1 // SPDX-License-Identifier: GPL-2.0 2 /* 3 * Copyright (C) 2008 Oracle. All rights reserved. 4 */ 5 6 #include <linux/sched.h> 7 #include <linux/pagemap.h> 8 #include <linux/spinlock.h> 9 #include <linux/page-flags.h> 10 #include <asm/bug.h> 11 #include "misc.h" 12 #include "ctree.h" 13 #include "extent_io.h" 14 #include "locking.h" 15 #include "accessors.h" 16 17 /* 18 * Lockdep class keys for extent_buffer->lock's in this root. For a given 19 * eb, the lockdep key is determined by the btrfs_root it belongs to and 20 * the level the eb occupies in the tree. 21 * 22 * Different roots are used for different purposes and may nest inside each 23 * other and they require separate keysets. As lockdep keys should be 24 * static, assign keysets according to the purpose of the root as indicated 25 * by btrfs_root->root_key.objectid. This ensures that all special purpose 26 * roots have separate keysets. 27 * 28 * Lock-nesting across peer nodes is always done with the immediate parent 29 * node locked thus preventing deadlock. As lockdep doesn't know this, use 30 * subclass to avoid triggering lockdep warning in such cases. 31 * 32 * The key is set by the readpage_end_io_hook after the buffer has passed 33 * csum validation but before the pages are unlocked. It is also set by 34 * btrfs_init_new_buffer on freshly allocated blocks. 35 * 36 * We also add a check to make sure the highest level of the tree is the 37 * same as our lockdep setup here. If BTRFS_MAX_LEVEL changes, this code 38 * needs update as well. 39 */ 40 #ifdef CONFIG_DEBUG_LOCK_ALLOC 41 #if BTRFS_MAX_LEVEL != 8 42 #error 43 #endif 44 45 #define DEFINE_LEVEL(stem, level) \ 46 .names[level] = "btrfs-" stem "-0" #level, 47 48 #define DEFINE_NAME(stem) \ 49 DEFINE_LEVEL(stem, 0) \ 50 DEFINE_LEVEL(stem, 1) \ 51 DEFINE_LEVEL(stem, 2) \ 52 DEFINE_LEVEL(stem, 3) \ 53 DEFINE_LEVEL(stem, 4) \ 54 DEFINE_LEVEL(stem, 5) \ 55 DEFINE_LEVEL(stem, 6) \ 56 DEFINE_LEVEL(stem, 7) 57 58 static struct btrfs_lockdep_keyset { 59 u64 id; /* root objectid */ 60 /* Longest entry: btrfs-block-group-00 */ 61 char names[BTRFS_MAX_LEVEL][24]; 62 struct lock_class_key keys[BTRFS_MAX_LEVEL]; 63 } btrfs_lockdep_keysets[] = { 64 { .id = BTRFS_ROOT_TREE_OBJECTID, DEFINE_NAME("root") }, 65 { .id = BTRFS_EXTENT_TREE_OBJECTID, DEFINE_NAME("extent") }, 66 { .id = BTRFS_CHUNK_TREE_OBJECTID, DEFINE_NAME("chunk") }, 67 { .id = BTRFS_DEV_TREE_OBJECTID, DEFINE_NAME("dev") }, 68 { .id = BTRFS_CSUM_TREE_OBJECTID, DEFINE_NAME("csum") }, 69 { .id = BTRFS_QUOTA_TREE_OBJECTID, DEFINE_NAME("quota") }, 70 { .id = BTRFS_TREE_LOG_OBJECTID, DEFINE_NAME("log") }, 71 { .id = BTRFS_TREE_RELOC_OBJECTID, DEFINE_NAME("treloc") }, 72 { .id = BTRFS_DATA_RELOC_TREE_OBJECTID, DEFINE_NAME("dreloc") }, 73 { .id = BTRFS_UUID_TREE_OBJECTID, DEFINE_NAME("uuid") }, 74 { .id = BTRFS_FREE_SPACE_TREE_OBJECTID, DEFINE_NAME("free-space") }, 75 { .id = BTRFS_BLOCK_GROUP_TREE_OBJECTID, DEFINE_NAME("block-group") }, 76 { .id = 0, DEFINE_NAME("tree") }, 77 }; 78 79 #undef DEFINE_LEVEL 80 #undef DEFINE_NAME 81 82 void btrfs_set_buffer_lockdep_class(u64 objectid, struct extent_buffer *eb, int level) 83 { 84 struct btrfs_lockdep_keyset *ks; 85 86 BUG_ON(level >= ARRAY_SIZE(ks->keys)); 87 88 /* Find the matching keyset, id 0 is the default entry */ 89 for (ks = btrfs_lockdep_keysets; ks->id; ks++) 90 if (ks->id == objectid) 91 break; 92 93 lockdep_set_class_and_name(&eb->lock, &ks->keys[level], ks->names[level]); 94 } 95 96 void btrfs_maybe_reset_lockdep_class(struct btrfs_root *root, struct extent_buffer *eb) 97 { 98 if (test_bit(BTRFS_ROOT_RESET_LOCKDEP_CLASS, &root->state)) 99 btrfs_set_buffer_lockdep_class(root->root_key.objectid, 100 eb, btrfs_header_level(eb)); 101 } 102 103 #endif 104 105 /* 106 * Extent buffer locking 107 * ===================== 108 * 109 * We use a rw_semaphore for tree locking, and the semantics are exactly the 110 * same: 111 * 112 * - reader/writer exclusion 113 * - writer/writer exclusion 114 * - reader/reader sharing 115 * - try-lock semantics for readers and writers 116 * 117 * The rwsem implementation does opportunistic spinning which reduces number of 118 * times the locking task needs to sleep. 119 */ 120 121 /* 122 * __btrfs_tree_read_lock - lock extent buffer for read 123 * @eb: the eb to be locked 124 * @nest: the nesting level to be used for lockdep 125 * 126 * This takes the read lock on the extent buffer, using the specified nesting 127 * level for lockdep purposes. 128 */ 129 void __btrfs_tree_read_lock(struct extent_buffer *eb, enum btrfs_lock_nesting nest) 130 { 131 u64 start_ns = 0; 132 133 if (trace_btrfs_tree_read_lock_enabled()) 134 start_ns = ktime_get_ns(); 135 136 down_read_nested(&eb->lock, nest); 137 trace_btrfs_tree_read_lock(eb, start_ns); 138 } 139 140 void btrfs_tree_read_lock(struct extent_buffer *eb) 141 { 142 __btrfs_tree_read_lock(eb, BTRFS_NESTING_NORMAL); 143 } 144 145 /* 146 * Try-lock for read. 147 * 148 * Return 1 if the rwlock has been taken, 0 otherwise 149 */ 150 int btrfs_try_tree_read_lock(struct extent_buffer *eb) 151 { 152 if (down_read_trylock(&eb->lock)) { 153 trace_btrfs_try_tree_read_lock(eb); 154 return 1; 155 } 156 return 0; 157 } 158 159 /* 160 * Try-lock for write. 161 * 162 * Return 1 if the rwlock has been taken, 0 otherwise 163 */ 164 int btrfs_try_tree_write_lock(struct extent_buffer *eb) 165 { 166 if (down_write_trylock(&eb->lock)) { 167 eb->lock_owner = current->pid; 168 trace_btrfs_try_tree_write_lock(eb); 169 return 1; 170 } 171 return 0; 172 } 173 174 /* 175 * Release read lock. 176 */ 177 void btrfs_tree_read_unlock(struct extent_buffer *eb) 178 { 179 trace_btrfs_tree_read_unlock(eb); 180 up_read(&eb->lock); 181 } 182 183 /* 184 * __btrfs_tree_lock - lock eb for write 185 * @eb: the eb to lock 186 * @nest: the nesting to use for the lock 187 * 188 * Returns with the eb->lock write locked. 189 */ 190 void __btrfs_tree_lock(struct extent_buffer *eb, enum btrfs_lock_nesting nest) 191 __acquires(&eb->lock) 192 { 193 u64 start_ns = 0; 194 195 if (trace_btrfs_tree_lock_enabled()) 196 start_ns = ktime_get_ns(); 197 198 down_write_nested(&eb->lock, nest); 199 eb->lock_owner = current->pid; 200 trace_btrfs_tree_lock(eb, start_ns); 201 } 202 203 void btrfs_tree_lock(struct extent_buffer *eb) 204 { 205 __btrfs_tree_lock(eb, BTRFS_NESTING_NORMAL); 206 } 207 208 /* 209 * Release the write lock. 210 */ 211 void btrfs_tree_unlock(struct extent_buffer *eb) 212 { 213 trace_btrfs_tree_unlock(eb); 214 eb->lock_owner = 0; 215 up_write(&eb->lock); 216 } 217 218 /* 219 * This releases any locks held in the path starting at level and going all the 220 * way up to the root. 221 * 222 * btrfs_search_slot will keep the lock held on higher nodes in a few corner 223 * cases, such as COW of the block at slot zero in the node. This ignores 224 * those rules, and it should only be called when there are no more updates to 225 * be done higher up in the tree. 226 */ 227 void btrfs_unlock_up_safe(struct btrfs_path *path, int level) 228 { 229 int i; 230 231 if (path->keep_locks) 232 return; 233 234 for (i = level; i < BTRFS_MAX_LEVEL; i++) { 235 if (!path->nodes[i]) 236 continue; 237 if (!path->locks[i]) 238 continue; 239 btrfs_tree_unlock_rw(path->nodes[i], path->locks[i]); 240 path->locks[i] = 0; 241 } 242 } 243 244 /* 245 * Loop around taking references on and locking the root node of the tree until 246 * we end up with a lock on the root node. 247 * 248 * Return: root extent buffer with write lock held 249 */ 250 struct extent_buffer *btrfs_lock_root_node(struct btrfs_root *root) 251 { 252 struct extent_buffer *eb; 253 254 while (1) { 255 eb = btrfs_root_node(root); 256 257 btrfs_maybe_reset_lockdep_class(root, eb); 258 btrfs_tree_lock(eb); 259 if (eb == root->node) 260 break; 261 btrfs_tree_unlock(eb); 262 free_extent_buffer(eb); 263 } 264 return eb; 265 } 266 267 /* 268 * Loop around taking references on and locking the root node of the tree until 269 * we end up with a lock on the root node. 270 * 271 * Return: root extent buffer with read lock held 272 */ 273 struct extent_buffer *btrfs_read_lock_root_node(struct btrfs_root *root) 274 { 275 struct extent_buffer *eb; 276 277 while (1) { 278 eb = btrfs_root_node(root); 279 280 btrfs_maybe_reset_lockdep_class(root, eb); 281 btrfs_tree_read_lock(eb); 282 if (eb == root->node) 283 break; 284 btrfs_tree_read_unlock(eb); 285 free_extent_buffer(eb); 286 } 287 return eb; 288 } 289 290 /* 291 * Loop around taking references on and locking the root node of the tree in 292 * nowait mode until we end up with a lock on the root node or returning to 293 * avoid blocking. 294 * 295 * Return: root extent buffer with read lock held or -EAGAIN. 296 */ 297 struct extent_buffer *btrfs_try_read_lock_root_node(struct btrfs_root *root) 298 { 299 struct extent_buffer *eb; 300 301 while (1) { 302 eb = btrfs_root_node(root); 303 if (!btrfs_try_tree_read_lock(eb)) { 304 free_extent_buffer(eb); 305 return ERR_PTR(-EAGAIN); 306 } 307 if (eb == root->node) 308 break; 309 btrfs_tree_read_unlock(eb); 310 free_extent_buffer(eb); 311 } 312 return eb; 313 } 314 315 /* 316 * DREW locks 317 * ========== 318 * 319 * DREW stands for double-reader-writer-exclusion lock. It's used in situation 320 * where you want to provide A-B exclusion but not AA or BB. 321 * 322 * Currently implementation gives more priority to reader. If a reader and a 323 * writer both race to acquire their respective sides of the lock the writer 324 * would yield its lock as soon as it detects a concurrent reader. Additionally 325 * if there are pending readers no new writers would be allowed to come in and 326 * acquire the lock. 327 */ 328 329 void btrfs_drew_lock_init(struct btrfs_drew_lock *lock) 330 { 331 atomic_set(&lock->readers, 0); 332 atomic_set(&lock->writers, 0); 333 init_waitqueue_head(&lock->pending_readers); 334 init_waitqueue_head(&lock->pending_writers); 335 } 336 337 /* Return true if acquisition is successful, false otherwise */ 338 bool btrfs_drew_try_write_lock(struct btrfs_drew_lock *lock) 339 { 340 if (atomic_read(&lock->readers)) 341 return false; 342 343 atomic_inc(&lock->writers); 344 345 /* Ensure writers count is updated before we check for pending readers */ 346 smp_mb__after_atomic(); 347 if (atomic_read(&lock->readers)) { 348 btrfs_drew_write_unlock(lock); 349 return false; 350 } 351 352 return true; 353 } 354 355 void btrfs_drew_write_lock(struct btrfs_drew_lock *lock) 356 { 357 while (true) { 358 if (btrfs_drew_try_write_lock(lock)) 359 return; 360 wait_event(lock->pending_writers, !atomic_read(&lock->readers)); 361 } 362 } 363 364 void btrfs_drew_write_unlock(struct btrfs_drew_lock *lock) 365 { 366 atomic_dec(&lock->writers); 367 cond_wake_up(&lock->pending_readers); 368 } 369 370 void btrfs_drew_read_lock(struct btrfs_drew_lock *lock) 371 { 372 atomic_inc(&lock->readers); 373 374 /* 375 * Ensure the pending reader count is perceieved BEFORE this reader 376 * goes to sleep in case of active writers. This guarantees new writers 377 * won't be allowed and that the current reader will be woken up when 378 * the last active writer finishes its jobs. 379 */ 380 smp_mb__after_atomic(); 381 382 wait_event(lock->pending_readers, atomic_read(&lock->writers) == 0); 383 } 384 385 void btrfs_drew_read_unlock(struct btrfs_drew_lock *lock) 386 { 387 /* 388 * atomic_dec_and_test implies a full barrier, so woken up writers 389 * are guaranteed to see the decrement 390 */ 391 if (atomic_dec_and_test(&lock->readers)) 392 wake_up(&lock->pending_writers); 393 } 394