[deliverable/linux.git] / fs / sync.c

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

#include <linux/kernel.h>
#include <linux/file.h>
#include <linux/fs.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 "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)
		return 0;

	/* Avoid doing twice syncing and cache pruning for quota sync */
	if (!wait) {
		writeout_quota_sb(sb, -1);
		writeback_inodes_sb(sb);
	} else {
		sync_quota_sb(sb, -1);
		sync_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>
	8	#include <linux/module.h>
914e2637	9	#include <linux/sched.h>
f79e2abb AM	10	#include <linux/writeback.h>
	11	#include <linux/syscalls.h>
	12	#include <linux/linkage.h>
	13	#include <linux/pagemap.h>
cf9a2ae8 DH	14	#include <linux/quotaops.h>
cf9a2ae8 DH	15	#include <linux/buffer_head.h>
5a3e5cb8	16	#include "internal.h"
f79e2abb AM	17
	18	#define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\| \
	19	SYNC_FILE_RANGE_WAIT_AFTER)
	20
c15c54f5	21	/*
d8a8559c JA	22	* Do the filesystem syncing work. For simple filesystems
	23	* writeback_inodes_sb(sb) just dirties buffers with inodes so we have to
	24	* submit IO for these buffers via __sync_blockdev(). This also speeds up the
	25	* wait == 1 case since in that case write_inode() functions do
	26	* sync_dirty_buffer() and thus effectively write one block at a time.
c15c54f5	27	*/
60b0680f	28	static int __sync_filesystem(struct super_block *sb, int wait)
c15c54f5	29	{
32a88aa1 JA	30	/*
	31	* This should be safe, as we require bdi backing to actually
	32	* write out data in the first place
	33	*/
	34	if (!sb->s_bdi)
	35	return 0;
	36
c3f8a40c	37	/* Avoid doing twice syncing and cache pruning for quota sync */
d8a8559c	38	if (!wait) {
c3f8a40c	39	writeout_quota_sb(sb, -1);
d8a8559c JA	40	writeback_inodes_sb(sb);
d8a8559c JA	41	} else {
c3f8a40c	42	sync_quota_sb(sb, -1);
d8a8559c JA	43	sync_inodes_sb(sb);
d8a8559c JA	44	}
c15c54f5 JK	45	if (sb->s_op->sync_fs)
	46	sb->s_op->sync_fs(sb, wait);
	47	return __sync_blockdev(sb->s_bdev, wait);
	48	}
	49
	50	/*
	51	* Write out and wait upon all dirty data associated with this
	52	* superblock. Filesystem data as well as the underlying block
	53	* device. Takes the superblock lock.
	54	*/
60b0680f	55	int sync_filesystem(struct super_block *sb)
c15c54f5 JK	56	{
	57	int ret;
	58
5af7926f CH	59	/*
	60	* We need to be protected against the filesystem going from
	61	* r/o to r/w or vice versa.
	62	*/
	63	WARN_ON(!rwsem_is_locked(&sb->s_umount));
	64
	65	/*
	66	* No point in syncing out anything if the filesystem is read-only.
	67	*/
	68	if (sb->s_flags & MS_RDONLY)
	69	return 0;
	70
60b0680f	71	ret = __sync_filesystem(sb, 0);
c15c54f5 JK	72	if (ret < 0)
c15c54f5 JK	73	return ret;
60b0680f	74	return __sync_filesystem(sb, 1);
c15c54f5	75	}
60b0680f	76	EXPORT_SYMBOL_GPL(sync_filesystem);
c15c54f5 JK	77
	78	/*
	79	* Sync all the data for all the filesystems (called by sys_sync() and
	80	* emergency sync)
	81	*
	82	* This operation is careful to avoid the livelock which could easily happen
	83	* if two or more filesystems are being continuously dirtied. s_need_sync
	84	* is used only here. We set it against all filesystems and then clear it as
	85	* we sync them. So redirtied filesystems are skipped.
	86	*
	87	* But if process A is currently running sync_filesystems and then process B
	88	* calls sync_filesystems as well, process B will set all the s_need_sync
	89	* flags again, which will cause process A to resync everything. Fix that with
	90	* a local mutex.
	91	*/
	92	static void sync_filesystems(int wait)
	93	{
	94	struct super_block *sb;
	95	static DEFINE_MUTEX(mutex);
	96
	97	mutex_lock(&mutex); /* Could be down_interruptible */
	98	spin_lock(&sb_lock);
5af7926f	99	list_for_each_entry(sb, &super_blocks, s_list)
c15c54f5	100	sb->s_need_sync = 1;
c15c54f5 JK	101
	102	restart:
	103	list_for_each_entry(sb, &super_blocks, s_list) {
	104	if (!sb->s_need_sync)
	105	continue;
	106	sb->s_need_sync = 0;
c15c54f5 JK	107	sb->s_count++;
c15c54f5 JK	108	spin_unlock(&sb_lock);
5af7926f	109
c15c54f5	110	down_read(&sb->s_umount);
32a88aa1	111	if (!(sb->s_flags & MS_RDONLY) && sb->s_root && sb->s_bdi)
60b0680f	112	__sync_filesystem(sb, wait);
c15c54f5	113	up_read(&sb->s_umount);
5af7926f	114
c15c54f5 JK	115	/* restart only when sb is no longer on the list */
	116	spin_lock(&sb_lock);
	117	if (__put_super_and_need_restart(sb))
	118	goto restart;
	119	}
	120	spin_unlock(&sb_lock);
	121	mutex_unlock(&mutex);
	122	}
	123
3beab0b4 ZY	124	/*
	125	* sync everything. Start out by waking pdflush, because that writes back
	126	* all queues in parallel.
	127	*/
5cee5815	128	SYSCALL_DEFINE0(sync)
cf9a2ae8	129	{
03ba3782	130	wakeup_flusher_threads(0);
5cee5815 JK	131	sync_filesystems(0);
5cee5815 JK	132	sync_filesystems(1);
cf9a2ae8 DH	133	if (unlikely(laptop_mode))
cf9a2ae8 DH	134	laptop_sync_completion();
cf9a2ae8 DH	135	return 0;
	136	}
	137
a2a9537a JA	138	static void do_sync_work(struct work_struct *work)
a2a9537a JA	139	{
5cee5815 JK	140	/*
	141	* Sync twice to reduce the possibility we skipped some inodes / pages
	142	* because they were temporarily locked
	143	*/
	144	sync_filesystems(0);
	145	sync_filesystems(0);
	146	printk("Emergency Sync complete\n");
a2a9537a JA	147	kfree(work);
	148	}
	149
cf9a2ae8 DH	150	void emergency_sync(void)
cf9a2ae8 DH	151	{
a2a9537a JA	152	struct work_struct *work;
	153
	154	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	155	if (work) {
	156	INIT_WORK(work, do_sync_work);
	157	schedule_work(work);
	158	}
cf9a2ae8 DH	159	}
	160
	161	/*
	162	* Generic function to fsync a file.
	163	*
	164	* filp may be NULL if called via the msync of a vma.
	165	*/
	166	int file_fsync(struct file filp, struct dentry dentry, int datasync)
	167	{
	168	struct inode * inode = dentry->d_inode;
	169	struct super_block * sb;
	170	int ret, err;
	171
	172	/* sync the inode to buffers */
	173	ret = write_inode_now(inode, 0);
	174
	175	/* sync the superblock to buffers */
	176	sb = inode->i_sb;
762873c2	177	if (sb->s_dirt && sb->s_op->write_super)
cf9a2ae8	178	sb->s_op->write_super(sb);
cf9a2ae8 DH	179
	180	/* .. finally sync the buffers to disk */
	181	err = sync_blockdev(sb->s_bdev);
	182	if (!ret)
	183	ret = err;
	184	return ret;
	185	}
1fe72eaa	186	EXPORT_SYMBOL(file_fsync);
cf9a2ae8	187
4c728ef5	188	/**
148f948b	189	* vfs_fsync_range - helper to sync a range of data & metadata to disk
4c728ef5 CH	190	* @file: file to sync
4c728ef5 CH	191	* @dentry: dentry of @file
148f948b JK	192	* @start: offset in bytes of the beginning of data range to sync
	193	* @end: offset in bytes of the end of data range (inclusive)
	194	* @datasync: perform only datasync
4c728ef5	195	*
148f948b JK	196	* Write back data in range @start..@end and metadata for @file to disk. If
	197	* @datasync is set only metadata needed to access modified file data is
	198	* written.
4c728ef5 CH	199	*
	200	* In case this function is called from nfsd @file may be %NULL and
	201	* only @dentry is set. This can only happen when the filesystem
	202	* implements the export_operations API.
	203	*/
148f948b JK	204	int vfs_fsync_range(struct file file, struct dentry dentry, loff_t start,
148f948b JK	205	loff_t end, int datasync)
cf9a2ae8	206	{
4c728ef5 CH	207	const struct file_operations *fop;
	208	struct address_space *mapping;
	209	int err, ret;
	210
	211	/*
	212	* Get mapping and operations from the file in case we have
	213	* as file, or get the default values for them in case we
	214	* don't have a struct file available. Damn nfsd..
	215	*/
	216	if (file) {
	217	mapping = file->f_mapping;
	218	fop = file->f_op;
	219	} else {
	220	mapping = dentry->d_inode->i_mapping;
	221	fop = dentry->d_inode->i_fop;
	222	}
cf9a2ae8	223
4c728ef5	224	if (!fop \|\| !fop->fsync) {
cf9a2ae8 DH	225	ret = -EINVAL;
	226	goto out;
	227	}
	228
2daea67e	229	ret = filemap_write_and_wait_range(mapping, start, end);
cf9a2ae8 DH	230
	231	/*
	232	* We need to protect against concurrent writers, which could cause
	233	* livelocks in fsync_buffers_list().
	234	*/
	235	mutex_lock(&mapping->host->i_mutex);
4c728ef5	236	err = fop->fsync(file, dentry, datasync);
cf9a2ae8 DH	237	if (!ret)
	238	ret = err;
	239	mutex_unlock(&mapping->host->i_mutex);
148f948b	240
cf9a2ae8 DH	241	out:
	242	return ret;
	243	}
148f948b JK	244	EXPORT_SYMBOL(vfs_fsync_range);
	245
	246	/**
	247	* vfs_fsync - perform a fsync or fdatasync on a file
	248	* @file: file to sync
	249	* @dentry: dentry of @file
	250	* @datasync: only perform a fdatasync operation
	251	*
	252	* Write back data and metadata for @file to disk. If @datasync is
	253	* set only metadata needed to access modified file data is written.
	254	*
	255	* In case this function is called from nfsd @file may be %NULL and
	256	* only @dentry is set. This can only happen when the filesystem
	257	* implements the export_operations API.
	258	*/
	259	int vfs_fsync(struct file file, struct dentry dentry, int datasync)
	260	{
	261	return vfs_fsync_range(file, dentry, 0, LLONG_MAX, datasync);
	262	}
4c728ef5	263	EXPORT_SYMBOL(vfs_fsync);
cf9a2ae8	264
4c728ef5	265	static int do_fsync(unsigned int fd, int datasync)
cf9a2ae8 DH	266	{
	267	struct file *file;
	268	int ret = -EBADF;
	269
	270	file = fget(fd);
	271	if (file) {
4c728ef5	272	ret = vfs_fsync(file, file->f_path.dentry, datasync);
cf9a2ae8 DH	273	fput(file);
	274	}
	275	return ret;
	276	}
	277
a5f8fa9e	278	SYSCALL_DEFINE1(fsync, unsigned int, fd)
cf9a2ae8	279	{
4c728ef5	280	return do_fsync(fd, 0);
cf9a2ae8 DH	281	}
cf9a2ae8 DH	282
a5f8fa9e	283	SYSCALL_DEFINE1(fdatasync, unsigned int, fd)
cf9a2ae8	284	{
4c728ef5	285	return do_fsync(fd, 1);
cf9a2ae8 DH	286	}
cf9a2ae8 DH	287
148f948b JK	288	/**
	289	* generic_write_sync - perform syncing after a write if file / inode is sync
	290	* @file: file to which the write happened
	291	* @pos: offset where the write started
	292	* @count: length of the write
	293	*
	294	* This is just a simple wrapper about our general syncing function.
	295	*/
	296	int generic_write_sync(struct file *file, loff_t pos, loff_t count)
	297	{
6b2f3d1f	298	if (!(file->f_flags & O_DSYNC) && !IS_SYNC(file->f_mapping->host))
148f948b JK	299	return 0;
148f948b JK	300	return vfs_fsync_range(file, file->f_path.dentry, pos,
6b2f3d1f CH	301	pos + count - 1,
6b2f3d1f CH	302	(file->f_flags & __O_SYNC) ? 0 : 1);
148f948b JK	303	}
	304	EXPORT_SYMBOL(generic_write_sync);
	305
f79e2abb AM	306	/*
	307	* sys_sync_file_range() permits finely controlled syncing over a segment of
	308	* a file in the range offset .. (offset+nbytes-1) inclusive. If nbytes is
	309	* zero then sys_sync_file_range() will operate from offset out to EOF.
	310	*
	311	* The flag bits are:
	312	*
	313	* SYNC_FILE_RANGE_WAIT_BEFORE: wait upon writeout of all pages in the range
	314	* before performing the write.
	315	*
	316	* SYNC_FILE_RANGE_WRITE: initiate writeout of all those dirty pages in the
cce77081 PM	317	* range which are not presently under writeback. Note that this may block for
cce77081 PM	318	* significant periods due to exhaustion of disk request structures.
f79e2abb AM	319	*
	320	* SYNC_FILE_RANGE_WAIT_AFTER: wait upon writeout of all pages in the range
	321	* after performing the write.
	322	*
	323	* Useful combinations of the flag bits are:
	324	*
	325	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE: ensures that all pages
	326	* in the range which were dirty on entry to sys_sync_file_range() are placed
	327	* under writeout. This is a start-write-for-data-integrity operation.
	328	*
	329	* SYNC_FILE_RANGE_WRITE: start writeout of all dirty pages in the range which
	330	* are not presently under writeout. This is an asynchronous flush-to-disk
	331	* operation. Not suitable for data integrity operations.
	332	*
	333	* SYNC_FILE_RANGE_WAIT_BEFORE (or SYNC_FILE_RANGE_WAIT_AFTER): wait for
	334	* completion of writeout of all pages in the range. This will be used after an
	335	* earlier SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE operation to wait
	336	* for that operation to complete and to return the result.
	337	*
	338	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\|SYNC_FILE_RANGE_WAIT_AFTER:
	339	* a traditional sync() operation. This is a write-for-data-integrity operation
	340	* which will ensure that all pages in the range which were dirty on entry to
	341	* sys_sync_file_range() are committed to disk.
	342	*
	343	*
	344	* SYNC_FILE_RANGE_WAIT_BEFORE and SYNC_FILE_RANGE_WAIT_AFTER will detect any
	345	* I/O errors or ENOSPC conditions and will return those to the caller, after
	346	* clearing the EIO and ENOSPC flags in the address_space.
	347	*
	348	* It should be noted that none of these operations write out the file's
	349	* metadata. So unless the application is strictly performing overwrites of
	350	* already-instantiated disk blocks, there are no guarantees here that the data
	351	* will be available after a crash.
	352	*/
6673e0c3 HC	353	SYSCALL_DEFINE(sync_file_range)(int fd, loff_t offset, loff_t nbytes,
6673e0c3 HC	354	unsigned int flags)
f79e2abb AM	355	{
	356	int ret;
	357	struct file *file;
7a0ad10c	358	struct address_space *mapping;
f79e2abb AM	359	loff_t endbyte; /* inclusive */
	360	int fput_needed;
	361	umode_t i_mode;
	362
	363	ret = -EINVAL;
	364	if (flags & ~VALID_FLAGS)
	365	goto out;
	366
	367	endbyte = offset + nbytes;
	368
	369	if ((s64)offset < 0)
	370	goto out;
	371	if ((s64)endbyte < 0)
	372	goto out;
	373	if (endbyte < offset)
	374	goto out;
	375
	376	if (sizeof(pgoff_t) == 4) {
	377	if (offset >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	378	/*
	379	* The range starts outside a 32 bit machine's
	380	* pagecache addressing capabilities. Let it "succeed"
	381	*/
	382	ret = 0;
	383	goto out;
	384	}
	385	if (endbyte >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	386	/*
	387	* Out to EOF
	388	*/
	389	nbytes = 0;
	390	}
	391	}
	392
	393	if (nbytes == 0)
111ebb6e	394	endbyte = LLONG_MAX;
f79e2abb AM	395	else
	396	endbyte--; /* inclusive */
	397
	398	ret = -EBADF;
	399	file = fget_light(fd, &fput_needed);
	400	if (!file)
	401	goto out;
	402
0f7fc9e4	403	i_mode = file->f_path.dentry->d_inode->i_mode;
f79e2abb AM	404	ret = -ESPIPE;
	405	if (!S_ISREG(i_mode) && !S_ISBLK(i_mode) && !S_ISDIR(i_mode) &&
	406	!S_ISLNK(i_mode))
	407	goto out_put;
	408
7a0ad10c CH	409	mapping = file->f_mapping;
	410	if (!mapping) {
	411	ret = -EINVAL;
	412	goto out_put;
	413	}
	414
	415	ret = 0;
	416	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
	417	ret = filemap_fdatawait_range(mapping, offset, endbyte);
	418	if (ret < 0)
	419	goto out_put;
	420	}
	421
	422	if (flags & SYNC_FILE_RANGE_WRITE) {
	423	ret = filemap_fdatawrite_range(mapping, offset, endbyte);
	424	if (ret < 0)
	425	goto out_put;
	426	}
	427
	428	if (flags & SYNC_FILE_RANGE_WAIT_AFTER)
	429	ret = filemap_fdatawait_range(mapping, offset, endbyte);
	430
f79e2abb AM	431	out_put:
	432	fput_light(file, fput_needed);
	433	out:
	434	return ret;
	435	}
6673e0c3 HC	436	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	437	asmlinkage long SyS_sync_file_range(long fd, loff_t offset, loff_t nbytes,
	438	long flags)
	439	{
	440	return SYSC_sync_file_range((int) fd, offset, nbytes,
	441	(unsigned int) flags);
	442	}
	443	SYSCALL_ALIAS(sys_sync_file_range, SyS_sync_file_range);
	444	#endif
f79e2abb	445
edd5cd4a DW	446	/* It would be nice if people remember that not all the world's an i386
edd5cd4a DW	447	when they introduce new system calls */
6673e0c3 HC	448	SYSCALL_DEFINE(sync_file_range2)(int fd, unsigned int flags,
6673e0c3 HC	449	loff_t offset, loff_t nbytes)
edd5cd4a DW	450	{
	451	return sys_sync_file_range(fd, offset, nbytes, flags);
	452	}
6673e0c3 HC	453	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	454	asmlinkage long SyS_sync_file_range2(long fd, long flags,
	455	loff_t offset, loff_t nbytes)
	456	{
	457	return SYSC_sync_file_range2((int) fd, (unsigned int) flags,
	458	offset, nbytes);
	459	}
	460	SYSCALL_ALIAS(sys_sync_file_range2, SyS_sync_file_range2);
	461	#endif