[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;
		}

		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;
		}
		if (blk_special_request(vbr->req))
			vbr->req->errors = (error != 0);

		__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_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;
	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 (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);
}

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;
}

/* 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,
				  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);
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
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	}
4cb2ea28	73	if (blk_special_request(vbr->req))
4cb2ea28	74	vbr->req->errors = (error != 0);
1cde26f9	75
40cbbb78	76	__blk_end_request_all(vbr->req, error);
e467cde2 RR	77	list_del(&vbr->list);
	78	mempool_free(vbr, vblk->pool);
	79	}
	80	/* In case queue is stopped waiting for more buffers. */
	81	blk_start_queue(vblk->disk->queue);
	82	spin_unlock_irqrestore(&vblk->lock, flags);
e467cde2 RR	83	}
	84
	85	static bool do_req(struct request_queue q, struct virtio_blk vblk,
	86	struct request *req)
	87	{
1cde26f9	88	unsigned long num, out = 0, in = 0;
e467cde2 RR	89	struct virtblk_req *vbr;
	90
	91	vbr = mempool_alloc(vblk->pool, GFP_ATOMIC);
	92	if (!vbr)
	93	/* When another request finishes we'll try again. */
	94	return false;
	95
	96	vbr->req = req;
f1b0ef06 CH	97	switch (req->cmd_type) {
f1b0ef06 CH	98	case REQ_TYPE_FS:
e467cde2	99	vbr->out_hdr.type = 0;
83096ebf	100	vbr->out_hdr.sector = blk_rq_pos(vbr->req);
766ca442	101	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
f1b0ef06 CH	102	break;
f1b0ef06 CH	103	case REQ_TYPE_BLOCK_PC:
e467cde2 RR	104	vbr->out_hdr.type = VIRTIO_BLK_T_SCSI_CMD;
e467cde2 RR	105	vbr->out_hdr.sector = 0;
766ca442	106	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
f1b0ef06	107	break;
4cb2ea28	108	case REQ_TYPE_SPECIAL:
	109	vbr->out_hdr.type = VIRTIO_BLK_T_GET_ID;
	110	vbr->out_hdr.sector = 0;
	111	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	112	break;
f1b0ef06 CH	113	case REQ_TYPE_LINUX_BLOCK:
	114	if (req->cmd[0] == REQ_LB_OP_FLUSH) {
	115	vbr->out_hdr.type = VIRTIO_BLK_T_FLUSH;
	116	vbr->out_hdr.sector = 0;
	117	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	118	break;
	119	}
	120	/FALLTHRU/
	121	default:
e467cde2 RR	122	/* We don't put anything else in the queue. */
	123	BUG();
	124	}
	125
	126	if (blk_barrier_rq(vbr->req))
	127	vbr->out_hdr.type \|= VIRTIO_BLK_T_BARRIER;
	128
1cde26f9	129	sg_set_buf(&vblk->sg[out++], &vbr->out_hdr, sizeof(vbr->out_hdr));
e467cde2	130
1cde26f9 HR	131	/*
	132	* If this is a packet command we need a couple of additional headers.
	133	* Behind the normal outhdr we put a segment with the scsi command
	134	* block, and before the normal inhdr we put the sense data and the
	135	* inhdr with additional status information before the normal inhdr.
	136	*/
	137	if (blk_pc_request(vbr->req))
	138	sg_set_buf(&vblk->sg[out++], vbr->req->cmd, vbr->req->cmd_len);
	139
	140	num = blk_rq_map_sg(q, vbr->req, vblk->sg + out);
	141
	142	if (blk_pc_request(vbr->req)) {
	143	sg_set_buf(&vblk->sg[num + out + in++], vbr->req->sense, 96);
	144	sg_set_buf(&vblk->sg[num + out + in++], &vbr->in_hdr,
	145	sizeof(vbr->in_hdr));
	146	}
	147
	148	sg_set_buf(&vblk->sg[num + out + in++], &vbr->status,
	149	sizeof(vbr->status));
	150
	151	if (num) {
	152	if (rq_data_dir(vbr->req) == WRITE) {
	153	vbr->out_hdr.type \|= VIRTIO_BLK_T_OUT;
	154	out += num;
	155	} else {
	156	vbr->out_hdr.type \|= VIRTIO_BLK_T_IN;
	157	in += num;
	158	}
e467cde2 RR	159	}
e467cde2 RR	160
09ec6b69	161	if (virtqueue_add_buf(vblk->vq, vblk->sg, out, in, vbr) < 0) {
e467cde2 RR	162	mempool_free(vbr, vblk->pool);
	163	return false;
	164	}
	165
	166	list_add_tail(&vbr->list, &vblk->reqs);
	167	return true;
	168	}
	169
	170	static void do_virtblk_request(struct request_queue *q)
	171	{
6c3b46f7	172	struct virtio_blk *vblk = q->queuedata;
e467cde2 RR	173	struct request *req;
	174	unsigned int issued = 0;
	175
9934c8c0	176	while ((req = blk_peek_request(q)) != NULL) {
0864b79a	177	BUG_ON(req->nr_phys_segments + 2 > vblk->sg_elems);
e467cde2 RR	178
	179	/* If this request fails, stop queue and wait for something to
	180	finish to restart it. */
	181	if (!do_req(q, vblk, req)) {
	182	blk_stop_queue(q);
	183	break;
	184	}
9934c8c0	185	blk_start_request(req);
e467cde2 RR	186	issued++;
	187	}
	188
	189	if (issued)
09ec6b69	190	virtqueue_kick(vblk->vq);
e467cde2 RR	191	}
e467cde2 RR	192
f1b0ef06 CH	193	static void virtblk_prepare_flush(struct request_queue q, struct request req)
	194	{
	195	req->cmd_type = REQ_TYPE_LINUX_BLOCK;
	196	req->cmd[0] = REQ_LB_OP_FLUSH;
	197	}
	198
4cb2ea28	199	/* return id (s/n) string for disk to id_str
	200	*/
	201	static int virtblk_get_id(struct gendisk disk, char id_str)
	202	{
	203	struct virtio_blk *vblk = disk->private_data;
	204	struct request *req;
	205	struct bio *bio;
	206
	207	bio = bio_map_kern(vblk->disk->queue, id_str, VIRTIO_BLK_ID_BYTES,
	208	GFP_KERNEL);
	209	if (IS_ERR(bio))
	210	return PTR_ERR(bio);
	211
	212	req = blk_make_request(vblk->disk->queue, bio, GFP_KERNEL);
	213	if (IS_ERR(req)) {
	214	bio_put(bio);
	215	return PTR_ERR(req);
	216	}
	217
	218	req->cmd_type = REQ_TYPE_SPECIAL;
	219	return blk_execute_rq(vblk->disk->queue, vblk->disk, req, false);
	220	}
	221
4e109852	222	static int virtblk_ioctl(struct block_device *bdev, fmode_t mode,
e467cde2 RR	223	unsigned cmd, unsigned long data)
e467cde2 RR	224	{
1cde26f9 HR	225	struct gendisk *disk = bdev->bd_disk;
	226	struct virtio_blk *vblk = disk->private_data;
	227
bdb4a130	228	if (cmd == 0x56424944) { /* 'VBID' */
234f2725	229	void __user usr_data = (void __user )data;
	230	char id_str[VIRTIO_BLK_ID_BYTES];
	231	int err;
	232
	233	err = virtblk_get_id(disk, id_str);
	234	if (!err && copy_to_user(usr_data, id_str, VIRTIO_BLK_ID_BYTES))
	235	err = -EFAULT;
	236	return err;
	237	}
1cde26f9 HR	238	/*
	239	* Only allow the generic SCSI ioctls if the host can support it.
	240	*/
	241	if (!virtio_has_feature(vblk->vdev, VIRTIO_BLK_F_SCSI))
d9ecdea7	242	return -ENOTTY;
1cde26f9	243
3225beab RR	244	return scsi_cmd_ioctl(disk->queue, disk, mode, cmd,
3225beab RR	245	(void __user *)data);
e467cde2 RR	246	}
e467cde2 RR	247
135da0b0 CB	248	/* We provide getgeo only to please some old bootloader/partitioning tools */
	249	static int virtblk_getgeo(struct block_device bd, struct hd_geometry geo)
	250	{
48e4043d RH	251	struct virtio_blk *vblk = bd->bd_disk->private_data;
	252	struct virtio_blk_geometry vgeo;
	253	int err;
	254
	255	/* see if the host passed in geometry config */
	256	err = virtio_config_val(vblk->vdev, VIRTIO_BLK_F_GEOMETRY,
	257	offsetof(struct virtio_blk_config, geometry),
	258	&vgeo);
	259
	260	if (!err) {
	261	geo->heads = vgeo.heads;
	262	geo->sectors = vgeo.sectors;
	263	geo->cylinders = vgeo.cylinders;
	264	} else {
	265	/* some standard values, similar to sd */
	266	geo->heads = 1 << 6;
	267	geo->sectors = 1 << 5;
	268	geo->cylinders = get_capacity(bd->bd_disk) >> 11;
	269	}
135da0b0 CB	270	return 0;
	271	}
	272
83d5cde4	273	static const struct block_device_operations virtblk_fops = {
4e109852	274	.locked_ioctl = virtblk_ioctl,
135da0b0 CB	275	.owner = THIS_MODULE,
135da0b0 CB	276	.getgeo = virtblk_getgeo,
e467cde2 RR	277	};
e467cde2 RR	278
d50ed907 CB	279	static int index_to_minor(int index)
	280	{
	281	return index << PART_BITS;
	282	}
	283
98e94444	284	static int __devinit virtblk_probe(struct virtio_device *vdev)
e467cde2 RR	285	{
e467cde2 RR	286	struct virtio_blk *vblk;
69740c8b	287	struct request_queue *q;
4f3bf19c	288	int err;
e467cde2	289	u64 cap;
69740c8b CH	290	u32 v, blk_size, sg_elems, opt_io_size;
	291	u16 min_io_size;
	292	u8 physical_block_exp, alignment_offset;
e467cde2	293
d50ed907	294	if (index_to_minor(index) >= 1 << MINORBITS)
4f3bf19c CB	295	return -ENOSPC;
4f3bf19c CB	296
0864b79a RR	297	/* We need to know how many segments before we allocate. */
	298	err = virtio_config_val(vdev, VIRTIO_BLK_F_SEG_MAX,
	299	offsetof(struct virtio_blk_config, seg_max),
	300	&sg_elems);
a5b365a6 CH	301
	302	/* We need at least one SG element, whatever they say. */
	303	if (err \|\| !sg_elems)
0864b79a RR	304	sg_elems = 1;
	305
	306	/* We need an extra sg elements at head and tail. */
	307	sg_elems += 2;
	308	vdev->priv = vblk = kmalloc(sizeof(*vblk) +
	309	sizeof(vblk->sg[0]) * sg_elems, GFP_KERNEL);
e467cde2 RR	310	if (!vblk) {
	311	err = -ENOMEM;
	312	goto out;
	313	}
	314
	315	INIT_LIST_HEAD(&vblk->reqs);
	316	spin_lock_init(&vblk->lock);
	317	vblk->vdev = vdev;
0864b79a RR	318	vblk->sg_elems = sg_elems;
0864b79a RR	319	sg_init_table(vblk->sg, vblk->sg_elems);
e467cde2 RR	320
e467cde2 RR	321	/* We expect one virtqueue, for output. */
d2a7ddda	322	vblk->vq = virtio_find_single_vq(vdev, blk_done, "requests");
e467cde2 RR	323	if (IS_ERR(vblk->vq)) {
	324	err = PTR_ERR(vblk->vq);
	325	goto out_free_vblk;
	326	}
	327
	328	vblk->pool = mempool_create_kmalloc_pool(1,sizeof(struct virtblk_req));
	329	if (!vblk->pool) {
	330	err = -ENOMEM;
	331	goto out_free_vq;
	332	}
	333
e467cde2	334	/* FIXME: How many partitions? How long is a piece of string? */
4f3bf19c	335	vblk->disk = alloc_disk(1 << PART_BITS);
e467cde2 RR	336	if (!vblk->disk) {
e467cde2 RR	337	err = -ENOMEM;
4f3bf19c	338	goto out_mempool;
e467cde2 RR	339	}
e467cde2 RR	340
69740c8b CH	341	q = vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
69740c8b CH	342	if (!q) {
e467cde2 RR	343	err = -ENOMEM;
	344	goto out_put_disk;
	345	}
	346
69740c8b	347	q->queuedata = vblk;
7d116b62	348
d50ed907 CB	349	if (index < 26) {
	350	sprintf(vblk->disk->disk_name, "vd%c", 'a' + index % 26);
	351	} else if (index < (26 + 1) * 26) {
	352	sprintf(vblk->disk->disk_name, "vd%c%c",
	353	'a' + index / 26 - 1, 'a' + index % 26);
	354	} else {
	355	const unsigned int m1 = (index / 26 - 1) / 26 - 1;
	356	const unsigned int m2 = (index / 26 - 1) % 26;
	357	const unsigned int m3 = index % 26;
	358	sprintf(vblk->disk->disk_name, "vd%c%c%c",
	359	'a' + m1, 'a' + m2, 'a' + m3);
	360	}
	361
e467cde2	362	vblk->disk->major = major;
d50ed907	363	vblk->disk->first_minor = index_to_minor(index);
e467cde2 RR	364	vblk->disk->private_data = vblk;
e467cde2 RR	365	vblk->disk->fops = &virtblk_fops;
c4839346	366	vblk->disk->driverfs_dev = &vdev->dev;
d50ed907	367	index++;
4f3bf19c	368
e467cde2	369	/* If barriers are supported, tell block layer that queue is ordered */
f1b0ef06	370	if (virtio_has_feature(vdev, VIRTIO_BLK_F_FLUSH))
69740c8b	371	blk_queue_ordered(q, QUEUE_ORDERED_DRAIN_FLUSH,
f1b0ef06 CH	372	virtblk_prepare_flush);
f1b0ef06 CH	373	else if (virtio_has_feature(vdev, VIRTIO_BLK_F_BARRIER))
69740c8b	374	blk_queue_ordered(q, QUEUE_ORDERED_TAG, NULL);
e467cde2	375
3ef53609 CB	376	/* If disk is read-only in the host, the guest should obey */
	377	if (virtio_has_feature(vdev, VIRTIO_BLK_F_RO))
	378	set_disk_ro(vblk->disk, 1);
	379
a586d4f6	380	/* Host must always specify the capacity. */
72e61eb4 RR	381	vdev->config->get(vdev, offsetof(struct virtio_blk_config, capacity),
72e61eb4 RR	382	&cap, sizeof(cap));
e467cde2 RR	383
	384	/* If capacity is too big, truncate with warning. */
	385	if ((sector_t)cap != cap) {
	386	dev_warn(&vdev->dev, "Capacity %llu too large: truncating\n",
	387	(unsigned long long)cap);
	388	cap = (sector_t)-1;
	389	}
	390	set_capacity(vblk->disk, cap);
	391
0864b79a	392	/* We can handle whatever the host told us to handle. */
ee714f2d	393	blk_queue_max_segments(q, vblk->sg_elems-2);
0864b79a	394
4eff3cae	395	/* No need to bounce any requests */
69740c8b	396	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
4eff3cae	397
4b7f7e20	398	/* No real sector limit. */
ee714f2d	399	blk_queue_max_hw_sectors(q, -1U);
4b7f7e20	400
a586d4f6 RR	401	/* Host can optionally specify maximum segment size and number of
	402	* segments. */
	403	err = virtio_config_val(vdev, VIRTIO_BLK_F_SIZE_MAX,
	404	offsetof(struct virtio_blk_config, size_max),
	405	&v);
e467cde2	406	if (!err)
69740c8b	407	blk_queue_max_segment_size(q, v);
4b7f7e20	408	else
69740c8b	409	blk_queue_max_segment_size(q, -1U);
e467cde2	410
066f4d82 CB	411	/* Host can optionally specify the block size of the device */
	412	err = virtio_config_val(vdev, VIRTIO_BLK_F_BLK_SIZE,
	413	offsetof(struct virtio_blk_config, blk_size),
	414	&blk_size);
	415	if (!err)
69740c8b CH	416	blk_queue_logical_block_size(q, blk_size);
	417	else
	418	blk_size = queue_logical_block_size(q);
	419
	420	/* Use topology information if available */
	421	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	422	offsetof(struct virtio_blk_config, physical_block_exp),
	423	&physical_block_exp);
	424	if (!err && physical_block_exp)
	425	blk_queue_physical_block_size(q,
	426	blk_size * (1 << physical_block_exp));
	427
	428	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	429	offsetof(struct virtio_blk_config, alignment_offset),
	430	&alignment_offset);
	431	if (!err && alignment_offset)
	432	blk_queue_alignment_offset(q, blk_size * alignment_offset);
	433
	434	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	435	offsetof(struct virtio_blk_config, min_io_size),
	436	&min_io_size);
	437	if (!err && min_io_size)
	438	blk_queue_io_min(q, blk_size * min_io_size);
	439
	440	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	441	offsetof(struct virtio_blk_config, opt_io_size),
	442	&opt_io_size);
	443	if (!err && opt_io_size)
	444	blk_queue_io_opt(q, blk_size * opt_io_size);
	445
066f4d82	446
e467cde2 RR	447	add_disk(vblk->disk);
	448	return 0;
	449
	450	out_put_disk:
	451	put_disk(vblk->disk);
e467cde2 RR	452	out_mempool:
	453	mempool_destroy(vblk->pool);
	454	out_free_vq:
d2a7ddda	455	vdev->config->del_vqs(vdev);
e467cde2 RR	456	out_free_vblk:
	457	kfree(vblk);
	458	out:
	459	return err;
	460	}
	461
98e94444	462	static void __devexit virtblk_remove(struct virtio_device *vdev)
e467cde2 RR	463	{
e467cde2 RR	464	struct virtio_blk *vblk = vdev->priv;
e467cde2	465
6e5aa7ef	466	/* Nothing should be pending. */
e467cde2	467	BUG_ON(!list_empty(&vblk->reqs));
6e5aa7ef RR	468
	469	/* Stop all the virtqueues. */
	470	vdev->config->reset(vdev);
	471
ac9d463a	472	del_gendisk(vblk->disk);
e467cde2 RR	473	blk_cleanup_queue(vblk->disk->queue);
e467cde2 RR	474	put_disk(vblk->disk);
e467cde2	475	mempool_destroy(vblk->pool);
d2a7ddda	476	vdev->config->del_vqs(vdev);
e467cde2 RR	477	kfree(vblk);
	478	}
	479
47483e25	480	static const struct virtio_device_id id_table[] = {
e467cde2 RR	481	{ VIRTIO_ID_BLOCK, VIRTIO_DEV_ANY_ID },
	482	{ 0 },
	483	};
	484
c45a6816 RR	485	static unsigned int features[] = {
c45a6816 RR	486	VIRTIO_BLK_F_BARRIER, VIRTIO_BLK_F_SEG_MAX, VIRTIO_BLK_F_SIZE_MAX,
066f4d82	487	VIRTIO_BLK_F_GEOMETRY, VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
69740c8b	488	VIRTIO_BLK_F_SCSI, VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY
c45a6816 RR	489	};
c45a6816 RR	490
4fbfff76 RM	491	/*
	492	* virtio_blk causes spurious section mismatch warning by
	493	* simultaneously referring to a __devinit and a __devexit function.
	494	* Use __refdata to avoid this warning.
	495	*/
	496	static struct virtio_driver __refdata virtio_blk = {
c45a6816 RR	497	.feature_table = features,
c45a6816 RR	498	.feature_table_size = ARRAY_SIZE(features),
e467cde2 RR	499	.driver.name = KBUILD_MODNAME,
	500	.driver.owner = THIS_MODULE,
	501	.id_table = id_table,
	502	.probe = virtblk_probe,
	503	.remove = __devexit_p(virtblk_remove),
	504	};
	505
	506	static int __init init(void)
	507	{
4f3bf19c CB	508	major = register_blkdev(0, "virtblk");
	509	if (major < 0)
	510	return major;
e467cde2 RR	511	return register_virtio_driver(&virtio_blk);
	512	}
	513
	514	static void __exit fini(void)
	515	{
4f3bf19c	516	unregister_blkdev(major, "virtblk");
e467cde2 RR	517	unregister_virtio_driver(&virtio_blk);
	518	}
	519	module_init(init);
	520	module_exit(fini);
	521
	522	MODULE_DEVICE_TABLE(virtio, id_table);
	523	MODULE_DESCRIPTION("Virtio block driver");
	524	MODULE_LICENSE("GPL");