[thirdparty/kernel/linux.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 >> SECTOR_SHIFT;
	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, NULL, 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;
}

enum {
	DIO_MULTI_BIO		= 1,
	DIO_SHOULD_DIRTY	= 2,
	DIO_IS_SYNC		= 4,
};

struct blkdev_dio {
	union {
		struct kiocb		*iocb;
		struct task_struct	*waiter;
	};
	size_t			size;
	atomic_t		ref;
	unsigned int		flags;
	struct bio		bio ____cacheline_aligned_in_smp;
};

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->flags & DIO_SHOULD_DIRTY;

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

	if (!(dio->flags & DIO_MULTI_BIO) || atomic_dec_and_test(&dio->ref)) {
		if (!(dio->flags & 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->flags & 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);
	is_sync = is_sync_kiocb(iocb);
	if (is_sync) {
		dio->flags = DIO_IS_SYNC;
		dio->waiter = current;
		bio_get(bio);
	} else {
		dio->flags = 0;
		dio->iocb = iocb;
	}

	dio->size = 0;
	if (is_read && iter_is_iovec(iter))
		dio->flags |= DIO_SHOULD_DIRTY;

	/*
	 * 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 >> SECTOR_SHIFT;
		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->flags & 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->flags & 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->flags |= DIO_MULTI_BIO;
			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, NULL, 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 (unlikely(pos + iov_iter_count(to) > size)) {
		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);

	if (unlikely(shorted))
		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 >> SECTOR_SHIFT,
					     len >> SECTOR_SHIFT, GFP_KERNEL,
					     BLKDEV_ZERO_NOUNMAP);
		break;
	case FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE:
		error = blkdev_issue_zeroout(bdev, start >> SECTOR_SHIFT,
					     len >> SECTOR_SHIFT, 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 >> SECTOR_SHIFT,
					     len >> SECTOR_SHIFT, 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);
6549a874	79	bio.bi_iter.bi_sector = pos >> SECTOR_SHIFT;
cd82cca7 CH	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;
5a72e899	108	if (!(iocb->ki_flags & IOCB_HIPRI) \|\| !bio_poll(&bio, NULL, 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
09ce8744 JA	126	enum {
	127	DIO_MULTI_BIO = 1,
	128	DIO_SHOULD_DIRTY = 2,
	129	DIO_IS_SYNC = 4,
	130	};
	131
cd82cca7 CH	132	struct blkdev_dio {
	133	union {
	134	struct kiocb *iocb;
	135	struct task_struct *waiter;
	136	};
	137	size_t size;
	138	atomic_t ref;
09ce8744	139	unsigned int flags;
6155631a	140	struct bio bio ____cacheline_aligned_in_smp;
cd82cca7 CH	141	};
	142
	143	static struct bio_set blkdev_dio_pool;
	144
cd82cca7 CH	145	static void blkdev_bio_end_io(struct bio *bio)
	146	{
	147	struct blkdev_dio *dio = bio->bi_private;
09ce8744	148	bool should_dirty = dio->flags & DIO_SHOULD_DIRTY;
cd82cca7 CH	149
	150	if (bio->bi_status && !dio->bio.bi_status)
	151	dio->bio.bi_status = bio->bi_status;
	152
09ce8744 JA	153	if (!(dio->flags & DIO_MULTI_BIO) \|\| atomic_dec_and_test(&dio->ref)) {
09ce8744 JA	154	if (!(dio->flags & DIO_IS_SYNC)) {
cd82cca7 CH	155	struct kiocb *iocb = dio->iocb;
	156	ssize_t ret;
	157
3e08773c CH	158	WRITE_ONCE(iocb->private, NULL);
3e08773c CH	159
cd82cca7 CH	160	if (likely(!dio->bio.bi_status)) {
	161	ret = dio->size;
	162	iocb->ki_pos += ret;
	163	} else {
	164	ret = blk_status_to_errno(dio->bio.bi_status);
	165	}
	166
	167	dio->iocb->ki_complete(iocb, ret, 0);
09ce8744	168	if (dio->flags & DIO_MULTI_BIO)
cd82cca7 CH	169	bio_put(&dio->bio);
	170	} else {
	171	struct task_struct *waiter = dio->waiter;
	172
	173	WRITE_ONCE(dio->waiter, NULL);
	174	blk_wake_io_task(waiter);
	175	}
	176	}
	177
	178	if (should_dirty) {
	179	bio_check_pages_dirty(bio);
	180	} else {
	181	bio_release_pages(bio, false);
	182	bio_put(bio);
	183	}
	184	}
	185
	186	static ssize_t __blkdev_direct_IO(struct kiocb iocb, struct iov_iter iter,
	187	unsigned int nr_pages)
	188	{
fac7c6d5	189	struct block_device *bdev = iocb->ki_filp->private_data;
cd82cca7 CH	190	struct blk_plug plug;
	191	struct blkdev_dio *dio;
	192	struct bio *bio;
71fc3f5e	193	bool do_poll = (iocb->ki_flags & IOCB_HIPRI);
cd82cca7 CH	194	bool is_read = (iov_iter_rw(iter) == READ), is_sync;
cd82cca7 CH	195	loff_t pos = iocb->ki_pos;
cd82cca7 CH	196	int ret = 0;
	197
	198	if ((pos \| iov_iter_alignment(iter)) &
	199	(bdev_logical_block_size(bdev) - 1))
	200	return -EINVAL;
	201
	202	bio = bio_alloc_kiocb(iocb, nr_pages, &blkdev_dio_pool);
	203
	204	dio = container_of(bio, struct blkdev_dio, bio);
09ce8744 JA	205	is_sync = is_sync_kiocb(iocb);
	206	if (is_sync) {
	207	dio->flags = DIO_IS_SYNC;
cd82cca7 CH	208	dio->waiter = current;
	209	bio_get(bio);
	210	} else {
09ce8744	211	dio->flags = 0;
cd82cca7 CH	212	dio->iocb = iocb;
	213	}
	214
	215	dio->size = 0;
09ce8744 JA	216	if (is_read && iter_is_iovec(iter))
09ce8744 JA	217	dio->flags \|= DIO_SHOULD_DIRTY;
cd82cca7 CH	218
	219	/*
	220	* Don't plug for HIPRI/polled IO, as those should go straight
	221	* to issue
	222	*/
71fc3f5e	223	if (!(iocb->ki_flags & IOCB_HIPRI))
cd82cca7 CH	224	blk_start_plug(&plug);
	225
	226	for (;;) {
	227	bio_set_dev(bio, bdev);
6549a874	228	bio->bi_iter.bi_sector = pos >> SECTOR_SHIFT;
cd82cca7 CH	229	bio->bi_write_hint = iocb->ki_hint;
	230	bio->bi_private = dio;
	231	bio->bi_end_io = blkdev_bio_end_io;
	232	bio->bi_ioprio = iocb->ki_ioprio;
	233
	234	ret = bio_iov_iter_get_pages(bio, iter);
	235	if (unlikely(ret)) {
	236	bio->bi_status = BLK_STS_IOERR;
	237	bio_endio(bio);
	238	break;
	239	}
	240
	241	if (is_read) {
	242	bio->bi_opf = REQ_OP_READ;
09ce8744	243	if (dio->flags & DIO_SHOULD_DIRTY)
cd82cca7 CH	244	bio_set_pages_dirty(bio);
	245	} else {
	246	bio->bi_opf = dio_bio_write_op(iocb);
	247	task_io_account_write(bio->bi_iter.bi_size);
	248	}
	249	if (iocb->ki_flags & IOCB_NOWAIT)
	250	bio->bi_opf \|= REQ_NOWAIT;
	251
	252	dio->size += bio->bi_iter.bi_size;
	253	pos += bio->bi_iter.bi_size;
	254
	255	nr_pages = bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS);
	256	if (!nr_pages) {
71fc3f5e	257	if (do_poll)
cd82cca7	258	bio_set_polled(bio, iocb);
3e08773c	259	submit_bio(bio);
71fc3f5e	260	if (do_poll)
3e08773c	261	WRITE_ONCE(iocb->private, bio);
cd82cca7 CH	262	break;
cd82cca7 CH	263	}
09ce8744	264	if (!(dio->flags & DIO_MULTI_BIO)) {
cd82cca7 CH	265	/*
	266	* AIO needs an extra reference to ensure the dio
	267	* structure which is embedded into the first bio
	268	* stays around.
	269	*/
	270	if (!is_sync)
	271	bio_get(bio);
09ce8744	272	dio->flags \|= DIO_MULTI_BIO;
cd82cca7	273	atomic_set(&dio->ref, 2);
71fc3f5e	274	do_poll = false;
cd82cca7 CH	275	} else {
	276	atomic_inc(&dio->ref);
	277	}
	278
	279	submit_bio(bio);
	280	bio = bio_alloc(GFP_KERNEL, nr_pages);
	281	}
	282
71fc3f5e	283	if (!(iocb->ki_flags & IOCB_HIPRI))
cd82cca7 CH	284	blk_finish_plug(&plug);
	285
	286	if (!is_sync)
	287	return -EIOCBQUEUED;
	288
	289	for (;;) {
	290	set_current_state(TASK_UNINTERRUPTIBLE);
	291	if (!READ_ONCE(dio->waiter))
	292	break;
	293
5a72e899	294	if (!do_poll \|\| !bio_poll(bio, NULL, 0))
cd82cca7 CH	295	blk_io_schedule();
	296	}
	297	__set_current_state(TASK_RUNNING);
	298
	299	if (!ret)
	300	ret = blk_status_to_errno(dio->bio.bi_status);
	301	if (likely(!ret))
	302	ret = dio->size;
	303
	304	bio_put(&dio->bio);
	305	return ret;
	306	}
	307
	308	static ssize_t blkdev_direct_IO(struct kiocb iocb, struct iov_iter iter)
	309	{
	310	unsigned int nr_pages;
	311
	312	if (!iov_iter_count(iter))
	313	return 0;
	314
	315	nr_pages = bio_iov_vecs_to_alloc(iter, BIO_MAX_VECS + 1);
	316	if (is_sync_kiocb(iocb) && nr_pages <= BIO_MAX_VECS)
	317	return __blkdev_direct_IO_simple(iocb, iter, nr_pages);
	318
	319	return __blkdev_direct_IO(iocb, iter, bio_max_segs(nr_pages));
	320	}
	321
	322	static int blkdev_writepage(struct page page, struct writeback_control wbc)
	323	{
	324	return block_write_full_page(page, blkdev_get_block, wbc);
	325	}
	326
	327	static int blkdev_readpage(struct file * file, struct page * page)
	328	{
	329	return block_read_full_page(page, blkdev_get_block);
	330	}
	331
	332	static void blkdev_readahead(struct readahead_control *rac)
	333	{
	334	mpage_readahead(rac, blkdev_get_block);
	335	}
	336
	337	static int blkdev_write_begin(struct file file, struct address_space mapping,
	338	loff_t pos, unsigned len, unsigned flags, struct page **pagep,
	339	void **fsdata)
	340	{
	341	return block_write_begin(mapping, pos, len, flags, pagep,
	342	blkdev_get_block);
	343	}
	344
	345	static int blkdev_write_end(struct file file, struct address_space mapping,
	346	loff_t pos, unsigned len, unsigned copied, struct page *page,
	347	void *fsdata)
	348	{
	349	int ret;
	350	ret = block_write_end(file, mapping, pos, len, copied, page, fsdata);
	351
	352	unlock_page(page);
	353	put_page(page);
	354
	355	return ret;
	356	}
	357
	358	static int blkdev_writepages(struct address_space *mapping,
359	struct writeback_control *wbc)
360	{
361	return generic_writepages(mapping, wbc);
362	}
363
364	const struct address_space_operations def_blk_aops = {
365	.set_page_dirty = __set_page_dirty_buffers,
366	.readpage = blkdev_readpage,
367	.readahead = blkdev_readahead,
368	.writepage = blkdev_writepage,
369	.write_begin = blkdev_write_begin,
370	.write_end = blkdev_write_end,
371	.writepages = blkdev_writepages,
372	.direct_IO = blkdev_direct_IO,
373	.migratepage = buffer_migrate_page_norefs,
374	.is_dirty_writeback = buffer_check_dirty_writeback,
375	};
376
377	/*
378	* for a block special file file_inode(file)->i_size is zero
379	* so we compute the size by hand (just as in block_read/write above)
380	*/
381	static loff_t blkdev_llseek(struct file *file, loff_t offset, int whence)
382	{
383	struct inode *bd_inode = bdev_file_inode(file);
384	loff_t retval;
385
386	inode_lock(bd_inode);
387	retval = fixed_size_llseek(file, offset, whence, i_size_read(bd_inode));
388	inode_unlock(bd_inode);
389	return retval;
390	}
391
392	static int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
393	int datasync)
394	{
fac7c6d5	395	struct block_device *bdev = filp->private_data;
cd82cca7 CH	396	int error;
	397
	398	error = file_write_and_wait_range(filp, start, end);
	399	if (error)
	400	return error;
	401
	402	/*
	403	* There is no need to serialise calls to blkdev_issue_flush with
	404	* i_mutex and doing so causes performance issues with concurrent
	405	* O_SYNC writers to a block device.
	406	*/
	407	error = blkdev_issue_flush(bdev);
	408	if (error == -EOPNOTSUPP)
	409	error = 0;
	410
	411	return error;
	412	}
	413
	414	static int blkdev_open(struct inode inode, struct file filp)
	415	{
	416	struct block_device *bdev;
	417
	418	/*
	419	* Preserve backwards compatibility and allow large file access
	420	* even if userspace doesn't ask for it explicitly. Some mkfs
	421	* binary needs it. We might want to drop this workaround
	422	* during an unstable branch.
	423	*/
	424	filp->f_flags \|= O_LARGEFILE;
	425	filp->f_mode \|= FMODE_NOWAIT \| FMODE_BUF_RASYNC;
	426
	427	if (filp->f_flags & O_NDELAY)
	428	filp->f_mode \|= FMODE_NDELAY;
	429	if (filp->f_flags & O_EXCL)
	430	filp->f_mode \|= FMODE_EXCL;
	431	if ((filp->f_flags & O_ACCMODE) == 3)
	432	filp->f_mode \|= FMODE_WRITE_IOCTL;
	433
	434	bdev = blkdev_get_by_dev(inode->i_rdev, filp->f_mode, filp);
	435	if (IS_ERR(bdev))
	436	return PTR_ERR(bdev);
fac7c6d5 PB	437
fac7c6d5 PB	438	filp->private_data = bdev;
cd82cca7 CH	439	filp->f_mapping = bdev->bd_inode->i_mapping;
	440	filp->f_wb_err = filemap_sample_wb_err(filp->f_mapping);
	441	return 0;
	442	}
	443
	444	static int blkdev_close(struct inode inode, struct file filp)
	445	{
fac7c6d5	446	struct block_device *bdev = filp->private_data;
cd82cca7 CH	447
	448	blkdev_put(bdev, filp->f_mode);
	449	return 0;
	450	}
	451
cd82cca7 CH	452	/*
	453	* Write data to the block device. Only intended for the block device itself
	454	* and the raw driver which basically is a fake block device.
	455	*
	456	* Does not take i_mutex for the write and thus is not for general purpose
	457	* use.
	458	*/
	459	static ssize_t blkdev_write_iter(struct kiocb iocb, struct iov_iter from)
	460	{
fac7c6d5 PB	461	struct block_device *bdev = iocb->ki_filp->private_data;
fac7c6d5 PB	462	struct inode *bd_inode = bdev->bd_inode;
cd82cca7 CH	463	loff_t size = i_size_read(bd_inode);
	464	struct blk_plug plug;
	465	size_t shorted = 0;
	466	ssize_t ret;
	467
fac7c6d5	468	if (bdev_read_only(bdev))
cd82cca7 CH	469	return -EPERM;
	470
	471	if (IS_SWAPFILE(bd_inode) && !is_hibernate_resume_dev(bd_inode->i_rdev))
	472	return -ETXTBSY;
	473
	474	if (!iov_iter_count(from))
	475	return 0;
	476
	477	if (iocb->ki_pos >= size)
	478	return -ENOSPC;
	479
	480	if ((iocb->ki_flags & (IOCB_NOWAIT \| IOCB_DIRECT)) == IOCB_NOWAIT)
	481	return -EOPNOTSUPP;
	482
	483	size -= iocb->ki_pos;
	484	if (iov_iter_count(from) > size) {
	485	shorted = iov_iter_count(from) - size;
	486	iov_iter_truncate(from, size);
	487	}
	488
	489	blk_start_plug(&plug);
	490	ret = __generic_file_write_iter(iocb, from);
	491	if (ret > 0)
	492	ret = generic_write_sync(iocb, ret);
	493	iov_iter_reexpand(from, iov_iter_count(from) + shorted);
	494	blk_finish_plug(&plug);
	495	return ret;
	496	}
	497
	498	static ssize_t blkdev_read_iter(struct kiocb iocb, struct iov_iter to)
	499	{
fac7c6d5 PB	500	struct block_device *bdev = iocb->ki_filp->private_data;
fac7c6d5 PB	501	loff_t size = i_size_read(bdev->bd_inode);
cd82cca7 CH	502	loff_t pos = iocb->ki_pos;
	503	size_t shorted = 0;
	504	ssize_t ret;
	505
6450fe1f PB	506	if (unlikely(pos + iov_iter_count(to) > size)) {
	507	if (pos >= size)
	508	return 0;
	509	size -= pos;
	510	if (iov_iter_count(to) > size) {
	511	shorted = iov_iter_count(to) - size;
	512	iov_iter_truncate(to, size);
	513	}
cd82cca7 CH	514	}
	515
	516	ret = generic_file_read_iter(iocb, to);
6450fe1f PB	517
	518	if (unlikely(shorted))
	519	iov_iter_reexpand(to, iov_iter_count(to) + shorted);
cd82cca7 CH	520	return ret;
	521	}
	522
	523	#define BLKDEV_FALLOC_FL_SUPPORTED \
	524	(FALLOC_FL_KEEP_SIZE \| FALLOC_FL_PUNCH_HOLE \| \
	525	FALLOC_FL_ZERO_RANGE \| FALLOC_FL_NO_HIDE_STALE)
	526
	527	static long blkdev_fallocate(struct file *file, int mode, loff_t start,
	528	loff_t len)
	529	{
f278eb3d ML	530	struct inode *inode = bdev_file_inode(file);
f278eb3d ML	531	struct block_device *bdev = I_BDEV(inode);
cd82cca7 CH	532	loff_t end = start + len - 1;
	533	loff_t isize;
	534	int error;
	535
	536	/* Fail if we don't recognize the flags. */
	537	if (mode & ~BLKDEV_FALLOC_FL_SUPPORTED)
	538	return -EOPNOTSUPP;
	539
	540	/* Don't go off the end of the device. */
	541	isize = i_size_read(bdev->bd_inode);
	542	if (start >= isize)
	543	return -EINVAL;
	544	if (end >= isize) {
	545	if (mode & FALLOC_FL_KEEP_SIZE) {
	546	len = isize - start;
	547	end = start + len - 1;
	548	} else
	549	return -EINVAL;
	550	}
	551
	552	/*
	553	* Don't allow IO that isn't aligned to logical block size.
	554	*/
	555	if ((start \| len) & (bdev_logical_block_size(bdev) - 1))
	556	return -EINVAL;
	557
f278eb3d ML	558	filemap_invalidate_lock(inode->i_mapping);
f278eb3d ML	559
cd82cca7 CH	560	/* Invalidate the page cache, including dirty pages. */
	561	error = truncate_bdev_range(bdev, file->f_mode, start, end);
	562	if (error)
f278eb3d	563	goto fail;
cd82cca7 CH	564
	565	switch (mode) {
	566	case FALLOC_FL_ZERO_RANGE:
	567	case FALLOC_FL_ZERO_RANGE \| FALLOC_FL_KEEP_SIZE:
6549a874 PB	568	error = blkdev_issue_zeroout(bdev, start >> SECTOR_SHIFT,
	569	len >> SECTOR_SHIFT, GFP_KERNEL,
	570	BLKDEV_ZERO_NOUNMAP);
cd82cca7 CH	571	break;
cd82cca7 CH	572	case FALLOC_FL_PUNCH_HOLE \| FALLOC_FL_KEEP_SIZE:
6549a874 PB	573	error = blkdev_issue_zeroout(bdev, start >> SECTOR_SHIFT,
	574	len >> SECTOR_SHIFT, GFP_KERNEL,
	575	BLKDEV_ZERO_NOFALLBACK);
cd82cca7 CH	576	break;
cd82cca7 CH	577	case FALLOC_FL_PUNCH_HOLE \| FALLOC_FL_KEEP_SIZE \| FALLOC_FL_NO_HIDE_STALE:
6549a874 PB	578	error = blkdev_issue_discard(bdev, start >> SECTOR_SHIFT,
6549a874 PB	579	len >> SECTOR_SHIFT, GFP_KERNEL, 0);
cd82cca7 CH	580	break;
cd82cca7 CH	581	default:
f278eb3d	582	error = -EOPNOTSUPP;
cd82cca7	583	}
cd82cca7	584
f278eb3d ML	585	fail:
	586	filemap_invalidate_unlock(inode->i_mapping);
	587	return error;
cd82cca7 CH	588	}
	589
	590	const struct file_operations def_blk_fops = {
	591	.open = blkdev_open,
	592	.release = blkdev_close,
	593	.llseek = blkdev_llseek,
	594	.read_iter = blkdev_read_iter,
	595	.write_iter = blkdev_write_iter,
3e08773c	596	.iopoll = iocb_bio_iopoll,
cd82cca7 CH	597	.mmap = generic_file_mmap,
cd82cca7 CH	598	.fsync = blkdev_fsync,
8a709512	599	.unlocked_ioctl = blkdev_ioctl,
cd82cca7 CH	600	#ifdef CONFIG_COMPAT
	601	.compat_ioctl = compat_blkdev_ioctl,
	602	#endif
	603	.splice_read = generic_file_splice_read,
	604	.splice_write = iter_file_splice_write,
	605	.fallocate = blkdev_fallocate,
	606	};
	607
	608	static __init int blkdev_init(void)
	609	{
	610	return bioset_init(&blkdev_dio_pool, 4,
	611	offsetof(struct blkdev_dio, bio),
	612	BIOSET_NEED_BVECS\|BIOSET_PERCPU_CACHE);
	613	}
	614	module_init(blkdev_init);