[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/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>

#define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \
			SYNC_FILE_RANGE_WAIT_AFTER)

/*
 * sync everything.  Start out by waking pdflush, because that writes back
 * all queues in parallel.
 */
static void do_sync(unsigned long wait)
{
	wakeup_pdflush(0);
	sync_inodes(0);		/* All mappings, inodes and their blockdevs */
	DQUOT_SYNC(NULL);
	sync_supers();		/* Write the superblocks */
	sync_filesystems(0);	/* Start syncing the filesystems */
	sync_filesystems(wait);	/* Waitingly sync the filesystems */
	sync_inodes(wait);	/* Mappings, inodes and blockdevs, again. */
	if (!wait)
		printk("Emergency Sync complete\n");
	if (unlikely(laptop_mode))
		laptop_sync_completion();
}

asmlinkage long sys_sync(void)
{
	do_sync(1);
	return 0;
}

void emergency_sync(void)
{
	pdflush_operation(do_sync, 0);
}

/*
 * 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_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;
}

long do_fsync(struct file *file, int datasync)
{
	int ret;
	int err;
	struct address_space *mapping = file->f_mapping;

	if (!file->f_op || !file->f_op->fsync) {
		/* Why?  We can still call filemap_fdatawrite */
		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 = file->f_op->fsync(file, file->f_path.dentry, datasync);
	if (!ret)
		ret = err;
	mutex_unlock(&mapping->host->i_mutex);
	err = filemap_fdatawait(mapping);
	if (!ret)
		ret = err;
out:
	return ret;
}

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

	file = fget(fd);
	if (file) {
		ret = do_fsync(file, datasync);
		fput(file);
	}
	return ret;
}

asmlinkage long sys_fsync(unsigned int fd)
{
	return __do_fsync(fd, 0);
}

asmlinkage long sys_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.
 *
 * 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.
 */
asmlinkage long sys_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;
}

/*
 * `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_NONE);
		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>
f79e2abb AM	16
	17	#define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\| \
	18	SYNC_FILE_RANGE_WAIT_AFTER)
	19
cf9a2ae8 DH	20	/*
	21	* sync everything. Start out by waking pdflush, because that writes back
	22	* all queues in parallel.
	23	*/
	24	static void do_sync(unsigned long wait)
	25	{
	26	wakeup_pdflush(0);
	27	sync_inodes(0); /* All mappings, inodes and their blockdevs */
	28	DQUOT_SYNC(NULL);
	29	sync_supers(); /* Write the superblocks */
	30	sync_filesystems(0); /* Start syncing the filesystems */
	31	sync_filesystems(wait); /* Waitingly sync the filesystems */
	32	sync_inodes(wait); /* Mappings, inodes and blockdevs, again. */
	33	if (!wait)
	34	printk("Emergency Sync complete\n");
	35	if (unlikely(laptop_mode))
	36	laptop_sync_completion();
	37	}
	38
	39	asmlinkage long sys_sync(void)
	40	{
	41	do_sync(1);
	42	return 0;
	43	}
	44
	45	void emergency_sync(void)
	46	{
	47	pdflush_operation(do_sync, 0);
	48	}
	49
	50	/*
	51	* Generic function to fsync a file.
	52	*
	53	* filp may be NULL if called via the msync of a vma.
	54	*/
	55	int file_fsync(struct file filp, struct dentry dentry, int datasync)
	56	{
	57	struct inode * inode = dentry->d_inode;
	58	struct super_block * sb;
	59	int ret, err;
	60
	61	/* sync the inode to buffers */
	62	ret = write_inode_now(inode, 0);
	63
	64	/* sync the superblock to buffers */
	65	sb = inode->i_sb;
	66	lock_super(sb);
	67	if (sb->s_op->write_super)
	68	sb->s_op->write_super(sb);
	69	unlock_super(sb);
	70
	71	/* .. finally sync the buffers to disk */
	72	err = sync_blockdev(sb->s_bdev);
	73	if (!ret)
	74	ret = err;
	75	return ret;
	76	}
	77
	78	long do_fsync(struct file *file, int datasync)
	79	{
	80	int ret;
	81	int err;
	82	struct address_space *mapping = file->f_mapping;
	83
84	if (!file->f_op \|\| !file->f_op->fsync) {
85	/* Why? We can still call filemap_fdatawrite */
86	ret = -EINVAL;
87	goto out;
88	}
89
90	ret = filemap_fdatawrite(mapping);
91
92	/*
93	* We need to protect against concurrent writers, which could cause
94	* livelocks in fsync_buffers_list().
95	*/
96	mutex_lock(&mapping->host->i_mutex);
0f7fc9e4	97	err = file->f_op->fsync(file, file->f_path.dentry, datasync);
cf9a2ae8 DH	98	if (!ret)
	99	ret = err;
	100	mutex_unlock(&mapping->host->i_mutex);
	101	err = filemap_fdatawait(mapping);
	102	if (!ret)
	103	ret = err;
	104	out:
	105	return ret;
	106	}
	107
	108	static long __do_fsync(unsigned int fd, int datasync)
	109	{
	110	struct file *file;
	111	int ret = -EBADF;
	112
	113	file = fget(fd);
	114	if (file) {
	115	ret = do_fsync(file, datasync);
	116	fput(file);
	117	}
	118	return ret;
	119	}
	120
	121	asmlinkage long sys_fsync(unsigned int fd)
	122	{
	123	return __do_fsync(fd, 0);
	124	}
	125
	126	asmlinkage long sys_fdatasync(unsigned int fd)
	127	{
	128	return __do_fsync(fd, 1);
	129	}
	130
f79e2abb AM	131	/*
	132	* sys_sync_file_range() permits finely controlled syncing over a segment of
	133	* a file in the range offset .. (offset+nbytes-1) inclusive. If nbytes is
	134	* zero then sys_sync_file_range() will operate from offset out to EOF.
	135	*
	136	* The flag bits are:
	137	*
	138	* SYNC_FILE_RANGE_WAIT_BEFORE: wait upon writeout of all pages in the range
	139	* before performing the write.
	140	*
	141	* SYNC_FILE_RANGE_WRITE: initiate writeout of all those dirty pages in the
	142	* range which are not presently under writeback.
	143	*
	144	* SYNC_FILE_RANGE_WAIT_AFTER: wait upon writeout of all pages in the range
	145	* after performing the write.
	146	*
	147	* Useful combinations of the flag bits are:
	148	*
	149	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE: ensures that all pages
	150	* in the range which were dirty on entry to sys_sync_file_range() are placed
	151	* under writeout. This is a start-write-for-data-integrity operation.
	152	*
	153	* SYNC_FILE_RANGE_WRITE: start writeout of all dirty pages in the range which
	154	* are not presently under writeout. This is an asynchronous flush-to-disk
	155	* operation. Not suitable for data integrity operations.
	156	*
	157	* SYNC_FILE_RANGE_WAIT_BEFORE (or SYNC_FILE_RANGE_WAIT_AFTER): wait for
	158	* completion of writeout of all pages in the range. This will be used after an
	159	* earlier SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE operation to wait
	160	* for that operation to complete and to return the result.
	161	*
	162	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\|SYNC_FILE_RANGE_WAIT_AFTER:
	163	* a traditional sync() operation. This is a write-for-data-integrity operation
	164	* which will ensure that all pages in the range which were dirty on entry to
	165	* sys_sync_file_range() are committed to disk.
	166	*
	167	*
	168	* SYNC_FILE_RANGE_WAIT_BEFORE and SYNC_FILE_RANGE_WAIT_AFTER will detect any
	169	* I/O errors or ENOSPC conditions and will return those to the caller, after
	170	* clearing the EIO and ENOSPC flags in the address_space.
	171	*
	172	* It should be noted that none of these operations write out the file's
	173	* metadata. So unless the application is strictly performing overwrites of
	174	* already-instantiated disk blocks, there are no guarantees here that the data
	175	* will be available after a crash.
	176	*/
	177	asmlinkage long sys_sync_file_range(int fd, loff_t offset, loff_t nbytes,
5246d050	178	unsigned int flags)
f79e2abb AM	179	{
	180	int ret;
	181	struct file *file;
	182	loff_t endbyte; /* inclusive */
	183	int fput_needed;
	184	umode_t i_mode;
	185
	186	ret = -EINVAL;
	187	if (flags & ~VALID_FLAGS)
	188	goto out;
	189
	190	endbyte = offset + nbytes;
	191
	192	if ((s64)offset < 0)
	193	goto out;
	194	if ((s64)endbyte < 0)
	195	goto out;
	196	if (endbyte < offset)
	197	goto out;
	198
	199	if (sizeof(pgoff_t) == 4) {
	200	if (offset >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	201	/*
	202	* The range starts outside a 32 bit machine's
	203	* pagecache addressing capabilities. Let it "succeed"
	204	*/
	205	ret = 0;
	206	goto out;
	207	}
	208	if (endbyte >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	209	/*
	210	* Out to EOF
	211	*/
	212	nbytes = 0;
	213	}
	214	}
	215
	216	if (nbytes == 0)
111ebb6e	217	endbyte = LLONG_MAX;
f79e2abb AM	218	else
	219	endbyte--; /* inclusive */
	220
	221	ret = -EBADF;
	222	file = fget_light(fd, &fput_needed);
	223	if (!file)
	224	goto out;
	225
0f7fc9e4	226	i_mode = file->f_path.dentry->d_inode->i_mode;
f79e2abb AM	227	ret = -ESPIPE;
	228	if (!S_ISREG(i_mode) && !S_ISBLK(i_mode) && !S_ISDIR(i_mode) &&
	229	!S_ISLNK(i_mode))
	230	goto out_put;
	231
ef51c976	232	ret = do_sync_mapping_range(file->f_mapping, offset, endbyte, flags);
f79e2abb AM	233	out_put:
	234	fput_light(file, fput_needed);
	235	out:
	236	return ret;
	237	}
	238
	239	/*
	240	* `endbyte' is inclusive
	241	*/
5b04aa3a MF	242	int do_sync_mapping_range(struct address_space *mapping, loff_t offset,
5b04aa3a MF	243	loff_t endbyte, unsigned int flags)
f79e2abb AM	244	{
f79e2abb AM	245	int ret;
f79e2abb	246
f79e2abb AM	247	if (!mapping) {
	248	ret = -EINVAL;
	249	goto out;
	250	}
	251
	252	ret = 0;
	253	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
	254	ret = wait_on_page_writeback_range(mapping,
	255	offset >> PAGE_CACHE_SHIFT,
	256	endbyte >> PAGE_CACHE_SHIFT);
	257	if (ret < 0)
	258	goto out;
	259	}
	260
	261	if (flags & SYNC_FILE_RANGE_WRITE) {
	262	ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
	263	WB_SYNC_NONE);
	264	if (ret < 0)
	265	goto out;
	266	}
	267
	268	if (flags & SYNC_FILE_RANGE_WAIT_AFTER) {
	269	ret = wait_on_page_writeback_range(mapping,
	270	offset >> PAGE_CACHE_SHIFT,
	271	endbyte >> PAGE_CACHE_SHIFT);
	272	}
	273	out:
	274	return ret;
	275	}
5b04aa3a	276	EXPORT_SYMBOL_GPL(do_sync_mapping_range);