[mirror_ubuntu-bionic-kernel.git] / drivers / block / virtio_blk.c

//#define DEBUG
#include <linux/spinlock.h>
#include <linux/blkdev.h>
#include <linux/hdreg.h>
#include <linux/virtio.h>
#include <linux/virtio_blk.h>
#include <linux/scatterlist.h>

#define VIRTIO_MAX_SG	(3+MAX_PHYS_SEGMENTS)
#define PART_BITS 4

static unsigned char virtblk_index = 'a';
static int major, minor;

struct virtio_blk
{
	spinlock_t lock;

	struct virtio_device *vdev;
	struct virtqueue *vq;

	/* The disk structure for the kernel. */
	struct gendisk *disk;

	/* Request tracking. */
	struct list_head reqs;

	mempool_t *pool;

	/* Scatterlist: can be too big for stack. */
	struct scatterlist sg[VIRTIO_MAX_SG];
};

struct virtblk_req
{
	struct list_head list;
	struct request *req;
	struct virtio_blk_outhdr out_hdr;
	struct virtio_blk_inhdr in_hdr;
};

static void blk_done(struct virtqueue *vq)
{
	struct virtio_blk *vblk = vq->vdev->priv;
	struct virtblk_req *vbr;
	unsigned int len;
	unsigned long flags;

	spin_lock_irqsave(&vblk->lock, flags);
	while ((vbr = vblk->vq->vq_ops->get_buf(vblk->vq, &len)) != NULL) {
		int uptodate;
		switch (vbr->in_hdr.status) {
		case VIRTIO_BLK_S_OK:
			uptodate = 1;
			break;
		case VIRTIO_BLK_S_UNSUPP:
			uptodate = -ENOTTY;
			break;
		default:
			uptodate = 0;
			break;
		}

		end_dequeued_request(vbr->req, uptodate);
		list_del(&vbr->list);
		mempool_free(vbr, vblk->pool);
	}
	/* In case queue is stopped waiting for more buffers. */
	blk_start_queue(vblk->disk->queue);
	spin_unlock_irqrestore(&vblk->lock, flags);
}

static bool do_req(struct request_queue *q, struct virtio_blk *vblk,
		   struct request *req)
{
	unsigned long num, out, in;
	struct virtblk_req *vbr;

	vbr = mempool_alloc(vblk->pool, GFP_ATOMIC);
	if (!vbr)
		/* When another request finishes we'll try again. */
		return false;

	vbr->req = req;
	if (blk_fs_request(vbr->req)) {
		vbr->out_hdr.type = 0;
		vbr->out_hdr.sector = vbr->req->sector;
		vbr->out_hdr.ioprio = vbr->req->ioprio;
	} else if (blk_pc_request(vbr->req)) {
		vbr->out_hdr.type = VIRTIO_BLK_T_SCSI_CMD;
		vbr->out_hdr.sector = 0;
		vbr->out_hdr.ioprio = vbr->req->ioprio;
	} else {
		/* We don't put anything else in the queue. */
		BUG();
	}

	if (blk_barrier_rq(vbr->req))
		vbr->out_hdr.type |= VIRTIO_BLK_T_BARRIER;

	/* This init could be done at vblk creation time */
	sg_init_table(vblk->sg, VIRTIO_MAX_SG);
	sg_set_buf(&vblk->sg[0], &vbr->out_hdr, sizeof(vbr->out_hdr));
	num = blk_rq_map_sg(q, vbr->req, vblk->sg+1);
	sg_set_buf(&vblk->sg[num+1], &vbr->in_hdr, sizeof(vbr->in_hdr));

	if (rq_data_dir(vbr->req) == WRITE) {
		vbr->out_hdr.type |= VIRTIO_BLK_T_OUT;
		out = 1 + num;
		in = 1;
	} else {
		vbr->out_hdr.type |= VIRTIO_BLK_T_IN;
		out = 1;
		in = 1 + num;
	}

	if (vblk->vq->vq_ops->add_buf(vblk->vq, vblk->sg, out, in, vbr)) {
		mempool_free(vbr, vblk->pool);
		return false;
	}

	list_add_tail(&vbr->list, &vblk->reqs);
	return true;
}

static void do_virtblk_request(struct request_queue *q)
{
	struct virtio_blk *vblk = NULL;
	struct request *req;
	unsigned int issued = 0;

	while ((req = elv_next_request(q)) != NULL) {
		vblk = req->rq_disk->private_data;
		BUG_ON(req->nr_phys_segments > ARRAY_SIZE(vblk->sg));

		/* If this request fails, stop queue and wait for something to
		   finish to restart it. */
		if (!do_req(q, vblk, req)) {
			blk_stop_queue(q);
			break;
		}
		blkdev_dequeue_request(req);
		issued++;
	}

	if (issued)
		vblk->vq->vq_ops->kick(vblk->vq);
}

static int virtblk_ioctl(struct inode *inode, struct file *filp,
			 unsigned cmd, unsigned long data)
{
	return scsi_cmd_ioctl(filp, inode->i_bdev->bd_disk->queue,
			      inode->i_bdev->bd_disk, cmd,
			      (void __user *)data);
}

/* We provide getgeo only to please some old bootloader/partitioning tools */
static int virtblk_getgeo(struct block_device *bd, struct hd_geometry *geo)
{
	/* some standard values, similar to sd */
	geo->heads = 1 << 6;
	geo->sectors = 1 << 5;
	geo->cylinders = get_capacity(bd->bd_disk) >> 11;
	return 0;
}

static struct block_device_operations virtblk_fops = {
	.ioctl  = virtblk_ioctl,
	.owner  = THIS_MODULE,
	.getgeo = virtblk_getgeo,
};

static int virtblk_probe(struct virtio_device *vdev)
{
	struct virtio_blk *vblk;
	int err;
	u64 cap;
	u32 v;

	if (minor >= 1 << MINORBITS)
		return -ENOSPC;

	vdev->priv = vblk = kmalloc(sizeof(*vblk), GFP_KERNEL);
	if (!vblk) {
		err = -ENOMEM;
		goto out;
	}

	INIT_LIST_HEAD(&vblk->reqs);
	spin_lock_init(&vblk->lock);
	vblk->vdev = vdev;

	/* We expect one virtqueue, for output. */
	vblk->vq = vdev->config->find_vq(vdev, 0, blk_done);
	if (IS_ERR(vblk->vq)) {
		err = PTR_ERR(vblk->vq);
		goto out_free_vblk;
	}

	vblk->pool = mempool_create_kmalloc_pool(1,sizeof(struct virtblk_req));
	if (!vblk->pool) {
		err = -ENOMEM;
		goto out_free_vq;
	}

	/* FIXME: How many partitions?  How long is a piece of string? */
	vblk->disk = alloc_disk(1 << PART_BITS);
	if (!vblk->disk) {
		err = -ENOMEM;
		goto out_mempool;
	}

	vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
	if (!vblk->disk->queue) {
		err = -ENOMEM;
		goto out_put_disk;
	}

	sprintf(vblk->disk->disk_name, "vd%c", virtblk_index++);
	vblk->disk->major = major;
	vblk->disk->first_minor = minor;
	vblk->disk->private_data = vblk;
	vblk->disk->fops = &virtblk_fops;

	minor += 1 << PART_BITS;

	/* If barriers are supported, tell block layer that queue is ordered */
	if (vdev->config->feature(vdev, VIRTIO_BLK_F_BARRIER))
		blk_queue_ordered(vblk->disk->queue, QUEUE_ORDERED_TAG, NULL);

	/* Host must always specify the capacity. */
	__virtio_config_val(vdev, offsetof(struct virtio_blk_config, capacity),
			    &cap);

	/* If capacity is too big, truncate with warning. */
	if ((sector_t)cap != cap) {
		dev_warn(&vdev->dev, "Capacity %llu too large: truncating\n",
			 (unsigned long long)cap);
		cap = (sector_t)-1;
	}
	set_capacity(vblk->disk, cap);

	/* Host can optionally specify maximum segment size and number of
	 * segments. */
	err = virtio_config_val(vdev, VIRTIO_BLK_F_SIZE_MAX,
				offsetof(struct virtio_blk_config, size_max),
				&v);
	if (!err)
		blk_queue_max_segment_size(vblk->disk->queue, v);

	err = virtio_config_val(vdev, VIRTIO_BLK_F_SEG_MAX,
				offsetof(struct virtio_blk_config, seg_max),
				&v);
	if (!err)
		blk_queue_max_hw_segments(vblk->disk->queue, v);

	add_disk(vblk->disk);
	return 0;

out_put_disk:
	put_disk(vblk->disk);
out_mempool:
	mempool_destroy(vblk->pool);
out_free_vq:
	vdev->config->del_vq(vblk->vq);
out_free_vblk:
	kfree(vblk);
out:
	return err;
}

static void virtblk_remove(struct virtio_device *vdev)
{
	struct virtio_blk *vblk = vdev->priv;
	int major = vblk->disk->major;

	/* Nothing should be pending. */
	BUG_ON(!list_empty(&vblk->reqs));

	/* Stop all the virtqueues. */
	vdev->config->reset(vdev);

	blk_cleanup_queue(vblk->disk->queue);
	put_disk(vblk->disk);
	unregister_blkdev(major, "virtblk");
	mempool_destroy(vblk->pool);
	vdev->config->del_vq(vblk->vq);
	kfree(vblk);
}

static struct virtio_device_id id_table[] = {
	{ VIRTIO_ID_BLOCK, VIRTIO_DEV_ANY_ID },
	{ 0 },
};

static struct virtio_driver virtio_blk = {
	.driver.name =	KBUILD_MODNAME,
	.driver.owner =	THIS_MODULE,
	.id_table =	id_table,
	.probe =	virtblk_probe,
	.remove =	__devexit_p(virtblk_remove),
};

static int __init init(void)
{
	major = register_blkdev(0, "virtblk");
	if (major < 0)
		return major;
	return register_virtio_driver(&virtio_blk);
}

static void __exit fini(void)
{
	unregister_blkdev(major, "virtblk");
	unregister_virtio_driver(&virtio_blk);
}
module_init(init);
module_exit(fini);

MODULE_DEVICE_TABLE(virtio, id_table);
MODULE_DESCRIPTION("Virtio block driver");
MODULE_LICENSE("GPL");
Commit	Line	Data
e467cde2 RR	1	//#define DEBUG
	2	#include <linux/spinlock.h>
	3	#include <linux/blkdev.h>
	4	#include <linux/hdreg.h>
	5	#include <linux/virtio.h>
	6	#include <linux/virtio_blk.h>
3d1266c7 JA	7	#include <linux/scatterlist.h>
	8
	9	#define VIRTIO_MAX_SG (3+MAX_PHYS_SEGMENTS)
4f3bf19c	10	#define PART_BITS 4
e467cde2 RR	11
e467cde2 RR	12	static unsigned char virtblk_index = 'a';
4f3bf19c CB	13	static int major, minor;
4f3bf19c CB	14
e467cde2 RR	15	struct virtio_blk
	16	{
	17	spinlock_t lock;
	18
	19	struct virtio_device *vdev;
	20	struct virtqueue *vq;
	21
	22	/* The disk structure for the kernel. */
	23	struct gendisk *disk;
	24
	25	/* Request tracking. */
	26	struct list_head reqs;
	27
	28	mempool_t *pool;
	29
	30	/* Scatterlist: can be too big for stack. */
3d1266c7	31	struct scatterlist sg[VIRTIO_MAX_SG];
e467cde2 RR	32	};
	33
	34	struct virtblk_req
	35	{
	36	struct list_head list;
	37	struct request *req;
	38	struct virtio_blk_outhdr out_hdr;
	39	struct virtio_blk_inhdr in_hdr;
	40	};
	41
18445c4d	42	static void blk_done(struct virtqueue *vq)
e467cde2 RR	43	{
	44	struct virtio_blk *vblk = vq->vdev->priv;
	45	struct virtblk_req *vbr;
	46	unsigned int len;
	47	unsigned long flags;
	48
	49	spin_lock_irqsave(&vblk->lock, flags);
	50	while ((vbr = vblk->vq->vq_ops->get_buf(vblk->vq, &len)) != NULL) {
	51	int uptodate;
	52	switch (vbr->in_hdr.status) {
	53	case VIRTIO_BLK_S_OK:
	54	uptodate = 1;
	55	break;
	56	case VIRTIO_BLK_S_UNSUPP:
	57	uptodate = -ENOTTY;
	58	break;
	59	default:
	60	uptodate = 0;
	61	break;
	62	}
	63
	64	end_dequeued_request(vbr->req, uptodate);
	65	list_del(&vbr->list);
	66	mempool_free(vbr, vblk->pool);
	67	}
	68	/* In case queue is stopped waiting for more buffers. */
	69	blk_start_queue(vblk->disk->queue);
	70	spin_unlock_irqrestore(&vblk->lock, flags);
e467cde2 RR	71	}
	72
	73	static bool do_req(struct request_queue q, struct virtio_blk vblk,
	74	struct request *req)
	75	{
	76	unsigned long num, out, in;
	77	struct virtblk_req *vbr;
	78
	79	vbr = mempool_alloc(vblk->pool, GFP_ATOMIC);
	80	if (!vbr)
	81	/* When another request finishes we'll try again. */
	82	return false;
	83
	84	vbr->req = req;
	85	if (blk_fs_request(vbr->req)) {
	86	vbr->out_hdr.type = 0;
	87	vbr->out_hdr.sector = vbr->req->sector;
	88	vbr->out_hdr.ioprio = vbr->req->ioprio;
	89	} else if (blk_pc_request(vbr->req)) {
	90	vbr->out_hdr.type = VIRTIO_BLK_T_SCSI_CMD;
	91	vbr->out_hdr.sector = 0;
	92	vbr->out_hdr.ioprio = vbr->req->ioprio;
	93	} else {
	94	/* We don't put anything else in the queue. */
	95	BUG();
	96	}
	97
	98	if (blk_barrier_rq(vbr->req))
	99	vbr->out_hdr.type \|= VIRTIO_BLK_T_BARRIER;
	100
3d1266c7 JA	101	/* This init could be done at vblk creation time */
3d1266c7 JA	102	sg_init_table(vblk->sg, VIRTIO_MAX_SG);
e467cde2 RR	103	sg_set_buf(&vblk->sg[0], &vbr->out_hdr, sizeof(vbr->out_hdr));
	104	num = blk_rq_map_sg(q, vbr->req, vblk->sg+1);
	105	sg_set_buf(&vblk->sg[num+1], &vbr->in_hdr, sizeof(vbr->in_hdr));
	106
	107	if (rq_data_dir(vbr->req) == WRITE) {
	108	vbr->out_hdr.type \|= VIRTIO_BLK_T_OUT;
	109	out = 1 + num;
	110	in = 1;
	111	} else {
	112	vbr->out_hdr.type \|= VIRTIO_BLK_T_IN;
	113	out = 1;
	114	in = 1 + num;
	115	}
	116
	117	if (vblk->vq->vq_ops->add_buf(vblk->vq, vblk->sg, out, in, vbr)) {
	118	mempool_free(vbr, vblk->pool);
	119	return false;
	120	}
	121
	122	list_add_tail(&vbr->list, &vblk->reqs);
	123	return true;
	124	}
	125
	126	static void do_virtblk_request(struct request_queue *q)
	127	{
	128	struct virtio_blk *vblk = NULL;
	129	struct request *req;
	130	unsigned int issued = 0;
	131
	132	while ((req = elv_next_request(q)) != NULL) {
	133	vblk = req->rq_disk->private_data;
	134	BUG_ON(req->nr_phys_segments > ARRAY_SIZE(vblk->sg));
	135
	136	/* If this request fails, stop queue and wait for something to
	137	finish to restart it. */
	138	if (!do_req(q, vblk, req)) {
	139	blk_stop_queue(q);
	140	break;
	141	}
	142	blkdev_dequeue_request(req);
	143	issued++;
	144	}
	145
	146	if (issued)
	147	vblk->vq->vq_ops->kick(vblk->vq);
	148	}
	149
	150	static int virtblk_ioctl(struct inode inode, struct file filp,
	151	unsigned cmd, unsigned long data)
	152	{
	153	return scsi_cmd_ioctl(filp, inode->i_bdev->bd_disk->queue,
	154	inode->i_bdev->bd_disk, cmd,
	155	(void __user *)data);
	156	}
	157
135da0b0 CB	158	/* We provide getgeo only to please some old bootloader/partitioning tools */
	159	static int virtblk_getgeo(struct block_device bd, struct hd_geometry geo)
	160	{
	161	/* some standard values, similar to sd */
	162	geo->heads = 1 << 6;
	163	geo->sectors = 1 << 5;
	164	geo->cylinders = get_capacity(bd->bd_disk) >> 11;
	165	return 0;
	166	}
	167
e467cde2	168	static struct block_device_operations virtblk_fops = {
135da0b0 CB	169	.ioctl = virtblk_ioctl,
	170	.owner = THIS_MODULE,
	171	.getgeo = virtblk_getgeo,
e467cde2 RR	172	};
	173
	174	static int virtblk_probe(struct virtio_device *vdev)
	175	{
	176	struct virtio_blk *vblk;
4f3bf19c	177	int err;
e467cde2 RR	178	u64 cap;
	179	u32 v;
	180
4f3bf19c CB	181	if (minor >= 1 << MINORBITS)
	182	return -ENOSPC;
	183
e467cde2 RR	184	vdev->priv = vblk = kmalloc(sizeof(*vblk), GFP_KERNEL);
	185	if (!vblk) {
	186	err = -ENOMEM;
	187	goto out;
	188	}
	189
	190	INIT_LIST_HEAD(&vblk->reqs);
	191	spin_lock_init(&vblk->lock);
	192	vblk->vdev = vdev;
	193
	194	/* We expect one virtqueue, for output. */
a586d4f6	195	vblk->vq = vdev->config->find_vq(vdev, 0, blk_done);
e467cde2 RR	196	if (IS_ERR(vblk->vq)) {
	197	err = PTR_ERR(vblk->vq);
	198	goto out_free_vblk;
	199	}
	200
	201	vblk->pool = mempool_create_kmalloc_pool(1,sizeof(struct virtblk_req));
	202	if (!vblk->pool) {
	203	err = -ENOMEM;
	204	goto out_free_vq;
	205	}
	206
e467cde2	207	/* FIXME: How many partitions? How long is a piece of string? */
4f3bf19c	208	vblk->disk = alloc_disk(1 << PART_BITS);
e467cde2 RR	209	if (!vblk->disk) {
e467cde2 RR	210	err = -ENOMEM;
4f3bf19c	211	goto out_mempool;
e467cde2 RR	212	}
	213
	214	vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
	215	if (!vblk->disk->queue) {
	216	err = -ENOMEM;
	217	goto out_put_disk;
	218	}
	219
	220	sprintf(vblk->disk->disk_name, "vd%c", virtblk_index++);
	221	vblk->disk->major = major;
4f3bf19c	222	vblk->disk->first_minor = minor;
e467cde2 RR	223	vblk->disk->private_data = vblk;
	224	vblk->disk->fops = &virtblk_fops;
	225
4f3bf19c CB	226	minor += 1 << PART_BITS;
4f3bf19c CB	227
e467cde2	228	/* If barriers are supported, tell block layer that queue is ordered */
a586d4f6	229	if (vdev->config->feature(vdev, VIRTIO_BLK_F_BARRIER))
e467cde2 RR	230	blk_queue_ordered(vblk->disk->queue, QUEUE_ORDERED_TAG, NULL);
e467cde2 RR	231
a586d4f6 RR	232	/* Host must always specify the capacity. */
	233	__virtio_config_val(vdev, offsetof(struct virtio_blk_config, capacity),
	234	&cap);
e467cde2 RR	235
	236	/* If capacity is too big, truncate with warning. */
	237	if ((sector_t)cap != cap) {
	238	dev_warn(&vdev->dev, "Capacity %llu too large: truncating\n",
	239	(unsigned long long)cap);
	240	cap = (sector_t)-1;
	241	}
	242	set_capacity(vblk->disk, cap);
	243
a586d4f6 RR	244	/* Host can optionally specify maximum segment size and number of
	245	* segments. */
	246	err = virtio_config_val(vdev, VIRTIO_BLK_F_SIZE_MAX,
	247	offsetof(struct virtio_blk_config, size_max),
	248	&v);
e467cde2 RR	249	if (!err)
e467cde2 RR	250	blk_queue_max_segment_size(vblk->disk->queue, v);
e467cde2	251
a586d4f6 RR	252	err = virtio_config_val(vdev, VIRTIO_BLK_F_SEG_MAX,
	253	offsetof(struct virtio_blk_config, seg_max),
	254	&v);
e467cde2 RR	255	if (!err)
e467cde2 RR	256	blk_queue_max_hw_segments(vblk->disk->queue, v);
e467cde2 RR	257
	258	add_disk(vblk->disk);
	259	return 0;
	260
	261	out_put_disk:
	262	put_disk(vblk->disk);
e467cde2 RR	263	out_mempool:
	264	mempool_destroy(vblk->pool);
	265	out_free_vq:
	266	vdev->config->del_vq(vblk->vq);
	267	out_free_vblk:
	268	kfree(vblk);
	269	out:
	270	return err;
	271	}
	272
	273	static void virtblk_remove(struct virtio_device *vdev)
	274	{
	275	struct virtio_blk *vblk = vdev->priv;
	276	int major = vblk->disk->major;
	277
6e5aa7ef	278	/* Nothing should be pending. */
e467cde2	279	BUG_ON(!list_empty(&vblk->reqs));
6e5aa7ef RR	280
	281	/* Stop all the virtqueues. */
	282	vdev->config->reset(vdev);
	283
e467cde2 RR	284	blk_cleanup_queue(vblk->disk->queue);
	285	put_disk(vblk->disk);
	286	unregister_blkdev(major, "virtblk");
	287	mempool_destroy(vblk->pool);
74b2553f	288	vdev->config->del_vq(vblk->vq);
e467cde2 RR	289	kfree(vblk);
	290	}
	291
	292	static struct virtio_device_id id_table[] = {
	293	{ VIRTIO_ID_BLOCK, VIRTIO_DEV_ANY_ID },
	294	{ 0 },
	295	};
	296
	297	static struct virtio_driver virtio_blk = {
	298	.driver.name = KBUILD_MODNAME,
	299	.driver.owner = THIS_MODULE,
	300	.id_table = id_table,
	301	.probe = virtblk_probe,
	302	.remove = __devexit_p(virtblk_remove),
	303	};
	304
	305	static int __init init(void)
	306	{
4f3bf19c CB	307	major = register_blkdev(0, "virtblk");
	308	if (major < 0)
	309	return major;
e467cde2 RR	310	return register_virtio_driver(&virtio_blk);
	311	}
	312
	313	static void __exit fini(void)
	314	{
4f3bf19c	315	unregister_blkdev(major, "virtblk");
e467cde2 RR	316	unregister_virtio_driver(&virtio_blk);
	317	}
	318	module_init(init);
	319	module_exit(fini);
	320
	321	MODULE_DEVICE_TABLE(virtio, id_table);
	322	MODULE_DESCRIPTION("Virtio block driver");
	323	MODULE_LICENSE("GPL");