summaryrefslogtreecommitdiffstats
path: root/fs/btrfs/locking.c
blob: 93146b4952769b9d994e7ec9d8a46d65198c52d9 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
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
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
// SPDX-License-Identifier: GPL-2.0
/*
 * Copyright (C) 2008 Oracle.  All rights reserved.
 */

#include <linux/sched.h>
#include <linux/pagemap.h>
#include <linux/spinlock.h>
#include <linux/page-flags.h>
#include <asm/bug.h>
#include "misc.h"
#include "ctree.h"
#include "extent_io.h"
#include "locking.h"

#ifdef CONFIG_BTRFS_DEBUG
static inline void btrfs_assert_spinning_writers_get(struct extent_buffer *eb)
{
	WARN_ON(eb->spinning_writers);
	eb->spinning_writers++;
}

static inline void btrfs_assert_spinning_writers_put(struct extent_buffer *eb)
{
	WARN_ON(eb->spinning_writers != 1);
	eb->spinning_writers--;
}

static inline void btrfs_assert_no_spinning_writers(struct extent_buffer *eb)
{
	WARN_ON(eb->spinning_writers);
}

static inline void btrfs_assert_spinning_readers_get(struct extent_buffer *eb)
{
	atomic_inc(&eb->spinning_readers);
}

static inline void btrfs_assert_spinning_readers_put(struct extent_buffer *eb)
{
	WARN_ON(atomic_read(&eb->spinning_readers) == 0);
	atomic_dec(&eb->spinning_readers);
}

static inline void btrfs_assert_tree_read_locks_get(struct extent_buffer *eb)
{
	atomic_inc(&eb->read_locks);
}

static inline void btrfs_assert_tree_read_locks_put(struct extent_buffer *eb)
{
	atomic_dec(&eb->read_locks);
}

static inline void btrfs_assert_tree_read_locked(struct extent_buffer *eb)
{
	BUG_ON(!atomic_read(&eb->read_locks));
}

static inline void btrfs_assert_tree_write_locks_get(struct extent_buffer *eb)
{
	eb->write_locks++;
}

static inline void btrfs_assert_tree_write_locks_put(struct extent_buffer *eb)
{
	eb->write_locks--;
}

#else
static void btrfs_assert_spinning_writers_get(struct extent_buffer *eb) { }
static void btrfs_assert_spinning_writers_put(struct extent_buffer *eb) { }
static void btrfs_assert_no_spinning_writers(struct extent_buffer *eb) { }
static void btrfs_assert_spinning_readers_put(struct extent_buffer *eb) { }
static void btrfs_assert_spinning_readers_get(struct extent_buffer *eb) { }
static void btrfs_assert_tree_read_locked(struct extent_buffer *eb) { }
static void btrfs_assert_tree_read_locks_get(struct extent_buffer *eb) { }
static void btrfs_assert_tree_read_locks_put(struct extent_buffer *eb) { }
static void btrfs_assert_tree_write_locks_get(struct extent_buffer *eb) { }
static void btrfs_assert_tree_write_locks_put(struct extent_buffer *eb) { }
#endif

void btrfs_set_lock_blocking_read(struct extent_buffer *eb)
{
	trace_btrfs_set_lock_blocking_read(eb);
	/*
	 * No lock is required.  The lock owner may change if we have a read
	 * lock, but it won't change to or away from us.  If we have the write
	 * lock, we are the owner and it'll never change.
	 */
	if (eb->lock_nested && current->pid == eb->lock_owner)
		return;
	btrfs_assert_tree_read_locked(eb);
	atomic_inc(&eb->blocking_readers);
	btrfs_assert_spinning_readers_put(eb);
	read_unlock(&eb->lock);
}

void btrfs_set_lock_blocking_write(struct extent_buffer *eb)
{
	trace_btrfs_set_lock_blocking_write(eb);
	/*
	 * No lock is required.  The lock owner may change if we have a read
	 * lock, but it won't change to or away from us.  If we have the write
	 * lock, we are the owner and it'll never change.
	 */
	if (eb->lock_nested && current->pid == eb->lock_owner)
		return;
	if (eb->blocking_writers == 0) {
		btrfs_assert_spinning_writers_put(eb);
		btrfs_assert_tree_locked(eb);
		eb->blocking_writers++;
		write_unlock(&eb->lock);
	}
}

/*
 * take a spinning read lock.  This will wait for any blocking
 * writers
 */
void btrfs_tree_read_lock(struct extent_buffer *eb)
{
	u64 start_ns = 0;

	if (trace_btrfs_tree_read_lock_enabled())
		start_ns = ktime_get_ns();
again:
	read_lock(&eb->lock);
	BUG_ON(eb->blocking_writers == 0 &&
	       current->pid == eb->lock_owner);
	if (eb->blocking_writers && current->pid == eb->lock_owner) {
		/*
		 * This extent is already write-locked by our thread. We allow
		 * an additional read lock to be added because it's for the same
		 * thread. btrfs_find_all_roots() depends on this as it may be
		 * called on a partly (write-)locked tree.
		 */
		BUG_ON(eb->lock_nested);
		eb->lock_nested = true;
		read_unlock(&eb->lock);
		trace_btrfs_tree_read_lock(eb, start_ns);
		return;
	}
	if (eb->blocking_writers) {
		read_unlock(&eb->lock);
		wait_event(eb->write_lock_wq,
			   eb->blocking_writers == 0);
		goto again;
	}
	btrfs_assert_tree_read_locks_get(eb);
	btrfs_assert_spinning_readers_get(eb);
	trace_btrfs_tree_read_lock(eb, start_ns);
}

/*
 * take a spinning read lock.
 * returns 1 if we get the read lock and 0 if we don't
 * this won't wait for blocking writers
 */
int btrfs_tree_read_lock_atomic(struct extent_buffer *eb)
{
	if (eb->blocking_writers)
		return 0;

	read_lock(&eb->lock);
	if (eb->blocking_writers) {
		read_unlock(&eb->lock);
		return 0;
	}
	btrfs_assert_tree_read_locks_get(eb);
	btrfs_assert_spinning_readers_get(eb);
	trace_btrfs_tree_read_lock_atomic(eb);
	return 1;
}

/*
 * returns 1 if we get the read lock and 0 if we don't
 * this won't wait for blocking writers
 */
int btrfs_try_tree_read_lock(struct extent_buffer *eb)
{
	if (eb->blocking_writers)
		return 0;

	if (!read_trylock(&eb->lock))
		return 0;

	if (eb->blocking_writers) {
		read_unlock(&eb->lock);
		return 0;
	}
	btrfs_assert_tree_read_locks_get(eb);
	btrfs_assert_spinning_readers_get(eb);
	trace_btrfs_try_tree_read_lock(eb);
	return 1;
}

/*
 * returns 1 if we get the read lock and 0 if we don't
 * this won't wait for blocking writers or readers
 */
int btrfs_try_tree_write_lock(struct extent_buffer *eb)
{
	if (eb->blocking_writers || atomic_read(&eb->blocking_readers))
		return 0;

	write_lock(&eb->lock);
	if (eb->blocking_writers || atomic_read(&eb->blocking_readers)) {
		write_unlock(&eb->lock);
		return 0;
	}
	btrfs_assert_tree_write_locks_get(eb);
	btrfs_assert_spinning_writers_get(eb);
	eb->lock_owner = current->pid;
	trace_btrfs_try_tree_write_lock(eb);
	return 1;
}

/*
 * drop a spinning read lock
 */
void btrfs_tree_read_unlock(struct extent_buffer *eb)
{
	trace_btrfs_tree_read_unlock(eb);
	/*
	 * if we're nested, we have the write lock.  No new locking
	 * is needed as long as we are the lock owner.
	 * The write unlock will do a barrier for us, and the lock_nested
	 * field only matters to the lock owner.
	 */
	if (eb->lock_nested && current->pid == eb->lock_owner) {
		eb->lock_nested = false;
		return;
	}
	btrfs_assert_tree_read_locked(eb);
	btrfs_assert_spinning_readers_put(eb);
	btrfs_assert_tree_read_locks_put(eb);
	read_unlock(&eb->lock);
}

/*
 * drop a blocking read lock
 */
void btrfs_tree_read_unlock_blocking(struct extent_buffer *eb)
{
	trace_btrfs_tree_read_unlock_blocking(eb);
	/*
	 * if we're nested, we have the write lock.  No new locking
	 * is needed as long as we are the lock owner.
	 * The write unlock will do a barrier for us, and the lock_nested
	 * field only matters to the lock owner.
	 */
	if (eb->lock_nested && current->pid == eb->lock_owner) {
		eb->lock_nested = false;
		return;
	}
	btrfs_assert_tree_read_locked(eb);
	WARN_ON(atomic_read(&eb->blocking_readers) == 0);
	/* atomic_dec_and_test implies a barrier */
	if (atomic_dec_and_test(&eb->blocking_readers))
		cond_wake_up_nomb(&eb->read_lock_wq);
	btrfs_assert_tree_read_locks_put(eb);
}

/*
 * take a spinning write lock.  This will wait for both
 * blocking readers or writers
 */
void btrfs_tree_lock(struct extent_buffer *eb)
{
	u64 start_ns = 0;

	if (trace_btrfs_tree_lock_enabled())
		start_ns = ktime_get_ns();

	WARN_ON(eb->lock_owner == current->pid);
again:
	wait_event(eb->read_lock_wq, atomic_read(&eb->blocking_readers) == 0);
	wait_event(eb->write_lock_wq, eb->blocking_writers == 0);
	write_lock(&eb->lock);
	if (atomic_read(&eb->blocking_readers) || eb->blocking_writers) {
		write_unlock(&eb->lock);
		goto again;
	}
	btrfs_assert_spinning_writers_get(eb);
	btrfs_assert_tree_write_locks_get(eb);
	eb->lock_owner = current->pid;
	trace_btrfs_tree_lock(eb, start_ns);
}

/*
 * drop a spinning or a blocking write lock.
 */
void btrfs_tree_unlock(struct extent_buffer *eb)
{
	int blockers = eb->blocking_writers;

	BUG_ON(blockers > 1);

	btrfs_assert_tree_locked(eb);
	trace_btrfs_tree_unlock(eb);
	eb->lock_owner = 0;
	btrfs_assert_tree_write_locks_put(eb);

	if (blockers) {
		btrfs_assert_no_spinning_writers(eb);
		eb->blocking_writers--;
		/*
		 * We need to order modifying blocking_writers above with
		 * actually waking up the sleepers to ensure they see the
		 * updated value of blocking_writers
		 */
		cond_wake_up(&eb->write_lock_wq);
	} else {
		btrfs_assert_spinning_writers_put(eb);
		write_unlock(&eb->lock);
	}
}

/*
 * Set all locked nodes in the path to blocking locks.  This should be done
 * before scheduling
 */
void btrfs_set_path_blocking(struct btrfs_path *p)
{
	int i;

	for (i = 0; i < BTRFS_MAX_LEVEL; i++) {
		if (!p->nodes[i] || !p->locks[i])
			continue;
		/*
		 * If we currently have a spinning reader or writer lock this
		 * will bump the count of blocking holders and drop the
		 * spinlock.
		 */
		if (p->locks[i] == BTRFS_READ_LOCK) {
			btrfs_set_lock_blocking_read(p->nodes[i]);
			p->locks[i] = BTRFS_READ_LOCK_BLOCKING;
		} else if (p->locks[i] == BTRFS_WRITE_LOCK) {
			btrfs_set_lock_blocking_write(p->nodes[i]);
			p->locks[i] = BTRFS_WRITE_LOCK_BLOCKING;
		}
	}
}

/*
 * This releases any locks held in the path starting at level and going all the
 * way up to the root.
 *
 * btrfs_search_slot will keep the lock held on higher nodes in a few corner
 * cases, such as COW of the block at slot zero in the node.  This ignores
 * those rules, and it should only be called when there are no more updates to
 * be done higher up in the tree.
 */
void btrfs_unlock_up_safe(struct btrfs_path *path, int level)
{
	int i;

	if (path->keep_locks)
		return;

	for (i = level; i < BTRFS_MAX_LEVEL; i++) {
		if (!path->nodes[i])
			continue;
		if (!path->locks[i])
			continue;
		btrfs_tree_unlock_rw(path->nodes[i], path->locks[i]);
		path->locks[i] = 0;
	}
}