[linux-2.6-block.git] / block / fops.c

// SPDX-License-Identifier: GPL-2.0-only
/*
 * Copyright (C) 1991, 1992  Linus Torvalds
 * Copyright (C) 2001  Andrea Arcangeli <andrea@suse.de> SuSE
 * Copyright (C) 2016 - 2020 Christoph Hellwig
 */
#include <linux/init.h>
#include <linux/mm.h>
#include <linux/blkdev.h>
#include <linux/buffer_head.h>
#include <linux/mpage.h>
#include <linux/uio.h>
#include <linux/namei.h>
#include <linux/task_io_accounting_ops.h>
#include <linux/falloc.h>
#include <linux/suspend.h>
#include <linux/fs.h>
#include "blk.h"

static inline struct inode *bdev_file_inode(struct file *file)
{
	return file->f_mapping->host;
}

static int blkdev_get_block(struct inode *inode, sector_t iblock,
		struct buffer_head *bh, int create)
{
	bh->b_bdev = I_BDEV(inode);
	bh->b_blocknr = iblock;
	set_buffer_mapped(bh);
	return 0;
}

static unsigned int dio_bio_write_op(struct kiocb *iocb)
{
	unsigned int op = REQ_OP_WRITE | REQ_SYNC | REQ_IDLE;

	/* avoid the need for a I/O completion work item */
	if (iocb->ki_flags & IOCB_DSYNC)
		op |= REQ_FUA;
	return op;
}

#define DIO_INLINE_BIO_VECS 4

static void blkdev_bio_end_io_simple(struct bio *bio)
{
	struct task_struct *waiter = bio->bi_private;

	WRITE_ONCE(bio->bi_private, NULL);
	blk_wake_io_task(waiter);
}

static ssize_t __blkdev_direct_IO_simple(struct kiocb *iocb,
		struct iov_iter *iter, unsigned int nr_pages)
{
	struct block_device *bdev = iocb->ki_filp->private_data;
	struct bio_vec inline_vecs[DIO_INLINE_BIO_VECS], *vecs;
	loff_t pos = iocb->ki_pos;
	bool should_dirty = false;
	struct bio bio;
	ssize_t ret;

	if ((pos | iov_iter_alignment(iter)) &
	    (bdev_logical_block_size(bdev) - 1))
		return -EINVAL;

	if (nr_pages <= DIO_INLINE_BIO_VECS)
		vecs = inline_vecs;
	else {
		vecs = kmalloc_array(nr_pages, sizeof(struct bio_vec),
				     GFP_KERNEL);
		if (!vecs)
			return -ENOMEM;
	}

	bio_init(&bio, vecs, nr_pages);
	bio_set_dev(&bio, bdev);
	bio.bi_iter.bi_sector = pos >> 9;
	bio.bi_write_hint = iocb->ki_hint;
	bio.bi_private = current;
	bio.bi_end_io = blkdev_bio_end_io_simple;
	bio.bi_ioprio = iocb->ki_ioprio;

	ret = bio_iov_iter_get_pages(&bio, iter);
	if (unlikely(ret))
		goto out;
	ret = bio.bi_iter.bi_size;

	if (iov_iter_rw(iter) == READ) {
		bio.bi_opf = REQ_OP_READ;
		if (iter_is_iovec(iter))
			should_dirty = true;
	} else {
		bio.bi_opf = dio_bio_write_op(iocb);
		task_io_account_write(ret);
	}
	if (iocb->ki_flags & IOCB_NOWAIT)
		bio.bi_opf |= REQ_NOWAIT;
	if (iocb->ki_flags & IOCB_HIPRI)
		bio_set_polled(&bio, iocb);

	submit_bio(&bio);
	for (;;) {
		set_current_state(TASK_UNINTERRUPTIBLE);
		if (!READ_ONCE(bio.bi_private))
			break;
		if (!(iocb->ki_flags & IOCB_HIPRI) || !bio_poll(&bio, 0))
			blk_io_schedule();
	}
	__set_current_state(TASK_RUNNING);

	bio_release_pages(&bio, should_dirty);
	if (unlikely(bio.bi_status))
		ret = blk_status_to_errno(bio.bi_status);

out:
	if (vecs != inline_vecs)
		kfree(vecs);

	bio_uninit(&bio);

	return ret;
}

struct blkdev_dio {
	union {
		struct kiocb		*iocb;
		struct task_struct	*waiter;
	};
	size_t			size;
	atomic_t		ref;
	bool			multi_bio : 1;
	bool			should_dirty : 1;
	bool			is_sync : 1;
	struct bio		bio;
};

static struct bio_set blkdev_dio_pool;

static void blkdev_bio_end_io(struct bio *bio)
{
	struct blkdev_dio *dio = bio->bi_private;
	bool should_dirty = dio->should_dirty;

	if (bio->bi_status && !dio->bio.bi_status)
		dio->bio.bi_status = bio->bi_status;

	if (!dio->multi_bio || atomic_dec_and_test(&dio->ref)) {
		if (!dio->is_sync) {
			struct kiocb *iocb = dio->iocb;
			ssize_t ret;

			WRITE_ONCE(iocb->private, NULL);

			if (likely(!dio->bio.bi_status)) {
				ret = dio->size;
				iocb->ki_pos += ret;
			} else {
				ret = blk_status_to_errno(dio->bio.bi_status);
			}

			dio->iocb->ki_complete(iocb, ret, 0);
			if (dio->multi_bio)
				bio_put(&dio->bio);
		} else {
			struct task_struct *waiter = dio->waiter;

			WRITE_ONCE(dio->waiter, NULL);
			blk_wake_io_task(waiter);
		}
	}

	if (should_dirty) {
		bio_check_pages_dirty(bio);
	} else {
		bio_release_pages(bio, false);
		bio_put(bio);
	}
}

static ssize_t __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter,
		unsigned int nr_pages)
{
	struct block_device *bdev = iocb->ki_filp->private_data;
	struct blk_plug plug;
	struct blkdev_dio *dio;
	struct bio *bio;
	bool do_poll = (iocb->ki_flags & IOCB_HIPRI);
	bool is_read = (iov_iter_rw(iter) == READ), is_sync;
	loff_t pos = iocb->ki_pos;
	int ret = 0;

	if ((pos | iov_iter_alignment(iter)) &
	    (bdev_logical_block_size(bdev) - 1))
		return -EINVAL;

	bio = bio_alloc_kiocb(iocb, nr_pages, &blkdev_dio_pool);

	dio = container_of(bio, struct blkdev_dio, bio);
	dio->is_sync = is_sync = is_sync_kiocb(iocb);
	if (dio->is_sync) {
		dio->waiter = current;
		bio_get(bio);
	} else {
		dio->iocb = iocb;
	}

	dio->size = 0;
	dio->multi_bio = false;
	dio->should_dirty = is_read && iter_is_iovec(iter);

	/*
	 * Don't plug for HIPRI/polled IO, as those should go straight
	 * to issue
	 */
	if (!(iocb->ki_flags & IOCB_HIPRI))
		blk_start_plug(&plug);

	for (;;) {
		bio_set_dev(bio, bdev);
		bio->bi_iter.bi_sector = pos >> 9;
		bio->bi_write_hint = iocb->ki_hint;
		bio->bi_private = dio;
		bio->bi_end_io = blkdev_bio_end_io;
		bio->bi_ioprio = iocb->ki_ioprio;

		ret = bio_iov_iter_get_pages(bio, iter);
		if (unlikely(ret)) {
			bio->bi_status = BLK_STS_IOERR;
			bio_endio(bio);
			break;
		}

		if (is_read) {
			bio->bi_opf = REQ_OP_READ;
			if (dio->should_dirty)
				bio_set_pages_dirty(bio);
		} else {
			bio->bi_opf = dio_bio_write_op(iocb);
			task_io_account_write(bio->bi_iter.bi_size);
		}
		if (iocb->ki_flags & IOCB_NOWAIT)
			bio->bi_opf |= REQ_NOWAIT;

		dio->size += bio->bi_iter.bi_size;
		pos += bio->bi_iter.bi_size;

		nr_pages = bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS);
		if (!nr_pages) {
			if (do_poll)
				bio_set_polled(bio, iocb);
			submit_bio(bio);
			if (do_poll)
				WRITE_ONCE(iocb->private, bio);
			break;
		}
		if (!dio->multi_bio) {
			/*
			 * AIO needs an extra reference to ensure the dio
			 * structure which is embedded into the first bio
			 * stays around.
			 */
			if (!is_sync)
				bio_get(bio);
			dio->multi_bio = true;
			atomic_set(&dio->ref, 2);
			do_poll = false;
		} else {
			atomic_inc(&dio->ref);
		}

		submit_bio(bio);
		bio = bio_alloc(GFP_KERNEL, nr_pages);
	}

	if (!(iocb->ki_flags & IOCB_HIPRI))
		blk_finish_plug(&plug);

	if (!is_sync)
		return -EIOCBQUEUED;

	for (;;) {
		set_current_state(TASK_UNINTERRUPTIBLE);
		if (!READ_ONCE(dio->waiter))
			break;

		if (!do_poll || !bio_poll(bio, 0))
			blk_io_schedule();
	}
	__set_current_state(TASK_RUNNING);

	if (!ret)
		ret = blk_status_to_errno(dio->bio.bi_status);
	if (likely(!ret))
		ret = dio->size;

	bio_put(&dio->bio);
	return ret;
}

static ssize_t blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter)
{
	unsigned int nr_pages;

	if (!iov_iter_count(iter))
		return 0;

	nr_pages = bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS + 1);
	if (is_sync_kiocb(iocb) && nr_pages <= BIO_MAX_VECS)
		return __blkdev_direct_IO_simple(iocb, iter, nr_pages);

	return __blkdev_direct_IO(iocb, iter, bio_max_segs(nr_pages));
}

static int blkdev_writepage(struct page *page, struct writeback_control *wbc)
{
	return block_write_full_page(page, blkdev_get_block, wbc);
}

static int blkdev_readpage(struct file * file, struct page * page)
{
	return block_read_full_page(page, blkdev_get_block);
}

static void blkdev_readahead(struct readahead_control *rac)
{
	mpage_readahead(rac, blkdev_get_block);
}

static int blkdev_write_begin(struct file *file, struct address_space *mapping,
		loff_t pos, unsigned len, unsigned flags, struct page **pagep,
		void **fsdata)
{
	return block_write_begin(mapping, pos, len, flags, pagep,
				 blkdev_get_block);
}

static int blkdev_write_end(struct file *file, struct address_space *mapping,
		loff_t pos, unsigned len, unsigned copied, struct page *page,
		void *fsdata)
{
	int ret;
	ret = block_write_end(file, mapping, pos, len, copied, page, fsdata);

	unlock_page(page);
	put_page(page);

	return ret;
}

static int blkdev_writepages(struct address_space *mapping,
			     struct writeback_control *wbc)
{
	return generic_writepages(mapping, wbc);
}

const struct address_space_operations def_blk_aops = {
	.set_page_dirty	= __set_page_dirty_buffers,
	.readpage	= blkdev_readpage,
	.readahead	= blkdev_readahead,
	.writepage	= blkdev_writepage,
	.write_begin	= blkdev_write_begin,
	.write_end	= blkdev_write_end,
	.writepages	= blkdev_writepages,
	.direct_IO	= blkdev_direct_IO,
	.migratepage	= buffer_migrate_page_norefs,
	.is_dirty_writeback = buffer_check_dirty_writeback,
};

/*
 * for a block special file file_inode(file)->i_size is zero
 * so we compute the size by hand (just as in block_read/write above)
 */
static loff_t blkdev_llseek(struct file *file, loff_t offset, int whence)
{
	struct inode *bd_inode = bdev_file_inode(file);
	loff_t retval;

	inode_lock(bd_inode);
	retval = fixed_size_llseek(file, offset, whence, i_size_read(bd_inode));
	inode_unlock(bd_inode);
	return retval;
}

static int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
		int datasync)
{
	struct block_device *bdev = filp->private_data;
	int error;

	error = file_write_and_wait_range(filp, start, end);
	if (error)
		return error;

	/*
	 * There is no need to serialise calls to blkdev_issue_flush with
	 * i_mutex and doing so causes performance issues with concurrent
	 * O_SYNC writers to a block device.
	 */
	error = blkdev_issue_flush(bdev);
	if (error == -EOPNOTSUPP)
		error = 0;

	return error;
}

static int blkdev_open(struct inode *inode, struct file *filp)
{
	struct block_device *bdev;

	/*
	 * Preserve backwards compatibility and allow large file access
	 * even if userspace doesn't ask for it explicitly. Some mkfs
	 * binary needs it. We might want to drop this workaround
	 * during an unstable branch.
	 */
	filp->f_flags |= O_LARGEFILE;
	filp->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;

	if (filp->f_flags & O_NDELAY)
		filp->f_mode |= FMODE_NDELAY;
	if (filp->f_flags & O_EXCL)
		filp->f_mode |= FMODE_EXCL;
	if ((filp->f_flags & O_ACCMODE) == 3)
		filp->f_mode |= FMODE_WRITE_IOCTL;

	bdev = blkdev_get_by_dev(inode->i_rdev, filp->f_mode, filp);
	if (IS_ERR(bdev))
		return PTR_ERR(bdev);

	filp->private_data = bdev;
	filp->f_mapping = bdev->bd_inode->i_mapping;
	filp->f_wb_err = filemap_sample_wb_err(filp->f_mapping);
	return 0;
}

static int blkdev_close(struct inode *inode, struct file *filp)
{
	struct block_device *bdev = filp->private_data;

	blkdev_put(bdev, filp->f_mode);
	return 0;
}

/*
 * Write data to the block device.  Only intended for the block device itself
 * and the raw driver which basically is a fake block device.
 *
 * Does not take i_mutex for the write and thus is not for general purpose
 * use.
 */
static ssize_t blkdev_write_iter(struct kiocb *iocb, struct iov_iter *from)
{
	struct block_device *bdev = iocb->ki_filp->private_data;
	struct inode *bd_inode = bdev->bd_inode;
	loff_t size = i_size_read(bd_inode);
	struct blk_plug plug;
	size_t shorted = 0;
	ssize_t ret;

	if (bdev_read_only(bdev))
		return -EPERM;

	if (IS_SWAPFILE(bd_inode) && !is_hibernate_resume_dev(bd_inode->i_rdev))
		return -ETXTBSY;

	if (!iov_iter_count(from))
		return 0;

	if (iocb->ki_pos >= size)
		return -ENOSPC;

	if ((iocb->ki_flags & (IOCB_NOWAIT | IOCB_DIRECT)) == IOCB_NOWAIT)
		return -EOPNOTSUPP;

	size -= iocb->ki_pos;
	if (iov_iter_count(from) > size) {
		shorted = iov_iter_count(from) - size;
		iov_iter_truncate(from, size);
	}

	blk_start_plug(&plug);
	ret = __generic_file_write_iter(iocb, from);
	if (ret > 0)
		ret = generic_write_sync(iocb, ret);
	iov_iter_reexpand(from, iov_iter_count(from) + shorted);
	blk_finish_plug(&plug);
	return ret;
}

static ssize_t blkdev_read_iter(struct kiocb *iocb, struct iov_iter *to)
{
	struct block_device *bdev = iocb->ki_filp->private_data;
	loff_t size = i_size_read(bdev->bd_inode);
	loff_t pos = iocb->ki_pos;
	size_t shorted = 0;
	ssize_t ret;

	if (pos >= size)
		return 0;

	size -= pos;
	if (iov_iter_count(to) > size) {
		shorted = iov_iter_count(to) - size;
		iov_iter_truncate(to, size);
	}

	ret = generic_file_read_iter(iocb, to);
	iov_iter_reexpand(to, iov_iter_count(to) + shorted);
	return ret;
}

#define	BLKDEV_FALLOC_FL_SUPPORTED					\
		(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE |		\
		 FALLOC_FL_ZERO_RANGE | FALLOC_FL_NO_HIDE_STALE)

static long blkdev_fallocate(struct file *file, int mode, loff_t start,
			     loff_t len)
{
	struct inode *inode = bdev_file_inode(file);
	struct block_device *bdev = I_BDEV(inode);
	loff_t end = start + len - 1;
	loff_t isize;
	int error;

	/* Fail if we don't recognize the flags. */
	if (mode & ~BLKDEV_FALLOC_FL_SUPPORTED)
		return -EOPNOTSUPP;

	/* Don't go off the end of the device. */
	isize = i_size_read(bdev->bd_inode);
	if (start >= isize)
		return -EINVAL;
	if (end >= isize) {
		if (mode & FALLOC_FL_KEEP_SIZE) {
			len = isize - start;
			end = start + len - 1;
		} else
			return -EINVAL;
	}

	/*
	 * Don't allow IO that isn't aligned to logical block size.
	 */
	if ((start | len) & (bdev_logical_block_size(bdev) - 1))
		return -EINVAL;

	filemap_invalidate_lock(inode->i_mapping);

	/* Invalidate the page cache, including dirty pages. */
	error = truncate_bdev_range(bdev, file->f_mode, start, end);
	if (error)
		goto fail;

	switch (mode) {
	case FALLOC_FL_ZERO_RANGE:
	case FALLOC_FL_ZERO_RANGE | FALLOC_FL_KEEP_SIZE:
		error = blkdev_issue_zeroout(bdev, start >> 9, len >> 9,
					    GFP_KERNEL, BLKDEV_ZERO_NOUNMAP);
		break;
	case FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE:
		error = blkdev_issue_zeroout(bdev, start >> 9, len >> 9,
					     GFP_KERNEL, BLKDEV_ZERO_NOFALLBACK);
		break;
	case FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE | FALLOC_FL_NO_HIDE_STALE:
		error = blkdev_issue_discard(bdev, start >> 9, len >> 9,
					     GFP_KERNEL, 0);
		break;
	default:
		error = -EOPNOTSUPP;
	}

 fail:
	filemap_invalidate_unlock(inode->i_mapping);
	return error;
}

const struct file_operations def_blk_fops = {
	.open		= blkdev_open,
	.release	= blkdev_close,
	.llseek		= blkdev_llseek,
	.read_iter	= blkdev_read_iter,
	.write_iter	= blkdev_write_iter,
	.iopoll		= iocb_bio_iopoll,
	.mmap		= generic_file_mmap,
	.fsync		= blkdev_fsync,
	.unlocked_ioctl	= blkdev_ioctl,
#ifdef CONFIG_COMPAT
	.compat_ioctl	= compat_blkdev_ioctl,
#endif
	.splice_read	= generic_file_splice_read,
	.splice_write	= iter_file_splice_write,
	.fallocate	= blkdev_fallocate,
};

static __init int blkdev_init(void)
{
	return bioset_init(&blkdev_dio_pool, 4,
				offsetof(struct blkdev_dio, bio),
				BIOSET_NEED_BVECS|BIOSET_PERCPU_CACHE);
}
module_init(blkdev_init);
Commit	Line	Data
cd82cca7 CH	1	// SPDX-License-Identifier: GPL-2.0-only
	2	/*
	3	* Copyright (C) 1991, 1992 Linus Torvalds
	4	* Copyright (C) 2001 Andrea Arcangeli <andrea@suse.de> SuSE
	5	* Copyright (C) 2016 - 2020 Christoph Hellwig
	6	*/
	7	#include <linux/init.h>
	8	#include <linux/mm.h>
	9	#include <linux/blkdev.h>
	10	#include <linux/buffer_head.h>
	11	#include <linux/mpage.h>
	12	#include <linux/uio.h>
	13	#include <linux/namei.h>
	14	#include <linux/task_io_accounting_ops.h>
	15	#include <linux/falloc.h>
	16	#include <linux/suspend.h>
f278eb3d	17	#include <linux/fs.h>
cd82cca7 CH	18	#include "blk.h"
cd82cca7 CH	19
fac7c6d5	20	static inline struct inode bdev_file_inode(struct file file)
cd82cca7 CH	21	{
	22	return file->f_mapping->host;
	23	}
	24
	25	static int blkdev_get_block(struct inode *inode, sector_t iblock,
	26	struct buffer_head *bh, int create)
	27	{
	28	bh->b_bdev = I_BDEV(inode);
	29	bh->b_blocknr = iblock;
	30	set_buffer_mapped(bh);
	31	return 0;
	32	}
	33
	34	static unsigned int dio_bio_write_op(struct kiocb *iocb)
	35	{
	36	unsigned int op = REQ_OP_WRITE \| REQ_SYNC \| REQ_IDLE;
	37
	38	/* avoid the need for a I/O completion work item */
	39	if (iocb->ki_flags & IOCB_DSYNC)
	40	op \|= REQ_FUA;
	41	return op;
	42	}
	43
	44	#define DIO_INLINE_BIO_VECS 4
	45
	46	static void blkdev_bio_end_io_simple(struct bio *bio)
	47	{
	48	struct task_struct *waiter = bio->bi_private;
	49
	50	WRITE_ONCE(bio->bi_private, NULL);
	51	blk_wake_io_task(waiter);
	52	}
	53
	54	static ssize_t __blkdev_direct_IO_simple(struct kiocb *iocb,
	55	struct iov_iter *iter, unsigned int nr_pages)
	56	{
fac7c6d5	57	struct block_device *bdev = iocb->ki_filp->private_data;
cd82cca7 CH	58	struct bio_vec inline_vecs[DIO_INLINE_BIO_VECS], *vecs;
	59	loff_t pos = iocb->ki_pos;
	60	bool should_dirty = false;
	61	struct bio bio;
	62	ssize_t ret;
cd82cca7 CH	63
	64	if ((pos \| iov_iter_alignment(iter)) &
	65	(bdev_logical_block_size(bdev) - 1))
	66	return -EINVAL;
	67
	68	if (nr_pages <= DIO_INLINE_BIO_VECS)
	69	vecs = inline_vecs;
	70	else {
	71	vecs = kmalloc_array(nr_pages, sizeof(struct bio_vec),
	72	GFP_KERNEL);
	73	if (!vecs)
	74	return -ENOMEM;
	75	}
	76
	77	bio_init(&bio, vecs, nr_pages);
	78	bio_set_dev(&bio, bdev);
	79	bio.bi_iter.bi_sector = pos >> 9;
	80	bio.bi_write_hint = iocb->ki_hint;
	81	bio.bi_private = current;
	82	bio.bi_end_io = blkdev_bio_end_io_simple;
	83	bio.bi_ioprio = iocb->ki_ioprio;
	84
	85	ret = bio_iov_iter_get_pages(&bio, iter);
	86	if (unlikely(ret))
	87	goto out;
	88	ret = bio.bi_iter.bi_size;
	89
	90	if (iov_iter_rw(iter) == READ) {
	91	bio.bi_opf = REQ_OP_READ;
	92	if (iter_is_iovec(iter))
	93	should_dirty = true;
	94	} else {
	95	bio.bi_opf = dio_bio_write_op(iocb);
	96	task_io_account_write(ret);
	97	}
	98	if (iocb->ki_flags & IOCB_NOWAIT)
	99	bio.bi_opf \|= REQ_NOWAIT;
	100	if (iocb->ki_flags & IOCB_HIPRI)
	101	bio_set_polled(&bio, iocb);
	102
3e08773c	103	submit_bio(&bio);
cd82cca7 CH	104	for (;;) {
	105	set_current_state(TASK_UNINTERRUPTIBLE);
	106	if (!READ_ONCE(bio.bi_private))
	107	break;
3e08773c	108	if (!(iocb->ki_flags & IOCB_HIPRI) \|\| !bio_poll(&bio, 0))
cd82cca7 CH	109	blk_io_schedule();
	110	}
	111	__set_current_state(TASK_RUNNING);
	112
	113	bio_release_pages(&bio, should_dirty);
	114	if (unlikely(bio.bi_status))
	115	ret = blk_status_to_errno(bio.bi_status);
	116
	117	out:
	118	if (vecs != inline_vecs)
	119	kfree(vecs);
	120
	121	bio_uninit(&bio);
	122
	123	return ret;
	124	}
	125
	126	struct blkdev_dio {
	127	union {
	128	struct kiocb *iocb;
	129	struct task_struct *waiter;
	130	};
	131	size_t size;
	132	atomic_t ref;
	133	bool multi_bio : 1;
	134	bool should_dirty : 1;
	135	bool is_sync : 1;
	136	struct bio bio;
	137	};
	138
	139	static struct bio_set blkdev_dio_pool;
	140
cd82cca7 CH	141	static void blkdev_bio_end_io(struct bio *bio)
	142	{
	143	struct blkdev_dio *dio = bio->bi_private;
	144	bool should_dirty = dio->should_dirty;
	145
	146	if (bio->bi_status && !dio->bio.bi_status)
	147	dio->bio.bi_status = bio->bi_status;
	148
	149	if (!dio->multi_bio \|\| atomic_dec_and_test(&dio->ref)) {
	150	if (!dio->is_sync) {
	151	struct kiocb *iocb = dio->iocb;
	152	ssize_t ret;
	153
3e08773c CH	154	WRITE_ONCE(iocb->private, NULL);
3e08773c CH	155
cd82cca7 CH	156	if (likely(!dio->bio.bi_status)) {
	157	ret = dio->size;
	158	iocb->ki_pos += ret;
	159	} else {
	160	ret = blk_status_to_errno(dio->bio.bi_status);
	161	}
	162
	163	dio->iocb->ki_complete(iocb, ret, 0);
	164	if (dio->multi_bio)
	165	bio_put(&dio->bio);
	166	} else {
	167	struct task_struct *waiter = dio->waiter;
	168
	169	WRITE_ONCE(dio->waiter, NULL);
	170	blk_wake_io_task(waiter);
	171	}
	172	}
	173
	174	if (should_dirty) {
	175	bio_check_pages_dirty(bio);
	176	} else {
	177	bio_release_pages(bio, false);
	178	bio_put(bio);
	179	}
	180	}
	181
	182	static ssize_t __blkdev_direct_IO(struct kiocb iocb, struct iov_iter iter,
	183	unsigned int nr_pages)
	184	{
fac7c6d5	185	struct block_device *bdev = iocb->ki_filp->private_data;
cd82cca7 CH	186	struct blk_plug plug;
	187	struct blkdev_dio *dio;
	188	struct bio *bio;
71fc3f5e	189	bool do_poll = (iocb->ki_flags & IOCB_HIPRI);
cd82cca7 CH	190	bool is_read = (iov_iter_rw(iter) == READ), is_sync;
cd82cca7 CH	191	loff_t pos = iocb->ki_pos;
cd82cca7 CH	192	int ret = 0;
	193
	194	if ((pos \| iov_iter_alignment(iter)) &
	195	(bdev_logical_block_size(bdev) - 1))
	196	return -EINVAL;
	197
	198	bio = bio_alloc_kiocb(iocb, nr_pages, &blkdev_dio_pool);
	199
	200	dio = container_of(bio, struct blkdev_dio, bio);
	201	dio->is_sync = is_sync = is_sync_kiocb(iocb);
	202	if (dio->is_sync) {
	203	dio->waiter = current;
	204	bio_get(bio);
	205	} else {
	206	dio->iocb = iocb;
	207	}
	208
	209	dio->size = 0;
	210	dio->multi_bio = false;
	211	dio->should_dirty = is_read && iter_is_iovec(iter);
	212
	213	/*
	214	* Don't plug for HIPRI/polled IO, as those should go straight
	215	* to issue
	216	*/
71fc3f5e	217	if (!(iocb->ki_flags & IOCB_HIPRI))
cd82cca7 CH	218	blk_start_plug(&plug);
	219
	220	for (;;) {
	221	bio_set_dev(bio, bdev);
	222	bio->bi_iter.bi_sector = pos >> 9;
	223	bio->bi_write_hint = iocb->ki_hint;
	224	bio->bi_private = dio;
	225	bio->bi_end_io = blkdev_bio_end_io;
	226	bio->bi_ioprio = iocb->ki_ioprio;
	227
	228	ret = bio_iov_iter_get_pages(bio, iter);
	229	if (unlikely(ret)) {
	230	bio->bi_status = BLK_STS_IOERR;
	231	bio_endio(bio);
	232	break;
	233	}
	234
	235	if (is_read) {
	236	bio->bi_opf = REQ_OP_READ;
	237	if (dio->should_dirty)
	238	bio_set_pages_dirty(bio);
	239	} else {
	240	bio->bi_opf = dio_bio_write_op(iocb);
	241	task_io_account_write(bio->bi_iter.bi_size);
	242	}
	243	if (iocb->ki_flags & IOCB_NOWAIT)
	244	bio->bi_opf \|= REQ_NOWAIT;
	245
	246	dio->size += bio->bi_iter.bi_size;
	247	pos += bio->bi_iter.bi_size;
	248
	249	nr_pages = bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS);
	250	if (!nr_pages) {
71fc3f5e	251	if (do_poll)
cd82cca7	252	bio_set_polled(bio, iocb);
3e08773c	253	submit_bio(bio);
71fc3f5e	254	if (do_poll)
3e08773c	255	WRITE_ONCE(iocb->private, bio);
cd82cca7 CH	256	break;
cd82cca7 CH	257	}
cd82cca7 CH	258	if (!dio->multi_bio) {
	259	/*
	260	* AIO needs an extra reference to ensure the dio
	261	* structure which is embedded into the first bio
	262	* stays around.
	263	*/
	264	if (!is_sync)
	265	bio_get(bio);
	266	dio->multi_bio = true;
	267	atomic_set(&dio->ref, 2);
71fc3f5e	268	do_poll = false;
cd82cca7 CH	269	} else {
	270	atomic_inc(&dio->ref);
	271	}
	272
	273	submit_bio(bio);
	274	bio = bio_alloc(GFP_KERNEL, nr_pages);
	275	}
	276
71fc3f5e	277	if (!(iocb->ki_flags & IOCB_HIPRI))
cd82cca7 CH	278	blk_finish_plug(&plug);
	279
	280	if (!is_sync)
	281	return -EIOCBQUEUED;
	282
	283	for (;;) {
	284	set_current_state(TASK_UNINTERRUPTIBLE);
	285	if (!READ_ONCE(dio->waiter))
	286	break;
	287
3e08773c	288	if (!do_poll \|\| !bio_poll(bio, 0))
cd82cca7 CH	289	blk_io_schedule();
	290	}
	291	__set_current_state(TASK_RUNNING);
	292
	293	if (!ret)
	294	ret = blk_status_to_errno(dio->bio.bi_status);
	295	if (likely(!ret))
	296	ret = dio->size;
	297
	298	bio_put(&dio->bio);
	299	return ret;
	300	}
	301
	302	static ssize_t blkdev_direct_IO(struct kiocb iocb, struct iov_iter iter)
	303	{
	304	unsigned int nr_pages;
	305
	306	if (!iov_iter_count(iter))
	307	return 0;
	308
	309	nr_pages = bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS + 1);
	310	if (is_sync_kiocb(iocb) && nr_pages <= BIO_MAX_VECS)
	311	return __blkdev_direct_IO_simple(iocb, iter, nr_pages);
	312
	313	return __blkdev_direct_IO(iocb, iter, bio_max_segs(nr_pages));
	314	}
	315
	316	static int blkdev_writepage(struct page page, struct writeback_control wbc)
	317	{
	318	return block_write_full_page(page, blkdev_get_block, wbc);
	319	}
	320
	321	static int blkdev_readpage(struct file * file, struct page * page)
	322	{
	323	return block_read_full_page(page, blkdev_get_block);
	324	}
	325
	326	static void blkdev_readahead(struct readahead_control *rac)
	327	{
	328	mpage_readahead(rac, blkdev_get_block);
	329	}
	330
	331	static int blkdev_write_begin(struct file file, struct address_space mapping,
	332	loff_t pos, unsigned len, unsigned flags, struct page **pagep,
	333	void **fsdata)
	334	{
	335	return block_write_begin(mapping, pos, len, flags, pagep,
	336	blkdev_get_block);
	337	}
	338
	339	static int blkdev_write_end(struct file file, struct address_space mapping,
	340	loff_t pos, unsigned len, unsigned copied, struct page *page,
	341	void *fsdata)
	342	{
	343	int ret;
	344	ret = block_write_end(file, mapping, pos, len, copied, page, fsdata);
	345
	346	unlock_page(page);
	347	put_page(page);
	348
	349	return ret;
	350	}
	351
	352	static int blkdev_writepages(struct address_space *mapping,
353	struct writeback_control *wbc)
354	{
355	return generic_writepages(mapping, wbc);
356	}
357
358	const struct address_space_operations def_blk_aops = {
359	.set_page_dirty = __set_page_dirty_buffers,
360	.readpage = blkdev_readpage,
361	.readahead = blkdev_readahead,
362	.writepage = blkdev_writepage,
363	.write_begin = blkdev_write_begin,
364	.write_end = blkdev_write_end,
365	.writepages = blkdev_writepages,
366	.direct_IO = blkdev_direct_IO,
367	.migratepage = buffer_migrate_page_norefs,
368	.is_dirty_writeback = buffer_check_dirty_writeback,
369	};
370
371	/*
372	* for a block special file file_inode(file)->i_size is zero
373	* so we compute the size by hand (just as in block_read/write above)
374	*/
375	static loff_t blkdev_llseek(struct file *file, loff_t offset, int whence)
376	{
377	struct inode *bd_inode = bdev_file_inode(file);
378	loff_t retval;
379
380	inode_lock(bd_inode);
381	retval = fixed_size_llseek(file, offset, whence, i_size_read(bd_inode));
382	inode_unlock(bd_inode);
383	return retval;
384	}
385
386	static int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
387	int datasync)
388	{
fac7c6d5	389	struct block_device *bdev = filp->private_data;
cd82cca7 CH	390	int error;
	391
	392	error = file_write_and_wait_range(filp, start, end);
	393	if (error)
	394	return error;
	395
	396	/*
	397	* There is no need to serialise calls to blkdev_issue_flush with
	398	* i_mutex and doing so causes performance issues with concurrent
	399	* O_SYNC writers to a block device.
	400	*/
	401	error = blkdev_issue_flush(bdev);
	402	if (error == -EOPNOTSUPP)
	403	error = 0;
	404
	405	return error;
	406	}
	407
	408	static int blkdev_open(struct inode inode, struct file filp)
	409	{
	410	struct block_device *bdev;
	411
	412	/*
	413	* Preserve backwards compatibility and allow large file access
	414	* even if userspace doesn't ask for it explicitly. Some mkfs
	415	* binary needs it. We might want to drop this workaround
	416	* during an unstable branch.
	417	*/
	418	filp->f_flags \|= O_LARGEFILE;
	419	filp->f_mode \|= FMODE_NOWAIT \| FMODE_BUF_RASYNC;
	420
	421	if (filp->f_flags & O_NDELAY)
	422	filp->f_mode \|= FMODE_NDELAY;
	423	if (filp->f_flags & O_EXCL)
	424	filp->f_mode \|= FMODE_EXCL;
	425	if ((filp->f_flags & O_ACCMODE) == 3)
	426	filp->f_mode \|= FMODE_WRITE_IOCTL;
	427
	428	bdev = blkdev_get_by_dev(inode->i_rdev, filp->f_mode, filp);
	429	if (IS_ERR(bdev))
	430	return PTR_ERR(bdev);
fac7c6d5 PB	431
fac7c6d5 PB	432	filp->private_data = bdev;
cd82cca7 CH	433	filp->f_mapping = bdev->bd_inode->i_mapping;
	434	filp->f_wb_err = filemap_sample_wb_err(filp->f_mapping);
	435	return 0;
	436	}
	437
	438	static int blkdev_close(struct inode inode, struct file filp)
	439	{
fac7c6d5	440	struct block_device *bdev = filp->private_data;
cd82cca7 CH	441
	442	blkdev_put(bdev, filp->f_mode);
	443	return 0;
	444	}
	445
cd82cca7 CH	446	/*
	447	* Write data to the block device. Only intended for the block device itself
	448	* and the raw driver which basically is a fake block device.
	449	*
	450	* Does not take i_mutex for the write and thus is not for general purpose
	451	* use.
	452	*/
	453	static ssize_t blkdev_write_iter(struct kiocb iocb, struct iov_iter from)
	454	{
fac7c6d5 PB	455	struct block_device *bdev = iocb->ki_filp->private_data;
fac7c6d5 PB	456	struct inode *bd_inode = bdev->bd_inode;
cd82cca7 CH	457	loff_t size = i_size_read(bd_inode);
	458	struct blk_plug plug;
	459	size_t shorted = 0;
	460	ssize_t ret;
	461
fac7c6d5	462	if (bdev_read_only(bdev))
cd82cca7 CH	463	return -EPERM;
	464
	465	if (IS_SWAPFILE(bd_inode) && !is_hibernate_resume_dev(bd_inode->i_rdev))
	466	return -ETXTBSY;
	467
	468	if (!iov_iter_count(from))
	469	return 0;
	470
	471	if (iocb->ki_pos >= size)
	472	return -ENOSPC;
	473
	474	if ((iocb->ki_flags & (IOCB_NOWAIT \| IOCB_DIRECT)) == IOCB_NOWAIT)
	475	return -EOPNOTSUPP;
	476
	477	size -= iocb->ki_pos;
	478	if (iov_iter_count(from) > size) {
	479	shorted = iov_iter_count(from) - size;
	480	iov_iter_truncate(from, size);
	481	}
	482
	483	blk_start_plug(&plug);
	484	ret = __generic_file_write_iter(iocb, from);
	485	if (ret > 0)
	486	ret = generic_write_sync(iocb, ret);
	487	iov_iter_reexpand(from, iov_iter_count(from) + shorted);
	488	blk_finish_plug(&plug);
	489	return ret;
	490	}
	491
	492	static ssize_t blkdev_read_iter(struct kiocb iocb, struct iov_iter to)
	493	{
fac7c6d5 PB	494	struct block_device *bdev = iocb->ki_filp->private_data;
fac7c6d5 PB	495	loff_t size = i_size_read(bdev->bd_inode);
cd82cca7 CH	496	loff_t pos = iocb->ki_pos;
	497	size_t shorted = 0;
	498	ssize_t ret;
	499
	500	if (pos >= size)
	501	return 0;
	502
	503	size -= pos;
	504	if (iov_iter_count(to) > size) {
	505	shorted = iov_iter_count(to) - size;
	506	iov_iter_truncate(to, size);
	507	}
	508
	509	ret = generic_file_read_iter(iocb, to);
	510	iov_iter_reexpand(to, iov_iter_count(to) + shorted);
	511	return ret;
	512	}
	513
	514	#define BLKDEV_FALLOC_FL_SUPPORTED \
	515	(FALLOC_FL_KEEP_SIZE \| FALLOC_FL_PUNCH_HOLE \| \
	516	FALLOC_FL_ZERO_RANGE \| FALLOC_FL_NO_HIDE_STALE)
	517
	518	static long blkdev_fallocate(struct file *file, int mode, loff_t start,
	519	loff_t len)
	520	{
f278eb3d ML	521	struct inode *inode = bdev_file_inode(file);
f278eb3d ML	522	struct block_device *bdev = I_BDEV(inode);
cd82cca7 CH	523	loff_t end = start + len - 1;
	524	loff_t isize;
	525	int error;
	526
	527	/* Fail if we don't recognize the flags. */
	528	if (mode & ~BLKDEV_FALLOC_FL_SUPPORTED)
	529	return -EOPNOTSUPP;
	530
	531	/* Don't go off the end of the device. */
	532	isize = i_size_read(bdev->bd_inode);
	533	if (start >= isize)
	534	return -EINVAL;
	535	if (end >= isize) {
	536	if (mode & FALLOC_FL_KEEP_SIZE) {
	537	len = isize - start;
	538	end = start + len - 1;
	539	} else
	540	return -EINVAL;
	541	}
	542
	543	/*
	544	* Don't allow IO that isn't aligned to logical block size.
	545	*/
	546	if ((start \| len) & (bdev_logical_block_size(bdev) - 1))
	547	return -EINVAL;
	548
f278eb3d ML	549	filemap_invalidate_lock(inode->i_mapping);
f278eb3d ML	550
cd82cca7 CH	551	/* Invalidate the page cache, including dirty pages. */
	552	error = truncate_bdev_range(bdev, file->f_mode, start, end);
	553	if (error)
f278eb3d	554	goto fail;
cd82cca7 CH	555
	556	switch (mode) {
	557	case FALLOC_FL_ZERO_RANGE:
	558	case FALLOC_FL_ZERO_RANGE \| FALLOC_FL_KEEP_SIZE:
	559	error = blkdev_issue_zeroout(bdev, start >> 9, len >> 9,
	560	GFP_KERNEL, BLKDEV_ZERO_NOUNMAP);
	561	break;
	562	case FALLOC_FL_PUNCH_HOLE \| FALLOC_FL_KEEP_SIZE:
	563	error = blkdev_issue_zeroout(bdev, start >> 9, len >> 9,
	564	GFP_KERNEL, BLKDEV_ZERO_NOFALLBACK);
	565	break;
	566	case FALLOC_FL_PUNCH_HOLE \| FALLOC_FL_KEEP_SIZE \| FALLOC_FL_NO_HIDE_STALE:
	567	error = blkdev_issue_discard(bdev, start >> 9, len >> 9,
	568	GFP_KERNEL, 0);
	569	break;
	570	default:
f278eb3d	571	error = -EOPNOTSUPP;
cd82cca7	572	}
cd82cca7	573
f278eb3d ML	574	fail:
	575	filemap_invalidate_unlock(inode->i_mapping);
	576	return error;
cd82cca7 CH	577	}
	578
	579	const struct file_operations def_blk_fops = {
	580	.open = blkdev_open,
	581	.release = blkdev_close,
	582	.llseek = blkdev_llseek,
	583	.read_iter = blkdev_read_iter,
	584	.write_iter = blkdev_write_iter,
3e08773c	585	.iopoll = iocb_bio_iopoll,
cd82cca7 CH	586	.mmap = generic_file_mmap,
cd82cca7 CH	587	.fsync = blkdev_fsync,
8a709512	588	.unlocked_ioctl = blkdev_ioctl,
cd82cca7 CH	589	#ifdef CONFIG_COMPAT
	590	.compat_ioctl = compat_blkdev_ioctl,
	591	#endif
	592	.splice_read = generic_file_splice_read,
	593	.splice_write = iter_file_splice_write,
	594	.fallocate = blkdev_fallocate,
	595	};
	596
	597	static __init int blkdev_init(void)
	598	{
	599	return bioset_init(&blkdev_dio_pool, 4,
	600	offsetof(struct blkdev_dio, bio),
	601	BIOSET_NEED_BVECS\|BIOSET_PERCPU_CACHE);
	602	}
	603	module_init(blkdev_init);