[net-next-2.6.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);

static void sync_one_sb(struct super_block *sb, void *arg)
{
	if (!(sb->s_flags & MS_RDONLY) && sb->s_bdi)
		__sync_filesystem(sb, *(int *)arg);
}
/*
 * Sync all the data for all the filesystems (called by sys_sync() and
 * emergency sync)
 */
static void sync_filesystems(int wait)
{
	iterate_supers(sync_one_sb, &wait);
}

/*
 * 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.
 */
int file_fsync(struct file *filp, int datasync)
{
	struct inode *inode = filp->f_mapping->host;
	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
 * @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.
 */
int vfs_fsync_range(struct file *file, loff_t start, loff_t end, int datasync)
{
	struct address_space *mapping = file->f_mapping;
	int err, ret;

	if (!file->f_op || !file->f_op->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 = file->f_op->fsync(file, 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
 * @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.
 */
int vfs_fsync(struct file *file, int datasync)
{
	return vfs_fsync_range(file, 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, 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, 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	44	else
0e3c9a22	45	writeback_inodes_sb(sb);
5fb324ad	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	79
01a05b33 AV	80	static void sync_one_sb(struct super_block sb, void arg)
	81	{
	82	if (!(sb->s_flags & MS_RDONLY) && sb->s_bdi)
	83	__sync_filesystem(sb, (int )arg);
	84	}
c15c54f5 JK	85	/*
	86	* Sync all the data for all the filesystems (called by sys_sync() and
	87	* emergency sync)
c15c54f5 JK	88	*/
	89	static void sync_filesystems(int wait)
	90	{
01a05b33	91	iterate_supers(sync_one_sb, &wait);
c15c54f5 JK	92	}
c15c54f5 JK	93
3beab0b4 ZY	94	/*
	95	* sync everything. Start out by waking pdflush, because that writes back
	96	* all queues in parallel.
	97	*/
5cee5815	98	SYSCALL_DEFINE0(sync)
cf9a2ae8	99	{
03ba3782	100	wakeup_flusher_threads(0);
5cee5815 JK	101	sync_filesystems(0);
5cee5815 JK	102	sync_filesystems(1);
cf9a2ae8 DH	103	if (unlikely(laptop_mode))
cf9a2ae8 DH	104	laptop_sync_completion();
cf9a2ae8 DH	105	return 0;
	106	}
	107
a2a9537a JA	108	static void do_sync_work(struct work_struct *work)
a2a9537a JA	109	{
5cee5815 JK	110	/*
	111	* Sync twice to reduce the possibility we skipped some inodes / pages
	112	* because they were temporarily locked
	113	*/
	114	sync_filesystems(0);
	115	sync_filesystems(0);
	116	printk("Emergency Sync complete\n");
a2a9537a JA	117	kfree(work);
	118	}
	119
cf9a2ae8 DH	120	void emergency_sync(void)
cf9a2ae8 DH	121	{
a2a9537a JA	122	struct work_struct *work;
	123
	124	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	125	if (work) {
	126	INIT_WORK(work, do_sync_work);
	127	schedule_work(work);
	128	}
cf9a2ae8 DH	129	}
	130
	131	/*
	132	* Generic function to fsync a file.
cf9a2ae8	133	*/
7ea80859	134	int file_fsync(struct file *filp, int datasync)
cf9a2ae8	135	{
7ea80859	136	struct inode *inode = filp->f_mapping->host;
cf9a2ae8 DH	137	struct super_block * sb;
	138	int ret, err;
	139
	140	/* sync the inode to buffers */
	141	ret = write_inode_now(inode, 0);
	142
	143	/* sync the superblock to buffers */
	144	sb = inode->i_sb;
762873c2	145	if (sb->s_dirt && sb->s_op->write_super)
cf9a2ae8	146	sb->s_op->write_super(sb);
cf9a2ae8 DH	147
	148	/* .. finally sync the buffers to disk */
	149	err = sync_blockdev(sb->s_bdev);
	150	if (!ret)
	151	ret = err;
	152	return ret;
	153	}
1fe72eaa	154	EXPORT_SYMBOL(file_fsync);
cf9a2ae8	155
4c728ef5	156	/**
148f948b	157	* vfs_fsync_range - helper to sync a range of data & metadata to disk
4c728ef5	158	* @file: file to sync
148f948b JK	159	* @start: offset in bytes of the beginning of data range to sync
	160	* @end: offset in bytes of the end of data range (inclusive)
	161	* @datasync: perform only datasync
4c728ef5	162	*
148f948b JK	163	* Write back data in range @start..@end and metadata for @file to disk. If
	164	* @datasync is set only metadata needed to access modified file data is
	165	* written.
4c728ef5	166	*/
8018ab05	167	int vfs_fsync_range(struct file *file, loff_t start, loff_t end, int datasync)
cf9a2ae8	168	{
8018ab05	169	struct address_space *mapping = file->f_mapping;
4c728ef5 CH	170	int err, ret;
4c728ef5 CH	171
8018ab05	172	if (!file->f_op \|\| !file->f_op->fsync) {
cf9a2ae8 DH	173	ret = -EINVAL;
	174	goto out;
	175	}
	176
2daea67e	177	ret = filemap_write_and_wait_range(mapping, start, end);
cf9a2ae8 DH	178
	179	/*
	180	* We need to protect against concurrent writers, which could cause
	181	* livelocks in fsync_buffers_list().
	182	*/
	183	mutex_lock(&mapping->host->i_mutex);
7ea80859	184	err = file->f_op->fsync(file, datasync);
cf9a2ae8 DH	185	if (!ret)
	186	ret = err;
	187	mutex_unlock(&mapping->host->i_mutex);
148f948b	188
cf9a2ae8 DH	189	out:
	190	return ret;
	191	}
148f948b JK	192	EXPORT_SYMBOL(vfs_fsync_range);
	193
	194	/**
	195	* vfs_fsync - perform a fsync or fdatasync on a file
	196	* @file: file to sync
148f948b JK	197	* @datasync: only perform a fdatasync operation
	198	*
	199	* Write back data and metadata for @file to disk. If @datasync is
	200	* set only metadata needed to access modified file data is written.
148f948b	201	*/
8018ab05	202	int vfs_fsync(struct file *file, int datasync)
148f948b	203	{
8018ab05	204	return vfs_fsync_range(file, 0, LLONG_MAX, datasync);
148f948b	205	}
4c728ef5	206	EXPORT_SYMBOL(vfs_fsync);
cf9a2ae8	207
4c728ef5	208	static int do_fsync(unsigned int fd, int datasync)
cf9a2ae8 DH	209	{
	210	struct file *file;
	211	int ret = -EBADF;
	212
	213	file = fget(fd);
	214	if (file) {
8018ab05	215	ret = vfs_fsync(file, datasync);
cf9a2ae8 DH	216	fput(file);
	217	}
	218	return ret;
	219	}
	220
a5f8fa9e	221	SYSCALL_DEFINE1(fsync, unsigned int, fd)
cf9a2ae8	222	{
4c728ef5	223	return do_fsync(fd, 0);
cf9a2ae8 DH	224	}
cf9a2ae8 DH	225
a5f8fa9e	226	SYSCALL_DEFINE1(fdatasync, unsigned int, fd)
cf9a2ae8	227	{
4c728ef5	228	return do_fsync(fd, 1);
cf9a2ae8 DH	229	}
cf9a2ae8 DH	230
148f948b JK	231	/**
	232	* generic_write_sync - perform syncing after a write if file / inode is sync
	233	* @file: file to which the write happened
	234	* @pos: offset where the write started
	235	* @count: length of the write
	236	*
	237	* This is just a simple wrapper about our general syncing function.
	238	*/
	239	int generic_write_sync(struct file *file, loff_t pos, loff_t count)
	240	{
6b2f3d1f	241	if (!(file->f_flags & O_DSYNC) && !IS_SYNC(file->f_mapping->host))
148f948b	242	return 0;
8018ab05	243	return vfs_fsync_range(file, pos, pos + count - 1,
6b2f3d1f	244	(file->f_flags & __O_SYNC) ? 0 : 1);
148f948b JK	245	}
	246	EXPORT_SYMBOL(generic_write_sync);
	247
f79e2abb AM	248	/*
	249	* sys_sync_file_range() permits finely controlled syncing over a segment of
	250	* a file in the range offset .. (offset+nbytes-1) inclusive. If nbytes is
	251	* zero then sys_sync_file_range() will operate from offset out to EOF.
	252	*
	253	* The flag bits are:
	254	*
	255	* SYNC_FILE_RANGE_WAIT_BEFORE: wait upon writeout of all pages in the range
	256	* before performing the write.
	257	*
	258	* SYNC_FILE_RANGE_WRITE: initiate writeout of all those dirty pages in the
cce77081 PM	259	* range which are not presently under writeback. Note that this may block for
cce77081 PM	260	* significant periods due to exhaustion of disk request structures.
f79e2abb AM	261	*
	262	* SYNC_FILE_RANGE_WAIT_AFTER: wait upon writeout of all pages in the range
	263	* after performing the write.
	264	*
	265	* Useful combinations of the flag bits are:
	266	*
	267	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE: ensures that all pages
	268	* in the range which were dirty on entry to sys_sync_file_range() are placed
	269	* under writeout. This is a start-write-for-data-integrity operation.
	270	*
	271	* SYNC_FILE_RANGE_WRITE: start writeout of all dirty pages in the range which
	272	* are not presently under writeout. This is an asynchronous flush-to-disk
	273	* operation. Not suitable for data integrity operations.
	274	*
	275	* SYNC_FILE_RANGE_WAIT_BEFORE (or SYNC_FILE_RANGE_WAIT_AFTER): wait for
	276	* completion of writeout of all pages in the range. This will be used after an
	277	* earlier SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE operation to wait
	278	* for that operation to complete and to return the result.
	279	*
	280	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\|SYNC_FILE_RANGE_WAIT_AFTER:
	281	* a traditional sync() operation. This is a write-for-data-integrity operation
	282	* which will ensure that all pages in the range which were dirty on entry to
	283	* sys_sync_file_range() are committed to disk.
	284	*
	285	*
	286	* SYNC_FILE_RANGE_WAIT_BEFORE and SYNC_FILE_RANGE_WAIT_AFTER will detect any
	287	* I/O errors or ENOSPC conditions and will return those to the caller, after
	288	* clearing the EIO and ENOSPC flags in the address_space.
	289	*
	290	* It should be noted that none of these operations write out the file's
	291	* metadata. So unless the application is strictly performing overwrites of
	292	* already-instantiated disk blocks, there are no guarantees here that the data
	293	* will be available after a crash.
	294	*/
6673e0c3 HC	295	SYSCALL_DEFINE(sync_file_range)(int fd, loff_t offset, loff_t nbytes,
6673e0c3 HC	296	unsigned int flags)
f79e2abb AM	297	{
	298	int ret;
	299	struct file *file;
7a0ad10c	300	struct address_space *mapping;
f79e2abb AM	301	loff_t endbyte; /* inclusive */
	302	int fput_needed;
	303	umode_t i_mode;
	304
	305	ret = -EINVAL;
	306	if (flags & ~VALID_FLAGS)
	307	goto out;
	308
	309	endbyte = offset + nbytes;
	310
	311	if ((s64)offset < 0)
	312	goto out;
	313	if ((s64)endbyte < 0)
	314	goto out;
	315	if (endbyte < offset)
	316	goto out;
	317
	318	if (sizeof(pgoff_t) == 4) {
	319	if (offset >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	320	/*
	321	* The range starts outside a 32 bit machine's
	322	* pagecache addressing capabilities. Let it "succeed"
	323	*/
	324	ret = 0;
	325	goto out;
	326	}
	327	if (endbyte >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	328	/*
	329	* Out to EOF
	330	*/
	331	nbytes = 0;
	332	}
	333	}
	334
	335	if (nbytes == 0)
111ebb6e	336	endbyte = LLONG_MAX;
f79e2abb AM	337	else
	338	endbyte--; /* inclusive */
	339
	340	ret = -EBADF;
	341	file = fget_light(fd, &fput_needed);
	342	if (!file)
	343	goto out;
	344
0f7fc9e4	345	i_mode = file->f_path.dentry->d_inode->i_mode;
f79e2abb AM	346	ret = -ESPIPE;
	347	if (!S_ISREG(i_mode) && !S_ISBLK(i_mode) && !S_ISDIR(i_mode) &&
	348	!S_ISLNK(i_mode))
	349	goto out_put;
	350
7a0ad10c CH	351	mapping = file->f_mapping;
	352	if (!mapping) {
	353	ret = -EINVAL;
	354	goto out_put;
	355	}
	356
	357	ret = 0;
	358	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
	359	ret = filemap_fdatawait_range(mapping, offset, endbyte);
	360	if (ret < 0)
	361	goto out_put;
	362	}
	363
	364	if (flags & SYNC_FILE_RANGE_WRITE) {
	365	ret = filemap_fdatawrite_range(mapping, offset, endbyte);
	366	if (ret < 0)
	367	goto out_put;
	368	}
	369
	370	if (flags & SYNC_FILE_RANGE_WAIT_AFTER)
	371	ret = filemap_fdatawait_range(mapping, offset, endbyte);
	372
f79e2abb AM	373	out_put:
	374	fput_light(file, fput_needed);
	375	out:
	376	return ret;
	377	}
6673e0c3 HC	378	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	379	asmlinkage long SyS_sync_file_range(long fd, loff_t offset, loff_t nbytes,
	380	long flags)
	381	{
	382	return SYSC_sync_file_range((int) fd, offset, nbytes,
	383	(unsigned int) flags);
	384	}
	385	SYSCALL_ALIAS(sys_sync_file_range, SyS_sync_file_range);
	386	#endif
f79e2abb	387
edd5cd4a DW	388	/* It would be nice if people remember that not all the world's an i386
edd5cd4a DW	389	when they introduce new system calls */
6673e0c3 HC	390	SYSCALL_DEFINE(sync_file_range2)(int fd, unsigned int flags,
6673e0c3 HC	391	loff_t offset, loff_t nbytes)
edd5cd4a DW	392	{
	393	return sys_sync_file_range(fd, offset, nbytes, flags);
	394	}
6673e0c3 HC	395	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	396	asmlinkage long SyS_sync_file_range2(long fd, long flags,
	397	loff_t offset, loff_t nbytes)
	398	{
	399	return SYSC_sync_file_range2((int) fd, (unsigned int) flags,
	400	offset, nbytes);
	401	}
	402	SYSCALL_ALIAS(sys_sync_file_range2, SyS_sync_file_range2);
	403	#endif