drivers/md/bcache/btree.c

   1 /*
   2  * Copyright (C) 2010 Kent Overstreet <kent.overstreet@gmail.com>
   3  *
   4  * Uses a block device as cache for other block devices; optimized for SSDs.
   5  * All allocation is done in buckets, which should match the erase block size
   6  * of the device.
   7  *
   8  * Buckets containing cached data are kept on a heap sorted by priority;
   9  * bucket priority is increased on cache hit, and periodically all the buckets
  10  * on the heap have their priority scaled down. This currently is just used as
  11  * an LRU but in the future should allow for more intelligent heuristics.
  12  *
  13  * Buckets have an 8 bit counter; freeing is accomplished by incrementing the
  14  * counter. Garbage collection is used to remove stale pointers.
  15  *
  16  * Indexing is done via a btree; nodes are not necessarily fully sorted, rather
  17  * as keys are inserted we only sort the pages that have not yet been written.
  18  * When garbage collection is run, we resort the entire node.
  19  *
  20  * All configuration is done via sysfs; see Documentation/bcache.txt.
  21  */
  22
  23 #include "bcache.h"
  24 #include "btree.h"
  25 #include "debug.h"
  26 #include "extents.h"
  27
  28 #include <linux/slab.h>
  29 #include <linux/bitops.h>
  30 #include <linux/freezer.h>
  31 #include <linux/hash.h>
  32 #include <linux/kthread.h>
  33 #include <linux/prefetch.h>
  34 #include <linux/random.h>
  35 #include <linux/rcupdate.h>
  36 #include <trace/events/bcache.h>
  37
  38 /*
  39  * Todo:
  40  * register_bcache: Return errors out to userspace correctly
  41  *
  42  * Writeback: don't undirty key until after a cache flush
  43  *
  44  * Create an iterator for key pointers
  45  *
  46  * On btree write error, mark bucket such that it won't be freed from the cache
  47  *
  48  * Journalling:
  49  *   Check for bad keys in replay
  50  *   Propagate barriers
  51  *   Refcount journal entries in journal_replay
  52  *
  53  * Garbage collection:
  54  *   Finish incremental gc
  55  *   Gc should free old UUIDs, data for invalid UUIDs
  56  *
  57  * Provide a way to list backing device UUIDs we have data cached for, and
  58  * probably how long it's been since we've seen them, and a way to invalidate
  59  * dirty data for devices that will never be attached again
  60  *
  61  * Keep 1 min/5 min/15 min statistics of how busy a block device has been, so
  62  * that based on that and how much dirty data we have we can keep writeback
  63  * from being starved
  64  *
  65  * Add a tracepoint or somesuch to watch for writeback starvation
  66  *
  67  * When btree depth > 1 and splitting an interior node, we have to make sure
  68  * alloc_bucket() cannot fail. This should be true but is not completely
  69  * obvious.
  70  *
  71  * Plugging?
  72  *
  73  * If data write is less than hard sector size of ssd, round up offset in open
  74  * bucket to the next whole sector
  75  *
  76  * Superblock needs to be fleshed out for multiple cache devices
  77  *
  78  * Add a sysfs tunable for the number of writeback IOs in flight
  79  *
  80  * Add a sysfs tunable for the number of open data buckets
  81  *
  82  * IO tracking: Can we track when one process is doing io on behalf of another?
  83  * IO tracking: Don't use just an average, weigh more recent stuff higher
  84  *
  85  * Test module load/unload
  86  */
  87
  88 #define MAX_NEED_GC             64
  89 #define MAX_SAVE_PRIO           72
  90
  91 #define PTR_DIRTY_BIT           (((uint64_t) 1 << 36))
  92
  93 #define PTR_HASH(c, k)                                                  \
  94         (((k)->ptr[0] >> c->bucket_bits) | PTR_GEN(k, 0))
  95
  96 #define insert_lock(s, b)       ((b)->level <= (s)->lock)
  97
  98 /*
  99  * These macros are for recursing down the btree - they handle the details of
 100  * locking and looking up nodes in the cache for you. They're best treated as
 101  * mere syntax when reading code that uses them.
 102  *
 103  * op->lock determines whether we take a read or a write lock at a given depth.
 104  * If you've got a read lock and find that you need a write lock (i.e. you're
 105  * going to have to split), set op->lock and return -EINTR; btree_root() will
 106  * call you again and you'll have the correct lock.
 107  */
 108
 109 /**
 110  * btree - recurse down the btree on a specified key
 111  * @fn:         function to call, which will be passed the child node
 112  * @key:        key to recurse on
 113  * @b:          parent btree node
 114  * @op:         pointer to struct btree_op
 115  */
 116 #define btree(fn, key, b, op, ...)                                      \
 117 ({                                                                      \
 118         int _r, l = (b)->level - 1;                                     \
 119         bool _w = l <= (op)->lock;                                      \
 120         struct btree *_child = bch_btree_node_get((b)->c, op, key, l, _w);\
 121         if (!IS_ERR(_child)) {                                          \
 122                 _child->parent = (b);                                   \
 123                 _r = bch_btree_ ## fn(_child, op, ##__VA_ARGS__);       \
 124                 rw_unlock(_w, _child);                                  \
 125         } else                                                          \
 126                 _r = PTR_ERR(_child);                                   \
 127         _r;                                                             \
 128 })
 129
 130 /**
 131  * btree_root - call a function on the root of the btree
 132  * @fn:         function to call, which will be passed the child node
 133  * @c:          cache set
 134  * @op:         pointer to struct btree_op
 135  */
 136 #define btree_root(fn, c, op, ...)                                      \
 137 ({                                                                      \
 138         int _r = -EINTR;                                                \
 139         do {                                                            \
 140                 struct btree *_b = (c)->root;                           \
 141                 bool _w = insert_lock(op, _b);                          \
 142                 rw_lock(_w, _b, _b->level);                             \
 143                 if (_b == (c)->root &&                                  \
 144                     _w == insert_lock(op, _b)) {                        \
 145                         _b->parent = NULL;                              \
 146                         _r = bch_btree_ ## fn(_b, op, ##__VA_ARGS__);   \
 147                 }                                                       \
 148                 rw_unlock(_w, _b);                                      \
 149                 bch_cannibalize_unlock(c);                              \
 150                 if (_r == -EINTR)                                       \
 151                         schedule();                                     \
 152         } while (_r == -EINTR);                                         \
 153                                                                         \
 154         finish_wait(&(c)->btree_cache_wait, &(op)->wait);               \
 155         _r;                                                             \
 156 })
 157
 158 static inline struct bset *write_block(struct btree *b)
 159 {
 160         return ((void *) btree_bset_first(b)) + b->written * block_bytes(b->c);
 161 }
 162
 163 static void bch_btree_init_next(struct btree *b)
 164 {
 165         /* If not a leaf node, always sort */
 166         if (b->level && b->keys.nsets)
 167                 bch_btree_sort(&b->keys, &b->c->sort);
 168         else
 169                 bch_btree_sort_lazy(&b->keys, &b->c->sort);
 170
 171         if (b->written < btree_blocks(b))
 172                 bch_bset_init_next(&b->keys, write_block(b),
 173                                    bset_magic(&b->c->sb));
 174
 175 }
 176
 177 /* Btree key manipulation */
 178
 179 void bkey_put(struct cache_set *c, struct bkey *k)
 180 {
 181         unsigned i;
 182
 183         for (i = 0; i < KEY_PTRS(k); i++)
 184                 if (ptr_available(c, k, i))
 185                         atomic_dec_bug(&PTR_BUCKET(c, k, i)->pin);
 186 }
 187
 188 /* Btree IO */
 189
 190 static uint64_t btree_csum_set(struct btree *b, struct bset *i)
 191 {
 192         uint64_t crc = b->key.ptr[0];
 193         void *data = (void *) i + 8, *end = bset_bkey_last(i);
 194
 195         crc = bch_crc64_update(crc, data, end - data);
 196         return crc ^ 0xffffffffffffffffULL;
 197 }
 198
 199 void bch_btree_node_read_done(struct btree *b)
 200 {
 201         const char *err = "bad btree header";
 202         struct bset *i = btree_bset_first(b);
 203         struct btree_iter *iter;
 204
 205         iter = mempool_alloc(b->c->fill_iter, GFP_NOWAIT);
 206         iter->size = b->c->sb.bucket_size / b->c->sb.block_size;
 207         iter->used = 0;
 208
 209 #ifdef CONFIG_BCACHE_DEBUG
 210         iter->b = &b->keys;
 211 #endif
 212
 213         if (!i->seq)
 214                 goto err;
 215
 216         for (;
 217              b->written < btree_blocks(b) && i->seq == b->keys.set[0].data->seq;
 218              i = write_block(b)) {
 219                 err = "unsupported bset version";
 220                 if (i->version > BCACHE_BSET_VERSION)
 221                         goto err;
 222
 223                 err = "bad btree header";
 224                 if (b->written + set_blocks(i, block_bytes(b->c)) >
 225                     btree_blocks(b))
 226                         goto err;
 227
 228                 err = "bad magic";
 229                 if (i->magic != bset_magic(&b->c->sb))
 230                         goto err;
 231
 232                 err = "bad checksum";
 233                 switch (i->version) {
 234                 case 0:
 235                         if (i->csum != csum_set(i))
 236                                 goto err;
 237                         break;
 238                 case BCACHE_BSET_VERSION:
 239                         if (i->csum != btree_csum_set(b, i))
 240                                 goto err;
 241                         break;
 242                 }
 243
 244                 err = "empty set";
 245                 if (i != b->keys.set[0].data && !i->keys)
 246                         goto err;
 247
 248                 bch_btree_iter_push(iter, i->start, bset_bkey_last(i));
 249
 250                 b->written += set_blocks(i, block_bytes(b->c));
 251         }
 252
 253         err = "corrupted btree";
 254         for (i = write_block(b);
 255              bset_sector_offset(&b->keys, i) < KEY_SIZE(&b->key);
 256              i = ((void *) i) + block_bytes(b->c))
 257                 if (i->seq == b->keys.set[0].data->seq)
 258                         goto err;
 259
 260         bch_btree_sort_and_fix_extents(&b->keys, iter, &b->c->sort);
 261
 262         i = b->keys.set[0].data;
 263         err = "short btree key";
 264         if (b->keys.set[0].size &&
 265             bkey_cmp(&b->key, &b->keys.set[0].end) < 0)
 266                 goto err;
 267
 268         if (b->written < btree_blocks(b))
 269                 bch_bset_init_next(&b->keys, write_block(b),
 270                                    bset_magic(&b->c->sb));
 271 out:
 272         mempool_free(iter, b->c->fill_iter);
 273         return;
 274 err:
 275         set_btree_node_io_error(b);
 276         bch_cache_set_error(b->c, "%s at bucket %zu, block %u, %u keys",
 277                             err, PTR_BUCKET_NR(b->c, &b->key, 0),
 278                             bset_block_offset(b, i), i->keys);
 279         goto out;
 280 }
 281
 282 static void btree_node_read_endio(struct bio *bio, int error)
 283 {
 284         struct closure *cl = bio->bi_private;
 285         closure_put(cl);
 286 }
 287
 288 static void bch_btree_node_read(struct btree *b)
 289 {
 290         uint64_t start_time = local_clock();
 291         struct closure cl;
 292         struct bio *bio;
 293
 294         trace_bcache_btree_read(b);
 295
 296         closure_init_stack(&cl);
 297
 298         bio = bch_bbio_alloc(b->c);
 299         bio->bi_rw      = REQ_META|READ_SYNC;
 300         bio->bi_iter.bi_size = KEY_SIZE(&b->key) << 9;
 301         bio->bi_end_io  = btree_node_read_endio;
 302         bio->bi_private = &cl;
 303
 304         bch_bio_map(bio, b->keys.set[0].data);
 305
 306         bch_submit_bbio(bio, b->c, &b->key, 0);
 307         closure_sync(&cl);
 308
 309         if (!test_bit(BIO_UPTODATE, &bio->bi_flags))
 310                 set_btree_node_io_error(b);
 311
 312         bch_bbio_free(bio, b->c);
 313
 314         if (btree_node_io_error(b))
 315                 goto err;
 316
 317         bch_btree_node_read_done(b);
 318         bch_time_stats_update(&b->c->btree_read_time, start_time);
 319
 320         return;
 321 err:
 322         bch_cache_set_error(b->c, "io error reading bucket %zu",
 323                             PTR_BUCKET_NR(b->c, &b->key, 0));
 324 }
 325
 326 static void btree_complete_write(struct btree *b, struct btree_write *w)
 327 {
 328         if (w->prio_blocked &&
 329             !atomic_sub_return(w->prio_blocked, &b->c->prio_blocked))
 330                 wake_up_allocators(b->c);
 331
 332         if (w->journal) {
 333                 atomic_dec_bug(w->journal);
 334                 __closure_wake_up(&b->c->journal.wait);
 335         }
 336
 337         w->prio_blocked = 0;
 338         w->journal      = NULL;
 339 }
 340
 341 static void btree_node_write_unlock(struct closure *cl)
 342 {
 343         struct btree *b = container_of(cl, struct btree, io);
 344
 345         up(&b->io_mutex);
 346 }
 347
 348 static void __btree_node_write_done(struct closure *cl)
 349 {
 350         struct btree *b = container_of(cl, struct btree, io);
 351         struct btree_write *w = btree_prev_write(b);
 352
 353         bch_bbio_free(b->bio, b->c);
 354         b->bio = NULL;
 355         btree_complete_write(b, w);
 356
 357         if (btree_node_dirty(b))
 358                 schedule_delayed_work(&b->work, 30 * HZ);
 359
 360         closure_return_with_destructor(cl, btree_node_write_unlock);
 361 }
 362
 363 static void btree_node_write_done(struct closure *cl)
 364 {
 365         struct btree *b = container_of(cl, struct btree, io);
 366         struct bio_vec *bv;
 367         int n;
 368
 369         bio_for_each_segment_all(bv, b->bio, n)
 370                 __free_page(bv->bv_page);
 371
 372         __btree_node_write_done(cl);
 373 }
 374
 375 static void btree_node_write_endio(struct bio *bio, int error)
 376 {
 377         struct closure *cl = bio->bi_private;
 378         struct btree *b = container_of(cl, struct btree, io);
 379
 380         if (error)
 381                 set_btree_node_io_error(b);
 382
 383         bch_bbio_count_io_errors(b->c, bio, error, "writing btree");
 384         closure_put(cl);
 385 }
 386
 387 static void do_btree_node_write(struct btree *b)
 388 {
 389         struct closure *cl = &b->io;
 390         struct bset *i = btree_bset_last(b);
 391         BKEY_PADDED(key) k;
 392
 393         i->version      = BCACHE_BSET_VERSION;
 394         i->csum         = btree_csum_set(b, i);
 395
 396         BUG_ON(b->bio);
 397         b->bio = bch_bbio_alloc(b->c);
 398
 399         b->bio->bi_end_io       = btree_node_write_endio;
 400         b->bio->bi_private      = cl;
 401         b->bio->bi_rw           = REQ_META|WRITE_SYNC|REQ_FUA;
 402         b->bio->bi_iter.bi_size = roundup(set_bytes(i), block_bytes(b->c));
 403         bch_bio_map(b->bio, i);
 404
 405         /*
 406          * If we're appending to a leaf node, we don't technically need FUA -
 407          * this write just needs to be persisted before the next journal write,
 408          * which will be marked FLUSH|FUA.
 409          *
 410          * Similarly if we're writing a new btree root - the pointer is going to
 411          * be in the next journal entry.
 412          *
 413          * But if we're writing a new btree node (that isn't a root) or
 414          * appending to a non leaf btree node, we need either FUA or a flush
 415          * when we write the parent with the new pointer. FUA is cheaper than a
 416          * flush, and writes appending to leaf nodes aren't blocking anything so
 417          * just make all btree node writes FUA to keep things sane.
 418          */
 419
 420         bkey_copy(&k.key, &b->key);
 421         SET_PTR_OFFSET(&k.key, 0, PTR_OFFSET(&k.key, 0) +
 422                        bset_sector_offset(&b->keys, i));
 423
 424         if (!bio_alloc_pages(b->bio, GFP_NOIO)) {
 425                 int j;
 426                 struct bio_vec *bv;
 427                 void *base = (void *) ((unsigned long) i & ~(PAGE_SIZE - 1));
 428
 429                 bio_for_each_segment_all(bv, b->bio, j)
 430                         memcpy(page_address(bv->bv_page),
 431                                base + j * PAGE_SIZE, PAGE_SIZE);
 432
 433                 bch_submit_bbio(b->bio, b->c, &k.key, 0);
 434
 435                 continue_at(cl, btree_node_write_done, NULL);
 436         } else {
 437                 b->bio->bi_vcnt = 0;
 438                 bch_bio_map(b->bio, i);
 439
 440                 bch_submit_bbio(b->bio, b->c, &k.key, 0);
 441
 442                 closure_sync(cl);
 443                 continue_at_nobarrier(cl, __btree_node_write_done, NULL);
 444         }
 445 }
 446
 447 void __bch_btree_node_write(struct btree *b, struct closure *parent)
 448 {
 449         struct bset *i = btree_bset_last(b);
 450
 451         lockdep_assert_held(&b->write_lock);
 452
 453         trace_bcache_btree_write(b);
 454
 455         BUG_ON(current->bio_list);
 456         BUG_ON(b->written >= btree_blocks(b));
 457         BUG_ON(b->written && !i->keys);
 458         BUG_ON(btree_bset_first(b)->seq != i->seq);
 459         bch_check_keys(&b->keys, "writing");
 460
 461         cancel_delayed_work(&b->work);
 462
 463         /* If caller isn't waiting for write, parent refcount is cache set */
 464         down(&b->io_mutex);
 465         closure_init(&b->io, parent ?: &b->c->cl);
 466
 467         clear_bit(BTREE_NODE_dirty,      &b->flags);
 468         change_bit(BTREE_NODE_write_idx, &b->flags);
 469
 470         do_btree_node_write(b);
 471
 472         atomic_long_add(set_blocks(i, block_bytes(b->c)) * b->c->sb.block_size,
 473                         &PTR_CACHE(b->c, &b->key, 0)->btree_sectors_written);
 474
 475         b->written += set_blocks(i, block_bytes(b->c));
 476 }
 477
 478 void bch_btree_node_write(struct btree *b, struct closure *parent)
 479 {
 480         unsigned nsets = b->keys.nsets;
 481
 482         lockdep_assert_held(&b->lock);
 483
 484         __bch_btree_node_write(b, parent);
 485
 486         /*
 487          * do verify if there was more than one set initially (i.e. we did a
 488          * sort) and we sorted down to a single set:
 489          */
 490         if (nsets && !b->keys.nsets)
 491                 bch_btree_verify(b);
 492
 493         bch_btree_init_next(b);
 494 }
 495
 496 static void bch_btree_node_write_sync(struct btree *b)
 497 {
 498         struct closure cl;
 499
 500         closure_init_stack(&cl);
 501
 502         mutex_lock(&b->write_lock);
 503         bch_btree_node_write(b, &cl);
 504         mutex_unlock(&b->write_lock);
 505
 506         closure_sync(&cl);
 507 }
 508
 509 static void btree_node_write_work(struct work_struct *w)
 510 {
 511         struct btree *b = container_of(to_delayed_work(w), struct btree, work);
 512
 513         mutex_lock(&b->write_lock);
 514         if (btree_node_dirty(b))
 515                 __bch_btree_node_write(b, NULL);
 516         mutex_unlock(&b->write_lock);
 517 }
 518
 519 static void bch_btree_leaf_dirty(struct btree *b, atomic_t *journal_ref)
 520 {
 521         struct bset *i = btree_bset_last(b);
 522         struct btree_write *w = btree_current_write(b);
 523
 524         lockdep_assert_held(&b->write_lock);
 525
 526         BUG_ON(!b->written);
 527         BUG_ON(!i->keys);
 528
 529         if (!btree_node_dirty(b))
 530                 schedule_delayed_work(&b->work, 30 * HZ);
 531
 532         set_btree_node_dirty(b);
 533
 534         if (journal_ref) {
 535                 if (w->journal &&
 536                     journal_pin_cmp(b->c, w->journal, journal_ref)) {
 537                         atomic_dec_bug(w->journal);
 538                         w->journal = NULL;
 539                 }
 540
 541                 if (!w->journal) {
 542                         w->journal = journal_ref;
 543                         atomic_inc(w->journal);
 544                 }
 545         }
 546
 547         /* Force write if set is too big */
 548         if (set_bytes(i) > PAGE_SIZE - 48 &&
 549             !current->bio_list)
 550                 bch_btree_node_write(b, NULL);
 551 }
 552
 553 /*
 554  * Btree in memory cache - allocation/freeing
 555  * mca -> memory cache
 556  */
 557
 558 #define mca_reserve(c)  (((c->root && c->root->level)           \
 559                           ? c->root->level : 1) * 8 + 16)
 560 #define mca_can_free(c)                                         \
 561         max_t(int, 0, c->btree_cache_used - mca_reserve(c))
 562
 563 static void mca_data_free(struct btree *b)
 564 {
 565         BUG_ON(b->io_mutex.count != 1);
 566
 567         bch_btree_keys_free(&b->keys);
 568
 569         b->c->btree_cache_used--;
 570         list_move(&b->list, &b->c->btree_cache_freed);
 571 }
 572
 573 static void mca_bucket_free(struct btree *b)
 574 {
 575         BUG_ON(btree_node_dirty(b));
 576
 577         b->key.ptr[0] = 0;
 578         hlist_del_init_rcu(&b->hash);
 579         list_move(&b->list, &b->c->btree_cache_freeable);
 580 }
 581
 582 static unsigned btree_order(struct bkey *k)
 583 {
 584         return ilog2(KEY_SIZE(k) / PAGE_SECTORS ?: 1);
 585 }
 586
 587 static void mca_data_alloc(struct btree *b, struct bkey *k, gfp_t gfp)
 588 {
 589         if (!bch_btree_keys_alloc(&b->keys,
 590                                   max_t(unsigned,
 591                                         ilog2(b->c->btree_pages),
 592                                         btree_order(k)),
 593                                   gfp)) {
 594                 b->c->btree_cache_used++;
 595                 list_move(&b->list, &b->c->btree_cache);
 596         } else {
 597                 list_move(&b->list, &b->c->btree_cache_freed);
 598         }
 599 }
 600
 601 static struct btree *mca_bucket_alloc(struct cache_set *c,
 602                                       struct bkey *k, gfp_t gfp)
 603 {
 604         struct btree *b = kzalloc(sizeof(struct btree), gfp);
 605         if (!b)
 606                 return NULL;
 607
 608         init_rwsem(&b->lock);
 609         lockdep_set_novalidate_class(&b->lock);
 610         mutex_init(&b->write_lock);
 611         lockdep_set_novalidate_class(&b->write_lock);
 612         INIT_LIST_HEAD(&b->list);
 613         INIT_DELAYED_WORK(&b->work, btree_node_write_work);
 614         b->c = c;
 615         sema_init(&b->io_mutex, 1);
 616
 617         mca_data_alloc(b, k, gfp);
 618         return b;
 619 }
 620
 621 static int mca_reap(struct btree *b, unsigned min_order, bool flush)
 622 {
 623         struct closure cl;
 624
 625         closure_init_stack(&cl);
 626         lockdep_assert_held(&b->c->bucket_lock);
 627
 628         if (!down_write_trylock(&b->lock))
 629                 return -ENOMEM;
 630
 631         BUG_ON(btree_node_dirty(b) && !b->keys.set[0].data);
 632
 633         if (b->keys.page_order < min_order)
 634                 goto out_unlock;
 635
 636         if (!flush) {
 637                 if (btree_node_dirty(b))
 638                         goto out_unlock;
 639
 640                 if (down_trylock(&b->io_mutex))
 641                         goto out_unlock;
 642                 up(&b->io_mutex);
 643         }
 644
 645         mutex_lock(&b->write_lock);
 646         if (btree_node_dirty(b))
 647                 __bch_btree_node_write(b, &cl);
 648         mutex_unlock(&b->write_lock);
 649
 650         closure_sync(&cl);
 651
 652         /* wait for any in flight btree write */
 653         down(&b->io_mutex);
 654         up(&b->io_mutex);
 655
 656         return 0;
 657 out_unlock:
 658         rw_unlock(true, b);
 659         return -ENOMEM;
 660 }
 661
 662 static unsigned long bch_mca_scan(struct shrinker *shrink,
 663                                   struct shrink_control *sc)
 664 {
 665         struct cache_set *c = container_of(shrink, struct cache_set, shrink);
 666         struct btree *b, *t;
 667         unsigned long i, nr = sc->nr_to_scan;
 668         unsigned long freed = 0;
 669
 670         if (c->shrinker_disabled)
 671                 return SHRINK_STOP;
 672
 673         if (c->btree_cache_alloc_lock)
 674                 return SHRINK_STOP;
 675
 676         /* Return -1 if we can't do anything right now */
 677         if (sc->gfp_mask & __GFP_IO)
 678                 mutex_lock(&c->bucket_lock);
 679         else if (!mutex_trylock(&c->bucket_lock))
 680                 return -1;
 681
 682         /*
 683          * It's _really_ critical that we don't free too many btree nodes - we
 684          * have to always leave ourselves a reserve. The reserve is how we
 685          * guarantee that allocating memory for a new btree node can always
 686          * succeed, so that inserting keys into the btree can always succeed and
 687          * IO can always make forward progress:
 688          */
 689         nr /= c->btree_pages;
 690         nr = min_t(unsigned long, nr, mca_can_free(c));
 691
 692         i = 0;
 693         list_for_each_entry_safe(b, t, &c->btree_cache_freeable, list) {
 694                 if (freed >= nr)
 695                         break;
 696
 697                 if (++i > 3 &&
 698                     !mca_reap(b, 0, false)) {
 699                         mca_data_free(b);
 700                         rw_unlock(true, b);
 701                         freed++;
 702                 }
 703         }
 704
 705         for (i = 0; (nr--) && i < c->btree_cache_used; i++) {
 706                 if (list_empty(&c->btree_cache))
 707                         goto out;
 708
 709                 b = list_first_entry(&c->btree_cache, struct btree, list);
 710                 list_rotate_left(&c->btree_cache);
 711
 712                 if (!b->accessed &&
 713                     !mca_reap(b, 0, false)) {
 714                         mca_bucket_free(b);
 715                         mca_data_free(b);
 716                         rw_unlock(true, b);
 717                         freed++;
 718                 } else
 719                         b->accessed = 0;
 720         }
 721 out:
 722         mutex_unlock(&c->bucket_lock);
 723         return freed;
 724 }
 725
 726 static unsigned long bch_mca_count(struct shrinker *shrink,
 727                                    struct shrink_control *sc)
 728 {
 729         struct cache_set *c = container_of(shrink, struct cache_set, shrink);
 730
 731         if (c->shrinker_disabled)
 732                 return 0;
 733
 734         if (c->btree_cache_alloc_lock)
 735                 return 0;
 736
 737         return mca_can_free(c) * c->btree_pages;
 738 }
 739
 740 void bch_btree_cache_free(struct cache_set *c)
 741 {
 742         struct btree *b;
 743         struct closure cl;
 744         closure_init_stack(&cl);
 745
 746         if (c->shrink.list.next)
 747                 unregister_shrinker(&c->shrink);
 748
 749         mutex_lock(&c->bucket_lock);
 750
 751 #ifdef CONFIG_BCACHE_DEBUG
 752         if (c->verify_data)
 753                 list_move(&c->verify_data->list, &c->btree_cache);
 754
 755         free_pages((unsigned long) c->verify_ondisk, ilog2(bucket_pages(c)));
 756 #endif
 757
 758         list_splice(&c->btree_cache_freeable,
 759                     &c->btree_cache);
 760
 761         while (!list_empty(&c->btree_cache)) {
 762                 b = list_first_entry(&c->btree_cache, struct btree, list);
 763
 764                 if (btree_node_dirty(b))
 765                         btree_complete_write(b, btree_current_write(b));
 766                 clear_bit(BTREE_NODE_dirty, &b->flags);
 767
 768                 mca_data_free(b);
 769         }
 770
 771         while (!list_empty(&c->btree_cache_freed)) {
 772                 b = list_first_entry(&c->btree_cache_freed,
 773                                      struct btree, list);
 774                 list_del(&b->list);
 775                 cancel_delayed_work_sync(&b->work);
 776                 kfree(b);
 777         }
 778
 779         mutex_unlock(&c->bucket_lock);
 780 }
 781
 782 int bch_btree_cache_alloc(struct cache_set *c)
 783 {
 784         unsigned i;
 785
 786         for (i = 0; i < mca_reserve(c); i++)
 787                 if (!mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL))
 788                         return -ENOMEM;
 789
 790         list_splice_init(&c->btree_cache,
 791                          &c->btree_cache_freeable);
 792
 793 #ifdef CONFIG_BCACHE_DEBUG
 794         mutex_init(&c->verify_lock);
 795
 796         c->verify_ondisk = (void *)
 797                 __get_free_pages(GFP_KERNEL, ilog2(bucket_pages(c)));
 798
 799         c->verify_data = mca_bucket_alloc(c, &ZERO_KEY, GFP_KERNEL);
 800
 801         if (c->verify_data &&
 802             c->verify_data->keys.set->data)
 803                 list_del_init(&c->verify_data->list);
 804         else
 805                 c->verify_data = NULL;
 806 #endif
 807
 808         c->shrink.count_objects = bch_mca_count;
 809         c->shrink.scan_objects = bch_mca_scan;
 810         c->shrink.seeks = 4;
 811         c->shrink.batch = c->btree_pages * 2;
 812         register_shrinker(&c->shrink);
 813
 814         return 0;
 815 }
 816
 817 /* Btree in memory cache - hash table */
 818
 819 static struct hlist_head *mca_hash(struct cache_set *c, struct bkey *k)
 820 {
 821         return &c->bucket_hash[hash_32(PTR_HASH(c, k), BUCKET_HASH_BITS)];
 822 }
 823
 824 static struct btree *mca_find(struct cache_set *c, struct bkey *k)
 825 {
 826         struct btree *b;
 827
 828         rcu_read_lock();
 829         hlist_for_each_entry_rcu(b, mca_hash(c, k), hash)
 830                 if (PTR_HASH(c, &b->key) == PTR_HASH(c, k))
 831                         goto out;
 832         b = NULL;
 833 out:
 834         rcu_read_unlock();
 835         return b;
 836 }
 837
 838 static int mca_cannibalize_lock(struct cache_set *c, struct btree_op *op)
 839 {
 840         struct task_struct *old;
 841
 842         old = cmpxchg(&c->btree_cache_alloc_lock, NULL, current);
 843         if (old && old != current) {
 844                 if (op)
 845                         prepare_to_wait(&c->btree_cache_wait, &op->wait,
 846                                         TASK_UNINTERRUPTIBLE);
 847                 return -EINTR;
 848         }
 849
 850         return 0;
 851 }
 852
 853 static struct btree *mca_cannibalize(struct cache_set *c, struct btree_op *op,
 854                                      struct bkey *k)
 855 {
 856         struct btree *b;
 857
 858         trace_bcache_btree_cache_cannibalize(c);
 859
 860         if (mca_cannibalize_lock(c, op))
 861                 return ERR_PTR(-EINTR);
 862
 863         list_for_each_entry_reverse(b, &c->btree_cache, list)
 864                 if (!mca_reap(b, btree_order(k), false))
 865                         return b;
 866
 867         list_for_each_entry_reverse(b, &c->btree_cache, list)
 868                 if (!mca_reap(b, btree_order(k), true))
 869                         return b;
 870
 871         WARN(1, "btree cache cannibalize failed\n");
 872         return ERR_PTR(-ENOMEM);
 873 }
 874
 875 /*
 876  * We can only have one thread cannibalizing other cached btree nodes at a time,
 877  * or we'll deadlock. We use an open coded mutex to ensure that, which a
 878  * cannibalize_bucket() will take. This means every time we unlock the root of
 879  * the btree, we need to release this lock if we have it held.
 880  */
 881 static void bch_cannibalize_unlock(struct cache_set *c)
 882 {
 883         if (c->btree_cache_alloc_lock == current) {
 884                 c->btree_cache_alloc_lock = NULL;
 885                 wake_up(&c->btree_cache_wait);
 886         }
 887 }
 888
 889 static struct btree *mca_alloc(struct cache_set *c, struct btree_op *op,
 890                                struct bkey *k, int level)
 891 {
 892         struct btree *b;
 893
 894         BUG_ON(current->bio_list);
 895
 896         lockdep_assert_held(&c->bucket_lock);
 897
 898         if (mca_find(c, k))
 899                 return NULL;
 900
 901         /* btree_free() doesn't free memory; it sticks the node on the end of
 902          * the list. Check if there's any freed nodes there:
 903          */
 904         list_for_each_entry(b, &c->btree_cache_freeable, list)
 905                 if (!mca_reap(b, btree_order(k), false))
 906                         goto out;
 907
 908         /* We never free struct btree itself, just the memory that holds the on
 909          * disk node. Check the freed list before allocating a new one:
 910          */
 911         list_for_each_entry(b, &c->btree_cache_freed, list)
 912                 if (!mca_reap(b, 0, false)) {
 913                         mca_data_alloc(b, k, __GFP_NOWARN|GFP_NOIO);
 914                         if (!b->keys.set[0].data)
 915                                 goto err;
 916                         else
 917                                 goto out;
 918                 }
 919
 920         b = mca_bucket_alloc(c, k, __GFP_NOWARN|GFP_NOIO);
 921         if (!b)
 922                 goto err;
 923
 924         BUG_ON(!down_write_trylock(&b->lock));
 925         if (!b->keys.set->data)
 926                 goto err;
 927 out:
 928         BUG_ON(b->io_mutex.count != 1);
 929
 930         bkey_copy(&b->key, k);
 931         list_move(&b->list, &c->btree_cache);
 932         hlist_del_init_rcu(&b->hash);
 933         hlist_add_head_rcu(&b->hash, mca_hash(c, k));
 934
 935         lock_set_subclass(&b->lock.dep_map, level + 1, _THIS_IP_);
 936         b->parent       = (void *) ~0UL;
 937         b->flags        = 0;
 938         b->written      = 0;
 939         b->level        = level;
 940
 941         if (!b->level)
 942                 bch_btree_keys_init(&b->keys, &bch_extent_keys_ops,
 943                                     &b->c->expensive_debug_checks);
 944         else
 945                 bch_btree_keys_init(&b->keys, &bch_btree_keys_ops,
 946                                     &b->c->expensive_debug_checks);
 947
 948         return b;
 949 err:
 950         if (b)
 951                 rw_unlock(true, b);
 952
 953         b = mca_cannibalize(c, op, k);
 954         if (!IS_ERR(b))
 955                 goto out;
 956
 957         return b;
 958 }
 959
 960 /**
 961  * bch_btree_node_get - find a btree node in the cache and lock it, reading it
 962  * in from disk if necessary.
 963  *
 964  * If IO is necessary and running under generic_make_request, returns -EAGAIN.
 965  *
 966  * The btree node will have either a read or a write lock held, depending on
 967  * level and op->lock.
 968  */
 969 struct btree *bch_btree_node_get(struct cache_set *c, struct btree_op *op,
 970                                  struct bkey *k, int level, bool write)
 971 {
 972         int i = 0;
 973         struct btree *b;
 974
 975         BUG_ON(level < 0);
 976 retry:
 977         b = mca_find(c, k);
 978
 979         if (!b) {
 980                 if (current->bio_list)
 981                         return ERR_PTR(-EAGAIN);
 982
 983                 mutex_lock(&c->bucket_lock);
 984                 b = mca_alloc(c, op, k, level);
 985                 mutex_unlock(&c->bucket_lock);
 986
 987                 if (!b)
 988                         goto retry;
 989                 if (IS_ERR(b))
 990                         return b;
 991
 992                 bch_btree_node_read(b);
 993
 994                 if (!write)
 995                         downgrade_write(&b->lock);
 996         } else {
 997                 rw_lock(write, b, level);
 998                 if (PTR_HASH(c, &b->key) != PTR_HASH(c, k)) {
 999                         rw_unlock(write, b);
1000                         goto retry;
1001                 }
1002                 BUG_ON(b->level != level);
1003         }
1004
1005         b->accessed = 1;
1006
1007         for (; i <= b->keys.nsets && b->keys.set[i].size; i++) {
1008                 prefetch(b->keys.set[i].tree);
1009                 prefetch(b->keys.set[i].data);
1010         }
1011
1012         for (; i <= b->keys.nsets; i++)
1013                 prefetch(b->keys.set[i].data);
1014
1015         if (btree_node_io_error(b)) {
1016                 rw_unlock(write, b);
1017                 return ERR_PTR(-EIO);
1018         }
1019
1020         BUG_ON(!b->written);
1021
1022         return b;
1023 }
1024
1025 static void btree_node_prefetch(struct cache_set *c, struct bkey *k, int level)
1026 {
1027         struct btree *b;
1028
1029         mutex_lock(&c->bucket_lock);
1030         b = mca_alloc(c, NULL, k, level);
1031         mutex_unlock(&c->bucket_lock);
1032
1033         if (!IS_ERR_OR_NULL(b)) {
1034                 bch_btree_node_read(b);
1035                 rw_unlock(true, b);
1036         }
1037 }
1038
1039 /* Btree alloc */
1040
1041 static void btree_node_free(struct btree *b)
1042 {
1043         trace_bcache_btree_node_free(b);
1044
1045         BUG_ON(b == b->c->root);
1046
1047         mutex_lock(&b->write_lock);
1048
1049         if (btree_node_dirty(b))
1050                 btree_complete_write(b, btree_current_write(b));
1051         clear_bit(BTREE_NODE_dirty, &b->flags);
1052
1053         mutex_unlock(&b->write_lock);
1054
1055         cancel_delayed_work(&b->work);
1056
1057         mutex_lock(&b->c->bucket_lock);
1058         bch_bucket_free(b->c, &b->key);
1059         mca_bucket_free(b);
1060         mutex_unlock(&b->c->bucket_lock);
1061 }
1062
1063 struct btree *bch_btree_node_alloc(struct cache_set *c, struct btree_op *op,
1064                                    int level)
1065 {
1066         BKEY_PADDED(key) k;
1067         struct btree *b = ERR_PTR(-EAGAIN);
1068
1069         mutex_lock(&c->bucket_lock);
1070 retry:
1071         if (__bch_bucket_alloc_set(c, RESERVE_BTREE, &k.key, 1, op != NULL))
1072                 goto err;
1073
1074         bkey_put(c, &k.key);
1075         SET_KEY_SIZE(&k.key, c->btree_pages * PAGE_SECTORS);
1076
1077         b = mca_alloc(c, op, &k.key, level);
1078         if (IS_ERR(b))
1079                 goto err_free;
1080
1081         if (!b) {
1082                 cache_bug(c,
1083                         "Tried to allocate bucket that was in btree cache");
1084                 goto retry;
1085         }
1086
1087         b->accessed = 1;
1088         bch_bset_init_next(&b->keys, b->keys.set->data, bset_magic(&b->c->sb));
1089
1090         mutex_unlock(&c->bucket_lock);
1091
1092         trace_bcache_btree_node_alloc(b);
1093         return b;
1094 err_free:
1095         bch_bucket_free(c, &k.key);
1096 err:
1097         mutex_unlock(&c->bucket_lock);
1098
1099         trace_bcache_btree_node_alloc_fail(b);
1100         return b;
1101 }
1102
1103 static struct btree *btree_node_alloc_replacement(struct btree *b,
1104                                                   struct btree_op *op)
1105 {
1106         struct btree *n = bch_btree_node_alloc(b->c, op, b->level);
1107         if (!IS_ERR_OR_NULL(n)) {
1108                 mutex_lock(&n->write_lock);
1109                 bch_btree_sort_into(&b->keys, &n->keys, &b->c->sort);
1110                 bkey_copy_key(&n->key, &b->key);
1111                 mutex_unlock(&n->write_lock);
1112         }
1113
1114         return n;
1115 }
1116
1117 static void make_btree_freeing_key(struct btree *b, struct bkey *k)
1118 {
1119         unsigned i;
1120
1121         mutex_lock(&b->c->bucket_lock);
1122
1123         atomic_inc(&b->c->prio_blocked);
1124
1125         bkey_copy(k, &b->key);
1126         bkey_copy_key(k, &ZERO_KEY);
1127
1128         for (i = 0; i < KEY_PTRS(k); i++)
1129                 SET_PTR_GEN(k, i,
1130                             bch_inc_gen(PTR_CACHE(b->c, &b->key, i),
1131                                         PTR_BUCKET(b->c, &b->key, i)));
1132
1133         mutex_unlock(&b->c->bucket_lock);
1134 }
1135
1136 static int btree_check_reserve(struct btree *b, struct btree_op *op)
1137 {
1138         struct cache_set *c = b->c;
1139         struct cache *ca;
1140         unsigned i, reserve = (c->root->level - b->level) * 2 + 1;
1141
1142         mutex_lock(&c->bucket_lock);
1143
1144         for_each_cache(ca, c, i)
1145                 if (fifo_used(&ca->free[RESERVE_BTREE]) < reserve) {
1146                         if (op)
1147                                 prepare_to_wait(&c->btree_cache_wait, &op->wait,
1148                                                 TASK_UNINTERRUPTIBLE);
1149                         mutex_unlock(&c->bucket_lock);
1150                         return -EINTR;
1151                 }
1152
1153         mutex_unlock(&c->bucket_lock);
1154
1155         return mca_cannibalize_lock(b->c, op);
1156 }
1157
1158 /* Garbage collection */
1159
1160 static uint8_t __bch_btree_mark_key(struct cache_set *c, int level,
1161                                     struct bkey *k)
1162 {
1163         uint8_t stale = 0;
1164         unsigned i;
1165         struct bucket *g;
1166
1167         /*
1168          * ptr_invalid() can't return true for the keys that mark btree nodes as
1169          * freed, but since ptr_bad() returns true we'll never actually use them
1170          * for anything and thus we don't want mark their pointers here
1171          */
1172         if (!bkey_cmp(k, &ZERO_KEY))
1173                 return stale;
1174
1175         for (i = 0; i < KEY_PTRS(k); i++) {
1176                 if (!ptr_available(c, k, i))
1177                         continue;
1178
1179                 g = PTR_BUCKET(c, k, i);
1180
1181                 if (gen_after(g->last_gc, PTR_GEN(k, i)))
1182                         g->last_gc = PTR_GEN(k, i);
1183
1184                 if (ptr_stale(c, k, i)) {
1185                         stale = max(stale, ptr_stale(c, k, i));
1186                         continue;
1187                 }
1188
1189                 cache_bug_on(GC_MARK(g) &&
1190                              (GC_MARK(g) == GC_MARK_METADATA) != (level != 0),
1191                              c, "inconsistent ptrs: mark = %llu, level = %i",
1192                              GC_MARK(g), level);
1193
1194                 if (level)
1195                         SET_GC_MARK(g, GC_MARK_METADATA);
1196                 else if (KEY_DIRTY(k))
1197                         SET_GC_MARK(g, GC_MARK_DIRTY);
1198                 else if (!GC_MARK(g))
1199                         SET_GC_MARK(g, GC_MARK_RECLAIMABLE);
1200
1201                 /* guard against overflow */
1202                 SET_GC_SECTORS_USED(g, min_t(unsigned,
1203                                              GC_SECTORS_USED(g) + KEY_SIZE(k),
1204                                              MAX_GC_SECTORS_USED));
1205
1206                 BUG_ON(!GC_SECTORS_USED(g));
1207         }
1208
1209         return stale;
1210 }
1211
1212 #define btree_mark_key(b, k)    __bch_btree_mark_key(b->c, b->level, k)
1213
1214 void bch_initial_mark_key(struct cache_set *c, int level, struct bkey *k)
1215 {
1216         unsigned i;
1217
1218         for (i = 0; i < KEY_PTRS(k); i++)
1219                 if (ptr_available(c, k, i) &&
1220                     !ptr_stale(c, k, i)) {
1221                         struct bucket *b = PTR_BUCKET(c, k, i);
1222
1223                         b->gen = PTR_GEN(k, i);
1224
1225                         if (level && bkey_cmp(k, &ZERO_KEY))
1226                                 b->prio = BTREE_PRIO;
1227                         else if (!level && b->prio == BTREE_PRIO)
1228                                 b->prio = INITIAL_PRIO;
1229                 }
1230
1231         __bch_btree_mark_key(c, level, k);
1232 }
1233
1234 static bool btree_gc_mark_node(struct btree *b, struct gc_stat *gc)
1235 {
1236         uint8_t stale = 0;
1237         unsigned keys = 0, good_keys = 0;
1238         struct bkey *k;
1239         struct btree_iter iter;
1240         struct bset_tree *t;
1241
1242         gc->nodes++;
1243
1244         for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid) {
1245                 stale = max(stale, btree_mark_key(b, k));
1246                 keys++;
1247
1248                 if (bch_ptr_bad(&b->keys, k))
1249                         continue;
1250
1251                 gc->key_bytes += bkey_u64s(k);
1252                 gc->nkeys++;
1253                 good_keys++;
1254
1255                 gc->data += KEY_SIZE(k);
1256         }
1257
1258         for (t = b->keys.set; t <= &b->keys.set[b->keys.nsets]; t++)
1259                 btree_bug_on(t->size &&
1260                              bset_written(&b->keys, t) &&
1261                              bkey_cmp(&b->key, &t->end) < 0,
1262                              b, "found short btree key in gc");
1263
1264         if (b->c->gc_always_rewrite)
1265                 return true;
1266
1267         if (stale > 10)
1268                 return true;
1269
1270         if ((keys - good_keys) * 2 > keys)
1271                 return true;
1272
1273         return false;
1274 }
1275
1276 #define GC_MERGE_NODES  4U
1277
1278 struct gc_merge_info {
1279         struct btree    *b;
1280         unsigned        keys;
1281 };
1282
1283 static int bch_btree_insert_node(struct btree *, struct btree_op *,
1284                                  struct keylist *, atomic_t *, struct bkey *);
1285
1286 static int btree_gc_coalesce(struct btree *b, struct btree_op *op,
1287                              struct gc_stat *gc, struct gc_merge_info *r)
1288 {
1289         unsigned i, nodes = 0, keys = 0, blocks;
1290         struct btree *new_nodes[GC_MERGE_NODES];
1291         struct keylist keylist;
1292         struct closure cl;
1293         struct bkey *k;
1294
1295         bch_keylist_init(&keylist);
1296
1297         if (btree_check_reserve(b, NULL))
1298                 return 0;
1299
1300         memset(new_nodes, 0, sizeof(new_nodes));
1301         closure_init_stack(&cl);
1302
1303         while (nodes < GC_MERGE_NODES && !IS_ERR_OR_NULL(r[nodes].b))
1304                 keys += r[nodes++].keys;
1305
1306         blocks = btree_default_blocks(b->c) * 2 / 3;
1307
1308         if (nodes < 2 ||
1309             __set_blocks(b->keys.set[0].data, keys,
1310                          block_bytes(b->c)) > blocks * (nodes - 1))
1311                 return 0;
1312
1313         for (i = 0; i < nodes; i++) {
1314                 new_nodes[i] = btree_node_alloc_replacement(r[i].b, NULL);
1315                 if (IS_ERR_OR_NULL(new_nodes[i]))
1316                         goto out_nocoalesce;
1317         }
1318
1319         /*
1320          * We have to check the reserve here, after we've allocated our new
1321          * nodes, to make sure the insert below will succeed - we also check
1322          * before as an optimization to potentially avoid a bunch of expensive
1323          * allocs/sorts
1324          */
1325         if (btree_check_reserve(b, NULL))
1326                 goto out_nocoalesce;
1327
1328         for (i = 0; i < nodes; i++)
1329                 mutex_lock(&new_nodes[i]->write_lock);
1330
1331         for (i = nodes - 1; i > 0; --i) {
1332                 struct bset *n1 = btree_bset_first(new_nodes[i]);
1333                 struct bset *n2 = btree_bset_first(new_nodes[i - 1]);
1334                 struct bkey *k, *last = NULL;
1335
1336                 keys = 0;
1337
1338                 if (i > 1) {
1339                         for (k = n2->start;
1340                              k < bset_bkey_last(n2);
1341                              k = bkey_next(k)) {
1342                                 if (__set_blocks(n1, n1->keys + keys +
1343                                                  bkey_u64s(k),
1344                                                  block_bytes(b->c)) > blocks)
1345                                         break;
1346
1347                                 last = k;
1348                                 keys += bkey_u64s(k);
1349                         }
1350                 } else {
1351                         /*
1352                          * Last node we're not getting rid of - we're getting
1353                          * rid of the node at r[0]. Have to try and fit all of
1354                          * the remaining keys into this node; we can't ensure
1355                          * they will always fit due to rounding and variable
1356                          * length keys (shouldn't be possible in practice,
1357                          * though)
1358                          */
1359                         if (__set_blocks(n1, n1->keys + n2->keys,
1360                                          block_bytes(b->c)) >
1361                             btree_blocks(new_nodes[i]))
1362                                 goto out_nocoalesce;
1363
1364                         keys = n2->keys;
1365                         /* Take the key of the node we're getting rid of */
1366                         last = &r->b->key;
1367                 }
1368
1369                 BUG_ON(__set_blocks(n1, n1->keys + keys, block_bytes(b->c)) >
1370                        btree_blocks(new_nodes[i]));
1371
1372                 if (last)
1373                         bkey_copy_key(&new_nodes[i]->key, last);
1374
1375                 memcpy(bset_bkey_last(n1),
1376                        n2->start,
1377                        (void *) bset_bkey_idx(n2, keys) - (void *) n2->start);
1378
1379                 n1->keys += keys;
1380                 r[i].keys = n1->keys;
1381
1382                 memmove(n2->start,
1383                         bset_bkey_idx(n2, keys),
1384                         (void *) bset_bkey_last(n2) -
1385                         (void *) bset_bkey_idx(n2, keys));
1386
1387                 n2->keys -= keys;
1388
1389                 if (__bch_keylist_realloc(&keylist,
1390                                           bkey_u64s(&new_nodes[i]->key)))
1391                         goto out_nocoalesce;
1392
1393                 bch_btree_node_write(new_nodes[i], &cl);
1394                 bch_keylist_add(&keylist, &new_nodes[i]->key);
1395         }
1396
1397         for (i = 0; i < nodes; i++)
1398                 mutex_unlock(&new_nodes[i]->write_lock);
1399
1400         closure_sync(&cl);
1401
1402         /* We emptied out this node */
1403         BUG_ON(btree_bset_first(new_nodes[0])->keys);
1404         btree_node_free(new_nodes[0]);
1405         rw_unlock(true, new_nodes[0]);
1406
1407         for (i = 0; i < nodes; i++) {
1408                 if (__bch_keylist_realloc(&keylist, bkey_u64s(&r[i].b->key)))
1409                         goto out_nocoalesce;
1410
1411                 make_btree_freeing_key(r[i].b, keylist.top);
1412                 bch_keylist_push(&keylist);
1413         }
1414
1415         bch_btree_insert_node(b, op, &keylist, NULL, NULL);
1416         BUG_ON(!bch_keylist_empty(&keylist));
1417
1418         for (i = 0; i < nodes; i++) {
1419                 btree_node_free(r[i].b);
1420                 rw_unlock(true, r[i].b);
1421
1422                 r[i].b = new_nodes[i];
1423         }
1424
1425         memmove(r, r + 1, sizeof(r[0]) * (nodes - 1));
1426         r[nodes - 1].b = ERR_PTR(-EINTR);
1427
1428         trace_bcache_btree_gc_coalesce(nodes);
1429         gc->nodes--;
1430
1431         bch_keylist_free(&keylist);
1432
1433         /* Invalidated our iterator */
1434         return -EINTR;
1435
1436 out_nocoalesce:
1437         closure_sync(&cl);
1438         bch_keylist_free(&keylist);
1439
1440         while ((k = bch_keylist_pop(&keylist)))
1441                 if (!bkey_cmp(k, &ZERO_KEY))
1442                         atomic_dec(&b->c->prio_blocked);
1443
1444         for (i = 0; i < nodes; i++)
1445                 if (!IS_ERR_OR_NULL(new_nodes[i])) {
1446                         btree_node_free(new_nodes[i]);
1447                         rw_unlock(true, new_nodes[i]);
1448                 }
1449         return 0;
1450 }
1451
1452 static int btree_gc_rewrite_node(struct btree *b, struct btree_op *op,
1453                                  struct btree *replace)
1454 {
1455         struct keylist keys;
1456         struct btree *n;
1457
1458         if (btree_check_reserve(b, NULL))
1459                 return 0;
1460
1461         n = btree_node_alloc_replacement(replace, NULL);
1462
1463         /* recheck reserve after allocating replacement node */
1464         if (btree_check_reserve(b, NULL)) {
1465                 btree_node_free(n);
1466                 rw_unlock(true, n);
1467                 return 0;
1468         }
1469
1470         bch_btree_node_write_sync(n);
1471
1472         bch_keylist_init(&keys);
1473         bch_keylist_add(&keys, &n->key);
1474
1475         make_btree_freeing_key(replace, keys.top);
1476         bch_keylist_push(&keys);
1477
1478         bch_btree_insert_node(b, op, &keys, NULL, NULL);
1479         BUG_ON(!bch_keylist_empty(&keys));
1480
1481         btree_node_free(replace);
1482         rw_unlock(true, n);
1483
1484         /* Invalidated our iterator */
1485         return -EINTR;
1486 }
1487
1488 static unsigned btree_gc_count_keys(struct btree *b)
1489 {
1490         struct bkey *k;
1491         struct btree_iter iter;
1492         unsigned ret = 0;
1493
1494         for_each_key_filter(&b->keys, k, &iter, bch_ptr_bad)
1495                 ret += bkey_u64s(k);
1496
1497         return ret;
1498 }
1499
1500 static int btree_gc_recurse(struct btree *b, struct btree_op *op,
1501                             struct closure *writes, struct gc_stat *gc)
1502 {
1503         int ret = 0;
1504         bool should_rewrite;
1505         struct bkey *k;
1506         struct btree_iter iter;
1507         struct gc_merge_info r[GC_MERGE_NODES];
1508         struct gc_merge_info *i, *last = r + ARRAY_SIZE(r) - 1;
1509
1510         bch_btree_iter_init(&b->keys, &iter, &b->c->gc_done);
1511
1512         for (i = r; i < r + ARRAY_SIZE(r); i++)
1513                 i->b = ERR_PTR(-EINTR);
1514
1515         while (1) {
1516                 k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad);
1517                 if (k) {
1518                         r->b = bch_btree_node_get(b->c, op, k, b->level - 1,
1519                                                   true);
1520                         if (IS_ERR(r->b)) {
1521                                 ret = PTR_ERR(r->b);
1522                                 break;
1523                         }
1524
1525                         r->keys = btree_gc_count_keys(r->b);
1526
1527                         ret = btree_gc_coalesce(b, op, gc, r);
1528                         if (ret)
1529                                 break;
1530                 }
1531
1532                 if (!last->b)
1533                         break;
1534
1535                 if (!IS_ERR(last->b)) {
1536                         should_rewrite = btree_gc_mark_node(last->b, gc);
1537                         if (should_rewrite) {
1538                                 ret = btree_gc_rewrite_node(b, op, last->b);
1539                                 if (ret)
1540                                         break;
1541                         }
1542
1543                         if (last->b->level) {
1544                                 ret = btree_gc_recurse(last->b, op, writes, gc);
1545                                 if (ret)
1546                                         break;
1547                         }
1548
1549                         bkey_copy_key(&b->c->gc_done, &last->b->key);
1550
1551                         /*
1552                          * Must flush leaf nodes before gc ends, since replace
1553                          * operations aren't journalled
1554                          */
1555                         mutex_lock(&last->b->write_lock);
1556                         if (btree_node_dirty(last->b))
1557                                 bch_btree_node_write(last->b, writes);
1558                         mutex_unlock(&last->b->write_lock);
1559                         rw_unlock(true, last->b);
1560                 }
1561
1562                 memmove(r + 1, r, sizeof(r[0]) * (GC_MERGE_NODES - 1));
1563                 r->b = NULL;
1564
1565                 if (need_resched()) {
1566                         ret = -EAGAIN;
1567                         break;
1568                 }
1569         }
1570
1571         for (i = r; i < r + ARRAY_SIZE(r); i++)
1572                 if (!IS_ERR_OR_NULL(i->b)) {
1573                         mutex_lock(&i->b->write_lock);
1574                         if (btree_node_dirty(i->b))
1575                                 bch_btree_node_write(i->b, writes);
1576                         mutex_unlock(&i->b->write_lock);
1577                         rw_unlock(true, i->b);
1578                 }
1579
1580         return ret;
1581 }
1582
1583 static int bch_btree_gc_root(struct btree *b, struct btree_op *op,
1584                              struct closure *writes, struct gc_stat *gc)
1585 {
1586         struct btree *n = NULL;
1587         int ret = 0;
1588         bool should_rewrite;
1589
1590         should_rewrite = btree_gc_mark_node(b, gc);
1591         if (should_rewrite) {
1592                 n = btree_node_alloc_replacement(b, NULL);
1593
1594                 if (!IS_ERR_OR_NULL(n)) {
1595                         bch_btree_node_write_sync(n);
1596
1597                         bch_btree_set_root(n);
1598                         btree_node_free(b);
1599                         rw_unlock(true, n);
1600
1601                         return -EINTR;
1602                 }
1603         }
1604
1605         __bch_btree_mark_key(b->c, b->level + 1, &b->key);
1606
1607         if (b->level) {
1608                 ret = btree_gc_recurse(b, op, writes, gc);
1609                 if (ret)
1610                         return ret;
1611         }
1612
1613         bkey_copy_key(&b->c->gc_done, &b->key);
1614
1615         return ret;
1616 }
1617
1618 static void btree_gc_start(struct cache_set *c)
1619 {
1620         struct cache *ca;
1621         struct bucket *b;
1622         unsigned i;
1623
1624         if (!c->gc_mark_valid)
1625                 return;
1626
1627         mutex_lock(&c->bucket_lock);
1628
1629         c->gc_mark_valid = 0;
1630         c->gc_done = ZERO_KEY;
1631
1632         for_each_cache(ca, c, i)
1633                 for_each_bucket(b, ca) {
1634                         b->last_gc = b->gen;
1635                         if (!atomic_read(&b->pin)) {
1636                                 SET_GC_MARK(b, 0);
1637                                 SET_GC_SECTORS_USED(b, 0);
1638                         }
1639                 }
1640
1641         mutex_unlock(&c->bucket_lock);
1642 }
1643
1644 static size_t bch_btree_gc_finish(struct cache_set *c)
1645 {
1646         size_t available = 0;
1647         struct bucket *b;
1648         struct cache *ca;
1649         unsigned i;
1650
1651         mutex_lock(&c->bucket_lock);
1652
1653         set_gc_sectors(c);
1654         c->gc_mark_valid = 1;
1655         c->need_gc      = 0;
1656
1657         for (i = 0; i < KEY_PTRS(&c->uuid_bucket); i++)
1658                 SET_GC_MARK(PTR_BUCKET(c, &c->uuid_bucket, i),
1659                             GC_MARK_METADATA);
1660
1661         /* don't reclaim buckets to which writeback keys point */
1662         rcu_read_lock();
1663         for (i = 0; i < c->nr_uuids; i++) {
1664                 struct bcache_device *d = c->devices[i];
1665                 struct cached_dev *dc;
1666                 struct keybuf_key *w, *n;
1667                 unsigned j;
1668
1669                 if (!d || UUID_FLASH_ONLY(&c->uuids[i]))
1670                         continue;
1671                 dc = container_of(d, struct cached_dev, disk);
1672
1673                 spin_lock(&dc->writeback_keys.lock);
1674                 rbtree_postorder_for_each_entry_safe(w, n,
1675                                         &dc->writeback_keys.keys, node)
1676                         for (j = 0; j < KEY_PTRS(&w->key); j++)
1677                                 SET_GC_MARK(PTR_BUCKET(c, &w->key, j),
1678                                             GC_MARK_DIRTY);
1679                 spin_unlock(&dc->writeback_keys.lock);
1680         }
1681         rcu_read_unlock();
1682
1683         for_each_cache(ca, c, i) {
1684                 uint64_t *i;
1685
1686                 ca->invalidate_needs_gc = 0;
1687
1688                 for (i = ca->sb.d; i < ca->sb.d + ca->sb.keys; i++)
1689                         SET_GC_MARK(ca->buckets + *i, GC_MARK_METADATA);
1690
1691                 for (i = ca->prio_buckets;
1692                      i < ca->prio_buckets + prio_buckets(ca) * 2; i++)
1693                         SET_GC_MARK(ca->buckets + *i, GC_MARK_METADATA);
1694
1695                 for_each_bucket(b, ca) {
1696                         c->need_gc      = max(c->need_gc, bucket_gc_gen(b));
1697
1698                         if (atomic_read(&b->pin))
1699                                 continue;
1700
1701                         BUG_ON(!GC_MARK(b) && GC_SECTORS_USED(b));
1702
1703                         if (!GC_MARK(b) || GC_MARK(b) == GC_MARK_RECLAIMABLE)
1704                                 available++;
1705                 }
1706         }
1707
1708         mutex_unlock(&c->bucket_lock);
1709         return available;
1710 }
1711
1712 static void bch_btree_gc(struct cache_set *c)
1713 {
1714         int ret;
1715         unsigned long available;
1716         struct gc_stat stats;
1717         struct closure writes;
1718         struct btree_op op;
1719         uint64_t start_time = local_clock();
1720
1721         trace_bcache_gc_start(c);
1722
1723         memset(&stats, 0, sizeof(struct gc_stat));
1724         closure_init_stack(&writes);
1725         bch_btree_op_init(&op, SHRT_MAX);
1726
1727         btree_gc_start(c);
1728
1729         do {
1730                 ret = btree_root(gc_root, c, &op, &writes, &stats);
1731                 closure_sync(&writes);
1732
1733                 if (ret && ret != -EAGAIN)
1734                         pr_warn("gc failed!");
1735         } while (ret);
1736
1737         available = bch_btree_gc_finish(c);
1738         wake_up_allocators(c);
1739
1740         bch_time_stats_update(&c->btree_gc_time, start_time);
1741
1742         stats.key_bytes *= sizeof(uint64_t);
1743         stats.data      <<= 9;
1744         stats.in_use    = (c->nbuckets - available) * 100 / c->nbuckets;
1745         memcpy(&c->gc_stats, &stats, sizeof(struct gc_stat));
1746
1747         trace_bcache_gc_end(c);
1748
1749         bch_moving_gc(c);
1750 }
1751
1752 static int bch_gc_thread(void *arg)
1753 {
1754         struct cache_set *c = arg;
1755         struct cache *ca;
1756         unsigned i;
1757
1758         while (1) {
1759 again:
1760                 bch_btree_gc(c);
1761
1762                 set_current_state(TASK_INTERRUPTIBLE);
1763                 if (kthread_should_stop())
1764                         break;
1765
1766                 mutex_lock(&c->bucket_lock);
1767
1768                 for_each_cache(ca, c, i)
1769                         if (ca->invalidate_needs_gc) {
1770                                 mutex_unlock(&c->bucket_lock);
1771                                 set_current_state(TASK_RUNNING);
1772                                 goto again;
1773                         }
1774
1775                 mutex_unlock(&c->bucket_lock);
1776
1777                 try_to_freeze();
1778                 schedule();
1779         }
1780
1781         return 0;
1782 }
1783
1784 int bch_gc_thread_start(struct cache_set *c)
1785 {
1786         c->gc_thread = kthread_create(bch_gc_thread, c, "bcache_gc");
1787         if (IS_ERR(c->gc_thread))
1788                 return PTR_ERR(c->gc_thread);
1789
1790         set_task_state(c->gc_thread, TASK_INTERRUPTIBLE);
1791         return 0;
1792 }
1793
1794 /* Initial partial gc */
1795
1796 static int bch_btree_check_recurse(struct btree *b, struct btree_op *op)
1797 {
1798         int ret = 0;
1799         struct bkey *k, *p = NULL;
1800         struct btree_iter iter;
1801
1802         for_each_key_filter(&b->keys, k, &iter, bch_ptr_invalid)
1803                 bch_initial_mark_key(b->c, b->level, k);
1804
1805         bch_initial_mark_key(b->c, b->level + 1, &b->key);
1806
1807         if (b->level) {
1808                 bch_btree_iter_init(&b->keys, &iter, NULL);
1809
1810                 do {
1811                         k = bch_btree_iter_next_filter(&iter, &b->keys,
1812                                                        bch_ptr_bad);
1813                         if (k)
1814                                 btree_node_prefetch(b->c, k, b->level - 1);
1815
1816                         if (p)
1817                                 ret = btree(check_recurse, p, b, op);
1818
1819                         p = k;
1820                 } while (p && !ret);
1821         }
1822
1823         return ret;
1824 }
1825
1826 int bch_btree_check(struct cache_set *c)
1827 {
1828         struct btree_op op;
1829
1830         bch_btree_op_init(&op, SHRT_MAX);
1831
1832         return btree_root(check_recurse, c, &op);
1833 }
1834
1835 void bch_initial_gc_finish(struct cache_set *c)
1836 {
1837         struct cache *ca;
1838         struct bucket *b;
1839         unsigned i;
1840
1841         bch_btree_gc_finish(c);
1842
1843         mutex_lock(&c->bucket_lock);
1844
1845         /*
1846          * We need to put some unused buckets directly on the prio freelist in
1847          * order to get the allocator thread started - it needs freed buckets in
1848          * order to rewrite the prios and gens, and it needs to rewrite prios
1849          * and gens in order to free buckets.
1850          *
1851          * This is only safe for buckets that have no live data in them, which
1852          * there should always be some of.
1853          */
1854         for_each_cache(ca, c, i) {
1855                 for_each_bucket(b, ca) {
1856                         if (fifo_full(&ca->free[RESERVE_PRIO]))
1857                                 break;
1858
1859                         if (bch_can_invalidate_bucket(ca, b) &&
1860                             !GC_MARK(b)) {
1861                                 __bch_invalidate_one_bucket(ca, b);
1862                                 fifo_push(&ca->free[RESERVE_PRIO],
1863                                           b - ca->buckets);
1864                         }
1865                 }
1866         }
1867
1868         mutex_unlock(&c->bucket_lock);
1869 }
1870
1871 /* Btree insertion */
1872
1873 static bool btree_insert_key(struct btree *b, struct bkey *k,
1874                              struct bkey *replace_key)
1875 {
1876         unsigned status;
1877
1878         BUG_ON(bkey_cmp(k, &b->key) > 0);
1879
1880         status = bch_btree_insert_key(&b->keys, k, replace_key);
1881         if (status != BTREE_INSERT_STATUS_NO_INSERT) {
1882                 bch_check_keys(&b->keys, "%u for %s", status,
1883                                replace_key ? "replace" : "insert");
1884
1885                 trace_bcache_btree_insert_key(b, k, replace_key != NULL,
1886                                               status);
1887                 return true;
1888         } else
1889                 return false;
1890 }
1891
1892 static size_t insert_u64s_remaining(struct btree *b)
1893 {
1894         long ret = bch_btree_keys_u64s_remaining(&b->keys);
1895
1896         /*
1897          * Might land in the middle of an existing extent and have to split it
1898          */
1899         if (b->keys.ops->is_extents)
1900                 ret -= KEY_MAX_U64S;
1901
1902         return max(ret, 0L);
1903 }
1904
1905 static bool bch_btree_insert_keys(struct btree *b, struct btree_op *op,
1906                                   struct keylist *insert_keys,
1907                                   struct bkey *replace_key)
1908 {
1909         bool ret = false;
1910         int oldsize = bch_count_data(&b->keys);
1911
1912         while (!bch_keylist_empty(insert_keys)) {
1913                 struct bkey *k = insert_keys->keys;
1914
1915                 if (bkey_u64s(k) > insert_u64s_remaining(b))
1916                         break;
1917
1918                 if (bkey_cmp(k, &b->key) <= 0) {
1919                         if (!b->level)
1920                                 bkey_put(b->c, k);
1921
1922                         ret |= btree_insert_key(b, k, replace_key);
1923                         bch_keylist_pop_front(insert_keys);
1924                 } else if (bkey_cmp(&START_KEY(k), &b->key) < 0) {
1925                         BKEY_PADDED(key) temp;
1926                         bkey_copy(&temp.key, insert_keys->keys);
1927
1928                         bch_cut_back(&b->key, &temp.key);
1929                         bch_cut_front(&b->key, insert_keys->keys);
1930
1931                         ret |= btree_insert_key(b, &temp.key, replace_key);
1932                         break;
1933                 } else {
1934                         break;
1935                 }
1936         }
1937
1938         if (!ret)
1939                 op->insert_collision = true;
1940
1941         BUG_ON(!bch_keylist_empty(insert_keys) && b->level);
1942
1943         BUG_ON(bch_count_data(&b->keys) < oldsize);
1944         return ret;
1945 }
1946
1947 static int btree_split(struct btree *b, struct btree_op *op,
1948                        struct keylist *insert_keys,
1949                        struct bkey *replace_key)
1950 {
1951         bool split;
1952         struct btree *n1, *n2 = NULL, *n3 = NULL;
1953         uint64_t start_time = local_clock();
1954         struct closure cl;
1955         struct keylist parent_keys;
1956
1957         closure_init_stack(&cl);
1958         bch_keylist_init(&parent_keys);
1959
1960         if (btree_check_reserve(b, op)) {
1961                 if (!b->level)
1962                         return -EINTR;
1963                 else
1964                         WARN(1, "insufficient reserve for split\n");
1965         }
1966
1967         n1 = btree_node_alloc_replacement(b, op);
1968         if (IS_ERR(n1))
1969                 goto err;
1970
1971         split = set_blocks(btree_bset_first(n1),
1972                            block_bytes(n1->c)) > (btree_blocks(b) * 4) / 5;
1973
1974         if (split) {
1975                 unsigned keys = 0;
1976
1977                 trace_bcache_btree_node_split(b, btree_bset_first(n1)->keys);
1978
1979                 n2 = bch_btree_node_alloc(b->c, op, b->level);
1980                 if (IS_ERR(n2))
1981                         goto err_free1;
1982
1983                 if (!b->parent) {
1984                         n3 = bch_btree_node_alloc(b->c, op, b->level + 1);
1985                         if (IS_ERR(n3))
1986                                 goto err_free2;
1987                 }
1988
1989                 mutex_lock(&n1->write_lock);
1990                 mutex_lock(&n2->write_lock);
1991
1992                 bch_btree_insert_keys(n1, op, insert_keys, replace_key);
1993
1994                 /*
1995                  * Has to be a linear search because we don't have an auxiliary
1996                  * search tree yet
1997                  */
1998
1999                 while (keys < (btree_bset_first(n1)->keys * 3) / 5)
2000                         keys += bkey_u64s(bset_bkey_idx(btree_bset_first(n1),
2001                                                         keys));
2002
2003                 bkey_copy_key(&n1->key,
2004                               bset_bkey_idx(btree_bset_first(n1), keys));
2005                 keys += bkey_u64s(bset_bkey_idx(btree_bset_first(n1), keys));
2006
2007                 btree_bset_first(n2)->keys = btree_bset_first(n1)->keys - keys;
2008                 btree_bset_first(n1)->keys = keys;
2009
2010                 memcpy(btree_bset_first(n2)->start,
2011                        bset_bkey_last(btree_bset_first(n1)),
2012                        btree_bset_first(n2)->keys * sizeof(uint64_t));
2013
2014                 bkey_copy_key(&n2->key, &b->key);
2015
2016                 bch_keylist_add(&parent_keys, &n2->key);
2017                 bch_btree_node_write(n2, &cl);
2018                 mutex_unlock(&n2->write_lock);
2019                 rw_unlock(true, n2);
2020         } else {
2021                 trace_bcache_btree_node_compact(b, btree_bset_first(n1)->keys);
2022
2023                 mutex_lock(&n1->write_lock);
2024                 bch_btree_insert_keys(n1, op, insert_keys, replace_key);
2025         }
2026
2027         bch_keylist_add(&parent_keys, &n1->key);
2028         bch_btree_node_write(n1, &cl);
2029         mutex_unlock(&n1->write_lock);
2030
2031         if (n3) {
2032                 /* Depth increases, make a new root */
2033                 mutex_lock(&n3->write_lock);
2034                 bkey_copy_key(&n3->key, &MAX_KEY);
2035                 bch_btree_insert_keys(n3, op, &parent_keys, NULL);
2036                 bch_btree_node_write(n3, &cl);
2037                 mutex_unlock(&n3->write_lock);
2038
2039                 closure_sync(&cl);
2040                 bch_btree_set_root(n3);
2041                 rw_unlock(true, n3);
2042         } else if (!b->parent) {
2043                 /* Root filled up but didn't need to be split */
2044                 closure_sync(&cl);
2045                 bch_btree_set_root(n1);
2046         } else {
2047                 /* Split a non root node */
2048                 closure_sync(&cl);
2049                 make_btree_freeing_key(b, parent_keys.top);
2050                 bch_keylist_push(&parent_keys);
2051
2052                 bch_btree_insert_node(b->parent, op, &parent_keys, NULL, NULL);
2053                 BUG_ON(!bch_keylist_empty(&parent_keys));
2054         }
2055
2056         btree_node_free(b);
2057         rw_unlock(true, n1);
2058
2059         bch_time_stats_update(&b->c->btree_split_time, start_time);
2060
2061         return 0;
2062 err_free2:
2063         bkey_put(b->c, &n2->key);
2064         btree_node_free(n2);
2065         rw_unlock(true, n2);
2066 err_free1:
2067         bkey_put(b->c, &n1->key);
2068         btree_node_free(n1);
2069         rw_unlock(true, n1);
2070 err:
2071         WARN(1, "bcache: btree split failed (level %u)", b->level);
2072
2073         if (n3 == ERR_PTR(-EAGAIN) ||
2074             n2 == ERR_PTR(-EAGAIN) ||
2075             n1 == ERR_PTR(-EAGAIN))
2076                 return -EAGAIN;
2077
2078         return -ENOMEM;
2079 }
2080
2081 static int bch_btree_insert_node(struct btree *b, struct btree_op *op,
2082                                  struct keylist *insert_keys,
2083                                  atomic_t *journal_ref,
2084                                  struct bkey *replace_key)
2085 {
2086         struct closure cl;
2087
2088         BUG_ON(b->level && replace_key);
2089
2090         closure_init_stack(&cl);
2091
2092         mutex_lock(&b->write_lock);
2093
2094         if (write_block(b) != btree_bset_last(b) &&
2095             b->keys.last_set_unwritten)
2096                 bch_btree_init_next(b); /* just wrote a set */
2097
2098         if (bch_keylist_nkeys(insert_keys) > insert_u64s_remaining(b)) {
2099                 mutex_unlock(&b->write_lock);
2100                 goto split;
2101         }
2102
2103         BUG_ON(write_block(b) != btree_bset_last(b));
2104
2105         if (bch_btree_insert_keys(b, op, insert_keys, replace_key)) {
2106                 if (!b->level)
2107                         bch_btree_leaf_dirty(b, journal_ref);
2108                 else
2109                         bch_btree_node_write(b, &cl);
2110         }
2111
2112         mutex_unlock(&b->write_lock);
2113
2114         /* wait for btree node write if necessary, after unlock */
2115         closure_sync(&cl);
2116
2117         return 0;
2118 split:
2119         if (current->bio_list) {
2120                 op->lock = b->c->root->level + 1;
2121                 return -EAGAIN;
2122         } else if (op->lock <= b->c->root->level) {
2123                 op->lock = b->c->root->level + 1;
2124                 return -EINTR;
2125         } else {
2126                 /* Invalidated all iterators */
2127                 int ret = btree_split(b, op, insert_keys, replace_key);
2128
2129                 if (bch_keylist_empty(insert_keys))
2130                         return 0;
2131                 else if (!ret)
2132                         return -EINTR;
2133                 return ret;
2134         }
2135 }
2136
2137 int bch_btree_insert_check_key(struct btree *b, struct btree_op *op,
2138                                struct bkey *check_key)
2139 {
2140         int ret = -EINTR;
2141         uint64_t btree_ptr = b->key.ptr[0];
2142         unsigned long seq = b->seq;
2143         struct keylist insert;
2144         bool upgrade = op->lock == -1;
2145
2146         bch_keylist_init(&insert);
2147
2148         if (upgrade) {
2149                 rw_unlock(false, b);
2150                 rw_lock(true, b, b->level);
2151
2152                 if (b->key.ptr[0] != btree_ptr ||
2153                     b->seq != seq + 1)
2154                         goto out;
2155         }
2156
2157         SET_KEY_PTRS(check_key, 1);
2158         get_random_bytes(&check_key->ptr[0], sizeof(uint64_t));
2159
2160         SET_PTR_DEV(check_key, 0, PTR_CHECK_DEV);
2161
2162         bch_keylist_add(&insert, check_key);
2163
2164         ret = bch_btree_insert_node(b, op, &insert, NULL, NULL);
2165
2166         BUG_ON(!ret && !bch_keylist_empty(&insert));
2167 out:
2168         if (upgrade)
2169                 downgrade_write(&b->lock);
2170         return ret;
2171 }
2172
2173 struct btree_insert_op {
2174         struct btree_op op;
2175         struct keylist  *keys;
2176         atomic_t        *journal_ref;
2177         struct bkey     *replace_key;
2178 };
2179
2180 static int btree_insert_fn(struct btree_op *b_op, struct btree *b)
2181 {
2182         struct btree_insert_op *op = container_of(b_op,
2183                                         struct btree_insert_op, op);
2184
2185         int ret = bch_btree_insert_node(b, &op->op, op->keys,
2186                                         op->journal_ref, op->replace_key);
2187         if (ret && !bch_keylist_empty(op->keys))
2188                 return ret;
2189         else
2190                 return MAP_DONE;
2191 }
2192
2193 int bch_btree_insert(struct cache_set *c, struct keylist *keys,
2194                      atomic_t *journal_ref, struct bkey *replace_key)
2195 {
2196         struct btree_insert_op op;
2197         int ret = 0;
2198
2199         BUG_ON(current->bio_list);
2200         BUG_ON(bch_keylist_empty(keys));
2201
2202         bch_btree_op_init(&op.op, 0);
2203         op.keys         = keys;
2204         op.journal_ref  = journal_ref;
2205         op.replace_key  = replace_key;
2206
2207         while (!ret && !bch_keylist_empty(keys)) {
2208                 op.op.lock = 0;
2209                 ret = bch_btree_map_leaf_nodes(&op.op, c,
2210                                                &START_KEY(keys->keys),
2211                                                btree_insert_fn);
2212         }
2213
2214         if (ret) {
2215                 struct bkey *k;
2216
2217                 pr_err("error %i", ret);
2218
2219                 while ((k = bch_keylist_pop(keys)))
2220                         bkey_put(c, k);
2221         } else if (op.op.insert_collision)
2222                 ret = -ESRCH;
2223
2224         return ret;
2225 }
2226
2227 void bch_btree_set_root(struct btree *b)
2228 {
2229         unsigned i;
2230         struct closure cl;
2231
2232         closure_init_stack(&cl);
2233
2234         trace_bcache_btree_set_root(b);
2235
2236         BUG_ON(!b->written);
2237
2238         for (i = 0; i < KEY_PTRS(&b->key); i++)
2239                 BUG_ON(PTR_BUCKET(b->c, &b->key, i)->prio != BTREE_PRIO);
2240
2241         mutex_lock(&b->c->bucket_lock);
2242         list_del_init(&b->list);
2243         mutex_unlock(&b->c->bucket_lock);
2244
2245         b->c->root = b;
2246
2247         bch_journal_meta(b->c, &cl);
2248         closure_sync(&cl);
2249 }
2250
2251 /* Map across nodes or keys */
2252
2253 static int bch_btree_map_nodes_recurse(struct btree *b, struct btree_op *op,
2254                                        struct bkey *from,
2255                                        btree_map_nodes_fn *fn, int flags)
2256 {
2257         int ret = MAP_CONTINUE;
2258
2259         if (b->level) {
2260                 struct bkey *k;
2261                 struct btree_iter iter;
2262
2263                 bch_btree_iter_init(&b->keys, &iter, from);
2264
2265                 while ((k = bch_btree_iter_next_filter(&iter, &b->keys,
2266                                                        bch_ptr_bad))) {
2267                         ret = btree(map_nodes_recurse, k, b,
2268                                     op, from, fn, flags);
2269                         from = NULL;
2270
2271                         if (ret != MAP_CONTINUE)
2272                                 return ret;
2273                 }
2274         }
2275
2276         if (!b->level || flags == MAP_ALL_NODES)
2277                 ret = fn(op, b);
2278
2279         return ret;
2280 }
2281
2282 int __bch_btree_map_nodes(struct btree_op *op, struct cache_set *c,
2283                           struct bkey *from, btree_map_nodes_fn *fn, int flags)
2284 {
2285         return btree_root(map_nodes_recurse, c, op, from, fn, flags);
2286 }
2287
2288 static int bch_btree_map_keys_recurse(struct btree *b, struct btree_op *op,
2289                                       struct bkey *from, btree_map_keys_fn *fn,
2290                                       int flags)
2291 {
2292         int ret = MAP_CONTINUE;
2293         struct bkey *k;
2294         struct btree_iter iter;
2295
2296         bch_btree_iter_init(&b->keys, &iter, from);
2297
2298         while ((k = bch_btree_iter_next_filter(&iter, &b->keys, bch_ptr_bad))) {
2299                 ret = !b->level
2300                         ? fn(op, b, k)
2301                         : btree(map_keys_recurse, k, b, op, from, fn, flags);
2302                 from = NULL;
2303
2304                 if (ret != MAP_CONTINUE)
2305                         return ret;
2306         }
2307
2308         if (!b->level && (flags & MAP_END_KEY))
2309                 ret = fn(op, b, &KEY(KEY_INODE(&b->key),
2310                                      KEY_OFFSET(&b->key), 0));
2311
2312         return ret;
2313 }
2314
2315 int bch_btree_map_keys(struct btree_op *op, struct cache_set *c,
2316                        struct bkey *from, btree_map_keys_fn *fn, int flags)
2317 {
2318         return btree_root(map_keys_recurse, c, op, from, fn, flags);
2319 }
2320
2321 /* Keybuf code */
2322
2323 static inline int keybuf_cmp(struct keybuf_key *l, struct keybuf_key *r)
2324 {
2325         /* Overlapping keys compare equal */
2326         if (bkey_cmp(&l->key, &START_KEY(&r->key)) <= 0)
2327                 return -1;
2328         if (bkey_cmp(&START_KEY(&l->key), &r->key) >= 0)
2329                 return 1;
2330         return 0;
2331 }
2332
2333 static inline int keybuf_nonoverlapping_cmp(struct keybuf_key *l,
2334                                             struct keybuf_key *r)
2335 {
2336         return clamp_t(int64_t, bkey_cmp(&l->key, &r->key), -1, 1);
2337 }
2338
2339 struct refill {
2340         struct btree_op op;
2341         unsigned        nr_found;
2342         struct keybuf   *buf;
2343         struct bkey     *end;
2344         keybuf_pred_fn  *pred;
2345 };
2346
2347 static int refill_keybuf_fn(struct btree_op *op, struct btree *b,
2348                             struct bkey *k)
2349 {
2350         struct refill *refill = container_of(op, struct refill, op);
2351         struct keybuf *buf = refill->buf;
2352         int ret = MAP_CONTINUE;
2353
2354         if (bkey_cmp(k, refill->end) >= 0) {
2355                 ret = MAP_DONE;
2356                 goto out;
2357         }
2358
2359         if (!KEY_SIZE(k)) /* end key */
2360                 goto out;
2361
2362         if (refill->pred(buf, k)) {
2363                 struct keybuf_key *w;
2364
2365                 spin_lock(&buf->lock);
2366
2367                 w = array_alloc(&buf->freelist);
2368                 if (!w) {
2369                         spin_unlock(&buf->lock);
2370                         return MAP_DONE;
2371                 }
2372
2373                 w->private = NULL;
2374                 bkey_copy(&w->key, k);
2375
2376                 if (RB_INSERT(&buf->keys, w, node, keybuf_cmp))
2377                         array_free(&buf->freelist, w);
2378                 else
2379                         refill->nr_found++;
2380
2381                 if (array_freelist_empty(&buf->freelist))
2382                         ret = MAP_DONE;
2383
2384                 spin_unlock(&buf->lock);
2385         }
2386 out:
2387         buf->last_scanned = *k;
2388         return ret;
2389 }
2390
2391 void bch_refill_keybuf(struct cache_set *c, struct keybuf *buf,
2392                        struct bkey *end, keybuf_pred_fn *pred)
2393 {
2394         struct bkey start = buf->last_scanned;
2395         struct refill refill;
2396
2397         cond_resched();
2398
2399         bch_btree_op_init(&refill.op, -1);
2400         refill.nr_found = 0;
2401         refill.buf      = buf;
2402         refill.end      = end;
2403         refill.pred     = pred;
2404
2405         bch_btree_map_keys(&refill.op, c, &buf->last_scanned,
2406                            refill_keybuf_fn, MAP_END_KEY);
2407
2408         trace_bcache_keyscan(refill.nr_found,
2409                              KEY_INODE(&start), KEY_OFFSET(&start),
2410                              KEY_INODE(&buf->last_scanned),
2411                              KEY_OFFSET(&buf->last_scanned));
2412
2413         spin_lock(&buf->lock);
2414
2415         if (!RB_EMPTY_ROOT(&buf->keys)) {
2416                 struct keybuf_key *w;
2417                 w = RB_FIRST(&buf->keys, struct keybuf_key, node);
2418                 buf->start      = START_KEY(&w->key);
2419
2420                 w = RB_LAST(&buf->keys, struct keybuf_key, node);
2421                 buf->end        = w->key;
2422         } else {
2423                 buf->start      = MAX_KEY;
2424                 buf->end        = MAX_KEY;
2425         }
2426
2427         spin_unlock(&buf->lock);
2428 }
2429
2430 static void __bch_keybuf_del(struct keybuf *buf, struct keybuf_key *w)
2431 {
2432         rb_erase(&w->node, &buf->keys);
2433         array_free(&buf->freelist, w);
2434 }
2435
2436 void bch_keybuf_del(struct keybuf *buf, struct keybuf_key *w)
2437 {
2438         spin_lock(&buf->lock);
2439         __bch_keybuf_del(buf, w);
2440         spin_unlock(&buf->lock);
2441 }
2442
2443 bool bch_keybuf_check_overlapping(struct keybuf *buf, struct bkey *start,
2444                                   struct bkey *end)
2445 {
2446         bool ret = false;
2447         struct keybuf_key *p, *w, s;
2448         s.key = *start;
2449
2450         if (bkey_cmp(end, &buf->start) <= 0 ||
2451             bkey_cmp(start, &buf->end) >= 0)
2452                 return false;
2453
2454         spin_lock(&buf->lock);
2455         w = RB_GREATER(&buf->keys, s, node, keybuf_nonoverlapping_cmp);
2456
2457         while (w && bkey_cmp(&START_KEY(&w->key), end) < 0) {
2458                 p = w;
2459                 w = RB_NEXT(w, node);
2460
2461                 if (p->private)
2462                         ret = true;
2463                 else
2464                         __bch_keybuf_del(buf, p);
2465         }
2466
2467         spin_unlock(&buf->lock);
2468         return ret;
2469 }
2470
2471 struct keybuf_key *bch_keybuf_next(struct keybuf *buf)
2472 {
2473         struct keybuf_key *w;
2474         spin_lock(&buf->lock);
2475
2476         w = RB_FIRST(&buf->keys, struct keybuf_key, node);
2477
2478         while (w && w->private)
2479                 w = RB_NEXT(w, node);
2480
2481         if (w)
2482                 w->private = ERR_PTR(-EINTR);
2483
2484         spin_unlock(&buf->lock);
2485         return w;
2486 }
2487
2488 struct keybuf_key *bch_keybuf_next_rescan(struct cache_set *c,
2489                                           struct keybuf *buf,
2490                                           struct bkey *end,
2491                                           keybuf_pred_fn *pred)
2492 {
2493         struct keybuf_key *ret;
2494
2495         while (1) {
2496                 ret = bch_keybuf_next(buf);
2497                 if (ret)
2498                         break;
2499
2500                 if (bkey_cmp(&buf->last_scanned, end) >= 0) {
2501                         pr_debug("scan finished");
2502                         break;
2503                 }
2504
2505                 bch_refill_keybuf(c, buf, end, pred);
2506         }
2507
2508         return ret;
2509 }
2510
2511 void bch_keybuf_init(struct keybuf *buf)
2512 {
2513         buf->last_scanned       = MAX_KEY;
2514         buf->keys               = RB_ROOT;
2515
2516         spin_lock_init(&buf->lock);
2517         array_allocator_init(&buf->freelist);
2518 }