sync.c 9.94 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
2 3 4 5 6 7 8
/*
 * High-level sync()-related operations
 */

#include <linux/kernel.h>
#include <linux/file.h>
#include <linux/fs.h>
9
#include <linux/slab.h>
10
#include <linux/export.h>
11
#include <linux/namei.h>
12
#include <linux/sched.h>
13 14 15 16
#include <linux/writeback.h>
#include <linux/syscalls.h>
#include <linux/linkage.h>
#include <linux/pagemap.h>
17
#include <linux/quotaops.h>
18
#include <linux/backing-dev.h>
19
#include "internal.h"
20 21 22 23

#define VALID_FLAGS (SYNC_FILE_RANGE_WAIT_BEFORE|SYNC_FILE_RANGE_WRITE| \
			SYNC_FILE_RANGE_WAIT_AFTER)

24
/*
25 26 27 28 29
 * 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.
30
 */
31
static int __sync_filesystem(struct super_block *sb, int wait)
32
{
33
	if (wait)
34
		sync_inodes_sb(sb);
35
	else
36
		writeback_inodes_sb(sb, WB_REASON_SYNC);
37

38 39 40 41 42 43 44 45 46 47
	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.
 */
48
int sync_filesystem(struct super_block *sb)
49 50 51
{
	int ret;

52 53 54 55 56 57 58 59 60
	/*
	 * 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.
	 */
61
	if (sb_rdonly(sb))
62 63
		return 0;

64
	ret = __sync_filesystem(sb, 0);
65 66
	if (ret < 0)
		return ret;
67
	return __sync_filesystem(sb, 1);
68
}
69
EXPORT_SYMBOL(sync_filesystem);
70

71
static void sync_inodes_one_sb(struct super_block *sb, void *arg)
Al Viro's avatar
Al Viro committed
72
{
73
	if (!sb_rdonly(sb))
74
		sync_inodes_sb(sb);
Al Viro's avatar
Al Viro committed
75
}
76 77 78

static void sync_fs_one_sb(struct super_block *sb, void *arg)
{
79
	if (!sb_rdonly(sb) && sb->s_op->sync_fs)
80 81 82
		sb->s_op->sync_fs(sb, *(int *)arg);
}

83
static void fdatawrite_one_bdev(struct block_device *bdev, void *arg)
84
{
85
	filemap_fdatawrite(bdev->bd_inode->i_mapping);
86 87
}

88
static void fdatawait_one_bdev(struct block_device *bdev, void *arg)
89
{
90 91 92 93 94 95
	/*
	 * We keep the error status of individual mapping so that
	 * applications can catch the writeback error using fsync(2).
	 * See filemap_fdatawait_keep_errors() for details.
	 */
	filemap_fdatawait_keep_errors(bdev->bd_inode->i_mapping);
96 97
}

98
/*
99 100 101 102 103 104 105 106
 * Sync everything. We start by waking flusher threads so that most of
 * writeback runs on all devices in parallel. Then we sync all inodes reliably
 * which effectively also waits for all flusher threads to finish doing
 * writeback. At this point all data is on disk so metadata should be stable
 * and we tell filesystems to sync their metadata via ->sync_fs() calls.
 * Finally, we writeout all block devices because some filesystems (e.g. ext2)
 * just write metadata (such as inodes or bitmaps) to block device page cache
 * and do not sync it on their own in ->sync_fs().
107
 */
108
void ksys_sync(void)
109
{
110 111
	int nowait = 0, wait = 1;

112
	wakeup_flusher_threads(WB_REASON_SYNC);
113
	iterate_supers(sync_inodes_one_sb, NULL);
114
	iterate_supers(sync_fs_one_sb, &nowait);
115
	iterate_supers(sync_fs_one_sb, &wait);
116 117
	iterate_bdevs(fdatawrite_one_bdev, NULL);
	iterate_bdevs(fdatawait_one_bdev, NULL);
118 119
	if (unlikely(laptop_mode))
		laptop_sync_completion();
120 121 122 123 124
}

SYSCALL_DEFINE0(sync)
{
	ksys_sync();
125 126 127
	return 0;
}

128 129
static void do_sync_work(struct work_struct *work)
{
130 131
	int nowait = 0;

132 133 134 135
	/*
	 * Sync twice to reduce the possibility we skipped some inodes / pages
	 * because they were temporarily locked
	 */
136 137
	iterate_supers(sync_inodes_one_sb, &nowait);
	iterate_supers(sync_fs_one_sb, &nowait);
138
	iterate_bdevs(fdatawrite_one_bdev, NULL);
139 140
	iterate_supers(sync_inodes_one_sb, &nowait);
	iterate_supers(sync_fs_one_sb, &nowait);
141
	iterate_bdevs(fdatawrite_one_bdev, NULL);
142
	printk("Emergency Sync complete\n");
143 144 145
	kfree(work);
}

146 147
void emergency_sync(void)
{
148 149 150 151 152 153 154
	struct work_struct *work;

	work = kmalloc(sizeof(*work), GFP_ATOMIC);
	if (work) {
		INIT_WORK(work, do_sync_work);
		schedule_work(work);
	}
155 156
}

157 158 159 160 161
/*
 * sync a single super
 */
SYSCALL_DEFINE1(syncfs, int, fd)
{
162
	struct fd f = fdget(fd);
163 164 165
	struct super_block *sb;
	int ret;

166
	if (!f.file)
167
		return -EBADF;
Al Viro's avatar
Al Viro committed
168
	sb = f.file->f_path.dentry->d_sb;
169 170 171 172 173

	down_read(&sb->s_umount);
	ret = sync_filesystem(sb);
	up_read(&sb->s_umount);

174
	fdput(f);
175 176 177
	return ret;
}

Christoph Hellwig's avatar
Christoph Hellwig committed
178
/**
179
 * vfs_fsync_range - helper to sync a range of data & metadata to disk
Christoph Hellwig's avatar
Christoph Hellwig committed
180
 * @file:		file to sync
181 182 183
 * @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
Christoph Hellwig's avatar
Christoph Hellwig committed
184
 *
185 186 187
 * 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.
Christoph Hellwig's avatar
Christoph Hellwig committed
188
 */
189
int vfs_fsync_range(struct file *file, loff_t start, loff_t end, int datasync)
190
{
191 192
	struct inode *inode = file->f_mapping->host;

Al Viro's avatar
Al Viro committed
193
	if (!file->f_op->fsync)
194
		return -EINVAL;
195
	if (!datasync && (inode->i_state & I_DIRTY_TIME))
196
		mark_inode_dirty_sync(inode);
197
	return file->f_op->fsync(file, start, end, datasync);
198
}
199 200 201 202 203 204 205 206 207 208
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.
 */
209
int vfs_fsync(struct file *file, int datasync)
210
{
211
	return vfs_fsync_range(file, 0, LLONG_MAX, datasync);
212
}
Christoph Hellwig's avatar
Christoph Hellwig committed
213
EXPORT_SYMBOL(vfs_fsync);
214

Christoph Hellwig's avatar
Christoph Hellwig committed
215
static int do_fsync(unsigned int fd, int datasync)
216
{
217
	struct fd f = fdget(fd);
218 219
	int ret = -EBADF;

220 221 222
	if (f.file) {
		ret = vfs_fsync(f.file, datasync);
		fdput(f);
223 224 225 226
	}
	return ret;
}

227
SYSCALL_DEFINE1(fsync, unsigned int, fd)
228
{
Christoph Hellwig's avatar
Christoph Hellwig committed
229
	return do_fsync(fd, 0);
230 231
}

232
SYSCALL_DEFINE1(fdatasync, unsigned int, fd)
233
{
Christoph Hellwig's avatar
Christoph Hellwig committed
234
	return do_fsync(fd, 1);
235 236
}

237 238 239 240 241 242 243 244 245 246 247
/*
 * 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
248 249
 * range which are not presently under writeback. Note that this may block for
 * significant periods due to exhaustion of disk request structures.
250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283
 *
 * 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.
 */
284 285
int ksys_sync_file_range(int fd, loff_t offset, loff_t nbytes,
			 unsigned int flags)
286 287
{
	int ret;
288
	struct fd f;
289
	struct address_space *mapping;
290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
	loff_t endbyte;			/* inclusive */
	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) {
307
		if (offset >= (0x100000000ULL << PAGE_SHIFT)) {
308 309 310 311 312 313 314
			/*
			 * The range starts outside a 32 bit machine's
			 * pagecache addressing capabilities.  Let it "succeed"
			 */
			ret = 0;
			goto out;
		}
315
		if (endbyte >= (0x100000000ULL << PAGE_SHIFT)) {
316 317 318 319 320 321 322 323
			/*
			 * Out to EOF
			 */
			nbytes = 0;
		}
	}

	if (nbytes == 0)
324
		endbyte = LLONG_MAX;
325 326 327 328
	else
		endbyte--;		/* inclusive */

	ret = -EBADF;
329 330
	f = fdget(fd);
	if (!f.file)
331 332
		goto out;

Al Viro's avatar
Al Viro committed
333
	i_mode = file_inode(f.file)->i_mode;
334 335 336 337 338
	ret = -ESPIPE;
	if (!S_ISREG(i_mode) && !S_ISBLK(i_mode) && !S_ISDIR(i_mode) &&
			!S_ISLNK(i_mode))
		goto out_put;

339
	mapping = f.file->f_mapping;
340 341
	ret = 0;
	if (flags & SYNC_FILE_RANGE_WAIT_BEFORE) {
342
		ret = file_fdatawait_range(f.file, offset, endbyte);
343 344 345 346 347
		if (ret < 0)
			goto out_put;
	}

	if (flags & SYNC_FILE_RANGE_WRITE) {
348 349
		ret = __filemap_fdatawrite_range(mapping, offset, endbyte,
						 WB_SYNC_NONE);
350 351 352 353 354
		if (ret < 0)
			goto out_put;
	}

	if (flags & SYNC_FILE_RANGE_WAIT_AFTER)
355
		ret = file_fdatawait_range(f.file, offset, endbyte);
356

357
out_put:
358
	fdput(f);
359 360 361 362
out:
	return ret;
}

363 364 365 366 367 368
SYSCALL_DEFINE4(sync_file_range, int, fd, loff_t, offset, loff_t, nbytes,
				unsigned int, flags)
{
	return ksys_sync_file_range(fd, offset, nbytes, flags);
}

369 370
/* It would be nice if people remember that not all the world's an i386
   when they introduce new system calls */
371 372
SYSCALL_DEFINE4(sync_file_range2, int, fd, unsigned int, flags,
				 loff_t, offset, loff_t, nbytes)
373
{
374
	return ksys_sync_file_range(fd, offset, nbytes, flags);
375
}