summaryrefslogtreecommitdiffstats
path: root/fs/bcachefs/journal_types.h
blob: 19183fcf7ad7fd031ba263da226ecb21e9d04685 (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
/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _BCACHEFS_JOURNAL_TYPES_H
#define _BCACHEFS_JOURNAL_TYPES_H

#include <linux/cache.h>
#include <linux/workqueue.h>

#include "alloc_types.h"
#include "super_types.h"
#include "fifo.h"

#define JOURNAL_BUF_BITS	2
#define JOURNAL_BUF_NR		(1U << JOURNAL_BUF_BITS)
#define JOURNAL_BUF_MASK	(JOURNAL_BUF_NR - 1)

/*
 * We put JOURNAL_BUF_NR of these in struct journal; we used them for writes to
 * the journal that are being staged or in flight.
 */
struct journal_buf {
	struct closure		io;
	struct jset		*data;

	__BKEY_PADDED(key, BCH_REPLICAS_MAX);
	struct bch_devs_list	devs_written;

	struct closure_waitlist	wait;
	u64			last_seq;	/* copy of data->last_seq */
	long			expires;
	u64			flush_time;

	unsigned		buf_size;	/* size in bytes of @data */
	unsigned		sectors;	/* maximum size for current entry */
	unsigned		disk_sectors;	/* maximum size entry could have been, if
						   buf_size was bigger */
	unsigned		u64s_reserved;
	bool			noflush:1;	/* write has already been kicked off, and was noflush */
	bool			must_flush:1;	/* something wants a flush */
	bool			separate_flush:1;
	bool			need_flush_to_write_buffer:1;
	bool			write_started:1;
	bool			write_allocated:1;
	bool			write_done:1;
	u8			idx;
};

/*
 * Something that makes a journal entry dirty - i.e. a btree node that has to be
 * flushed:
 */

enum journal_pin_type {
	JOURNAL_PIN_btree,
	JOURNAL_PIN_key_cache,
	JOURNAL_PIN_other,
	JOURNAL_PIN_NR,
};

struct journal_entry_pin_list {
	struct list_head		list[JOURNAL_PIN_NR];
	struct list_head		flushed;
	atomic_t			count;
	struct bch_devs_list		devs;
};

struct journal;
struct journal_entry_pin;
typedef int (*journal_pin_flush_fn)(struct journal *j,
				struct journal_entry_pin *, u64);

struct journal_entry_pin {
	struct list_head		list;
	journal_pin_flush_fn		flush;
	u64				seq;
};

struct journal_res {
	bool			ref;
	u8			idx;
	u16			u64s;
	u32			offset;
	u64			seq;
};

union journal_res_state {
	struct {
		atomic64_t	counter;
	};

	struct {
		u64		v;
	};

	struct {
		u64		cur_entry_offset:20,
				idx:2,
				unwritten_idx:2,
				buf0_count:10,
				buf1_count:10,
				buf2_count:10,
				buf3_count:10;
	};
};

/* bytes: */
#define JOURNAL_ENTRY_SIZE_MIN		(64U << 10) /* 64k */
#define JOURNAL_ENTRY_SIZE_MAX		(4U  << 20) /* 4M */

/*
 * We stash some journal state as sentinal values in cur_entry_offset:
 * note - cur_entry_offset is in units of u64s
 */
#define JOURNAL_ENTRY_OFFSET_MAX	((1U << 20) - 1)

#define JOURNAL_ENTRY_CLOSED_VAL	(JOURNAL_ENTRY_OFFSET_MAX - 1)
#define JOURNAL_ENTRY_ERROR_VAL		(JOURNAL_ENTRY_OFFSET_MAX)

struct journal_space {
	/* Units of 512 bytes sectors: */
	unsigned	next_entry; /* How big the next journal entry can be */
	unsigned	total;
};

enum journal_space_from {
	journal_space_discarded,
	journal_space_clean_ondisk,
	journal_space_clean,
	journal_space_total,
	journal_space_nr,
};

#define JOURNAL_FLAGS()			\
	x(replay_done)			\
	x(running)			\
	x(may_skip_flush)		\
	x(need_flush_write)		\
	x(space_low)

enum journal_flags {
#define x(n)	JOURNAL_##n,
	JOURNAL_FLAGS()
#undef x
};

/* Reasons we may fail to get a journal reservation: */
#define JOURNAL_ERRORS()		\
	x(ok)				\
	x(retry)			\
	x(blocked)			\
	x(max_in_flight)		\
	x(journal_full)			\
	x(journal_pin_full)		\
	x(journal_stuck)		\
	x(insufficient_devices)

enum journal_errors {
#define x(n)	JOURNAL_ERR_##n,
	JOURNAL_ERRORS()
#undef x
};

typedef DARRAY(u64)		darray_u64;

struct journal_bio {
	struct bch_dev		*ca;
	unsigned		buf_idx;

	struct bio		bio;
};

/* Embedded in struct bch_fs */
struct journal {
	/* Fastpath stuff up front: */
	struct {

	union journal_res_state reservations;
	enum bch_watermark	watermark;

	} __aligned(SMP_CACHE_BYTES);

	unsigned long		flags;

	/* Max size of current journal entry */
	unsigned		cur_entry_u64s;
	unsigned		cur_entry_sectors;

	/* Reserved space in journal entry to be used just prior to write */
	unsigned		entry_u64s_reserved;


	/*
	 * 0, or -ENOSPC if waiting on journal reclaim, or -EROFS if
	 * insufficient devices:
	 */
	enum journal_errors	cur_entry_error;

	unsigned		buf_size_want;
	/*
	 * We may queue up some things to be journalled (log messages) before
	 * the journal has actually started - stash them here:
	 */
	darray_u64		early_journal_entries;

	/*
	 * Protects journal_buf->data, when accessing without a jorunal
	 * reservation: for synchronization between the btree write buffer code
	 * and the journal write path:
	 */
	struct mutex		buf_lock;
	/*
	 * Two journal entries -- one is currently open for new entries, the
	 * other is possibly being written out.
	 */
	struct journal_buf	buf[JOURNAL_BUF_NR];

	spinlock_t		lock;

	/* if nonzero, we may not open a new journal entry: */
	unsigned		blocked;

	/* Used when waiting because the journal was full */
	wait_queue_head_t	wait;
	struct closure_waitlist	async_wait;

	struct delayed_work	write_work;
	struct workqueue_struct *wq;

	/* Sequence number of most recent journal entry (last entry in @pin) */
	atomic64_t		seq;

	/* seq, last_seq from the most recent journal entry successfully written */
	u64			seq_ondisk;
	u64			flushed_seq_ondisk;
	u64			last_seq_ondisk;
	u64			err_seq;
	u64			last_empty_seq;
	u64			oldest_seq_found_ondisk;

	/*
	 * FIFO of journal entries whose btree updates have not yet been
	 * written out.
	 *
	 * Each entry is a reference count. The position in the FIFO is the
	 * entry's sequence number relative to @seq.
	 *
	 * The journal entry itself holds a reference count, put when the
	 * journal entry is written out. Each btree node modified by the journal
	 * entry also holds a reference count, put when the btree node is
	 * written.
	 *
	 * When a reference count reaches zero, the journal entry is no longer
	 * needed. When all journal entries in the oldest journal bucket are no
	 * longer needed, the bucket can be discarded and reused.
	 */
	struct {
		u64 front, back, size, mask;
		struct journal_entry_pin_list *data;
	}			pin;

	struct journal_space	space[journal_space_nr];

	u64			replay_journal_seq;
	u64			replay_journal_seq_end;

	struct write_point	wp;
	spinlock_t		err_lock;

	struct mutex		reclaim_lock;
	/*
	 * Used for waiting until journal reclaim has freed up space in the
	 * journal:
	 */
	wait_queue_head_t	reclaim_wait;
	struct task_struct	*reclaim_thread;
	bool			reclaim_kicked;
	unsigned long		next_reclaim;
	u64			nr_direct_reclaim;
	u64			nr_background_reclaim;

	unsigned long		last_flushed;
	struct journal_entry_pin *flush_in_progress;
	bool			flush_in_progress_dropped;
	wait_queue_head_t	pin_flush_wait;

	/* protects advancing ja->discard_idx: */
	struct mutex		discard_lock;
	bool			can_discard;

	unsigned long		last_flush_write;

	u64			write_start_time;

	u64			nr_flush_writes;
	u64			nr_noflush_writes;
	u64			entry_bytes_written;

	struct bch2_time_stats	*flush_write_time;
	struct bch2_time_stats	*noflush_write_time;
	struct bch2_time_stats	*flush_seq_time;

#ifdef CONFIG_DEBUG_LOCK_ALLOC
	struct lockdep_map	res_map;
#endif
} __aligned(SMP_CACHE_BYTES);

/*
 * Embedded in struct bch_dev. First three fields refer to the array of journal
 * buckets, in bch_sb.
 */
struct journal_device {
	/*
	 * For each journal bucket, contains the max sequence number of the
	 * journal writes it contains - so we know when a bucket can be reused.
	 */
	u64			*bucket_seq;

	unsigned		sectors_free;

	/*
	 * discard_idx <= dirty_idx_ondisk <= dirty_idx <= cur_idx:
	 */
	unsigned		discard_idx;		/* Next bucket to discard */
	unsigned		dirty_idx_ondisk;
	unsigned		dirty_idx;
	unsigned		cur_idx;		/* Journal bucket we're currently writing to */
	unsigned		nr;

	u64			*buckets;

	/* Bio for journal reads/writes to this device */
	struct journal_bio	*bio[JOURNAL_BUF_NR];

	/* for bch_journal_read_device */
	struct closure		read;
	u64			highest_seq_found;
};

/*
 * journal_entry_res - reserve space in every journal entry:
 */
struct journal_entry_res {
	unsigned		u64s;
};

#endif /* _BCACHEFS_JOURNAL_TYPES_H */