[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>
#include "internal.h"

#define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \
			SYNC_FILE_RANGE_WAIT_AFTER)

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
5cee5815	21	SYSCALL_DEFINE0(sync)
cf9a2ae8	22	{
5cee5815 JK	23	sync_filesystems(0);
5cee5815 JK	24	sync_filesystems(1);
cf9a2ae8 DH	25	if (unlikely(laptop_mode))
cf9a2ae8 DH	26	laptop_sync_completion();
cf9a2ae8 DH	27	return 0;
	28	}
	29
a2a9537a JA	30	static void do_sync_work(struct work_struct *work)
a2a9537a JA	31	{
5cee5815 JK	32	/*
	33	* Sync twice to reduce the possibility we skipped some inodes / pages
	34	* because they were temporarily locked
	35	*/
	36	sync_filesystems(0);
	37	sync_filesystems(0);
	38	printk("Emergency Sync complete\n");
a2a9537a JA	39	kfree(work);
	40	}
	41
cf9a2ae8 DH	42	void emergency_sync(void)
cf9a2ae8 DH	43	{
a2a9537a JA	44	struct work_struct *work;
	45
	46	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	47	if (work) {
	48	INIT_WORK(work, do_sync_work);
	49	schedule_work(work);
	50	}
cf9a2ae8 DH	51	}
	52
	53	/*
	54	* Generic function to fsync a file.
	55	*
	56	* filp may be NULL if called via the msync of a vma.
	57	*/
	58	int file_fsync(struct file filp, struct dentry dentry, int datasync)
	59	{
	60	struct inode * inode = dentry->d_inode;
	61	struct super_block * sb;
	62	int ret, err;
	63
	64	/* sync the inode to buffers */
	65	ret = write_inode_now(inode, 0);
	66
	67	/* sync the superblock to buffers */
	68	sb = inode->i_sb;
	69	lock_super(sb);
762873c2	70	if (sb->s_dirt && sb->s_op->write_super)
cf9a2ae8 DH	71	sb->s_op->write_super(sb);
	72	unlock_super(sb);
	73
	74	/* .. finally sync the buffers to disk */
	75	err = sync_blockdev(sb->s_bdev);
	76	if (!ret)
	77	ret = err;
	78	return ret;
	79	}
	80
4c728ef5 CH	81	/**
	82	* vfs_fsync - perform a fsync or fdatasync on a file
	83	* @file: file to sync
	84	* @dentry: dentry of @file
	85	* @data: only perform a fdatasync operation
	86	*
	87	* Write back data and metadata for @file to disk. If @datasync is
	88	* set only metadata needed to access modified file data is written.
	89	*
	90	* In case this function is called from nfsd @file may be %NULL and
	91	* only @dentry is set. This can only happen when the filesystem
	92	* implements the export_operations API.
	93	*/
	94	int vfs_fsync(struct file file, struct dentry dentry, int datasync)
cf9a2ae8	95	{
4c728ef5 CH	96	const struct file_operations *fop;
	97	struct address_space *mapping;
	98	int err, ret;
	99
	100	/*
	101	* Get mapping and operations from the file in case we have
	102	* as file, or get the default values for them in case we
	103	* don't have a struct file available. Damn nfsd..
	104	*/
	105	if (file) {
	106	mapping = file->f_mapping;
	107	fop = file->f_op;
	108	} else {
	109	mapping = dentry->d_inode->i_mapping;
	110	fop = dentry->d_inode->i_fop;
	111	}
cf9a2ae8	112
4c728ef5	113	if (!fop \|\| !fop->fsync) {
cf9a2ae8 DH	114	ret = -EINVAL;
	115	goto out;
	116	}
	117
	118	ret = filemap_fdatawrite(mapping);
	119
	120	/*
	121	* We need to protect against concurrent writers, which could cause
	122	* livelocks in fsync_buffers_list().
	123	*/
	124	mutex_lock(&mapping->host->i_mutex);
4c728ef5	125	err = fop->fsync(file, dentry, datasync);
cf9a2ae8 DH	126	if (!ret)
	127	ret = err;
	128	mutex_unlock(&mapping->host->i_mutex);
	129	err = filemap_fdatawait(mapping);
	130	if (!ret)
	131	ret = err;
	132	out:
	133	return ret;
	134	}
4c728ef5	135	EXPORT_SYMBOL(vfs_fsync);
cf9a2ae8	136
4c728ef5	137	static int do_fsync(unsigned int fd, int datasync)
cf9a2ae8 DH	138	{
	139	struct file *file;
	140	int ret = -EBADF;
	141
	142	file = fget(fd);
	143	if (file) {
4c728ef5	144	ret = vfs_fsync(file, file->f_path.dentry, datasync);
cf9a2ae8 DH	145	fput(file);
	146	}
	147	return ret;
	148	}
	149
a5f8fa9e	150	SYSCALL_DEFINE1(fsync, unsigned int, fd)
cf9a2ae8	151	{
4c728ef5	152	return do_fsync(fd, 0);
cf9a2ae8 DH	153	}
cf9a2ae8 DH	154
a5f8fa9e	155	SYSCALL_DEFINE1(fdatasync, unsigned int, fd)
cf9a2ae8	156	{
4c728ef5	157	return do_fsync(fd, 1);
cf9a2ae8 DH	158	}
cf9a2ae8 DH	159
f79e2abb AM	160	/*
	161	* sys_sync_file_range() permits finely controlled syncing over a segment of
	162	* a file in the range offset .. (offset+nbytes-1) inclusive. If nbytes is
	163	* zero then sys_sync_file_range() will operate from offset out to EOF.
	164	*
	165	* The flag bits are:
	166	*
	167	* SYNC_FILE_RANGE_WAIT_BEFORE: wait upon writeout of all pages in the range
	168	* before performing the write.
	169	*
	170	* SYNC_FILE_RANGE_WRITE: initiate writeout of all those dirty pages in the
cce77081 PM	171	* range which are not presently under writeback. Note that this may block for
cce77081 PM	172	* significant periods due to exhaustion of disk request structures.
f79e2abb AM	173	*
	174	* SYNC_FILE_RANGE_WAIT_AFTER: wait upon writeout of all pages in the range
	175	* after performing the write.
	176	*
	177	* Useful combinations of the flag bits are:
	178	*
	179	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE: ensures that all pages
	180	* in the range which were dirty on entry to sys_sync_file_range() are placed
	181	* under writeout. This is a start-write-for-data-integrity operation.
	182	*
	183	* SYNC_FILE_RANGE_WRITE: start writeout of all dirty pages in the range which
	184	* are not presently under writeout. This is an asynchronous flush-to-disk
	185	* operation. Not suitable for data integrity operations.
	186	*
	187	* SYNC_FILE_RANGE_WAIT_BEFORE (or SYNC_FILE_RANGE_WAIT_AFTER): wait for
	188	* completion of writeout of all pages in the range. This will be used after an
	189	* earlier SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE operation to wait
	190	* for that operation to complete and to return the result.
	191	*
	192	* SYNC_FILE_RANGE_WAIT_BEFORE\|SYNC_FILE_RANGE_WRITE\|SYNC_FILE_RANGE_WAIT_AFTER:
	193	* a traditional sync() operation. This is a write-for-data-integrity operation
	194	* which will ensure that all pages in the range which were dirty on entry to
	195	* sys_sync_file_range() are committed to disk.
	196	*
	197	*
	198	* SYNC_FILE_RANGE_WAIT_BEFORE and SYNC_FILE_RANGE_WAIT_AFTER will detect any
	199	* I/O errors or ENOSPC conditions and will return those to the caller, after
	200	* clearing the EIO and ENOSPC flags in the address_space.
	201	*
	202	* It should be noted that none of these operations write out the file's
	203	* metadata. So unless the application is strictly performing overwrites of
	204	* already-instantiated disk blocks, there are no guarantees here that the data
	205	* will be available after a crash.
	206	*/
6673e0c3 HC	207	SYSCALL_DEFINE(sync_file_range)(int fd, loff_t offset, loff_t nbytes,
6673e0c3 HC	208	unsigned int flags)
f79e2abb AM	209	{
	210	int ret;
	211	struct file *file;
	212	loff_t endbyte; /* inclusive */
	213	int fput_needed;
	214	umode_t i_mode;
	215
	216	ret = -EINVAL;
	217	if (flags & ~VALID_FLAGS)
	218	goto out;
	219
	220	endbyte = offset + nbytes;
	221
	222	if ((s64)offset < 0)
	223	goto out;
	224	if ((s64)endbyte < 0)
	225	goto out;
	226	if (endbyte < offset)
	227	goto out;
	228
	229	if (sizeof(pgoff_t) == 4) {
	230	if (offset >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	231	/*
	232	* The range starts outside a 32 bit machine's
	233	* pagecache addressing capabilities. Let it "succeed"
	234	*/
	235	ret = 0;
	236	goto out;
	237	}
	238	if (endbyte >= (0x100000000ULL << PAGE_CACHE_SHIFT)) {
	239	/*
	240	* Out to EOF
	241	*/
	242	nbytes = 0;
	243	}
	244	}
	245
	246	if (nbytes == 0)
111ebb6e	247	endbyte = LLONG_MAX;
f79e2abb AM	248	else
	249	endbyte--; /* inclusive */
	250
	251	ret = -EBADF;
	252	file = fget_light(fd, &fput_needed);
	253	if (!file)
	254	goto out;
	255
0f7fc9e4	256	i_mode = file->f_path.dentry->d_inode->i_mode;
f79e2abb AM	257	ret = -ESPIPE;
	258	if (!S_ISREG(i_mode) && !S_ISBLK(i_mode) && !S_ISDIR(i_mode) &&
	259	!S_ISLNK(i_mode))
	260	goto out_put;
	261
ef51c976	262	ret = do_sync_mapping_range(file->f_mapping, offset, endbyte, flags);
f79e2abb AM	263	out_put:
	264	fput_light(file, fput_needed);
	265	out:
	266	return ret;
	267	}
6673e0c3 HC	268	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	269	asmlinkage long SyS_sync_file_range(long fd, loff_t offset, loff_t nbytes,
	270	long flags)
	271	{
	272	return SYSC_sync_file_range((int) fd, offset, nbytes,
	273	(unsigned int) flags);
	274	}
	275	SYSCALL_ALIAS(sys_sync_file_range, SyS_sync_file_range);
	276	#endif
f79e2abb	277
edd5cd4a DW	278	/* It would be nice if people remember that not all the world's an i386
edd5cd4a DW	279	when they introduce new system calls */
6673e0c3 HC	280	SYSCALL_DEFINE(sync_file_range2)(int fd, unsigned int flags,
6673e0c3 HC	281	loff_t offset, loff_t nbytes)
edd5cd4a DW	282	{
	283	return sys_sync_file_range(fd, offset, nbytes, flags);
	284	}
6673e0c3 HC	285	#ifdef CONFIG_HAVE_SYSCALL_WRAPPERS
	286	asmlinkage long SyS_sync_file_range2(long fd, long flags,
	287	loff_t offset, loff_t nbytes)
	288	{
	289	return SYSC_sync_file_range2((int) fd, (unsigned int) flags,
	290	offset, nbytes);
	291	}
	292	SYSCALL_ALIAS(sys_sync_file_range2, SyS_sync_file_range2);
	293	#endif
edd5cd4a	294
f79e2abb AM	295	/*
	296	* `endbyte' is inclusive
	297	*/
5b04aa3a MF	298	int do_sync_mapping_range(struct address_space *mapping, loff_t offset,
5b04aa3a MF	299	loff_t endbyte, unsigned int flags)
f79e2abb AM	300	{
f79e2abb AM	301	int ret;
f79e2abb	302
f79e2abb AM	303	if (!mapping) {
	304	ret = -EINVAL;
	305	goto out;
	306	}
	307
	308	ret = 0;
	309	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
	310	ret = wait_on_page_writeback_range(mapping,
	311	offset >> PAGE_CACHE_SHIFT,
	312	endbyte >> PAGE_CACHE_SHIFT);
	313	if (ret < 0)
	314	goto out;
	315	}
	316
	317	if (flags & SYNC_FILE_RANGE_WRITE) {
	318	ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
ee53a891	319	WB_SYNC_ALL);
f79e2abb AM	320	if (ret < 0)
	321	goto out;
	322	}
	323
	324	if (flags & SYNC_FILE_RANGE_WAIT_AFTER) {
	325	ret = wait_on_page_writeback_range(mapping,
	326	offset >> PAGE_CACHE_SHIFT,
	327	endbyte >> PAGE_CACHE_SHIFT);
	328	}
	329	out:
	330	return ret;
	331	}
5b04aa3a	332	EXPORT_SYMBOL_GPL(do_sync_mapping_range);