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
|
/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _BCACHEFS_BTREE_IO_H
#define _BCACHEFS_BTREE_IO_H
#include "bkey_methods.h"
#include "bset.h"
#include "btree_locking.h"
#include "checksum.h"
#include "extents.h"
#include "io_types.h"
struct bch_fs;
struct btree_write;
struct btree;
struct btree_iter;
static inline bool btree_node_dirty(struct btree *b)
{
return test_bit(BTREE_NODE_dirty, &b->flags);
}
static inline void set_btree_node_dirty(struct bch_fs *c, struct btree *b)
{
if (!test_and_set_bit(BTREE_NODE_dirty, &b->flags))
atomic_inc(&c->btree_cache.dirty);
}
static inline void clear_btree_node_dirty(struct bch_fs *c, struct btree *b)
{
if (test_and_clear_bit(BTREE_NODE_dirty, &b->flags))
atomic_dec(&c->btree_cache.dirty);
}
struct btree_read_bio {
struct bch_fs *c;
u64 start_time;
unsigned have_ioref:1;
struct extent_ptr_decoded pick;
struct work_struct work;
struct bio bio;
};
struct btree_write_bio {
struct work_struct work;
void *data;
unsigned bytes;
struct bch_write_bio wbio;
};
static inline void btree_node_io_unlock(struct btree *b)
{
EBUG_ON(!btree_node_write_in_flight(b));
clear_btree_node_write_in_flight(b);
wake_up_bit(&b->flags, BTREE_NODE_write_in_flight);
}
static inline void btree_node_io_lock(struct btree *b)
{
wait_on_bit_lock_io(&b->flags, BTREE_NODE_write_in_flight,
TASK_UNINTERRUPTIBLE);
}
static inline void btree_node_wait_on_io(struct btree *b)
{
wait_on_bit_io(&b->flags, BTREE_NODE_write_in_flight,
TASK_UNINTERRUPTIBLE);
}
static inline bool btree_node_may_write(struct btree *b)
{
return list_empty_careful(&b->write_blocked) &&
(!b->written || !b->will_make_reachable);
}
enum compact_mode {
COMPACT_LAZY,
COMPACT_ALL,
};
bool bch2_compact_whiteouts(struct bch_fs *, struct btree *,
enum compact_mode);
static inline bool should_compact_bset_lazy(struct btree *b,
struct bset_tree *t)
{
unsigned total_u64s = bset_u64s(t);
unsigned dead_u64s = bset_dead_u64s(b, t);
return dead_u64s > 64 && dead_u64s * 3 > total_u64s;
}
static inline bool bch2_maybe_compact_whiteouts(struct bch_fs *c, struct btree *b)
{
struct bset_tree *t;
for_each_bset(b, t)
if (should_compact_bset_lazy(b, t))
return bch2_compact_whiteouts(c, b, COMPACT_LAZY);
return false;
}
static inline struct nonce btree_nonce(struct bset *i, unsigned offset)
{
return (struct nonce) {{
[0] = cpu_to_le32(offset),
[1] = ((__le32 *) &i->seq)[0],
[2] = ((__le32 *) &i->seq)[1],
[3] = ((__le32 *) &i->journal_seq)[0]^BCH_NONCE_BTREE,
}};
}
static inline void bset_encrypt(struct bch_fs *c, struct bset *i, unsigned offset)
{
struct nonce nonce = btree_nonce(i, offset);
if (!offset) {
struct btree_node *bn = container_of(i, struct btree_node, keys);
unsigned bytes = (void *) &bn->keys - (void *) &bn->flags;
bch2_encrypt(c, BSET_CSUM_TYPE(i), nonce, &bn->flags,
bytes);
nonce = nonce_add(nonce, round_up(bytes, CHACHA_BLOCK_SIZE));
}
bch2_encrypt(c, BSET_CSUM_TYPE(i), nonce, i->_data,
vstruct_end(i) - (void *) i->_data);
}
void bch2_btree_sort_into(struct bch_fs *, struct btree *, struct btree *);
void bch2_btree_build_aux_trees(struct btree *);
void bch2_btree_init_next(struct bch_fs *, struct btree *,
struct btree_iter *);
int bch2_btree_node_read_done(struct bch_fs *, struct bch_dev *,
struct btree *, bool);
void bch2_btree_node_read(struct bch_fs *, struct btree *, bool);
int bch2_btree_root_read(struct bch_fs *, enum btree_id,
const struct bkey_i *, unsigned);
void bch2_btree_complete_write(struct bch_fs *, struct btree *,
struct btree_write *);
void bch2_btree_write_error_work(struct work_struct *);
void __bch2_btree_node_write(struct bch_fs *, struct btree *);
bool bch2_btree_post_write_cleanup(struct bch_fs *, struct btree *);
void bch2_btree_node_write(struct bch_fs *, struct btree *,
enum six_lock_type);
static inline void btree_node_write_if_need(struct bch_fs *c, struct btree *b,
enum six_lock_type lock_held)
{
while (b->written &&
btree_node_need_write(b) &&
btree_node_may_write(b)) {
if (!btree_node_write_in_flight(b)) {
bch2_btree_node_write(c, b, lock_held);
break;
}
six_unlock_type(&b->c.lock, lock_held);
btree_node_wait_on_io(b);
btree_node_lock_type(c, b, lock_held);
}
}
#define bch2_btree_node_write_cond(_c, _b, cond) \
do { \
unsigned long old, new, v = READ_ONCE((_b)->flags); \
\
do { \
old = new = v; \
\
if (!(old & (1 << BTREE_NODE_dirty)) || !(cond)) \
break; \
\
new |= (1 << BTREE_NODE_need_write); \
} while ((v = cmpxchg(&(_b)->flags, old, new)) != old); \
\
btree_node_write_if_need(_c, _b, SIX_LOCK_read); \
} while (0)
void bch2_btree_flush_all_reads(struct bch_fs *);
void bch2_btree_flush_all_writes(struct bch_fs *);
void bch2_dirty_btree_nodes_to_text(struct printbuf *, struct bch_fs *);
static inline void compat_bformat(unsigned level, enum btree_id btree_id,
unsigned version, unsigned big_endian,
int write, struct bkey_format *f)
{
if (version < bcachefs_metadata_version_inode_btree_change &&
btree_id == BTREE_ID_inodes) {
swap(f->bits_per_field[BKEY_FIELD_INODE],
f->bits_per_field[BKEY_FIELD_OFFSET]);
swap(f->field_offset[BKEY_FIELD_INODE],
f->field_offset[BKEY_FIELD_OFFSET]);
}
if (version < bcachefs_metadata_version_snapshot &&
(level || btree_type_has_snapshots(btree_id))) {
u64 max_packed =
~(~0ULL << f->bits_per_field[BKEY_FIELD_SNAPSHOT]);
f->field_offset[BKEY_FIELD_SNAPSHOT] = write
? 0
: U32_MAX - max_packed;
}
}
static inline void compat_bpos(unsigned level, enum btree_id btree_id,
unsigned version, unsigned big_endian,
int write, struct bpos *p)
{
if (big_endian != CPU_BIG_ENDIAN)
bch2_bpos_swab(p);
if (version < bcachefs_metadata_version_inode_btree_change &&
btree_id == BTREE_ID_inodes)
swap(p->inode, p->offset);
}
static inline void compat_btree_node(unsigned level, enum btree_id btree_id,
unsigned version, unsigned big_endian,
int write,
struct btree_node *bn)
{
if (version < bcachefs_metadata_version_inode_btree_change &&
btree_node_type_is_extents(btree_id) &&
bpos_cmp(bn->min_key, POS_MIN) &&
write)
bn->min_key = bpos_nosnap_predecessor(bn->min_key);
if (version < bcachefs_metadata_version_snapshot &&
write)
bn->max_key.snapshot = 0;
compat_bpos(level, btree_id, version, big_endian, write, &bn->min_key);
compat_bpos(level, btree_id, version, big_endian, write, &bn->max_key);
if (version < bcachefs_metadata_version_snapshot &&
!write)
bn->max_key.snapshot = U32_MAX;
if (version < bcachefs_metadata_version_inode_btree_change &&
btree_node_type_is_extents(btree_id) &&
bpos_cmp(bn->min_key, POS_MIN) &&
!write)
bn->min_key = bpos_nosnap_successor(bn->min_key);
}
#endif /* _BCACHEFS_BTREE_IO_H */
|
