[net-next-2.6.git] / drivers / block / virtio_blk.c

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

#define PART_BITS 4

static int major, index;

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;

	/* What host tells us, plus 2 for header & tailer. */
	unsigned int sg_elems;

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

struct virtblk_req
{
	struct list_head list;
	struct request *req;
	struct virtio_blk_outhdr out_hdr;
	struct virtio_scsi_inhdr in_hdr;
	u8 status;
};

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 = virtqueue_get_buf(vblk->vq, &len)) != NULL) {
		int error;

		switch (vbr->status) {
		case VIRTIO_BLK_S_OK:
			error = 0;
			break;
		case VIRTIO_BLK_S_UNSUPP:
			error = -ENOTTY;
			break;
		default:
			error = -EIO;
			break;
		}

		switch (vbr->req->cmd_type) {
		case REQ_TYPE_BLOCK_PC:
			vbr->req->resid_len = vbr->in_hdr.residual;
			vbr->req->sense_len = vbr->in_hdr.sense_len;
			vbr->req->errors = vbr->in_hdr.errors;
			break;
		case REQ_TYPE_SPECIAL:
			vbr->req->errors = (error != 0);
			break;
		default:
			break;
		}

		__blk_end_request_all(vbr->req, error);
		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 = 0, in = 0;
	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 (req->cmd_flags & REQ_FLUSH) {
		vbr->out_hdr.type = VIRTIO_BLK_T_FLUSH;
		vbr->out_hdr.sector = 0;
		vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	} else {
		switch (req->cmd_type) {
		case REQ_TYPE_FS:
			vbr->out_hdr.type = 0;
			vbr->out_hdr.sector = blk_rq_pos(vbr->req);
			vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
			break;
		case REQ_TYPE_BLOCK_PC:
			vbr->out_hdr.type = VIRTIO_BLK_T_SCSI_CMD;
			vbr->out_hdr.sector = 0;
			vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
			break;
		case REQ_TYPE_SPECIAL:
			vbr->out_hdr.type = VIRTIO_BLK_T_GET_ID;
			vbr->out_hdr.sector = 0;
			vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
			break;
		default:
			/* We don't put anything else in the queue. */
			BUG();
		}
	}

	if (vbr->req->cmd_flags & REQ_HARDBARRIER)
		vbr->out_hdr.type |= VIRTIO_BLK_T_BARRIER;

	sg_set_buf(&vblk->sg[out++], &vbr->out_hdr, sizeof(vbr->out_hdr));

	/*
	 * If this is a packet command we need a couple of additional headers.
	 * Behind the normal outhdr we put a segment with the scsi command
	 * block, and before the normal inhdr we put the sense data and the
	 * inhdr with additional status information before the normal inhdr.
	 */
	if (vbr->req->cmd_type == REQ_TYPE_BLOCK_PC)
		sg_set_buf(&vblk->sg[out++], vbr->req->cmd, vbr->req->cmd_len);

	num = blk_rq_map_sg(q, vbr->req, vblk->sg + out);

	if (vbr->req->cmd_type == REQ_TYPE_BLOCK_PC) {
		sg_set_buf(&vblk->sg[num + out + in++], vbr->req->sense, 96);
		sg_set_buf(&vblk->sg[num + out + in++], &vbr->in_hdr,
			   sizeof(vbr->in_hdr));
	}

	sg_set_buf(&vblk->sg[num + out + in++], &vbr->status,
		   sizeof(vbr->status));

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

	if (virtqueue_add_buf(vblk->vq, vblk->sg, out, in, vbr) < 0) {
		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 = q->queuedata;
	struct request *req;
	unsigned int issued = 0;

	while ((req = blk_peek_request(q)) != NULL) {
		BUG_ON(req->nr_phys_segments + 2 > vblk->sg_elems);

		/* 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;
		}
		blk_start_request(req);
		issued++;
	}

	if (issued)
		virtqueue_kick(vblk->vq);
}

/* return id (s/n) string for *disk to *id_str
 */
static int virtblk_get_id(struct gendisk *disk, char *id_str)
{
	struct virtio_blk *vblk = disk->private_data;
	struct request *req;
	struct bio *bio;

	bio = bio_map_kern(vblk->disk->queue, id_str, VIRTIO_BLK_ID_BYTES,
			   GFP_KERNEL);
	if (IS_ERR(bio))
		return PTR_ERR(bio);

	req = blk_make_request(vblk->disk->queue, bio, GFP_KERNEL);
	if (IS_ERR(req)) {
		bio_put(bio);
		return PTR_ERR(req);
	}

	req->cmd_type = REQ_TYPE_SPECIAL;
	return blk_execute_rq(vblk->disk->queue, vblk->disk, req, false);
}

static int virtblk_ioctl(struct block_device *bdev, fmode_t mode,
			 unsigned cmd, unsigned long data)
{
	struct gendisk *disk = bdev->bd_disk;
	struct virtio_blk *vblk = disk->private_data;

	if (cmd == 0x56424944) { /* 'VBID' */
		void __user *usr_data = (void __user *)data;
		char id_str[VIRTIO_BLK_ID_BYTES];
		int err;

		err = virtblk_get_id(disk, id_str);
		if (!err && copy_to_user(usr_data, id_str, VIRTIO_BLK_ID_BYTES))
			err = -EFAULT;
		return err;
	}
	/*
	 * Only allow the generic SCSI ioctls if the host can support it.
	 */
	if (!virtio_has_feature(vblk->vdev, VIRTIO_BLK_F_SCSI))
		return -ENOTTY;

	return scsi_cmd_ioctl(disk->queue, disk, mode, 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)
{
	struct virtio_blk *vblk = bd->bd_disk->private_data;
	struct virtio_blk_geometry vgeo;
	int err;

	/* see if the host passed in geometry config */
	err = virtio_config_val(vblk->vdev, VIRTIO_BLK_F_GEOMETRY,
				offsetof(struct virtio_blk_config, geometry),
				&vgeo);

	if (!err) {
		geo->heads = vgeo.heads;
		geo->sectors = vgeo.sectors;
		geo->cylinders = vgeo.cylinders;
	} else {
		/* 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 const struct block_device_operations virtblk_fops = {
	.locked_ioctl = virtblk_ioctl,
	.owner  = THIS_MODULE,
	.getgeo = virtblk_getgeo,
};

static int index_to_minor(int index)
{
	return index << PART_BITS;
}

static int __devinit virtblk_probe(struct virtio_device *vdev)
{
	struct virtio_blk *vblk;
	struct request_queue *q;
	int err;
	u64 cap;
	u32 v, blk_size, sg_elems, opt_io_size;
	u16 min_io_size;
	u8 physical_block_exp, alignment_offset;

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

	/* We need to know how many segments before we allocate. */
	err = virtio_config_val(vdev, VIRTIO_BLK_F_SEG_MAX,
				offsetof(struct virtio_blk_config, seg_max),
				&sg_elems);

	/* We need at least one SG element, whatever they say. */
	if (err || !sg_elems)
		sg_elems = 1;

	/* We need an extra sg elements at head and tail. */
	sg_elems += 2;
	vdev->priv = vblk = kmalloc(sizeof(*vblk) +
				    sizeof(vblk->sg[0]) * sg_elems, GFP_KERNEL);
	if (!vblk) {
		err = -ENOMEM;
		goto out;
	}

	INIT_LIST_HEAD(&vblk->reqs);
	spin_lock_init(&vblk->lock);
	vblk->vdev = vdev;
	vblk->sg_elems = sg_elems;
	sg_init_table(vblk->sg, vblk->sg_elems);

	/* We expect one virtqueue, for output. */
	vblk->vq = virtio_find_single_vq(vdev, blk_done, "requests");
	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;
	}

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

	q->queuedata = vblk;

	if (index < 26) {
		sprintf(vblk->disk->disk_name, "vd%c", 'a' + index % 26);
	} else if (index < (26 + 1) * 26) {
		sprintf(vblk->disk->disk_name, "vd%c%c",
			'a' + index / 26 - 1, 'a' + index % 26);
	} else {
		const unsigned int m1 = (index / 26 - 1) / 26 - 1;
		const unsigned int m2 = (index / 26 - 1) % 26;
		const unsigned int m3 =  index % 26;
		sprintf(vblk->disk->disk_name, "vd%c%c%c",
			'a' + m1, 'a' + m2, 'a' + m3);
	}

	vblk->disk->major = major;
	vblk->disk->first_minor = index_to_minor(index);
	vblk->disk->private_data = vblk;
	vblk->disk->fops = &virtblk_fops;
	vblk->disk->driverfs_dev = &vdev->dev;
	index++;

	/* If barriers are supported, tell block layer that queue is ordered */
	if (virtio_has_feature(vdev, VIRTIO_BLK_F_FLUSH))
		blk_queue_ordered(q, QUEUE_ORDERED_DRAIN_FLUSH);
	else if (virtio_has_feature(vdev, VIRTIO_BLK_F_BARRIER))
		blk_queue_ordered(q, QUEUE_ORDERED_TAG);

	/* If disk is read-only in the host, the guest should obey */
	if (virtio_has_feature(vdev, VIRTIO_BLK_F_RO))
		set_disk_ro(vblk->disk, 1);

	/* Host must always specify the capacity. */
	vdev->config->get(vdev, offsetof(struct virtio_blk_config, capacity),
			  &cap, sizeof(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);

	/* We can handle whatever the host told us to handle. */
	blk_queue_max_segments(q, vblk->sg_elems-2);

	/* No need to bounce any requests */
	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);

	/* No real sector limit. */
	blk_queue_max_hw_sectors(q, -1U);

	/* 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(q, v);
	else
		blk_queue_max_segment_size(q, -1U);

	/* Host can optionally specify the block size of the device */
	err = virtio_config_val(vdev, VIRTIO_BLK_F_BLK_SIZE,
				offsetof(struct virtio_blk_config, blk_size),
				&blk_size);
	if (!err)
		blk_queue_logical_block_size(q, blk_size);
	else
		blk_size = queue_logical_block_size(q);

	/* Use topology information if available */
	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
			offsetof(struct virtio_blk_config, physical_block_exp),
			&physical_block_exp);
	if (!err && physical_block_exp)
		blk_queue_physical_block_size(q,
				blk_size * (1 << physical_block_exp));

	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
			offsetof(struct virtio_blk_config, alignment_offset),
			&alignment_offset);
	if (!err && alignment_offset)
		blk_queue_alignment_offset(q, blk_size * alignment_offset);

	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
			offsetof(struct virtio_blk_config, min_io_size),
			&min_io_size);
	if (!err && min_io_size)
		blk_queue_io_min(q, blk_size * min_io_size);

	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
			offsetof(struct virtio_blk_config, opt_io_size),
			&opt_io_size);
	if (!err && opt_io_size)
		blk_queue_io_opt(q, blk_size * opt_io_size);


	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_vqs(vdev);
out_free_vblk:
	kfree(vblk);
out:
	return err;
}

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

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

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

	del_gendisk(vblk->disk);
	blk_cleanup_queue(vblk->disk->queue);
	put_disk(vblk->disk);
	mempool_destroy(vblk->pool);
	vdev->config->del_vqs(vdev);
	kfree(vblk);
}

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

static unsigned int features[] = {
	VIRTIO_BLK_F_BARRIER, VIRTIO_BLK_F_SEG_MAX, VIRTIO_BLK_F_SIZE_MAX,
	VIRTIO_BLK_F_GEOMETRY, VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
	VIRTIO_BLK_F_SCSI, VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY
};

/*
 * virtio_blk causes spurious section mismatch warning by
 * simultaneously referring to a __devinit and a __devexit function.
 * Use __refdata to avoid this warning.
 */
static struct virtio_driver __refdata virtio_blk = {
	.feature_table = features,
	.feature_table_size = ARRAY_SIZE(features),
	.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
e467cde2 RR	2	#include <linux/spinlock.h>
5a0e3ad6	3	#include <linux/slab.h>
e467cde2 RR	4	#include <linux/blkdev.h>
	5	#include <linux/hdreg.h>
	6	#include <linux/virtio.h>
	7	#include <linux/virtio_blk.h>
3d1266c7 JA	8	#include <linux/scatterlist.h>
3d1266c7 JA	9
4f3bf19c	10	#define PART_BITS 4
e467cde2	11
d50ed907	12	static int major, index;
4f3bf19c	13
e467cde2 RR	14	struct virtio_blk
	15	{
	16	spinlock_t lock;
	17
	18	struct virtio_device *vdev;
	19	struct virtqueue *vq;
	20
	21	/* The disk structure for the kernel. */
	22	struct gendisk *disk;
	23
	24	/* Request tracking. */
	25	struct list_head reqs;
	26
	27	mempool_t *pool;
	28
0864b79a RR	29	/* What host tells us, plus 2 for header & tailer. */
	30	unsigned int sg_elems;
	31
e467cde2	32	/* Scatterlist: can be too big for stack. */
0864b79a	33	struct scatterlist sg[/sg_elems/];
e467cde2 RR	34	};
	35
	36	struct virtblk_req
	37	{
	38	struct list_head list;
	39	struct request *req;
	40	struct virtio_blk_outhdr out_hdr;
1cde26f9	41	struct virtio_scsi_inhdr in_hdr;
cb38fa23	42	u8 status;
e467cde2 RR	43	};
e467cde2 RR	44
18445c4d	45	static void blk_done(struct virtqueue *vq)
e467cde2 RR	46	{
	47	struct virtio_blk *vblk = vq->vdev->priv;
	48	struct virtblk_req *vbr;
	49	unsigned int len;
	50	unsigned long flags;
	51
	52	spin_lock_irqsave(&vblk->lock, flags);
09ec6b69	53	while ((vbr = virtqueue_get_buf(vblk->vq, &len)) != NULL) {
8316982a	54	int error;
1cde26f9	55
cb38fa23	56	switch (vbr->status) {
e467cde2	57	case VIRTIO_BLK_S_OK:
8316982a	58	error = 0;
e467cde2 RR	59	break;
e467cde2 RR	60	case VIRTIO_BLK_S_UNSUPP:
8316982a	61	error = -ENOTTY;
e467cde2 RR	62	break;
e467cde2 RR	63	default:
8316982a	64	error = -EIO;
e467cde2 RR	65	break;
	66	}
	67
33659ebb CH	68	switch (vbr->req->cmd_type) {
33659ebb CH	69	case REQ_TYPE_BLOCK_PC:
1cde26f9 HR	70	vbr->req->resid_len = vbr->in_hdr.residual;
	71	vbr->req->sense_len = vbr->in_hdr.sense_len;
	72	vbr->req->errors = vbr->in_hdr.errors;
33659ebb CH	73	break;
33659ebb CH	74	case REQ_TYPE_SPECIAL:
4cb2ea28	75	vbr->req->errors = (error != 0);
33659ebb	76	break;
15fa6e81 JA	77	default:
15fa6e81 JA	78	break;
33659ebb	79	}
1cde26f9	80
40cbbb78	81	__blk_end_request_all(vbr->req, error);
e467cde2 RR	82	list_del(&vbr->list);
	83	mempool_free(vbr, vblk->pool);
	84	}
	85	/* In case queue is stopped waiting for more buffers. */
	86	blk_start_queue(vblk->disk->queue);
	87	spin_unlock_irqrestore(&vblk->lock, flags);
e467cde2 RR	88	}
	89
	90	static bool do_req(struct request_queue q, struct virtio_blk vblk,
	91	struct request *req)
	92	{
1cde26f9	93	unsigned long num, out = 0, in = 0;
e467cde2 RR	94	struct virtblk_req *vbr;
	95
	96	vbr = mempool_alloc(vblk->pool, GFP_ATOMIC);
	97	if (!vbr)
	98	/* When another request finishes we'll try again. */
	99	return false;
	100
	101	vbr->req = req;
dd40e456 FT	102
	103	if (req->cmd_flags & REQ_FLUSH) {
	104	vbr->out_hdr.type = VIRTIO_BLK_T_FLUSH;
4cb2ea28	105	vbr->out_hdr.sector = 0;
4cb2ea28	106	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
dd40e456 FT	107	} else {
	108	switch (req->cmd_type) {
	109	case REQ_TYPE_FS:
	110	vbr->out_hdr.type = 0;
	111	vbr->out_hdr.sector = blk_rq_pos(vbr->req);
	112	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	113	break;
	114	case REQ_TYPE_BLOCK_PC:
	115	vbr->out_hdr.type = VIRTIO_BLK_T_SCSI_CMD;
f1b0ef06 CH	116	vbr->out_hdr.sector = 0;
	117	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	118	break;
dd40e456 FT	119	case REQ_TYPE_SPECIAL:
	120	vbr->out_hdr.type = VIRTIO_BLK_T_GET_ID;
	121	vbr->out_hdr.sector = 0;
	122	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	123	break;
	124	default:
	125	/* We don't put anything else in the queue. */
	126	BUG();
f1b0ef06	127	}
e467cde2 RR	128	}
e467cde2 RR	129
33659ebb	130	if (vbr->req->cmd_flags & REQ_HARDBARRIER)
e467cde2 RR	131	vbr->out_hdr.type \|= VIRTIO_BLK_T_BARRIER;
e467cde2 RR	132
1cde26f9	133	sg_set_buf(&vblk->sg[out++], &vbr->out_hdr, sizeof(vbr->out_hdr));
e467cde2	134
1cde26f9 HR	135	/*
	136	* If this is a packet command we need a couple of additional headers.
	137	* Behind the normal outhdr we put a segment with the scsi command
	138	* block, and before the normal inhdr we put the sense data and the
	139	* inhdr with additional status information before the normal inhdr.
	140	*/
33659ebb	141	if (vbr->req->cmd_type == REQ_TYPE_BLOCK_PC)
1cde26f9 HR	142	sg_set_buf(&vblk->sg[out++], vbr->req->cmd, vbr->req->cmd_len);
	143
	144	num = blk_rq_map_sg(q, vbr->req, vblk->sg + out);
	145
33659ebb	146	if (vbr->req->cmd_type == REQ_TYPE_BLOCK_PC) {
1cde26f9 HR	147	sg_set_buf(&vblk->sg[num + out + in++], vbr->req->sense, 96);
	148	sg_set_buf(&vblk->sg[num + out + in++], &vbr->in_hdr,
	149	sizeof(vbr->in_hdr));
	150	}
	151
	152	sg_set_buf(&vblk->sg[num + out + in++], &vbr->status,
	153	sizeof(vbr->status));
	154
	155	if (num) {
	156	if (rq_data_dir(vbr->req) == WRITE) {
	157	vbr->out_hdr.type \|= VIRTIO_BLK_T_OUT;
	158	out += num;
	159	} else {
	160	vbr->out_hdr.type \|= VIRTIO_BLK_T_IN;
	161	in += num;
	162	}
e467cde2 RR	163	}
e467cde2 RR	164
09ec6b69	165	if (virtqueue_add_buf(vblk->vq, vblk->sg, out, in, vbr) < 0) {
e467cde2 RR	166	mempool_free(vbr, vblk->pool);
	167	return false;
	168	}
	169
	170	list_add_tail(&vbr->list, &vblk->reqs);
	171	return true;
	172	}
	173
	174	static void do_virtblk_request(struct request_queue *q)
	175	{
6c3b46f7	176	struct virtio_blk *vblk = q->queuedata;
e467cde2 RR	177	struct request *req;
	178	unsigned int issued = 0;
	179
9934c8c0	180	while ((req = blk_peek_request(q)) != NULL) {
0864b79a	181	BUG_ON(req->nr_phys_segments + 2 > vblk->sg_elems);
e467cde2 RR	182
	183	/* If this request fails, stop queue and wait for something to
	184	finish to restart it. */
	185	if (!do_req(q, vblk, req)) {
	186	blk_stop_queue(q);
	187	break;
	188	}
9934c8c0	189	blk_start_request(req);
e467cde2 RR	190	issued++;
	191	}
	192
	193	if (issued)
09ec6b69	194	virtqueue_kick(vblk->vq);
e467cde2 RR	195	}
e467cde2 RR	196
4cb2ea28	197	/* return id (s/n) string for disk to id_str
	198	*/
	199	static int virtblk_get_id(struct gendisk disk, char id_str)
	200	{
	201	struct virtio_blk *vblk = disk->private_data;
	202	struct request *req;
	203	struct bio *bio;
	204
	205	bio = bio_map_kern(vblk->disk->queue, id_str, VIRTIO_BLK_ID_BYTES,
	206	GFP_KERNEL);
	207	if (IS_ERR(bio))
	208	return PTR_ERR(bio);
	209
	210	req = blk_make_request(vblk->disk->queue, bio, GFP_KERNEL);
	211	if (IS_ERR(req)) {
	212	bio_put(bio);
	213	return PTR_ERR(req);
	214	}
	215
	216	req->cmd_type = REQ_TYPE_SPECIAL;
	217	return blk_execute_rq(vblk->disk->queue, vblk->disk, req, false);
	218	}
	219
4e109852	220	static int virtblk_ioctl(struct block_device *bdev, fmode_t mode,
e467cde2 RR	221	unsigned cmd, unsigned long data)
e467cde2 RR	222	{
1cde26f9 HR	223	struct gendisk *disk = bdev->bd_disk;
	224	struct virtio_blk *vblk = disk->private_data;
	225
bdb4a130	226	if (cmd == 0x56424944) { /* 'VBID' */
234f2725	227	void __user usr_data = (void __user )data;
	228	char id_str[VIRTIO_BLK_ID_BYTES];
	229	int err;
	230
	231	err = virtblk_get_id(disk, id_str);
	232	if (!err && copy_to_user(usr_data, id_str, VIRTIO_BLK_ID_BYTES))
	233	err = -EFAULT;
	234	return err;
	235	}
1cde26f9 HR	236	/*
	237	* Only allow the generic SCSI ioctls if the host can support it.
	238	*/
	239	if (!virtio_has_feature(vblk->vdev, VIRTIO_BLK_F_SCSI))
d9ecdea7	240	return -ENOTTY;
1cde26f9	241
3225beab RR	242	return scsi_cmd_ioctl(disk->queue, disk, mode, cmd,
3225beab RR	243	(void __user *)data);
e467cde2 RR	244	}
e467cde2 RR	245
135da0b0 CB	246	/* We provide getgeo only to please some old bootloader/partitioning tools */
	247	static int virtblk_getgeo(struct block_device bd, struct hd_geometry geo)
	248	{
48e4043d RH	249	struct virtio_blk *vblk = bd->bd_disk->private_data;
	250	struct virtio_blk_geometry vgeo;
	251	int err;
	252
	253	/* see if the host passed in geometry config */
	254	err = virtio_config_val(vblk->vdev, VIRTIO_BLK_F_GEOMETRY,
	255	offsetof(struct virtio_blk_config, geometry),
	256	&vgeo);
	257
	258	if (!err) {
	259	geo->heads = vgeo.heads;
	260	geo->sectors = vgeo.sectors;
	261	geo->cylinders = vgeo.cylinders;
	262	} else {
	263	/* some standard values, similar to sd */
	264	geo->heads = 1 << 6;
	265	geo->sectors = 1 << 5;
	266	geo->cylinders = get_capacity(bd->bd_disk) >> 11;
	267	}
135da0b0 CB	268	return 0;
	269	}
	270
83d5cde4	271	static const struct block_device_operations virtblk_fops = {
4e109852	272	.locked_ioctl = virtblk_ioctl,
135da0b0 CB	273	.owner = THIS_MODULE,
135da0b0 CB	274	.getgeo = virtblk_getgeo,
e467cde2 RR	275	};
e467cde2 RR	276
d50ed907 CB	277	static int index_to_minor(int index)
	278	{
	279	return index << PART_BITS;
	280	}
	281
98e94444	282	static int __devinit virtblk_probe(struct virtio_device *vdev)
e467cde2 RR	283	{
e467cde2 RR	284	struct virtio_blk *vblk;
69740c8b	285	struct request_queue *q;
4f3bf19c	286	int err;
e467cde2	287	u64 cap;
69740c8b CH	288	u32 v, blk_size, sg_elems, opt_io_size;
	289	u16 min_io_size;
	290	u8 physical_block_exp, alignment_offset;
e467cde2	291
d50ed907	292	if (index_to_minor(index) >= 1 << MINORBITS)
4f3bf19c CB	293	return -ENOSPC;
4f3bf19c CB	294
0864b79a RR	295	/* We need to know how many segments before we allocate. */
	296	err = virtio_config_val(vdev, VIRTIO_BLK_F_SEG_MAX,
	297	offsetof(struct virtio_blk_config, seg_max),
	298	&sg_elems);
a5b365a6 CH	299
	300	/* We need at least one SG element, whatever they say. */
	301	if (err \|\| !sg_elems)
0864b79a RR	302	sg_elems = 1;
	303
	304	/* We need an extra sg elements at head and tail. */
	305	sg_elems += 2;
	306	vdev->priv = vblk = kmalloc(sizeof(*vblk) +
	307	sizeof(vblk->sg[0]) * sg_elems, GFP_KERNEL);
e467cde2 RR	308	if (!vblk) {
	309	err = -ENOMEM;
	310	goto out;
	311	}
	312
	313	INIT_LIST_HEAD(&vblk->reqs);
	314	spin_lock_init(&vblk->lock);
	315	vblk->vdev = vdev;
0864b79a RR	316	vblk->sg_elems = sg_elems;
0864b79a RR	317	sg_init_table(vblk->sg, vblk->sg_elems);
e467cde2 RR	318
e467cde2 RR	319	/* We expect one virtqueue, for output. */
d2a7ddda	320	vblk->vq = virtio_find_single_vq(vdev, blk_done, "requests");
e467cde2 RR	321	if (IS_ERR(vblk->vq)) {
	322	err = PTR_ERR(vblk->vq);
	323	goto out_free_vblk;
	324	}
	325
	326	vblk->pool = mempool_create_kmalloc_pool(1,sizeof(struct virtblk_req));
	327	if (!vblk->pool) {
	328	err = -ENOMEM;
	329	goto out_free_vq;
	330	}
	331
e467cde2	332	/* FIXME: How many partitions? How long is a piece of string? */
4f3bf19c	333	vblk->disk = alloc_disk(1 << PART_BITS);
e467cde2 RR	334	if (!vblk->disk) {
e467cde2 RR	335	err = -ENOMEM;
4f3bf19c	336	goto out_mempool;
e467cde2 RR	337	}
e467cde2 RR	338
69740c8b CH	339	q = vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
69740c8b CH	340	if (!q) {
e467cde2 RR	341	err = -ENOMEM;
	342	goto out_put_disk;
	343	}
	344
69740c8b	345	q->queuedata = vblk;
7d116b62	346
d50ed907 CB	347	if (index < 26) {
	348	sprintf(vblk->disk->disk_name, "vd%c", 'a' + index % 26);
	349	} else if (index < (26 + 1) * 26) {
	350	sprintf(vblk->disk->disk_name, "vd%c%c",
	351	'a' + index / 26 - 1, 'a' + index % 26);
	352	} else {
	353	const unsigned int m1 = (index / 26 - 1) / 26 - 1;
	354	const unsigned int m2 = (index / 26 - 1) % 26;
	355	const unsigned int m3 = index % 26;
	356	sprintf(vblk->disk->disk_name, "vd%c%c%c",
	357	'a' + m1, 'a' + m2, 'a' + m3);
	358	}
	359
e467cde2	360	vblk->disk->major = major;
d50ed907	361	vblk->disk->first_minor = index_to_minor(index);
e467cde2 RR	362	vblk->disk->private_data = vblk;
e467cde2 RR	363	vblk->disk->fops = &virtblk_fops;
c4839346	364	vblk->disk->driverfs_dev = &vdev->dev;
d50ed907	365	index++;
4f3bf19c	366
e467cde2	367	/* If barriers are supported, tell block layer that queue is ordered */
f1b0ef06	368	if (virtio_has_feature(vdev, VIRTIO_BLK_F_FLUSH))
00fff265	369	blk_queue_ordered(q, QUEUE_ORDERED_DRAIN_FLUSH);
f1b0ef06	370	else if (virtio_has_feature(vdev, VIRTIO_BLK_F_BARRIER))
00fff265	371	blk_queue_ordered(q, QUEUE_ORDERED_TAG);
e467cde2	372
3ef53609 CB	373	/* If disk is read-only in the host, the guest should obey */
	374	if (virtio_has_feature(vdev, VIRTIO_BLK_F_RO))
	375	set_disk_ro(vblk->disk, 1);
	376
a586d4f6	377	/* Host must always specify the capacity. */
72e61eb4 RR	378	vdev->config->get(vdev, offsetof(struct virtio_blk_config, capacity),
72e61eb4 RR	379	&cap, sizeof(cap));
e467cde2 RR	380
	381	/* If capacity is too big, truncate with warning. */
	382	if ((sector_t)cap != cap) {
	383	dev_warn(&vdev->dev, "Capacity %llu too large: truncating\n",
	384	(unsigned long long)cap);
	385	cap = (sector_t)-1;
	386	}
	387	set_capacity(vblk->disk, cap);
	388
0864b79a	389	/* We can handle whatever the host told us to handle. */
ee714f2d	390	blk_queue_max_segments(q, vblk->sg_elems-2);
0864b79a	391
4eff3cae	392	/* No need to bounce any requests */
69740c8b	393	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
4eff3cae	394
4b7f7e20	395	/* No real sector limit. */
ee714f2d	396	blk_queue_max_hw_sectors(q, -1U);
4b7f7e20	397
a586d4f6 RR	398	/* Host can optionally specify maximum segment size and number of
	399	* segments. */
	400	err = virtio_config_val(vdev, VIRTIO_BLK_F_SIZE_MAX,
	401	offsetof(struct virtio_blk_config, size_max),
	402	&v);
e467cde2	403	if (!err)
69740c8b	404	blk_queue_max_segment_size(q, v);
4b7f7e20	405	else
69740c8b	406	blk_queue_max_segment_size(q, -1U);
e467cde2	407
066f4d82 CB	408	/* Host can optionally specify the block size of the device */
	409	err = virtio_config_val(vdev, VIRTIO_BLK_F_BLK_SIZE,
	410	offsetof(struct virtio_blk_config, blk_size),
	411	&blk_size);
	412	if (!err)
69740c8b CH	413	blk_queue_logical_block_size(q, blk_size);
	414	else
	415	blk_size = queue_logical_block_size(q);
	416
	417	/* Use topology information if available */
	418	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	419	offsetof(struct virtio_blk_config, physical_block_exp),
	420	&physical_block_exp);
	421	if (!err && physical_block_exp)
	422	blk_queue_physical_block_size(q,
	423	blk_size * (1 << physical_block_exp));
	424
	425	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	426	offsetof(struct virtio_blk_config, alignment_offset),
	427	&alignment_offset);
	428	if (!err && alignment_offset)
	429	blk_queue_alignment_offset(q, blk_size * alignment_offset);
	430
	431	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	432	offsetof(struct virtio_blk_config, min_io_size),
	433	&min_io_size);
	434	if (!err && min_io_size)
	435	blk_queue_io_min(q, blk_size * min_io_size);
	436
	437	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	438	offsetof(struct virtio_blk_config, opt_io_size),
	439	&opt_io_size);
	440	if (!err && opt_io_size)
	441	blk_queue_io_opt(q, blk_size * opt_io_size);
	442
066f4d82	443
e467cde2 RR	444	add_disk(vblk->disk);
	445	return 0;
	446
	447	out_put_disk:
	448	put_disk(vblk->disk);
e467cde2 RR	449	out_mempool:
	450	mempool_destroy(vblk->pool);
	451	out_free_vq:
d2a7ddda	452	vdev->config->del_vqs(vdev);
e467cde2 RR	453	out_free_vblk:
	454	kfree(vblk);
	455	out:
	456	return err;
	457	}
	458
98e94444	459	static void __devexit virtblk_remove(struct virtio_device *vdev)
e467cde2 RR	460	{
e467cde2 RR	461	struct virtio_blk *vblk = vdev->priv;
e467cde2	462
6e5aa7ef	463	/* Nothing should be pending. */
e467cde2	464	BUG_ON(!list_empty(&vblk->reqs));
6e5aa7ef RR	465
	466	/* Stop all the virtqueues. */
	467	vdev->config->reset(vdev);
	468
ac9d463a	469	del_gendisk(vblk->disk);
e467cde2 RR	470	blk_cleanup_queue(vblk->disk->queue);
e467cde2 RR	471	put_disk(vblk->disk);
e467cde2	472	mempool_destroy(vblk->pool);
d2a7ddda	473	vdev->config->del_vqs(vdev);
e467cde2 RR	474	kfree(vblk);
	475	}
	476
47483e25	477	static const struct virtio_device_id id_table[] = {
e467cde2 RR	478	{ VIRTIO_ID_BLOCK, VIRTIO_DEV_ANY_ID },
	479	{ 0 },
	480	};
	481
c45a6816 RR	482	static unsigned int features[] = {
c45a6816 RR	483	VIRTIO_BLK_F_BARRIER, VIRTIO_BLK_F_SEG_MAX, VIRTIO_BLK_F_SIZE_MAX,
066f4d82	484	VIRTIO_BLK_F_GEOMETRY, VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
69740c8b	485	VIRTIO_BLK_F_SCSI, VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY
c45a6816 RR	486	};
c45a6816 RR	487
4fbfff76 RM	488	/*
	489	* virtio_blk causes spurious section mismatch warning by
	490	* simultaneously referring to a __devinit and a __devexit function.
	491	* Use __refdata to avoid this warning.
	492	*/
	493	static struct virtio_driver __refdata virtio_blk = {
c45a6816 RR	494	.feature_table = features,
c45a6816 RR	495	.feature_table_size = ARRAY_SIZE(features),
e467cde2 RR	496	.driver.name = KBUILD_MODNAME,
	497	.driver.owner = THIS_MODULE,
	498	.id_table = id_table,
	499	.probe = virtblk_probe,
	500	.remove = __devexit_p(virtblk_remove),
	501	};
	502
	503	static int __init init(void)
	504	{
4f3bf19c CB	505	major = register_blkdev(0, "virtblk");
	506	if (major < 0)
	507	return major;
e467cde2 RR	508	return register_virtio_driver(&virtio_blk);
	509	}
	510
	511	static void __exit fini(void)
	512	{
4f3bf19c	513	unregister_blkdev(major, "virtblk");
e467cde2 RR	514	unregister_virtio_driver(&virtio_blk);
	515	}
	516	module_init(init);
	517	module_exit(fini);
	518
	519	MODULE_DEVICE_TABLE(virtio, id_table);
	520	MODULE_DESCRIPTION("Virtio block driver");
	521	MODULE_LICENSE("GPL");