[linux-2.6-block.git] / fs / sync.c

/*
 * High-level sync()-related operations
 */

#include <linux/kernel.h>
#include <linux/file.h>
#include <linux/fs.h>
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/sched.h>
#include <linux/writeback.h>
#include <linux/syscalls.h>
#include <linux/linkage.h>
#include <linux/pagemap.h>
#include <linux/quotaops.h>
#include <linux/buffer_head.h>
#include <linux/backing-dev.h>
#include "internal.h"

#define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \
			SYNC_FILE_RANGE_WAIT_AFTER)

/*
 * Do the filesystem syncing work. For simple filesystems
 * writeback_inodes_sb(sb) just dirties buffers with inodes so we have to
 * submit IO for these buffers via __sync_blockdev(). This also speeds up the
 * wait == 1 case since in that case write_inode() functions do
 * sync_dirty_buffer() and thus effectively write one block at a time.
 */
static int __sync_filesystem(struct super_block *sb, int wait)
{
	/*
	 * This should be safe, as we require bdi backing to actually
	 * write out data in the first place
	 */
	if (!sb->s_bdi || sb->s_bdi == &noop_backing_dev_info)
		return 0;

	if (sb->s_qcop && sb->s_qcop->quota_sync)
		sb->s_qcop->quota_sync(sb, -1, wait);

	if (wait)
		sync_inodes_sb(sb);
	else
		writeback_inodes_sb(sb);

	if (sb->s_op->sync_fs)
		sb->s_op->sync_fs(sb, wait);
	return __sync_blockdev(sb->s_bdev, wait);
}

/*
 * Write out and wait upon all dirty data associated with this
 * superblock.  Filesystem data as well as the underlying block
 * device.  Takes the superblock lock.
 */
int sync_filesystem(struct super_block *sb)
{
	int ret;

	/*
	 * We need to be protected against the filesystem going from
	 * r/o to r/w or vice versa.
	 */
	WARN_ON(!rwsem_is_locked(&sb->s_umount));

	/*
	 * No point in syncing out anything if the filesystem is read-only.
	 */
	if (sb->s_flags & MS_RDONLY)
		return 0;

	ret = __sync_filesystem(sb, 0);
	if (ret < 0)
		return ret;
	return __sync_filesystem(sb, 1);
}
EXPORT_SYMBOL_GPL(sync_filesystem);

/*
 * Sync all the data for all the filesystems (called by sys_sync() and
 * emergency sync)
 *
 * This operation is careful to avoid the livelock which could easily happen
 * if two or more filesystems are being continuously dirtied.  s_need_sync
 * is used only here.  We set it against all filesystems and then clear it as
 * we sync them.  So redirtied filesystems are skipped.
 *
 * But if process A is currently running sync_filesystems and then process B
 * calls sync_filesystems as well, process B will set all the s_need_sync
 * flags again, which will cause process A to resync everything.  Fix that with
 * a local mutex.
 */
static void sync_filesystems(int wait)
{
	struct super_block *sb;
	static DEFINE_MUTEX(mutex);

	mutex_lock(&mutex);		/* Could be down_interruptible */
	spin_lock(&sb_lock);
	list_for_each_entry(sb, &super_blocks, s_list)
		sb->s_need_sync = 1;

restart:
	list_for_each_entry(sb, &super_blocks, s_list) {
		if (!sb->s_need_sync)
			continue;
		sb->s_need_sync = 0;
		sb->s_count++;
		spin_unlock(&sb_lock);

		down_read(&sb->s_umount);
		if (!(sb->s_flags & MS_RDONLY) && sb->s_root && sb->s_bdi)
			__sync_filesystem(sb, wait);
		up_read(&sb->s_umount);

		/* restart only when sb is no longer on the list */
		spin_lock(&sb_lock);
		if (__put_super_and_need_restart(sb))
			goto restart;
	}
	spin_unlock(&sb_lock);
	mutex_unlock(&mutex);
}

/*
 * sync everything.  Start out by waking pdflush, because that writes back
 * all queues in parallel.
 */
SYSCALL_DEFINE0(sync)
{
	wakeup_flusher_threads(0);
	sync_filesystems(0);
	sync_filesystems(1);
	if (unlikely(laptop_mode))
		laptop_sync_completion();
	return 0;
}

static void do_sync_work(struct work_struct *work)
{
	/*
	 * Sync twice to reduce the possibility we skipped some inodes / pages
	 * because they were temporarily locked
	 */
	sync_filesystems(0);
	sync_filesystems(0);
	printk("Emergency Sync complete\n");
	kfree(work);
}

void emergency_sync(void)
{
	struct work_struct *work;

	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	if (work) {
		INIT_WORK(work, do_sync_work);
		schedule_work(work);
	}
}

/*
 * Generic function to fsync a file.
 *
 * filp may be NULL if called via the msync of a vma.
 */
int file_fsync(struct file *filp, struct dentry *dentry, int datasync)
{
	struct inode * inode = dentry->d_inode;
	struct super_block * sb;
	int ret, err;

	/* sync the inode to buffers */
	ret = write_inode_now(inode, 0);

	/* sync the superblock to buffers */
	sb = inode->i_sb;
	if (sb->s_dirt && sb->s_op->write_super)
		sb->s_op->write_super(sb);

	/* .. finally sync the buffers to disk */
	err = sync_blockdev(sb->s_bdev);
	if (!ret)
		ret = err;
	return ret;
}
EXPORT_SYMBOL(file_fsync);

/**
 * vfs_fsync_range - helper to sync a range of data & metadata to disk
 * @file:		file to sync
 * @dentry:		dentry of @file
 * @start:		offset in bytes of the beginning of data range to sync
 * @end:		offset in bytes of the end of data range (inclusive)
 * @datasync:		perform only datasync
 *
 * Write back data in range @start..@end and metadata for @file to disk.  If
 * @datasync is set only metadata needed to access modified file data is
 * written.
 *
 * In case this function is called from nfsd @file may be %NULL and
 * only @dentry is set.  This can only happen when the filesystem
 * implements the export_operations API.
 */
int vfs_fsync_range(struct file *file, struct dentry *dentry, loff_t start,
		    loff_t end, int datasync)
{
	const struct file_operations *fop;
	struct address_space *mapping;
	int err, ret;

	/*
	 * Get mapping and operations from the file in case we have
	 * as file, or get the default values for them in case we
	 * don't have a struct file available.  Damn nfsd..
	 */
	if (file) {
		mapping = file->f_mapping;
		fop = file->f_op;
	} else {
		mapping = dentry->d_inode->i_mapping;
		fop = dentry->d_inode->i_fop;
	}

	if (!fop || !fop->fsync) {
		ret = -EINVAL;
		goto out;
	}

	ret = filemap_write_and_wait_range(mapping, start, end);

	/*
	 * We need to protect against concurrent writers, which could cause
	 * livelocks in fsync_buffers_list().
	 */
	mutex_lock(&mapping->host->i_mutex);
	err = fop->fsync(file, dentry, datasync);
	if (!ret)
		ret = err;
	mutex_unlock(&mapping->host->i_mutex);

out:
	return ret;
}
EXPORT_SYMBOL(vfs_fsync_range);

/**
 * vfs_fsync - perform a fsync or fdatasync on a file
 * @file:		file to sync
 * @dentry:		dentry of @file
 * @datasync:		only perform a fdatasync operation
 *
 * Write back data and metadata for @file to disk.  If @datasync is
 * set only metadata needed to access modified file data is written.
 *
 * In case this function is called from nfsd @file may be %NULL and
 * only @dentry is set.  This can only happen when the filesystem
 * implements the export_operations API.
 */
int vfs_fsync(struct file *file, struct dentry *dentry, int datasync)
{
	return vfs_fsync_range(file, dentry, 0, LLONG_MAX, datasync);
}
EXPORT_SYMBOL(vfs_fsync);

static int do_fsync(unsigned int fd, int datasync)
{
	struct file *file;
	int ret = -EBADF;

	file = fget(fd);
	if (file) {
		ret = vfs_fsync(file, file->f_path.dentry, datasync);
		fput(file);
	}
	return ret;
}

SYSCALL_DEFINE1(fsync, unsigned int, fd)
{
	return do_fsync(fd, 0);
}

SYSCALL_DEFINE1(fdatasync, unsigned int, fd)
{
	return do_fsync(fd, 1);
}

/**
 * generic_write_sync - perform syncing after a write if file / inode is sync
 * @file:	file to which the write happened
 * @pos:	offset where the write started
 * @count:	length of the write
 *
 * This is just a simple wrapper about our general syncing function.
 */
int generic_write_sync(struct file *file, loff_t pos, loff_t count)
{
	if (!(file->f_flags & O_DSYNC) && !IS_SYNC(file->f_mapping->host))
		return 0;
	return vfs_fsync_range(file, file->f_path.dentry, pos,
			       pos + count - 1,
			       (file->f_flags & __O_SYNC) ? 0 : 1);
}
EXPORT_SYMBOL(generic_write_sync);

/*
 * sys_sync_file_range() permits finely controlled syncing over a segment of
 * a file in the range offset .. (offset+nbytes-1) inclusive.  If nbytes is
 * zero then sys_sync_file_range() will operate from offset out to EOF.
 *
 * The flag bits are:
 *
 * SYNC_FILE_RANGE_WAIT_BEFORE: wait upon writeout of all pages in the range
 * before performing the write.
 *
 * SYNC_FILE_RANGE_WRITE: initiate writeout of all those dirty pages in the
 * range which are not presently under writeback. Note that this may block for
 * significant periods due to exhaustion of disk request structures.
 *
 * SYNC_FILE_RANGE_WAIT_AFTER: wait upon writeout of all pages in the range
 * after performing the write.
 *
 * Useful combinations of the flag bits are:
 *
 * SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE: ensures that all pages
 * in the range which were dirty on entry to sys_sync_file_range() are placed
 * under writeout.  This is a start-write-for-data-integrity operation.
 *
 * SYNC_FILE_RANGE_WRITE: start writeout of all dirty pages in the range which
 * are not presently under writeout.  This is an asynchronous flush-to-disk
 * operation.  Not suitable for data integrity operations.
 *
 * SYNC_FILE_RANGE_WAIT_BEFORE (or SYNC_FILE_RANGE_WAIT_AFTER): wait for
 * completion of writeout of all pages in the range.  This will be used after an
 * earlier SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE operation to wait
 * for that operation to complete and to return the result.
 *
 * SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE|SYNC_FILE_RANGE_WAIT_AFTER:
 * a traditional sync() operation.  This is a write-for-data-integrity operation
 * which will ensure that all pages in the range which were dirty on entry to
 * sys_sync_file_range() are committed to disk.
 *
 *
 * SYNC_FILE_RANGE_WAIT_BEFORE and SYNC_FILE_RANGE_WAIT_AFTER will detect any
 * I/O errors or ENOSPC conditions and will return those to the caller, after
 * clearing the EIO and ENOSPC flags in the address_space.
 *
 * It should be noted that none of these operations write out the file's
 * metadata.  So unless the application is strictly performing overwrites of
 * already-instantiated disk blocks, there are no guarantees here that the data
 * will be available after a crash.
 */
SYSCALL_DEFINE(sync_file_range)(int fd, loff_t offset, loff_t nbytes,
				unsigned int flags)
{
	int ret;
	struct file *file;
	struct address_space *mapping;
	loff_t endbyte;			/* inclusive */
	int fput_needed;
	umode_t i_mode;

	ret = -EINVAL;
	if (flags & ~VALID_FLAGS)
		goto out;

	endbyte = offset + nbytes;

	if ((s64)offset < 0)
		goto out;
	if ((s64)endbyte < 0)
		goto out;
	if (endbyte < offset)
		goto out;

	if (sizeof(pgoff_t) == 4) {
		if (offset >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
			/*
			 * The range starts outside a 32 bit machine's
			 * pagecache addressing capabilities.  Let it "succeed"
			 */
			ret = 0;
			goto out;
		}
		if (endbyte >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
			/*
			 * Out to EOF
			 */
			nbytes = 0;
		}
	}

	if (nbytes == 0)
		endbyte = LLONG_MAX;
	else
		endbyte--;		/* inclusive */

	ret = -EBADF;
	file = fget_light(fd, &fput_needed);
	if (!file)
		goto out;

	i_mode = file->f_path.dentry->d_inode->i_mode;
	ret = -ESPIPE;
	if (!S_ISREG(i_mode) && !S_ISBLK(i_mode) && !S_ISDIR(i_mode) &&
			!S_ISLNK(i_mode))
		goto out_put;

	mapping = file->f_mapping;
	if (!mapping) {
		ret = -EINVAL;
		goto out_put;
	}

	ret = 0;
	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
		ret = filemap_fdatawait_range(mapping, offset, endbyte);
		if (ret < 0)
			goto out_put;
	}

	if (flags & SYNC_FILE_RANGE_WRITE) {
		ret = filemap_fdatawrite_range(mapping, offset, endbyte);
		if (ret < 0)
			goto out_put;
	}

	if (flags & SYNC_FILE_RANGE_WAIT_AFTER)
		ret = filemap_fdatawait_range(mapping, offset, endbyte);

out_put:
	fput_light(file, fput_needed);
out:
	return ret;
}
#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
asmlinkage long SyS_sync_file_range(long fd, loff_t offset, loff_t nbytes,
				    long flags)
{
	return SYSC_sync_file_range((int) fd, offset, nbytes,
				    (unsigned int) flags);
}
SYSCALL_ALIAS(sys_sync_file_range, SyS_sync_file_range);
#endif

/* It would be nice if people remember that not all the world's an i386
   when they introduce new system calls */
SYSCALL_DEFINE(sync_file_range2)(int fd, unsigned int flags,
				 loff_t offset, loff_t nbytes)
{
	return sys_sync_file_range(fd, offset, nbytes, flags);
}
#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
asmlinkage long SyS_sync_file_range2(long fd, long flags,
				     loff_t offset, loff_t nbytes)
{
	return SYSC_sync_file_range2((int) fd, (unsigned int) flags,
				     offset, nbytes);
}
SYSCALL_ALIAS(sys_sync_file_range2, SyS_sync_file_range2);
#endif
Commit	Line	Data
f79e2abb AM	1	/*
	2	* High-level sync()-related operations
	3	*/
	4
	5	#include <linux/kernel.h>
	6	#include <linux/file.h>
	7	#include <linux/fs.h>
5a0e3ad6	8	#include <linux/slab.h>
f79e2abb	9	#include <linux/module.h>
914e2637	10	#include <linux/sched.h>
f79e2abb AM	11	#include <linux/writeback.h>
	12	#include <linux/syscalls.h>
	13	#include <linux/linkage.h>
	14	#include <linux/pagemap.h>
cf9a2ae8 DH	15	#include <linux/quotaops.h>
cf9a2ae8 DH	16	#include <linux/buffer_head.h>
5129a469	17	#include <linux/backing-dev.h>
5a3e5cb8	18	#include "internal.h"
f79e2abb AM	19
	20	#define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\| \
	21	SYNC_FILE_RANGE_WAIT_AFTER)
	22
c15c54f5	23	/*
d8a8559c JA	24	* Do the filesystem syncing work. For simple filesystems
	25	* writeback_inodes_sb(sb) just dirties buffers with inodes so we have to
	26	* submit IO for these buffers via __sync_blockdev(). This also speeds up the
	27	* wait == 1 case since in that case write_inode() functions do
	28	* sync_dirty_buffer() and thus effectively write one block at a time.
c15c54f5	29	*/
60b0680f	30	static int __sync_filesystem(struct super_block *sb, int wait)
c15c54f5	31	{
32a88aa1 JA	32	/*
	33	* This should be safe, as we require bdi backing to actually
	34	* write out data in the first place
	35	*/
5129a469	36	if (!sb->s_bdi \|\| sb->s_bdi == &noop_backing_dev_info)
32a88aa1 JA	37	return 0;
32a88aa1 JA	38
5fb324ad CH	39	if (sb->s_qcop && sb->s_qcop->quota_sync)
	40	sb->s_qcop->quota_sync(sb, -1, wait);
	41
	42	if (wait)
d8a8559c	43	sync_inodes_sb(sb);
5fb324ad CH	44	else
	45	writeback_inodes_sb(sb);
	46
c15c54f5 JK	47	if (sb->s_op->sync_fs)
	48	sb->s_op->sync_fs(sb, wait);
	49	return __sync_blockdev(sb->s_bdev, wait);
	50	}
	51
	52	/*
	53	* Write out and wait upon all dirty data associated with this
	54	* superblock. Filesystem data as well as the underlying block
	55	* device. Takes the superblock lock.
	56	*/
60b0680f	57	int sync_filesystem(struct super_block *sb)
c15c54f5 JK	58	{
	59	int ret;
	60
5af7926f CH	61	/*
	62	* We need to be protected against the filesystem going from
	63	* r/o to r/w or vice versa.
	64	*/
	65	WARN_ON(!rwsem_is_locked(&sb->s_umount));
	66
	67	/*
	68	* No point in syncing out anything if the filesystem is read-only.
	69	*/
	70	if (sb->s_flags & MS_RDONLY)
	71	return 0;
	72
60b0680f	73	ret = __sync_filesystem(sb, 0);
c15c54f5 JK	74	if (ret < 0)
c15c54f5 JK	75	return ret;
60b0680f	76	return __sync_filesystem(sb, 1);
c15c54f5	77	}
60b0680f	78	EXPORT_SYMBOL_GPL(sync_filesystem);
c15c54f5 JK	79
	80	/*
	81	* Sync all the data for all the filesystems (called by sys_sync() and
	82	* emergency sync)
	83	*
	84	* This operation is careful to avoid the livelock which could easily happen
	85	* if two or more filesystems are being continuously dirtied. s_need_sync
	86	* is used only here. We set it against all filesystems and then clear it as
	87	* we sync them. So redirtied filesystems are skipped.
	88	*
	89	* But if process A is currently running sync_filesystems and then process B
	90	* calls sync_filesystems as well, process B will set all the s_need_sync
	91	* flags again, which will cause process A to resync everything. Fix that with
	92	* a local mutex.
	93	*/
	94	static void sync_filesystems(int wait)
	95	{
	96	struct super_block *sb;
	97	static DEFINE_MUTEX(mutex);
	98
	99	mutex_lock(&mutex); /* Could be down_interruptible */
	100	spin_lock(&sb_lock);
5af7926f	101	list_for_each_entry(sb, &super_blocks, s_list)
c15c54f5	102	sb->s_need_sync = 1;
c15c54f5 JK	103
	104	restart:
	105	list_for_each_entry(sb, &super_blocks, s_list) {
	106	if (!sb->s_need_sync)
	107	continue;
	108	sb->s_need_sync = 0;
c15c54f5 JK	109	sb->s_count++;
c15c54f5 JK	110	spin_unlock(&sb_lock);
5af7926f	111
c15c54f5	112	down_read(&sb->s_umount);
32a88aa1	113	if (!(sb->s_flags & MS_RDONLY) && sb->s_root && sb->s_bdi)
60b0680f	114	__sync_filesystem(sb, wait);
c15c54f5	115	up_read(&sb->s_umount);
5af7926f	116
c15c54f5 JK	117	/* restart only when sb is no longer on the list */
	118	spin_lock(&sb_lock);
	119	if (__put_super_and_need_restart(sb))
	120	goto restart;
	121	}
	122	spin_unlock(&sb_lock);
	123	mutex_unlock(&mutex);
	124	}
	125
3beab0b4 ZY	126	/*
	127	* sync everything. Start out by waking pdflush, because that writes back
	128	* all queues in parallel.
	129	*/
5cee5815	130	SYSCALL_DEFINE0(sync)
cf9a2ae8	131	{
03ba3782	132	wakeup_flusher_threads(0);
5cee5815 JK	133	sync_filesystems(0);
5cee5815 JK	134	sync_filesystems(1);
cf9a2ae8 DH	135	if (unlikely(laptop_mode))
cf9a2ae8 DH	136	laptop_sync_completion();
cf9a2ae8 DH	137	return 0;
	138	}
	139
a2a9537a JA	140	static void do_sync_work(struct work_struct *work)
a2a9537a JA	141	{
5cee5815 JK	142	/*
	143	* Sync twice to reduce the possibility we skipped some inodes / pages
	144	* because they were temporarily locked
	145	*/
	146	sync_filesystems(0);
	147	sync_filesystems(0);
	148	printk("Emergency Sync complete\n");
a2a9537a JA	149	kfree(work);
	150	}
	151
cf9a2ae8 DH	152	void emergency_sync(void)
cf9a2ae8 DH	153	{
a2a9537a JA	154	struct work_struct *work;
	155
	156	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	157	if (work) {
	158	INIT_WORK(work, do_sync_work);
	159	schedule_work(work);
	160	}
cf9a2ae8 DH	161	}
	162
	163	/*
	164	* Generic function to fsync a file.
	165	*
	166	* filp may be NULL if called via the msync of a vma.
	167	*/
	168	int file_fsync(struct file filp, struct dentry dentry, int datasync)
	169	{
	170	struct inode * inode = dentry->d_inode;
	171	struct super_block * sb;
	172	int ret, err;
	173
	174	/* sync the inode to buffers */
	175	ret = write_inode_now(inode, 0);
	176
	177	/* sync the superblock to buffers */
	178	sb = inode->i_sb;
762873c2	179	if (sb->s_dirt && sb->s_op->write_super)
cf9a2ae8	180	sb->s_op->write_super(sb);
cf9a2ae8 DH	181
	182	/* .. finally sync the buffers to disk */
	183	err = sync_blockdev(sb->s_bdev);
	184	if (!ret)
	185	ret = err;
	186	return ret;
	187	}
1fe72eaa	188	EXPORT_SYMBOL(file_fsync);
cf9a2ae8	189
4c728ef5	190	/**
148f948b	191	* vfs_fsync_range - helper to sync a range of data & metadata to disk
4c728ef5 CH	192	* @file: file to sync
4c728ef5 CH	193	* @dentry: dentry of @file
148f948b JK	194	* @start: offset in bytes of the beginning of data range to sync
	195	* @end: offset in bytes of the end of data range (inclusive)
	196	* @datasync: perform only datasync
4c728ef5	197	*
148f948b JK	198	* Write back data in range @start..@end and metadata for @file to disk. If
	199	* @datasync is set only metadata needed to access modified file data is
	200	* written.
4c728ef5 CH	201	*
	202	* In case this function is called from nfsd @file may be %NULL and
	203	* only @dentry is set. This can only happen when the filesystem
	204	* implements the export_operations API.
	205	*/
148f948b JK	206	int vfs_fsync_range(struct file file, struct dentry dentry, loff_t start,
148f948b JK	207	loff_t end, int datasync)
cf9a2ae8	208	{
4c728ef5 CH	209	const struct file_operations *fop;
	210	struct address_space *mapping;
	211	int err, ret;
	212
	213	/*
	214	* Get mapping and operations from the file in case we have
	215	* as file, or get the default values for them in case we
	216	* don't have a struct file available. Damn nfsd..
	217	*/
	218	if (file) {
	219	mapping = file->f_mapping;
	220	fop = file->f_op;
	221	} else {
	222	mapping = dentry->d_inode->i_mapping;
	223	fop = dentry->d_inode->i_fop;
	224	}
cf9a2ae8	225
4c728ef5	226	if (!fop \|\| !fop->fsync) {
cf9a2ae8 DH	227	ret = -EINVAL;
	228	goto out;
	229	}
	230
2daea67e	231	ret = filemap_write_and_wait_range(mapping, start, end);
cf9a2ae8 DH	232
	233	/*
	234	* We need to protect against concurrent writers, which could cause
	235	* livelocks in fsync_buffers_list().
	236	*/
	237	mutex_lock(&mapping->host->i_mutex);
4c728ef5	238	err = fop->fsync(file, dentry, datasync);
cf9a2ae8 DH	239	if (!ret)
	240	ret = err;
	241	mutex_unlock(&mapping->host->i_mutex);
148f948b	242
cf9a2ae8 DH	243	out:
	244	return ret;
	245	}
148f948b JK	246	EXPORT_SYMBOL(vfs_fsync_range);
	247
	248	/**
	249	* vfs_fsync - perform a fsync or fdatasync on a file
	250	* @file: file to sync
	251	* @dentry: dentry of @file
	252	* @datasync: only perform a fdatasync operation
	253	*
	254	* Write back data and metadata for @file to disk. If @datasync is
	255	* set only metadata needed to access modified file data is written.
	256	*
	257	* In case this function is called from nfsd @file may be %NULL and
	258	* only @dentry is set. This can only happen when the filesystem
	259	* implements the export_operations API.
	260	*/
	261	int vfs_fsync(struct file file, struct dentry dentry, int datasync)
	262	{
	263	return vfs_fsync_range(file, dentry, 0, LLONG_MAX, datasync);
	264	}
4c728ef5	265	EXPORT_SYMBOL(vfs_fsync);
cf9a2ae8	266
4c728ef5	267	static int do_fsync(unsigned int fd, int datasync)
cf9a2ae8 DH	268	{
	269	struct file *file;
	270	int ret = -EBADF;
	271
	272	file = fget(fd);
	273	if (file) {
4c728ef5	274	ret = vfs_fsync(file, file->f_path.dentry, datasync);
cf9a2ae8 DH	275	fput(file);
	276	}
	277	return ret;
	278	}
	279
a5f8fa9e	280	SYSCALL_DEFINE1(fsync, unsigned int, fd)
cf9a2ae8	281	{
4c728ef5	282	return do_fsync(fd, 0);
cf9a2ae8 DH	283	}
cf9a2ae8 DH	284
a5f8fa9e	285	SYSCALL_DEFINE1(fdatasync, unsigned int, fd)
cf9a2ae8	286	{
4c728ef5	287	return do_fsync(fd, 1);
cf9a2ae8 DH	288	}
cf9a2ae8 DH	289
148f948b JK	290	/**
	291	* generic_write_sync - perform syncing after a write if file / inode is sync
	292	* @file: file to which the write happened
	293	* @pos: offset where the write started
	294	* @count: length of the write
	295	*
	296	* This is just a simple wrapper about our general syncing function.
	297	*/
	298	int generic_write_sync(struct file *file, loff_t pos, loff_t count)
	299	{
6b2f3d1f	300	if (!(file->f_flags & O_DSYNC) && !IS_SYNC(file->f_mapping->host))
148f948b JK	301	return 0;
148f948b JK	302	return vfs_fsync_range(file, file->f_path.dentry, pos,
6b2f3d1f CH	303	pos + count - 1,
6b2f3d1f CH	304	(file->f_flags & __O_SYNC) ? 0 : 1);
148f948b JK	305	}
	306	EXPORT_SYMBOL(generic_write_sync);
	307
f79e2abb AM	308	/*
	309	* sys_sync_file_range() permits finely controlled syncing over a segment of
	310	* a file in the range offset .. (offset+nbytes-1) inclusive. If nbytes is
	311	* zero then sys_sync_file_range() will operate from offset out to EOF.
	312	*
	313	* The flag bits are:
	314	*
	315	* SYNC_FILE_RANGE_WAIT_BEFORE: wait upon writeout of all pages in the range
	316	* before performing the write.
	317	*
	318	* SYNC_FILE_RANGE_WRITE: initiate writeout of all those dirty pages in the
cce77081 PM	319	* range which are not presently under writeback. Note that this may block for
cce77081 PM	320	* significant periods due to exhaustion of disk request structures.
f79e2abb AM	321	*
	322	* SYNC_FILE_RANGE_WAIT_AFTER: wait upon writeout of all pages in the range
	323	* after performing the write.
	324	*
	325	* Useful combinations of the flag bits are:
	326	*
	327	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE: ensures that all pages
	328	* in the range which were dirty on entry to sys_sync_file_range() are placed
	329	* under writeout. This is a start-write-for-data-integrity operation.
	330	*
	331	* SYNC_FILE_RANGE_WRITE: start writeout of all dirty pages in the range which
	332	* are not presently under writeout. This is an asynchronous flush-to-disk
	333	* operation. Not suitable for data integrity operations.
	334	*
	335	* SYNC_FILE_RANGE_WAIT_BEFORE (or SYNC_FILE_RANGE_WAIT_AFTER): wait for
	336	* completion of writeout of all pages in the range. This will be used after an
	337	* earlier SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE operation to wait
	338	* for that operation to complete and to return the result.
	339	*
	340	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\|SYNC_FILE_RANGE_WAIT_AFTER:
	341	* a traditional sync() operation. This is a write-for-data-integrity operation
	342	* which will ensure that all pages in the range which were dirty on entry to
	343	* sys_sync_file_range() are committed to disk.
	344	*
	345	*
	346	* SYNC_FILE_RANGE_WAIT_BEFORE and SYNC_FILE_RANGE_WAIT_AFTER will detect any
	347	* I/O errors or ENOSPC conditions and will return those to the caller, after
	348	* clearing the EIO and ENOSPC flags in the address_space.
	349	*
	350	* It should be noted that none of these operations write out the file's
	351	* metadata. So unless the application is strictly performing overwrites of
	352	* already-instantiated disk blocks, there are no guarantees here that the data
	353	* will be available after a crash.
	354	*/
6673e0c3 HC	355	SYSCALL_DEFINE(sync_file_range)(int fd, loff_t offset, loff_t nbytes,
6673e0c3 HC	356	unsigned int flags)
f79e2abb AM	357	{
	358	int ret;
	359	struct file *file;
7a0ad10c	360	struct address_space *mapping;
f79e2abb AM	361	loff_t endbyte; /* inclusive */
	362	int fput_needed;
	363	umode_t i_mode;
	364
	365	ret = -EINVAL;
	366	if (flags & ~VALID_FLAGS)
	367	goto out;
	368
	369	endbyte = offset + nbytes;
	370
	371	if ((s64)offset < 0)
	372	goto out;
	373	if ((s64)endbyte < 0)
	374	goto out;
	375	if (endbyte < offset)
	376	goto out;
	377
	378	if (sizeof(pgoff_t) == 4) {
	379	if (offset >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	380	/*
	381	* The range starts outside a 32 bit machine's
	382	* pagecache addressing capabilities. Let it "succeed"
	383	*/
	384	ret = 0;
	385	goto out;
	386	}
	387	if (endbyte >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	388	/*
	389	* Out to EOF
	390	*/
	391	nbytes = 0;
	392	}
	393	}
	394
	395	if (nbytes == 0)
111ebb6e	396	endbyte = LLONG_MAX;
f79e2abb AM	397	else
	398	endbyte--; /* inclusive */
	399
	400	ret = -EBADF;
	401	file = fget_light(fd, &fput_needed);
	402	if (!file)
	403	goto out;
	404
0f7fc9e4	405	i_mode = file->f_path.dentry->d_inode->i_mode;
f79e2abb AM	406	ret = -ESPIPE;
	407	if (!S_ISREG(i_mode) && !S_ISBLK(i_mode) && !S_ISDIR(i_mode) &&
	408	!S_ISLNK(i_mode))
	409	goto out_put;
	410
7a0ad10c CH	411	mapping = file->f_mapping;
	412	if (!mapping) {
	413	ret = -EINVAL;
	414	goto out_put;
	415	}
	416
	417	ret = 0;
	418	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
	419	ret = filemap_fdatawait_range(mapping, offset, endbyte);
	420	if (ret < 0)
	421	goto out_put;
	422	}
	423
	424	if (flags & SYNC_FILE_RANGE_WRITE) {
	425	ret = filemap_fdatawrite_range(mapping, offset, endbyte);
	426	if (ret < 0)
	427	goto out_put;
	428	}
	429
	430	if (flags & SYNC_FILE_RANGE_WAIT_AFTER)
	431	ret = filemap_fdatawait_range(mapping, offset, endbyte);
	432
f79e2abb AM	433	out_put:
	434	fput_light(file, fput_needed);
	435	out:
	436	return ret;
	437	}
6673e0c3 HC	438	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	439	asmlinkage long SyS_sync_file_range(long fd, loff_t offset, loff_t nbytes,
	440	long flags)
	441	{
	442	return SYSC_sync_file_range((int) fd, offset, nbytes,
	443	(unsigned int) flags);
	444	}
	445	SYSCALL_ALIAS(sys_sync_file_range, SyS_sync_file_range);
	446	#endif
f79e2abb	447
edd5cd4a DW	448	/* It would be nice if people remember that not all the world's an i386
edd5cd4a DW	449	when they introduce new system calls */
6673e0c3 HC	450	SYSCALL_DEFINE(sync_file_range2)(int fd, unsigned int flags,
6673e0c3 HC	451	loff_t offset, loff_t nbytes)
edd5cd4a DW	452	{
	453	return sys_sync_file_range(fd, offset, nbytes, flags);
	454	}
6673e0c3 HC	455	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	456	asmlinkage long SyS_sync_file_range2(long fd, long flags,
	457	loff_t offset, loff_t nbytes)
	458	{
	459	return SYSC_sync_file_range2((int) fd, (unsigned int) flags,
	460	offset, nbytes);
	461	}
	462	SYSCALL_ALIAS(sys_sync_file_range2, SyS_sync_file_range2);
	463	#endif