[linux.git] / block / ioctl.c

#include <linux/capability.h>
#include <linux/blkdev.h>
#include <linux/export.h>
#include <linux/gfp.h>
#include <linux/blkpg.h>
#include <linux/hdreg.h>
#include <linux/badblocks.h>
#include <linux/backing-dev.h>
#include <linux/fs.h>
#include <linux/blktrace_api.h>
#include <linux/pr.h>
#include <asm/uaccess.h>

static int blkpg_ioctl(struct block_device *bdev, struct blkpg_ioctl_arg __user *arg)
{
	struct block_device *bdevp;
	struct gendisk *disk;
	struct hd_struct *part, *lpart;
	struct blkpg_ioctl_arg a;
	struct blkpg_partition p;
	struct disk_part_iter piter;
	long long start, length;
	int partno;

	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;
	if (copy_from_user(&a, arg, sizeof(struct blkpg_ioctl_arg)))
		return -EFAULT;
	if (copy_from_user(&p, a.data, sizeof(struct blkpg_partition)))
		return -EFAULT;
	disk = bdev->bd_disk;
	if (bdev != bdev->bd_contains)
		return -EINVAL;
	partno = p.pno;
	if (partno <= 0)
		return -EINVAL;
	switch (a.op) {
		case BLKPG_ADD_PARTITION:
			start = p.start >> 9;
			length = p.length >> 9;
			/* check for fit in a hd_struct */
			if (sizeof(sector_t) == sizeof(long) &&
			    sizeof(long long) > sizeof(long)) {
				long pstart = start, plength = length;
				if (pstart != start || plength != length
				    || pstart < 0 || plength < 0 || partno > 65535)
					return -EINVAL;
			}

			mutex_lock(&bdev->bd_mutex);

			/* overlap? */
			disk_part_iter_init(&piter, disk,
					    DISK_PITER_INCL_EMPTY);
			while ((part = disk_part_iter_next(&piter))) {
				if (!(start + length <= part->start_sect ||
				      start >= part->start_sect + part->nr_sects)) {
					disk_part_iter_exit(&piter);
					mutex_unlock(&bdev->bd_mutex);
					return -EBUSY;
				}
			}
			disk_part_iter_exit(&piter);

			/* all seems OK */
			part = add_partition(disk, partno, start, length,
					     ADDPART_FLAG_NONE, NULL);
			mutex_unlock(&bdev->bd_mutex);
			return PTR_ERR_OR_ZERO(part);
		case BLKPG_DEL_PARTITION:
			part = disk_get_part(disk, partno);
			if (!part)
				return -ENXIO;

			bdevp = bdget(part_devt(part));
			disk_put_part(part);
			if (!bdevp)
				return -ENOMEM;

			mutex_lock(&bdevp->bd_mutex);
			if (bdevp->bd_openers) {
				mutex_unlock(&bdevp->bd_mutex);
				bdput(bdevp);
				return -EBUSY;
			}
			/* all seems OK */
			fsync_bdev(bdevp);
			invalidate_bdev(bdevp);

			mutex_lock_nested(&bdev->bd_mutex, 1);
			delete_partition(disk, partno);
			mutex_unlock(&bdev->bd_mutex);
			mutex_unlock(&bdevp->bd_mutex);
			bdput(bdevp);

			return 0;
		case BLKPG_RESIZE_PARTITION:
			start = p.start >> 9;
			/* new length of partition in bytes */
			length = p.length >> 9;
			/* check for fit in a hd_struct */
			if (sizeof(sector_t) == sizeof(long) &&
			    sizeof(long long) > sizeof(long)) {
				long pstart = start, plength = length;
				if (pstart != start || plength != length
				    || pstart < 0 || plength < 0)
					return -EINVAL;
			}
			part = disk_get_part(disk, partno);
			if (!part)
				return -ENXIO;
			bdevp = bdget(part_devt(part));
			if (!bdevp) {
				disk_put_part(part);
				return -ENOMEM;
			}
			mutex_lock(&bdevp->bd_mutex);
			mutex_lock_nested(&bdev->bd_mutex, 1);
			if (start != part->start_sect) {
				mutex_unlock(&bdevp->bd_mutex);
				mutex_unlock(&bdev->bd_mutex);
				bdput(bdevp);
				disk_put_part(part);
				return -EINVAL;
			}
			/* overlap? */
			disk_part_iter_init(&piter, disk,
					    DISK_PITER_INCL_EMPTY);
			while ((lpart = disk_part_iter_next(&piter))) {
				if (lpart->partno != partno &&
				   !(start + length <= lpart->start_sect ||
				   start >= lpart->start_sect + lpart->nr_sects)
				   ) {
					disk_part_iter_exit(&piter);
					mutex_unlock(&bdevp->bd_mutex);
					mutex_unlock(&bdev->bd_mutex);
					bdput(bdevp);
					disk_put_part(part);
					return -EBUSY;
				}
			}
			disk_part_iter_exit(&piter);
			part_nr_sects_write(part, (sector_t)length);
			i_size_write(bdevp->bd_inode, p.length);
			mutex_unlock(&bdevp->bd_mutex);
			mutex_unlock(&bdev->bd_mutex);
			bdput(bdevp);
			disk_put_part(part);
			return 0;
		default:
			return -EINVAL;
	}
}

/*
 * This is an exported API for the block driver, and will not
 * acquire bd_mutex. This API should be used in case that
 * caller has held bd_mutex already.
 */
int __blkdev_reread_part(struct block_device *bdev)
{
	struct gendisk *disk = bdev->bd_disk;

	if (!disk_part_scan_enabled(disk) || bdev != bdev->bd_contains)
		return -EINVAL;
	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;

	lockdep_assert_held(&bdev->bd_mutex);

	return rescan_partitions(disk, bdev);
}
EXPORT_SYMBOL(__blkdev_reread_part);

/*
 * This is an exported API for the block driver, and will
 * try to acquire bd_mutex. If bd_mutex has been held already
 * in current context, please call __blkdev_reread_part().
 *
 * Make sure the held locks in current context aren't required
 * in open()/close() handler and I/O path for avoiding ABBA deadlock:
 * - bd_mutex is held before calling block driver's open/close
 *   handler
 * - reading partition table may submit I/O to the block device
 */
int blkdev_reread_part(struct block_device *bdev)
{
	int res;

	mutex_lock(&bdev->bd_mutex);
	res = __blkdev_reread_part(bdev);
	mutex_unlock(&bdev->bd_mutex);

	return res;
}
EXPORT_SYMBOL(blkdev_reread_part);

static int blk_ioctl_discard(struct block_device *bdev, fmode_t mode,
		unsigned long arg, unsigned long flags)
{
	uint64_t range[2];
	uint64_t start, len;

	if (!(mode & FMODE_WRITE))
		return -EBADF;

	if (copy_from_user(range, (void __user *)arg, sizeof(range)))
		return -EFAULT;

	start = range[0];
	len = range[1];

	if (start & 511)
		return -EINVAL;
	if (len & 511)
		return -EINVAL;
	start >>= 9;
	len >>= 9;

	if (start + len > (i_size_read(bdev->bd_inode) >> 9))
		return -EINVAL;
	return blkdev_issue_discard(bdev, start, len, GFP_KERNEL, flags);
}

static int blk_ioctl_zeroout(struct block_device *bdev, fmode_t mode,
		unsigned long arg)
{
	uint64_t range[2];
	uint64_t start, len;

	if (!(mode & FMODE_WRITE))
		return -EBADF;

	if (copy_from_user(range, (void __user *)arg, sizeof(range)))
		return -EFAULT;

	start = range[0];
	len = range[1];

	if (start & 511)
		return -EINVAL;
	if (len & 511)
		return -EINVAL;
	start >>= 9;
	len >>= 9;

	if (start + len > (i_size_read(bdev->bd_inode) >> 9))
		return -EINVAL;

	return blkdev_issue_zeroout(bdev, start, len, GFP_KERNEL, false);
}

static int put_ushort(unsigned long arg, unsigned short val)
{
	return put_user(val, (unsigned short __user *)arg);
}

static int put_int(unsigned long arg, int val)
{
	return put_user(val, (int __user *)arg);
}

static int put_uint(unsigned long arg, unsigned int val)
{
	return put_user(val, (unsigned int __user *)arg);
}

static int put_long(unsigned long arg, long val)
{
	return put_user(val, (long __user *)arg);
}

static int put_ulong(unsigned long arg, unsigned long val)
{
	return put_user(val, (unsigned long __user *)arg);
}

static int put_u64(unsigned long arg, u64 val)
{
	return put_user(val, (u64 __user *)arg);
}

int __blkdev_driver_ioctl(struct block_device *bdev, fmode_t mode,
			unsigned cmd, unsigned long arg)
{
	struct gendisk *disk = bdev->bd_disk;

	if (disk->fops->ioctl)
		return disk->fops->ioctl(bdev, mode, cmd, arg);

	return -ENOTTY;
}
/*
 * For the record: _GPL here is only because somebody decided to slap it
 * on the previous export.  Sheer idiocy, since it wasn't copyrightable
 * at all and could be open-coded without any exports by anybody who cares.
 */
EXPORT_SYMBOL_GPL(__blkdev_driver_ioctl);

static int blkdev_pr_register(struct block_device *bdev,
		struct pr_registration __user *arg)
{
	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	struct pr_registration reg;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;
	if (!ops || !ops->pr_register)
		return -EOPNOTSUPP;
	if (copy_from_user(&reg, arg, sizeof(reg)))
		return -EFAULT;

	if (reg.flags & ~PR_FL_IGNORE_KEY)
		return -EOPNOTSUPP;
	return ops->pr_register(bdev, reg.old_key, reg.new_key, reg.flags);
}

static int blkdev_pr_reserve(struct block_device *bdev,
		struct pr_reservation __user *arg)
{
	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	struct pr_reservation rsv;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;
	if (!ops || !ops->pr_reserve)
		return -EOPNOTSUPP;
	if (copy_from_user(&rsv, arg, sizeof(rsv)))
		return -EFAULT;

	if (rsv.flags & ~PR_FL_IGNORE_KEY)
		return -EOPNOTSUPP;
	return ops->pr_reserve(bdev, rsv.key, rsv.type, rsv.flags);
}

static int blkdev_pr_release(struct block_device *bdev,
		struct pr_reservation __user *arg)
{
	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	struct pr_reservation rsv;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;
	if (!ops || !ops->pr_release)
		return -EOPNOTSUPP;
	if (copy_from_user(&rsv, arg, sizeof(rsv)))
		return -EFAULT;

	if (rsv.flags)
		return -EOPNOTSUPP;
	return ops->pr_release(bdev, rsv.key, rsv.type);
}

static int blkdev_pr_preempt(struct block_device *bdev,
		struct pr_preempt __user *arg, bool abort)
{
	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	struct pr_preempt p;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;
	if (!ops || !ops->pr_preempt)
		return -EOPNOTSUPP;
	if (copy_from_user(&p, arg, sizeof(p)))
		return -EFAULT;

	if (p.flags)
		return -EOPNOTSUPP;
	return ops->pr_preempt(bdev, p.old_key, p.new_key, p.type, abort);
}

static int blkdev_pr_clear(struct block_device *bdev,
		struct pr_clear __user *arg)
{
	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	struct pr_clear c;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;
	if (!ops || !ops->pr_clear)
		return -EOPNOTSUPP;
	if (copy_from_user(&c, arg, sizeof(c)))
		return -EFAULT;

	if (c.flags)
		return -EOPNOTSUPP;
	return ops->pr_clear(bdev, c.key);
}

/*
 * Is it an unrecognized ioctl? The correct returns are either
 * ENOTTY (final) or ENOIOCTLCMD ("I don't know this one, try a
 * fallback"). ENOIOCTLCMD gets turned into ENOTTY by the ioctl
 * code before returning.
 *
 * Confused drivers sometimes return EINVAL, which is wrong. It
 * means "I understood the ioctl command, but the parameters to
 * it were wrong".
 *
 * We should aim to just fix the broken drivers, the EINVAL case
 * should go away.
 */
static inline int is_unrecognized_ioctl(int ret)
{
	return	ret == -EINVAL ||
		ret == -ENOTTY ||
		ret == -ENOIOCTLCMD;
}

#ifdef CONFIG_FS_DAX
bool blkdev_dax_capable(struct block_device *bdev)
{
	struct gendisk *disk = bdev->bd_disk;

	if (!disk->fops->direct_access)
		return false;

	/*
	 * If the partition is not aligned on a page boundary, we can't
	 * do dax I/O to it.
	 */
	if ((bdev->bd_part->start_sect % (PAGE_SIZE / 512))
			|| (bdev->bd_part->nr_sects % (PAGE_SIZE / 512)))
		return false;

	/*
	 * If the device has known bad blocks, force all I/O through the
	 * driver / page cache.
	 *
	 * TODO: support finer grained dax error handling
	 */
	if (disk->bb && disk->bb->count)
		return false;

	return true;
}

static int blkdev_daxset(struct block_device *bdev, unsigned long argp)
{
	unsigned long arg;
	int rc = 0;

	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;

	if (get_user(arg, (int __user *)(argp)))
		return -EFAULT;
	arg = !!arg;
	if (arg == !!(bdev->bd_inode->i_flags & S_DAX))
		return 0;

	if (arg)
		arg = S_DAX;

	if (arg && !blkdev_dax_capable(bdev))
		return -ENOTTY;

	mutex_lock(&bdev->bd_inode->i_mutex);
	if (bdev->bd_map_count == 0)
		inode_set_flags(bdev->bd_inode, arg, S_DAX);
	else
		rc = -EBUSY;
	mutex_unlock(&bdev->bd_inode->i_mutex);
	return rc;
}
#else
static int blkdev_daxset(struct block_device *bdev, int arg)
{
	if (arg)
		return -ENOTTY;
	return 0;
}
#endif

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

	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;

	ret = __blkdev_driver_ioctl(bdev, mode, cmd, arg);
	if (!is_unrecognized_ioctl(ret))
		return ret;

	fsync_bdev(bdev);
	invalidate_bdev(bdev);
	return 0;
}

static int blkdev_roset(struct block_device *bdev, fmode_t mode,
		unsigned cmd, unsigned long arg)
{
	int ret, n;

	ret = __blkdev_driver_ioctl(bdev, mode, cmd, arg);
	if (!is_unrecognized_ioctl(ret))
		return ret;
	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;
	if (get_user(n, (int __user *)arg))
		return -EFAULT;
	set_device_ro(bdev, n);
	return 0;
}

static int blkdev_getgeo(struct block_device *bdev,
		struct hd_geometry __user *argp)
{
	struct gendisk *disk = bdev->bd_disk;
	struct hd_geometry geo;
	int ret;

	if (!argp)
		return -EINVAL;
	if (!disk->fops->getgeo)
		return -ENOTTY;

	/*
	 * We need to set the startsect first, the driver may
	 * want to override it.
	 */
	memset(&geo, 0, sizeof(geo));
	geo.start = get_start_sect(bdev);
	ret = disk->fops->getgeo(bdev, &geo);
	if (ret)
		return ret;
	if (copy_to_user(argp, &geo, sizeof(geo)))
		return -EFAULT;
	return 0;
}

/* set the logical block size */
static int blkdev_bszset(struct block_device *bdev, fmode_t mode,
		int __user *argp)
{
	int ret, n;

	if (!capable(CAP_SYS_ADMIN))
		return -EACCES;
	if (!argp)
		return -EINVAL;
	if (get_user(n, argp))
		return -EFAULT;

	if (!(mode & FMODE_EXCL)) {
		bdgrab(bdev);
		if (blkdev_get(bdev, mode | FMODE_EXCL, &bdev) < 0)
			return -EBUSY;
	}

	ret = set_blocksize(bdev, n);
	if (!(mode & FMODE_EXCL))
		blkdev_put(bdev, mode | FMODE_EXCL);
	return ret;
}

/*
 * always keep this in sync with compat_blkdev_ioctl()
 */
int blkdev_ioctl(struct block_device *bdev, fmode_t mode, unsigned cmd,
			unsigned long arg)
{
	struct backing_dev_info *bdi;
	void __user *argp = (void __user *)arg;
	loff_t size;
	unsigned int max_sectors;

	switch (cmd) {
	case BLKFLSBUF:
		return blkdev_flushbuf(bdev, mode, cmd, arg);
	case BLKROSET:
		return blkdev_roset(bdev, mode, cmd, arg);
	case BLKDISCARD:
		return blk_ioctl_discard(bdev, mode, arg, 0);
	case BLKSECDISCARD:
		return blk_ioctl_discard(bdev, mode, arg,
				BLKDEV_DISCARD_SECURE);
	case BLKZEROOUT:
		return blk_ioctl_zeroout(bdev, mode, arg);
	case HDIO_GETGEO:
		return blkdev_getgeo(bdev, argp);
	case BLKRAGET:
	case BLKFRAGET:
		if (!arg)
			return -EINVAL;
		bdi = blk_get_backing_dev_info(bdev);
		return put_long(arg, (bdi->ra_pages * PAGE_CACHE_SIZE) / 512);
	case BLKROGET:
		return put_int(arg, bdev_read_only(bdev) != 0);
	case BLKBSZGET: /* get block device soft block size (cf. BLKSSZGET) */
		return put_int(arg, block_size(bdev));
	case BLKSSZGET: /* get block device logical block size */
		return put_int(arg, bdev_logical_block_size(bdev));
	case BLKPBSZGET: /* get block device physical block size */
		return put_uint(arg, bdev_physical_block_size(bdev));
	case BLKIOMIN:
		return put_uint(arg, bdev_io_min(bdev));
	case BLKIOOPT:
		return put_uint(arg, bdev_io_opt(bdev));
	case BLKALIGNOFF:
		return put_int(arg, bdev_alignment_offset(bdev));
	case BLKDISCARDZEROES:
		return put_uint(arg, bdev_discard_zeroes_data(bdev));
	case BLKSECTGET:
		max_sectors = min_t(unsigned int, USHRT_MAX,
				    queue_max_sectors(bdev_get_queue(bdev)));
		return put_ushort(arg, max_sectors);
	case BLKROTATIONAL:
		return put_ushort(arg, !blk_queue_nonrot(bdev_get_queue(bdev)));
	case BLKRASET:
	case BLKFRASET:
		if(!capable(CAP_SYS_ADMIN))
			return -EACCES;
		bdi = blk_get_backing_dev_info(bdev);
		bdi->ra_pages = (arg * 512) / PAGE_CACHE_SIZE;
		return 0;
	case BLKBSZSET:
		return blkdev_bszset(bdev, mode, argp);
	case BLKPG:
		return blkpg_ioctl(bdev, argp);
	case BLKRRPART:
		return blkdev_reread_part(bdev);
	case BLKGETSIZE:
		size = i_size_read(bdev->bd_inode);
		if ((size >> 9) > ~0UL)
			return -EFBIG;
		return put_ulong(arg, size >> 9);
	case BLKGETSIZE64:
		return put_u64(arg, i_size_read(bdev->bd_inode));
	case BLKTRACESTART:
	case BLKTRACESTOP:
	case BLKTRACESETUP:
	case BLKTRACETEARDOWN:
		return blk_trace_ioctl(bdev, cmd, argp);
	case BLKDAXSET:
		return blkdev_daxset(bdev, arg);
	case BLKDAXGET:
		return put_int(arg, !!(bdev->bd_inode->i_flags & S_DAX));
		break;
	case IOC_PR_REGISTER:
		return blkdev_pr_register(bdev, argp);
	case IOC_PR_RESERVE:
		return blkdev_pr_reserve(bdev, argp);
	case IOC_PR_RELEASE:
		return blkdev_pr_release(bdev, argp);
	case IOC_PR_PREEMPT:
		return blkdev_pr_preempt(bdev, argp, false);
	case IOC_PR_PREEMPT_ABORT:
		return blkdev_pr_preempt(bdev, argp, true);
	case IOC_PR_CLEAR:
		return blkdev_pr_clear(bdev, argp);
	default:
		return __blkdev_driver_ioctl(bdev, mode, cmd, arg);
	}
}
EXPORT_SYMBOL_GPL(blkdev_ioctl);
Commit	Line	Data
c59ede7b	1	#include <linux/capability.h>
1da177e4	2	#include <linux/blkdev.h>
d5decd3b	3	#include <linux/export.h>
5a0e3ad6	4	#include <linux/gfp.h>
1da177e4	5	#include <linux/blkpg.h>
a885c8c4	6	#include <linux/hdreg.h>
57f7f317	7	#include <linux/badblocks.h>
1da177e4	8	#include <linux/backing-dev.h>
ff01bb48	9	#include <linux/fs.h>
2056a782	10	#include <linux/blktrace_api.h>
bbd3e064	11	#include <linux/pr.h>
1da177e4 LT	12	#include <asm/uaccess.h>
	13
	14	static int blkpg_ioctl(struct block_device bdev, struct blkpg_ioctl_arg __user arg)
	15	{
	16	struct block_device *bdevp;
	17	struct gendisk *disk;
c83f6bf9	18	struct hd_struct part, lpart;
1da177e4 LT	19	struct blkpg_ioctl_arg a;
1da177e4 LT	20	struct blkpg_partition p;
e71bf0d0	21	struct disk_part_iter piter;
1da177e4	22	long long start, length;
cf771cb5	23	int partno;
1da177e4 LT	24
	25	if (!capable(CAP_SYS_ADMIN))
	26	return -EACCES;
	27	if (copy_from_user(&a, arg, sizeof(struct blkpg_ioctl_arg)))
	28	return -EFAULT;
	29	if (copy_from_user(&p, a.data, sizeof(struct blkpg_partition)))
	30	return -EFAULT;
	31	disk = bdev->bd_disk;
	32	if (bdev != bdev->bd_contains)
	33	return -EINVAL;
cf771cb5	34	partno = p.pno;
540eed56	35	if (partno <= 0)
1da177e4 LT	36	return -EINVAL;
	37	switch (a.op) {
	38	case BLKPG_ADD_PARTITION:
	39	start = p.start >> 9;
	40	length = p.length >> 9;
c83f6bf9 VG	41	/* check for fit in a hd_struct */
c83f6bf9 VG	42	if (sizeof(sector_t) == sizeof(long) &&
1da177e4 LT	43	sizeof(long long) > sizeof(long)) {
	44	long pstart = start, plength = length;
	45	if (pstart != start \|\| plength != length
2bd6efad	46	\|\| pstart < 0 \|\| plength < 0 \|\| partno > 65535)
1da177e4 LT	47	return -EINVAL;
1da177e4 LT	48	}
88e34126	49
c039e313	50	mutex_lock(&bdev->bd_mutex);
88e34126	51
1da177e4	52	/* overlap? */
e71bf0d0 TH	53	disk_part_iter_init(&piter, disk,
	54	DISK_PITER_INCL_EMPTY);
	55	while ((part = disk_part_iter_next(&piter))) {
	56	if (!(start + length <= part->start_sect \|\|
	57	start >= part->start_sect + part->nr_sects)) {
	58	disk_part_iter_exit(&piter);
c039e313	59	mutex_unlock(&bdev->bd_mutex);
1da177e4 LT	60	return -EBUSY;
	61	}
	62	}
e71bf0d0 TH	63	disk_part_iter_exit(&piter);
e71bf0d0 TH	64
1da177e4	65	/* all seems OK */
ba32929a	66	part = add_partition(disk, partno, start, length,
6d1d8050	67	ADDPART_FLAG_NONE, NULL);
c039e313	68	mutex_unlock(&bdev->bd_mutex);
c7d1ba41	69	return PTR_ERR_OR_ZERO(part);
1da177e4	70	case BLKPG_DEL_PARTITION:
e71bf0d0 TH	71	part = disk_get_part(disk, partno);
e71bf0d0 TH	72	if (!part)
1da177e4	73	return -ENXIO;
e71bf0d0 TH	74
	75	bdevp = bdget(part_devt(part));
	76	disk_put_part(part);
1da177e4 LT	77	if (!bdevp)
1da177e4 LT	78	return -ENOMEM;
e71bf0d0	79
2e7b651d	80	mutex_lock(&bdevp->bd_mutex);
1da177e4	81	if (bdevp->bd_openers) {
c039e313	82	mutex_unlock(&bdevp->bd_mutex);
1da177e4 LT	83	bdput(bdevp);
	84	return -EBUSY;
	85	}
	86	/* all seems OK */
	87	fsync_bdev(bdevp);
f98393a6	88	invalidate_bdev(bdevp);
1da177e4	89
6d740cd5	90	mutex_lock_nested(&bdev->bd_mutex, 1);
cf771cb5	91	delete_partition(disk, partno);
c039e313 AV	92	mutex_unlock(&bdev->bd_mutex);
c039e313 AV	93	mutex_unlock(&bdevp->bd_mutex);
1da177e4 LT	94	bdput(bdevp);
1da177e4 LT	95
c83f6bf9 VG	96	return 0;
	97	case BLKPG_RESIZE_PARTITION:
	98	start = p.start >> 9;
	99	/* new length of partition in bytes */
	100	length = p.length >> 9;
	101	/* check for fit in a hd_struct */
	102	if (sizeof(sector_t) == sizeof(long) &&
	103	sizeof(long long) > sizeof(long)) {
	104	long pstart = start, plength = length;
	105	if (pstart != start \|\| plength != length
	106	\|\| pstart < 0 \|\| plength < 0)
	107	return -EINVAL;
	108	}
	109	part = disk_get_part(disk, partno);
	110	if (!part)
	111	return -ENXIO;
	112	bdevp = bdget(part_devt(part));
	113	if (!bdevp) {
	114	disk_put_part(part);
	115	return -ENOMEM;
	116	}
	117	mutex_lock(&bdevp->bd_mutex);
	118	mutex_lock_nested(&bdev->bd_mutex, 1);
	119	if (start != part->start_sect) {
	120	mutex_unlock(&bdevp->bd_mutex);
	121	mutex_unlock(&bdev->bd_mutex);
	122	bdput(bdevp);
	123	disk_put_part(part);
	124	return -EINVAL;
	125	}
	126	/* overlap? */
	127	disk_part_iter_init(&piter, disk,
	128	DISK_PITER_INCL_EMPTY);
	129	while ((lpart = disk_part_iter_next(&piter))) {
	130	if (lpart->partno != partno &&
	131	!(start + length <= lpart->start_sect \|\|
	132	start >= lpart->start_sect + lpart->nr_sects)
	133	) {
	134	disk_part_iter_exit(&piter);
	135	mutex_unlock(&bdevp->bd_mutex);
	136	mutex_unlock(&bdev->bd_mutex);
	137	bdput(bdevp);
	138	disk_put_part(part);
	139	return -EBUSY;
	140	}
	141	}
	142	disk_part_iter_exit(&piter);
	143	part_nr_sects_write(part, (sector_t)length);
	144	i_size_write(bdevp->bd_inode, p.length);
	145	mutex_unlock(&bdevp->bd_mutex);
	146	mutex_unlock(&bdev->bd_mutex);
	147	bdput(bdevp);
	148	disk_put_part(part);
1da177e4 LT	149	return 0;
	150	default:
	151	return -EINVAL;
	152	}
	153	}
	154
be324177 JW	155	/*
	156	* This is an exported API for the block driver, and will not
	157	* acquire bd_mutex. This API should be used in case that
	158	* caller has held bd_mutex already.
	159	*/
	160	int __blkdev_reread_part(struct block_device *bdev)
1da177e4 LT	161	{
1da177e4 LT	162	struct gendisk *disk = bdev->bd_disk;
1da177e4	163
d27769ec	164	if (!disk_part_scan_enabled(disk) \|\| bdev != bdev->bd_contains)
1da177e4 LT	165	return -EINVAL;
	166	if (!capable(CAP_SYS_ADMIN))
	167	return -EACCES;
be324177 JW	168
	169	lockdep_assert_held(&bdev->bd_mutex);
	170
	171	return rescan_partitions(disk, bdev);
	172	}
	173	EXPORT_SYMBOL(__blkdev_reread_part);
	174
	175	/*
	176	* This is an exported API for the block driver, and will
	177	* try to acquire bd_mutex. If bd_mutex has been held already
	178	* in current context, please call __blkdev_reread_part().
b04a5636 ML	179	*
	180	* Make sure the held locks in current context aren't required
	181	* in open()/close() handler and I/O path for avoiding ABBA deadlock:
	182	* - bd_mutex is held before calling block driver's open/close
	183	* handler
	184	* - reading partition table may submit I/O to the block device
be324177 JW	185	*/
	186	int blkdev_reread_part(struct block_device *bdev)
	187	{
	188	int res;
	189
b04a5636	190	mutex_lock(&bdev->bd_mutex);
be324177	191	res = __blkdev_reread_part(bdev);
c039e313	192	mutex_unlock(&bdev->bd_mutex);
be324177	193
1da177e4 LT	194	return res;
1da177e4 LT	195	}
be324177	196	EXPORT_SYMBOL(blkdev_reread_part);
1da177e4	197
d8e4bb81 CH	198	static int blk_ioctl_discard(struct block_device *bdev, fmode_t mode,
d8e4bb81 CH	199	unsigned long arg, unsigned long flags)
d30a2605	200	{
d8e4bb81 CH	201	uint64_t range[2];
	202	uint64_t start, len;
	203
	204	if (!(mode & FMODE_WRITE))
	205	return -EBADF;
	206
	207	if (copy_from_user(range, (void __user *)arg, sizeof(range)))
	208	return -EFAULT;
	209
	210	start = range[0];
	211	len = range[1];
8d57a98c	212
d30a2605 DW	213	if (start & 511)
	214	return -EINVAL;
	215	if (len & 511)
	216	return -EINVAL;
	217	start >>= 9;
	218	len >>= 9;
	219
77304d2a	220	if (start + len > (i_size_read(bdev->bd_inode) >> 9))
d30a2605	221	return -EINVAL;
8d57a98c	222	return blkdev_issue_discard(bdev, start, len, GFP_KERNEL, flags);
d30a2605 DW	223	}
d30a2605 DW	224
d8e4bb81 CH	225	static int blk_ioctl_zeroout(struct block_device *bdev, fmode_t mode,
d8e4bb81 CH	226	unsigned long arg)
66ba32dc	227	{
d8e4bb81 CH	228	uint64_t range[2];
	229	uint64_t start, len;
	230
	231	if (!(mode & FMODE_WRITE))
	232	return -EBADF;
	233
	234	if (copy_from_user(range, (void __user *)arg, sizeof(range)))
	235	return -EFAULT;
	236
	237	start = range[0];
	238	len = range[1];
	239
66ba32dc MP	240	if (start & 511)
	241	return -EINVAL;
	242	if (len & 511)
	243	return -EINVAL;
	244	start >>= 9;
	245	len >>= 9;
	246
	247	if (start + len > (i_size_read(bdev->bd_inode) >> 9))
	248	return -EINVAL;
	249
d93ba7a5	250	return blkdev_issue_zeroout(bdev, start, len, GFP_KERNEL, false);
66ba32dc MP	251	}
66ba32dc MP	252
1da177e4 LT	253	static int put_ushort(unsigned long arg, unsigned short val)
	254	{
	255	return put_user(val, (unsigned short __user *)arg);
	256	}
	257
	258	static int put_int(unsigned long arg, int val)
	259	{
	260	return put_user(val, (int __user *)arg);
	261	}
	262
ac481c20 MP	263	static int put_uint(unsigned long arg, unsigned int val)
	264	{
	265	return put_user(val, (unsigned int __user *)arg);
	266	}
	267
1da177e4 LT	268	static int put_long(unsigned long arg, long val)
	269	{
	270	return put_user(val, (long __user *)arg);
	271	}
	272
	273	static int put_ulong(unsigned long arg, unsigned long val)
	274	{
	275	return put_user(val, (unsigned long __user *)arg);
	276	}
	277
	278	static int put_u64(unsigned long arg, u64 val)
	279	{
	280	return put_user(val, (u64 __user *)arg);
	281	}
	282
633a08b8 AV	283	int __blkdev_driver_ioctl(struct block_device *bdev, fmode_t mode,
	284	unsigned cmd, unsigned long arg)
	285	{
	286	struct gendisk *disk = bdev->bd_disk;
d4430d62 AV	287
	288	if (disk->fops->ioctl)
	289	return disk->fops->ioctl(bdev, mode, cmd, arg);
633a08b8	290
633a08b8 AV	291	return -ENOTTY;
	292	}
	293	/*
	294	* For the record: _GPL here is only because somebody decided to slap it
	295	* on the previous export. Sheer idiocy, since it wasn't copyrightable
	296	* at all and could be open-coded without any exports by anybody who cares.
	297	*/
	298	EXPORT_SYMBOL_GPL(__blkdev_driver_ioctl);
	299
bbd3e064 CH	300	static int blkdev_pr_register(struct block_device *bdev,
	301	struct pr_registration __user *arg)
	302	{
	303	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	304	struct pr_registration reg;
	305
	306	if (!capable(CAP_SYS_ADMIN))
	307	return -EPERM;
	308	if (!ops \|\| !ops->pr_register)
	309	return -EOPNOTSUPP;
	310	if (copy_from_user(&reg, arg, sizeof(reg)))
	311	return -EFAULT;
	312
	313	if (reg.flags & ~PR_FL_IGNORE_KEY)
	314	return -EOPNOTSUPP;
	315	return ops->pr_register(bdev, reg.old_key, reg.new_key, reg.flags);
	316	}
	317
	318	static int blkdev_pr_reserve(struct block_device *bdev,
	319	struct pr_reservation __user *arg)
	320	{
	321	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	322	struct pr_reservation rsv;
	323
	324	if (!capable(CAP_SYS_ADMIN))
	325	return -EPERM;
	326	if (!ops \|\| !ops->pr_reserve)
	327	return -EOPNOTSUPP;
	328	if (copy_from_user(&rsv, arg, sizeof(rsv)))
	329	return -EFAULT;
	330
	331	if (rsv.flags & ~PR_FL_IGNORE_KEY)
	332	return -EOPNOTSUPP;
	333	return ops->pr_reserve(bdev, rsv.key, rsv.type, rsv.flags);
	334	}
	335
	336	static int blkdev_pr_release(struct block_device *bdev,
	337	struct pr_reservation __user *arg)
	338	{
	339	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	340	struct pr_reservation rsv;
	341
	342	if (!capable(CAP_SYS_ADMIN))
	343	return -EPERM;
	344	if (!ops \|\| !ops->pr_release)
	345	return -EOPNOTSUPP;
	346	if (copy_from_user(&rsv, arg, sizeof(rsv)))
	347	return -EFAULT;
	348
	349	if (rsv.flags)
	350	return -EOPNOTSUPP;
	351	return ops->pr_release(bdev, rsv.key, rsv.type);
	352	}
	353
	354	static int blkdev_pr_preempt(struct block_device *bdev,
	355	struct pr_preempt __user *arg, bool abort)
	356	{
	357	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
	358	struct pr_preempt p;
	359
	360	if (!capable(CAP_SYS_ADMIN))
	361	return -EPERM;
	362	if (!ops \|\| !ops->pr_preempt)
	363	return -EOPNOTSUPP;
364	if (copy_from_user(&p, arg, sizeof(p)))
365	return -EFAULT;
366
367	if (p.flags)
368	return -EOPNOTSUPP;
369	return ops->pr_preempt(bdev, p.old_key, p.new_key, p.type, abort);
370	}
371
372	static int blkdev_pr_clear(struct block_device *bdev,
373	struct pr_clear __user *arg)
374	{
375	const struct pr_ops *ops = bdev->bd_disk->fops->pr_ops;
376	struct pr_clear c;
377
378	if (!capable(CAP_SYS_ADMIN))
379	return -EPERM;
380	if (!ops \|\| !ops->pr_clear)
381	return -EOPNOTSUPP;
382	if (copy_from_user(&c, arg, sizeof(c)))
383	return -EFAULT;
384
385	if (c.flags)
386	return -EOPNOTSUPP;
387	return ops->pr_clear(bdev, c.key);
388	}
389
07d106d0 LT	390	/*
	391	* Is it an unrecognized ioctl? The correct returns are either
	392	* ENOTTY (final) or ENOIOCTLCMD ("I don't know this one, try a
	393	* fallback"). ENOIOCTLCMD gets turned into ENOTTY by the ioctl
	394	* code before returning.
	395	*
	396	* Confused drivers sometimes return EINVAL, which is wrong. It
	397	* means "I understood the ioctl command, but the parameters to
	398	* it were wrong".
	399	*
	400	* We should aim to just fix the broken drivers, the EINVAL case
	401	* should go away.
	402	*/
	403	static inline int is_unrecognized_ioctl(int ret)
	404	{
	405	return ret == -EINVAL \|\|
	406	ret == -ENOTTY \|\|
	407	ret == -ENOIOCTLCMD;
	408	}
	409
5a023cdb DW	410	#ifdef CONFIG_FS_DAX
	411	bool blkdev_dax_capable(struct block_device *bdev)
	412	{
	413	struct gendisk *disk = bdev->bd_disk;
	414
	415	if (!disk->fops->direct_access)
	416	return false;
	417
	418	/*
	419	* If the partition is not aligned on a page boundary, we can't
	420	* do dax I/O to it.
	421	*/
	422	if ((bdev->bd_part->start_sect % (PAGE_SIZE / 512))
	423	\|\| (bdev->bd_part->nr_sects % (PAGE_SIZE / 512)))
	424	return false;
	425
57f7f317 DW	426	/*
	427	* If the device has known bad blocks, force all I/O through the
	428	* driver / page cache.
	429	*
	430	* TODO: support finer grained dax error handling
	431	*/
	432	if (disk->bb && disk->bb->count)
	433	return false;
	434
5a023cdb DW	435	return true;
	436	}
	437
	438	static int blkdev_daxset(struct block_device *bdev, unsigned long argp)
	439	{
	440	unsigned long arg;
	441	int rc = 0;
	442
	443	if (!capable(CAP_SYS_ADMIN))
	444	return -EACCES;
	445
	446	if (get_user(arg, (int __user *)(argp)))
	447	return -EFAULT;
	448	arg = !!arg;
	449	if (arg == !!(bdev->bd_inode->i_flags & S_DAX))
	450	return 0;
	451
	452	if (arg)
	453	arg = S_DAX;
	454
	455	if (arg && !blkdev_dax_capable(bdev))
	456	return -ENOTTY;
	457
	458	mutex_lock(&bdev->bd_inode->i_mutex);
	459	if (bdev->bd_map_count == 0)
	460	inode_set_flags(bdev->bd_inode, arg, S_DAX);
	461	else
	462	rc = -EBUSY;
	463	mutex_unlock(&bdev->bd_inode->i_mutex);
	464	return rc;
	465	}
	466	#else
	467	static int blkdev_daxset(struct block_device *bdev, int arg)
	468	{
	469	if (arg)
	470	return -ENOTTY;
	471	return 0;
	472	}
	473	#endif
	474
d8e4bb81 CH	475	static int blkdev_flushbuf(struct block_device *bdev, fmode_t mode,
d8e4bb81 CH	476	unsigned cmd, unsigned long arg)
bb93e3a5	477	{
d8e4bb81	478	int ret;
bb93e3a5	479
d8e4bb81 CH	480	if (!capable(CAP_SYS_ADMIN))
d8e4bb81 CH	481	return -EACCES;
bb93e3a5	482
d8e4bb81 CH	483	ret = __blkdev_driver_ioctl(bdev, mode, cmd, arg);
	484	if (!is_unrecognized_ioctl(ret))
	485	return ret;
bb93e3a5	486
d8e4bb81 CH	487	fsync_bdev(bdev);
	488	invalidate_bdev(bdev);
	489	return 0;
	490	}
d30a2605	491
d8e4bb81 CH	492	static int blkdev_roset(struct block_device *bdev, fmode_t mode,
	493	unsigned cmd, unsigned long arg)
	494	{
	495	int ret, n;
d30a2605	496
d8e4bb81 CH	497	ret = __blkdev_driver_ioctl(bdev, mode, cmd, arg);
	498	if (!is_unrecognized_ioctl(ret))
	499	return ret;
	500	if (!capable(CAP_SYS_ADMIN))
	501	return -EACCES;
	502	if (get_user(n, (int __user *)arg))
	503	return -EFAULT;
	504	set_device_ro(bdev, n);
	505	return 0;
	506	}
d30a2605	507
d8e4bb81 CH	508	static int blkdev_getgeo(struct block_device *bdev,
	509	struct hd_geometry __user *argp)
	510	{
	511	struct gendisk *disk = bdev->bd_disk;
	512	struct hd_geometry geo;
	513	int ret;
d30a2605	514
d8e4bb81 CH	515	if (!argp)
	516	return -EINVAL;
	517	if (!disk->fops->getgeo)
	518	return -ENOTTY;
	519
	520	/*
	521	* We need to set the startsect first, the driver may
	522	* want to override it.
	523	*/
	524	memset(&geo, 0, sizeof(geo));
	525	geo.start = get_start_sect(bdev);
	526	ret = disk->fops->getgeo(bdev, &geo);
	527	if (ret)
	528	return ret;
	529	if (copy_to_user(argp, &geo, sizeof(geo)))
	530	return -EFAULT;
	531	return 0;
	532	}
66ba32dc	533
d8e4bb81 CH	534	/* set the logical block size */
	535	static int blkdev_bszset(struct block_device *bdev, fmode_t mode,
	536	int __user *argp)
	537	{
	538	int ret, n;
66ba32dc	539
d8e4bb81 CH	540	if (!capable(CAP_SYS_ADMIN))
	541	return -EACCES;
	542	if (!argp)
	543	return -EINVAL;
	544	if (get_user(n, argp))
	545	return -EFAULT;
66ba32dc	546
d8e4bb81 CH	547	if (!(mode & FMODE_EXCL)) {
	548	bdgrab(bdev);
	549	if (blkdev_get(bdev, mode \| FMODE_EXCL, &bdev) < 0)
	550	return -EBUSY;
66ba32dc	551	}
d30a2605	552
d8e4bb81 CH	553	ret = set_blocksize(bdev, n);
	554	if (!(mode & FMODE_EXCL))
	555	blkdev_put(bdev, mode \| FMODE_EXCL);
	556	return ret;
	557	}
a885c8c4	558
d8e4bb81 CH	559	/*
	560	* always keep this in sync with compat_blkdev_ioctl()
	561	*/
	562	int blkdev_ioctl(struct block_device *bdev, fmode_t mode, unsigned cmd,
	563	unsigned long arg)
	564	{
	565	struct backing_dev_info *bdi;
	566	void __user argp = (void __user )arg;
	567	loff_t size;
	568	unsigned int max_sectors;
	569
	570	switch (cmd) {
	571	case BLKFLSBUF:
	572	return blkdev_flushbuf(bdev, mode, cmd, arg);
	573	case BLKROSET:
	574	return blkdev_roset(bdev, mode, cmd, arg);
	575	case BLKDISCARD:
	576	return blk_ioctl_discard(bdev, mode, arg, 0);
	577	case BLKSECDISCARD:
	578	return blk_ioctl_discard(bdev, mode, arg,
	579	BLKDEV_DISCARD_SECURE);
	580	case BLKZEROOUT:
	581	return blk_ioctl_zeroout(bdev, mode, arg);
	582	case HDIO_GETGEO:
	583	return blkdev_getgeo(bdev, argp);
45048d09 AV	584	case BLKRAGET:
	585	case BLKFRAGET:
	586	if (!arg)
	587	return -EINVAL;
	588	bdi = blk_get_backing_dev_info(bdev);
45048d09 AV	589	return put_long(arg, (bdi->ra_pages * PAGE_CACHE_SIZE) / 512);
	590	case BLKROGET:
	591	return put_int(arg, bdev_read_only(bdev) != 0);
ac481c20	592	case BLKBSZGET: /* get block device soft block size (cf. BLKSSZGET) */
45048d09	593	return put_int(arg, block_size(bdev));
ac481c20	594	case BLKSSZGET: /* get block device logical block size */
e1defc4f	595	return put_int(arg, bdev_logical_block_size(bdev));
ac481c20 MP	596	case BLKPBSZGET: /* get block device physical block size */
	597	return put_uint(arg, bdev_physical_block_size(bdev));
	598	case BLKIOMIN:
	599	return put_uint(arg, bdev_io_min(bdev));
	600	case BLKIOOPT:
	601	return put_uint(arg, bdev_io_opt(bdev));
	602	case BLKALIGNOFF:
	603	return put_int(arg, bdev_alignment_offset(bdev));
98262f27 MP	604	case BLKDISCARDZEROES:
98262f27 MP	605	return put_uint(arg, bdev_discard_zeroes_data(bdev));
45048d09	606	case BLKSECTGET:
63f26496 AM	607	max_sectors = min_t(unsigned int, USHRT_MAX,
	608	queue_max_sectors(bdev_get_queue(bdev)));
	609	return put_ushort(arg, max_sectors);
ef00f59c MP	610	case BLKROTATIONAL:
ef00f59c MP	611	return put_ushort(arg, !blk_queue_nonrot(bdev_get_queue(bdev)));
45048d09 AV	612	case BLKRASET:
	613	case BLKFRASET:
	614	if(!capable(CAP_SYS_ADMIN))
	615	return -EACCES;
	616	bdi = blk_get_backing_dev_info(bdev);
45048d09	617	bdi->ra_pages = (arg * 512) / PAGE_CACHE_SIZE;
45048d09 AV	618	return 0;
45048d09 AV	619	case BLKBSZSET:
d8e4bb81	620	return blkdev_bszset(bdev, mode, argp);
45048d09	621	case BLKPG:
d8e4bb81	622	return blkpg_ioctl(bdev, argp);
45048d09	623	case BLKRRPART:
d8e4bb81	624	return blkdev_reread_part(bdev);
45048d09	625	case BLKGETSIZE:
77304d2a	626	size = i_size_read(bdev->bd_inode);
45048d09 AV	627	if ((size >> 9) > ~0UL)
	628	return -EFBIG;
	629	return put_ulong(arg, size >> 9);
	630	case BLKGETSIZE64:
77304d2a	631	return put_u64(arg, i_size_read(bdev->bd_inode));
45048d09 AV	632	case BLKTRACESTART:
	633	case BLKTRACESTOP:
	634	case BLKTRACESETUP:
	635	case BLKTRACETEARDOWN:
d8e4bb81	636	return blk_trace_ioctl(bdev, cmd, argp);
5a023cdb DW	637	case BLKDAXSET:
	638	return blkdev_daxset(bdev, arg);
	639	case BLKDAXGET:
	640	return put_int(arg, !!(bdev->bd_inode->i_flags & S_DAX));
	641	break;
bbd3e064 CH	642	case IOC_PR_REGISTER:
	643	return blkdev_pr_register(bdev, argp);
	644	case IOC_PR_RESERVE:
	645	return blkdev_pr_reserve(bdev, argp);
	646	case IOC_PR_RELEASE:
	647	return blkdev_pr_release(bdev, argp);
	648	case IOC_PR_PREEMPT:
	649	return blkdev_pr_preempt(bdev, argp, false);
	650	case IOC_PR_PREEMPT_ABORT:
	651	return blkdev_pr_preempt(bdev, argp, true);
	652	case IOC_PR_CLEAR:
	653	return blkdev_pr_clear(bdev, argp);
45048d09	654	default:
d8e4bb81	655	return __blkdev_driver_ioctl(bdev, mode, cmd, arg);
45048d09	656	}
1da177e4	657	}
68f66feb	658	EXPORT_SYMBOL_GPL(blkdev_ioctl);