git.sesse.net Git - bcachefs-tools-debian/blob - libbcachefs/btree_cache.c

   1 // SPDX-License-Identifier: GPL-2.0
   2
   3 #include "bcachefs.h"
   4 #include "bkey_buf.h"
   5 #include "btree_cache.h"
   6 #include "btree_io.h"
   7 #include "btree_iter.h"
   8 #include "btree_locking.h"
   9 #include "debug.h"
  10 #include "errcode.h"
  11 #include "error.h"
  12
  13 #include <linux/prefetch.h>
  14 #include <linux/sched/mm.h>
  15 #include <linux/seq_buf.h>
  16 #include <trace/events/bcachefs.h>
  17
  18 #define BTREE_CACHE_NOT_FREED_INCREMENT(counter) \
  19 do {                                             \
  20         if (shrinker_counter)                    \
  21                 bc->not_freed_##counter++;       \
  22 } while (0)
  23
  24 const char * const bch2_btree_node_flags[] = {
  25 #define x(f)    #f,
  26         BTREE_FLAGS()
  27 #undef x
  28         NULL
  29 };
  30
  31 void bch2_recalc_btree_reserve(struct bch_fs *c)
  32 {
  33         unsigned i, reserve = 16;
  34
  35         if (!c->btree_roots[0].b)
  36                 reserve += 8;
  37
  38         for (i = 0; i < BTREE_ID_NR; i++)
  39                 if (c->btree_roots[i].b)
  40                         reserve += min_t(unsigned, 1,
  41                                          c->btree_roots[i].b->c.level) * 8;
  42
  43         c->btree_cache.reserve = reserve;
  44 }
  45
  46 static inline unsigned btree_cache_can_free(struct btree_cache *bc)
  47 {
  48         return max_t(int, 0, bc->used - bc->reserve);
  49 }
  50
  51 static void btree_node_to_freedlist(struct btree_cache *bc, struct btree *b)
  52 {
  53         if (b->c.lock.readers)
  54                 list_move(&b->list, &bc->freed_pcpu);
  55         else
  56                 list_move(&b->list, &bc->freed_nonpcpu);
  57 }
  58
  59 static void btree_node_data_free(struct bch_fs *c, struct btree *b)
  60 {
  61         struct btree_cache *bc = &c->btree_cache;
  62
  63         EBUG_ON(btree_node_write_in_flight(b));
  64
  65         kvpfree(b->data, btree_bytes(c));
  66         b->data = NULL;
  67 #ifdef __KERNEL__
  68         vfree(b->aux_data);
  69 #else
  70         munmap(b->aux_data, btree_aux_data_bytes(b));
  71 #endif
  72         b->aux_data = NULL;
  73
  74         bc->used--;
  75
  76         btree_node_to_freedlist(bc, b);
  77 }
  78
  79 static int bch2_btree_cache_cmp_fn(struct rhashtable_compare_arg *arg,
  80                                    const void *obj)
  81 {
  82         const struct btree *b = obj;
  83         const u64 *v = arg->key;
  84
  85         return b->hash_val == *v ? 0 : 1;
  86 }
  87
  88 static const struct rhashtable_params bch_btree_cache_params = {
  89         .head_offset    = offsetof(struct btree, hash),
  90         .key_offset     = offsetof(struct btree, hash_val),
  91         .key_len        = sizeof(u64),
  92         .obj_cmpfn      = bch2_btree_cache_cmp_fn,
  93 };
  94
  95 static int btree_node_data_alloc(struct bch_fs *c, struct btree *b, gfp_t gfp)
  96 {
  97         BUG_ON(b->data || b->aux_data);
  98
  99         b->data = kvpmalloc(btree_bytes(c), gfp);
 100         if (!b->data)
 101                 return -ENOMEM;
 102 #ifdef __KERNEL__
 103         b->aux_data = vmalloc_exec(btree_aux_data_bytes(b), gfp);
 104 #else
 105         b->aux_data = mmap(NULL, btree_aux_data_bytes(b),
 106                            PROT_READ|PROT_WRITE|PROT_EXEC,
 107                            MAP_PRIVATE|MAP_ANONYMOUS, 0, 0);
 108         if (b->aux_data == MAP_FAILED)
 109                 b->aux_data = NULL;
 110 #endif
 111         if (!b->aux_data) {
 112                 kvpfree(b->data, btree_bytes(c));
 113                 b->data = NULL;
 114                 return -ENOMEM;
 115         }
 116
 117         return 0;
 118 }
 119
 120 static struct btree *__btree_node_mem_alloc(struct bch_fs *c, gfp_t gfp)
 121 {
 122         struct btree *b;
 123
 124         b = kzalloc(sizeof(struct btree), gfp);
 125         if (!b)
 126                 return NULL;
 127
 128         bkey_btree_ptr_init(&b->key);
 129         __six_lock_init(&b->c.lock, "b->c.lock", &bch2_btree_node_lock_key);
 130 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 131         lockdep_set_no_check_recursion(&b->c.lock.dep_map);
 132 #endif
 133         INIT_LIST_HEAD(&b->list);
 134         INIT_LIST_HEAD(&b->write_blocked);
 135         b->byte_order = ilog2(btree_bytes(c));
 136         return b;
 137 }
 138
 139 struct btree *__bch2_btree_node_mem_alloc(struct bch_fs *c)
 140 {
 141         struct btree_cache *bc = &c->btree_cache;
 142         struct btree *b;
 143
 144         b = __btree_node_mem_alloc(c, GFP_KERNEL);
 145         if (!b)
 146                 return NULL;
 147
 148         if (btree_node_data_alloc(c, b, GFP_KERNEL)) {
 149                 kfree(b);
 150                 return NULL;
 151         }
 152
 153         bc->used++;
 154         list_add(&b->list, &bc->freeable);
 155         return b;
 156 }
 157
 158 /* Btree in memory cache - hash table */
 159
 160 void bch2_btree_node_hash_remove(struct btree_cache *bc, struct btree *b)
 161 {
 162         int ret = rhashtable_remove_fast(&bc->table, &b->hash, bch_btree_cache_params);
 163
 164         BUG_ON(ret);
 165
 166         /* Cause future lookups for this node to fail: */
 167         b->hash_val = 0;
 168 }
 169
 170 int __bch2_btree_node_hash_insert(struct btree_cache *bc, struct btree *b)
 171 {
 172         BUG_ON(b->hash_val);
 173         b->hash_val = btree_ptr_hash_val(&b->key);
 174
 175         return rhashtable_lookup_insert_fast(&bc->table, &b->hash,
 176                                              bch_btree_cache_params);
 177 }
 178
 179 int bch2_btree_node_hash_insert(struct btree_cache *bc, struct btree *b,
 180                                 unsigned level, enum btree_id id)
 181 {
 182         int ret;
 183
 184         b->c.level      = level;
 185         b->c.btree_id   = id;
 186
 187         mutex_lock(&bc->lock);
 188         ret = __bch2_btree_node_hash_insert(bc, b);
 189         if (!ret)
 190                 list_add_tail(&b->list, &bc->live);
 191         mutex_unlock(&bc->lock);
 192
 193         return ret;
 194 }
 195
 196 __flatten
 197 static inline struct btree *btree_cache_find(struct btree_cache *bc,
 198                                      const struct bkey_i *k)
 199 {
 200         u64 v = btree_ptr_hash_val(k);
 201
 202         return rhashtable_lookup_fast(&bc->table, &v, bch_btree_cache_params);
 203 }
 204
 205 /*
 206  * this version is for btree nodes that have already been freed (we're not
 207  * reaping a real btree node)
 208  */
 209 static int __btree_node_reclaim(struct bch_fs *c, struct btree *b, bool flush, bool shrinker_counter)
 210 {
 211         struct btree_cache *bc = &c->btree_cache;
 212         int ret = 0;
 213
 214         lockdep_assert_held(&bc->lock);
 215 wait_on_io:
 216         if (b->flags & ((1U << BTREE_NODE_dirty)|
 217                         (1U << BTREE_NODE_read_in_flight)|
 218                         (1U << BTREE_NODE_write_in_flight))) {
 219                 if (!flush) {
 220                         if (btree_node_dirty(b))
 221                                 BTREE_CACHE_NOT_FREED_INCREMENT(dirty);
 222                         else if (btree_node_read_in_flight(b))
 223                                 BTREE_CACHE_NOT_FREED_INCREMENT(read_in_flight);
 224                         else if (btree_node_write_in_flight(b))
 225                                 BTREE_CACHE_NOT_FREED_INCREMENT(write_in_flight);
 226                         return -ENOMEM;
 227                 }
 228
 229                 /* XXX: waiting on IO with btree cache lock held */
 230                 bch2_btree_node_wait_on_read(b);
 231                 bch2_btree_node_wait_on_write(b);
 232         }
 233
 234         if (!six_trylock_intent(&b->c.lock)) {
 235                 BTREE_CACHE_NOT_FREED_INCREMENT(lock_intent);
 236                 return -ENOMEM;
 237         }
 238
 239         if (!six_trylock_write(&b->c.lock)) {
 240                 BTREE_CACHE_NOT_FREED_INCREMENT(lock_write);
 241                 goto out_unlock_intent;
 242         }
 243
 244         /* recheck under lock */
 245         if (b->flags & ((1U << BTREE_NODE_read_in_flight)|
 246                         (1U << BTREE_NODE_write_in_flight))) {
 247                 if (!flush) {
 248                         if (btree_node_read_in_flight(b))
 249                                 BTREE_CACHE_NOT_FREED_INCREMENT(read_in_flight);
 250                         else if (btree_node_write_in_flight(b))
 251                                 BTREE_CACHE_NOT_FREED_INCREMENT(write_in_flight);
 252                         goto out_unlock;
 253                 }
 254                 six_unlock_write(&b->c.lock);
 255                 six_unlock_intent(&b->c.lock);
 256                 goto wait_on_io;
 257         }
 258
 259         if (btree_node_noevict(b)) {
 260                 BTREE_CACHE_NOT_FREED_INCREMENT(noevict);
 261                 goto out_unlock;
 262         }
 263         if (btree_node_write_blocked(b)) {
 264                 BTREE_CACHE_NOT_FREED_INCREMENT(write_blocked);
 265                 goto out_unlock;
 266         }
 267         if (btree_node_will_make_reachable(b)) {
 268                 BTREE_CACHE_NOT_FREED_INCREMENT(will_make_reachable);
 269                 goto out_unlock;
 270         }
 271
 272         if (btree_node_dirty(b)) {
 273                 if (!flush) {
 274                         BTREE_CACHE_NOT_FREED_INCREMENT(dirty);
 275                         goto out_unlock;
 276                 }
 277                 /*
 278                  * Using the underscore version because we don't want to compact
 279                  * bsets after the write, since this node is about to be evicted
 280                  * - unless btree verify mode is enabled, since it runs out of
 281                  * the post write cleanup:
 282                  */
 283                 if (bch2_verify_btree_ondisk)
 284                         bch2_btree_node_write(c, b, SIX_LOCK_intent,
 285                                               BTREE_WRITE_cache_reclaim);
 286                 else
 287                         __bch2_btree_node_write(c, b,
 288                                                 BTREE_WRITE_cache_reclaim);
 289
 290                 six_unlock_write(&b->c.lock);
 291                 six_unlock_intent(&b->c.lock);
 292                 goto wait_on_io;
 293         }
 294 out:
 295         if (b->hash_val && !ret)
 296                 trace_and_count(c, btree_cache_reap, c, b);
 297         return ret;
 298 out_unlock:
 299         six_unlock_write(&b->c.lock);
 300 out_unlock_intent:
 301         six_unlock_intent(&b->c.lock);
 302         ret = -ENOMEM;
 303         goto out;
 304 }
 305
 306 static int btree_node_reclaim(struct bch_fs *c, struct btree *b, bool shrinker_counter)
 307 {
 308         return __btree_node_reclaim(c, b, false, shrinker_counter);
 309 }
 310
 311 static int btree_node_write_and_reclaim(struct bch_fs *c, struct btree *b)
 312 {
 313         return __btree_node_reclaim(c, b, true, false);
 314 }
 315
 316 static unsigned long bch2_btree_cache_scan(struct shrinker *shrink,
 317                                            struct shrink_control *sc)
 318 {
 319         struct bch_fs *c = container_of(shrink, struct bch_fs,
 320                                         btree_cache.shrink);
 321         struct btree_cache *bc = &c->btree_cache;
 322         struct btree *b, *t;
 323         unsigned long nr = sc->nr_to_scan;
 324         unsigned long can_free = 0;
 325         unsigned long freed = 0;
 326         unsigned long touched = 0;
 327         unsigned i, flags;
 328         unsigned long ret = SHRINK_STOP;
 329         bool trigger_writes = atomic_read(&bc->dirty) + nr >=
 330                 bc->used * 3 / 4;
 331
 332         if (bch2_btree_shrinker_disabled)
 333                 return SHRINK_STOP;
 334
 335         mutex_lock(&bc->lock);
 336         flags = memalloc_nofs_save();
 337
 338         /*
 339          * It's _really_ critical that we don't free too many btree nodes - we
 340          * have to always leave ourselves a reserve. The reserve is how we
 341          * guarantee that allocating memory for a new btree node can always
 342          * succeed, so that inserting keys into the btree can always succeed and
 343          * IO can always make forward progress:
 344          */
 345         can_free = btree_cache_can_free(bc);
 346         nr = min_t(unsigned long, nr, can_free);
 347
 348         i = 0;
 349         list_for_each_entry_safe(b, t, &bc->freeable, list) {
 350                 /*
 351                  * Leave a few nodes on the freeable list, so that a btree split
 352                  * won't have to hit the system allocator:
 353                  */
 354                 if (++i <= 3)
 355                         continue;
 356
 357                 touched++;
 358
 359                 if (touched >= nr)
 360                         goto out;
 361
 362                 if (!btree_node_reclaim(c, b, true)) {
 363                         btree_node_data_free(c, b);
 364                         six_unlock_write(&b->c.lock);
 365                         six_unlock_intent(&b->c.lock);
 366                         freed++;
 367                         bc->freed++;
 368                 }
 369         }
 370 restart:
 371         list_for_each_entry_safe(b, t, &bc->live, list) {
 372                 touched++;
 373
 374                 if (btree_node_accessed(b)) {
 375                         clear_btree_node_accessed(b);
 376                         bc->not_freed_access_bit++;
 377                 } else if (!btree_node_reclaim(c, b, true)) {
 378                         freed++;
 379                         btree_node_data_free(c, b);
 380                         bc->freed++;
 381
 382                         bch2_btree_node_hash_remove(bc, b);
 383                         six_unlock_write(&b->c.lock);
 384                         six_unlock_intent(&b->c.lock);
 385
 386                         if (freed == nr)
 387                                 goto out_rotate;
 388                 } else if (trigger_writes &&
 389                            btree_node_dirty(b) &&
 390                            !btree_node_will_make_reachable(b) &&
 391                            !btree_node_write_blocked(b) &&
 392                            six_trylock_read(&b->c.lock)) {
 393                         list_move(&bc->live, &b->list);
 394                         mutex_unlock(&bc->lock);
 395                         __bch2_btree_node_write(c, b, BTREE_WRITE_cache_reclaim);
 396                         six_unlock_read(&b->c.lock);
 397                         if (touched >= nr)
 398                                 goto out_nounlock;
 399                         mutex_lock(&bc->lock);
 400                         goto restart;
 401                 }
 402
 403                 if (touched >= nr)
 404                         break;
 405         }
 406 out_rotate:
 407         if (&t->list != &bc->live)
 408                 list_move_tail(&bc->live, &t->list);
 409 out:
 410         mutex_unlock(&bc->lock);
 411 out_nounlock:
 412         ret = freed;
 413         memalloc_nofs_restore(flags);
 414         trace_and_count(c, btree_cache_scan, sc->nr_to_scan, can_free, ret);
 415         return ret;
 416 }
 417
 418 static unsigned long bch2_btree_cache_count(struct shrinker *shrink,
 419                                             struct shrink_control *sc)
 420 {
 421         struct bch_fs *c = container_of(shrink, struct bch_fs,
 422                                         btree_cache.shrink);
 423         struct btree_cache *bc = &c->btree_cache;
 424
 425         if (bch2_btree_shrinker_disabled)
 426                 return 0;
 427
 428         return btree_cache_can_free(bc);
 429 }
 430
 431 static void bch2_btree_cache_shrinker_to_text(struct seq_buf *s, struct shrinker *shrink)
 432 {
 433         struct bch_fs *c = container_of(shrink, struct bch_fs,
 434                                         btree_cache.shrink);
 435         char *cbuf;
 436         size_t buflen = seq_buf_get_buf(s, &cbuf);
 437         struct printbuf out = PRINTBUF_EXTERN(cbuf, buflen);
 438
 439         bch2_btree_cache_to_text(&out, &c->btree_cache);
 440         seq_buf_commit(s, out.pos);
 441 }
 442
 443 void bch2_fs_btree_cache_exit(struct bch_fs *c)
 444 {
 445         struct btree_cache *bc = &c->btree_cache;
 446         struct btree *b;
 447         unsigned i, flags;
 448
 449         if (bc->shrink.list.next)
 450                 unregister_shrinker(&bc->shrink);
 451
 452         /* vfree() can allocate memory: */
 453         flags = memalloc_nofs_save();
 454         mutex_lock(&bc->lock);
 455
 456         if (c->verify_data)
 457                 list_move(&c->verify_data->list, &bc->live);
 458
 459         kvpfree(c->verify_ondisk, btree_bytes(c));
 460
 461         for (i = 0; i < BTREE_ID_NR; i++)
 462                 if (c->btree_roots[i].b)
 463                         list_add(&c->btree_roots[i].b->list, &bc->live);
 464
 465         list_splice(&bc->freeable, &bc->live);
 466
 467         while (!list_empty(&bc->live)) {
 468                 b = list_first_entry(&bc->live, struct btree, list);
 469
 470                 BUG_ON(btree_node_read_in_flight(b) ||
 471                        btree_node_write_in_flight(b));
 472
 473                 if (btree_node_dirty(b))
 474                         bch2_btree_complete_write(c, b, btree_current_write(b));
 475                 clear_btree_node_dirty_acct(c, b);
 476
 477                 btree_node_data_free(c, b);
 478         }
 479
 480         BUG_ON(atomic_read(&c->btree_cache.dirty));
 481
 482         list_splice(&bc->freed_pcpu, &bc->freed_nonpcpu);
 483
 484         while (!list_empty(&bc->freed_nonpcpu)) {
 485                 b = list_first_entry(&bc->freed_nonpcpu, struct btree, list);
 486                 list_del(&b->list);
 487                 six_lock_pcpu_free(&b->c.lock);
 488                 kfree(b);
 489         }
 490
 491         mutex_unlock(&bc->lock);
 492         memalloc_nofs_restore(flags);
 493
 494         if (bc->table_init_done)
 495                 rhashtable_destroy(&bc->table);
 496 }
 497
 498 int bch2_fs_btree_cache_init(struct bch_fs *c)
 499 {
 500         struct btree_cache *bc = &c->btree_cache;
 501         unsigned i;
 502         int ret = 0;
 503
 504         pr_verbose_init(c->opts, "");
 505
 506         ret = rhashtable_init(&bc->table, &bch_btree_cache_params);
 507         if (ret)
 508                 goto out;
 509
 510         bc->table_init_done = true;
 511
 512         bch2_recalc_btree_reserve(c);
 513
 514         for (i = 0; i < bc->reserve; i++)
 515                 if (!__bch2_btree_node_mem_alloc(c)) {
 516                         ret = -ENOMEM;
 517                         goto out;
 518                 }
 519
 520         list_splice_init(&bc->live, &bc->freeable);
 521
 522         mutex_init(&c->verify_lock);
 523
 524         bc->shrink.count_objects        = bch2_btree_cache_count;
 525         bc->shrink.scan_objects         = bch2_btree_cache_scan;
 526         bc->shrink.to_text              = bch2_btree_cache_shrinker_to_text;
 527         bc->shrink.seeks                = 4;
 528         ret = register_shrinker(&bc->shrink, "%s/btree_cache", c->name);
 529 out:
 530         pr_verbose_init(c->opts, "ret %i", ret);
 531         return ret;
 532 }
 533
 534 void bch2_fs_btree_cache_init_early(struct btree_cache *bc)
 535 {
 536         mutex_init(&bc->lock);
 537         INIT_LIST_HEAD(&bc->live);
 538         INIT_LIST_HEAD(&bc->freeable);
 539         INIT_LIST_HEAD(&bc->freed_pcpu);
 540         INIT_LIST_HEAD(&bc->freed_nonpcpu);
 541 }
 542
 543 /*
 544  * We can only have one thread cannibalizing other cached btree nodes at a time,
 545  * or we'll deadlock. We use an open coded mutex to ensure that, which a
 546  * cannibalize_bucket() will take. This means every time we unlock the root of
 547  * the btree, we need to release this lock if we have it held.
 548  */
 549 void bch2_btree_cache_cannibalize_unlock(struct bch_fs *c)
 550 {
 551         struct btree_cache *bc = &c->btree_cache;
 552
 553         if (bc->alloc_lock == current) {
 554                 trace_and_count(c, btree_cache_cannibalize_unlock, c);
 555                 bc->alloc_lock = NULL;
 556                 closure_wake_up(&bc->alloc_wait);
 557         }
 558 }
 559
 560 int bch2_btree_cache_cannibalize_lock(struct bch_fs *c, struct closure *cl)
 561 {
 562         struct btree_cache *bc = &c->btree_cache;
 563         struct task_struct *old;
 564
 565         old = cmpxchg(&bc->alloc_lock, NULL, current);
 566         if (old == NULL || old == current)
 567                 goto success;
 568
 569         if (!cl) {
 570                 trace_and_count(c, btree_cache_cannibalize_lock_fail, c);
 571                 return -ENOMEM;
 572         }
 573
 574         closure_wait(&bc->alloc_wait, cl);
 575
 576         /* Try again, after adding ourselves to waitlist */
 577         old = cmpxchg(&bc->alloc_lock, NULL, current);
 578         if (old == NULL || old == current) {
 579                 /* We raced */
 580                 closure_wake_up(&bc->alloc_wait);
 581                 goto success;
 582         }
 583
 584         trace_and_count(c, btree_cache_cannibalize_lock_fail, c);
 585         return -BCH_ERR_btree_cache_cannibalize_lock_blocked;
 586
 587 success:
 588         trace_and_count(c, btree_cache_cannibalize_lock, c);
 589         return 0;
 590 }
 591
 592 static struct btree *btree_node_cannibalize(struct bch_fs *c)
 593 {
 594         struct btree_cache *bc = &c->btree_cache;
 595         struct btree *b;
 596
 597         list_for_each_entry_reverse(b, &bc->live, list)
 598                 if (!btree_node_reclaim(c, b, false))
 599                         return b;
 600
 601         while (1) {
 602                 list_for_each_entry_reverse(b, &bc->live, list)
 603                         if (!btree_node_write_and_reclaim(c, b))
 604                                 return b;
 605
 606                 /*
 607                  * Rare case: all nodes were intent-locked.
 608                  * Just busy-wait.
 609                  */
 610                 WARN_ONCE(1, "btree cache cannibalize failed\n");
 611                 cond_resched();
 612         }
 613 }
 614
 615 struct btree *bch2_btree_node_mem_alloc(struct bch_fs *c, bool pcpu_read_locks)
 616 {
 617         struct btree_cache *bc = &c->btree_cache;
 618         struct list_head *freed = pcpu_read_locks
 619                 ? &bc->freed_pcpu
 620                 : &bc->freed_nonpcpu;
 621         struct btree *b, *b2;
 622         u64 start_time = local_clock();
 623         unsigned flags;
 624
 625         flags = memalloc_nofs_save();
 626         mutex_lock(&bc->lock);
 627
 628         /*
 629          * We never free struct btree itself, just the memory that holds the on
 630          * disk node. Check the freed list before allocating a new one:
 631          */
 632         list_for_each_entry(b, freed, list)
 633                 if (!btree_node_reclaim(c, b, false)) {
 634                         list_del_init(&b->list);
 635                         goto got_node;
 636                 }
 637
 638         b = __btree_node_mem_alloc(c, __GFP_NOWARN);
 639         if (!b) {
 640                 mutex_unlock(&bc->lock);
 641                 b = __btree_node_mem_alloc(c, GFP_KERNEL);
 642                 if (!b)
 643                         goto err;
 644                 mutex_lock(&bc->lock);
 645         }
 646
 647         if (pcpu_read_locks)
 648                 six_lock_pcpu_alloc(&b->c.lock);
 649
 650         BUG_ON(!six_trylock_intent(&b->c.lock));
 651         BUG_ON(!six_trylock_write(&b->c.lock));
 652 got_node:
 653
 654         /*
 655          * btree_free() doesn't free memory; it sticks the node on the end of
 656          * the list. Check if there's any freed nodes there:
 657          */
 658         list_for_each_entry(b2, &bc->freeable, list)
 659                 if (!btree_node_reclaim(c, b2, false)) {
 660                         swap(b->data, b2->data);
 661                         swap(b->aux_data, b2->aux_data);
 662                         btree_node_to_freedlist(bc, b2);
 663                         six_unlock_write(&b2->c.lock);
 664                         six_unlock_intent(&b2->c.lock);
 665                         goto got_mem;
 666                 }
 667
 668         mutex_unlock(&bc->lock);
 669
 670         if (btree_node_data_alloc(c, b, __GFP_NOWARN|GFP_KERNEL))
 671                 goto err;
 672
 673         mutex_lock(&bc->lock);
 674         bc->used++;
 675 got_mem:
 676         mutex_unlock(&bc->lock);
 677
 678         BUG_ON(btree_node_hashed(b));
 679         BUG_ON(btree_node_dirty(b));
 680         BUG_ON(btree_node_write_in_flight(b));
 681 out:
 682         b->flags                = 0;
 683         b->written              = 0;
 684         b->nsets                = 0;
 685         b->sib_u64s[0]          = 0;
 686         b->sib_u64s[1]          = 0;
 687         b->whiteout_u64s        = 0;
 688         bch2_btree_keys_init(b);
 689         set_btree_node_accessed(b);
 690
 691         bch2_time_stats_update(&c->times[BCH_TIME_btree_node_mem_alloc],
 692                                start_time);
 693
 694         memalloc_nofs_restore(flags);
 695         return b;
 696 err:
 697         mutex_lock(&bc->lock);
 698
 699         /* Try to cannibalize another cached btree node: */
 700         if (bc->alloc_lock == current) {
 701                 b2 = btree_node_cannibalize(c);
 702                 bch2_btree_node_hash_remove(bc, b2);
 703
 704                 if (b) {
 705                         swap(b->data, b2->data);
 706                         swap(b->aux_data, b2->aux_data);
 707                         btree_node_to_freedlist(bc, b2);
 708                         six_unlock_write(&b2->c.lock);
 709                         six_unlock_intent(&b2->c.lock);
 710                 } else {
 711                         b = b2;
 712                         list_del_init(&b->list);
 713                 }
 714
 715                 mutex_unlock(&bc->lock);
 716
 717                 trace_and_count(c, btree_cache_cannibalize, c);
 718                 goto out;
 719         }
 720
 721         mutex_unlock(&bc->lock);
 722         memalloc_nofs_restore(flags);
 723         return ERR_PTR(-ENOMEM);
 724 }
 725
 726 /* Slowpath, don't want it inlined into btree_iter_traverse() */
 727 static noinline struct btree *bch2_btree_node_fill(struct bch_fs *c,
 728                                 struct btree_trans *trans,
 729                                 struct btree_path *path,
 730                                 const struct bkey_i *k,
 731                                 enum btree_id btree_id,
 732                                 unsigned level,
 733                                 enum six_lock_type lock_type,
 734                                 bool sync)
 735 {
 736         struct btree_cache *bc = &c->btree_cache;
 737         struct btree *b;
 738         u32 seq;
 739
 740         BUG_ON(level + 1 >= BTREE_MAX_DEPTH);
 741         /*
 742          * Parent node must be locked, else we could read in a btree node that's
 743          * been freed:
 744          */
 745         if (trans && !bch2_btree_node_relock(trans, path, level + 1)) {
 746                 trace_and_count(c, trans_restart_relock_parent_for_fill, trans, _THIS_IP_, path);
 747                 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_fill_relock));
 748         }
 749
 750         b = bch2_btree_node_mem_alloc(c, level != 0);
 751
 752         if (trans && b == ERR_PTR(-ENOMEM)) {
 753                 trans->memory_allocation_failure = true;
 754                 trace_and_count(c, trans_restart_memory_allocation_failure, trans, _THIS_IP_, path);
 755                 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_fill_mem_alloc_fail));
 756         }
 757
 758         if (IS_ERR(b))
 759                 return b;
 760
 761         /*
 762          * Btree nodes read in from disk should not have the accessed bit set
 763          * initially, so that linear scans don't thrash the cache:
 764          */
 765         clear_btree_node_accessed(b);
 766
 767         bkey_copy(&b->key, k);
 768         if (bch2_btree_node_hash_insert(bc, b, level, btree_id)) {
 769                 /* raced with another fill: */
 770
 771                 /* mark as unhashed... */
 772                 b->hash_val = 0;
 773
 774                 mutex_lock(&bc->lock);
 775                 list_add(&b->list, &bc->freeable);
 776                 mutex_unlock(&bc->lock);
 777
 778                 six_unlock_write(&b->c.lock);
 779                 six_unlock_intent(&b->c.lock);
 780                 return NULL;
 781         }
 782
 783         set_btree_node_read_in_flight(b);
 784
 785         six_unlock_write(&b->c.lock);
 786         seq = b->c.lock.state.seq;
 787         six_unlock_intent(&b->c.lock);
 788
 789         /* Unlock before doing IO: */
 790         if (trans && sync)
 791                 bch2_trans_unlock(trans);
 792
 793         bch2_btree_node_read(c, b, sync);
 794
 795         if (!sync)
 796                 return NULL;
 797
 798         if (trans) {
 799                 int ret = bch2_trans_relock(trans) ?:
 800                         bch2_btree_path_relock_intent(trans, path);
 801                 if (ret) {
 802                         BUG_ON(!trans->restarted);
 803                         return ERR_PTR(ret);
 804                 }
 805         }
 806
 807         if (!six_relock_type(&b->c.lock, lock_type, seq)) {
 808                 if (trans)
 809                         trace_and_count(c, trans_restart_relock_after_fill, trans, _THIS_IP_, path);
 810                 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_relock_after_fill));
 811         }
 812
 813         return b;
 814 }
 815
 816 static noinline void btree_bad_header(struct bch_fs *c, struct btree *b)
 817 {
 818         struct printbuf buf = PRINTBUF;
 819
 820         if (!test_bit(BCH_FS_INITIAL_GC_DONE, &c->flags))
 821                 return;
 822
 823         prt_printf(&buf,
 824                "btree node header doesn't match ptr\n"
 825                "btree %s level %u\n"
 826                "ptr: ",
 827                bch2_btree_ids[b->c.btree_id], b->c.level);
 828         bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key));
 829
 830         prt_printf(&buf, "\nheader: btree %s level %llu\n"
 831                "min ",
 832                bch2_btree_ids[BTREE_NODE_ID(b->data)],
 833                BTREE_NODE_LEVEL(b->data));
 834         bch2_bpos_to_text(&buf, b->data->min_key);
 835
 836         prt_printf(&buf, "\nmax ");
 837         bch2_bpos_to_text(&buf, b->data->max_key);
 838
 839         bch2_fs_inconsistent(c, "%s", buf.buf);
 840         printbuf_exit(&buf);
 841 }
 842
 843 static inline void btree_check_header(struct bch_fs *c, struct btree *b)
 844 {
 845         if (b->c.btree_id != BTREE_NODE_ID(b->data) ||
 846             b->c.level != BTREE_NODE_LEVEL(b->data) ||
 847             !bpos_eq(b->data->max_key, b->key.k.p) ||
 848             (b->key.k.type == KEY_TYPE_btree_ptr_v2 &&
 849              !bpos_eq(b->data->min_key,
 850                       bkey_i_to_btree_ptr_v2(&b->key)->v.min_key)))
 851                 btree_bad_header(c, b);
 852 }
 853
 854 static struct btree *__bch2_btree_node_get(struct btree_trans *trans, struct btree_path *path,
 855                                            const struct bkey_i *k, unsigned level,
 856                                            enum six_lock_type lock_type,
 857                                            unsigned long trace_ip)
 858 {
 859         struct bch_fs *c = trans->c;
 860         struct btree_cache *bc = &c->btree_cache;
 861         struct btree *b;
 862         struct bset_tree *t;
 863         int ret;
 864
 865         EBUG_ON(level >= BTREE_MAX_DEPTH);
 866 retry:
 867         b = btree_cache_find(bc, k);
 868         if (unlikely(!b)) {
 869                 /*
 870                  * We must have the parent locked to call bch2_btree_node_fill(),
 871                  * else we could read in a btree node from disk that's been
 872                  * freed:
 873                  */
 874                 b = bch2_btree_node_fill(c, trans, path, k, path->btree_id,
 875                                          level, lock_type, true);
 876
 877                 /* We raced and found the btree node in the cache */
 878                 if (!b)
 879                         goto retry;
 880
 881                 if (IS_ERR(b))
 882                         return b;
 883         } else {
 884                 if (btree_node_read_locked(path, level + 1))
 885                         btree_node_unlock(trans, path, level + 1);
 886
 887                 ret = btree_node_lock(trans, path, &b->c, level, lock_type, trace_ip);
 888                 if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
 889                         return ERR_PTR(ret);
 890
 891                 BUG_ON(ret);
 892
 893                 if (unlikely(b->hash_val != btree_ptr_hash_val(k) ||
 894                              b->c.level != level ||
 895                              race_fault())) {
 896                         six_unlock_type(&b->c.lock, lock_type);
 897                         if (bch2_btree_node_relock(trans, path, level + 1))
 898                                 goto retry;
 899
 900                         trace_and_count(c, trans_restart_btree_node_reused, trans, trace_ip, path);
 901                         return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_lock_node_reused));
 902                 }
 903
 904                 /* avoid atomic set bit if it's not needed: */
 905                 if (!btree_node_accessed(b))
 906                         set_btree_node_accessed(b);
 907         }
 908
 909         if (unlikely(btree_node_read_in_flight(b))) {
 910                 u32 seq = b->c.lock.state.seq;
 911
 912                 six_unlock_type(&b->c.lock, lock_type);
 913                 bch2_trans_unlock(trans);
 914
 915                 bch2_btree_node_wait_on_read(b);
 916
 917                 /*
 918                  * should_be_locked is not set on this path yet, so we need to
 919                  * relock it specifically:
 920                  */
 921                 if (trans) {
 922                         int ret = bch2_trans_relock(trans) ?:
 923                                 bch2_btree_path_relock_intent(trans, path);
 924                         if (ret) {
 925                                 BUG_ON(!trans->restarted);
 926                                 return ERR_PTR(ret);
 927                         }
 928                 }
 929
 930                 if (!six_relock_type(&b->c.lock, lock_type, seq))
 931                         goto retry;
 932         }
 933
 934         prefetch(b->aux_data);
 935
 936         for_each_bset(b, t) {
 937                 void *p = (u64 *) b->aux_data + t->aux_data_offset;
 938
 939                 prefetch(p + L1_CACHE_BYTES * 0);
 940                 prefetch(p + L1_CACHE_BYTES * 1);
 941                 prefetch(p + L1_CACHE_BYTES * 2);
 942         }
 943
 944         if (unlikely(btree_node_read_error(b))) {
 945                 six_unlock_type(&b->c.lock, lock_type);
 946                 return ERR_PTR(-EIO);
 947         }
 948
 949         EBUG_ON(b->c.btree_id != path->btree_id);
 950         EBUG_ON(BTREE_NODE_LEVEL(b->data) != level);
 951         btree_check_header(c, b);
 952
 953         return b;
 954 }
 955
 956 /**
 957  * bch_btree_node_get - find a btree node in the cache and lock it, reading it
 958  * in from disk if necessary.
 959  *
 960  * The btree node will have either a read or a write lock held, depending on
 961  * the @write parameter.
 962  */
 963 struct btree *bch2_btree_node_get(struct btree_trans *trans, struct btree_path *path,
 964                                   const struct bkey_i *k, unsigned level,
 965                                   enum six_lock_type lock_type,
 966                                   unsigned long trace_ip)
 967 {
 968         struct bch_fs *c = trans->c;
 969         struct btree *b;
 970         struct bset_tree *t;
 971         int ret;
 972
 973         EBUG_ON(level >= BTREE_MAX_DEPTH);
 974
 975         b = btree_node_mem_ptr(k);
 976
 977         /*
 978          * Check b->hash_val _before_ calling btree_node_lock() - this might not
 979          * be the node we want anymore, and trying to lock the wrong node could
 980          * cause an unneccessary transaction restart:
 981          */
 982         if (unlikely(!c->opts.btree_node_mem_ptr_optimization ||
 983                      !b ||
 984                      b->hash_val != btree_ptr_hash_val(k)))
 985                 return __bch2_btree_node_get(trans, path, k, level, lock_type, trace_ip);
 986
 987         if (btree_node_read_locked(path, level + 1))
 988                 btree_node_unlock(trans, path, level + 1);
 989
 990         ret = btree_node_lock(trans, path, &b->c, level, lock_type, trace_ip);
 991         if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
 992                 return ERR_PTR(ret);
 993
 994         BUG_ON(ret);
 995
 996         if (unlikely(b->hash_val != btree_ptr_hash_val(k) ||
 997                      b->c.level != level ||
 998                      race_fault())) {
 999                 six_unlock_type(&b->c.lock, lock_type);
1000                 if (bch2_btree_node_relock(trans, path, level + 1))
1001                         return __bch2_btree_node_get(trans, path, k, level, lock_type, trace_ip);
1002
1003                 trace_and_count(c, trans_restart_btree_node_reused, trans, trace_ip, path);
1004                 return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_lock_node_reused));
1005         }
1006
1007         if (unlikely(btree_node_read_in_flight(b))) {
1008                 u32 seq = b->c.lock.state.seq;
1009
1010                 six_unlock_type(&b->c.lock, lock_type);
1011                 bch2_trans_unlock(trans);
1012
1013                 bch2_btree_node_wait_on_read(b);
1014
1015                 /*
1016                  * should_be_locked is not set on this path yet, so we need to
1017                  * relock it specifically:
1018                  */
1019                 if (trans) {
1020                         int ret = bch2_trans_relock(trans) ?:
1021                                 bch2_btree_path_relock_intent(trans, path);
1022                         if (ret) {
1023                                 BUG_ON(!trans->restarted);
1024                                 return ERR_PTR(ret);
1025                         }
1026                 }
1027
1028                 if (!six_relock_type(&b->c.lock, lock_type, seq))
1029                         return __bch2_btree_node_get(trans, path, k, level, lock_type, trace_ip);
1030         }
1031
1032         prefetch(b->aux_data);
1033
1034         for_each_bset(b, t) {
1035                 void *p = (u64 *) b->aux_data + t->aux_data_offset;
1036
1037                 prefetch(p + L1_CACHE_BYTES * 0);
1038                 prefetch(p + L1_CACHE_BYTES * 1);
1039                 prefetch(p + L1_CACHE_BYTES * 2);
1040         }
1041
1042         /* avoid atomic set bit if it's not needed: */
1043         if (!btree_node_accessed(b))
1044                 set_btree_node_accessed(b);
1045
1046         if (unlikely(btree_node_read_error(b))) {
1047                 six_unlock_type(&b->c.lock, lock_type);
1048                 return ERR_PTR(-EIO);
1049         }
1050
1051         EBUG_ON(b->c.btree_id != path->btree_id);
1052         EBUG_ON(BTREE_NODE_LEVEL(b->data) != level);
1053         btree_check_header(c, b);
1054
1055         return b;
1056 }
1057
1058 struct btree *bch2_btree_node_get_noiter(struct btree_trans *trans,
1059                                          const struct bkey_i *k,
1060                                          enum btree_id btree_id,
1061                                          unsigned level,
1062                                          bool nofill)
1063 {
1064         struct bch_fs *c = trans->c;
1065         struct btree_cache *bc = &c->btree_cache;
1066         struct btree *b;
1067         struct bset_tree *t;
1068         int ret;
1069
1070         EBUG_ON(level >= BTREE_MAX_DEPTH);
1071
1072         if (c->opts.btree_node_mem_ptr_optimization) {
1073                 b = btree_node_mem_ptr(k);
1074                 if (b)
1075                         goto lock_node;
1076         }
1077 retry:
1078         b = btree_cache_find(bc, k);
1079         if (unlikely(!b)) {
1080                 if (nofill)
1081                         goto out;
1082
1083                 b = bch2_btree_node_fill(c, NULL, NULL, k, btree_id,
1084                                          level, SIX_LOCK_read, true);
1085
1086                 /* We raced and found the btree node in the cache */
1087                 if (!b)
1088                         goto retry;
1089
1090                 if (IS_ERR(b) &&
1091                     !bch2_btree_cache_cannibalize_lock(c, NULL))
1092                         goto retry;
1093
1094                 if (IS_ERR(b))
1095                         goto out;
1096         } else {
1097 lock_node:
1098                 ret = btree_node_lock_nopath(trans, &b->c, SIX_LOCK_read, _THIS_IP_);
1099                 if (bch2_err_matches(ret, BCH_ERR_transaction_restart))
1100                         return ERR_PTR(ret);
1101
1102                 BUG_ON(ret);
1103
1104                 if (unlikely(b->hash_val != btree_ptr_hash_val(k) ||
1105                              b->c.btree_id != btree_id ||
1106                              b->c.level != level)) {
1107                         six_unlock_read(&b->c.lock);
1108                         goto retry;
1109                 }
1110         }
1111
1112         /* XXX: waiting on IO with btree locks held: */
1113         __bch2_btree_node_wait_on_read(b);
1114
1115         prefetch(b->aux_data);
1116
1117         for_each_bset(b, t) {
1118                 void *p = (u64 *) b->aux_data + t->aux_data_offset;
1119
1120                 prefetch(p + L1_CACHE_BYTES * 0);
1121                 prefetch(p + L1_CACHE_BYTES * 1);
1122                 prefetch(p + L1_CACHE_BYTES * 2);
1123         }
1124
1125         /* avoid atomic set bit if it's not needed: */
1126         if (!btree_node_accessed(b))
1127                 set_btree_node_accessed(b);
1128
1129         if (unlikely(btree_node_read_error(b))) {
1130                 six_unlock_read(&b->c.lock);
1131                 b = ERR_PTR(-EIO);
1132                 goto out;
1133         }
1134
1135         EBUG_ON(b->c.btree_id != btree_id);
1136         EBUG_ON(BTREE_NODE_LEVEL(b->data) != level);
1137         btree_check_header(c, b);
1138 out:
1139         bch2_btree_cache_cannibalize_unlock(c);
1140         return b;
1141 }
1142
1143 int bch2_btree_node_prefetch(struct bch_fs *c,
1144                              struct btree_trans *trans,
1145                              struct btree_path *path,
1146                              const struct bkey_i *k,
1147                              enum btree_id btree_id, unsigned level)
1148 {
1149         struct btree_cache *bc = &c->btree_cache;
1150         struct btree *b;
1151
1152         BUG_ON(trans && !btree_node_locked(path, level + 1));
1153         BUG_ON(level >= BTREE_MAX_DEPTH);
1154
1155         b = btree_cache_find(bc, k);
1156         if (b)
1157                 return 0;
1158
1159         b = bch2_btree_node_fill(c, trans, path, k, btree_id,
1160                                  level, SIX_LOCK_read, false);
1161         return PTR_ERR_OR_ZERO(b);
1162 }
1163
1164 void bch2_btree_node_evict(struct btree_trans *trans, const struct bkey_i *k)
1165 {
1166         struct bch_fs *c = trans->c;
1167         struct btree_cache *bc = &c->btree_cache;
1168         struct btree *b;
1169
1170         b = btree_cache_find(bc, k);
1171         if (!b)
1172                 return;
1173 wait_on_io:
1174         /* not allowed to wait on io with btree locks held: */
1175
1176         /* XXX we're called from btree_gc which will be holding other btree
1177          * nodes locked
1178          */
1179         __bch2_btree_node_wait_on_read(b);
1180         __bch2_btree_node_wait_on_write(b);
1181
1182         btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_intent);
1183         btree_node_lock_nopath_nofail(trans, &b->c, SIX_LOCK_write);
1184
1185         if (btree_node_dirty(b)) {
1186                 __bch2_btree_node_write(c, b, BTREE_WRITE_cache_reclaim);
1187                 six_unlock_write(&b->c.lock);
1188                 six_unlock_intent(&b->c.lock);
1189                 goto wait_on_io;
1190         }
1191
1192         BUG_ON(btree_node_dirty(b));
1193
1194         mutex_lock(&bc->lock);
1195         btree_node_data_free(c, b);
1196         bch2_btree_node_hash_remove(bc, b);
1197         mutex_unlock(&bc->lock);
1198
1199         six_unlock_write(&b->c.lock);
1200         six_unlock_intent(&b->c.lock);
1201 }
1202
1203 void bch2_btree_node_to_text(struct printbuf *out, struct bch_fs *c,
1204                              struct btree *b)
1205 {
1206         const struct bkey_format *f = &b->format;
1207         struct bset_stats stats;
1208
1209         memset(&stats, 0, sizeof(stats));
1210
1211         bch2_btree_keys_stats(b, &stats);
1212
1213         prt_printf(out, "l %u ", b->c.level);
1214         bch2_bpos_to_text(out, b->data->min_key);
1215         prt_printf(out, " - ");
1216         bch2_bpos_to_text(out, b->data->max_key);
1217         prt_printf(out, ":\n"
1218                "    ptrs: ");
1219         bch2_val_to_text(out, c, bkey_i_to_s_c(&b->key));
1220
1221         prt_printf(out, "\n"
1222                "    format: u64s %u fields %u %u %u %u %u\n"
1223                "    unpack fn len: %u\n"
1224                "    bytes used %zu/%zu (%zu%% full)\n"
1225                "    sib u64s: %u, %u (merge threshold %u)\n"
1226                "    nr packed keys %u\n"
1227                "    nr unpacked keys %u\n"
1228                "    floats %zu\n"
1229                "    failed unpacked %zu\n",
1230                f->key_u64s,
1231                f->bits_per_field[0],
1232                f->bits_per_field[1],
1233                f->bits_per_field[2],
1234                f->bits_per_field[3],
1235                f->bits_per_field[4],
1236                b->unpack_fn_len,
1237                b->nr.live_u64s * sizeof(u64),
1238                btree_bytes(c) - sizeof(struct btree_node),
1239                b->nr.live_u64s * 100 / btree_max_u64s(c),
1240                b->sib_u64s[0],
1241                b->sib_u64s[1],
1242                c->btree_foreground_merge_threshold,
1243                b->nr.packed_keys,
1244                b->nr.unpacked_keys,
1245                stats.floats,
1246                stats.failed);
1247 }
1248
1249 void bch2_btree_cache_to_text(struct printbuf *out, struct btree_cache *bc)
1250 {
1251         prt_printf(out, "nr nodes:\t\t%u\n", bc->used);
1252         prt_printf(out, "nr dirty:\t\t%u\n", atomic_read(&bc->dirty));
1253         prt_printf(out, "cannibalize lock:\t%p\n", bc->alloc_lock);
1254
1255         prt_printf(out, "freed:\t\t\t\t%u\n", bc->freed);
1256         prt_printf(out, "not freed, dirty:\t\t%u\n", bc->not_freed_dirty);
1257         prt_printf(out, "not freed, write in flight:\t%u\n", bc->not_freed_write_in_flight);
1258         prt_printf(out, "not freed, read in flight:\t%u\n", bc->not_freed_read_in_flight);
1259         prt_printf(out, "not freed, lock intent failed:\t%u\n", bc->not_freed_lock_intent);
1260         prt_printf(out, "not freed, lock write failed:\t%u\n", bc->not_freed_lock_write);
1261         prt_printf(out, "not freed, access bit:\t\t%u\n", bc->not_freed_access_bit);
1262         prt_printf(out, "not freed, no evict failed:\t%u\n", bc->not_freed_noevict);
1263         prt_printf(out, "not freed, write blocked:\t%u\n", bc->not_freed_write_blocked);
1264         prt_printf(out, "not freed, will make reachable:\t%u\n", bc->not_freed_will_make_reachable);
1265
1266 }