[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/smp_lock.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_locked_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);
}

static int virtblk_ioctl(struct block_device *bdev, fmode_t mode,
			     unsigned int cmd, unsigned long param)
{
	int ret;

	lock_kernel();
	ret = virtblk_locked_ioctl(bdev, mode, cmd, param);
	unlock_kernel();

	return ret;
}

/* 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 = {
	.ioctl  = virtblk_ioctl,
	.owner  = THIS_MODULE,
	.getgeo = virtblk_getgeo,
};

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

static ssize_t virtblk_serial_show(struct device *dev,
				struct device_attribute *attr, char *buf)
{
	struct gendisk *disk = dev_to_disk(dev);
	int err;

	/* sysfs gives us a PAGE_SIZE buffer */
	BUILD_BUG_ON(PAGE_SIZE < VIRTIO_BLK_ID_BYTES);

	buf[VIRTIO_BLK_ID_BYTES] = '\0';
	err = virtblk_get_id(disk, buf);
	if (!err)
		return strlen(buf);

	if (err == -EIO) /* Unsupported? Make it empty. */
		return 0;

	return err;
}
DEVICE_ATTR(serial, S_IRUGO, virtblk_serial_show, NULL);

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 (virtio_has_feature(vdev, VIRTIO_BLK_F_FLUSH)) {
		/*
		 * If the FLUSH feature is supported we do have support for
		 * flushing a volatile write cache on the host.  Use that
		 * to implement write barrier support.
		 */
		blk_queue_ordered(q, QUEUE_ORDERED_DRAIN_FLUSH);
	} else if (virtio_has_feature(vdev, VIRTIO_BLK_F_BARRIER)) {
		/*
		 * If the BARRIER feature is supported the host expects us
		 * to order request by tags.  This implies there is not
		 * volatile write cache on the host, and that the host
		 * never re-orders outstanding I/O.  This feature is not
		 * useful for real life scenarious and deprecated.
		 */
		blk_queue_ordered(q, QUEUE_ORDERED_TAG);
	} else {
		/*
		 * If the FLUSH feature is not supported we must assume that
		 * the host does not perform any kind of volatile write
		 * caching. We still need to drain the queue to provider
		 * proper barrier semantics.
		 */
		blk_queue_ordered(q, QUEUE_ORDERED_DRAIN);
	}

	/* 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);
	err = device_create_file(disk_to_dev(vblk->disk), &dev_attr_serial);
	if (err)
		goto out_del_disk;

	return 0;

out_del_disk:
	del_gendisk(vblk->disk);
	blk_cleanup_queue(vblk->disk->queue);
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	4	#include <linux/blkdev.h>
8a6cfeb6	5	#include <linux/smp_lock.h>
e467cde2 RR	6	#include <linux/hdreg.h>
	7	#include <linux/virtio.h>
	8	#include <linux/virtio_blk.h>
3d1266c7 JA	9	#include <linux/scatterlist.h>
3d1266c7 JA	10
4f3bf19c	11	#define PART_BITS 4
e467cde2	12
d50ed907	13	static int major, index;
4f3bf19c	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
0864b79a RR	30	/* What host tells us, plus 2 for header & tailer. */
	31	unsigned int sg_elems;
	32
e467cde2	33	/* Scatterlist: can be too big for stack. */
0864b79a	34	struct scatterlist sg[/sg_elems/];
e467cde2 RR	35	};
	36
	37	struct virtblk_req
	38	{
	39	struct list_head list;
	40	struct request *req;
	41	struct virtio_blk_outhdr out_hdr;
1cde26f9	42	struct virtio_scsi_inhdr in_hdr;
cb38fa23	43	u8 status;
e467cde2 RR	44	};
e467cde2 RR	45
18445c4d	46	static void blk_done(struct virtqueue *vq)
e467cde2 RR	47	{
	48	struct virtio_blk *vblk = vq->vdev->priv;
	49	struct virtblk_req *vbr;
	50	unsigned int len;
	51	unsigned long flags;
	52
	53	spin_lock_irqsave(&vblk->lock, flags);
09ec6b69	54	while ((vbr = virtqueue_get_buf(vblk->vq, &len)) != NULL) {
8316982a	55	int error;
1cde26f9	56
cb38fa23	57	switch (vbr->status) {
e467cde2	58	case VIRTIO_BLK_S_OK:
8316982a	59	error = 0;
e467cde2 RR	60	break;
e467cde2 RR	61	case VIRTIO_BLK_S_UNSUPP:
8316982a	62	error = -ENOTTY;
e467cde2 RR	63	break;
e467cde2 RR	64	default:
8316982a	65	error = -EIO;
e467cde2 RR	66	break;
	67	}
	68
33659ebb CH	69	switch (vbr->req->cmd_type) {
33659ebb CH	70	case REQ_TYPE_BLOCK_PC:
1cde26f9 HR	71	vbr->req->resid_len = vbr->in_hdr.residual;
	72	vbr->req->sense_len = vbr->in_hdr.sense_len;
	73	vbr->req->errors = vbr->in_hdr.errors;
33659ebb CH	74	break;
33659ebb CH	75	case REQ_TYPE_SPECIAL:
4cb2ea28	76	vbr->req->errors = (error != 0);
33659ebb	77	break;
15fa6e81 JA	78	default:
15fa6e81 JA	79	break;
33659ebb	80	}
1cde26f9	81
40cbbb78	82	__blk_end_request_all(vbr->req, error);
e467cde2 RR	83	list_del(&vbr->list);
	84	mempool_free(vbr, vblk->pool);
	85	}
	86	/* In case queue is stopped waiting for more buffers. */
	87	blk_start_queue(vblk->disk->queue);
	88	spin_unlock_irqrestore(&vblk->lock, flags);
e467cde2 RR	89	}
	90
	91	static bool do_req(struct request_queue q, struct virtio_blk vblk,
	92	struct request *req)
	93	{
1cde26f9	94	unsigned long num, out = 0, in = 0;
e467cde2 RR	95	struct virtblk_req *vbr;
	96
	97	vbr = mempool_alloc(vblk->pool, GFP_ATOMIC);
	98	if (!vbr)
	99	/* When another request finishes we'll try again. */
	100	return false;
	101
	102	vbr->req = req;
dd40e456 FT	103
	104	if (req->cmd_flags & REQ_FLUSH) {
	105	vbr->out_hdr.type = VIRTIO_BLK_T_FLUSH;
4cb2ea28	106	vbr->out_hdr.sector = 0;
4cb2ea28	107	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
dd40e456 FT	108	} else {
	109	switch (req->cmd_type) {
	110	case REQ_TYPE_FS:
	111	vbr->out_hdr.type = 0;
	112	vbr->out_hdr.sector = blk_rq_pos(vbr->req);
	113	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	114	break;
	115	case REQ_TYPE_BLOCK_PC:
	116	vbr->out_hdr.type = VIRTIO_BLK_T_SCSI_CMD;
f1b0ef06 CH	117	vbr->out_hdr.sector = 0;
	118	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	119	break;
dd40e456 FT	120	case REQ_TYPE_SPECIAL:
	121	vbr->out_hdr.type = VIRTIO_BLK_T_GET_ID;
	122	vbr->out_hdr.sector = 0;
	123	vbr->out_hdr.ioprio = req_get_ioprio(vbr->req);
	124	break;
	125	default:
	126	/* We don't put anything else in the queue. */
	127	BUG();
f1b0ef06	128	}
e467cde2 RR	129	}
e467cde2 RR	130
33659ebb	131	if (vbr->req->cmd_flags & REQ_HARDBARRIER)
e467cde2 RR	132	vbr->out_hdr.type \|= VIRTIO_BLK_T_BARRIER;
e467cde2 RR	133
1cde26f9	134	sg_set_buf(&vblk->sg[out++], &vbr->out_hdr, sizeof(vbr->out_hdr));
e467cde2	135
1cde26f9 HR	136	/*
	137	* If this is a packet command we need a couple of additional headers.
	138	* Behind the normal outhdr we put a segment with the scsi command
	139	* block, and before the normal inhdr we put the sense data and the
	140	* inhdr with additional status information before the normal inhdr.
	141	*/
33659ebb	142	if (vbr->req->cmd_type == REQ_TYPE_BLOCK_PC)
1cde26f9 HR	143	sg_set_buf(&vblk->sg[out++], vbr->req->cmd, vbr->req->cmd_len);
	144
	145	num = blk_rq_map_sg(q, vbr->req, vblk->sg + out);
	146
33659ebb	147	if (vbr->req->cmd_type == REQ_TYPE_BLOCK_PC) {
1cde26f9 HR	148	sg_set_buf(&vblk->sg[num + out + in++], vbr->req->sense, 96);
	149	sg_set_buf(&vblk->sg[num + out + in++], &vbr->in_hdr,
	150	sizeof(vbr->in_hdr));
	151	}
	152
	153	sg_set_buf(&vblk->sg[num + out + in++], &vbr->status,
	154	sizeof(vbr->status));
	155
	156	if (num) {
	157	if (rq_data_dir(vbr->req) == WRITE) {
	158	vbr->out_hdr.type \|= VIRTIO_BLK_T_OUT;
	159	out += num;
	160	} else {
	161	vbr->out_hdr.type \|= VIRTIO_BLK_T_IN;
	162	in += num;
	163	}
e467cde2 RR	164	}
e467cde2 RR	165
09ec6b69	166	if (virtqueue_add_buf(vblk->vq, vblk->sg, out, in, vbr) < 0) {
e467cde2 RR	167	mempool_free(vbr, vblk->pool);
	168	return false;
	169	}
	170
	171	list_add_tail(&vbr->list, &vblk->reqs);
	172	return true;
	173	}
	174
	175	static void do_virtblk_request(struct request_queue *q)
	176	{
6c3b46f7	177	struct virtio_blk *vblk = q->queuedata;
e467cde2 RR	178	struct request *req;
	179	unsigned int issued = 0;
	180
9934c8c0	181	while ((req = blk_peek_request(q)) != NULL) {
0864b79a	182	BUG_ON(req->nr_phys_segments + 2 > vblk->sg_elems);
e467cde2 RR	183
	184	/* If this request fails, stop queue and wait for something to
	185	finish to restart it. */
	186	if (!do_req(q, vblk, req)) {
	187	blk_stop_queue(q);
	188	break;
	189	}
9934c8c0	190	blk_start_request(req);
e467cde2 RR	191	issued++;
	192	}
	193
	194	if (issued)
09ec6b69	195	virtqueue_kick(vblk->vq);
e467cde2 RR	196	}
e467cde2 RR	197
4cb2ea28	198	/* return id (s/n) string for disk to id_str
	199	*/
	200	static int virtblk_get_id(struct gendisk disk, char id_str)
	201	{
	202	struct virtio_blk *vblk = disk->private_data;
	203	struct request *req;
	204	struct bio *bio;
	205
	206	bio = bio_map_kern(vblk->disk->queue, id_str, VIRTIO_BLK_ID_BYTES,
	207	GFP_KERNEL);
	208	if (IS_ERR(bio))
	209	return PTR_ERR(bio);
	210
	211	req = blk_make_request(vblk->disk->queue, bio, GFP_KERNEL);
	212	if (IS_ERR(req)) {
	213	bio_put(bio);
	214	return PTR_ERR(req);
	215	}
	216
	217	req->cmd_type = REQ_TYPE_SPECIAL;
	218	return blk_execute_rq(vblk->disk->queue, vblk->disk, req, false);
	219	}
	220
8a6cfeb6	221	static int virtblk_locked_ioctl(struct block_device *bdev, fmode_t mode,
e467cde2 RR	222	unsigned cmd, unsigned long data)
e467cde2 RR	223	{
1cde26f9 HR	224	struct gendisk *disk = bdev->bd_disk;
	225	struct virtio_blk *vblk = disk->private_data;
	226
	227	/*
	228	* Only allow the generic SCSI ioctls if the host can support it.
	229	*/
	230	if (!virtio_has_feature(vblk->vdev, VIRTIO_BLK_F_SCSI))
d9ecdea7	231	return -ENOTTY;
1cde26f9	232
3225beab RR	233	return scsi_cmd_ioctl(disk->queue, disk, mode, cmd,
3225beab RR	234	(void __user *)data);
e467cde2 RR	235	}
e467cde2 RR	236
8a6cfeb6 AB	237	static int virtblk_ioctl(struct block_device *bdev, fmode_t mode,
	238	unsigned int cmd, unsigned long param)
	239	{
	240	int ret;
	241
	242	lock_kernel();
	243	ret = virtblk_locked_ioctl(bdev, mode, cmd, param);
	244	unlock_kernel();
	245
	246	return ret;
	247	}
	248
135da0b0 CB	249	/* We provide getgeo only to please some old bootloader/partitioning tools */
	250	static int virtblk_getgeo(struct block_device bd, struct hd_geometry geo)
	251	{
48e4043d RH	252	struct virtio_blk *vblk = bd->bd_disk->private_data;
	253	struct virtio_blk_geometry vgeo;
	254	int err;
	255
	256	/* see if the host passed in geometry config */
	257	err = virtio_config_val(vblk->vdev, VIRTIO_BLK_F_GEOMETRY,
	258	offsetof(struct virtio_blk_config, geometry),
	259	&vgeo);
	260
	261	if (!err) {
	262	geo->heads = vgeo.heads;
	263	geo->sectors = vgeo.sectors;
	264	geo->cylinders = vgeo.cylinders;
	265	} else {
	266	/* some standard values, similar to sd */
	267	geo->heads = 1 << 6;
	268	geo->sectors = 1 << 5;
	269	geo->cylinders = get_capacity(bd->bd_disk) >> 11;
	270	}
135da0b0 CB	271	return 0;
	272	}
	273
83d5cde4	274	static const struct block_device_operations virtblk_fops = {
8a6cfeb6	275	.ioctl = virtblk_ioctl,
135da0b0 CB	276	.owner = THIS_MODULE,
135da0b0 CB	277	.getgeo = virtblk_getgeo,
e467cde2 RR	278	};
e467cde2 RR	279
d50ed907 CB	280	static int index_to_minor(int index)
	281	{
	282	return index << PART_BITS;
	283	}
	284
a5eb9e4f RH	285	static ssize_t virtblk_serial_show(struct device *dev,
	286	struct device_attribute attr, char buf)
	287	{
	288	struct gendisk *disk = dev_to_disk(dev);
	289	int err;
	290
	291	/* sysfs gives us a PAGE_SIZE buffer */
	292	BUILD_BUG_ON(PAGE_SIZE < VIRTIO_BLK_ID_BYTES);
	293
	294	buf[VIRTIO_BLK_ID_BYTES] = '\0';
	295	err = virtblk_get_id(disk, buf);
	296	if (!err)
	297	return strlen(buf);
	298
	299	if (err == -EIO) /* Unsupported? Make it empty. */
	300	return 0;
	301
	302	return err;
	303	}
	304	DEVICE_ATTR(serial, S_IRUGO, virtblk_serial_show, NULL);
	305
98e94444	306	static int __devinit virtblk_probe(struct virtio_device *vdev)
e467cde2 RR	307	{
e467cde2 RR	308	struct virtio_blk *vblk;
69740c8b	309	struct request_queue *q;
4f3bf19c	310	int err;
e467cde2	311	u64 cap;
69740c8b CH	312	u32 v, blk_size, sg_elems, opt_io_size;
	313	u16 min_io_size;
	314	u8 physical_block_exp, alignment_offset;
e467cde2	315
d50ed907	316	if (index_to_minor(index) >= 1 << MINORBITS)
4f3bf19c CB	317	return -ENOSPC;
4f3bf19c CB	318
0864b79a RR	319	/* We need to know how many segments before we allocate. */
	320	err = virtio_config_val(vdev, VIRTIO_BLK_F_SEG_MAX,
	321	offsetof(struct virtio_blk_config, seg_max),
	322	&sg_elems);
a5b365a6 CH	323
	324	/* We need at least one SG element, whatever they say. */
	325	if (err \|\| !sg_elems)
0864b79a RR	326	sg_elems = 1;
	327
	328	/* We need an extra sg elements at head and tail. */
	329	sg_elems += 2;
	330	vdev->priv = vblk = kmalloc(sizeof(*vblk) +
	331	sizeof(vblk->sg[0]) * sg_elems, GFP_KERNEL);
e467cde2 RR	332	if (!vblk) {
	333	err = -ENOMEM;
	334	goto out;
	335	}
	336
	337	INIT_LIST_HEAD(&vblk->reqs);
	338	spin_lock_init(&vblk->lock);
	339	vblk->vdev = vdev;
0864b79a RR	340	vblk->sg_elems = sg_elems;
0864b79a RR	341	sg_init_table(vblk->sg, vblk->sg_elems);
e467cde2 RR	342
e467cde2 RR	343	/* We expect one virtqueue, for output. */
d2a7ddda	344	vblk->vq = virtio_find_single_vq(vdev, blk_done, "requests");
e467cde2 RR	345	if (IS_ERR(vblk->vq)) {
	346	err = PTR_ERR(vblk->vq);
	347	goto out_free_vblk;
	348	}
	349
	350	vblk->pool = mempool_create_kmalloc_pool(1,sizeof(struct virtblk_req));
	351	if (!vblk->pool) {
	352	err = -ENOMEM;
	353	goto out_free_vq;
	354	}
	355
e467cde2	356	/* FIXME: How many partitions? How long is a piece of string? */
4f3bf19c	357	vblk->disk = alloc_disk(1 << PART_BITS);
e467cde2 RR	358	if (!vblk->disk) {
e467cde2 RR	359	err = -ENOMEM;
4f3bf19c	360	goto out_mempool;
e467cde2 RR	361	}
e467cde2 RR	362
69740c8b CH	363	q = vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
69740c8b CH	364	if (!q) {
e467cde2 RR	365	err = -ENOMEM;
	366	goto out_put_disk;
	367	}
	368
69740c8b	369	q->queuedata = vblk;
7d116b62	370
d50ed907 CB	371	if (index < 26) {
	372	sprintf(vblk->disk->disk_name, "vd%c", 'a' + index % 26);
	373	} else if (index < (26 + 1) * 26) {
	374	sprintf(vblk->disk->disk_name, "vd%c%c",
	375	'a' + index / 26 - 1, 'a' + index % 26);
	376	} else {
	377	const unsigned int m1 = (index / 26 - 1) / 26 - 1;
	378	const unsigned int m2 = (index / 26 - 1) % 26;
	379	const unsigned int m3 = index % 26;
	380	sprintf(vblk->disk->disk_name, "vd%c%c%c",
	381	'a' + m1, 'a' + m2, 'a' + m3);
	382	}
	383
e467cde2	384	vblk->disk->major = major;
d50ed907	385	vblk->disk->first_minor = index_to_minor(index);
e467cde2 RR	386	vblk->disk->private_data = vblk;
e467cde2 RR	387	vblk->disk->fops = &virtblk_fops;
c4839346	388	vblk->disk->driverfs_dev = &vdev->dev;
d50ed907	389	index++;
4f3bf19c	390
10bc310c CH	391	if (virtio_has_feature(vdev, VIRTIO_BLK_F_FLUSH)) {
	392	/*
	393	* If the FLUSH feature is supported we do have support for
	394	* flushing a volatile write cache on the host. Use that
	395	* to implement write barrier support.
	396	*/
00fff265	397	blk_queue_ordered(q, QUEUE_ORDERED_DRAIN_FLUSH);
10bc310c CH	398	} else if (virtio_has_feature(vdev, VIRTIO_BLK_F_BARRIER)) {
	399	/*
	400	* If the BARRIER feature is supported the host expects us
	401	* to order request by tags. This implies there is not
	402	* volatile write cache on the host, and that the host
	403	* never re-orders outstanding I/O. This feature is not
	404	* useful for real life scenarious and deprecated.
	405	*/
00fff265	406	blk_queue_ordered(q, QUEUE_ORDERED_TAG);
10bc310c CH	407	} else {
	408	/*
	409	* If the FLUSH feature is not supported we must assume that
	410	* the host does not perform any kind of volatile write
	411	* caching. We still need to drain the queue to provider
	412	* proper barrier semantics.
	413	*/
2f9e825d	414	blk_queue_ordered(q, QUEUE_ORDERED_DRAIN);
10bc310c	415	}
e467cde2	416
3ef53609 CB	417	/* If disk is read-only in the host, the guest should obey */
	418	if (virtio_has_feature(vdev, VIRTIO_BLK_F_RO))
	419	set_disk_ro(vblk->disk, 1);
	420
a586d4f6	421	/* Host must always specify the capacity. */
72e61eb4 RR	422	vdev->config->get(vdev, offsetof(struct virtio_blk_config, capacity),
72e61eb4 RR	423	&cap, sizeof(cap));
e467cde2 RR	424
	425	/* If capacity is too big, truncate with warning. */
	426	if ((sector_t)cap != cap) {
	427	dev_warn(&vdev->dev, "Capacity %llu too large: truncating\n",
	428	(unsigned long long)cap);
	429	cap = (sector_t)-1;
	430	}
	431	set_capacity(vblk->disk, cap);
	432
0864b79a	433	/* We can handle whatever the host told us to handle. */
ee714f2d	434	blk_queue_max_segments(q, vblk->sg_elems-2);
0864b79a	435
4eff3cae	436	/* No need to bounce any requests */
69740c8b	437	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
4eff3cae	438
4b7f7e20	439	/* No real sector limit. */
ee714f2d	440	blk_queue_max_hw_sectors(q, -1U);
4b7f7e20	441
a586d4f6 RR	442	/* Host can optionally specify maximum segment size and number of
	443	* segments. */
	444	err = virtio_config_val(vdev, VIRTIO_BLK_F_SIZE_MAX,
	445	offsetof(struct virtio_blk_config, size_max),
	446	&v);
e467cde2	447	if (!err)
69740c8b	448	blk_queue_max_segment_size(q, v);
4b7f7e20	449	else
69740c8b	450	blk_queue_max_segment_size(q, -1U);
e467cde2	451
066f4d82 CB	452	/* Host can optionally specify the block size of the device */
	453	err = virtio_config_val(vdev, VIRTIO_BLK_F_BLK_SIZE,
	454	offsetof(struct virtio_blk_config, blk_size),
	455	&blk_size);
	456	if (!err)
69740c8b CH	457	blk_queue_logical_block_size(q, blk_size);
	458	else
	459	blk_size = queue_logical_block_size(q);
	460
	461	/* Use topology information if available */
	462	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	463	offsetof(struct virtio_blk_config, physical_block_exp),
	464	&physical_block_exp);
	465	if (!err && physical_block_exp)
	466	blk_queue_physical_block_size(q,
	467	blk_size * (1 << physical_block_exp));
	468
	469	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	470	offsetof(struct virtio_blk_config, alignment_offset),
	471	&alignment_offset);
	472	if (!err && alignment_offset)
	473	blk_queue_alignment_offset(q, blk_size * alignment_offset);
	474
	475	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	476	offsetof(struct virtio_blk_config, min_io_size),
	477	&min_io_size);
	478	if (!err && min_io_size)
	479	blk_queue_io_min(q, blk_size * min_io_size);
	480
	481	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
	482	offsetof(struct virtio_blk_config, opt_io_size),
	483	&opt_io_size);
	484	if (!err && opt_io_size)
	485	blk_queue_io_opt(q, blk_size * opt_io_size);
	486
066f4d82	487
e467cde2	488	add_disk(vblk->disk);
a5eb9e4f RH	489	err = device_create_file(disk_to_dev(vblk->disk), &dev_attr_serial);
	490	if (err)
	491	goto out_del_disk;
	492
e467cde2 RR	493	return 0;
e467cde2 RR	494
a5eb9e4f RH	495	out_del_disk:
	496	del_gendisk(vblk->disk);
	497	blk_cleanup_queue(vblk->disk->queue);
e467cde2 RR	498	out_put_disk:
e467cde2 RR	499	put_disk(vblk->disk);
e467cde2 RR	500	out_mempool:
	501	mempool_destroy(vblk->pool);
	502	out_free_vq:
d2a7ddda	503	vdev->config->del_vqs(vdev);
e467cde2 RR	504	out_free_vblk:
	505	kfree(vblk);
	506	out:
	507	return err;
	508	}
	509
98e94444	510	static void __devexit virtblk_remove(struct virtio_device *vdev)
e467cde2 RR	511	{
e467cde2 RR	512	struct virtio_blk *vblk = vdev->priv;
e467cde2	513
6e5aa7ef	514	/* Nothing should be pending. */
e467cde2	515	BUG_ON(!list_empty(&vblk->reqs));
6e5aa7ef RR	516
	517	/* Stop all the virtqueues. */
	518	vdev->config->reset(vdev);
	519
ac9d463a	520	del_gendisk(vblk->disk);
e467cde2 RR	521	blk_cleanup_queue(vblk->disk->queue);
e467cde2 RR	522	put_disk(vblk->disk);
e467cde2	523	mempool_destroy(vblk->pool);
d2a7ddda	524	vdev->config->del_vqs(vdev);
e467cde2 RR	525	kfree(vblk);
	526	}
	527
47483e25	528	static const struct virtio_device_id id_table[] = {
e467cde2 RR	529	{ VIRTIO_ID_BLOCK, VIRTIO_DEV_ANY_ID },
	530	{ 0 },
	531	};
	532
c45a6816 RR	533	static unsigned int features[] = {
c45a6816 RR	534	VIRTIO_BLK_F_BARRIER, VIRTIO_BLK_F_SEG_MAX, VIRTIO_BLK_F_SIZE_MAX,
066f4d82	535	VIRTIO_BLK_F_GEOMETRY, VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
69740c8b	536	VIRTIO_BLK_F_SCSI, VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY
c45a6816 RR	537	};
c45a6816 RR	538
4fbfff76 RM	539	/*
	540	* virtio_blk causes spurious section mismatch warning by
	541	* simultaneously referring to a __devinit and a __devexit function.
	542	* Use __refdata to avoid this warning.
	543	*/
	544	static struct virtio_driver __refdata virtio_blk = {
c45a6816 RR	545	.feature_table = features,
c45a6816 RR	546	.feature_table_size = ARRAY_SIZE(features),
e467cde2 RR	547	.driver.name = KBUILD_MODNAME,
	548	.driver.owner = THIS_MODULE,
	549	.id_table = id_table,
	550	.probe = virtblk_probe,
	551	.remove = __devexit_p(virtblk_remove),
	552	};
	553
	554	static int __init init(void)
	555	{
4f3bf19c CB	556	major = register_blkdev(0, "virtblk");
	557	if (major < 0)
	558	return major;
e467cde2 RR	559	return register_virtio_driver(&virtio_blk);
	560	}
	561
	562	static void __exit fini(void)
	563	{
4f3bf19c	564	unregister_blkdev(major, "virtblk");
e467cde2 RR	565	unregister_virtio_driver(&virtio_blk);
	566	}
	567	module_init(init);
	568	module_exit(fini);
	569
	570	MODULE_DEVICE_TABLE(virtio, id_table);
	571	MODULE_DESCRIPTION("Virtio block driver");
	572	MODULE_LICENSE("GPL");