[deliverable/linux.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 = vblk->vq->vq_ops->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;
		}

		if (blk_pc_request(vbr->req)) {
			vbr->req->resid_len = vbr->in_hdr.residual;
			vbr->req->sense_len = vbr->in_hdr.sense_len;
			vbr->req->errors = vbr->in_hdr.errors;
		}

		__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;
	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_LINUX_BLOCK:
		if (req->cmd[0] == REQ_LB_OP_FLUSH) {
			vbr->out_hdr.type = VIRTIO_BLK_T_FLUSH;
			vbr->out_hdr.sector = 0;
			vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
			break;
		}
		/*FALLTHRU*/
	default:
		/* We don't put anything else in the queue. */
		BUG();
	}

	if (blk_barrier_rq(vbr->req))
		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 (blk_pc_request(vbr->req))
		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 (blk_pc_request(vbr->req)) {
		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 (vblk->vq->vq_ops->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)
		vblk->vq->vq_ops->kick(vblk->vq);
}

static void virtblk_prepare_flush(struct request_queue *q, struct request *req)
{
	req->cmd_type = REQ_TYPE_LINUX_BLOCK;
	req->cmd[0] = REQ_LB_OP_FLUSH;
}

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;

	/*
	 * 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);
	if (err)
		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,
				  virtblk_prepare_flush);
	else if (virtio_has_feature(vdev, VIRTIO_BLK_F_BARRIER))
		blk_queue_ordered(q, QUEUE_ORDERED_TAG, NULL);

	/* 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);
	53	while ((vbr = vblk->vq->vq_ops->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
1cde26f9 HR	68	if (blk_pc_request(vbr->req)) {
	69	vbr->req->resid_len = vbr->in_hdr.residual;
	70	vbr->req->sense_len = vbr->in_hdr.sense_len;
	71	vbr->req->errors = vbr->in_hdr.errors;
	72	}
	73
40cbbb78	74	__blk_end_request_all(vbr->req, error);
e467cde2 RR	75	list_del(&vbr->list);
	76	mempool_free(vbr, vblk->pool);
	77	}
	78	/* In case queue is stopped waiting for more buffers. */
	79	blk_start_queue(vblk->disk->queue);
	80	spin_unlock_irqrestore(&vblk->lock, flags);
e467cde2 RR	81	}
	82
	83	static bool do_req(struct request_queue q, struct virtio_blk vblk,
	84	struct request *req)
	85	{
1cde26f9	86	unsigned long num, out = 0, in = 0;
e467cde2 RR	87	struct virtblk_req *vbr;
	88
	89	vbr = mempool_alloc(vblk->pool, GFP_ATOMIC);
	90	if (!vbr)
	91	/* When another request finishes we'll try again. */
	92	return false;
	93
	94	vbr->req = req;
f1b0ef06 CH	95	switch (req->cmd_type) {
f1b0ef06 CH	96	case REQ_TYPE_FS:
e467cde2	97	vbr->out_hdr.type = 0;
83096ebf	98	vbr->out_hdr.sector = blk_rq_pos(vbr->req);
766ca442	99	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
f1b0ef06 CH	100	break;
f1b0ef06 CH	101	case REQ_TYPE_BLOCK_PC:
e467cde2 RR	102	vbr->out_hdr.type = VIRTIO_BLK_T_SCSI_CMD;
e467cde2 RR	103	vbr->out_hdr.sector = 0;
766ca442	104	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
f1b0ef06 CH	105	break;
	106	case REQ_TYPE_LINUX_BLOCK:
	107	if (req->cmd[0] == REQ_LB_OP_FLUSH) {
	108	vbr->out_hdr.type = VIRTIO_BLK_T_FLUSH;
	109	vbr->out_hdr.sector = 0;
	110	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	111	break;
	112	}
	113	/FALLTHRU/
	114	default:
e467cde2 RR	115	/* We don't put anything else in the queue. */
	116	BUG();
	117	}
	118
	119	if (blk_barrier_rq(vbr->req))
	120	vbr->out_hdr.type \|= VIRTIO_BLK_T_BARRIER;
	121
1cde26f9	122	sg_set_buf(&vblk->sg[out++], &vbr->out_hdr, sizeof(vbr->out_hdr));
e467cde2	123
1cde26f9 HR	124	/*
	125	* If this is a packet command we need a couple of additional headers.
	126	* Behind the normal outhdr we put a segment with the scsi command
	127	* block, and before the normal inhdr we put the sense data and the
	128	* inhdr with additional status information before the normal inhdr.
	129	*/
	130	if (blk_pc_request(vbr->req))
	131	sg_set_buf(&vblk->sg[out++], vbr->req->cmd, vbr->req->cmd_len);
	132
	133	num = blk_rq_map_sg(q, vbr->req, vblk->sg + out);
	134
	135	if (blk_pc_request(vbr->req)) {
	136	sg_set_buf(&vblk->sg[num + out + in++], vbr->req->sense, 96);
	137	sg_set_buf(&vblk->sg[num + out + in++], &vbr->in_hdr,
	138	sizeof(vbr->in_hdr));
	139	}
	140
	141	sg_set_buf(&vblk->sg[num + out + in++], &vbr->status,
	142	sizeof(vbr->status));
	143
	144	if (num) {
	145	if (rq_data_dir(vbr->req) == WRITE) {
	146	vbr->out_hdr.type \|= VIRTIO_BLK_T_OUT;
	147	out += num;
	148	} else {
	149	vbr->out_hdr.type \|= VIRTIO_BLK_T_IN;
	150	in += num;
	151	}
e467cde2 RR	152	}
e467cde2 RR	153
3c1b27d5	154	if (vblk->vq->vq_ops->add_buf(vblk->vq, vblk->sg, out, in, vbr) < 0) {
e467cde2 RR	155	mempool_free(vbr, vblk->pool);
	156	return false;
	157	}
	158
	159	list_add_tail(&vbr->list, &vblk->reqs);
	160	return true;
	161	}
	162
	163	static void do_virtblk_request(struct request_queue *q)
	164	{
6c3b46f7	165	struct virtio_blk *vblk = q->queuedata;
e467cde2 RR	166	struct request *req;
	167	unsigned int issued = 0;
	168
9934c8c0	169	while ((req = blk_peek_request(q)) != NULL) {
0864b79a	170	BUG_ON(req->nr_phys_segments + 2 > vblk->sg_elems);
e467cde2 RR	171
	172	/* If this request fails, stop queue and wait for something to
	173	finish to restart it. */
	174	if (!do_req(q, vblk, req)) {
	175	blk_stop_queue(q);
	176	break;
	177	}
9934c8c0	178	blk_start_request(req);
e467cde2 RR	179	issued++;
	180	}
	181
	182	if (issued)
	183	vblk->vq->vq_ops->kick(vblk->vq);
	184	}
	185
f1b0ef06 CH	186	static void virtblk_prepare_flush(struct request_queue q, struct request req)
	187	{
	188	req->cmd_type = REQ_TYPE_LINUX_BLOCK;
	189	req->cmd[0] = REQ_LB_OP_FLUSH;
	190	}
	191
4e109852	192	static int virtblk_ioctl(struct block_device *bdev, fmode_t mode,
e467cde2 RR	193	unsigned cmd, unsigned long data)
e467cde2 RR	194	{
1cde26f9 HR	195	struct gendisk *disk = bdev->bd_disk;
	196	struct virtio_blk *vblk = disk->private_data;
	197
	198	/*
	199	* Only allow the generic SCSI ioctls if the host can support it.
	200	*/
	201	if (!virtio_has_feature(vblk->vdev, VIRTIO_BLK_F_SCSI))
d9ecdea7	202	return -ENOTTY;
1cde26f9	203
3225beab RR	204	return scsi_cmd_ioctl(disk->queue, disk, mode, cmd,
3225beab RR	205	(void __user *)data);
e467cde2 RR	206	}
e467cde2 RR	207
135da0b0 CB	208	/* We provide getgeo only to please some old bootloader/partitioning tools */
	209	static int virtblk_getgeo(struct block_device bd, struct hd_geometry geo)
	210	{
48e4043d RH	211	struct virtio_blk *vblk = bd->bd_disk->private_data;
	212	struct virtio_blk_geometry vgeo;
	213	int err;
	214
	215	/* see if the host passed in geometry config */
	216	err = virtio_config_val(vblk->vdev, VIRTIO_BLK_F_GEOMETRY,
	217	offsetof(struct virtio_blk_config, geometry),
	218	&vgeo);
	219
	220	if (!err) {
	221	geo->heads = vgeo.heads;
	222	geo->sectors = vgeo.sectors;
	223	geo->cylinders = vgeo.cylinders;
	224	} else {
	225	/* some standard values, similar to sd */
	226	geo->heads = 1 << 6;
	227	geo->sectors = 1 << 5;
	228	geo->cylinders = get_capacity(bd->bd_disk) >> 11;
	229	}
135da0b0 CB	230	return 0;
	231	}
	232
83d5cde4	233	static const struct block_device_operations virtblk_fops = {
4e109852	234	.locked_ioctl = virtblk_ioctl,
135da0b0 CB	235	.owner = THIS_MODULE,
135da0b0 CB	236	.getgeo = virtblk_getgeo,
e467cde2 RR	237	};
e467cde2 RR	238
d50ed907 CB	239	static int index_to_minor(int index)
	240	{
	241	return index << PART_BITS;
	242	}
	243
98e94444	244	static int __devinit virtblk_probe(struct virtio_device *vdev)
e467cde2 RR	245	{
e467cde2 RR	246	struct virtio_blk *vblk;
69740c8b	247	struct request_queue *q;
4f3bf19c	248	int err;
e467cde2	249	u64 cap;
69740c8b CH	250	u32 v, blk_size, sg_elems, opt_io_size;
	251	u16 min_io_size;
	252	u8 physical_block_exp, alignment_offset;
e467cde2	253
d50ed907	254	if (index_to_minor(index) >= 1 << MINORBITS)
4f3bf19c CB	255	return -ENOSPC;
4f3bf19c CB	256
0864b79a RR	257	/* We need to know how many segments before we allocate. */
	258	err = virtio_config_val(vdev, VIRTIO_BLK_F_SEG_MAX,
	259	offsetof(struct virtio_blk_config, seg_max),
	260	&sg_elems);
	261	if (err)
	262	sg_elems = 1;
	263
	264	/* We need an extra sg elements at head and tail. */
	265	sg_elems += 2;
	266	vdev->priv = vblk = kmalloc(sizeof(*vblk) +
	267	sizeof(vblk->sg[0]) * sg_elems, GFP_KERNEL);
e467cde2 RR	268	if (!vblk) {
	269	err = -ENOMEM;
	270	goto out;
	271	}
	272
	273	INIT_LIST_HEAD(&vblk->reqs);
	274	spin_lock_init(&vblk->lock);
	275	vblk->vdev = vdev;
0864b79a RR	276	vblk->sg_elems = sg_elems;
0864b79a RR	277	sg_init_table(vblk->sg, vblk->sg_elems);
e467cde2 RR	278
e467cde2 RR	279	/* We expect one virtqueue, for output. */
d2a7ddda	280	vblk->vq = virtio_find_single_vq(vdev, blk_done, "requests");
e467cde2 RR	281	if (IS_ERR(vblk->vq)) {
	282	err = PTR_ERR(vblk->vq);
	283	goto out_free_vblk;
	284	}
	285
	286	vblk->pool = mempool_create_kmalloc_pool(1,sizeof(struct virtblk_req));
	287	if (!vblk->pool) {
	288	err = -ENOMEM;
	289	goto out_free_vq;
	290	}
	291
e467cde2	292	/* FIXME: How many partitions? How long is a piece of string? */
4f3bf19c	293	vblk->disk = alloc_disk(1 << PART_BITS);
e467cde2 RR	294	if (!vblk->disk) {
e467cde2 RR	295	err = -ENOMEM;
4f3bf19c	296	goto out_mempool;
e467cde2 RR	297	}
e467cde2 RR	298
69740c8b CH	299	q = vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
69740c8b CH	300	if (!q) {
e467cde2 RR	301	err = -ENOMEM;
	302	goto out_put_disk;
	303	}
	304
69740c8b	305	q->queuedata = vblk;
7d116b62	306
d50ed907 CB	307	if (index < 26) {
	308	sprintf(vblk->disk->disk_name, "vd%c", 'a' + index % 26);
	309	} else if (index < (26 + 1) * 26) {
	310	sprintf(vblk->disk->disk_name, "vd%c%c",
	311	'a' + index / 26 - 1, 'a' + index % 26);
	312	} else {
	313	const unsigned int m1 = (index / 26 - 1) / 26 - 1;
	314	const unsigned int m2 = (index / 26 - 1) % 26;
	315	const unsigned int m3 = index % 26;
	316	sprintf(vblk->disk->disk_name, "vd%c%c%c",
	317	'a' + m1, 'a' + m2, 'a' + m3);
	318	}
	319
e467cde2	320	vblk->disk->major = major;
d50ed907	321	vblk->disk->first_minor = index_to_minor(index);
e467cde2 RR	322	vblk->disk->private_data = vblk;
e467cde2 RR	323	vblk->disk->fops = &virtblk_fops;
c4839346	324	vblk->disk->driverfs_dev = &vdev->dev;
d50ed907	325	index++;
4f3bf19c	326
e467cde2	327	/* If barriers are supported, tell block layer that queue is ordered */
f1b0ef06	328	if (virtio_has_feature(vdev, VIRTIO_BLK_F_FLUSH))
69740c8b	329	blk_queue_ordered(q, QUEUE_ORDERED_DRAIN_FLUSH,
f1b0ef06 CH	330	virtblk_prepare_flush);
f1b0ef06 CH	331	else if (virtio_has_feature(vdev, VIRTIO_BLK_F_BARRIER))
69740c8b	332	blk_queue_ordered(q, QUEUE_ORDERED_TAG, NULL);
e467cde2	333
3ef53609 CB	334	/* If disk is read-only in the host, the guest should obey */
	335	if (virtio_has_feature(vdev, VIRTIO_BLK_F_RO))
	336	set_disk_ro(vblk->disk, 1);
	337
a586d4f6	338	/* Host must always specify the capacity. */
72e61eb4 RR	339	vdev->config->get(vdev, offsetof(struct virtio_blk_config, capacity),
72e61eb4 RR	340	&cap, sizeof(cap));
e467cde2 RR	341
	342	/* If capacity is too big, truncate with warning. */
	343	if ((sector_t)cap != cap) {
	344	dev_warn(&vdev->dev, "Capacity %llu too large: truncating\n",
	345	(unsigned long long)cap);
	346	cap = (sector_t)-1;
	347	}
	348	set_capacity(vblk->disk, cap);
	349
0864b79a	350	/* We can handle whatever the host told us to handle. */
ee714f2d	351	blk_queue_max_segments(q, vblk->sg_elems-2);
0864b79a	352
4eff3cae	353	/* No need to bounce any requests */
69740c8b	354	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
4eff3cae	355
4b7f7e20	356	/* No real sector limit. */
ee714f2d	357	blk_queue_max_hw_sectors(q, -1U);
4b7f7e20	358
a586d4f6 RR	359	/* Host can optionally specify maximum segment size and number of
	360	* segments. */
	361	err = virtio_config_val(vdev, VIRTIO_BLK_F_SIZE_MAX,
	362	offsetof(struct virtio_blk_config, size_max),
	363	&v);
e467cde2	364	if (!err)
69740c8b	365	blk_queue_max_segment_size(q, v);
4b7f7e20	366	else
69740c8b	367	blk_queue_max_segment_size(q, -1U);
e467cde2	368
066f4d82 CB	369	/* Host can optionally specify the block size of the device */
	370	err = virtio_config_val(vdev, VIRTIO_BLK_F_BLK_SIZE,
	371	offsetof(struct virtio_blk_config, blk_size),
	372	&blk_size);
	373	if (!err)
69740c8b CH	374	blk_queue_logical_block_size(q, blk_size);
	375	else
	376	blk_size = queue_logical_block_size(q);
	377
	378	/* Use topology information if available */
	379	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	380	offsetof(struct virtio_blk_config, physical_block_exp),
	381	&physical_block_exp);
	382	if (!err && physical_block_exp)
	383	blk_queue_physical_block_size(q,
	384	blk_size * (1 << physical_block_exp));
	385
	386	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	387	offsetof(struct virtio_blk_config, alignment_offset),
	388	&alignment_offset);
	389	if (!err && alignment_offset)
	390	blk_queue_alignment_offset(q, blk_size * alignment_offset);
	391
	392	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	393	offsetof(struct virtio_blk_config, min_io_size),
	394	&min_io_size);
	395	if (!err && min_io_size)
	396	blk_queue_io_min(q, blk_size * min_io_size);
	397
	398	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	399	offsetof(struct virtio_blk_config, opt_io_size),
	400	&opt_io_size);
	401	if (!err && opt_io_size)
	402	blk_queue_io_opt(q, blk_size * opt_io_size);
	403
066f4d82	404
e467cde2 RR	405	add_disk(vblk->disk);
	406	return 0;
	407
	408	out_put_disk:
	409	put_disk(vblk->disk);
e467cde2 RR	410	out_mempool:
	411	mempool_destroy(vblk->pool);
	412	out_free_vq:
d2a7ddda	413	vdev->config->del_vqs(vdev);
e467cde2 RR	414	out_free_vblk:
	415	kfree(vblk);
	416	out:
	417	return err;
	418	}
	419
98e94444	420	static void __devexit virtblk_remove(struct virtio_device *vdev)
e467cde2 RR	421	{
e467cde2 RR	422	struct virtio_blk *vblk = vdev->priv;
e467cde2	423
6e5aa7ef	424	/* Nothing should be pending. */
e467cde2	425	BUG_ON(!list_empty(&vblk->reqs));
6e5aa7ef RR	426
	427	/* Stop all the virtqueues. */
	428	vdev->config->reset(vdev);
	429
ac9d463a	430	del_gendisk(vblk->disk);
e467cde2 RR	431	blk_cleanup_queue(vblk->disk->queue);
e467cde2 RR	432	put_disk(vblk->disk);
e467cde2	433	mempool_destroy(vblk->pool);
d2a7ddda	434	vdev->config->del_vqs(vdev);
e467cde2 RR	435	kfree(vblk);
	436	}
	437
47483e25	438	static const struct virtio_device_id id_table[] = {
e467cde2 RR	439	{ VIRTIO_ID_BLOCK, VIRTIO_DEV_ANY_ID },
	440	{ 0 },
	441	};
	442
c45a6816 RR	443	static unsigned int features[] = {
c45a6816 RR	444	VIRTIO_BLK_F_BARRIER, VIRTIO_BLK_F_SEG_MAX, VIRTIO_BLK_F_SIZE_MAX,
066f4d82	445	VIRTIO_BLK_F_GEOMETRY, VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
69740c8b	446	VIRTIO_BLK_F_SCSI, VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY
c45a6816 RR	447	};
c45a6816 RR	448
4fbfff76 RM	449	/*
	450	* virtio_blk causes spurious section mismatch warning by
	451	* simultaneously referring to a __devinit and a __devexit function.
	452	* Use __refdata to avoid this warning.
	453	*/
	454	static struct virtio_driver __refdata virtio_blk = {
c45a6816 RR	455	.feature_table = features,
c45a6816 RR	456	.feature_table_size = ARRAY_SIZE(features),
e467cde2 RR	457	.driver.name = KBUILD_MODNAME,
	458	.driver.owner = THIS_MODULE,
	459	.id_table = id_table,
	460	.probe = virtblk_probe,
	461	.remove = __devexit_p(virtblk_remove),
	462	};
	463
	464	static int __init init(void)
	465	{
4f3bf19c CB	466	major = register_blkdev(0, "virtblk");
	467	if (major < 0)
	468	return major;
e467cde2 RR	469	return register_virtio_driver(&virtio_blk);
	470	}
	471
	472	static void __exit fini(void)
	473	{
4f3bf19c	474	unregister_blkdev(major, "virtblk");
e467cde2 RR	475	unregister_virtio_driver(&virtio_blk);
	476	}
	477	module_init(init);
	478	module_exit(fini);
	479
	480	MODULE_DEVICE_TABLE(virtio, id_table);
	481	MODULE_DESCRIPTION("Virtio block driver");
	482	MODULE_LICENSE("GPL");