;
15f9c890	182	vq->vring.desc[i].addr = sg_phys(sg);
0a8a69dd RR	183	vq->vring.desc[i].len = sg->length;
	184	prev = i;
	185	sg++;
	186	}
	187	for (; in; i = vq->vring.desc[i].next, in--) {
	188	vq->vring.desc[i].flags = VRING_DESC_F_NEXT\|VRING_DESC_F_WRITE;
15f9c890	189	vq->vring.desc[i].addr = sg_phys(sg);
0a8a69dd RR	190	vq->vring.desc[i].len = sg->length;
	191	prev = i;
	192	sg++;
	193	}
	194	/* Last one doesn't continue. */
	195	vq->vring.desc[prev].flags &= ~VRING_DESC_F_NEXT;
	196
	197	/* Update free pointer */
	198	vq->free_head = i;
	199
9fa29b9d	200	add_head:
0a8a69dd RR	201	/* Set token. */
	202	vq->data[head] = data;
	203
	204	/* Put entry in available array (but don't update avail->idx until they
	205	* do sync). FIXME: avoid modulus here? */
	206	avail = (vq->vring.avail->idx + vq->num_added++) % vq->vring.num;
	207	vq->vring.avail->ring[avail] = head;
	208
	209	pr_debug("Added buffer head %i to %p\n", head, vq);
	210	END_USE(vq);
3c1b27d5 RR	211
	212	/* If we're indirect, we can fit many (assuming not OOM). */
	213	if (vq->indirect)
	214	return vq->num_free ? vq->vring.num : 0;
	215	return vq->num_free;
0a8a69dd RR	216	}
	217
	218	static void vring_kick(struct virtqueue *_vq)
	219	{
	220	struct vring_virtqueue *vq = to_vvq(_vq);
	221	START_USE(vq);
	222	/* Descriptors and available array need to be set before we expose the
	223	* new available array entries. */
	224	wmb();
	225
	226	vq->vring.avail->idx += vq->num_added;
	227	vq->num_added = 0;
	228
	229	/* Need to update avail index before checking if we should notify */
	230	mb();
	231
	232	if (!(vq->vring.used->flags & VRING_USED_F_NO_NOTIFY))
	233	/* Prod other side to tell it about changes. */
	234	vq->notify(&vq->vq);
	235
	236	END_USE(vq);
	237	}
	238
	239	static void detach_buf(struct vring_virtqueue *vq, unsigned int head)
	240	{
	241	unsigned int i;
	242
	243	/* Clear data ptr. */
	244	vq->data[head] = NULL;
	245
	246	/* Put back on free list: find end */
	247	i = head;
9fa29b9d MM	248
	249	/* Free the indirect table */
	250	if (vq->vring.desc[i].flags & VRING_DESC_F_INDIRECT)
	251	kfree(phys_to_virt(vq->vring.desc[i].addr));
	252
0a8a69dd RR	253	while (vq->vring.desc[i].flags & VRING_DESC_F_NEXT) {
	254	i = vq->vring.desc[i].next;
	255	vq->num_free++;
	256	}
	257
	258	vq->vring.desc[i].next = vq->free_head;
	259	vq->free_head = head;
	260	/* Plus final descriptor */
	261	vq->num_free++;
	262	}
	263
0a8a69dd RR	264	static inline bool more_used(const struct vring_virtqueue *vq)
	265	{
	266	return vq->last_used_idx != vq->vring.used->idx;
	267	}
	268
	269	static void vring_get_buf(struct virtqueue _vq, unsigned int *len)
	270	{
	271	struct vring_virtqueue *vq = to_vvq(_vq);
	272	void *ret;
	273	unsigned int i;
	274
	275	START_USE(vq);
	276
5ef82752 RR	277	if (unlikely(vq->broken)) {
	278	END_USE(vq);
	279	return NULL;
	280	}
	281
0a8a69dd RR	282	if (!more_used(vq)) {
	283	pr_debug("No more buffers in queue\n");
	284	END_USE(vq);
	285	return NULL;
	286	}
	287
2d61ba95 MT	288	/* Only get used array entries after they have been exposed by host. */
	289	rmb();
	290
0a8a69dd RR	291	i = vq->vring.used->ring[vq->last_used_idx%vq->vring.num].id;
	292	*len = vq->vring.used->ring[vq->last_used_idx%vq->vring.num].len;
	293
	294	if (unlikely(i >= vq->vring.num)) {
	295	BAD_RING(vq, "id %u out of range\n", i);
	296	return NULL;
	297	}
	298	if (unlikely(!vq->data[i])) {
	299	BAD_RING(vq, "id %u is not a head!\n", i);
	300	return NULL;
	301	}
	302
	303	/* detach_buf clears data, so grab it now. */
	304	ret = vq->data[i];
	305	detach_buf(vq, i);
	306	vq->last_used_idx++;
	307	END_USE(vq);
	308	return ret;
	309	}
	310
18445c4d RR	311	static void vring_disable_cb(struct virtqueue *_vq)
	312	{
	313	struct vring_virtqueue *vq = to_vvq(_vq);
	314
18445c4d	315	vq->vring.avail->flags \|= VRING_AVAIL_F_NO_INTERRUPT;
18445c4d RR	316	}
	317
	318	static bool vring_enable_cb(struct virtqueue *_vq)
0a8a69dd RR	319	{
	320	struct vring_virtqueue *vq = to_vvq(_vq);
	321
	322	START_USE(vq);
0a8a69dd RR	323
	324	/* We optimistically turn back on interrupts, then check if there was
	325	* more to do. */
	326	vq->vring.avail->flags &= ~VRING_AVAIL_F_NO_INTERRUPT;
	327	mb();
	328	if (unlikely(more_used(vq))) {
0a8a69dd RR	329	END_USE(vq);
	330	return false;
	331	}
	332
	333	END_USE(vq);
	334	return true;
	335	}
	336
	337	irqreturn_t vring_interrupt(int irq, void *_vq)
	338	{
	339	struct vring_virtqueue *vq = to_vvq(_vq);
	340
	341	if (!more_used(vq)) {
	342	pr_debug("virtqueue interrupt with no work for %p\n", vq);
	343	return IRQ_NONE;
	344	}
	345
	346	if (unlikely(vq->broken))
	347	return IRQ_HANDLED;
	348
	349	pr_debug("virtqueue callback for %p (%p)\n", vq, vq->vq.callback);
18445c4d RR	350	if (vq->vq.callback)
18445c4d RR	351	vq->vq.callback(&vq->vq);
0a8a69dd RR	352
	353	return IRQ_HANDLED;
	354	}
c6fd4701	355	EXPORT_SYMBOL_GPL(vring_interrupt);
0a8a69dd RR	356
	357	static struct virtqueue_ops vring_vq_ops = {
	358	.add_buf = vring_add_buf,
	359	.get_buf = vring_get_buf,
	360	.kick = vring_kick,
18445c4d RR	361	.disable_cb = vring_disable_cb,
18445c4d RR	362	.enable_cb = vring_enable_cb,
0a8a69dd RR	363	};
	364
	365	struct virtqueue *vring_new_virtqueue(unsigned int num,
87c7d57c	366	unsigned int vring_align,
0a8a69dd RR	367	struct virtio_device *vdev,
	368	void *pages,
	369	void (notify)(struct virtqueue ),
9499f5e7 RR	370	void (callback)(struct virtqueue ),
9499f5e7 RR	371	const char *name)
0a8a69dd RR	372	{
	373	struct vring_virtqueue *vq;
	374	unsigned int i;
	375
42b36cc0 RR	376	/* We assume num is a power of 2. */
	377	if (num & (num - 1)) {
	378	dev_warn(&vdev->dev, "Bad virtqueue length %u\n", num);
	379	return NULL;
	380	}
	381
0a8a69dd RR	382	vq = kmalloc(sizeof(vq) + sizeof(void )*num, GFP_KERNEL);
	383	if (!vq)
	384	return NULL;
	385
87c7d57c	386	vring_init(&vq->vring, num, pages, vring_align);
0a8a69dd RR	387	vq->vq.callback = callback;
	388	vq->vq.vdev = vdev;
	389	vq->vq.vq_ops = &vring_vq_ops;
9499f5e7	390	vq->vq.name = name;
0a8a69dd RR	391	vq->notify = notify;
	392	vq->broken = false;
	393	vq->last_used_idx = 0;
	394	vq->num_added = 0;
9499f5e7	395	list_add_tail(&vq->vq.list, &vdev->vqs);
0a8a69dd RR	396	#ifdef DEBUG
	397	vq->in_use = false;
	398	#endif
	399
9fa29b9d MM	400	vq->indirect = virtio_has_feature(vdev, VIRTIO_RING_F_INDIRECT_DESC);
9fa29b9d MM	401
0a8a69dd RR	402	/* No callback? Tell other side not to bother us. */
	403	if (!callback)
	404	vq->vring.avail->flags \|= VRING_AVAIL_F_NO_INTERRUPT;
	405
	406	/* Put everything in free lists. */
	407	vq->num_free = num;
	408	vq->free_head = 0;
	409	for (i = 0; i < num-1; i++)
	410	vq->vring.desc[i].next = i+1;
	411
	412	return &vq->vq;
	413	}
c6fd4701	414	EXPORT_SYMBOL_GPL(vring_new_virtqueue);
0a8a69dd RR	415
	416	void vring_del_virtqueue(struct virtqueue *vq)
	417	{
9499f5e7	418	list_del(&vq->list);
0a8a69dd RR	419	kfree(to_vvq(vq));
0a8a69dd RR	420	}
c6fd4701	421	EXPORT_SYMBOL_GPL(vring_del_virtqueue);
0a8a69dd	422
e34f8725 RR	423	/* Manipulates transport-specific feature bits. */
	424	void vring_transport_features(struct virtio_device *vdev)
	425	{
	426	unsigned int i;
	427
	428	for (i = VIRTIO_TRANSPORT_F_START; i < VIRTIO_TRANSPORT_F_END; i++) {
	429	switch (i) {
9fa29b9d MM	430	case VIRTIO_RING_F_INDIRECT_DESC:
9fa29b9d MM	431	break;
e34f8725 RR	432	default:
	433	/* We don't understand this bit. */
	434	clear_bit(i, vdev->features);
	435	}
	436	}
	437	}
	438	EXPORT_SYMBOL_GPL(vring_transport_features);
	439
c6fd4701	440	MODULE_LICENSE("GPL");