[mirror_ubuntu-eoan-kernel.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 sync_inodes_sb(sb, 0)
 * 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)
{
	/* Avoid doing twice syncing and cache pruning for quota sync */
	if (!wait)
		writeout_quota_sb(sb, -1);
	else
		sync_quota_sb(sb, -1);
	sync_inodes_sb(sb, wait);
	lock_super(sb);
	if (sb->s_dirt && sb->s_op->write_super)
		sb->s_op->write_super(sb);
	unlock_super(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;

	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) {
		if (sb->s_flags & MS_RDONLY)
			continue;
		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;
		if (sb->s_flags & MS_RDONLY)
			continue;	/* hm.  Was remounted r/o meanwhile */
		sb->s_count++;
		spin_unlock(&sb_lock);
		down_read(&sb->s_umount);
		if (sb->s_root)
			__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);
}

SYSCALL_DEFINE0(sync)
{
	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;
	lock_super(sb);
	if (sb->s_dirt && sb->s_op->write_super)
		sb->s_op->write_super(sb);
	unlock_super(sb);

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

/**
 * vfs_fsync - perform a fsync or fdatasync on a file
 * @file:		file to sync
 * @dentry:		dentry of @file
 * @data:		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)
{
	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_fdatawrite(mapping);

	/*
	 * 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);
	err = filemap_fdatawait(mapping);
	if (!ret)
		ret = err;
out:
	return ret;
}
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);
}

/*
 * 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;
	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;

	ret = do_sync_mapping_range(file->f_mapping, offset, endbyte, flags);
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

/*
 * `endbyte' is inclusive
 */
int do_sync_mapping_range(struct address_space *mapping, loff_t offset,
			  loff_t endbyte, unsigned int flags)
{
	int ret;

	if (!mapping) {
		ret = -EINVAL;
		goto out;
	}

	ret = 0;
	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
		ret = wait_on_page_writeback_range(mapping,
					offset >> PAGE_CACHE_SHIFT,
					endbyte >> PAGE_CACHE_SHIFT);
		if (ret < 0)
			goto out;
	}

	if (flags & SYNC_FILE_RANGE_WRITE) {
		ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
						WB_SYNC_ALL);
		if (ret < 0)
			goto out;
	}

	if (flags & SYNC_FILE_RANGE_WAIT_AFTER) {
		ret = wait_on_page_writeback_range(mapping,
					offset >> PAGE_CACHE_SHIFT,
					endbyte >> PAGE_CACHE_SHIFT);
	}
out:
	return ret;
}
EXPORT_SYMBOL_GPL(do_sync_mapping_range);
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 JK	21	/*
	22	* Do the filesystem syncing work. For simple filesystems sync_inodes_sb(sb, 0)
	23	* just dirties buffers with inodes so we have to submit IO for these buffers
	24	* via __sync_blockdev(). This also speeds up the wait == 1 case since in that
	25	* case write_inode() functions do sync_dirty_buffer() and thus effectively
	26	* write one block at a time.
	27	*/
60b0680f	28	static int __sync_filesystem(struct super_block *sb, int wait)
c15c54f5	29	{
c3f8a40c JK	30	/* Avoid doing twice syncing and cache pruning for quota sync */
	31	if (!wait)
	32	writeout_quota_sb(sb, -1);
	33	else
	34	sync_quota_sb(sb, -1);
c15c54f5 JK	35	sync_inodes_sb(sb, wait);
	36	lock_super(sb);
	37	if (sb->s_dirt && sb->s_op->write_super)
	38	sb->s_op->write_super(sb);
	39	unlock_super(sb);
	40	if (sb->s_op->sync_fs)
	41	sb->s_op->sync_fs(sb, wait);
	42	return __sync_blockdev(sb->s_bdev, wait);
	43	}
	44
	45	/*
	46	* Write out and wait upon all dirty data associated with this
	47	* superblock. Filesystem data as well as the underlying block
	48	* device. Takes the superblock lock.
	49	*/
60b0680f	50	int sync_filesystem(struct super_block *sb)
c15c54f5 JK	51	{
	52	int ret;
	53
60b0680f	54	ret = __sync_filesystem(sb, 0);
c15c54f5 JK	55	if (ret < 0)
c15c54f5 JK	56	return ret;
60b0680f	57	return __sync_filesystem(sb, 1);
c15c54f5	58	}
60b0680f	59	EXPORT_SYMBOL_GPL(sync_filesystem);
c15c54f5 JK	60
	61	/*
	62	* Sync all the data for all the filesystems (called by sys_sync() and
	63	* emergency sync)
	64	*
	65	* This operation is careful to avoid the livelock which could easily happen
	66	* if two or more filesystems are being continuously dirtied. s_need_sync
	67	* is used only here. We set it against all filesystems and then clear it as
	68	* we sync them. So redirtied filesystems are skipped.
	69	*
	70	* But if process A is currently running sync_filesystems and then process B
	71	* calls sync_filesystems as well, process B will set all the s_need_sync
	72	* flags again, which will cause process A to resync everything. Fix that with
	73	* a local mutex.
	74	*/
	75	static void sync_filesystems(int wait)
	76	{
	77	struct super_block *sb;
	78	static DEFINE_MUTEX(mutex);
	79
	80	mutex_lock(&mutex); /* Could be down_interruptible */
	81	spin_lock(&sb_lock);
	82	list_for_each_entry(sb, &super_blocks, s_list) {
	83	if (sb->s_flags & MS_RDONLY)
	84	continue;
	85	sb->s_need_sync = 1;
	86	}
	87
	88	restart:
	89	list_for_each_entry(sb, &super_blocks, s_list) {
	90	if (!sb->s_need_sync)
	91	continue;
	92	sb->s_need_sync = 0;
	93	if (sb->s_flags & MS_RDONLY)
	94	continue; /* hm. Was remounted r/o meanwhile */
	95	sb->s_count++;
	96	spin_unlock(&sb_lock);
	97	down_read(&sb->s_umount);
	98	if (sb->s_root)
60b0680f	99	__sync_filesystem(sb, wait);
c15c54f5 JK	100	up_read(&sb->s_umount);
	101	/* restart only when sb is no longer on the list */
	102	spin_lock(&sb_lock);
	103	if (__put_super_and_need_restart(sb))
	104	goto restart;
	105	}
	106	spin_unlock(&sb_lock);
	107	mutex_unlock(&mutex);
	108	}
	109
5cee5815	110	SYSCALL_DEFINE0(sync)
cf9a2ae8	111	{
5cee5815 JK	112	sync_filesystems(0);
5cee5815 JK	113	sync_filesystems(1);
cf9a2ae8 DH	114	if (unlikely(laptop_mode))
cf9a2ae8 DH	115	laptop_sync_completion();
cf9a2ae8 DH	116	return 0;
	117	}
	118
a2a9537a JA	119	static void do_sync_work(struct work_struct *work)
a2a9537a JA	120	{
5cee5815 JK	121	/*
	122	* Sync twice to reduce the possibility we skipped some inodes / pages
	123	* because they were temporarily locked
	124	*/
	125	sync_filesystems(0);
	126	sync_filesystems(0);
	127	printk("Emergency Sync complete\n");
a2a9537a JA	128	kfree(work);
	129	}
	130
cf9a2ae8 DH	131	void emergency_sync(void)
cf9a2ae8 DH	132	{
a2a9537a JA	133	struct work_struct *work;
	134
	135	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	136	if (work) {
	137	INIT_WORK(work, do_sync_work);
	138	schedule_work(work);
	139	}
cf9a2ae8 DH	140	}
	141
	142	/*
	143	* Generic function to fsync a file.
	144	*
	145	* filp may be NULL if called via the msync of a vma.
	146	*/
	147	int file_fsync(struct file filp, struct dentry dentry, int datasync)
	148	{
	149	struct inode * inode = dentry->d_inode;
	150	struct super_block * sb;
	151	int ret, err;
	152
	153	/* sync the inode to buffers */
	154	ret = write_inode_now(inode, 0);
	155
	156	/* sync the superblock to buffers */
	157	sb = inode->i_sb;
	158	lock_super(sb);
762873c2	159	if (sb->s_dirt && sb->s_op->write_super)
cf9a2ae8 DH	160	sb->s_op->write_super(sb);
	161	unlock_super(sb);
	162
	163	/* .. finally sync the buffers to disk */
	164	err = sync_blockdev(sb->s_bdev);
	165	if (!ret)
	166	ret = err;
	167	return ret;
	168	}
	169
4c728ef5 CH	170	/**
	171	* vfs_fsync - perform a fsync or fdatasync on a file
	172	* @file: file to sync
	173	* @dentry: dentry of @file
	174	* @data: only perform a fdatasync operation
	175	*
	176	* Write back data and metadata for @file to disk. If @datasync is
	177	* set only metadata needed to access modified file data is written.
	178	*
	179	* In case this function is called from nfsd @file may be %NULL and
	180	* only @dentry is set. This can only happen when the filesystem
	181	* implements the export_operations API.
	182	*/
	183	int vfs_fsync(struct file file, struct dentry dentry, int datasync)
cf9a2ae8	184	{
4c728ef5 CH	185	const struct file_operations *fop;
	186	struct address_space *mapping;
	187	int err, ret;
	188
	189	/*
	190	* Get mapping and operations from the file in case we have
	191	* as file, or get the default values for them in case we
	192	* don't have a struct file available. Damn nfsd..
	193	*/
	194	if (file) {
	195	mapping = file->f_mapping;
	196	fop = file->f_op;
	197	} else {
	198	mapping = dentry->d_inode->i_mapping;
	199	fop = dentry->d_inode->i_fop;
	200	}
cf9a2ae8	201
4c728ef5	202	if (!fop \|\| !fop->fsync) {
cf9a2ae8 DH	203	ret = -EINVAL;
	204	goto out;
	205	}
	206
	207	ret = filemap_fdatawrite(mapping);
	208
	209	/*
	210	* We need to protect against concurrent writers, which could cause
	211	* livelocks in fsync_buffers_list().
	212	*/
	213	mutex_lock(&mapping->host->i_mutex);
4c728ef5	214	err = fop->fsync(file, dentry, datasync);
cf9a2ae8 DH	215	if (!ret)
	216	ret = err;
	217	mutex_unlock(&mapping->host->i_mutex);
	218	err = filemap_fdatawait(mapping);
	219	if (!ret)
	220	ret = err;
	221	out:
	222	return ret;
	223	}
4c728ef5	224	EXPORT_SYMBOL(vfs_fsync);
cf9a2ae8	225
4c728ef5	226	static int do_fsync(unsigned int fd, int datasync)
cf9a2ae8 DH	227	{
	228	struct file *file;
	229	int ret = -EBADF;
	230
	231	file = fget(fd);
	232	if (file) {
4c728ef5	233	ret = vfs_fsync(file, file->f_path.dentry, datasync);
cf9a2ae8 DH	234	fput(file);
	235	}
	236	return ret;
	237	}
	238
a5f8fa9e	239	SYSCALL_DEFINE1(fsync, unsigned int, fd)
cf9a2ae8	240	{
4c728ef5	241	return do_fsync(fd, 0);
cf9a2ae8 DH	242	}
cf9a2ae8 DH	243
a5f8fa9e	244	SYSCALL_DEFINE1(fdatasync, unsigned int, fd)
cf9a2ae8	245	{
4c728ef5	246	return do_fsync(fd, 1);
cf9a2ae8 DH	247	}
cf9a2ae8 DH	248
f79e2abb AM	249	/*
	250	* sys_sync_file_range() permits finely controlled syncing over a segment of
	251	* a file in the range offset .. (offset+nbytes-1) inclusive. If nbytes is
	252	* zero then sys_sync_file_range() will operate from offset out to EOF.
	253	*
	254	* The flag bits are:
	255	*
	256	* SYNC_FILE_RANGE_WAIT_BEFORE: wait upon writeout of all pages in the range
	257	* before performing the write.
	258	*
	259	* SYNC_FILE_RANGE_WRITE: initiate writeout of all those dirty pages in the
cce77081 PM	260	* range which are not presently under writeback. Note that this may block for
cce77081 PM	261	* significant periods due to exhaustion of disk request structures.
f79e2abb AM	262	*
	263	* SYNC_FILE_RANGE_WAIT_AFTER: wait upon writeout of all pages in the range
	264	* after performing the write.
	265	*
	266	* Useful combinations of the flag bits are:
	267	*
	268	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE: ensures that all pages
	269	* in the range which were dirty on entry to sys_sync_file_range() are placed
	270	* under writeout. This is a start-write-for-data-integrity operation.
	271	*
	272	* SYNC_FILE_RANGE_WRITE: start writeout of all dirty pages in the range which
	273	* are not presently under writeout. This is an asynchronous flush-to-disk
	274	* operation. Not suitable for data integrity operations.
	275	*
	276	* SYNC_FILE_RANGE_WAIT_BEFORE (or SYNC_FILE_RANGE_WAIT_AFTER): wait for
	277	* completion of writeout of all pages in the range. This will be used after an
	278	* earlier SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE operation to wait
	279	* for that operation to complete and to return the result.
	280	*
	281	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\|SYNC_FILE_RANGE_WAIT_AFTER:
	282	* a traditional sync() operation. This is a write-for-data-integrity operation
	283	* which will ensure that all pages in the range which were dirty on entry to
	284	* sys_sync_file_range() are committed to disk.
	285	*
	286	*
	287	* SYNC_FILE_RANGE_WAIT_BEFORE and SYNC_FILE_RANGE_WAIT_AFTER will detect any
	288	* I/O errors or ENOSPC conditions and will return those to the caller, after
	289	* clearing the EIO and ENOSPC flags in the address_space.
	290	*
	291	* It should be noted that none of these operations write out the file's
	292	* metadata. So unless the application is strictly performing overwrites of
	293	* already-instantiated disk blocks, there are no guarantees here that the data
	294	* will be available after a crash.
	295	*/
6673e0c3 HC	296	SYSCALL_DEFINE(sync_file_range)(int fd, loff_t offset, loff_t nbytes,
6673e0c3 HC	297	unsigned int flags)
f79e2abb AM	298	{
	299	int ret;
	300	struct file *file;
	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
ef51c976	351	ret = do_sync_mapping_range(file->f_mapping, offset, endbyte, flags);
f79e2abb AM	352	out_put:
	353	fput_light(file, fput_needed);
	354	out:
	355	return ret;
	356	}
6673e0c3 HC	357	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	358	asmlinkage long SyS_sync_file_range(long fd, loff_t offset, loff_t nbytes,
	359	long flags)
	360	{
	361	return SYSC_sync_file_range((int) fd, offset, nbytes,
	362	(unsigned int) flags);
	363	}
	364	SYSCALL_ALIAS(sys_sync_file_range, SyS_sync_file_range);
	365	#endif
f79e2abb	366
edd5cd4a DW	367	/* It would be nice if people remember that not all the world's an i386
edd5cd4a DW	368	when they introduce new system calls */
6673e0c3 HC	369	SYSCALL_DEFINE(sync_file_range2)(int fd, unsigned int flags,
6673e0c3 HC	370	loff_t offset, loff_t nbytes)
edd5cd4a DW	371	{
	372	return sys_sync_file_range(fd, offset, nbytes, flags);
	373	}
6673e0c3 HC	374	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	375	asmlinkage long SyS_sync_file_range2(long fd, long flags,
	376	loff_t offset, loff_t nbytes)
	377	{
	378	return SYSC_sync_file_range2((int) fd, (unsigned int) flags,
	379	offset, nbytes);
	380	}
	381	SYSCALL_ALIAS(sys_sync_file_range2, SyS_sync_file_range2);
	382	#endif
edd5cd4a	383
f79e2abb AM	384	/*
	385	* `endbyte' is inclusive
	386	*/
5b04aa3a MF	387	int do_sync_mapping_range(struct address_space *mapping, loff_t offset,
5b04aa3a MF	388	loff_t endbyte, unsigned int flags)
f79e2abb AM	389	{
f79e2abb AM	390	int ret;
f79e2abb	391
f79e2abb AM	392	if (!mapping) {
	393	ret = -EINVAL;
	394	goto out;
	395	}
	396
	397	ret = 0;
	398	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
	399	ret = wait_on_page_writeback_range(mapping,
	400	offset >> PAGE_CACHE_SHIFT,
	401	endbyte >> PAGE_CACHE_SHIFT);
	402	if (ret < 0)
	403	goto out;
	404	}
	405
	406	if (flags & SYNC_FILE_RANGE_WRITE) {
	407	ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
ee53a891	408	WB_SYNC_ALL);
f79e2abb AM	409	if (ret < 0)
	410	goto out;
	411	}
	412
	413	if (flags & SYNC_FILE_RANGE_WAIT_AFTER) {
	414	ret = wait_on_page_writeback_range(mapping,
	415	offset >> PAGE_CACHE_SHIFT,
	416	endbyte >> PAGE_CACHE_SHIFT);
	417	}
	418	out:
	419	return ret;
	420	}
5b04aa3a	421	EXPORT_SYMBOL_GPL(do_sync_mapping_range);