4 * The contents of this file are subject to the terms of the
5 * Common Development and Distribution License (the "License").
6 * You may not use this file except in compliance with the License.
8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
9 * or http://www.opensolaris.org/os/licensing.
10 * See the License for the specific language governing permissions
11 * and limitations under the License.
13 * When distributing Covered Code, include this CDDL HEADER in each
14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
15 * If applicable, add the following below this CDDL HEADER, with the
16 * fields enclosed by brackets "[]" replaced with your own identifying
17 * information: Portions Copyright [yyyy] [name of copyright owner]
22 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
23 * Copyright (c) 2018, Joyent, Inc.
24 * Copyright (c) 2011, 2018 by Delphix. All rights reserved.
25 * Copyright (c) 2014 by Saso Kiselkov. All rights reserved.
26 * Copyright 2017 Nexenta Systems, Inc. All rights reserved.
30 * DVA-based Adjustable Replacement Cache
32 * While much of the theory of operation used here is
33 * based on the self-tuning, low overhead replacement cache
34 * presented by Megiddo and Modha at FAST 2003, there are some
35 * significant differences:
37 * 1. The Megiddo and Modha model assumes any page is evictable.
38 * Pages in its cache cannot be "locked" into memory. This makes
39 * the eviction algorithm simple: evict the last page in the list.
40 * This also make the performance characteristics easy to reason
41 * about. Our cache is not so simple. At any given moment, some
42 * subset of the blocks in the cache are un-evictable because we
43 * have handed out a reference to them. Blocks are only evictable
44 * when there are no external references active. This makes
45 * eviction far more problematic: we choose to evict the evictable
46 * blocks that are the "lowest" in the list.
48 * There are times when it is not possible to evict the requested
49 * space. In these circumstances we are unable to adjust the cache
50 * size. To prevent the cache growing unbounded at these times we
51 * implement a "cache throttle" that slows the flow of new data
52 * into the cache until we can make space available.
54 * 2. The Megiddo and Modha model assumes a fixed cache size.
55 * Pages are evicted when the cache is full and there is a cache
56 * miss. Our model has a variable sized cache. It grows with
57 * high use, but also tries to react to memory pressure from the
58 * operating system: decreasing its size when system memory is
61 * 3. The Megiddo and Modha model assumes a fixed page size. All
62 * elements of the cache are therefore exactly the same size. So
63 * when adjusting the cache size following a cache miss, its simply
64 * a matter of choosing a single page to evict. In our model, we
65 * have variable sized cache blocks (rangeing from 512 bytes to
66 * 128K bytes). We therefore choose a set of blocks to evict to make
67 * space for a cache miss that approximates as closely as possible
68 * the space used by the new block.
70 * See also: "ARC: A Self-Tuning, Low Overhead Replacement Cache"
71 * by N. Megiddo & D. Modha, FAST 2003
77 * A new reference to a cache buffer can be obtained in two
78 * ways: 1) via a hash table lookup using the DVA as a key,
79 * or 2) via one of the ARC lists. The arc_read() interface
80 * uses method 1, while the internal ARC algorithms for
81 * adjusting the cache use method 2. We therefore provide two
82 * types of locks: 1) the hash table lock array, and 2) the
85 * Buffers do not have their own mutexes, rather they rely on the
86 * hash table mutexes for the bulk of their protection (i.e. most
87 * fields in the arc_buf_hdr_t are protected by these mutexes).
89 * buf_hash_find() returns the appropriate mutex (held) when it
90 * locates the requested buffer in the hash table. It returns
91 * NULL for the mutex if the buffer was not in the table.
93 * buf_hash_remove() expects the appropriate hash mutex to be
94 * already held before it is invoked.
96 * Each ARC state also has a mutex which is used to protect the
97 * buffer list associated with the state. When attempting to
98 * obtain a hash table lock while holding an ARC list lock you
99 * must use: mutex_tryenter() to avoid deadlock. Also note that
100 * the active state mutex must be held before the ghost state mutex.
102 * It as also possible to register a callback which is run when the
103 * arc_meta_limit is reached and no buffers can be safely evicted. In
104 * this case the arc user should drop a reference on some arc buffers so
105 * they can be reclaimed and the arc_meta_limit honored. For example,
106 * when using the ZPL each dentry holds a references on a znode. These
107 * dentries must be pruned before the arc buffer holding the znode can
110 * Note that the majority of the performance stats are manipulated
111 * with atomic operations.
113 * The L2ARC uses the l2ad_mtx on each vdev for the following:
115 * - L2ARC buflist creation
116 * - L2ARC buflist eviction
117 * - L2ARC write completion, which walks L2ARC buflists
118 * - ARC header destruction, as it removes from L2ARC buflists
119 * - ARC header release, as it removes from L2ARC buflists
125 * Every block that is in the ARC is tracked by an arc_buf_hdr_t structure.
126 * This structure can point either to a block that is still in the cache or to
127 * one that is only accessible in an L2 ARC device, or it can provide
128 * information about a block that was recently evicted. If a block is
129 * only accessible in the L2ARC, then the arc_buf_hdr_t only has enough
130 * information to retrieve it from the L2ARC device. This information is
131 * stored in the l2arc_buf_hdr_t sub-structure of the arc_buf_hdr_t. A block
132 * that is in this state cannot access the data directly.
134 * Blocks that are actively being referenced or have not been evicted
135 * are cached in the L1ARC. The L1ARC (l1arc_buf_hdr_t) is a structure within
136 * the arc_buf_hdr_t that will point to the data block in memory. A block can
137 * only be read by a consumer if it has an l1arc_buf_hdr_t. The L1ARC
138 * caches data in two ways -- in a list of ARC buffers (arc_buf_t) and
139 * also in the arc_buf_hdr_t's private physical data block pointer (b_pabd).
141 * The L1ARC's data pointer may or may not be uncompressed. The ARC has the
142 * ability to store the physical data (b_pabd) associated with the DVA of the
143 * arc_buf_hdr_t. Since the b_pabd is a copy of the on-disk physical block,
144 * it will match its on-disk compression characteristics. This behavior can be
145 * disabled by setting 'zfs_compressed_arc_enabled' to B_FALSE. When the
146 * compressed ARC functionality is disabled, the b_pabd will point to an
147 * uncompressed version of the on-disk data.
149 * Data in the L1ARC is not accessed by consumers of the ARC directly. Each
150 * arc_buf_hdr_t can have multiple ARC buffers (arc_buf_t) which reference it.
151 * Each ARC buffer (arc_buf_t) is being actively accessed by a specific ARC
152 * consumer. The ARC will provide references to this data and will keep it
153 * cached until it is no longer in use. The ARC caches only the L1ARC's physical
154 * data block and will evict any arc_buf_t that is no longer referenced. The
155 * amount of memory consumed by the arc_buf_ts' data buffers can be seen via the
156 * "overhead_size" kstat.
158 * Depending on the consumer, an arc_buf_t can be requested in uncompressed or
159 * compressed form. The typical case is that consumers will want uncompressed
160 * data, and when that happens a new data buffer is allocated where the data is
161 * decompressed for them to use. Currently the only consumer who wants
162 * compressed arc_buf_t's is "zfs send", when it streams data exactly as it
163 * exists on disk. When this happens, the arc_buf_t's data buffer is shared
164 * with the arc_buf_hdr_t.
166 * Here is a diagram showing an arc_buf_hdr_t referenced by two arc_buf_t's. The
167 * first one is owned by a compressed send consumer (and therefore references
168 * the same compressed data buffer as the arc_buf_hdr_t) and the second could be
169 * used by any other consumer (and has its own uncompressed copy of the data
184 * | b_buf +------------>+-----------+ arc_buf_t
185 * | b_pabd +-+ |b_next +---->+-----------+
186 * +-----------+ | |-----------| |b_next +-->NULL
187 * | |b_comp = T | +-----------+
188 * | |b_data +-+ |b_comp = F |
189 * | +-----------+ | |b_data +-+
190 * +->+------+ | +-----------+ |
192 * data | |<--------------+ | uncompressed
193 * +------+ compressed, | data
194 * shared +-->+------+
199 * When a consumer reads a block, the ARC must first look to see if the
200 * arc_buf_hdr_t is cached. If the hdr is cached then the ARC allocates a new
201 * arc_buf_t and either copies uncompressed data into a new data buffer from an
202 * existing uncompressed arc_buf_t, decompresses the hdr's b_pabd buffer into a
203 * new data buffer, or shares the hdr's b_pabd buffer, depending on whether the
204 * hdr is compressed and the desired compression characteristics of the
205 * arc_buf_t consumer. If the arc_buf_t ends up sharing data with the
206 * arc_buf_hdr_t and both of them are uncompressed then the arc_buf_t must be
207 * the last buffer in the hdr's b_buf list, however a shared compressed buf can
208 * be anywhere in the hdr's list.
210 * The diagram below shows an example of an uncompressed ARC hdr that is
211 * sharing its data with an arc_buf_t (note that the shared uncompressed buf is
212 * the last element in the buf list):
224 * | | arc_buf_t (shared)
225 * | b_buf +------------>+---------+ arc_buf_t
226 * | | |b_next +---->+---------+
227 * | b_pabd +-+ |---------| |b_next +-->NULL
228 * +-----------+ | | | +---------+
230 * | +---------+ | |b_data +-+
231 * +->+------+ | +---------+ |
233 * uncompressed | | | |
236 * | uncompressed | | |
239 * +---------------------------------+
241 * Writing to the ARC requires that the ARC first discard the hdr's b_pabd
242 * since the physical block is about to be rewritten. The new data contents
243 * will be contained in the arc_buf_t. As the I/O pipeline performs the write,
244 * it may compress the data before writing it to disk. The ARC will be called
245 * with the transformed data and will bcopy the transformed on-disk block into
246 * a newly allocated b_pabd. Writes are always done into buffers which have
247 * either been loaned (and hence are new and don't have other readers) or
248 * buffers which have been released (and hence have their own hdr, if there
249 * were originally other readers of the buf's original hdr). This ensures that
250 * the ARC only needs to update a single buf and its hdr after a write occurs.
252 * When the L2ARC is in use, it will also take advantage of the b_pabd. The
253 * L2ARC will always write the contents of b_pabd to the L2ARC. This means
254 * that when compressed ARC is enabled that the L2ARC blocks are identical
255 * to the on-disk block in the main data pool. This provides a significant
256 * advantage since the ARC can leverage the bp's checksum when reading from the
257 * L2ARC to determine if the contents are valid. However, if the compressed
258 * ARC is disabled, then the L2ARC's block must be transformed to look
259 * like the physical block in the main data pool before comparing the
260 * checksum and determining its validity.
262 * The L1ARC has a slightly different system for storing encrypted data.
263 * Raw (encrypted + possibly compressed) data has a few subtle differences from
264 * data that is just compressed. The biggest difference is that it is not
265 * possible to decrypt encrypted data (or visa versa) if the keys aren't loaded.
266 * The other difference is that encryption cannot be treated as a suggestion.
267 * If a caller would prefer compressed data, but they actually wind up with
268 * uncompressed data the worst thing that could happen is there might be a
269 * performance hit. If the caller requests encrypted data, however, we must be
270 * sure they actually get it or else secret information could be leaked. Raw
271 * data is stored in hdr->b_crypt_hdr.b_rabd. An encrypted header, therefore,
272 * may have both an encrypted version and a decrypted version of its data at
273 * once. When a caller needs a raw arc_buf_t, it is allocated and the data is
274 * copied out of this header. To avoid complications with b_pabd, raw buffers
280 #include <sys/spa_impl.h>
281 #include <sys/zio_compress.h>
282 #include <sys/zio_checksum.h>
283 #include <sys/zfs_context.h>
285 #include <sys/refcount.h>
286 #include <sys/vdev.h>
287 #include <sys/vdev_impl.h>
288 #include <sys/dsl_pool.h>
289 #include <sys/zio_checksum.h>
290 #include <sys/multilist.h>
293 #include <sys/fm/fs/zfs.h>
295 #include <sys/shrinker.h>
296 #include <sys/vmsystm.h>
298 #include <linux/page_compat.h>
300 #include <sys/callb.h>
301 #include <sys/kstat.h>
302 #include <sys/zthr.h>
303 #include <zfs_fletcher.h>
304 #include <sys/arc_impl.h>
305 #include <sys/trace_arc.h>
306 #include <sys/aggsum.h>
307 #include <sys/cityhash.h>
310 /* set with ZFS_DEBUG=watch, to enable watchpoints on frozen buffers */
311 boolean_t arc_watch
= B_FALSE
;
315 * This thread's job is to keep enough free memory in the system, by
316 * calling arc_kmem_reap_soon() plus arc_reduce_target_size(), which improves
317 * arc_available_memory().
319 static zthr_t
*arc_reap_zthr
;
322 * This thread's job is to keep arc_size under arc_c, by calling
323 * arc_adjust(), which improves arc_is_overflowing().
325 static zthr_t
*arc_adjust_zthr
;
327 static kmutex_t arc_adjust_lock
;
328 static kcondvar_t arc_adjust_waiters_cv
;
329 static boolean_t arc_adjust_needed
= B_FALSE
;
332 * The number of headers to evict in arc_evict_state_impl() before
333 * dropping the sublist lock and evicting from another sublist. A lower
334 * value means we're more likely to evict the "correct" header (i.e. the
335 * oldest header in the arc state), but comes with higher overhead
336 * (i.e. more invocations of arc_evict_state_impl()).
338 int zfs_arc_evict_batch_limit
= 10;
340 /* number of seconds before growing cache again */
341 static int arc_grow_retry
= 5;
344 * Minimum time between calls to arc_kmem_reap_soon().
346 int arc_kmem_cache_reap_retry_ms
= 1000;
348 /* shift of arc_c for calculating overflow limit in arc_get_data_impl */
349 int zfs_arc_overflow_shift
= 8;
351 /* shift of arc_c for calculating both min and max arc_p */
352 int arc_p_min_shift
= 4;
354 /* log2(fraction of arc to reclaim) */
355 static int arc_shrink_shift
= 7;
357 /* percent of pagecache to reclaim arc to */
359 static uint_t zfs_arc_pc_percent
= 0;
363 * log2(fraction of ARC which must be free to allow growing).
364 * I.e. If there is less than arc_c >> arc_no_grow_shift free memory,
365 * when reading a new block into the ARC, we will evict an equal-sized block
368 * This must be less than arc_shrink_shift, so that when we shrink the ARC,
369 * we will still not allow it to grow.
371 int arc_no_grow_shift
= 5;
375 * minimum lifespan of a prefetch block in clock ticks
376 * (initialized in arc_init())
378 static int arc_min_prefetch_ms
;
379 static int arc_min_prescient_prefetch_ms
;
382 * If this percent of memory is free, don't throttle.
384 int arc_lotsfree_percent
= 10;
387 * hdr_recl() uses this to determine if the arc is up and running.
389 static boolean_t arc_initialized
;
392 * The arc has filled available memory and has now warmed up.
394 static boolean_t arc_warm
;
397 * log2 fraction of the zio arena to keep free.
399 int arc_zio_arena_free_shift
= 2;
402 * These tunables are for performance analysis.
404 unsigned long zfs_arc_max
= 0;
405 unsigned long zfs_arc_min
= 0;
406 unsigned long zfs_arc_meta_limit
= 0;
407 unsigned long zfs_arc_meta_min
= 0;
408 unsigned long zfs_arc_dnode_limit
= 0;
409 unsigned long zfs_arc_dnode_reduce_percent
= 10;
410 int zfs_arc_grow_retry
= 0;
411 int zfs_arc_shrink_shift
= 0;
412 int zfs_arc_p_min_shift
= 0;
413 int zfs_arc_average_blocksize
= 8 * 1024; /* 8KB */
416 * ARC dirty data constraints for arc_tempreserve_space() throttle.
418 unsigned long zfs_arc_dirty_limit_percent
= 50; /* total dirty data limit */
419 unsigned long zfs_arc_anon_limit_percent
= 25; /* anon block dirty limit */
420 unsigned long zfs_arc_pool_dirty_percent
= 20; /* each pool's anon allowance */
423 * Enable or disable compressed arc buffers.
425 int zfs_compressed_arc_enabled
= B_TRUE
;
428 * ARC will evict meta buffers that exceed arc_meta_limit. This
429 * tunable make arc_meta_limit adjustable for different workloads.
431 unsigned long zfs_arc_meta_limit_percent
= 75;
434 * Percentage that can be consumed by dnodes of ARC meta buffers.
436 unsigned long zfs_arc_dnode_limit_percent
= 10;
439 * These tunables are Linux specific
441 unsigned long zfs_arc_sys_free
= 0;
442 int zfs_arc_min_prefetch_ms
= 0;
443 int zfs_arc_min_prescient_prefetch_ms
= 0;
444 int zfs_arc_p_dampener_disable
= 1;
445 int zfs_arc_meta_prune
= 10000;
446 int zfs_arc_meta_strategy
= ARC_STRATEGY_META_BALANCED
;
447 int zfs_arc_meta_adjust_restarts
= 4096;
448 int zfs_arc_lotsfree_percent
= 10;
451 static arc_state_t ARC_anon
;
452 static arc_state_t ARC_mru
;
453 static arc_state_t ARC_mru_ghost
;
454 static arc_state_t ARC_mfu
;
455 static arc_state_t ARC_mfu_ghost
;
456 static arc_state_t ARC_l2c_only
;
458 typedef struct arc_stats
{
459 kstat_named_t arcstat_hits
;
460 kstat_named_t arcstat_misses
;
461 kstat_named_t arcstat_demand_data_hits
;
462 kstat_named_t arcstat_demand_data_misses
;
463 kstat_named_t arcstat_demand_metadata_hits
;
464 kstat_named_t arcstat_demand_metadata_misses
;
465 kstat_named_t arcstat_prefetch_data_hits
;
466 kstat_named_t arcstat_prefetch_data_misses
;
467 kstat_named_t arcstat_prefetch_metadata_hits
;
468 kstat_named_t arcstat_prefetch_metadata_misses
;
469 kstat_named_t arcstat_mru_hits
;
470 kstat_named_t arcstat_mru_ghost_hits
;
471 kstat_named_t arcstat_mfu_hits
;
472 kstat_named_t arcstat_mfu_ghost_hits
;
473 kstat_named_t arcstat_deleted
;
475 * Number of buffers that could not be evicted because the hash lock
476 * was held by another thread. The lock may not necessarily be held
477 * by something using the same buffer, since hash locks are shared
478 * by multiple buffers.
480 kstat_named_t arcstat_mutex_miss
;
482 * Number of buffers skipped when updating the access state due to the
483 * header having already been released after acquiring the hash lock.
485 kstat_named_t arcstat_access_skip
;
487 * Number of buffers skipped because they have I/O in progress, are
488 * indirect prefetch buffers that have not lived long enough, or are
489 * not from the spa we're trying to evict from.
491 kstat_named_t arcstat_evict_skip
;
493 * Number of times arc_evict_state() was unable to evict enough
494 * buffers to reach its target amount.
496 kstat_named_t arcstat_evict_not_enough
;
497 kstat_named_t arcstat_evict_l2_cached
;
498 kstat_named_t arcstat_evict_l2_eligible
;
499 kstat_named_t arcstat_evict_l2_ineligible
;
500 kstat_named_t arcstat_evict_l2_skip
;
501 kstat_named_t arcstat_hash_elements
;
502 kstat_named_t arcstat_hash_elements_max
;
503 kstat_named_t arcstat_hash_collisions
;
504 kstat_named_t arcstat_hash_chains
;
505 kstat_named_t arcstat_hash_chain_max
;
506 kstat_named_t arcstat_p
;
507 kstat_named_t arcstat_c
;
508 kstat_named_t arcstat_c_min
;
509 kstat_named_t arcstat_c_max
;
510 /* Not updated directly; only synced in arc_kstat_update. */
511 kstat_named_t arcstat_size
;
513 * Number of compressed bytes stored in the arc_buf_hdr_t's b_pabd.
514 * Note that the compressed bytes may match the uncompressed bytes
515 * if the block is either not compressed or compressed arc is disabled.
517 kstat_named_t arcstat_compressed_size
;
519 * Uncompressed size of the data stored in b_pabd. If compressed
520 * arc is disabled then this value will be identical to the stat
523 kstat_named_t arcstat_uncompressed_size
;
525 * Number of bytes stored in all the arc_buf_t's. This is classified
526 * as "overhead" since this data is typically short-lived and will
527 * be evicted from the arc when it becomes unreferenced unless the
528 * zfs_keep_uncompressed_metadata or zfs_keep_uncompressed_level
529 * values have been set (see comment in dbuf.c for more information).
531 kstat_named_t arcstat_overhead_size
;
533 * Number of bytes consumed by internal ARC structures necessary
534 * for tracking purposes; these structures are not actually
535 * backed by ARC buffers. This includes arc_buf_hdr_t structures
536 * (allocated via arc_buf_hdr_t_full and arc_buf_hdr_t_l2only
537 * caches), and arc_buf_t structures (allocated via arc_buf_t
539 * Not updated directly; only synced in arc_kstat_update.
541 kstat_named_t arcstat_hdr_size
;
543 * Number of bytes consumed by ARC buffers of type equal to
544 * ARC_BUFC_DATA. This is generally consumed by buffers backing
545 * on disk user data (e.g. plain file contents).
546 * Not updated directly; only synced in arc_kstat_update.
548 kstat_named_t arcstat_data_size
;
550 * Number of bytes consumed by ARC buffers of type equal to
551 * ARC_BUFC_METADATA. This is generally consumed by buffers
552 * backing on disk data that is used for internal ZFS
553 * structures (e.g. ZAP, dnode, indirect blocks, etc).
554 * Not updated directly; only synced in arc_kstat_update.
556 kstat_named_t arcstat_metadata_size
;
558 * Number of bytes consumed by dmu_buf_impl_t objects.
559 * Not updated directly; only synced in arc_kstat_update.
561 kstat_named_t arcstat_dbuf_size
;
563 * Number of bytes consumed by dnode_t objects.
564 * Not updated directly; only synced in arc_kstat_update.
566 kstat_named_t arcstat_dnode_size
;
568 * Number of bytes consumed by bonus buffers.
569 * Not updated directly; only synced in arc_kstat_update.
571 kstat_named_t arcstat_bonus_size
;
573 * Total number of bytes consumed by ARC buffers residing in the
574 * arc_anon state. This includes *all* buffers in the arc_anon
575 * state; e.g. data, metadata, evictable, and unevictable buffers
576 * are all included in this value.
577 * Not updated directly; only synced in arc_kstat_update.
579 kstat_named_t arcstat_anon_size
;
581 * Number of bytes consumed by ARC buffers that meet the
582 * following criteria: backing buffers of type ARC_BUFC_DATA,
583 * residing in the arc_anon state, and are eligible for eviction
584 * (e.g. have no outstanding holds on the buffer).
585 * Not updated directly; only synced in arc_kstat_update.
587 kstat_named_t arcstat_anon_evictable_data
;
589 * Number of bytes consumed by ARC buffers that meet the
590 * following criteria: backing buffers of type ARC_BUFC_METADATA,
591 * residing in the arc_anon state, and are eligible for eviction
592 * (e.g. have no outstanding holds on the buffer).
593 * Not updated directly; only synced in arc_kstat_update.
595 kstat_named_t arcstat_anon_evictable_metadata
;
597 * Total number of bytes consumed by ARC buffers residing in the
598 * arc_mru state. This includes *all* buffers in the arc_mru
599 * state; e.g. data, metadata, evictable, and unevictable buffers
600 * are all included in this value.
601 * Not updated directly; only synced in arc_kstat_update.
603 kstat_named_t arcstat_mru_size
;
605 * Number of bytes consumed by ARC buffers that meet the
606 * following criteria: backing buffers of type ARC_BUFC_DATA,
607 * residing in the arc_mru state, and are eligible for eviction
608 * (e.g. have no outstanding holds on the buffer).
609 * Not updated directly; only synced in arc_kstat_update.
611 kstat_named_t arcstat_mru_evictable_data
;
613 * Number of bytes consumed by ARC buffers that meet the
614 * following criteria: backing buffers of type ARC_BUFC_METADATA,
615 * residing in the arc_mru state, and are eligible for eviction
616 * (e.g. have no outstanding holds on the buffer).
617 * Not updated directly; only synced in arc_kstat_update.
619 kstat_named_t arcstat_mru_evictable_metadata
;
621 * Total number of bytes that *would have been* consumed by ARC
622 * buffers in the arc_mru_ghost state. The key thing to note
623 * here, is the fact that this size doesn't actually indicate
624 * RAM consumption. The ghost lists only consist of headers and
625 * don't actually have ARC buffers linked off of these headers.
626 * Thus, *if* the headers had associated ARC buffers, these
627 * buffers *would have* consumed this number of bytes.
628 * Not updated directly; only synced in arc_kstat_update.
630 kstat_named_t arcstat_mru_ghost_size
;
632 * Number of bytes that *would have been* consumed by ARC
633 * buffers that are eligible for eviction, of type
634 * ARC_BUFC_DATA, and linked off the arc_mru_ghost state.
635 * Not updated directly; only synced in arc_kstat_update.
637 kstat_named_t arcstat_mru_ghost_evictable_data
;
639 * Number of bytes that *would have been* consumed by ARC
640 * buffers that are eligible for eviction, of type
641 * ARC_BUFC_METADATA, and linked off the arc_mru_ghost state.
642 * Not updated directly; only synced in arc_kstat_update.
644 kstat_named_t arcstat_mru_ghost_evictable_metadata
;
646 * Total number of bytes consumed by ARC buffers residing in the
647 * arc_mfu state. This includes *all* buffers in the arc_mfu
648 * state; e.g. data, metadata, evictable, and unevictable buffers
649 * are all included in this value.
650 * Not updated directly; only synced in arc_kstat_update.
652 kstat_named_t arcstat_mfu_size
;
654 * Number of bytes consumed by ARC buffers that are eligible for
655 * eviction, of type ARC_BUFC_DATA, and reside in the arc_mfu
657 * Not updated directly; only synced in arc_kstat_update.
659 kstat_named_t arcstat_mfu_evictable_data
;
661 * Number of bytes consumed by ARC buffers that are eligible for
662 * eviction, of type ARC_BUFC_METADATA, and reside in the
664 * Not updated directly; only synced in arc_kstat_update.
666 kstat_named_t arcstat_mfu_evictable_metadata
;
668 * Total number of bytes that *would have been* consumed by ARC
669 * buffers in the arc_mfu_ghost state. See the comment above
670 * arcstat_mru_ghost_size for more details.
671 * Not updated directly; only synced in arc_kstat_update.
673 kstat_named_t arcstat_mfu_ghost_size
;
675 * Number of bytes that *would have been* consumed by ARC
676 * buffers that are eligible for eviction, of type
677 * ARC_BUFC_DATA, and linked off the arc_mfu_ghost state.
678 * Not updated directly; only synced in arc_kstat_update.
680 kstat_named_t arcstat_mfu_ghost_evictable_data
;
682 * Number of bytes that *would have been* consumed by ARC
683 * buffers that are eligible for eviction, of type
684 * ARC_BUFC_METADATA, and linked off the arc_mru_ghost state.
685 * Not updated directly; only synced in arc_kstat_update.
687 kstat_named_t arcstat_mfu_ghost_evictable_metadata
;
688 kstat_named_t arcstat_l2_hits
;
689 kstat_named_t arcstat_l2_misses
;
690 kstat_named_t arcstat_l2_feeds
;
691 kstat_named_t arcstat_l2_rw_clash
;
692 kstat_named_t arcstat_l2_read_bytes
;
693 kstat_named_t arcstat_l2_write_bytes
;
694 kstat_named_t arcstat_l2_writes_sent
;
695 kstat_named_t arcstat_l2_writes_done
;
696 kstat_named_t arcstat_l2_writes_error
;
697 kstat_named_t arcstat_l2_writes_lock_retry
;
698 kstat_named_t arcstat_l2_evict_lock_retry
;
699 kstat_named_t arcstat_l2_evict_reading
;
700 kstat_named_t arcstat_l2_evict_l1cached
;
701 kstat_named_t arcstat_l2_free_on_write
;
702 kstat_named_t arcstat_l2_abort_lowmem
;
703 kstat_named_t arcstat_l2_cksum_bad
;
704 kstat_named_t arcstat_l2_io_error
;
705 kstat_named_t arcstat_l2_lsize
;
706 kstat_named_t arcstat_l2_psize
;
707 /* Not updated directly; only synced in arc_kstat_update. */
708 kstat_named_t arcstat_l2_hdr_size
;
709 kstat_named_t arcstat_memory_throttle_count
;
710 kstat_named_t arcstat_memory_direct_count
;
711 kstat_named_t arcstat_memory_indirect_count
;
712 kstat_named_t arcstat_memory_all_bytes
;
713 kstat_named_t arcstat_memory_free_bytes
;
714 kstat_named_t arcstat_memory_available_bytes
;
715 kstat_named_t arcstat_no_grow
;
716 kstat_named_t arcstat_tempreserve
;
717 kstat_named_t arcstat_loaned_bytes
;
718 kstat_named_t arcstat_prune
;
719 /* Not updated directly; only synced in arc_kstat_update. */
720 kstat_named_t arcstat_meta_used
;
721 kstat_named_t arcstat_meta_limit
;
722 kstat_named_t arcstat_dnode_limit
;
723 kstat_named_t arcstat_meta_max
;
724 kstat_named_t arcstat_meta_min
;
725 kstat_named_t arcstat_async_upgrade_sync
;
726 kstat_named_t arcstat_demand_hit_predictive_prefetch
;
727 kstat_named_t arcstat_demand_hit_prescient_prefetch
;
728 kstat_named_t arcstat_need_free
;
729 kstat_named_t arcstat_sys_free
;
730 kstat_named_t arcstat_raw_size
;
733 static arc_stats_t arc_stats
= {
734 { "hits", KSTAT_DATA_UINT64
},
735 { "misses", KSTAT_DATA_UINT64
},
736 { "demand_data_hits", KSTAT_DATA_UINT64
},
737 { "demand_data_misses", KSTAT_DATA_UINT64
},
738 { "demand_metadata_hits", KSTAT_DATA_UINT64
},
739 { "demand_metadata_misses", KSTAT_DATA_UINT64
},
740 { "prefetch_data_hits", KSTAT_DATA_UINT64
},
741 { "prefetch_data_misses", KSTAT_DATA_UINT64
},
742 { "prefetch_metadata_hits", KSTAT_DATA_UINT64
},
743 { "prefetch_metadata_misses", KSTAT_DATA_UINT64
},
744 { "mru_hits", KSTAT_DATA_UINT64
},
745 { "mru_ghost_hits", KSTAT_DATA_UINT64
},
746 { "mfu_hits", KSTAT_DATA_UINT64
},
747 { "mfu_ghost_hits", KSTAT_DATA_UINT64
},
748 { "deleted", KSTAT_DATA_UINT64
},
749 { "mutex_miss", KSTAT_DATA_UINT64
},
750 { "access_skip", KSTAT_DATA_UINT64
},
751 { "evict_skip", KSTAT_DATA_UINT64
},
752 { "evict_not_enough", KSTAT_DATA_UINT64
},
753 { "evict_l2_cached", KSTAT_DATA_UINT64
},
754 { "evict_l2_eligible", KSTAT_DATA_UINT64
},
755 { "evict_l2_ineligible", KSTAT_DATA_UINT64
},
756 { "evict_l2_skip", KSTAT_DATA_UINT64
},
757 { "hash_elements", KSTAT_DATA_UINT64
},
758 { "hash_elements_max", KSTAT_DATA_UINT64
},
759 { "hash_collisions", KSTAT_DATA_UINT64
},
760 { "hash_chains", KSTAT_DATA_UINT64
},
761 { "hash_chain_max", KSTAT_DATA_UINT64
},
762 { "p", KSTAT_DATA_UINT64
},
763 { "c", KSTAT_DATA_UINT64
},
764 { "c_min", KSTAT_DATA_UINT64
},
765 { "c_max", KSTAT_DATA_UINT64
},
766 { "size", KSTAT_DATA_UINT64
},
767 { "compressed_size", KSTAT_DATA_UINT64
},
768 { "uncompressed_size", KSTAT_DATA_UINT64
},
769 { "overhead_size", KSTAT_DATA_UINT64
},
770 { "hdr_size", KSTAT_DATA_UINT64
},
771 { "data_size", KSTAT_DATA_UINT64
},
772 { "metadata_size", KSTAT_DATA_UINT64
},
773 { "dbuf_size", KSTAT_DATA_UINT64
},
774 { "dnode_size", KSTAT_DATA_UINT64
},
775 { "bonus_size", KSTAT_DATA_UINT64
},
776 { "anon_size", KSTAT_DATA_UINT64
},
777 { "anon_evictable_data", KSTAT_DATA_UINT64
},
778 { "anon_evictable_metadata", KSTAT_DATA_UINT64
},
779 { "mru_size", KSTAT_DATA_UINT64
},
780 { "mru_evictable_data", KSTAT_DATA_UINT64
},
781 { "mru_evictable_metadata", KSTAT_DATA_UINT64
},
782 { "mru_ghost_size", KSTAT_DATA_UINT64
},
783 { "mru_ghost_evictable_data", KSTAT_DATA_UINT64
},
784 { "mru_ghost_evictable_metadata", KSTAT_DATA_UINT64
},
785 { "mfu_size", KSTAT_DATA_UINT64
},
786 { "mfu_evictable_data", KSTAT_DATA_UINT64
},
787 { "mfu_evictable_metadata", KSTAT_DATA_UINT64
},
788 { "mfu_ghost_size", KSTAT_DATA_UINT64
},
789 { "mfu_ghost_evictable_data", KSTAT_DATA_UINT64
},
790 { "mfu_ghost_evictable_metadata", KSTAT_DATA_UINT64
},
791 { "l2_hits", KSTAT_DATA_UINT64
},
792 { "l2_misses", KSTAT_DATA_UINT64
},
793 { "l2_feeds", KSTAT_DATA_UINT64
},
794 { "l2_rw_clash", KSTAT_DATA_UINT64
},
795 { "l2_read_bytes", KSTAT_DATA_UINT64
},
796 { "l2_write_bytes", KSTAT_DATA_UINT64
},
797 { "l2_writes_sent", KSTAT_DATA_UINT64
},
798 { "l2_writes_done", KSTAT_DATA_UINT64
},
799 { "l2_writes_error", KSTAT_DATA_UINT64
},
800 { "l2_writes_lock_retry", KSTAT_DATA_UINT64
},
801 { "l2_evict_lock_retry", KSTAT_DATA_UINT64
},
802 { "l2_evict_reading", KSTAT_DATA_UINT64
},
803 { "l2_evict_l1cached", KSTAT_DATA_UINT64
},
804 { "l2_free_on_write", KSTAT_DATA_UINT64
},
805 { "l2_abort_lowmem", KSTAT_DATA_UINT64
},
806 { "l2_cksum_bad", KSTAT_DATA_UINT64
},
807 { "l2_io_error", KSTAT_DATA_UINT64
},
808 { "l2_size", KSTAT_DATA_UINT64
},
809 { "l2_asize", KSTAT_DATA_UINT64
},
810 { "l2_hdr_size", KSTAT_DATA_UINT64
},
811 { "memory_throttle_count", KSTAT_DATA_UINT64
},
812 { "memory_direct_count", KSTAT_DATA_UINT64
},
813 { "memory_indirect_count", KSTAT_DATA_UINT64
},
814 { "memory_all_bytes", KSTAT_DATA_UINT64
},
815 { "memory_free_bytes", KSTAT_DATA_UINT64
},
816 { "memory_available_bytes", KSTAT_DATA_INT64
},
817 { "arc_no_grow", KSTAT_DATA_UINT64
},
818 { "arc_tempreserve", KSTAT_DATA_UINT64
},
819 { "arc_loaned_bytes", KSTAT_DATA_UINT64
},
820 { "arc_prune", KSTAT_DATA_UINT64
},
821 { "arc_meta_used", KSTAT_DATA_UINT64
},
822 { "arc_meta_limit", KSTAT_DATA_UINT64
},
823 { "arc_dnode_limit", KSTAT_DATA_UINT64
},
824 { "arc_meta_max", KSTAT_DATA_UINT64
},
825 { "arc_meta_min", KSTAT_DATA_UINT64
},
826 { "async_upgrade_sync", KSTAT_DATA_UINT64
},
827 { "demand_hit_predictive_prefetch", KSTAT_DATA_UINT64
},
828 { "demand_hit_prescient_prefetch", KSTAT_DATA_UINT64
},
829 { "arc_need_free", KSTAT_DATA_UINT64
},
830 { "arc_sys_free", KSTAT_DATA_UINT64
},
831 { "arc_raw_size", KSTAT_DATA_UINT64
}
834 #define ARCSTAT(stat) (arc_stats.stat.value.ui64)
836 #define ARCSTAT_INCR(stat, val) \
837 atomic_add_64(&arc_stats.stat.value.ui64, (val))
839 #define ARCSTAT_BUMP(stat) ARCSTAT_INCR(stat, 1)
840 #define ARCSTAT_BUMPDOWN(stat) ARCSTAT_INCR(stat, -1)
842 #define ARCSTAT_MAX(stat, val) { \
844 while ((val) > (m = arc_stats.stat.value.ui64) && \
845 (m != atomic_cas_64(&arc_stats.stat.value.ui64, m, (val)))) \
849 #define ARCSTAT_MAXSTAT(stat) \
850 ARCSTAT_MAX(stat##_max, arc_stats.stat.value.ui64)
853 * We define a macro to allow ARC hits/misses to be easily broken down by
854 * two separate conditions, giving a total of four different subtypes for
855 * each of hits and misses (so eight statistics total).
857 #define ARCSTAT_CONDSTAT(cond1, stat1, notstat1, cond2, stat2, notstat2, stat) \
860 ARCSTAT_BUMP(arcstat_##stat1##_##stat2##_##stat); \
862 ARCSTAT_BUMP(arcstat_##stat1##_##notstat2##_##stat); \
866 ARCSTAT_BUMP(arcstat_##notstat1##_##stat2##_##stat); \
868 ARCSTAT_BUMP(arcstat_##notstat1##_##notstat2##_##stat);\
873 static arc_state_t
*arc_anon
;
874 static arc_state_t
*arc_mru
;
875 static arc_state_t
*arc_mru_ghost
;
876 static arc_state_t
*arc_mfu
;
877 static arc_state_t
*arc_mfu_ghost
;
878 static arc_state_t
*arc_l2c_only
;
881 * There are several ARC variables that are critical to export as kstats --
882 * but we don't want to have to grovel around in the kstat whenever we wish to
883 * manipulate them. For these variables, we therefore define them to be in
884 * terms of the statistic variable. This assures that we are not introducing
885 * the possibility of inconsistency by having shadow copies of the variables,
886 * while still allowing the code to be readable.
888 #define arc_p ARCSTAT(arcstat_p) /* target size of MRU */
889 #define arc_c ARCSTAT(arcstat_c) /* target size of cache */
890 #define arc_c_min ARCSTAT(arcstat_c_min) /* min target cache size */
891 #define arc_c_max ARCSTAT(arcstat_c_max) /* max target cache size */
892 #define arc_no_grow ARCSTAT(arcstat_no_grow) /* do not grow cache size */
893 #define arc_tempreserve ARCSTAT(arcstat_tempreserve)
894 #define arc_loaned_bytes ARCSTAT(arcstat_loaned_bytes)
895 #define arc_meta_limit ARCSTAT(arcstat_meta_limit) /* max size for metadata */
896 #define arc_dnode_limit ARCSTAT(arcstat_dnode_limit) /* max size for dnodes */
897 #define arc_meta_min ARCSTAT(arcstat_meta_min) /* min size for metadata */
898 #define arc_meta_max ARCSTAT(arcstat_meta_max) /* max size of metadata */
899 #define arc_need_free ARCSTAT(arcstat_need_free) /* bytes to be freed */
900 #define arc_sys_free ARCSTAT(arcstat_sys_free) /* target system free bytes */
902 /* size of all b_rabd's in entire arc */
903 #define arc_raw_size ARCSTAT(arcstat_raw_size)
904 /* compressed size of entire arc */
905 #define arc_compressed_size ARCSTAT(arcstat_compressed_size)
906 /* uncompressed size of entire arc */
907 #define arc_uncompressed_size ARCSTAT(arcstat_uncompressed_size)
908 /* number of bytes in the arc from arc_buf_t's */
909 #define arc_overhead_size ARCSTAT(arcstat_overhead_size)
912 * There are also some ARC variables that we want to export, but that are
913 * updated so often that having the canonical representation be the statistic
914 * variable causes a performance bottleneck. We want to use aggsum_t's for these
915 * instead, but still be able to export the kstat in the same way as before.
916 * The solution is to always use the aggsum version, except in the kstat update
920 aggsum_t arc_meta_used
;
921 aggsum_t astat_data_size
;
922 aggsum_t astat_metadata_size
;
923 aggsum_t astat_dbuf_size
;
924 aggsum_t astat_dnode_size
;
925 aggsum_t astat_bonus_size
;
926 aggsum_t astat_hdr_size
;
927 aggsum_t astat_l2_hdr_size
;
929 static hrtime_t arc_growtime
;
930 static list_t arc_prune_list
;
931 static kmutex_t arc_prune_mtx
;
932 static taskq_t
*arc_prune_taskq
;
934 #define GHOST_STATE(state) \
935 ((state) == arc_mru_ghost || (state) == arc_mfu_ghost || \
936 (state) == arc_l2c_only)
938 #define HDR_IN_HASH_TABLE(hdr) ((hdr)->b_flags & ARC_FLAG_IN_HASH_TABLE)
939 #define HDR_IO_IN_PROGRESS(hdr) ((hdr)->b_flags & ARC_FLAG_IO_IN_PROGRESS)
940 #define HDR_IO_ERROR(hdr) ((hdr)->b_flags & ARC_FLAG_IO_ERROR)
941 #define HDR_PREFETCH(hdr) ((hdr)->b_flags & ARC_FLAG_PREFETCH)
942 #define HDR_PRESCIENT_PREFETCH(hdr) \
943 ((hdr)->b_flags & ARC_FLAG_PRESCIENT_PREFETCH)
944 #define HDR_COMPRESSION_ENABLED(hdr) \
945 ((hdr)->b_flags & ARC_FLAG_COMPRESSED_ARC)
947 #define HDR_L2CACHE(hdr) ((hdr)->b_flags & ARC_FLAG_L2CACHE)
948 #define HDR_L2_READING(hdr) \
949 (((hdr)->b_flags & ARC_FLAG_IO_IN_PROGRESS) && \
950 ((hdr)->b_flags & ARC_FLAG_HAS_L2HDR))
951 #define HDR_L2_WRITING(hdr) ((hdr)->b_flags & ARC_FLAG_L2_WRITING)
952 #define HDR_L2_EVICTED(hdr) ((hdr)->b_flags & ARC_FLAG_L2_EVICTED)
953 #define HDR_L2_WRITE_HEAD(hdr) ((hdr)->b_flags & ARC_FLAG_L2_WRITE_HEAD)
954 #define HDR_PROTECTED(hdr) ((hdr)->b_flags & ARC_FLAG_PROTECTED)
955 #define HDR_NOAUTH(hdr) ((hdr)->b_flags & ARC_FLAG_NOAUTH)
956 #define HDR_SHARED_DATA(hdr) ((hdr)->b_flags & ARC_FLAG_SHARED_DATA)
958 #define HDR_ISTYPE_METADATA(hdr) \
959 ((hdr)->b_flags & ARC_FLAG_BUFC_METADATA)
960 #define HDR_ISTYPE_DATA(hdr) (!HDR_ISTYPE_METADATA(hdr))
962 #define HDR_HAS_L1HDR(hdr) ((hdr)->b_flags & ARC_FLAG_HAS_L1HDR)
963 #define HDR_HAS_L2HDR(hdr) ((hdr)->b_flags & ARC_FLAG_HAS_L2HDR)
964 #define HDR_HAS_RABD(hdr) \
965 (HDR_HAS_L1HDR(hdr) && HDR_PROTECTED(hdr) && \
966 (hdr)->b_crypt_hdr.b_rabd != NULL)
967 #define HDR_ENCRYPTED(hdr) \
968 (HDR_PROTECTED(hdr) && DMU_OT_IS_ENCRYPTED((hdr)->b_crypt_hdr.b_ot))
969 #define HDR_AUTHENTICATED(hdr) \
970 (HDR_PROTECTED(hdr) && !DMU_OT_IS_ENCRYPTED((hdr)->b_crypt_hdr.b_ot))
972 /* For storing compression mode in b_flags */
973 #define HDR_COMPRESS_OFFSET (highbit64(ARC_FLAG_COMPRESS_0) - 1)
975 #define HDR_GET_COMPRESS(hdr) ((enum zio_compress)BF32_GET((hdr)->b_flags, \
976 HDR_COMPRESS_OFFSET, SPA_COMPRESSBITS))
977 #define HDR_SET_COMPRESS(hdr, cmp) BF32_SET((hdr)->b_flags, \
978 HDR_COMPRESS_OFFSET, SPA_COMPRESSBITS, (cmp));
980 #define ARC_BUF_LAST(buf) ((buf)->b_next == NULL)
981 #define ARC_BUF_SHARED(buf) ((buf)->b_flags & ARC_BUF_FLAG_SHARED)
982 #define ARC_BUF_COMPRESSED(buf) ((buf)->b_flags & ARC_BUF_FLAG_COMPRESSED)
983 #define ARC_BUF_ENCRYPTED(buf) ((buf)->b_flags & ARC_BUF_FLAG_ENCRYPTED)
989 #define HDR_FULL_CRYPT_SIZE ((int64_t)sizeof (arc_buf_hdr_t))
990 #define HDR_FULL_SIZE ((int64_t)offsetof(arc_buf_hdr_t, b_crypt_hdr))
991 #define HDR_L2ONLY_SIZE ((int64_t)offsetof(arc_buf_hdr_t, b_l1hdr))
994 * Hash table routines
997 #define HT_LOCK_ALIGN 64
998 #define HT_LOCK_PAD (P2NPHASE(sizeof (kmutex_t), (HT_LOCK_ALIGN)))
1003 unsigned char pad
[HT_LOCK_PAD
];
1007 #define BUF_LOCKS 8192
1008 typedef struct buf_hash_table
{
1010 arc_buf_hdr_t
**ht_table
;
1011 struct ht_lock ht_locks
[BUF_LOCKS
];
1014 static buf_hash_table_t buf_hash_table
;
1016 #define BUF_HASH_INDEX(spa, dva, birth) \
1017 (buf_hash(spa, dva, birth) & buf_hash_table.ht_mask)
1018 #define BUF_HASH_LOCK_NTRY(idx) (buf_hash_table.ht_locks[idx & (BUF_LOCKS-1)])
1019 #define BUF_HASH_LOCK(idx) (&(BUF_HASH_LOCK_NTRY(idx).ht_lock))
1020 #define HDR_LOCK(hdr) \
1021 (BUF_HASH_LOCK(BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth)))
1023 uint64_t zfs_crc64_table
[256];
1029 #define L2ARC_WRITE_SIZE (8 * 1024 * 1024) /* initial write max */
1030 #define L2ARC_HEADROOM 2 /* num of writes */
1033 * If we discover during ARC scan any buffers to be compressed, we boost
1034 * our headroom for the next scanning cycle by this percentage multiple.
1036 #define L2ARC_HEADROOM_BOOST 200
1037 #define L2ARC_FEED_SECS 1 /* caching interval secs */
1038 #define L2ARC_FEED_MIN_MS 200 /* min caching interval ms */
1041 * We can feed L2ARC from two states of ARC buffers, mru and mfu,
1042 * and each of the state has two types: data and metadata.
1044 #define L2ARC_FEED_TYPES 4
1046 #define l2arc_writes_sent ARCSTAT(arcstat_l2_writes_sent)
1047 #define l2arc_writes_done ARCSTAT(arcstat_l2_writes_done)
1049 /* L2ARC Performance Tunables */
1050 unsigned long l2arc_write_max
= L2ARC_WRITE_SIZE
; /* def max write size */
1051 unsigned long l2arc_write_boost
= L2ARC_WRITE_SIZE
; /* extra warmup write */
1052 unsigned long l2arc_headroom
= L2ARC_HEADROOM
; /* # of dev writes */
1053 unsigned long l2arc_headroom_boost
= L2ARC_HEADROOM_BOOST
;
1054 unsigned long l2arc_feed_secs
= L2ARC_FEED_SECS
; /* interval seconds */
1055 unsigned long l2arc_feed_min_ms
= L2ARC_FEED_MIN_MS
; /* min interval msecs */
1056 int l2arc_noprefetch
= B_TRUE
; /* don't cache prefetch bufs */
1057 int l2arc_feed_again
= B_TRUE
; /* turbo warmup */
1058 int l2arc_norw
= B_FALSE
; /* no reads during writes */
1063 static list_t L2ARC_dev_list
; /* device list */
1064 static list_t
*l2arc_dev_list
; /* device list pointer */
1065 static kmutex_t l2arc_dev_mtx
; /* device list mutex */
1066 static l2arc_dev_t
*l2arc_dev_last
; /* last device used */
1067 static list_t L2ARC_free_on_write
; /* free after write buf list */
1068 static list_t
*l2arc_free_on_write
; /* free after write list ptr */
1069 static kmutex_t l2arc_free_on_write_mtx
; /* mutex for list */
1070 static uint64_t l2arc_ndev
; /* number of devices */
1072 typedef struct l2arc_read_callback
{
1073 arc_buf_hdr_t
*l2rcb_hdr
; /* read header */
1074 blkptr_t l2rcb_bp
; /* original blkptr */
1075 zbookmark_phys_t l2rcb_zb
; /* original bookmark */
1076 int l2rcb_flags
; /* original flags */
1077 abd_t
*l2rcb_abd
; /* temporary buffer */
1078 } l2arc_read_callback_t
;
1080 typedef struct l2arc_data_free
{
1081 /* protected by l2arc_free_on_write_mtx */
1084 arc_buf_contents_t l2df_type
;
1085 list_node_t l2df_list_node
;
1086 } l2arc_data_free_t
;
1088 typedef enum arc_fill_flags
{
1089 ARC_FILL_LOCKED
= 1 << 0, /* hdr lock is held */
1090 ARC_FILL_COMPRESSED
= 1 << 1, /* fill with compressed data */
1091 ARC_FILL_ENCRYPTED
= 1 << 2, /* fill with encrypted data */
1092 ARC_FILL_NOAUTH
= 1 << 3, /* don't attempt to authenticate */
1093 ARC_FILL_IN_PLACE
= 1 << 4 /* fill in place (special case) */
1096 static kmutex_t l2arc_feed_thr_lock
;
1097 static kcondvar_t l2arc_feed_thr_cv
;
1098 static uint8_t l2arc_thread_exit
;
1100 static abd_t
*arc_get_data_abd(arc_buf_hdr_t
*, uint64_t, void *);
1101 static void *arc_get_data_buf(arc_buf_hdr_t
*, uint64_t, void *);
1102 static void arc_get_data_impl(arc_buf_hdr_t
*, uint64_t, void *);
1103 static void arc_free_data_abd(arc_buf_hdr_t
*, abd_t
*, uint64_t, void *);
1104 static void arc_free_data_buf(arc_buf_hdr_t
*, void *, uint64_t, void *);
1105 static void arc_free_data_impl(arc_buf_hdr_t
*hdr
, uint64_t size
, void *tag
);
1106 static void arc_hdr_free_abd(arc_buf_hdr_t
*, boolean_t
);
1107 static void arc_hdr_alloc_abd(arc_buf_hdr_t
*, boolean_t
);
1108 static void arc_access(arc_buf_hdr_t
*, kmutex_t
*);
1109 static boolean_t
arc_is_overflowing(void);
1110 static void arc_buf_watch(arc_buf_t
*);
1111 static void arc_tuning_update(void);
1112 static void arc_prune_async(int64_t);
1113 static uint64_t arc_all_memory(void);
1115 static arc_buf_contents_t
arc_buf_type(arc_buf_hdr_t
*);
1116 static uint32_t arc_bufc_to_flags(arc_buf_contents_t
);
1117 static inline void arc_hdr_set_flags(arc_buf_hdr_t
*hdr
, arc_flags_t flags
);
1118 static inline void arc_hdr_clear_flags(arc_buf_hdr_t
*hdr
, arc_flags_t flags
);
1120 static boolean_t
l2arc_write_eligible(uint64_t, arc_buf_hdr_t
*);
1121 static void l2arc_read_done(zio_t
*);
1125 * We use Cityhash for this. It's fast, and has good hash properties without
1126 * requiring any large static buffers.
1129 buf_hash(uint64_t spa
, const dva_t
*dva
, uint64_t birth
)
1131 return (cityhash4(spa
, dva
->dva_word
[0], dva
->dva_word
[1], birth
));
1134 #define HDR_EMPTY(hdr) \
1135 ((hdr)->b_dva.dva_word[0] == 0 && \
1136 (hdr)->b_dva.dva_word[1] == 0)
1138 #define HDR_EQUAL(spa, dva, birth, hdr) \
1139 ((hdr)->b_dva.dva_word[0] == (dva)->dva_word[0]) && \
1140 ((hdr)->b_dva.dva_word[1] == (dva)->dva_word[1]) && \
1141 ((hdr)->b_birth == birth) && ((hdr)->b_spa == spa)
1144 buf_discard_identity(arc_buf_hdr_t
*hdr
)
1146 hdr
->b_dva
.dva_word
[0] = 0;
1147 hdr
->b_dva
.dva_word
[1] = 0;
1151 static arc_buf_hdr_t
*
1152 buf_hash_find(uint64_t spa
, const blkptr_t
*bp
, kmutex_t
**lockp
)
1154 const dva_t
*dva
= BP_IDENTITY(bp
);
1155 uint64_t birth
= BP_PHYSICAL_BIRTH(bp
);
1156 uint64_t idx
= BUF_HASH_INDEX(spa
, dva
, birth
);
1157 kmutex_t
*hash_lock
= BUF_HASH_LOCK(idx
);
1160 mutex_enter(hash_lock
);
1161 for (hdr
= buf_hash_table
.ht_table
[idx
]; hdr
!= NULL
;
1162 hdr
= hdr
->b_hash_next
) {
1163 if (HDR_EQUAL(spa
, dva
, birth
, hdr
)) {
1168 mutex_exit(hash_lock
);
1174 * Insert an entry into the hash table. If there is already an element
1175 * equal to elem in the hash table, then the already existing element
1176 * will be returned and the new element will not be inserted.
1177 * Otherwise returns NULL.
1178 * If lockp == NULL, the caller is assumed to already hold the hash lock.
1180 static arc_buf_hdr_t
*
1181 buf_hash_insert(arc_buf_hdr_t
*hdr
, kmutex_t
**lockp
)
1183 uint64_t idx
= BUF_HASH_INDEX(hdr
->b_spa
, &hdr
->b_dva
, hdr
->b_birth
);
1184 kmutex_t
*hash_lock
= BUF_HASH_LOCK(idx
);
1185 arc_buf_hdr_t
*fhdr
;
1188 ASSERT(!DVA_IS_EMPTY(&hdr
->b_dva
));
1189 ASSERT(hdr
->b_birth
!= 0);
1190 ASSERT(!HDR_IN_HASH_TABLE(hdr
));
1192 if (lockp
!= NULL
) {
1194 mutex_enter(hash_lock
);
1196 ASSERT(MUTEX_HELD(hash_lock
));
1199 for (fhdr
= buf_hash_table
.ht_table
[idx
], i
= 0; fhdr
!= NULL
;
1200 fhdr
= fhdr
->b_hash_next
, i
++) {
1201 if (HDR_EQUAL(hdr
->b_spa
, &hdr
->b_dva
, hdr
->b_birth
, fhdr
))
1205 hdr
->b_hash_next
= buf_hash_table
.ht_table
[idx
];
1206 buf_hash_table
.ht_table
[idx
] = hdr
;
1207 arc_hdr_set_flags(hdr
, ARC_FLAG_IN_HASH_TABLE
);
1209 /* collect some hash table performance data */
1211 ARCSTAT_BUMP(arcstat_hash_collisions
);
1213 ARCSTAT_BUMP(arcstat_hash_chains
);
1215 ARCSTAT_MAX(arcstat_hash_chain_max
, i
);
1218 ARCSTAT_BUMP(arcstat_hash_elements
);
1219 ARCSTAT_MAXSTAT(arcstat_hash_elements
);
1225 buf_hash_remove(arc_buf_hdr_t
*hdr
)
1227 arc_buf_hdr_t
*fhdr
, **hdrp
;
1228 uint64_t idx
= BUF_HASH_INDEX(hdr
->b_spa
, &hdr
->b_dva
, hdr
->b_birth
);
1230 ASSERT(MUTEX_HELD(BUF_HASH_LOCK(idx
)));
1231 ASSERT(HDR_IN_HASH_TABLE(hdr
));
1233 hdrp
= &buf_hash_table
.ht_table
[idx
];
1234 while ((fhdr
= *hdrp
) != hdr
) {
1235 ASSERT3P(fhdr
, !=, NULL
);
1236 hdrp
= &fhdr
->b_hash_next
;
1238 *hdrp
= hdr
->b_hash_next
;
1239 hdr
->b_hash_next
= NULL
;
1240 arc_hdr_clear_flags(hdr
, ARC_FLAG_IN_HASH_TABLE
);
1242 /* collect some hash table performance data */
1243 ARCSTAT_BUMPDOWN(arcstat_hash_elements
);
1245 if (buf_hash_table
.ht_table
[idx
] &&
1246 buf_hash_table
.ht_table
[idx
]->b_hash_next
== NULL
)
1247 ARCSTAT_BUMPDOWN(arcstat_hash_chains
);
1251 * Global data structures and functions for the buf kmem cache.
1254 static kmem_cache_t
*hdr_full_cache
;
1255 static kmem_cache_t
*hdr_full_crypt_cache
;
1256 static kmem_cache_t
*hdr_l2only_cache
;
1257 static kmem_cache_t
*buf_cache
;
1264 #if defined(_KERNEL)
1266 * Large allocations which do not require contiguous pages
1267 * should be using vmem_free() in the linux kernel\
1269 vmem_free(buf_hash_table
.ht_table
,
1270 (buf_hash_table
.ht_mask
+ 1) * sizeof (void *));
1272 kmem_free(buf_hash_table
.ht_table
,
1273 (buf_hash_table
.ht_mask
+ 1) * sizeof (void *));
1275 for (i
= 0; i
< BUF_LOCKS
; i
++)
1276 mutex_destroy(&buf_hash_table
.ht_locks
[i
].ht_lock
);
1277 kmem_cache_destroy(hdr_full_cache
);
1278 kmem_cache_destroy(hdr_full_crypt_cache
);
1279 kmem_cache_destroy(hdr_l2only_cache
);
1280 kmem_cache_destroy(buf_cache
);
1284 * Constructor callback - called when the cache is empty
1285 * and a new buf is requested.
1289 hdr_full_cons(void *vbuf
, void *unused
, int kmflag
)
1291 arc_buf_hdr_t
*hdr
= vbuf
;
1293 bzero(hdr
, HDR_FULL_SIZE
);
1294 hdr
->b_l1hdr
.b_byteswap
= DMU_BSWAP_NUMFUNCS
;
1295 cv_init(&hdr
->b_l1hdr
.b_cv
, NULL
, CV_DEFAULT
, NULL
);
1296 zfs_refcount_create(&hdr
->b_l1hdr
.b_refcnt
);
1297 mutex_init(&hdr
->b_l1hdr
.b_freeze_lock
, NULL
, MUTEX_DEFAULT
, NULL
);
1298 list_link_init(&hdr
->b_l1hdr
.b_arc_node
);
1299 list_link_init(&hdr
->b_l2hdr
.b_l2node
);
1300 multilist_link_init(&hdr
->b_l1hdr
.b_arc_node
);
1301 arc_space_consume(HDR_FULL_SIZE
, ARC_SPACE_HDRS
);
1308 hdr_full_crypt_cons(void *vbuf
, void *unused
, int kmflag
)
1310 arc_buf_hdr_t
*hdr
= vbuf
;
1312 hdr_full_cons(vbuf
, unused
, kmflag
);
1313 bzero(&hdr
->b_crypt_hdr
, sizeof (hdr
->b_crypt_hdr
));
1314 arc_space_consume(sizeof (hdr
->b_crypt_hdr
), ARC_SPACE_HDRS
);
1321 hdr_l2only_cons(void *vbuf
, void *unused
, int kmflag
)
1323 arc_buf_hdr_t
*hdr
= vbuf
;
1325 bzero(hdr
, HDR_L2ONLY_SIZE
);
1326 arc_space_consume(HDR_L2ONLY_SIZE
, ARC_SPACE_L2HDRS
);
1333 buf_cons(void *vbuf
, void *unused
, int kmflag
)
1335 arc_buf_t
*buf
= vbuf
;
1337 bzero(buf
, sizeof (arc_buf_t
));
1338 mutex_init(&buf
->b_evict_lock
, NULL
, MUTEX_DEFAULT
, NULL
);
1339 arc_space_consume(sizeof (arc_buf_t
), ARC_SPACE_HDRS
);
1345 * Destructor callback - called when a cached buf is
1346 * no longer required.
1350 hdr_full_dest(void *vbuf
, void *unused
)
1352 arc_buf_hdr_t
*hdr
= vbuf
;
1354 ASSERT(HDR_EMPTY(hdr
));
1355 cv_destroy(&hdr
->b_l1hdr
.b_cv
);
1356 zfs_refcount_destroy(&hdr
->b_l1hdr
.b_refcnt
);
1357 mutex_destroy(&hdr
->b_l1hdr
.b_freeze_lock
);
1358 ASSERT(!multilist_link_active(&hdr
->b_l1hdr
.b_arc_node
));
1359 arc_space_return(HDR_FULL_SIZE
, ARC_SPACE_HDRS
);
1364 hdr_full_crypt_dest(void *vbuf
, void *unused
)
1366 arc_buf_hdr_t
*hdr
= vbuf
;
1368 hdr_full_dest(vbuf
, unused
);
1369 arc_space_return(sizeof (hdr
->b_crypt_hdr
), ARC_SPACE_HDRS
);
1374 hdr_l2only_dest(void *vbuf
, void *unused
)
1376 ASSERTV(arc_buf_hdr_t
*hdr
= vbuf
);
1378 ASSERT(HDR_EMPTY(hdr
));
1379 arc_space_return(HDR_L2ONLY_SIZE
, ARC_SPACE_L2HDRS
);
1384 buf_dest(void *vbuf
, void *unused
)
1386 arc_buf_t
*buf
= vbuf
;
1388 mutex_destroy(&buf
->b_evict_lock
);
1389 arc_space_return(sizeof (arc_buf_t
), ARC_SPACE_HDRS
);
1393 * Reclaim callback -- invoked when memory is low.
1397 hdr_recl(void *unused
)
1399 dprintf("hdr_recl called\n");
1401 * umem calls the reclaim func when we destroy the buf cache,
1402 * which is after we do arc_fini().
1404 if (arc_initialized
)
1405 zthr_wakeup(arc_reap_zthr
);
1411 uint64_t *ct
= NULL
;
1412 uint64_t hsize
= 1ULL << 12;
1416 * The hash table is big enough to fill all of physical memory
1417 * with an average block size of zfs_arc_average_blocksize (default 8K).
1418 * By default, the table will take up
1419 * totalmem * sizeof(void*) / 8K (1MB per GB with 8-byte pointers).
1421 while (hsize
* zfs_arc_average_blocksize
< arc_all_memory())
1424 buf_hash_table
.ht_mask
= hsize
- 1;
1425 #if defined(_KERNEL)
1427 * Large allocations which do not require contiguous pages
1428 * should be using vmem_alloc() in the linux kernel
1430 buf_hash_table
.ht_table
=
1431 vmem_zalloc(hsize
* sizeof (void*), KM_SLEEP
);
1433 buf_hash_table
.ht_table
=
1434 kmem_zalloc(hsize
* sizeof (void*), KM_NOSLEEP
);
1436 if (buf_hash_table
.ht_table
== NULL
) {
1437 ASSERT(hsize
> (1ULL << 8));
1442 hdr_full_cache
= kmem_cache_create("arc_buf_hdr_t_full", HDR_FULL_SIZE
,
1443 0, hdr_full_cons
, hdr_full_dest
, hdr_recl
, NULL
, NULL
, 0);
1444 hdr_full_crypt_cache
= kmem_cache_create("arc_buf_hdr_t_full_crypt",
1445 HDR_FULL_CRYPT_SIZE
, 0, hdr_full_crypt_cons
, hdr_full_crypt_dest
,
1446 hdr_recl
, NULL
, NULL
, 0);
1447 hdr_l2only_cache
= kmem_cache_create("arc_buf_hdr_t_l2only",
1448 HDR_L2ONLY_SIZE
, 0, hdr_l2only_cons
, hdr_l2only_dest
, hdr_recl
,
1450 buf_cache
= kmem_cache_create("arc_buf_t", sizeof (arc_buf_t
),
1451 0, buf_cons
, buf_dest
, NULL
, NULL
, NULL
, 0);
1453 for (i
= 0; i
< 256; i
++)
1454 for (ct
= zfs_crc64_table
+ i
, *ct
= i
, j
= 8; j
> 0; j
--)
1455 *ct
= (*ct
>> 1) ^ (-(*ct
& 1) & ZFS_CRC64_POLY
);
1457 for (i
= 0; i
< BUF_LOCKS
; i
++) {
1458 mutex_init(&buf_hash_table
.ht_locks
[i
].ht_lock
,
1459 NULL
, MUTEX_DEFAULT
, NULL
);
1463 #define ARC_MINTIME (hz>>4) /* 62 ms */
1466 * This is the size that the buf occupies in memory. If the buf is compressed,
1467 * it will correspond to the compressed size. You should use this method of
1468 * getting the buf size unless you explicitly need the logical size.
1471 arc_buf_size(arc_buf_t
*buf
)
1473 return (ARC_BUF_COMPRESSED(buf
) ?
1474 HDR_GET_PSIZE(buf
->b_hdr
) : HDR_GET_LSIZE(buf
->b_hdr
));
1478 arc_buf_lsize(arc_buf_t
*buf
)
1480 return (HDR_GET_LSIZE(buf
->b_hdr
));
1484 * This function will return B_TRUE if the buffer is encrypted in memory.
1485 * This buffer can be decrypted by calling arc_untransform().
1488 arc_is_encrypted(arc_buf_t
*buf
)
1490 return (ARC_BUF_ENCRYPTED(buf
) != 0);
1494 * Returns B_TRUE if the buffer represents data that has not had its MAC
1498 arc_is_unauthenticated(arc_buf_t
*buf
)
1500 return (HDR_NOAUTH(buf
->b_hdr
) != 0);
1504 arc_get_raw_params(arc_buf_t
*buf
, boolean_t
*byteorder
, uint8_t *salt
,
1505 uint8_t *iv
, uint8_t *mac
)
1507 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
1509 ASSERT(HDR_PROTECTED(hdr
));
1511 bcopy(hdr
->b_crypt_hdr
.b_salt
, salt
, ZIO_DATA_SALT_LEN
);
1512 bcopy(hdr
->b_crypt_hdr
.b_iv
, iv
, ZIO_DATA_IV_LEN
);
1513 bcopy(hdr
->b_crypt_hdr
.b_mac
, mac
, ZIO_DATA_MAC_LEN
);
1514 *byteorder
= (hdr
->b_l1hdr
.b_byteswap
== DMU_BSWAP_NUMFUNCS
) ?
1515 ZFS_HOST_BYTEORDER
: !ZFS_HOST_BYTEORDER
;
1519 * Indicates how this buffer is compressed in memory. If it is not compressed
1520 * the value will be ZIO_COMPRESS_OFF. It can be made normally readable with
1521 * arc_untransform() as long as it is also unencrypted.
1524 arc_get_compression(arc_buf_t
*buf
)
1526 return (ARC_BUF_COMPRESSED(buf
) ?
1527 HDR_GET_COMPRESS(buf
->b_hdr
) : ZIO_COMPRESS_OFF
);
1531 * Return the compression algorithm used to store this data in the ARC. If ARC
1532 * compression is enabled or this is an encrypted block, this will be the same
1533 * as what's used to store it on-disk. Otherwise, this will be ZIO_COMPRESS_OFF.
1535 static inline enum zio_compress
1536 arc_hdr_get_compress(arc_buf_hdr_t
*hdr
)
1538 return (HDR_COMPRESSION_ENABLED(hdr
) ?
1539 HDR_GET_COMPRESS(hdr
) : ZIO_COMPRESS_OFF
);
1542 static inline boolean_t
1543 arc_buf_is_shared(arc_buf_t
*buf
)
1545 boolean_t shared
= (buf
->b_data
!= NULL
&&
1546 buf
->b_hdr
->b_l1hdr
.b_pabd
!= NULL
&&
1547 abd_is_linear(buf
->b_hdr
->b_l1hdr
.b_pabd
) &&
1548 buf
->b_data
== abd_to_buf(buf
->b_hdr
->b_l1hdr
.b_pabd
));
1549 IMPLY(shared
, HDR_SHARED_DATA(buf
->b_hdr
));
1550 IMPLY(shared
, ARC_BUF_SHARED(buf
));
1551 IMPLY(shared
, ARC_BUF_COMPRESSED(buf
) || ARC_BUF_LAST(buf
));
1554 * It would be nice to assert arc_can_share() too, but the "hdr isn't
1555 * already being shared" requirement prevents us from doing that.
1562 * Free the checksum associated with this header. If there is no checksum, this
1566 arc_cksum_free(arc_buf_hdr_t
*hdr
)
1568 ASSERT(HDR_HAS_L1HDR(hdr
));
1570 mutex_enter(&hdr
->b_l1hdr
.b_freeze_lock
);
1571 if (hdr
->b_l1hdr
.b_freeze_cksum
!= NULL
) {
1572 kmem_free(hdr
->b_l1hdr
.b_freeze_cksum
, sizeof (zio_cksum_t
));
1573 hdr
->b_l1hdr
.b_freeze_cksum
= NULL
;
1575 mutex_exit(&hdr
->b_l1hdr
.b_freeze_lock
);
1579 * Return true iff at least one of the bufs on hdr is not compressed.
1580 * Encrypted buffers count as compressed.
1583 arc_hdr_has_uncompressed_buf(arc_buf_hdr_t
*hdr
)
1585 ASSERT(hdr
->b_l1hdr
.b_state
== arc_anon
||
1586 MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
1588 for (arc_buf_t
*b
= hdr
->b_l1hdr
.b_buf
; b
!= NULL
; b
= b
->b_next
) {
1589 if (!ARC_BUF_COMPRESSED(b
)) {
1598 * If we've turned on the ZFS_DEBUG_MODIFY flag, verify that the buf's data
1599 * matches the checksum that is stored in the hdr. If there is no checksum,
1600 * or if the buf is compressed, this is a no-op.
1603 arc_cksum_verify(arc_buf_t
*buf
)
1605 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
1608 if (!(zfs_flags
& ZFS_DEBUG_MODIFY
))
1611 if (ARC_BUF_COMPRESSED(buf
))
1614 ASSERT(HDR_HAS_L1HDR(hdr
));
1616 mutex_enter(&hdr
->b_l1hdr
.b_freeze_lock
);
1618 if (hdr
->b_l1hdr
.b_freeze_cksum
== NULL
|| HDR_IO_ERROR(hdr
)) {
1619 mutex_exit(&hdr
->b_l1hdr
.b_freeze_lock
);
1623 fletcher_2_native(buf
->b_data
, arc_buf_size(buf
), NULL
, &zc
);
1624 if (!ZIO_CHECKSUM_EQUAL(*hdr
->b_l1hdr
.b_freeze_cksum
, zc
))
1625 panic("buffer modified while frozen!");
1626 mutex_exit(&hdr
->b_l1hdr
.b_freeze_lock
);
1630 * This function makes the assumption that data stored in the L2ARC
1631 * will be transformed exactly as it is in the main pool. Because of
1632 * this we can verify the checksum against the reading process's bp.
1635 arc_cksum_is_equal(arc_buf_hdr_t
*hdr
, zio_t
*zio
)
1637 ASSERT(!BP_IS_EMBEDDED(zio
->io_bp
));
1638 VERIFY3U(BP_GET_PSIZE(zio
->io_bp
), ==, HDR_GET_PSIZE(hdr
));
1641 * Block pointers always store the checksum for the logical data.
1642 * If the block pointer has the gang bit set, then the checksum
1643 * it represents is for the reconstituted data and not for an
1644 * individual gang member. The zio pipeline, however, must be able to
1645 * determine the checksum of each of the gang constituents so it
1646 * treats the checksum comparison differently than what we need
1647 * for l2arc blocks. This prevents us from using the
1648 * zio_checksum_error() interface directly. Instead we must call the
1649 * zio_checksum_error_impl() so that we can ensure the checksum is
1650 * generated using the correct checksum algorithm and accounts for the
1651 * logical I/O size and not just a gang fragment.
1653 return (zio_checksum_error_impl(zio
->io_spa
, zio
->io_bp
,
1654 BP_GET_CHECKSUM(zio
->io_bp
), zio
->io_abd
, zio
->io_size
,
1655 zio
->io_offset
, NULL
) == 0);
1659 * Given a buf full of data, if ZFS_DEBUG_MODIFY is enabled this computes a
1660 * checksum and attaches it to the buf's hdr so that we can ensure that the buf
1661 * isn't modified later on. If buf is compressed or there is already a checksum
1662 * on the hdr, this is a no-op (we only checksum uncompressed bufs).
1665 arc_cksum_compute(arc_buf_t
*buf
)
1667 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
1669 if (!(zfs_flags
& ZFS_DEBUG_MODIFY
))
1672 ASSERT(HDR_HAS_L1HDR(hdr
));
1674 mutex_enter(&buf
->b_hdr
->b_l1hdr
.b_freeze_lock
);
1675 if (hdr
->b_l1hdr
.b_freeze_cksum
!= NULL
|| ARC_BUF_COMPRESSED(buf
)) {
1676 mutex_exit(&hdr
->b_l1hdr
.b_freeze_lock
);
1680 ASSERT(!ARC_BUF_ENCRYPTED(buf
));
1681 ASSERT(!ARC_BUF_COMPRESSED(buf
));
1682 hdr
->b_l1hdr
.b_freeze_cksum
= kmem_alloc(sizeof (zio_cksum_t
),
1684 fletcher_2_native(buf
->b_data
, arc_buf_size(buf
), NULL
,
1685 hdr
->b_l1hdr
.b_freeze_cksum
);
1686 mutex_exit(&hdr
->b_l1hdr
.b_freeze_lock
);
1692 arc_buf_sigsegv(int sig
, siginfo_t
*si
, void *unused
)
1694 panic("Got SIGSEGV at address: 0x%lx\n", (long)si
->si_addr
);
1700 arc_buf_unwatch(arc_buf_t
*buf
)
1704 ASSERT0(mprotect(buf
->b_data
, arc_buf_size(buf
),
1705 PROT_READ
| PROT_WRITE
));
1712 arc_buf_watch(arc_buf_t
*buf
)
1716 ASSERT0(mprotect(buf
->b_data
, arc_buf_size(buf
),
1721 static arc_buf_contents_t
1722 arc_buf_type(arc_buf_hdr_t
*hdr
)
1724 arc_buf_contents_t type
;
1725 if (HDR_ISTYPE_METADATA(hdr
)) {
1726 type
= ARC_BUFC_METADATA
;
1728 type
= ARC_BUFC_DATA
;
1730 VERIFY3U(hdr
->b_type
, ==, type
);
1735 arc_is_metadata(arc_buf_t
*buf
)
1737 return (HDR_ISTYPE_METADATA(buf
->b_hdr
) != 0);
1741 arc_bufc_to_flags(arc_buf_contents_t type
)
1745 /* metadata field is 0 if buffer contains normal data */
1747 case ARC_BUFC_METADATA
:
1748 return (ARC_FLAG_BUFC_METADATA
);
1752 panic("undefined ARC buffer type!");
1753 return ((uint32_t)-1);
1757 arc_buf_thaw(arc_buf_t
*buf
)
1759 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
1761 ASSERT3P(hdr
->b_l1hdr
.b_state
, ==, arc_anon
);
1762 ASSERT(!HDR_IO_IN_PROGRESS(hdr
));
1764 arc_cksum_verify(buf
);
1767 * Compressed buffers do not manipulate the b_freeze_cksum.
1769 if (ARC_BUF_COMPRESSED(buf
))
1772 ASSERT(HDR_HAS_L1HDR(hdr
));
1773 arc_cksum_free(hdr
);
1774 arc_buf_unwatch(buf
);
1778 arc_buf_freeze(arc_buf_t
*buf
)
1780 if (!(zfs_flags
& ZFS_DEBUG_MODIFY
))
1783 if (ARC_BUF_COMPRESSED(buf
))
1786 ASSERT(HDR_HAS_L1HDR(buf
->b_hdr
));
1787 arc_cksum_compute(buf
);
1791 * The arc_buf_hdr_t's b_flags should never be modified directly. Instead,
1792 * the following functions should be used to ensure that the flags are
1793 * updated in a thread-safe way. When manipulating the flags either
1794 * the hash_lock must be held or the hdr must be undiscoverable. This
1795 * ensures that we're not racing with any other threads when updating
1799 arc_hdr_set_flags(arc_buf_hdr_t
*hdr
, arc_flags_t flags
)
1801 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
1802 hdr
->b_flags
|= flags
;
1806 arc_hdr_clear_flags(arc_buf_hdr_t
*hdr
, arc_flags_t flags
)
1808 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
1809 hdr
->b_flags
&= ~flags
;
1813 * Setting the compression bits in the arc_buf_hdr_t's b_flags is
1814 * done in a special way since we have to clear and set bits
1815 * at the same time. Consumers that wish to set the compression bits
1816 * must use this function to ensure that the flags are updated in
1817 * thread-safe manner.
1820 arc_hdr_set_compress(arc_buf_hdr_t
*hdr
, enum zio_compress cmp
)
1822 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
1825 * Holes and embedded blocks will always have a psize = 0 so
1826 * we ignore the compression of the blkptr and set the
1827 * want to uncompress them. Mark them as uncompressed.
1829 if (!zfs_compressed_arc_enabled
|| HDR_GET_PSIZE(hdr
) == 0) {
1830 arc_hdr_clear_flags(hdr
, ARC_FLAG_COMPRESSED_ARC
);
1831 ASSERT(!HDR_COMPRESSION_ENABLED(hdr
));
1833 arc_hdr_set_flags(hdr
, ARC_FLAG_COMPRESSED_ARC
);
1834 ASSERT(HDR_COMPRESSION_ENABLED(hdr
));
1837 HDR_SET_COMPRESS(hdr
, cmp
);
1838 ASSERT3U(HDR_GET_COMPRESS(hdr
), ==, cmp
);
1842 * Looks for another buf on the same hdr which has the data decompressed, copies
1843 * from it, and returns true. If no such buf exists, returns false.
1846 arc_buf_try_copy_decompressed_data(arc_buf_t
*buf
)
1848 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
1849 boolean_t copied
= B_FALSE
;
1851 ASSERT(HDR_HAS_L1HDR(hdr
));
1852 ASSERT3P(buf
->b_data
, !=, NULL
);
1853 ASSERT(!ARC_BUF_COMPRESSED(buf
));
1855 for (arc_buf_t
*from
= hdr
->b_l1hdr
.b_buf
; from
!= NULL
;
1856 from
= from
->b_next
) {
1857 /* can't use our own data buffer */
1862 if (!ARC_BUF_COMPRESSED(from
)) {
1863 bcopy(from
->b_data
, buf
->b_data
, arc_buf_size(buf
));
1870 * There were no decompressed bufs, so there should not be a
1871 * checksum on the hdr either.
1873 EQUIV(!copied
, hdr
->b_l1hdr
.b_freeze_cksum
== NULL
);
1879 * Return the size of the block, b_pabd, that is stored in the arc_buf_hdr_t.
1882 arc_hdr_size(arc_buf_hdr_t
*hdr
)
1886 if (arc_hdr_get_compress(hdr
) != ZIO_COMPRESS_OFF
&&
1887 HDR_GET_PSIZE(hdr
) > 0) {
1888 size
= HDR_GET_PSIZE(hdr
);
1890 ASSERT3U(HDR_GET_LSIZE(hdr
), !=, 0);
1891 size
= HDR_GET_LSIZE(hdr
);
1897 arc_hdr_authenticate(arc_buf_hdr_t
*hdr
, spa_t
*spa
, uint64_t dsobj
)
1901 uint64_t lsize
= HDR_GET_LSIZE(hdr
);
1902 uint64_t psize
= HDR_GET_PSIZE(hdr
);
1903 void *tmpbuf
= NULL
;
1904 abd_t
*abd
= hdr
->b_l1hdr
.b_pabd
;
1906 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
1907 ASSERT(HDR_AUTHENTICATED(hdr
));
1908 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
1911 * The MAC is calculated on the compressed data that is stored on disk.
1912 * However, if compressed arc is disabled we will only have the
1913 * decompressed data available to us now. Compress it into a temporary
1914 * abd so we can verify the MAC. The performance overhead of this will
1915 * be relatively low, since most objects in an encrypted objset will
1916 * be encrypted (instead of authenticated) anyway.
1918 if (HDR_GET_COMPRESS(hdr
) != ZIO_COMPRESS_OFF
&&
1919 !HDR_COMPRESSION_ENABLED(hdr
)) {
1920 tmpbuf
= zio_buf_alloc(lsize
);
1921 abd
= abd_get_from_buf(tmpbuf
, lsize
);
1922 abd_take_ownership_of_buf(abd
, B_TRUE
);
1924 csize
= zio_compress_data(HDR_GET_COMPRESS(hdr
),
1925 hdr
->b_l1hdr
.b_pabd
, tmpbuf
, lsize
);
1926 ASSERT3U(csize
, <=, psize
);
1927 abd_zero_off(abd
, csize
, psize
- csize
);
1931 * Authentication is best effort. We authenticate whenever the key is
1932 * available. If we succeed we clear ARC_FLAG_NOAUTH.
1934 if (hdr
->b_crypt_hdr
.b_ot
== DMU_OT_OBJSET
) {
1935 ASSERT3U(HDR_GET_COMPRESS(hdr
), ==, ZIO_COMPRESS_OFF
);
1936 ASSERT3U(lsize
, ==, psize
);
1937 ret
= spa_do_crypt_objset_mac_abd(B_FALSE
, spa
, dsobj
, abd
,
1938 psize
, hdr
->b_l1hdr
.b_byteswap
!= DMU_BSWAP_NUMFUNCS
);
1940 ret
= spa_do_crypt_mac_abd(B_FALSE
, spa
, dsobj
, abd
, psize
,
1941 hdr
->b_crypt_hdr
.b_mac
);
1945 arc_hdr_clear_flags(hdr
, ARC_FLAG_NOAUTH
);
1946 else if (ret
!= ENOENT
)
1962 * This function will take a header that only has raw encrypted data in
1963 * b_crypt_hdr.b_rabd and decrypt it into a new buffer which is stored in
1964 * b_l1hdr.b_pabd. If designated in the header flags, this function will
1965 * also decompress the data.
1968 arc_hdr_decrypt(arc_buf_hdr_t
*hdr
, spa_t
*spa
, const zbookmark_phys_t
*zb
)
1973 boolean_t no_crypt
= B_FALSE
;
1974 boolean_t bswap
= (hdr
->b_l1hdr
.b_byteswap
!= DMU_BSWAP_NUMFUNCS
);
1976 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
1977 ASSERT(HDR_ENCRYPTED(hdr
));
1979 arc_hdr_alloc_abd(hdr
, B_FALSE
);
1981 ret
= spa_do_crypt_abd(B_FALSE
, spa
, zb
, hdr
->b_crypt_hdr
.b_ot
,
1982 B_FALSE
, bswap
, hdr
->b_crypt_hdr
.b_salt
, hdr
->b_crypt_hdr
.b_iv
,
1983 hdr
->b_crypt_hdr
.b_mac
, HDR_GET_PSIZE(hdr
), hdr
->b_l1hdr
.b_pabd
,
1984 hdr
->b_crypt_hdr
.b_rabd
, &no_crypt
);
1989 abd_copy(hdr
->b_l1hdr
.b_pabd
, hdr
->b_crypt_hdr
.b_rabd
,
1990 HDR_GET_PSIZE(hdr
));
1994 * If this header has disabled arc compression but the b_pabd is
1995 * compressed after decrypting it, we need to decompress the newly
1998 if (HDR_GET_COMPRESS(hdr
) != ZIO_COMPRESS_OFF
&&
1999 !HDR_COMPRESSION_ENABLED(hdr
)) {
2001 * We want to make sure that we are correctly honoring the
2002 * zfs_abd_scatter_enabled setting, so we allocate an abd here
2003 * and then loan a buffer from it, rather than allocating a
2004 * linear buffer and wrapping it in an abd later.
2006 cabd
= arc_get_data_abd(hdr
, arc_hdr_size(hdr
), hdr
);
2007 tmp
= abd_borrow_buf(cabd
, arc_hdr_size(hdr
));
2009 ret
= zio_decompress_data(HDR_GET_COMPRESS(hdr
),
2010 hdr
->b_l1hdr
.b_pabd
, tmp
, HDR_GET_PSIZE(hdr
),
2011 HDR_GET_LSIZE(hdr
));
2013 abd_return_buf(cabd
, tmp
, arc_hdr_size(hdr
));
2017 abd_return_buf_copy(cabd
, tmp
, arc_hdr_size(hdr
));
2018 arc_free_data_abd(hdr
, hdr
->b_l1hdr
.b_pabd
,
2019 arc_hdr_size(hdr
), hdr
);
2020 hdr
->b_l1hdr
.b_pabd
= cabd
;
2026 arc_hdr_free_abd(hdr
, B_FALSE
);
2028 arc_free_data_buf(hdr
, cabd
, arc_hdr_size(hdr
), hdr
);
2034 * This function is called during arc_buf_fill() to prepare the header's
2035 * abd plaintext pointer for use. This involves authenticated protected
2036 * data and decrypting encrypted data into the plaintext abd.
2039 arc_fill_hdr_crypt(arc_buf_hdr_t
*hdr
, kmutex_t
*hash_lock
, spa_t
*spa
,
2040 const zbookmark_phys_t
*zb
, boolean_t noauth
)
2044 ASSERT(HDR_PROTECTED(hdr
));
2046 if (hash_lock
!= NULL
)
2047 mutex_enter(hash_lock
);
2049 if (HDR_NOAUTH(hdr
) && !noauth
) {
2051 * The caller requested authenticated data but our data has
2052 * not been authenticated yet. Verify the MAC now if we can.
2054 ret
= arc_hdr_authenticate(hdr
, spa
, zb
->zb_objset
);
2057 } else if (HDR_HAS_RABD(hdr
) && hdr
->b_l1hdr
.b_pabd
== NULL
) {
2059 * If we only have the encrypted version of the data, but the
2060 * unencrypted version was requested we take this opportunity
2061 * to store the decrypted version in the header for future use.
2063 ret
= arc_hdr_decrypt(hdr
, spa
, zb
);
2068 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
2070 if (hash_lock
!= NULL
)
2071 mutex_exit(hash_lock
);
2076 if (hash_lock
!= NULL
)
2077 mutex_exit(hash_lock
);
2083 * This function is used by the dbuf code to decrypt bonus buffers in place.
2084 * The dbuf code itself doesn't have any locking for decrypting a shared dnode
2085 * block, so we use the hash lock here to protect against concurrent calls to
2089 arc_buf_untransform_in_place(arc_buf_t
*buf
, kmutex_t
*hash_lock
)
2091 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
2093 ASSERT(HDR_ENCRYPTED(hdr
));
2094 ASSERT3U(hdr
->b_crypt_hdr
.b_ot
, ==, DMU_OT_DNODE
);
2095 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
2096 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
2098 zio_crypt_copy_dnode_bonus(hdr
->b_l1hdr
.b_pabd
, buf
->b_data
,
2100 buf
->b_flags
&= ~ARC_BUF_FLAG_ENCRYPTED
;
2101 buf
->b_flags
&= ~ARC_BUF_FLAG_COMPRESSED
;
2102 hdr
->b_crypt_hdr
.b_ebufcnt
-= 1;
2106 * Given a buf that has a data buffer attached to it, this function will
2107 * efficiently fill the buf with data of the specified compression setting from
2108 * the hdr and update the hdr's b_freeze_cksum if necessary. If the buf and hdr
2109 * are already sharing a data buf, no copy is performed.
2111 * If the buf is marked as compressed but uncompressed data was requested, this
2112 * will allocate a new data buffer for the buf, remove that flag, and fill the
2113 * buf with uncompressed data. You can't request a compressed buf on a hdr with
2114 * uncompressed data, and (since we haven't added support for it yet) if you
2115 * want compressed data your buf must already be marked as compressed and have
2116 * the correct-sized data buffer.
2119 arc_buf_fill(arc_buf_t
*buf
, spa_t
*spa
, const zbookmark_phys_t
*zb
,
2120 arc_fill_flags_t flags
)
2123 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
2124 boolean_t hdr_compressed
=
2125 (arc_hdr_get_compress(hdr
) != ZIO_COMPRESS_OFF
);
2126 boolean_t compressed
= (flags
& ARC_FILL_COMPRESSED
) != 0;
2127 boolean_t encrypted
= (flags
& ARC_FILL_ENCRYPTED
) != 0;
2128 dmu_object_byteswap_t bswap
= hdr
->b_l1hdr
.b_byteswap
;
2129 kmutex_t
*hash_lock
= (flags
& ARC_FILL_LOCKED
) ? NULL
: HDR_LOCK(hdr
);
2131 ASSERT3P(buf
->b_data
, !=, NULL
);
2132 IMPLY(compressed
, hdr_compressed
|| ARC_BUF_ENCRYPTED(buf
));
2133 IMPLY(compressed
, ARC_BUF_COMPRESSED(buf
));
2134 IMPLY(encrypted
, HDR_ENCRYPTED(hdr
));
2135 IMPLY(encrypted
, ARC_BUF_ENCRYPTED(buf
));
2136 IMPLY(encrypted
, ARC_BUF_COMPRESSED(buf
));
2137 IMPLY(encrypted
, !ARC_BUF_SHARED(buf
));
2140 * If the caller wanted encrypted data we just need to copy it from
2141 * b_rabd and potentially byteswap it. We won't be able to do any
2142 * further transforms on it.
2145 ASSERT(HDR_HAS_RABD(hdr
));
2146 abd_copy_to_buf(buf
->b_data
, hdr
->b_crypt_hdr
.b_rabd
,
2147 HDR_GET_PSIZE(hdr
));
2152 * Adjust encrypted and authenticated headers to accomodate
2153 * the request if needed. Dnode blocks (ARC_FILL_IN_PLACE) are
2154 * allowed to fail decryption due to keys not being loaded
2155 * without being marked as an IO error.
2157 if (HDR_PROTECTED(hdr
)) {
2158 error
= arc_fill_hdr_crypt(hdr
, hash_lock
, spa
,
2159 zb
, !!(flags
& ARC_FILL_NOAUTH
));
2160 if (error
== EACCES
&& (flags
& ARC_FILL_IN_PLACE
) != 0) {
2162 } else if (error
!= 0) {
2163 if (hash_lock
!= NULL
)
2164 mutex_enter(hash_lock
);
2165 arc_hdr_set_flags(hdr
, ARC_FLAG_IO_ERROR
);
2166 if (hash_lock
!= NULL
)
2167 mutex_exit(hash_lock
);
2173 * There is a special case here for dnode blocks which are
2174 * decrypting their bonus buffers. These blocks may request to
2175 * be decrypted in-place. This is necessary because there may
2176 * be many dnodes pointing into this buffer and there is
2177 * currently no method to synchronize replacing the backing
2178 * b_data buffer and updating all of the pointers. Here we use
2179 * the hash lock to ensure there are no races. If the need
2180 * arises for other types to be decrypted in-place, they must
2181 * add handling here as well.
2183 if ((flags
& ARC_FILL_IN_PLACE
) != 0) {
2184 ASSERT(!hdr_compressed
);
2185 ASSERT(!compressed
);
2188 if (HDR_ENCRYPTED(hdr
) && ARC_BUF_ENCRYPTED(buf
)) {
2189 ASSERT3U(hdr
->b_crypt_hdr
.b_ot
, ==, DMU_OT_DNODE
);
2191 if (hash_lock
!= NULL
)
2192 mutex_enter(hash_lock
);
2193 arc_buf_untransform_in_place(buf
, hash_lock
);
2194 if (hash_lock
!= NULL
)
2195 mutex_exit(hash_lock
);
2197 /* Compute the hdr's checksum if necessary */
2198 arc_cksum_compute(buf
);
2204 if (hdr_compressed
== compressed
) {
2205 if (!arc_buf_is_shared(buf
)) {
2206 abd_copy_to_buf(buf
->b_data
, hdr
->b_l1hdr
.b_pabd
,
2210 ASSERT(hdr_compressed
);
2211 ASSERT(!compressed
);
2212 ASSERT3U(HDR_GET_LSIZE(hdr
), !=, HDR_GET_PSIZE(hdr
));
2215 * If the buf is sharing its data with the hdr, unlink it and
2216 * allocate a new data buffer for the buf.
2218 if (arc_buf_is_shared(buf
)) {
2219 ASSERT(ARC_BUF_COMPRESSED(buf
));
2221 /* We need to give the buf it's own b_data */
2222 buf
->b_flags
&= ~ARC_BUF_FLAG_SHARED
;
2224 arc_get_data_buf(hdr
, HDR_GET_LSIZE(hdr
), buf
);
2225 arc_hdr_clear_flags(hdr
, ARC_FLAG_SHARED_DATA
);
2227 /* Previously overhead was 0; just add new overhead */
2228 ARCSTAT_INCR(arcstat_overhead_size
, HDR_GET_LSIZE(hdr
));
2229 } else if (ARC_BUF_COMPRESSED(buf
)) {
2230 /* We need to reallocate the buf's b_data */
2231 arc_free_data_buf(hdr
, buf
->b_data
, HDR_GET_PSIZE(hdr
),
2234 arc_get_data_buf(hdr
, HDR_GET_LSIZE(hdr
), buf
);
2236 /* We increased the size of b_data; update overhead */
2237 ARCSTAT_INCR(arcstat_overhead_size
,
2238 HDR_GET_LSIZE(hdr
) - HDR_GET_PSIZE(hdr
));
2242 * Regardless of the buf's previous compression settings, it
2243 * should not be compressed at the end of this function.
2245 buf
->b_flags
&= ~ARC_BUF_FLAG_COMPRESSED
;
2248 * Try copying the data from another buf which already has a
2249 * decompressed version. If that's not possible, it's time to
2250 * bite the bullet and decompress the data from the hdr.
2252 if (arc_buf_try_copy_decompressed_data(buf
)) {
2253 /* Skip byteswapping and checksumming (already done) */
2254 ASSERT3P(hdr
->b_l1hdr
.b_freeze_cksum
, !=, NULL
);
2257 error
= zio_decompress_data(HDR_GET_COMPRESS(hdr
),
2258 hdr
->b_l1hdr
.b_pabd
, buf
->b_data
,
2259 HDR_GET_PSIZE(hdr
), HDR_GET_LSIZE(hdr
));
2262 * Absent hardware errors or software bugs, this should
2263 * be impossible, but log it anyway so we can debug it.
2267 "hdr %p, compress %d, psize %d, lsize %d",
2268 hdr
, arc_hdr_get_compress(hdr
),
2269 HDR_GET_PSIZE(hdr
), HDR_GET_LSIZE(hdr
));
2270 if (hash_lock
!= NULL
)
2271 mutex_enter(hash_lock
);
2272 arc_hdr_set_flags(hdr
, ARC_FLAG_IO_ERROR
);
2273 if (hash_lock
!= NULL
)
2274 mutex_exit(hash_lock
);
2275 return (SET_ERROR(EIO
));
2281 /* Byteswap the buf's data if necessary */
2282 if (bswap
!= DMU_BSWAP_NUMFUNCS
) {
2283 ASSERT(!HDR_SHARED_DATA(hdr
));
2284 ASSERT3U(bswap
, <, DMU_BSWAP_NUMFUNCS
);
2285 dmu_ot_byteswap
[bswap
].ob_func(buf
->b_data
, HDR_GET_LSIZE(hdr
));
2288 /* Compute the hdr's checksum if necessary */
2289 arc_cksum_compute(buf
);
2295 * If this function is being called to decrypt an encrypted buffer or verify an
2296 * authenticated one, the key must be loaded and a mapping must be made
2297 * available in the keystore via spa_keystore_create_mapping() or one of its
2301 arc_untransform(arc_buf_t
*buf
, spa_t
*spa
, const zbookmark_phys_t
*zb
,
2305 arc_fill_flags_t flags
= 0;
2308 flags
|= ARC_FILL_IN_PLACE
;
2310 ret
= arc_buf_fill(buf
, spa
, zb
, flags
);
2311 if (ret
== ECKSUM
) {
2313 * Convert authentication and decryption errors to EIO
2314 * (and generate an ereport) before leaving the ARC.
2316 ret
= SET_ERROR(EIO
);
2317 spa_log_error(spa
, zb
);
2318 zfs_ereport_post(FM_EREPORT_ZFS_AUTHENTICATION
,
2319 spa
, NULL
, zb
, NULL
, 0, 0);
2326 * Increment the amount of evictable space in the arc_state_t's refcount.
2327 * We account for the space used by the hdr and the arc buf individually
2328 * so that we can add and remove them from the refcount individually.
2331 arc_evictable_space_increment(arc_buf_hdr_t
*hdr
, arc_state_t
*state
)
2333 arc_buf_contents_t type
= arc_buf_type(hdr
);
2335 ASSERT(HDR_HAS_L1HDR(hdr
));
2337 if (GHOST_STATE(state
)) {
2338 ASSERT0(hdr
->b_l1hdr
.b_bufcnt
);
2339 ASSERT3P(hdr
->b_l1hdr
.b_buf
, ==, NULL
);
2340 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
2341 ASSERT(!HDR_HAS_RABD(hdr
));
2342 (void) zfs_refcount_add_many(&state
->arcs_esize
[type
],
2343 HDR_GET_LSIZE(hdr
), hdr
);
2347 ASSERT(!GHOST_STATE(state
));
2348 if (hdr
->b_l1hdr
.b_pabd
!= NULL
) {
2349 (void) zfs_refcount_add_many(&state
->arcs_esize
[type
],
2350 arc_hdr_size(hdr
), hdr
);
2352 if (HDR_HAS_RABD(hdr
)) {
2353 (void) zfs_refcount_add_many(&state
->arcs_esize
[type
],
2354 HDR_GET_PSIZE(hdr
), hdr
);
2357 for (arc_buf_t
*buf
= hdr
->b_l1hdr
.b_buf
; buf
!= NULL
;
2358 buf
= buf
->b_next
) {
2359 if (arc_buf_is_shared(buf
))
2361 (void) zfs_refcount_add_many(&state
->arcs_esize
[type
],
2362 arc_buf_size(buf
), buf
);
2367 * Decrement the amount of evictable space in the arc_state_t's refcount.
2368 * We account for the space used by the hdr and the arc buf individually
2369 * so that we can add and remove them from the refcount individually.
2372 arc_evictable_space_decrement(arc_buf_hdr_t
*hdr
, arc_state_t
*state
)
2374 arc_buf_contents_t type
= arc_buf_type(hdr
);
2376 ASSERT(HDR_HAS_L1HDR(hdr
));
2378 if (GHOST_STATE(state
)) {
2379 ASSERT0(hdr
->b_l1hdr
.b_bufcnt
);
2380 ASSERT3P(hdr
->b_l1hdr
.b_buf
, ==, NULL
);
2381 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
2382 ASSERT(!HDR_HAS_RABD(hdr
));
2383 (void) zfs_refcount_remove_many(&state
->arcs_esize
[type
],
2384 HDR_GET_LSIZE(hdr
), hdr
);
2388 ASSERT(!GHOST_STATE(state
));
2389 if (hdr
->b_l1hdr
.b_pabd
!= NULL
) {
2390 (void) zfs_refcount_remove_many(&state
->arcs_esize
[type
],
2391 arc_hdr_size(hdr
), hdr
);
2393 if (HDR_HAS_RABD(hdr
)) {
2394 (void) zfs_refcount_remove_many(&state
->arcs_esize
[type
],
2395 HDR_GET_PSIZE(hdr
), hdr
);
2398 for (arc_buf_t
*buf
= hdr
->b_l1hdr
.b_buf
; buf
!= NULL
;
2399 buf
= buf
->b_next
) {
2400 if (arc_buf_is_shared(buf
))
2402 (void) zfs_refcount_remove_many(&state
->arcs_esize
[type
],
2403 arc_buf_size(buf
), buf
);
2408 * Add a reference to this hdr indicating that someone is actively
2409 * referencing that memory. When the refcount transitions from 0 to 1,
2410 * we remove it from the respective arc_state_t list to indicate that
2411 * it is not evictable.
2414 add_reference(arc_buf_hdr_t
*hdr
, void *tag
)
2418 ASSERT(HDR_HAS_L1HDR(hdr
));
2419 if (!MUTEX_HELD(HDR_LOCK(hdr
))) {
2420 ASSERT(hdr
->b_l1hdr
.b_state
== arc_anon
);
2421 ASSERT(zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
));
2422 ASSERT3P(hdr
->b_l1hdr
.b_buf
, ==, NULL
);
2425 state
= hdr
->b_l1hdr
.b_state
;
2427 if ((zfs_refcount_add(&hdr
->b_l1hdr
.b_refcnt
, tag
) == 1) &&
2428 (state
!= arc_anon
)) {
2429 /* We don't use the L2-only state list. */
2430 if (state
!= arc_l2c_only
) {
2431 multilist_remove(state
->arcs_list
[arc_buf_type(hdr
)],
2433 arc_evictable_space_decrement(hdr
, state
);
2435 /* remove the prefetch flag if we get a reference */
2436 arc_hdr_clear_flags(hdr
, ARC_FLAG_PREFETCH
);
2441 * Remove a reference from this hdr. When the reference transitions from
2442 * 1 to 0 and we're not anonymous, then we add this hdr to the arc_state_t's
2443 * list making it eligible for eviction.
2446 remove_reference(arc_buf_hdr_t
*hdr
, kmutex_t
*hash_lock
, void *tag
)
2449 arc_state_t
*state
= hdr
->b_l1hdr
.b_state
;
2451 ASSERT(HDR_HAS_L1HDR(hdr
));
2452 ASSERT(state
== arc_anon
|| MUTEX_HELD(hash_lock
));
2453 ASSERT(!GHOST_STATE(state
));
2456 * arc_l2c_only counts as a ghost state so we don't need to explicitly
2457 * check to prevent usage of the arc_l2c_only list.
2459 if (((cnt
= zfs_refcount_remove(&hdr
->b_l1hdr
.b_refcnt
, tag
)) == 0) &&
2460 (state
!= arc_anon
)) {
2461 multilist_insert(state
->arcs_list
[arc_buf_type(hdr
)], hdr
);
2462 ASSERT3U(hdr
->b_l1hdr
.b_bufcnt
, >, 0);
2463 arc_evictable_space_increment(hdr
, state
);
2469 * Returns detailed information about a specific arc buffer. When the
2470 * state_index argument is set the function will calculate the arc header
2471 * list position for its arc state. Since this requires a linear traversal
2472 * callers are strongly encourage not to do this. However, it can be helpful
2473 * for targeted analysis so the functionality is provided.
2476 arc_buf_info(arc_buf_t
*ab
, arc_buf_info_t
*abi
, int state_index
)
2478 arc_buf_hdr_t
*hdr
= ab
->b_hdr
;
2479 l1arc_buf_hdr_t
*l1hdr
= NULL
;
2480 l2arc_buf_hdr_t
*l2hdr
= NULL
;
2481 arc_state_t
*state
= NULL
;
2483 memset(abi
, 0, sizeof (arc_buf_info_t
));
2488 abi
->abi_flags
= hdr
->b_flags
;
2490 if (HDR_HAS_L1HDR(hdr
)) {
2491 l1hdr
= &hdr
->b_l1hdr
;
2492 state
= l1hdr
->b_state
;
2494 if (HDR_HAS_L2HDR(hdr
))
2495 l2hdr
= &hdr
->b_l2hdr
;
2498 abi
->abi_bufcnt
= l1hdr
->b_bufcnt
;
2499 abi
->abi_access
= l1hdr
->b_arc_access
;
2500 abi
->abi_mru_hits
= l1hdr
->b_mru_hits
;
2501 abi
->abi_mru_ghost_hits
= l1hdr
->b_mru_ghost_hits
;
2502 abi
->abi_mfu_hits
= l1hdr
->b_mfu_hits
;
2503 abi
->abi_mfu_ghost_hits
= l1hdr
->b_mfu_ghost_hits
;
2504 abi
->abi_holds
= zfs_refcount_count(&l1hdr
->b_refcnt
);
2508 abi
->abi_l2arc_dattr
= l2hdr
->b_daddr
;
2509 abi
->abi_l2arc_hits
= l2hdr
->b_hits
;
2512 abi
->abi_state_type
= state
? state
->arcs_state
: ARC_STATE_ANON
;
2513 abi
->abi_state_contents
= arc_buf_type(hdr
);
2514 abi
->abi_size
= arc_hdr_size(hdr
);
2518 * Move the supplied buffer to the indicated state. The hash lock
2519 * for the buffer must be held by the caller.
2522 arc_change_state(arc_state_t
*new_state
, arc_buf_hdr_t
*hdr
,
2523 kmutex_t
*hash_lock
)
2525 arc_state_t
*old_state
;
2528 boolean_t update_old
, update_new
;
2529 arc_buf_contents_t buftype
= arc_buf_type(hdr
);
2532 * We almost always have an L1 hdr here, since we call arc_hdr_realloc()
2533 * in arc_read() when bringing a buffer out of the L2ARC. However, the
2534 * L1 hdr doesn't always exist when we change state to arc_anon before
2535 * destroying a header, in which case reallocating to add the L1 hdr is
2538 if (HDR_HAS_L1HDR(hdr
)) {
2539 old_state
= hdr
->b_l1hdr
.b_state
;
2540 refcnt
= zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
);
2541 bufcnt
= hdr
->b_l1hdr
.b_bufcnt
;
2542 update_old
= (bufcnt
> 0 || hdr
->b_l1hdr
.b_pabd
!= NULL
||
2545 old_state
= arc_l2c_only
;
2548 update_old
= B_FALSE
;
2550 update_new
= update_old
;
2552 ASSERT(MUTEX_HELD(hash_lock
));
2553 ASSERT3P(new_state
, !=, old_state
);
2554 ASSERT(!GHOST_STATE(new_state
) || bufcnt
== 0);
2555 ASSERT(old_state
!= arc_anon
|| bufcnt
<= 1);
2558 * If this buffer is evictable, transfer it from the
2559 * old state list to the new state list.
2562 if (old_state
!= arc_anon
&& old_state
!= arc_l2c_only
) {
2563 ASSERT(HDR_HAS_L1HDR(hdr
));
2564 multilist_remove(old_state
->arcs_list
[buftype
], hdr
);
2566 if (GHOST_STATE(old_state
)) {
2568 ASSERT3P(hdr
->b_l1hdr
.b_buf
, ==, NULL
);
2569 update_old
= B_TRUE
;
2571 arc_evictable_space_decrement(hdr
, old_state
);
2573 if (new_state
!= arc_anon
&& new_state
!= arc_l2c_only
) {
2575 * An L1 header always exists here, since if we're
2576 * moving to some L1-cached state (i.e. not l2c_only or
2577 * anonymous), we realloc the header to add an L1hdr
2580 ASSERT(HDR_HAS_L1HDR(hdr
));
2581 multilist_insert(new_state
->arcs_list
[buftype
], hdr
);
2583 if (GHOST_STATE(new_state
)) {
2585 ASSERT3P(hdr
->b_l1hdr
.b_buf
, ==, NULL
);
2586 update_new
= B_TRUE
;
2588 arc_evictable_space_increment(hdr
, new_state
);
2592 ASSERT(!HDR_EMPTY(hdr
));
2593 if (new_state
== arc_anon
&& HDR_IN_HASH_TABLE(hdr
))
2594 buf_hash_remove(hdr
);
2596 /* adjust state sizes (ignore arc_l2c_only) */
2598 if (update_new
&& new_state
!= arc_l2c_only
) {
2599 ASSERT(HDR_HAS_L1HDR(hdr
));
2600 if (GHOST_STATE(new_state
)) {
2604 * When moving a header to a ghost state, we first
2605 * remove all arc buffers. Thus, we'll have a
2606 * bufcnt of zero, and no arc buffer to use for
2607 * the reference. As a result, we use the arc
2608 * header pointer for the reference.
2610 (void) zfs_refcount_add_many(&new_state
->arcs_size
,
2611 HDR_GET_LSIZE(hdr
), hdr
);
2612 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
2613 ASSERT(!HDR_HAS_RABD(hdr
));
2615 uint32_t buffers
= 0;
2618 * Each individual buffer holds a unique reference,
2619 * thus we must remove each of these references one
2622 for (arc_buf_t
*buf
= hdr
->b_l1hdr
.b_buf
; buf
!= NULL
;
2623 buf
= buf
->b_next
) {
2624 ASSERT3U(bufcnt
, !=, 0);
2628 * When the arc_buf_t is sharing the data
2629 * block with the hdr, the owner of the
2630 * reference belongs to the hdr. Only
2631 * add to the refcount if the arc_buf_t is
2634 if (arc_buf_is_shared(buf
))
2637 (void) zfs_refcount_add_many(
2638 &new_state
->arcs_size
,
2639 arc_buf_size(buf
), buf
);
2641 ASSERT3U(bufcnt
, ==, buffers
);
2643 if (hdr
->b_l1hdr
.b_pabd
!= NULL
) {
2644 (void) zfs_refcount_add_many(
2645 &new_state
->arcs_size
,
2646 arc_hdr_size(hdr
), hdr
);
2649 if (HDR_HAS_RABD(hdr
)) {
2650 (void) zfs_refcount_add_many(
2651 &new_state
->arcs_size
,
2652 HDR_GET_PSIZE(hdr
), hdr
);
2657 if (update_old
&& old_state
!= arc_l2c_only
) {
2658 ASSERT(HDR_HAS_L1HDR(hdr
));
2659 if (GHOST_STATE(old_state
)) {
2661 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
2662 ASSERT(!HDR_HAS_RABD(hdr
));
2665 * When moving a header off of a ghost state,
2666 * the header will not contain any arc buffers.
2667 * We use the arc header pointer for the reference
2668 * which is exactly what we did when we put the
2669 * header on the ghost state.
2672 (void) zfs_refcount_remove_many(&old_state
->arcs_size
,
2673 HDR_GET_LSIZE(hdr
), hdr
);
2675 uint32_t buffers
= 0;
2678 * Each individual buffer holds a unique reference,
2679 * thus we must remove each of these references one
2682 for (arc_buf_t
*buf
= hdr
->b_l1hdr
.b_buf
; buf
!= NULL
;
2683 buf
= buf
->b_next
) {
2684 ASSERT3U(bufcnt
, !=, 0);
2688 * When the arc_buf_t is sharing the data
2689 * block with the hdr, the owner of the
2690 * reference belongs to the hdr. Only
2691 * add to the refcount if the arc_buf_t is
2694 if (arc_buf_is_shared(buf
))
2697 (void) zfs_refcount_remove_many(
2698 &old_state
->arcs_size
, arc_buf_size(buf
),
2701 ASSERT3U(bufcnt
, ==, buffers
);
2702 ASSERT(hdr
->b_l1hdr
.b_pabd
!= NULL
||
2705 if (hdr
->b_l1hdr
.b_pabd
!= NULL
) {
2706 (void) zfs_refcount_remove_many(
2707 &old_state
->arcs_size
, arc_hdr_size(hdr
),
2711 if (HDR_HAS_RABD(hdr
)) {
2712 (void) zfs_refcount_remove_many(
2713 &old_state
->arcs_size
, HDR_GET_PSIZE(hdr
),
2719 if (HDR_HAS_L1HDR(hdr
))
2720 hdr
->b_l1hdr
.b_state
= new_state
;
2723 * L2 headers should never be on the L2 state list since they don't
2724 * have L1 headers allocated.
2726 ASSERT(multilist_is_empty(arc_l2c_only
->arcs_list
[ARC_BUFC_DATA
]) &&
2727 multilist_is_empty(arc_l2c_only
->arcs_list
[ARC_BUFC_METADATA
]));
2731 arc_space_consume(uint64_t space
, arc_space_type_t type
)
2733 ASSERT(type
>= 0 && type
< ARC_SPACE_NUMTYPES
);
2738 case ARC_SPACE_DATA
:
2739 aggsum_add(&astat_data_size
, space
);
2741 case ARC_SPACE_META
:
2742 aggsum_add(&astat_metadata_size
, space
);
2744 case ARC_SPACE_BONUS
:
2745 aggsum_add(&astat_bonus_size
, space
);
2747 case ARC_SPACE_DNODE
:
2748 aggsum_add(&astat_dnode_size
, space
);
2750 case ARC_SPACE_DBUF
:
2751 aggsum_add(&astat_dbuf_size
, space
);
2753 case ARC_SPACE_HDRS
:
2754 aggsum_add(&astat_hdr_size
, space
);
2756 case ARC_SPACE_L2HDRS
:
2757 aggsum_add(&astat_l2_hdr_size
, space
);
2761 if (type
!= ARC_SPACE_DATA
)
2762 aggsum_add(&arc_meta_used
, space
);
2764 aggsum_add(&arc_size
, space
);
2768 arc_space_return(uint64_t space
, arc_space_type_t type
)
2770 ASSERT(type
>= 0 && type
< ARC_SPACE_NUMTYPES
);
2775 case ARC_SPACE_DATA
:
2776 aggsum_add(&astat_data_size
, -space
);
2778 case ARC_SPACE_META
:
2779 aggsum_add(&astat_metadata_size
, -space
);
2781 case ARC_SPACE_BONUS
:
2782 aggsum_add(&astat_bonus_size
, -space
);
2784 case ARC_SPACE_DNODE
:
2785 aggsum_add(&astat_dnode_size
, -space
);
2787 case ARC_SPACE_DBUF
:
2788 aggsum_add(&astat_dbuf_size
, -space
);
2790 case ARC_SPACE_HDRS
:
2791 aggsum_add(&astat_hdr_size
, -space
);
2793 case ARC_SPACE_L2HDRS
:
2794 aggsum_add(&astat_l2_hdr_size
, -space
);
2798 if (type
!= ARC_SPACE_DATA
) {
2799 ASSERT(aggsum_compare(&arc_meta_used
, space
) >= 0);
2801 * We use the upper bound here rather than the precise value
2802 * because the arc_meta_max value doesn't need to be
2803 * precise. It's only consumed by humans via arcstats.
2805 if (arc_meta_max
< aggsum_upper_bound(&arc_meta_used
))
2806 arc_meta_max
= aggsum_upper_bound(&arc_meta_used
);
2807 aggsum_add(&arc_meta_used
, -space
);
2810 ASSERT(aggsum_compare(&arc_size
, space
) >= 0);
2811 aggsum_add(&arc_size
, -space
);
2815 * Given a hdr and a buf, returns whether that buf can share its b_data buffer
2816 * with the hdr's b_pabd.
2819 arc_can_share(arc_buf_hdr_t
*hdr
, arc_buf_t
*buf
)
2822 * The criteria for sharing a hdr's data are:
2823 * 1. the buffer is not encrypted
2824 * 2. the hdr's compression matches the buf's compression
2825 * 3. the hdr doesn't need to be byteswapped
2826 * 4. the hdr isn't already being shared
2827 * 5. the buf is either compressed or it is the last buf in the hdr list
2829 * Criterion #5 maintains the invariant that shared uncompressed
2830 * bufs must be the final buf in the hdr's b_buf list. Reading this, you
2831 * might ask, "if a compressed buf is allocated first, won't that be the
2832 * last thing in the list?", but in that case it's impossible to create
2833 * a shared uncompressed buf anyway (because the hdr must be compressed
2834 * to have the compressed buf). You might also think that #3 is
2835 * sufficient to make this guarantee, however it's possible
2836 * (specifically in the rare L2ARC write race mentioned in
2837 * arc_buf_alloc_impl()) there will be an existing uncompressed buf that
2838 * is sharable, but wasn't at the time of its allocation. Rather than
2839 * allow a new shared uncompressed buf to be created and then shuffle
2840 * the list around to make it the last element, this simply disallows
2841 * sharing if the new buf isn't the first to be added.
2843 ASSERT3P(buf
->b_hdr
, ==, hdr
);
2844 boolean_t hdr_compressed
=
2845 arc_hdr_get_compress(hdr
) != ZIO_COMPRESS_OFF
;
2846 boolean_t buf_compressed
= ARC_BUF_COMPRESSED(buf
) != 0;
2847 return (!ARC_BUF_ENCRYPTED(buf
) &&
2848 buf_compressed
== hdr_compressed
&&
2849 hdr
->b_l1hdr
.b_byteswap
== DMU_BSWAP_NUMFUNCS
&&
2850 !HDR_SHARED_DATA(hdr
) &&
2851 (ARC_BUF_LAST(buf
) || ARC_BUF_COMPRESSED(buf
)));
2855 * Allocate a buf for this hdr. If you care about the data that's in the hdr,
2856 * or if you want a compressed buffer, pass those flags in. Returns 0 if the
2857 * copy was made successfully, or an error code otherwise.
2860 arc_buf_alloc_impl(arc_buf_hdr_t
*hdr
, spa_t
*spa
, const zbookmark_phys_t
*zb
,
2861 void *tag
, boolean_t encrypted
, boolean_t compressed
, boolean_t noauth
,
2862 boolean_t fill
, arc_buf_t
**ret
)
2865 arc_fill_flags_t flags
= ARC_FILL_LOCKED
;
2867 ASSERT(HDR_HAS_L1HDR(hdr
));
2868 ASSERT3U(HDR_GET_LSIZE(hdr
), >, 0);
2869 VERIFY(hdr
->b_type
== ARC_BUFC_DATA
||
2870 hdr
->b_type
== ARC_BUFC_METADATA
);
2871 ASSERT3P(ret
, !=, NULL
);
2872 ASSERT3P(*ret
, ==, NULL
);
2873 IMPLY(encrypted
, compressed
);
2875 hdr
->b_l1hdr
.b_mru_hits
= 0;
2876 hdr
->b_l1hdr
.b_mru_ghost_hits
= 0;
2877 hdr
->b_l1hdr
.b_mfu_hits
= 0;
2878 hdr
->b_l1hdr
.b_mfu_ghost_hits
= 0;
2879 hdr
->b_l1hdr
.b_l2_hits
= 0;
2881 buf
= *ret
= kmem_cache_alloc(buf_cache
, KM_PUSHPAGE
);
2884 buf
->b_next
= hdr
->b_l1hdr
.b_buf
;
2887 add_reference(hdr
, tag
);
2890 * We're about to change the hdr's b_flags. We must either
2891 * hold the hash_lock or be undiscoverable.
2893 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
2896 * Only honor requests for compressed bufs if the hdr is actually
2897 * compressed. This must be overriden if the buffer is encrypted since
2898 * encrypted buffers cannot be decompressed.
2901 buf
->b_flags
|= ARC_BUF_FLAG_COMPRESSED
;
2902 buf
->b_flags
|= ARC_BUF_FLAG_ENCRYPTED
;
2903 flags
|= ARC_FILL_COMPRESSED
| ARC_FILL_ENCRYPTED
;
2904 } else if (compressed
&&
2905 arc_hdr_get_compress(hdr
) != ZIO_COMPRESS_OFF
) {
2906 buf
->b_flags
|= ARC_BUF_FLAG_COMPRESSED
;
2907 flags
|= ARC_FILL_COMPRESSED
;
2912 flags
|= ARC_FILL_NOAUTH
;
2916 * If the hdr's data can be shared then we share the data buffer and
2917 * set the appropriate bit in the hdr's b_flags to indicate the hdr is
2918 * allocate a new buffer to store the buf's data.
2920 * There are two additional restrictions here because we're sharing
2921 * hdr -> buf instead of the usual buf -> hdr. First, the hdr can't be
2922 * actively involved in an L2ARC write, because if this buf is used by
2923 * an arc_write() then the hdr's data buffer will be released when the
2924 * write completes, even though the L2ARC write might still be using it.
2925 * Second, the hdr's ABD must be linear so that the buf's user doesn't
2926 * need to be ABD-aware.
2928 boolean_t can_share
= arc_can_share(hdr
, buf
) && !HDR_L2_WRITING(hdr
) &&
2929 hdr
->b_l1hdr
.b_pabd
!= NULL
&& abd_is_linear(hdr
->b_l1hdr
.b_pabd
);
2931 /* Set up b_data and sharing */
2933 buf
->b_data
= abd_to_buf(hdr
->b_l1hdr
.b_pabd
);
2934 buf
->b_flags
|= ARC_BUF_FLAG_SHARED
;
2935 arc_hdr_set_flags(hdr
, ARC_FLAG_SHARED_DATA
);
2938 arc_get_data_buf(hdr
, arc_buf_size(buf
), buf
);
2939 ARCSTAT_INCR(arcstat_overhead_size
, arc_buf_size(buf
));
2941 VERIFY3P(buf
->b_data
, !=, NULL
);
2943 hdr
->b_l1hdr
.b_buf
= buf
;
2944 hdr
->b_l1hdr
.b_bufcnt
+= 1;
2946 hdr
->b_crypt_hdr
.b_ebufcnt
+= 1;
2949 * If the user wants the data from the hdr, we need to either copy or
2950 * decompress the data.
2953 ASSERT3P(zb
, !=, NULL
);
2954 return (arc_buf_fill(buf
, spa
, zb
, flags
));
2960 static char *arc_onloan_tag
= "onloan";
2963 arc_loaned_bytes_update(int64_t delta
)
2965 atomic_add_64(&arc_loaned_bytes
, delta
);
2967 /* assert that it did not wrap around */
2968 ASSERT3S(atomic_add_64_nv(&arc_loaned_bytes
, 0), >=, 0);
2972 * Loan out an anonymous arc buffer. Loaned buffers are not counted as in
2973 * flight data by arc_tempreserve_space() until they are "returned". Loaned
2974 * buffers must be returned to the arc before they can be used by the DMU or
2978 arc_loan_buf(spa_t
*spa
, boolean_t is_metadata
, int size
)
2980 arc_buf_t
*buf
= arc_alloc_buf(spa
, arc_onloan_tag
,
2981 is_metadata
? ARC_BUFC_METADATA
: ARC_BUFC_DATA
, size
);
2983 arc_loaned_bytes_update(arc_buf_size(buf
));
2989 arc_loan_compressed_buf(spa_t
*spa
, uint64_t psize
, uint64_t lsize
,
2990 enum zio_compress compression_type
)
2992 arc_buf_t
*buf
= arc_alloc_compressed_buf(spa
, arc_onloan_tag
,
2993 psize
, lsize
, compression_type
);
2995 arc_loaned_bytes_update(arc_buf_size(buf
));
3001 arc_loan_raw_buf(spa_t
*spa
, uint64_t dsobj
, boolean_t byteorder
,
3002 const uint8_t *salt
, const uint8_t *iv
, const uint8_t *mac
,
3003 dmu_object_type_t ot
, uint64_t psize
, uint64_t lsize
,
3004 enum zio_compress compression_type
)
3006 arc_buf_t
*buf
= arc_alloc_raw_buf(spa
, arc_onloan_tag
, dsobj
,
3007 byteorder
, salt
, iv
, mac
, ot
, psize
, lsize
, compression_type
);
3009 atomic_add_64(&arc_loaned_bytes
, psize
);
3015 * Return a loaned arc buffer to the arc.
3018 arc_return_buf(arc_buf_t
*buf
, void *tag
)
3020 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
3022 ASSERT3P(buf
->b_data
, !=, NULL
);
3023 ASSERT(HDR_HAS_L1HDR(hdr
));
3024 (void) zfs_refcount_add(&hdr
->b_l1hdr
.b_refcnt
, tag
);
3025 (void) zfs_refcount_remove(&hdr
->b_l1hdr
.b_refcnt
, arc_onloan_tag
);
3027 arc_loaned_bytes_update(-arc_buf_size(buf
));
3030 /* Detach an arc_buf from a dbuf (tag) */
3032 arc_loan_inuse_buf(arc_buf_t
*buf
, void *tag
)
3034 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
3036 ASSERT3P(buf
->b_data
, !=, NULL
);
3037 ASSERT(HDR_HAS_L1HDR(hdr
));
3038 (void) zfs_refcount_add(&hdr
->b_l1hdr
.b_refcnt
, arc_onloan_tag
);
3039 (void) zfs_refcount_remove(&hdr
->b_l1hdr
.b_refcnt
, tag
);
3041 arc_loaned_bytes_update(arc_buf_size(buf
));
3045 l2arc_free_abd_on_write(abd_t
*abd
, size_t size
, arc_buf_contents_t type
)
3047 l2arc_data_free_t
*df
= kmem_alloc(sizeof (*df
), KM_SLEEP
);
3050 df
->l2df_size
= size
;
3051 df
->l2df_type
= type
;
3052 mutex_enter(&l2arc_free_on_write_mtx
);
3053 list_insert_head(l2arc_free_on_write
, df
);
3054 mutex_exit(&l2arc_free_on_write_mtx
);
3058 arc_hdr_free_on_write(arc_buf_hdr_t
*hdr
, boolean_t free_rdata
)
3060 arc_state_t
*state
= hdr
->b_l1hdr
.b_state
;
3061 arc_buf_contents_t type
= arc_buf_type(hdr
);
3062 uint64_t size
= (free_rdata
) ? HDR_GET_PSIZE(hdr
) : arc_hdr_size(hdr
);
3064 /* protected by hash lock, if in the hash table */
3065 if (multilist_link_active(&hdr
->b_l1hdr
.b_arc_node
)) {
3066 ASSERT(zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
));
3067 ASSERT(state
!= arc_anon
&& state
!= arc_l2c_only
);
3069 (void) zfs_refcount_remove_many(&state
->arcs_esize
[type
],
3072 (void) zfs_refcount_remove_many(&state
->arcs_size
, size
, hdr
);
3073 if (type
== ARC_BUFC_METADATA
) {
3074 arc_space_return(size
, ARC_SPACE_META
);
3076 ASSERT(type
== ARC_BUFC_DATA
);
3077 arc_space_return(size
, ARC_SPACE_DATA
);
3081 l2arc_free_abd_on_write(hdr
->b_crypt_hdr
.b_rabd
, size
, type
);
3083 l2arc_free_abd_on_write(hdr
->b_l1hdr
.b_pabd
, size
, type
);
3088 * Share the arc_buf_t's data with the hdr. Whenever we are sharing the
3089 * data buffer, we transfer the refcount ownership to the hdr and update
3090 * the appropriate kstats.
3093 arc_share_buf(arc_buf_hdr_t
*hdr
, arc_buf_t
*buf
)
3095 ASSERT(arc_can_share(hdr
, buf
));
3096 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
3097 ASSERT(!ARC_BUF_ENCRYPTED(buf
));
3098 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
3101 * Start sharing the data buffer. We transfer the
3102 * refcount ownership to the hdr since it always owns
3103 * the refcount whenever an arc_buf_t is shared.
3105 zfs_refcount_transfer_ownership_many(&hdr
->b_l1hdr
.b_state
->arcs_size
,
3106 arc_hdr_size(hdr
), buf
, hdr
);
3107 hdr
->b_l1hdr
.b_pabd
= abd_get_from_buf(buf
->b_data
, arc_buf_size(buf
));
3108 abd_take_ownership_of_buf(hdr
->b_l1hdr
.b_pabd
,
3109 HDR_ISTYPE_METADATA(hdr
));
3110 arc_hdr_set_flags(hdr
, ARC_FLAG_SHARED_DATA
);
3111 buf
->b_flags
|= ARC_BUF_FLAG_SHARED
;
3114 * Since we've transferred ownership to the hdr we need
3115 * to increment its compressed and uncompressed kstats and
3116 * decrement the overhead size.
3118 ARCSTAT_INCR(arcstat_compressed_size
, arc_hdr_size(hdr
));
3119 ARCSTAT_INCR(arcstat_uncompressed_size
, HDR_GET_LSIZE(hdr
));
3120 ARCSTAT_INCR(arcstat_overhead_size
, -arc_buf_size(buf
));
3124 arc_unshare_buf(arc_buf_hdr_t
*hdr
, arc_buf_t
*buf
)
3126 ASSERT(arc_buf_is_shared(buf
));
3127 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
3128 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
3131 * We are no longer sharing this buffer so we need
3132 * to transfer its ownership to the rightful owner.
3134 zfs_refcount_transfer_ownership_many(&hdr
->b_l1hdr
.b_state
->arcs_size
,
3135 arc_hdr_size(hdr
), hdr
, buf
);
3136 arc_hdr_clear_flags(hdr
, ARC_FLAG_SHARED_DATA
);
3137 abd_release_ownership_of_buf(hdr
->b_l1hdr
.b_pabd
);
3138 abd_put(hdr
->b_l1hdr
.b_pabd
);
3139 hdr
->b_l1hdr
.b_pabd
= NULL
;
3140 buf
->b_flags
&= ~ARC_BUF_FLAG_SHARED
;
3143 * Since the buffer is no longer shared between
3144 * the arc buf and the hdr, count it as overhead.
3146 ARCSTAT_INCR(arcstat_compressed_size
, -arc_hdr_size(hdr
));
3147 ARCSTAT_INCR(arcstat_uncompressed_size
, -HDR_GET_LSIZE(hdr
));
3148 ARCSTAT_INCR(arcstat_overhead_size
, arc_buf_size(buf
));
3152 * Remove an arc_buf_t from the hdr's buf list and return the last
3153 * arc_buf_t on the list. If no buffers remain on the list then return
3157 arc_buf_remove(arc_buf_hdr_t
*hdr
, arc_buf_t
*buf
)
3159 ASSERT(HDR_HAS_L1HDR(hdr
));
3160 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
3162 arc_buf_t
**bufp
= &hdr
->b_l1hdr
.b_buf
;
3163 arc_buf_t
*lastbuf
= NULL
;
3166 * Remove the buf from the hdr list and locate the last
3167 * remaining buffer on the list.
3169 while (*bufp
!= NULL
) {
3171 *bufp
= buf
->b_next
;
3174 * If we've removed a buffer in the middle of
3175 * the list then update the lastbuf and update
3178 if (*bufp
!= NULL
) {
3180 bufp
= &(*bufp
)->b_next
;
3184 ASSERT3P(lastbuf
, !=, buf
);
3185 IMPLY(hdr
->b_l1hdr
.b_bufcnt
> 0, lastbuf
!= NULL
);
3186 IMPLY(hdr
->b_l1hdr
.b_bufcnt
> 0, hdr
->b_l1hdr
.b_buf
!= NULL
);
3187 IMPLY(lastbuf
!= NULL
, ARC_BUF_LAST(lastbuf
));
3193 * Free up buf->b_data and pull the arc_buf_t off of the the arc_buf_hdr_t's
3197 arc_buf_destroy_impl(arc_buf_t
*buf
)
3199 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
3202 * Free up the data associated with the buf but only if we're not
3203 * sharing this with the hdr. If we are sharing it with the hdr, the
3204 * hdr is responsible for doing the free.
3206 if (buf
->b_data
!= NULL
) {
3208 * We're about to change the hdr's b_flags. We must either
3209 * hold the hash_lock or be undiscoverable.
3211 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)) || HDR_EMPTY(hdr
));
3213 arc_cksum_verify(buf
);
3214 arc_buf_unwatch(buf
);
3216 if (arc_buf_is_shared(buf
)) {
3217 arc_hdr_clear_flags(hdr
, ARC_FLAG_SHARED_DATA
);
3219 uint64_t size
= arc_buf_size(buf
);
3220 arc_free_data_buf(hdr
, buf
->b_data
, size
, buf
);
3221 ARCSTAT_INCR(arcstat_overhead_size
, -size
);
3225 ASSERT(hdr
->b_l1hdr
.b_bufcnt
> 0);
3226 hdr
->b_l1hdr
.b_bufcnt
-= 1;
3228 if (ARC_BUF_ENCRYPTED(buf
)) {
3229 hdr
->b_crypt_hdr
.b_ebufcnt
-= 1;
3232 * If we have no more encrypted buffers and we've
3233 * already gotten a copy of the decrypted data we can
3234 * free b_rabd to save some space.
3236 if (hdr
->b_crypt_hdr
.b_ebufcnt
== 0 &&
3237 HDR_HAS_RABD(hdr
) && hdr
->b_l1hdr
.b_pabd
!= NULL
&&
3238 !HDR_IO_IN_PROGRESS(hdr
)) {
3239 arc_hdr_free_abd(hdr
, B_TRUE
);
3244 arc_buf_t
*lastbuf
= arc_buf_remove(hdr
, buf
);
3246 if (ARC_BUF_SHARED(buf
) && !ARC_BUF_COMPRESSED(buf
)) {
3248 * If the current arc_buf_t is sharing its data buffer with the
3249 * hdr, then reassign the hdr's b_pabd to share it with the new
3250 * buffer at the end of the list. The shared buffer is always
3251 * the last one on the hdr's buffer list.
3253 * There is an equivalent case for compressed bufs, but since
3254 * they aren't guaranteed to be the last buf in the list and
3255 * that is an exceedingly rare case, we just allow that space be
3256 * wasted temporarily. We must also be careful not to share
3257 * encrypted buffers, since they cannot be shared.
3259 if (lastbuf
!= NULL
&& !ARC_BUF_ENCRYPTED(lastbuf
)) {
3260 /* Only one buf can be shared at once */
3261 VERIFY(!arc_buf_is_shared(lastbuf
));
3262 /* hdr is uncompressed so can't have compressed buf */
3263 VERIFY(!ARC_BUF_COMPRESSED(lastbuf
));
3265 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
3266 arc_hdr_free_abd(hdr
, B_FALSE
);
3269 * We must setup a new shared block between the
3270 * last buffer and the hdr. The data would have
3271 * been allocated by the arc buf so we need to transfer
3272 * ownership to the hdr since it's now being shared.
3274 arc_share_buf(hdr
, lastbuf
);
3276 } else if (HDR_SHARED_DATA(hdr
)) {
3278 * Uncompressed shared buffers are always at the end
3279 * of the list. Compressed buffers don't have the
3280 * same requirements. This makes it hard to
3281 * simply assert that the lastbuf is shared so
3282 * we rely on the hdr's compression flags to determine
3283 * if we have a compressed, shared buffer.
3285 ASSERT3P(lastbuf
, !=, NULL
);
3286 ASSERT(arc_buf_is_shared(lastbuf
) ||
3287 arc_hdr_get_compress(hdr
) != ZIO_COMPRESS_OFF
);
3291 * Free the checksum if we're removing the last uncompressed buf from
3294 if (!arc_hdr_has_uncompressed_buf(hdr
)) {
3295 arc_cksum_free(hdr
);
3298 /* clean up the buf */
3300 kmem_cache_free(buf_cache
, buf
);
3304 arc_hdr_alloc_abd(arc_buf_hdr_t
*hdr
, boolean_t alloc_rdata
)
3308 ASSERT3U(HDR_GET_LSIZE(hdr
), >, 0);
3309 ASSERT(HDR_HAS_L1HDR(hdr
));
3310 ASSERT(!HDR_SHARED_DATA(hdr
) || alloc_rdata
);
3311 IMPLY(alloc_rdata
, HDR_PROTECTED(hdr
));
3314 size
= HDR_GET_PSIZE(hdr
);
3315 ASSERT3P(hdr
->b_crypt_hdr
.b_rabd
, ==, NULL
);
3316 hdr
->b_crypt_hdr
.b_rabd
= arc_get_data_abd(hdr
, size
, hdr
);
3317 ASSERT3P(hdr
->b_crypt_hdr
.b_rabd
, !=, NULL
);
3318 ARCSTAT_INCR(arcstat_raw_size
, size
);
3320 size
= arc_hdr_size(hdr
);
3321 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
3322 hdr
->b_l1hdr
.b_pabd
= arc_get_data_abd(hdr
, size
, hdr
);
3323 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
3326 ARCSTAT_INCR(arcstat_compressed_size
, size
);
3327 ARCSTAT_INCR(arcstat_uncompressed_size
, HDR_GET_LSIZE(hdr
));
3331 arc_hdr_free_abd(arc_buf_hdr_t
*hdr
, boolean_t free_rdata
)
3333 uint64_t size
= (free_rdata
) ? HDR_GET_PSIZE(hdr
) : arc_hdr_size(hdr
);
3335 ASSERT(HDR_HAS_L1HDR(hdr
));
3336 ASSERT(hdr
->b_l1hdr
.b_pabd
!= NULL
|| HDR_HAS_RABD(hdr
));
3337 IMPLY(free_rdata
, HDR_HAS_RABD(hdr
));
3340 * If the hdr is currently being written to the l2arc then
3341 * we defer freeing the data by adding it to the l2arc_free_on_write
3342 * list. The l2arc will free the data once it's finished
3343 * writing it to the l2arc device.
3345 if (HDR_L2_WRITING(hdr
)) {
3346 arc_hdr_free_on_write(hdr
, free_rdata
);
3347 ARCSTAT_BUMP(arcstat_l2_free_on_write
);
3348 } else if (free_rdata
) {
3349 arc_free_data_abd(hdr
, hdr
->b_crypt_hdr
.b_rabd
, size
, hdr
);
3351 arc_free_data_abd(hdr
, hdr
->b_l1hdr
.b_pabd
, size
, hdr
);
3355 hdr
->b_crypt_hdr
.b_rabd
= NULL
;
3356 ARCSTAT_INCR(arcstat_raw_size
, -size
);
3358 hdr
->b_l1hdr
.b_pabd
= NULL
;
3361 if (hdr
->b_l1hdr
.b_pabd
== NULL
&& !HDR_HAS_RABD(hdr
))
3362 hdr
->b_l1hdr
.b_byteswap
= DMU_BSWAP_NUMFUNCS
;
3364 ARCSTAT_INCR(arcstat_compressed_size
, -size
);
3365 ARCSTAT_INCR(arcstat_uncompressed_size
, -HDR_GET_LSIZE(hdr
));
3368 static arc_buf_hdr_t
*
3369 arc_hdr_alloc(uint64_t spa
, int32_t psize
, int32_t lsize
,
3370 boolean_t
protected, enum zio_compress compression_type
,
3371 arc_buf_contents_t type
, boolean_t alloc_rdata
)
3375 VERIFY(type
== ARC_BUFC_DATA
|| type
== ARC_BUFC_METADATA
);
3377 hdr
= kmem_cache_alloc(hdr_full_crypt_cache
, KM_PUSHPAGE
);
3379 hdr
= kmem_cache_alloc(hdr_full_cache
, KM_PUSHPAGE
);
3382 ASSERT(HDR_EMPTY(hdr
));
3383 ASSERT3P(hdr
->b_l1hdr
.b_freeze_cksum
, ==, NULL
);
3384 HDR_SET_PSIZE(hdr
, psize
);
3385 HDR_SET_LSIZE(hdr
, lsize
);
3389 arc_hdr_set_flags(hdr
, arc_bufc_to_flags(type
) | ARC_FLAG_HAS_L1HDR
);
3390 arc_hdr_set_compress(hdr
, compression_type
);
3392 arc_hdr_set_flags(hdr
, ARC_FLAG_PROTECTED
);
3394 hdr
->b_l1hdr
.b_state
= arc_anon
;
3395 hdr
->b_l1hdr
.b_arc_access
= 0;
3396 hdr
->b_l1hdr
.b_bufcnt
= 0;
3397 hdr
->b_l1hdr
.b_buf
= NULL
;
3400 * Allocate the hdr's buffer. This will contain either
3401 * the compressed or uncompressed data depending on the block
3402 * it references and compressed arc enablement.
3404 arc_hdr_alloc_abd(hdr
, alloc_rdata
);
3405 ASSERT(zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
));
3411 * Transition between the two allocation states for the arc_buf_hdr struct.
3412 * The arc_buf_hdr struct can be allocated with (hdr_full_cache) or without
3413 * (hdr_l2only_cache) the fields necessary for the L1 cache - the smaller
3414 * version is used when a cache buffer is only in the L2ARC in order to reduce
3417 static arc_buf_hdr_t
*
3418 arc_hdr_realloc(arc_buf_hdr_t
*hdr
, kmem_cache_t
*old
, kmem_cache_t
*new)
3420 ASSERT(HDR_HAS_L2HDR(hdr
));
3422 arc_buf_hdr_t
*nhdr
;
3423 l2arc_dev_t
*dev
= hdr
->b_l2hdr
.b_dev
;
3425 ASSERT((old
== hdr_full_cache
&& new == hdr_l2only_cache
) ||
3426 (old
== hdr_l2only_cache
&& new == hdr_full_cache
));
3429 * if the caller wanted a new full header and the header is to be
3430 * encrypted we will actually allocate the header from the full crypt
3431 * cache instead. The same applies to freeing from the old cache.
3433 if (HDR_PROTECTED(hdr
) && new == hdr_full_cache
)
3434 new = hdr_full_crypt_cache
;
3435 if (HDR_PROTECTED(hdr
) && old
== hdr_full_cache
)
3436 old
= hdr_full_crypt_cache
;
3438 nhdr
= kmem_cache_alloc(new, KM_PUSHPAGE
);
3440 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)));
3441 buf_hash_remove(hdr
);
3443 bcopy(hdr
, nhdr
, HDR_L2ONLY_SIZE
);
3445 if (new == hdr_full_cache
|| new == hdr_full_crypt_cache
) {
3446 arc_hdr_set_flags(nhdr
, ARC_FLAG_HAS_L1HDR
);
3448 * arc_access and arc_change_state need to be aware that a
3449 * header has just come out of L2ARC, so we set its state to
3450 * l2c_only even though it's about to change.
3452 nhdr
->b_l1hdr
.b_state
= arc_l2c_only
;
3454 /* Verify previous threads set to NULL before freeing */
3455 ASSERT3P(nhdr
->b_l1hdr
.b_pabd
, ==, NULL
);
3456 ASSERT(!HDR_HAS_RABD(hdr
));
3458 ASSERT3P(hdr
->b_l1hdr
.b_buf
, ==, NULL
);
3459 ASSERT0(hdr
->b_l1hdr
.b_bufcnt
);
3460 ASSERT3P(hdr
->b_l1hdr
.b_freeze_cksum
, ==, NULL
);
3463 * If we've reached here, We must have been called from
3464 * arc_evict_hdr(), as such we should have already been
3465 * removed from any ghost list we were previously on
3466 * (which protects us from racing with arc_evict_state),
3467 * thus no locking is needed during this check.
3469 ASSERT(!multilist_link_active(&hdr
->b_l1hdr
.b_arc_node
));
3472 * A buffer must not be moved into the arc_l2c_only
3473 * state if it's not finished being written out to the
3474 * l2arc device. Otherwise, the b_l1hdr.b_pabd field
3475 * might try to be accessed, even though it was removed.
3477 VERIFY(!HDR_L2_WRITING(hdr
));
3478 VERIFY3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
3479 ASSERT(!HDR_HAS_RABD(hdr
));
3481 arc_hdr_clear_flags(nhdr
, ARC_FLAG_HAS_L1HDR
);
3484 * The header has been reallocated so we need to re-insert it into any
3487 (void) buf_hash_insert(nhdr
, NULL
);
3489 ASSERT(list_link_active(&hdr
->b_l2hdr
.b_l2node
));
3491 mutex_enter(&dev
->l2ad_mtx
);
3494 * We must place the realloc'ed header back into the list at
3495 * the same spot. Otherwise, if it's placed earlier in the list,
3496 * l2arc_write_buffers() could find it during the function's
3497 * write phase, and try to write it out to the l2arc.
3499 list_insert_after(&dev
->l2ad_buflist
, hdr
, nhdr
);
3500 list_remove(&dev
->l2ad_buflist
, hdr
);
3502 mutex_exit(&dev
->l2ad_mtx
);
3505 * Since we're using the pointer address as the tag when
3506 * incrementing and decrementing the l2ad_alloc refcount, we
3507 * must remove the old pointer (that we're about to destroy) and
3508 * add the new pointer to the refcount. Otherwise we'd remove
3509 * the wrong pointer address when calling arc_hdr_destroy() later.
3512 (void) zfs_refcount_remove_many(&dev
->l2ad_alloc
,
3513 arc_hdr_size(hdr
), hdr
);
3514 (void) zfs_refcount_add_many(&dev
->l2ad_alloc
,
3515 arc_hdr_size(nhdr
), nhdr
);
3517 buf_discard_identity(hdr
);
3518 kmem_cache_free(old
, hdr
);
3524 * This function allows an L1 header to be reallocated as a crypt
3525 * header and vice versa. If we are going to a crypt header, the
3526 * new fields will be zeroed out.
3528 static arc_buf_hdr_t
*
3529 arc_hdr_realloc_crypt(arc_buf_hdr_t
*hdr
, boolean_t need_crypt
)
3531 arc_buf_hdr_t
*nhdr
;
3533 kmem_cache_t
*ncache
, *ocache
;
3534 unsigned nsize
, osize
;
3537 * This function requires that hdr is in the arc_anon state.
3538 * Therefore it won't have any L2ARC data for us to worry
3541 ASSERT(HDR_HAS_L1HDR(hdr
));
3542 ASSERT(!HDR_HAS_L2HDR(hdr
));
3543 ASSERT3U(!!HDR_PROTECTED(hdr
), !=, need_crypt
);
3544 ASSERT3P(hdr
->b_l1hdr
.b_state
, ==, arc_anon
);
3545 ASSERT(!multilist_link_active(&hdr
->b_l1hdr
.b_arc_node
));
3546 ASSERT(!list_link_active(&hdr
->b_l2hdr
.b_l2node
));
3547 ASSERT3P(hdr
->b_hash_next
, ==, NULL
);
3550 ncache
= hdr_full_crypt_cache
;
3551 nsize
= sizeof (hdr
->b_crypt_hdr
);
3552 ocache
= hdr_full_cache
;
3553 osize
= HDR_FULL_SIZE
;
3555 ncache
= hdr_full_cache
;
3556 nsize
= HDR_FULL_SIZE
;
3557 ocache
= hdr_full_crypt_cache
;
3558 osize
= sizeof (hdr
->b_crypt_hdr
);
3561 nhdr
= kmem_cache_alloc(ncache
, KM_PUSHPAGE
);
3564 * Copy all members that aren't locks or condvars to the new header.
3565 * No lists are pointing to us (as we asserted above), so we don't
3566 * need to worry about the list nodes.
3568 nhdr
->b_dva
= hdr
->b_dva
;
3569 nhdr
->b_birth
= hdr
->b_birth
;
3570 nhdr
->b_type
= hdr
->b_type
;
3571 nhdr
->b_flags
= hdr
->b_flags
;
3572 nhdr
->b_psize
= hdr
->b_psize
;
3573 nhdr
->b_lsize
= hdr
->b_lsize
;
3574 nhdr
->b_spa
= hdr
->b_spa
;
3575 nhdr
->b_l1hdr
.b_freeze_cksum
= hdr
->b_l1hdr
.b_freeze_cksum
;
3576 nhdr
->b_l1hdr
.b_bufcnt
= hdr
->b_l1hdr
.b_bufcnt
;
3577 nhdr
->b_l1hdr
.b_byteswap
= hdr
->b_l1hdr
.b_byteswap
;
3578 nhdr
->b_l1hdr
.b_state
= hdr
->b_l1hdr
.b_state
;
3579 nhdr
->b_l1hdr
.b_arc_access
= hdr
->b_l1hdr
.b_arc_access
;
3580 nhdr
->b_l1hdr
.b_mru_hits
= hdr
->b_l1hdr
.b_mru_hits
;
3581 nhdr
->b_l1hdr
.b_mru_ghost_hits
= hdr
->b_l1hdr
.b_mru_ghost_hits
;
3582 nhdr
->b_l1hdr
.b_mfu_hits
= hdr
->b_l1hdr
.b_mfu_hits
;
3583 nhdr
->b_l1hdr
.b_mfu_ghost_hits
= hdr
->b_l1hdr
.b_mfu_ghost_hits
;
3584 nhdr
->b_l1hdr
.b_l2_hits
= hdr
->b_l1hdr
.b_l2_hits
;
3585 nhdr
->b_l1hdr
.b_acb
= hdr
->b_l1hdr
.b_acb
;
3586 nhdr
->b_l1hdr
.b_pabd
= hdr
->b_l1hdr
.b_pabd
;
3589 * This zfs_refcount_add() exists only to ensure that the individual
3590 * arc buffers always point to a header that is referenced, avoiding
3591 * a small race condition that could trigger ASSERTs.
3593 (void) zfs_refcount_add(&nhdr
->b_l1hdr
.b_refcnt
, FTAG
);
3594 nhdr
->b_l1hdr
.b_buf
= hdr
->b_l1hdr
.b_buf
;
3595 for (buf
= nhdr
->b_l1hdr
.b_buf
; buf
!= NULL
; buf
= buf
->b_next
) {
3596 mutex_enter(&buf
->b_evict_lock
);
3598 mutex_exit(&buf
->b_evict_lock
);
3601 zfs_refcount_transfer(&nhdr
->b_l1hdr
.b_refcnt
, &hdr
->b_l1hdr
.b_refcnt
);
3602 (void) zfs_refcount_remove(&nhdr
->b_l1hdr
.b_refcnt
, FTAG
);
3603 ASSERT0(zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
));
3606 arc_hdr_set_flags(nhdr
, ARC_FLAG_PROTECTED
);
3608 arc_hdr_clear_flags(nhdr
, ARC_FLAG_PROTECTED
);
3611 /* unset all members of the original hdr */
3612 bzero(&hdr
->b_dva
, sizeof (dva_t
));
3614 hdr
->b_type
= ARC_BUFC_INVALID
;
3619 hdr
->b_l1hdr
.b_freeze_cksum
= NULL
;
3620 hdr
->b_l1hdr
.b_buf
= NULL
;
3621 hdr
->b_l1hdr
.b_bufcnt
= 0;
3622 hdr
->b_l1hdr
.b_byteswap
= 0;
3623 hdr
->b_l1hdr
.b_state
= NULL
;
3624 hdr
->b_l1hdr
.b_arc_access
= 0;
3625 hdr
->b_l1hdr
.b_mru_hits
= 0;
3626 hdr
->b_l1hdr
.b_mru_ghost_hits
= 0;
3627 hdr
->b_l1hdr
.b_mfu_hits
= 0;
3628 hdr
->b_l1hdr
.b_mfu_ghost_hits
= 0;
3629 hdr
->b_l1hdr
.b_l2_hits
= 0;
3630 hdr
->b_l1hdr
.b_acb
= NULL
;
3631 hdr
->b_l1hdr
.b_pabd
= NULL
;
3633 if (ocache
== hdr_full_crypt_cache
) {
3634 ASSERT(!HDR_HAS_RABD(hdr
));
3635 hdr
->b_crypt_hdr
.b_ot
= DMU_OT_NONE
;
3636 hdr
->b_crypt_hdr
.b_ebufcnt
= 0;
3637 hdr
->b_crypt_hdr
.b_dsobj
= 0;
3638 bzero(hdr
->b_crypt_hdr
.b_salt
, ZIO_DATA_SALT_LEN
);
3639 bzero(hdr
->b_crypt_hdr
.b_iv
, ZIO_DATA_IV_LEN
);
3640 bzero(hdr
->b_crypt_hdr
.b_mac
, ZIO_DATA_MAC_LEN
);
3643 buf_discard_identity(hdr
);
3644 kmem_cache_free(ocache
, hdr
);
3650 * This function is used by the send / receive code to convert a newly
3651 * allocated arc_buf_t to one that is suitable for a raw encrypted write. It
3652 * is also used to allow the root objset block to be uupdated without altering
3653 * its embedded MACs. Both block types will always be uncompressed so we do not
3654 * have to worry about compression type or psize.
3657 arc_convert_to_raw(arc_buf_t
*buf
, uint64_t dsobj
, boolean_t byteorder
,
3658 dmu_object_type_t ot
, const uint8_t *salt
, const uint8_t *iv
,
3661 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
3663 ASSERT(ot
== DMU_OT_DNODE
|| ot
== DMU_OT_OBJSET
);
3664 ASSERT(HDR_HAS_L1HDR(hdr
));
3665 ASSERT3P(hdr
->b_l1hdr
.b_state
, ==, arc_anon
);
3667 buf
->b_flags
|= (ARC_BUF_FLAG_COMPRESSED
| ARC_BUF_FLAG_ENCRYPTED
);
3668 if (!HDR_PROTECTED(hdr
))
3669 hdr
= arc_hdr_realloc_crypt(hdr
, B_TRUE
);
3670 hdr
->b_crypt_hdr
.b_dsobj
= dsobj
;
3671 hdr
->b_crypt_hdr
.b_ot
= ot
;
3672 hdr
->b_l1hdr
.b_byteswap
= (byteorder
== ZFS_HOST_BYTEORDER
) ?
3673 DMU_BSWAP_NUMFUNCS
: DMU_OT_BYTESWAP(ot
);
3674 if (!arc_hdr_has_uncompressed_buf(hdr
))
3675 arc_cksum_free(hdr
);
3678 bcopy(salt
, hdr
->b_crypt_hdr
.b_salt
, ZIO_DATA_SALT_LEN
);
3680 bcopy(iv
, hdr
->b_crypt_hdr
.b_iv
, ZIO_DATA_IV_LEN
);
3682 bcopy(mac
, hdr
->b_crypt_hdr
.b_mac
, ZIO_DATA_MAC_LEN
);
3686 * Allocate a new arc_buf_hdr_t and arc_buf_t and return the buf to the caller.
3687 * The buf is returned thawed since we expect the consumer to modify it.
3690 arc_alloc_buf(spa_t
*spa
, void *tag
, arc_buf_contents_t type
, int32_t size
)
3692 arc_buf_hdr_t
*hdr
= arc_hdr_alloc(spa_load_guid(spa
), size
, size
,
3693 B_FALSE
, ZIO_COMPRESS_OFF
, type
, B_FALSE
);
3694 ASSERT(!MUTEX_HELD(HDR_LOCK(hdr
)));
3696 arc_buf_t
*buf
= NULL
;
3697 VERIFY0(arc_buf_alloc_impl(hdr
, spa
, NULL
, tag
, B_FALSE
, B_FALSE
,
3698 B_FALSE
, B_FALSE
, &buf
));
3705 * Allocate a compressed buf in the same manner as arc_alloc_buf. Don't use this
3706 * for bufs containing metadata.
3709 arc_alloc_compressed_buf(spa_t
*spa
, void *tag
, uint64_t psize
, uint64_t lsize
,
3710 enum zio_compress compression_type
)
3712 ASSERT3U(lsize
, >, 0);
3713 ASSERT3U(lsize
, >=, psize
);
3714 ASSERT3U(compression_type
, >, ZIO_COMPRESS_OFF
);
3715 ASSERT3U(compression_type
, <, ZIO_COMPRESS_FUNCTIONS
);
3717 arc_buf_hdr_t
*hdr
= arc_hdr_alloc(spa_load_guid(spa
), psize
, lsize
,
3718 B_FALSE
, compression_type
, ARC_BUFC_DATA
, B_FALSE
);
3719 ASSERT(!MUTEX_HELD(HDR_LOCK(hdr
)));
3721 arc_buf_t
*buf
= NULL
;
3722 VERIFY0(arc_buf_alloc_impl(hdr
, spa
, NULL
, tag
, B_FALSE
,
3723 B_TRUE
, B_FALSE
, B_FALSE
, &buf
));
3725 ASSERT3P(hdr
->b_l1hdr
.b_freeze_cksum
, ==, NULL
);
3727 if (!arc_buf_is_shared(buf
)) {
3729 * To ensure that the hdr has the correct data in it if we call
3730 * arc_untransform() on this buf before it's been written to
3731 * disk, it's easiest if we just set up sharing between the
3734 ASSERT(!abd_is_linear(hdr
->b_l1hdr
.b_pabd
));
3735 arc_hdr_free_abd(hdr
, B_FALSE
);
3736 arc_share_buf(hdr
, buf
);
3743 arc_alloc_raw_buf(spa_t
*spa
, void *tag
, uint64_t dsobj
, boolean_t byteorder
,
3744 const uint8_t *salt
, const uint8_t *iv
, const uint8_t *mac
,
3745 dmu_object_type_t ot
, uint64_t psize
, uint64_t lsize
,
3746 enum zio_compress compression_type
)
3750 arc_buf_contents_t type
= DMU_OT_IS_METADATA(ot
) ?
3751 ARC_BUFC_METADATA
: ARC_BUFC_DATA
;
3753 ASSERT3U(lsize
, >, 0);
3754 ASSERT3U(lsize
, >=, psize
);
3755 ASSERT3U(compression_type
, >=, ZIO_COMPRESS_OFF
);
3756 ASSERT3U(compression_type
, <, ZIO_COMPRESS_FUNCTIONS
);
3758 hdr
= arc_hdr_alloc(spa_load_guid(spa
), psize
, lsize
, B_TRUE
,
3759 compression_type
, type
, B_TRUE
);
3760 ASSERT(!MUTEX_HELD(HDR_LOCK(hdr
)));
3762 hdr
->b_crypt_hdr
.b_dsobj
= dsobj
;
3763 hdr
->b_crypt_hdr
.b_ot
= ot
;
3764 hdr
->b_l1hdr
.b_byteswap
= (byteorder
== ZFS_HOST_BYTEORDER
) ?
3765 DMU_BSWAP_NUMFUNCS
: DMU_OT_BYTESWAP(ot
);
3766 bcopy(salt
, hdr
->b_crypt_hdr
.b_salt
, ZIO_DATA_SALT_LEN
);
3767 bcopy(iv
, hdr
->b_crypt_hdr
.b_iv
, ZIO_DATA_IV_LEN
);
3768 bcopy(mac
, hdr
->b_crypt_hdr
.b_mac
, ZIO_DATA_MAC_LEN
);
3771 * This buffer will be considered encrypted even if the ot is not an
3772 * encrypted type. It will become authenticated instead in
3773 * arc_write_ready().
3776 VERIFY0(arc_buf_alloc_impl(hdr
, spa
, NULL
, tag
, B_TRUE
, B_TRUE
,
3777 B_FALSE
, B_FALSE
, &buf
));
3779 ASSERT3P(hdr
->b_l1hdr
.b_freeze_cksum
, ==, NULL
);
3785 arc_hdr_l2hdr_destroy(arc_buf_hdr_t
*hdr
)
3787 l2arc_buf_hdr_t
*l2hdr
= &hdr
->b_l2hdr
;
3788 l2arc_dev_t
*dev
= l2hdr
->b_dev
;
3789 uint64_t psize
= arc_hdr_size(hdr
);
3791 ASSERT(MUTEX_HELD(&dev
->l2ad_mtx
));
3792 ASSERT(HDR_HAS_L2HDR(hdr
));
3794 list_remove(&dev
->l2ad_buflist
, hdr
);
3796 ARCSTAT_INCR(arcstat_l2_psize
, -psize
);
3797 ARCSTAT_INCR(arcstat_l2_lsize
, -HDR_GET_LSIZE(hdr
));
3799 vdev_space_update(dev
->l2ad_vdev
, -psize
, 0, 0);
3801 (void) zfs_refcount_remove_many(&dev
->l2ad_alloc
, psize
, hdr
);
3802 arc_hdr_clear_flags(hdr
, ARC_FLAG_HAS_L2HDR
);
3806 arc_hdr_destroy(arc_buf_hdr_t
*hdr
)
3808 if (HDR_HAS_L1HDR(hdr
)) {
3809 ASSERT(hdr
->b_l1hdr
.b_buf
== NULL
||
3810 hdr
->b_l1hdr
.b_bufcnt
> 0);
3811 ASSERT(zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
));
3812 ASSERT3P(hdr
->b_l1hdr
.b_state
, ==, arc_anon
);
3814 ASSERT(!HDR_IO_IN_PROGRESS(hdr
));
3815 ASSERT(!HDR_IN_HASH_TABLE(hdr
));
3817 if (!HDR_EMPTY(hdr
))
3818 buf_discard_identity(hdr
);
3820 if (HDR_HAS_L2HDR(hdr
)) {
3821 l2arc_dev_t
*dev
= hdr
->b_l2hdr
.b_dev
;
3822 boolean_t buflist_held
= MUTEX_HELD(&dev
->l2ad_mtx
);
3825 mutex_enter(&dev
->l2ad_mtx
);
3828 * Even though we checked this conditional above, we
3829 * need to check this again now that we have the
3830 * l2ad_mtx. This is because we could be racing with
3831 * another thread calling l2arc_evict() which might have
3832 * destroyed this header's L2 portion as we were waiting
3833 * to acquire the l2ad_mtx. If that happens, we don't
3834 * want to re-destroy the header's L2 portion.
3836 if (HDR_HAS_L2HDR(hdr
))
3837 arc_hdr_l2hdr_destroy(hdr
);
3840 mutex_exit(&dev
->l2ad_mtx
);
3843 if (HDR_HAS_L1HDR(hdr
)) {
3844 arc_cksum_free(hdr
);
3846 while (hdr
->b_l1hdr
.b_buf
!= NULL
)
3847 arc_buf_destroy_impl(hdr
->b_l1hdr
.b_buf
);
3849 if (hdr
->b_l1hdr
.b_pabd
!= NULL
) {
3850 arc_hdr_free_abd(hdr
, B_FALSE
);
3853 if (HDR_HAS_RABD(hdr
))
3854 arc_hdr_free_abd(hdr
, B_TRUE
);
3857 ASSERT3P(hdr
->b_hash_next
, ==, NULL
);
3858 if (HDR_HAS_L1HDR(hdr
)) {
3859 ASSERT(!multilist_link_active(&hdr
->b_l1hdr
.b_arc_node
));
3860 ASSERT3P(hdr
->b_l1hdr
.b_acb
, ==, NULL
);
3862 if (!HDR_PROTECTED(hdr
)) {
3863 kmem_cache_free(hdr_full_cache
, hdr
);
3865 kmem_cache_free(hdr_full_crypt_cache
, hdr
);
3868 kmem_cache_free(hdr_l2only_cache
, hdr
);
3873 arc_buf_destroy(arc_buf_t
*buf
, void* tag
)
3875 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
3876 kmutex_t
*hash_lock
= HDR_LOCK(hdr
);
3878 if (hdr
->b_l1hdr
.b_state
== arc_anon
) {
3879 ASSERT3U(hdr
->b_l1hdr
.b_bufcnt
, ==, 1);
3880 ASSERT(!HDR_IO_IN_PROGRESS(hdr
));
3881 VERIFY0(remove_reference(hdr
, NULL
, tag
));
3882 arc_hdr_destroy(hdr
);
3886 mutex_enter(hash_lock
);
3887 ASSERT3P(hdr
, ==, buf
->b_hdr
);
3888 ASSERT(hdr
->b_l1hdr
.b_bufcnt
> 0);
3889 ASSERT3P(hash_lock
, ==, HDR_LOCK(hdr
));
3890 ASSERT3P(hdr
->b_l1hdr
.b_state
, !=, arc_anon
);
3891 ASSERT3P(buf
->b_data
, !=, NULL
);
3893 (void) remove_reference(hdr
, hash_lock
, tag
);
3894 arc_buf_destroy_impl(buf
);
3895 mutex_exit(hash_lock
);
3899 * Evict the arc_buf_hdr that is provided as a parameter. The resultant
3900 * state of the header is dependent on its state prior to entering this
3901 * function. The following transitions are possible:
3903 * - arc_mru -> arc_mru_ghost
3904 * - arc_mfu -> arc_mfu_ghost
3905 * - arc_mru_ghost -> arc_l2c_only
3906 * - arc_mru_ghost -> deleted
3907 * - arc_mfu_ghost -> arc_l2c_only
3908 * - arc_mfu_ghost -> deleted
3911 arc_evict_hdr(arc_buf_hdr_t
*hdr
, kmutex_t
*hash_lock
)
3913 arc_state_t
*evicted_state
, *state
;
3914 int64_t bytes_evicted
= 0;
3915 int min_lifetime
= HDR_PRESCIENT_PREFETCH(hdr
) ?
3916 arc_min_prescient_prefetch_ms
: arc_min_prefetch_ms
;
3918 ASSERT(MUTEX_HELD(hash_lock
));
3919 ASSERT(HDR_HAS_L1HDR(hdr
));
3921 state
= hdr
->b_l1hdr
.b_state
;
3922 if (GHOST_STATE(state
)) {
3923 ASSERT(!HDR_IO_IN_PROGRESS(hdr
));
3924 ASSERT3P(hdr
->b_l1hdr
.b_buf
, ==, NULL
);
3927 * l2arc_write_buffers() relies on a header's L1 portion
3928 * (i.e. its b_pabd field) during it's write phase.
3929 * Thus, we cannot push a header onto the arc_l2c_only
3930 * state (removing its L1 piece) until the header is
3931 * done being written to the l2arc.
3933 if (HDR_HAS_L2HDR(hdr
) && HDR_L2_WRITING(hdr
)) {
3934 ARCSTAT_BUMP(arcstat_evict_l2_skip
);
3935 return (bytes_evicted
);
3938 ARCSTAT_BUMP(arcstat_deleted
);
3939 bytes_evicted
+= HDR_GET_LSIZE(hdr
);
3941 DTRACE_PROBE1(arc__delete
, arc_buf_hdr_t
*, hdr
);
3943 if (HDR_HAS_L2HDR(hdr
)) {
3944 ASSERT(hdr
->b_l1hdr
.b_pabd
== NULL
);
3945 ASSERT(!HDR_HAS_RABD(hdr
));
3947 * This buffer is cached on the 2nd Level ARC;
3948 * don't destroy the header.
3950 arc_change_state(arc_l2c_only
, hdr
, hash_lock
);
3952 * dropping from L1+L2 cached to L2-only,
3953 * realloc to remove the L1 header.
3955 hdr
= arc_hdr_realloc(hdr
, hdr_full_cache
,
3958 arc_change_state(arc_anon
, hdr
, hash_lock
);
3959 arc_hdr_destroy(hdr
);
3961 return (bytes_evicted
);
3964 ASSERT(state
== arc_mru
|| state
== arc_mfu
);
3965 evicted_state
= (state
== arc_mru
) ? arc_mru_ghost
: arc_mfu_ghost
;
3967 /* prefetch buffers have a minimum lifespan */
3968 if (HDR_IO_IN_PROGRESS(hdr
) ||
3969 ((hdr
->b_flags
& (ARC_FLAG_PREFETCH
| ARC_FLAG_INDIRECT
)) &&
3970 ddi_get_lbolt() - hdr
->b_l1hdr
.b_arc_access
<
3971 MSEC_TO_TICK(min_lifetime
))) {
3972 ARCSTAT_BUMP(arcstat_evict_skip
);
3973 return (bytes_evicted
);
3976 ASSERT0(zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
));
3977 while (hdr
->b_l1hdr
.b_buf
) {
3978 arc_buf_t
*buf
= hdr
->b_l1hdr
.b_buf
;
3979 if (!mutex_tryenter(&buf
->b_evict_lock
)) {
3980 ARCSTAT_BUMP(arcstat_mutex_miss
);
3983 if (buf
->b_data
!= NULL
)
3984 bytes_evicted
+= HDR_GET_LSIZE(hdr
);
3985 mutex_exit(&buf
->b_evict_lock
);
3986 arc_buf_destroy_impl(buf
);
3989 if (HDR_HAS_L2HDR(hdr
)) {
3990 ARCSTAT_INCR(arcstat_evict_l2_cached
, HDR_GET_LSIZE(hdr
));
3992 if (l2arc_write_eligible(hdr
->b_spa
, hdr
)) {
3993 ARCSTAT_INCR(arcstat_evict_l2_eligible
,
3994 HDR_GET_LSIZE(hdr
));
3996 ARCSTAT_INCR(arcstat_evict_l2_ineligible
,
3997 HDR_GET_LSIZE(hdr
));
4001 if (hdr
->b_l1hdr
.b_bufcnt
== 0) {
4002 arc_cksum_free(hdr
);
4004 bytes_evicted
+= arc_hdr_size(hdr
);
4007 * If this hdr is being evicted and has a compressed
4008 * buffer then we discard it here before we change states.
4009 * This ensures that the accounting is updated correctly
4010 * in arc_free_data_impl().
4012 if (hdr
->b_l1hdr
.b_pabd
!= NULL
)
4013 arc_hdr_free_abd(hdr
, B_FALSE
);
4015 if (HDR_HAS_RABD(hdr
))
4016 arc_hdr_free_abd(hdr
, B_TRUE
);
4018 arc_change_state(evicted_state
, hdr
, hash_lock
);
4019 ASSERT(HDR_IN_HASH_TABLE(hdr
));
4020 arc_hdr_set_flags(hdr
, ARC_FLAG_IN_HASH_TABLE
);
4021 DTRACE_PROBE1(arc__evict
, arc_buf_hdr_t
*, hdr
);
4024 return (bytes_evicted
);
4028 arc_evict_state_impl(multilist_t
*ml
, int idx
, arc_buf_hdr_t
*marker
,
4029 uint64_t spa
, int64_t bytes
)
4031 multilist_sublist_t
*mls
;
4032 uint64_t bytes_evicted
= 0;
4034 kmutex_t
*hash_lock
;
4035 int evict_count
= 0;
4037 ASSERT3P(marker
, !=, NULL
);
4038 IMPLY(bytes
< 0, bytes
== ARC_EVICT_ALL
);
4040 mls
= multilist_sublist_lock(ml
, idx
);
4042 for (hdr
= multilist_sublist_prev(mls
, marker
); hdr
!= NULL
;
4043 hdr
= multilist_sublist_prev(mls
, marker
)) {
4044 if ((bytes
!= ARC_EVICT_ALL
&& bytes_evicted
>= bytes
) ||
4045 (evict_count
>= zfs_arc_evict_batch_limit
))
4049 * To keep our iteration location, move the marker
4050 * forward. Since we're not holding hdr's hash lock, we
4051 * must be very careful and not remove 'hdr' from the
4052 * sublist. Otherwise, other consumers might mistake the
4053 * 'hdr' as not being on a sublist when they call the
4054 * multilist_link_active() function (they all rely on
4055 * the hash lock protecting concurrent insertions and
4056 * removals). multilist_sublist_move_forward() was
4057 * specifically implemented to ensure this is the case
4058 * (only 'marker' will be removed and re-inserted).
4060 multilist_sublist_move_forward(mls
, marker
);
4063 * The only case where the b_spa field should ever be
4064 * zero, is the marker headers inserted by
4065 * arc_evict_state(). It's possible for multiple threads
4066 * to be calling arc_evict_state() concurrently (e.g.
4067 * dsl_pool_close() and zio_inject_fault()), so we must
4068 * skip any markers we see from these other threads.
4070 if (hdr
->b_spa
== 0)
4073 /* we're only interested in evicting buffers of a certain spa */
4074 if (spa
!= 0 && hdr
->b_spa
!= spa
) {
4075 ARCSTAT_BUMP(arcstat_evict_skip
);
4079 hash_lock
= HDR_LOCK(hdr
);
4082 * We aren't calling this function from any code path
4083 * that would already be holding a hash lock, so we're
4084 * asserting on this assumption to be defensive in case
4085 * this ever changes. Without this check, it would be
4086 * possible to incorrectly increment arcstat_mutex_miss
4087 * below (e.g. if the code changed such that we called
4088 * this function with a hash lock held).
4090 ASSERT(!MUTEX_HELD(hash_lock
));
4092 if (mutex_tryenter(hash_lock
)) {
4093 uint64_t evicted
= arc_evict_hdr(hdr
, hash_lock
);
4094 mutex_exit(hash_lock
);
4096 bytes_evicted
+= evicted
;
4099 * If evicted is zero, arc_evict_hdr() must have
4100 * decided to skip this header, don't increment
4101 * evict_count in this case.
4107 * If arc_size isn't overflowing, signal any
4108 * threads that might happen to be waiting.
4110 * For each header evicted, we wake up a single
4111 * thread. If we used cv_broadcast, we could
4112 * wake up "too many" threads causing arc_size
4113 * to significantly overflow arc_c; since
4114 * arc_get_data_impl() doesn't check for overflow
4115 * when it's woken up (it doesn't because it's
4116 * possible for the ARC to be overflowing while
4117 * full of un-evictable buffers, and the
4118 * function should proceed in this case).
4120 * If threads are left sleeping, due to not
4121 * using cv_broadcast here, they will be woken
4122 * up via cv_broadcast in arc_adjust_cb() just
4123 * before arc_adjust_zthr sleeps.
4125 mutex_enter(&arc_adjust_lock
);
4126 if (!arc_is_overflowing())
4127 cv_signal(&arc_adjust_waiters_cv
);
4128 mutex_exit(&arc_adjust_lock
);
4130 ARCSTAT_BUMP(arcstat_mutex_miss
);
4134 multilist_sublist_unlock(mls
);
4136 return (bytes_evicted
);
4140 * Evict buffers from the given arc state, until we've removed the
4141 * specified number of bytes. Move the removed buffers to the
4142 * appropriate evict state.
4144 * This function makes a "best effort". It skips over any buffers
4145 * it can't get a hash_lock on, and so, may not catch all candidates.
4146 * It may also return without evicting as much space as requested.
4148 * If bytes is specified using the special value ARC_EVICT_ALL, this
4149 * will evict all available (i.e. unlocked and evictable) buffers from
4150 * the given arc state; which is used by arc_flush().
4153 arc_evict_state(arc_state_t
*state
, uint64_t spa
, int64_t bytes
,
4154 arc_buf_contents_t type
)
4156 uint64_t total_evicted
= 0;
4157 multilist_t
*ml
= state
->arcs_list
[type
];
4159 arc_buf_hdr_t
**markers
;
4161 IMPLY(bytes
< 0, bytes
== ARC_EVICT_ALL
);
4163 num_sublists
= multilist_get_num_sublists(ml
);
4166 * If we've tried to evict from each sublist, made some
4167 * progress, but still have not hit the target number of bytes
4168 * to evict, we want to keep trying. The markers allow us to
4169 * pick up where we left off for each individual sublist, rather
4170 * than starting from the tail each time.
4172 markers
= kmem_zalloc(sizeof (*markers
) * num_sublists
, KM_SLEEP
);
4173 for (int i
= 0; i
< num_sublists
; i
++) {
4174 multilist_sublist_t
*mls
;
4176 markers
[i
] = kmem_cache_alloc(hdr_full_cache
, KM_SLEEP
);
4179 * A b_spa of 0 is used to indicate that this header is
4180 * a marker. This fact is used in arc_adjust_type() and
4181 * arc_evict_state_impl().
4183 markers
[i
]->b_spa
= 0;
4185 mls
= multilist_sublist_lock(ml
, i
);
4186 multilist_sublist_insert_tail(mls
, markers
[i
]);
4187 multilist_sublist_unlock(mls
);
4191 * While we haven't hit our target number of bytes to evict, or
4192 * we're evicting all available buffers.
4194 while (total_evicted
< bytes
|| bytes
== ARC_EVICT_ALL
) {
4195 int sublist_idx
= multilist_get_random_index(ml
);
4196 uint64_t scan_evicted
= 0;
4199 * Try to reduce pinned dnodes with a floor of arc_dnode_limit.
4200 * Request that 10% of the LRUs be scanned by the superblock
4203 if (type
== ARC_BUFC_DATA
&& aggsum_compare(&astat_dnode_size
,
4204 arc_dnode_limit
) > 0) {
4205 arc_prune_async((aggsum_upper_bound(&astat_dnode_size
) -
4206 arc_dnode_limit
) / sizeof (dnode_t
) /
4207 zfs_arc_dnode_reduce_percent
);
4211 * Start eviction using a randomly selected sublist,
4212 * this is to try and evenly balance eviction across all
4213 * sublists. Always starting at the same sublist
4214 * (e.g. index 0) would cause evictions to favor certain
4215 * sublists over others.
4217 for (int i
= 0; i
< num_sublists
; i
++) {
4218 uint64_t bytes_remaining
;
4219 uint64_t bytes_evicted
;
4221 if (bytes
== ARC_EVICT_ALL
)
4222 bytes_remaining
= ARC_EVICT_ALL
;
4223 else if (total_evicted
< bytes
)
4224 bytes_remaining
= bytes
- total_evicted
;
4228 bytes_evicted
= arc_evict_state_impl(ml
, sublist_idx
,
4229 markers
[sublist_idx
], spa
, bytes_remaining
);
4231 scan_evicted
+= bytes_evicted
;
4232 total_evicted
+= bytes_evicted
;
4234 /* we've reached the end, wrap to the beginning */
4235 if (++sublist_idx
>= num_sublists
)
4240 * If we didn't evict anything during this scan, we have
4241 * no reason to believe we'll evict more during another
4242 * scan, so break the loop.
4244 if (scan_evicted
== 0) {
4245 /* This isn't possible, let's make that obvious */
4246 ASSERT3S(bytes
, !=, 0);
4249 * When bytes is ARC_EVICT_ALL, the only way to
4250 * break the loop is when scan_evicted is zero.
4251 * In that case, we actually have evicted enough,
4252 * so we don't want to increment the kstat.
4254 if (bytes
!= ARC_EVICT_ALL
) {
4255 ASSERT3S(total_evicted
, <, bytes
);
4256 ARCSTAT_BUMP(arcstat_evict_not_enough
);
4263 for (int i
= 0; i
< num_sublists
; i
++) {
4264 multilist_sublist_t
*mls
= multilist_sublist_lock(ml
, i
);
4265 multilist_sublist_remove(mls
, markers
[i
]);
4266 multilist_sublist_unlock(mls
);
4268 kmem_cache_free(hdr_full_cache
, markers
[i
]);
4270 kmem_free(markers
, sizeof (*markers
) * num_sublists
);
4272 return (total_evicted
);
4276 * Flush all "evictable" data of the given type from the arc state
4277 * specified. This will not evict any "active" buffers (i.e. referenced).
4279 * When 'retry' is set to B_FALSE, the function will make a single pass
4280 * over the state and evict any buffers that it can. Since it doesn't
4281 * continually retry the eviction, it might end up leaving some buffers
4282 * in the ARC due to lock misses.
4284 * When 'retry' is set to B_TRUE, the function will continually retry the
4285 * eviction until *all* evictable buffers have been removed from the
4286 * state. As a result, if concurrent insertions into the state are
4287 * allowed (e.g. if the ARC isn't shutting down), this function might
4288 * wind up in an infinite loop, continually trying to evict buffers.
4291 arc_flush_state(arc_state_t
*state
, uint64_t spa
, arc_buf_contents_t type
,
4294 uint64_t evicted
= 0;
4296 while (zfs_refcount_count(&state
->arcs_esize
[type
]) != 0) {
4297 evicted
+= arc_evict_state(state
, spa
, ARC_EVICT_ALL
, type
);
4307 * Helper function for arc_prune_async() it is responsible for safely
4308 * handling the execution of a registered arc_prune_func_t.
4311 arc_prune_task(void *ptr
)
4313 arc_prune_t
*ap
= (arc_prune_t
*)ptr
;
4314 arc_prune_func_t
*func
= ap
->p_pfunc
;
4317 func(ap
->p_adjust
, ap
->p_private
);
4319 zfs_refcount_remove(&ap
->p_refcnt
, func
);
4323 * Notify registered consumers they must drop holds on a portion of the ARC
4324 * buffered they reference. This provides a mechanism to ensure the ARC can
4325 * honor the arc_meta_limit and reclaim otherwise pinned ARC buffers. This
4326 * is analogous to dnlc_reduce_cache() but more generic.
4328 * This operation is performed asynchronously so it may be safely called
4329 * in the context of the arc_reclaim_thread(). A reference is taken here
4330 * for each registered arc_prune_t and the arc_prune_task() is responsible
4331 * for releasing it once the registered arc_prune_func_t has completed.
4334 arc_prune_async(int64_t adjust
)
4338 mutex_enter(&arc_prune_mtx
);
4339 for (ap
= list_head(&arc_prune_list
); ap
!= NULL
;
4340 ap
= list_next(&arc_prune_list
, ap
)) {
4342 if (zfs_refcount_count(&ap
->p_refcnt
) >= 2)
4345 zfs_refcount_add(&ap
->p_refcnt
, ap
->p_pfunc
);
4346 ap
->p_adjust
= adjust
;
4347 if (taskq_dispatch(arc_prune_taskq
, arc_prune_task
,
4348 ap
, TQ_SLEEP
) == TASKQID_INVALID
) {
4349 zfs_refcount_remove(&ap
->p_refcnt
, ap
->p_pfunc
);
4352 ARCSTAT_BUMP(arcstat_prune
);
4354 mutex_exit(&arc_prune_mtx
);
4358 * Evict the specified number of bytes from the state specified,
4359 * restricting eviction to the spa and type given. This function
4360 * prevents us from trying to evict more from a state's list than
4361 * is "evictable", and to skip evicting altogether when passed a
4362 * negative value for "bytes". In contrast, arc_evict_state() will
4363 * evict everything it can, when passed a negative value for "bytes".
4366 arc_adjust_impl(arc_state_t
*state
, uint64_t spa
, int64_t bytes
,
4367 arc_buf_contents_t type
)
4371 if (bytes
> 0 && zfs_refcount_count(&state
->arcs_esize
[type
]) > 0) {
4372 delta
= MIN(zfs_refcount_count(&state
->arcs_esize
[type
]),
4374 return (arc_evict_state(state
, spa
, delta
, type
));
4381 * The goal of this function is to evict enough meta data buffers from the
4382 * ARC in order to enforce the arc_meta_limit. Achieving this is slightly
4383 * more complicated than it appears because it is common for data buffers
4384 * to have holds on meta data buffers. In addition, dnode meta data buffers
4385 * will be held by the dnodes in the block preventing them from being freed.
4386 * This means we can't simply traverse the ARC and expect to always find
4387 * enough unheld meta data buffer to release.
4389 * Therefore, this function has been updated to make alternating passes
4390 * over the ARC releasing data buffers and then newly unheld meta data
4391 * buffers. This ensures forward progress is maintained and meta_used
4392 * will decrease. Normally this is sufficient, but if required the ARC
4393 * will call the registered prune callbacks causing dentry and inodes to
4394 * be dropped from the VFS cache. This will make dnode meta data buffers
4395 * available for reclaim.
4398 arc_adjust_meta_balanced(uint64_t meta_used
)
4400 int64_t delta
, prune
= 0, adjustmnt
;
4401 uint64_t total_evicted
= 0;
4402 arc_buf_contents_t type
= ARC_BUFC_DATA
;
4403 int restarts
= MAX(zfs_arc_meta_adjust_restarts
, 0);
4407 * This slightly differs than the way we evict from the mru in
4408 * arc_adjust because we don't have a "target" value (i.e. no
4409 * "meta" arc_p). As a result, I think we can completely
4410 * cannibalize the metadata in the MRU before we evict the
4411 * metadata from the MFU. I think we probably need to implement a
4412 * "metadata arc_p" value to do this properly.
4414 adjustmnt
= meta_used
- arc_meta_limit
;
4416 if (adjustmnt
> 0 &&
4417 zfs_refcount_count(&arc_mru
->arcs_esize
[type
]) > 0) {
4418 delta
= MIN(zfs_refcount_count(&arc_mru
->arcs_esize
[type
]),
4420 total_evicted
+= arc_adjust_impl(arc_mru
, 0, delta
, type
);
4425 * We can't afford to recalculate adjustmnt here. If we do,
4426 * new metadata buffers can sneak into the MRU or ANON lists,
4427 * thus penalize the MFU metadata. Although the fudge factor is
4428 * small, it has been empirically shown to be significant for
4429 * certain workloads (e.g. creating many empty directories). As
4430 * such, we use the original calculation for adjustmnt, and
4431 * simply decrement the amount of data evicted from the MRU.
4434 if (adjustmnt
> 0 &&
4435 zfs_refcount_count(&arc_mfu
->arcs_esize
[type
]) > 0) {
4436 delta
= MIN(zfs_refcount_count(&arc_mfu
->arcs_esize
[type
]),
4438 total_evicted
+= arc_adjust_impl(arc_mfu
, 0, delta
, type
);
4441 adjustmnt
= meta_used
- arc_meta_limit
;
4443 if (adjustmnt
> 0 &&
4444 zfs_refcount_count(&arc_mru_ghost
->arcs_esize
[type
]) > 0) {
4445 delta
= MIN(adjustmnt
,
4446 zfs_refcount_count(&arc_mru_ghost
->arcs_esize
[type
]));
4447 total_evicted
+= arc_adjust_impl(arc_mru_ghost
, 0, delta
, type
);
4451 if (adjustmnt
> 0 &&
4452 zfs_refcount_count(&arc_mfu_ghost
->arcs_esize
[type
]) > 0) {
4453 delta
= MIN(adjustmnt
,
4454 zfs_refcount_count(&arc_mfu_ghost
->arcs_esize
[type
]));
4455 total_evicted
+= arc_adjust_impl(arc_mfu_ghost
, 0, delta
, type
);
4459 * If after attempting to make the requested adjustment to the ARC
4460 * the meta limit is still being exceeded then request that the
4461 * higher layers drop some cached objects which have holds on ARC
4462 * meta buffers. Requests to the upper layers will be made with
4463 * increasingly large scan sizes until the ARC is below the limit.
4465 if (meta_used
> arc_meta_limit
) {
4466 if (type
== ARC_BUFC_DATA
) {
4467 type
= ARC_BUFC_METADATA
;
4469 type
= ARC_BUFC_DATA
;
4471 if (zfs_arc_meta_prune
) {
4472 prune
+= zfs_arc_meta_prune
;
4473 arc_prune_async(prune
);
4482 return (total_evicted
);
4486 * Evict metadata buffers from the cache, such that arc_meta_used is
4487 * capped by the arc_meta_limit tunable.
4490 arc_adjust_meta_only(uint64_t meta_used
)
4492 uint64_t total_evicted
= 0;
4496 * If we're over the meta limit, we want to evict enough
4497 * metadata to get back under the meta limit. We don't want to
4498 * evict so much that we drop the MRU below arc_p, though. If
4499 * we're over the meta limit more than we're over arc_p, we
4500 * evict some from the MRU here, and some from the MFU below.
4502 target
= MIN((int64_t)(meta_used
- arc_meta_limit
),
4503 (int64_t)(zfs_refcount_count(&arc_anon
->arcs_size
) +
4504 zfs_refcount_count(&arc_mru
->arcs_size
) - arc_p
));
4506 total_evicted
+= arc_adjust_impl(arc_mru
, 0, target
, ARC_BUFC_METADATA
);
4509 * Similar to the above, we want to evict enough bytes to get us
4510 * below the meta limit, but not so much as to drop us below the
4511 * space allotted to the MFU (which is defined as arc_c - arc_p).
4513 target
= MIN((int64_t)(meta_used
- arc_meta_limit
),
4514 (int64_t)(zfs_refcount_count(&arc_mfu
->arcs_size
) -
4517 total_evicted
+= arc_adjust_impl(arc_mfu
, 0, target
, ARC_BUFC_METADATA
);
4519 return (total_evicted
);
4523 arc_adjust_meta(uint64_t meta_used
)
4525 if (zfs_arc_meta_strategy
== ARC_STRATEGY_META_ONLY
)
4526 return (arc_adjust_meta_only(meta_used
));
4528 return (arc_adjust_meta_balanced(meta_used
));
4532 * Return the type of the oldest buffer in the given arc state
4534 * This function will select a random sublist of type ARC_BUFC_DATA and
4535 * a random sublist of type ARC_BUFC_METADATA. The tail of each sublist
4536 * is compared, and the type which contains the "older" buffer will be
4539 static arc_buf_contents_t
4540 arc_adjust_type(arc_state_t
*state
)
4542 multilist_t
*data_ml
= state
->arcs_list
[ARC_BUFC_DATA
];
4543 multilist_t
*meta_ml
= state
->arcs_list
[ARC_BUFC_METADATA
];
4544 int data_idx
= multilist_get_random_index(data_ml
);
4545 int meta_idx
= multilist_get_random_index(meta_ml
);
4546 multilist_sublist_t
*data_mls
;
4547 multilist_sublist_t
*meta_mls
;
4548 arc_buf_contents_t type
;
4549 arc_buf_hdr_t
*data_hdr
;
4550 arc_buf_hdr_t
*meta_hdr
;
4553 * We keep the sublist lock until we're finished, to prevent
4554 * the headers from being destroyed via arc_evict_state().
4556 data_mls
= multilist_sublist_lock(data_ml
, data_idx
);
4557 meta_mls
= multilist_sublist_lock(meta_ml
, meta_idx
);
4560 * These two loops are to ensure we skip any markers that
4561 * might be at the tail of the lists due to arc_evict_state().
4564 for (data_hdr
= multilist_sublist_tail(data_mls
); data_hdr
!= NULL
;
4565 data_hdr
= multilist_sublist_prev(data_mls
, data_hdr
)) {
4566 if (data_hdr
->b_spa
!= 0)
4570 for (meta_hdr
= multilist_sublist_tail(meta_mls
); meta_hdr
!= NULL
;
4571 meta_hdr
= multilist_sublist_prev(meta_mls
, meta_hdr
)) {
4572 if (meta_hdr
->b_spa
!= 0)
4576 if (data_hdr
== NULL
&& meta_hdr
== NULL
) {
4577 type
= ARC_BUFC_DATA
;
4578 } else if (data_hdr
== NULL
) {
4579 ASSERT3P(meta_hdr
, !=, NULL
);
4580 type
= ARC_BUFC_METADATA
;
4581 } else if (meta_hdr
== NULL
) {
4582 ASSERT3P(data_hdr
, !=, NULL
);
4583 type
= ARC_BUFC_DATA
;
4585 ASSERT3P(data_hdr
, !=, NULL
);
4586 ASSERT3P(meta_hdr
, !=, NULL
);
4588 /* The headers can't be on the sublist without an L1 header */
4589 ASSERT(HDR_HAS_L1HDR(data_hdr
));
4590 ASSERT(HDR_HAS_L1HDR(meta_hdr
));
4592 if (data_hdr
->b_l1hdr
.b_arc_access
<
4593 meta_hdr
->b_l1hdr
.b_arc_access
) {
4594 type
= ARC_BUFC_DATA
;
4596 type
= ARC_BUFC_METADATA
;
4600 multilist_sublist_unlock(meta_mls
);
4601 multilist_sublist_unlock(data_mls
);
4607 * Evict buffers from the cache, such that arc_size is capped by arc_c.
4612 uint64_t total_evicted
= 0;
4615 uint64_t asize
= aggsum_value(&arc_size
);
4616 uint64_t ameta
= aggsum_value(&arc_meta_used
);
4619 * If we're over arc_meta_limit, we want to correct that before
4620 * potentially evicting data buffers below.
4622 total_evicted
+= arc_adjust_meta(ameta
);
4627 * If we're over the target cache size, we want to evict enough
4628 * from the list to get back to our target size. We don't want
4629 * to evict too much from the MRU, such that it drops below
4630 * arc_p. So, if we're over our target cache size more than
4631 * the MRU is over arc_p, we'll evict enough to get back to
4632 * arc_p here, and then evict more from the MFU below.
4634 target
= MIN((int64_t)(asize
- arc_c
),
4635 (int64_t)(zfs_refcount_count(&arc_anon
->arcs_size
) +
4636 zfs_refcount_count(&arc_mru
->arcs_size
) + ameta
- arc_p
));
4639 * If we're below arc_meta_min, always prefer to evict data.
4640 * Otherwise, try to satisfy the requested number of bytes to
4641 * evict from the type which contains older buffers; in an
4642 * effort to keep newer buffers in the cache regardless of their
4643 * type. If we cannot satisfy the number of bytes from this
4644 * type, spill over into the next type.
4646 if (arc_adjust_type(arc_mru
) == ARC_BUFC_METADATA
&&
4647 ameta
> arc_meta_min
) {
4648 bytes
= arc_adjust_impl(arc_mru
, 0, target
, ARC_BUFC_METADATA
);
4649 total_evicted
+= bytes
;
4652 * If we couldn't evict our target number of bytes from
4653 * metadata, we try to get the rest from data.
4658 arc_adjust_impl(arc_mru
, 0, target
, ARC_BUFC_DATA
);
4660 bytes
= arc_adjust_impl(arc_mru
, 0, target
, ARC_BUFC_DATA
);
4661 total_evicted
+= bytes
;
4664 * If we couldn't evict our target number of bytes from
4665 * data, we try to get the rest from metadata.
4670 arc_adjust_impl(arc_mru
, 0, target
, ARC_BUFC_METADATA
);
4674 * Re-sum ARC stats after the first round of evictions.
4676 asize
= aggsum_value(&arc_size
);
4677 ameta
= aggsum_value(&arc_meta_used
);
4683 * Now that we've tried to evict enough from the MRU to get its
4684 * size back to arc_p, if we're still above the target cache
4685 * size, we evict the rest from the MFU.
4687 target
= asize
- arc_c
;
4689 if (arc_adjust_type(arc_mfu
) == ARC_BUFC_METADATA
&&
4690 ameta
> arc_meta_min
) {
4691 bytes
= arc_adjust_impl(arc_mfu
, 0, target
, ARC_BUFC_METADATA
);
4692 total_evicted
+= bytes
;
4695 * If we couldn't evict our target number of bytes from
4696 * metadata, we try to get the rest from data.
4701 arc_adjust_impl(arc_mfu
, 0, target
, ARC_BUFC_DATA
);
4703 bytes
= arc_adjust_impl(arc_mfu
, 0, target
, ARC_BUFC_DATA
);
4704 total_evicted
+= bytes
;
4707 * If we couldn't evict our target number of bytes from
4708 * data, we try to get the rest from data.
4713 arc_adjust_impl(arc_mfu
, 0, target
, ARC_BUFC_METADATA
);
4717 * Adjust ghost lists
4719 * In addition to the above, the ARC also defines target values
4720 * for the ghost lists. The sum of the mru list and mru ghost
4721 * list should never exceed the target size of the cache, and
4722 * the sum of the mru list, mfu list, mru ghost list, and mfu
4723 * ghost list should never exceed twice the target size of the
4724 * cache. The following logic enforces these limits on the ghost
4725 * caches, and evicts from them as needed.
4727 target
= zfs_refcount_count(&arc_mru
->arcs_size
) +
4728 zfs_refcount_count(&arc_mru_ghost
->arcs_size
) - arc_c
;
4730 bytes
= arc_adjust_impl(arc_mru_ghost
, 0, target
, ARC_BUFC_DATA
);
4731 total_evicted
+= bytes
;
4736 arc_adjust_impl(arc_mru_ghost
, 0, target
, ARC_BUFC_METADATA
);
4739 * We assume the sum of the mru list and mfu list is less than
4740 * or equal to arc_c (we enforced this above), which means we
4741 * can use the simpler of the two equations below:
4743 * mru + mfu + mru ghost + mfu ghost <= 2 * arc_c
4744 * mru ghost + mfu ghost <= arc_c
4746 target
= zfs_refcount_count(&arc_mru_ghost
->arcs_size
) +
4747 zfs_refcount_count(&arc_mfu_ghost
->arcs_size
) - arc_c
;
4749 bytes
= arc_adjust_impl(arc_mfu_ghost
, 0, target
, ARC_BUFC_DATA
);
4750 total_evicted
+= bytes
;
4755 arc_adjust_impl(arc_mfu_ghost
, 0, target
, ARC_BUFC_METADATA
);
4757 return (total_evicted
);
4761 arc_flush(spa_t
*spa
, boolean_t retry
)
4766 * If retry is B_TRUE, a spa must not be specified since we have
4767 * no good way to determine if all of a spa's buffers have been
4768 * evicted from an arc state.
4770 ASSERT(!retry
|| spa
== 0);
4773 guid
= spa_load_guid(spa
);
4775 (void) arc_flush_state(arc_mru
, guid
, ARC_BUFC_DATA
, retry
);
4776 (void) arc_flush_state(arc_mru
, guid
, ARC_BUFC_METADATA
, retry
);
4778 (void) arc_flush_state(arc_mfu
, guid
, ARC_BUFC_DATA
, retry
);
4779 (void) arc_flush_state(arc_mfu
, guid
, ARC_BUFC_METADATA
, retry
);
4781 (void) arc_flush_state(arc_mru_ghost
, guid
, ARC_BUFC_DATA
, retry
);
4782 (void) arc_flush_state(arc_mru_ghost
, guid
, ARC_BUFC_METADATA
, retry
);
4784 (void) arc_flush_state(arc_mfu_ghost
, guid
, ARC_BUFC_DATA
, retry
);
4785 (void) arc_flush_state(arc_mfu_ghost
, guid
, ARC_BUFC_METADATA
, retry
);
4789 arc_reduce_target_size(int64_t to_free
)
4791 uint64_t asize
= aggsum_value(&arc_size
);
4794 if (c
> to_free
&& c
- to_free
> arc_c_min
) {
4795 arc_c
= c
- to_free
;
4796 atomic_add_64(&arc_p
, -(arc_p
>> arc_shrink_shift
));
4798 arc_c
= MAX(asize
, arc_c_min
);
4800 arc_p
= (arc_c
>> 1);
4801 ASSERT(arc_c
>= arc_c_min
);
4802 ASSERT((int64_t)arc_p
>= 0);
4807 if (asize
> arc_c
) {
4808 /* See comment in arc_adjust_cb_check() on why lock+flag */
4809 mutex_enter(&arc_adjust_lock
);
4810 arc_adjust_needed
= B_TRUE
;
4811 mutex_exit(&arc_adjust_lock
);
4812 zthr_wakeup(arc_adjust_zthr
);
4816 * Return maximum amount of memory that we could possibly use. Reduced
4817 * to half of all memory in user space which is primarily used for testing.
4820 arc_all_memory(void)
4823 #ifdef CONFIG_HIGHMEM
4824 return (ptob(totalram_pages
- totalhigh_pages
));
4826 return (ptob(totalram_pages
));
4827 #endif /* CONFIG_HIGHMEM */
4829 return (ptob(physmem
) / 2);
4830 #endif /* _KERNEL */
4834 * Return the amount of memory that is considered free. In user space
4835 * which is primarily used for testing we pretend that free memory ranges
4836 * from 0-20% of all memory.
4839 arc_free_memory(void)
4842 #ifdef CONFIG_HIGHMEM
4845 return (ptob(si
.freeram
- si
.freehigh
));
4847 return (ptob(nr_free_pages() +
4848 nr_inactive_file_pages() +
4849 nr_inactive_anon_pages() +
4850 nr_slab_reclaimable_pages()));
4852 #endif /* CONFIG_HIGHMEM */
4854 return (spa_get_random(arc_all_memory() * 20 / 100));
4855 #endif /* _KERNEL */
4858 typedef enum free_memory_reason_t
{
4863 FMR_PAGES_PP_MAXIMUM
,
4866 } free_memory_reason_t
;
4868 int64_t last_free_memory
;
4869 free_memory_reason_t last_free_reason
;
4873 * Additional reserve of pages for pp_reserve.
4875 int64_t arc_pages_pp_reserve
= 64;
4878 * Additional reserve of pages for swapfs.
4880 int64_t arc_swapfs_reserve
= 64;
4881 #endif /* _KERNEL */
4884 * Return the amount of memory that can be consumed before reclaim will be
4885 * needed. Positive if there is sufficient free memory, negative indicates
4886 * the amount of memory that needs to be freed up.
4889 arc_available_memory(void)
4891 int64_t lowest
= INT64_MAX
;
4892 free_memory_reason_t r
= FMR_UNKNOWN
;
4899 pgcnt_t needfree
= btop(arc_need_free
);
4900 pgcnt_t lotsfree
= btop(arc_sys_free
);
4901 pgcnt_t desfree
= 0;
4902 pgcnt_t freemem
= btop(arc_free_memory());
4906 n
= PAGESIZE
* (-needfree
);
4914 * check that we're out of range of the pageout scanner. It starts to
4915 * schedule paging if freemem is less than lotsfree and needfree.
4916 * lotsfree is the high-water mark for pageout, and needfree is the
4917 * number of needed free pages. We add extra pages here to make sure
4918 * the scanner doesn't start up while we're freeing memory.
4920 n
= PAGESIZE
* (freemem
- lotsfree
- needfree
- desfree
);
4928 * check to make sure that swapfs has enough space so that anon
4929 * reservations can still succeed. anon_resvmem() checks that the
4930 * availrmem is greater than swapfs_minfree, and the number of reserved
4931 * swap pages. We also add a bit of extra here just to prevent
4932 * circumstances from getting really dire.
4934 n
= PAGESIZE
* (availrmem
- swapfs_minfree
- swapfs_reserve
-
4935 desfree
- arc_swapfs_reserve
);
4938 r
= FMR_SWAPFS_MINFREE
;
4942 * Check that we have enough availrmem that memory locking (e.g., via
4943 * mlock(3C) or memcntl(2)) can still succeed. (pages_pp_maximum
4944 * stores the number of pages that cannot be locked; when availrmem
4945 * drops below pages_pp_maximum, page locking mechanisms such as
4946 * page_pp_lock() will fail.)
4948 n
= PAGESIZE
* (availrmem
- pages_pp_maximum
-
4949 arc_pages_pp_reserve
);
4952 r
= FMR_PAGES_PP_MAXIMUM
;
4958 * If we're on a 32-bit platform, it's possible that we'll exhaust the
4959 * kernel heap space before we ever run out of available physical
4960 * memory. Most checks of the size of the heap_area compare against
4961 * tune.t_minarmem, which is the minimum available real memory that we
4962 * can have in the system. However, this is generally fixed at 25 pages
4963 * which is so low that it's useless. In this comparison, we seek to
4964 * calculate the total heap-size, and reclaim if more than 3/4ths of the
4965 * heap is allocated. (Or, in the calculation, if less than 1/4th is
4968 n
= vmem_size(heap_arena
, VMEM_FREE
) -
4969 (vmem_size(heap_arena
, VMEM_FREE
| VMEM_ALLOC
) >> 2);
4977 * If zio data pages are being allocated out of a separate heap segment,
4978 * then enforce that the size of available vmem for this arena remains
4979 * above about 1/4th (1/(2^arc_zio_arena_free_shift)) free.
4981 * Note that reducing the arc_zio_arena_free_shift keeps more virtual
4982 * memory (in the zio_arena) free, which can avoid memory
4983 * fragmentation issues.
4985 if (zio_arena
!= NULL
) {
4986 n
= (int64_t)vmem_size(zio_arena
, VMEM_FREE
) -
4987 (vmem_size(zio_arena
, VMEM_ALLOC
) >>
4988 arc_zio_arena_free_shift
);
4995 /* Every 100 calls, free a small amount */
4996 if (spa_get_random(100) == 0)
4998 #endif /* _KERNEL */
5000 last_free_memory
= lowest
;
5001 last_free_reason
= r
;
5007 * Determine if the system is under memory pressure and is asking
5008 * to reclaim memory. A return value of B_TRUE indicates that the system
5009 * is under memory pressure and that the arc should adjust accordingly.
5012 arc_reclaim_needed(void)
5014 return (arc_available_memory() < 0);
5018 arc_kmem_reap_soon(void)
5021 kmem_cache_t
*prev_cache
= NULL
;
5022 kmem_cache_t
*prev_data_cache
= NULL
;
5023 extern kmem_cache_t
*zio_buf_cache
[];
5024 extern kmem_cache_t
*zio_data_buf_cache
[];
5025 extern kmem_cache_t
*range_seg_cache
;
5028 if ((aggsum_compare(&arc_meta_used
, arc_meta_limit
) >= 0) &&
5029 zfs_arc_meta_prune
) {
5031 * We are exceeding our meta-data cache limit.
5032 * Prune some entries to release holds on meta-data.
5034 arc_prune_async(zfs_arc_meta_prune
);
5038 * Reclaim unused memory from all kmem caches.
5044 for (i
= 0; i
< SPA_MAXBLOCKSIZE
>> SPA_MINBLOCKSHIFT
; i
++) {
5046 /* reach upper limit of cache size on 32-bit */
5047 if (zio_buf_cache
[i
] == NULL
)
5050 if (zio_buf_cache
[i
] != prev_cache
) {
5051 prev_cache
= zio_buf_cache
[i
];
5052 kmem_cache_reap_now(zio_buf_cache
[i
]);
5054 if (zio_data_buf_cache
[i
] != prev_data_cache
) {
5055 prev_data_cache
= zio_data_buf_cache
[i
];
5056 kmem_cache_reap_now(zio_data_buf_cache
[i
]);
5059 kmem_cache_reap_now(buf_cache
);
5060 kmem_cache_reap_now(hdr_full_cache
);
5061 kmem_cache_reap_now(hdr_l2only_cache
);
5062 kmem_cache_reap_now(range_seg_cache
);
5064 if (zio_arena
!= NULL
) {
5066 * Ask the vmem arena to reclaim unused memory from its
5069 vmem_qcache_reap(zio_arena
);
5075 arc_adjust_cb_check(void *arg
, zthr_t
*zthr
)
5078 * This is necessary in order to keep the kstat information
5079 * up to date for tools that display kstat data such as the
5080 * mdb ::arc dcmd and the Linux crash utility. These tools
5081 * typically do not call kstat's update function, but simply
5082 * dump out stats from the most recent update. Without
5083 * this call, these commands may show stale stats for the
5084 * anon, mru, mru_ghost, mfu, and mfu_ghost lists. Even
5085 * with this change, the data might be up to 1 second
5086 * out of date(the arc_adjust_zthr has a maximum sleep
5087 * time of 1 second); but that should suffice. The
5088 * arc_state_t structures can be queried directly if more
5089 * accurate information is needed.
5091 if (arc_ksp
!= NULL
)
5092 arc_ksp
->ks_update(arc_ksp
, KSTAT_READ
);
5095 * We have to rely on arc_get_data_impl() to tell us when to adjust,
5096 * rather than checking if we are overflowing here, so that we are
5097 * sure to not leave arc_get_data_impl() waiting on
5098 * arc_adjust_waiters_cv. If we have become "not overflowing" since
5099 * arc_get_data_impl() checked, we need to wake it up. We could
5100 * broadcast the CV here, but arc_get_data_impl() may have not yet
5101 * gone to sleep. We would need to use a mutex to ensure that this
5102 * function doesn't broadcast until arc_get_data_impl() has gone to
5103 * sleep (e.g. the arc_adjust_lock). However, the lock ordering of
5104 * such a lock would necessarily be incorrect with respect to the
5105 * zthr_lock, which is held before this function is called, and is
5106 * held by arc_get_data_impl() when it calls zthr_wakeup().
5108 return (arc_adjust_needed
);
5112 * Keep arc_size under arc_c by running arc_adjust which evicts data
5117 arc_adjust_cb(void *arg
, zthr_t
*zthr
)
5119 uint64_t evicted
= 0;
5120 fstrans_cookie_t cookie
= spl_fstrans_mark();
5122 /* Evict from cache */
5123 evicted
= arc_adjust();
5126 * If evicted is zero, we couldn't evict anything
5127 * via arc_adjust(). This could be due to hash lock
5128 * collisions, but more likely due to the majority of
5129 * arc buffers being unevictable. Therefore, even if
5130 * arc_size is above arc_c, another pass is unlikely to
5131 * be helpful and could potentially cause us to enter an
5132 * infinite loop. Additionally, zthr_iscancelled() is
5133 * checked here so that if the arc is shutting down, the
5134 * broadcast will wake any remaining arc adjust waiters.
5136 mutex_enter(&arc_adjust_lock
);
5137 arc_adjust_needed
= !zthr_iscancelled(arc_adjust_zthr
) &&
5138 evicted
> 0 && aggsum_compare(&arc_size
, arc_c
) > 0;
5139 if (!arc_adjust_needed
) {
5141 * We're either no longer overflowing, or we
5142 * can't evict anything more, so we should wake
5143 * arc_get_data_impl() sooner.
5145 cv_broadcast(&arc_adjust_waiters_cv
);
5148 mutex_exit(&arc_adjust_lock
);
5149 spl_fstrans_unmark(cookie
);
5154 arc_reap_cb_check(void *arg
, zthr_t
*zthr
)
5156 int64_t free_memory
= arc_available_memory();
5159 * If a kmem reap is already active, don't schedule more. We must
5160 * check for this because kmem_cache_reap_soon() won't actually
5161 * block on the cache being reaped (this is to prevent callers from
5162 * becoming implicitly blocked by a system-wide kmem reap -- which,
5163 * on a system with many, many full magazines, can take minutes).
5165 if (!kmem_cache_reap_active() && free_memory
< 0) {
5167 arc_no_grow
= B_TRUE
;
5170 * Wait at least zfs_grow_retry (default 5) seconds
5171 * before considering growing.
5173 arc_growtime
= gethrtime() + SEC2NSEC(arc_grow_retry
);
5175 } else if (free_memory
< arc_c
>> arc_no_grow_shift
) {
5176 arc_no_grow
= B_TRUE
;
5177 } else if (gethrtime() >= arc_growtime
) {
5178 arc_no_grow
= B_FALSE
;
5185 * Keep enough free memory in the system by reaping the ARC's kmem
5186 * caches. To cause more slabs to be reapable, we may reduce the
5187 * target size of the cache (arc_c), causing the arc_adjust_cb()
5188 * to free more buffers.
5192 arc_reap_cb(void *arg
, zthr_t
*zthr
)
5194 int64_t free_memory
;
5195 fstrans_cookie_t cookie
= spl_fstrans_mark();
5198 * Kick off asynchronous kmem_reap()'s of all our caches.
5200 arc_kmem_reap_soon();
5203 * Wait at least arc_kmem_cache_reap_retry_ms between
5204 * arc_kmem_reap_soon() calls. Without this check it is possible to
5205 * end up in a situation where we spend lots of time reaping
5206 * caches, while we're near arc_c_min. Waiting here also gives the
5207 * subsequent free memory check a chance of finding that the
5208 * asynchronous reap has already freed enough memory, and we don't
5209 * need to call arc_reduce_target_size().
5211 delay((hz
* arc_kmem_cache_reap_retry_ms
+ 999) / 1000);
5214 * Reduce the target size as needed to maintain the amount of free
5215 * memory in the system at a fraction of the arc_size (1/128th by
5216 * default). If oversubscribed (free_memory < 0) then reduce the
5217 * target arc_size by the deficit amount plus the fractional
5218 * amount. If free memory is positive but less then the fractional
5219 * amount, reduce by what is needed to hit the fractional amount.
5221 free_memory
= arc_available_memory();
5224 (arc_c
>> arc_shrink_shift
) - free_memory
;
5227 to_free
= MAX(to_free
, arc_need_free
);
5229 arc_reduce_target_size(to_free
);
5231 spl_fstrans_unmark(cookie
);
5236 * Determine the amount of memory eligible for eviction contained in the
5237 * ARC. All clean data reported by the ghost lists can always be safely
5238 * evicted. Due to arc_c_min, the same does not hold for all clean data
5239 * contained by the regular mru and mfu lists.
5241 * In the case of the regular mru and mfu lists, we need to report as
5242 * much clean data as possible, such that evicting that same reported
5243 * data will not bring arc_size below arc_c_min. Thus, in certain
5244 * circumstances, the total amount of clean data in the mru and mfu
5245 * lists might not actually be evictable.
5247 * The following two distinct cases are accounted for:
5249 * 1. The sum of the amount of dirty data contained by both the mru and
5250 * mfu lists, plus the ARC's other accounting (e.g. the anon list),
5251 * is greater than or equal to arc_c_min.
5252 * (i.e. amount of dirty data >= arc_c_min)
5254 * This is the easy case; all clean data contained by the mru and mfu
5255 * lists is evictable. Evicting all clean data can only drop arc_size
5256 * to the amount of dirty data, which is greater than arc_c_min.
5258 * 2. The sum of the amount of dirty data contained by both the mru and
5259 * mfu lists, plus the ARC's other accounting (e.g. the anon list),
5260 * is less than arc_c_min.
5261 * (i.e. arc_c_min > amount of dirty data)
5263 * 2.1. arc_size is greater than or equal arc_c_min.
5264 * (i.e. arc_size >= arc_c_min > amount of dirty data)
5266 * In this case, not all clean data from the regular mru and mfu
5267 * lists is actually evictable; we must leave enough clean data
5268 * to keep arc_size above arc_c_min. Thus, the maximum amount of
5269 * evictable data from the two lists combined, is exactly the
5270 * difference between arc_size and arc_c_min.
5272 * 2.2. arc_size is less than arc_c_min
5273 * (i.e. arc_c_min > arc_size > amount of dirty data)
5275 * In this case, none of the data contained in the mru and mfu
5276 * lists is evictable, even if it's clean. Since arc_size is
5277 * already below arc_c_min, evicting any more would only
5278 * increase this negative difference.
5281 arc_evictable_memory(void)
5283 int64_t asize
= aggsum_value(&arc_size
);
5284 uint64_t arc_clean
=
5285 zfs_refcount_count(&arc_mru
->arcs_esize
[ARC_BUFC_DATA
]) +
5286 zfs_refcount_count(&arc_mru
->arcs_esize
[ARC_BUFC_METADATA
]) +
5287 zfs_refcount_count(&arc_mfu
->arcs_esize
[ARC_BUFC_DATA
]) +
5288 zfs_refcount_count(&arc_mfu
->arcs_esize
[ARC_BUFC_METADATA
]);
5289 uint64_t arc_dirty
= MAX((int64_t)asize
- (int64_t)arc_clean
, 0);
5292 * Scale reported evictable memory in proportion to page cache, cap
5293 * at specified min/max.
5295 uint64_t min
= (ptob(nr_file_pages()) / 100) * zfs_arc_pc_percent
;
5296 min
= MAX(arc_c_min
, MIN(arc_c_max
, min
));
5298 if (arc_dirty
>= min
)
5301 return (MAX((int64_t)asize
- (int64_t)min
, 0));
5305 * If sc->nr_to_scan is zero, the caller is requesting a query of the
5306 * number of objects which can potentially be freed. If it is nonzero,
5307 * the request is to free that many objects.
5309 * Linux kernels >= 3.12 have the count_objects and scan_objects callbacks
5310 * in struct shrinker and also require the shrinker to return the number
5313 * Older kernels require the shrinker to return the number of freeable
5314 * objects following the freeing of nr_to_free.
5316 static spl_shrinker_t
5317 __arc_shrinker_func(struct shrinker
*shrink
, struct shrink_control
*sc
)
5321 /* The arc is considered warm once reclaim has occurred */
5322 if (unlikely(arc_warm
== B_FALSE
))
5325 /* Return the potential number of reclaimable pages */
5326 pages
= btop((int64_t)arc_evictable_memory());
5327 if (sc
->nr_to_scan
== 0)
5330 /* Not allowed to perform filesystem reclaim */
5331 if (!(sc
->gfp_mask
& __GFP_FS
))
5332 return (SHRINK_STOP
);
5334 /* Reclaim in progress */
5335 if (mutex_tryenter(&arc_adjust_lock
) == 0) {
5336 ARCSTAT_INCR(arcstat_need_free
, ptob(sc
->nr_to_scan
));
5340 mutex_exit(&arc_adjust_lock
);
5343 * Evict the requested number of pages by shrinking arc_c the
5347 arc_reduce_target_size(ptob(sc
->nr_to_scan
));
5348 if (current_is_kswapd())
5349 arc_kmem_reap_soon();
5350 #ifdef HAVE_SPLIT_SHRINKER_CALLBACK
5351 pages
= MAX((int64_t)pages
-
5352 (int64_t)btop(arc_evictable_memory()), 0);
5354 pages
= btop(arc_evictable_memory());
5357 * We've shrunk what we can, wake up threads.
5359 cv_broadcast(&arc_adjust_waiters_cv
);
5361 pages
= SHRINK_STOP
;
5364 * When direct reclaim is observed it usually indicates a rapid
5365 * increase in memory pressure. This occurs because the kswapd
5366 * threads were unable to asynchronously keep enough free memory
5367 * available. In this case set arc_no_grow to briefly pause arc
5368 * growth to avoid compounding the memory pressure.
5370 if (current_is_kswapd()) {
5371 ARCSTAT_BUMP(arcstat_memory_indirect_count
);
5373 arc_no_grow
= B_TRUE
;
5374 arc_kmem_reap_soon();
5375 ARCSTAT_BUMP(arcstat_memory_direct_count
);
5380 SPL_SHRINKER_CALLBACK_WRAPPER(arc_shrinker_func
);
5382 SPL_SHRINKER_DECLARE(arc_shrinker
, arc_shrinker_func
, DEFAULT_SEEKS
);
5383 #endif /* _KERNEL */
5386 * Adapt arc info given the number of bytes we are trying to add and
5387 * the state that we are coming from. This function is only called
5388 * when we are adding new content to the cache.
5391 arc_adapt(int bytes
, arc_state_t
*state
)
5394 uint64_t arc_p_min
= (arc_c
>> arc_p_min_shift
);
5395 int64_t mrug_size
= zfs_refcount_count(&arc_mru_ghost
->arcs_size
);
5396 int64_t mfug_size
= zfs_refcount_count(&arc_mfu_ghost
->arcs_size
);
5398 if (state
== arc_l2c_only
)
5403 * Adapt the target size of the MRU list:
5404 * - if we just hit in the MRU ghost list, then increase
5405 * the target size of the MRU list.
5406 * - if we just hit in the MFU ghost list, then increase
5407 * the target size of the MFU list by decreasing the
5408 * target size of the MRU list.
5410 if (state
== arc_mru_ghost
) {
5411 mult
= (mrug_size
>= mfug_size
) ? 1 : (mfug_size
/ mrug_size
);
5412 if (!zfs_arc_p_dampener_disable
)
5413 mult
= MIN(mult
, 10); /* avoid wild arc_p adjustment */
5415 arc_p
= MIN(arc_c
- arc_p_min
, arc_p
+ bytes
* mult
);
5416 } else if (state
== arc_mfu_ghost
) {
5419 mult
= (mfug_size
>= mrug_size
) ? 1 : (mrug_size
/ mfug_size
);
5420 if (!zfs_arc_p_dampener_disable
)
5421 mult
= MIN(mult
, 10);
5423 delta
= MIN(bytes
* mult
, arc_p
);
5424 arc_p
= MAX(arc_p_min
, arc_p
- delta
);
5426 ASSERT((int64_t)arc_p
>= 0);
5429 * Wake reap thread if we do not have any available memory
5431 if (arc_reclaim_needed()) {
5432 zthr_wakeup(arc_reap_zthr
);
5439 if (arc_c
>= arc_c_max
)
5443 * If we're within (2 * maxblocksize) bytes of the target
5444 * cache size, increment the target cache size
5446 ASSERT3U(arc_c
, >=, 2ULL << SPA_MAXBLOCKSHIFT
);
5447 if (aggsum_compare(&arc_size
, arc_c
- (2ULL << SPA_MAXBLOCKSHIFT
)) >=
5449 atomic_add_64(&arc_c
, (int64_t)bytes
);
5450 if (arc_c
> arc_c_max
)
5452 else if (state
== arc_anon
)
5453 atomic_add_64(&arc_p
, (int64_t)bytes
);
5457 ASSERT((int64_t)arc_p
>= 0);
5461 * Check if arc_size has grown past our upper threshold, determined by
5462 * zfs_arc_overflow_shift.
5465 arc_is_overflowing(void)
5467 /* Always allow at least one block of overflow */
5468 uint64_t overflow
= MAX(SPA_MAXBLOCKSIZE
,
5469 arc_c
>> zfs_arc_overflow_shift
);
5472 * We just compare the lower bound here for performance reasons. Our
5473 * primary goals are to make sure that the arc never grows without
5474 * bound, and that it can reach its maximum size. This check
5475 * accomplishes both goals. The maximum amount we could run over by is
5476 * 2 * aggsum_borrow_multiplier * NUM_CPUS * the average size of a block
5477 * in the ARC. In practice, that's in the tens of MB, which is low
5478 * enough to be safe.
5480 return (aggsum_lower_bound(&arc_size
) >= arc_c
+ overflow
);
5484 arc_get_data_abd(arc_buf_hdr_t
*hdr
, uint64_t size
, void *tag
)
5486 arc_buf_contents_t type
= arc_buf_type(hdr
);
5488 arc_get_data_impl(hdr
, size
, tag
);
5489 if (type
== ARC_BUFC_METADATA
) {
5490 return (abd_alloc(size
, B_TRUE
));
5492 ASSERT(type
== ARC_BUFC_DATA
);
5493 return (abd_alloc(size
, B_FALSE
));
5498 arc_get_data_buf(arc_buf_hdr_t
*hdr
, uint64_t size
, void *tag
)
5500 arc_buf_contents_t type
= arc_buf_type(hdr
);
5502 arc_get_data_impl(hdr
, size
, tag
);
5503 if (type
== ARC_BUFC_METADATA
) {
5504 return (zio_buf_alloc(size
));
5506 ASSERT(type
== ARC_BUFC_DATA
);
5507 return (zio_data_buf_alloc(size
));
5512 * Allocate a block and return it to the caller. If we are hitting the
5513 * hard limit for the cache size, we must sleep, waiting for the eviction
5514 * thread to catch up. If we're past the target size but below the hard
5515 * limit, we'll only signal the reclaim thread and continue on.
5518 arc_get_data_impl(arc_buf_hdr_t
*hdr
, uint64_t size
, void *tag
)
5520 arc_state_t
*state
= hdr
->b_l1hdr
.b_state
;
5521 arc_buf_contents_t type
= arc_buf_type(hdr
);
5523 arc_adapt(size
, state
);
5526 * If arc_size is currently overflowing, and has grown past our
5527 * upper limit, we must be adding data faster than the evict
5528 * thread can evict. Thus, to ensure we don't compound the
5529 * problem by adding more data and forcing arc_size to grow even
5530 * further past it's target size, we halt and wait for the
5531 * eviction thread to catch up.
5533 * It's also possible that the reclaim thread is unable to evict
5534 * enough buffers to get arc_size below the overflow limit (e.g.
5535 * due to buffers being un-evictable, or hash lock collisions).
5536 * In this case, we want to proceed regardless if we're
5537 * overflowing; thus we don't use a while loop here.
5539 if (arc_is_overflowing()) {
5540 mutex_enter(&arc_adjust_lock
);
5543 * Now that we've acquired the lock, we may no longer be
5544 * over the overflow limit, lets check.
5546 * We're ignoring the case of spurious wake ups. If that
5547 * were to happen, it'd let this thread consume an ARC
5548 * buffer before it should have (i.e. before we're under
5549 * the overflow limit and were signalled by the reclaim
5550 * thread). As long as that is a rare occurrence, it
5551 * shouldn't cause any harm.
5553 if (arc_is_overflowing()) {
5554 arc_adjust_needed
= B_TRUE
;
5555 zthr_wakeup(arc_adjust_zthr
);
5556 (void) cv_wait(&arc_adjust_waiters_cv
,
5559 mutex_exit(&arc_adjust_lock
);
5562 VERIFY3U(hdr
->b_type
, ==, type
);
5563 if (type
== ARC_BUFC_METADATA
) {
5564 arc_space_consume(size
, ARC_SPACE_META
);
5566 arc_space_consume(size
, ARC_SPACE_DATA
);
5570 * Update the state size. Note that ghost states have a
5571 * "ghost size" and so don't need to be updated.
5573 if (!GHOST_STATE(state
)) {
5575 (void) zfs_refcount_add_many(&state
->arcs_size
, size
, tag
);
5578 * If this is reached via arc_read, the link is
5579 * protected by the hash lock. If reached via
5580 * arc_buf_alloc, the header should not be accessed by
5581 * any other thread. And, if reached via arc_read_done,
5582 * the hash lock will protect it if it's found in the
5583 * hash table; otherwise no other thread should be
5584 * trying to [add|remove]_reference it.
5586 if (multilist_link_active(&hdr
->b_l1hdr
.b_arc_node
)) {
5587 ASSERT(zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
));
5588 (void) zfs_refcount_add_many(&state
->arcs_esize
[type
],
5593 * If we are growing the cache, and we are adding anonymous
5594 * data, and we have outgrown arc_p, update arc_p
5596 if (aggsum_compare(&arc_size
, arc_c
) < 0 &&
5597 hdr
->b_l1hdr
.b_state
== arc_anon
&&
5598 (zfs_refcount_count(&arc_anon
->arcs_size
) +
5599 zfs_refcount_count(&arc_mru
->arcs_size
) > arc_p
))
5600 arc_p
= MIN(arc_c
, arc_p
+ size
);
5605 arc_free_data_abd(arc_buf_hdr_t
*hdr
, abd_t
*abd
, uint64_t size
, void *tag
)
5607 arc_free_data_impl(hdr
, size
, tag
);
5612 arc_free_data_buf(arc_buf_hdr_t
*hdr
, void *buf
, uint64_t size
, void *tag
)
5614 arc_buf_contents_t type
= arc_buf_type(hdr
);
5616 arc_free_data_impl(hdr
, size
, tag
);
5617 if (type
== ARC_BUFC_METADATA
) {
5618 zio_buf_free(buf
, size
);
5620 ASSERT(type
== ARC_BUFC_DATA
);
5621 zio_data_buf_free(buf
, size
);
5626 * Free the arc data buffer.
5629 arc_free_data_impl(arc_buf_hdr_t
*hdr
, uint64_t size
, void *tag
)
5631 arc_state_t
*state
= hdr
->b_l1hdr
.b_state
;
5632 arc_buf_contents_t type
= arc_buf_type(hdr
);
5634 /* protected by hash lock, if in the hash table */
5635 if (multilist_link_active(&hdr
->b_l1hdr
.b_arc_node
)) {
5636 ASSERT(zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
));
5637 ASSERT(state
!= arc_anon
&& state
!= arc_l2c_only
);
5639 (void) zfs_refcount_remove_many(&state
->arcs_esize
[type
],
5642 (void) zfs_refcount_remove_many(&state
->arcs_size
, size
, tag
);
5644 VERIFY3U(hdr
->b_type
, ==, type
);
5645 if (type
== ARC_BUFC_METADATA
) {
5646 arc_space_return(size
, ARC_SPACE_META
);
5648 ASSERT(type
== ARC_BUFC_DATA
);
5649 arc_space_return(size
, ARC_SPACE_DATA
);
5654 * This routine is called whenever a buffer is accessed.
5655 * NOTE: the hash lock is dropped in this function.
5658 arc_access(arc_buf_hdr_t
*hdr
, kmutex_t
*hash_lock
)
5662 ASSERT(MUTEX_HELD(hash_lock
));
5663 ASSERT(HDR_HAS_L1HDR(hdr
));
5665 if (hdr
->b_l1hdr
.b_state
== arc_anon
) {
5667 * This buffer is not in the cache, and does not
5668 * appear in our "ghost" list. Add the new buffer
5672 ASSERT0(hdr
->b_l1hdr
.b_arc_access
);
5673 hdr
->b_l1hdr
.b_arc_access
= ddi_get_lbolt();
5674 DTRACE_PROBE1(new_state__mru
, arc_buf_hdr_t
*, hdr
);
5675 arc_change_state(arc_mru
, hdr
, hash_lock
);
5677 } else if (hdr
->b_l1hdr
.b_state
== arc_mru
) {
5678 now
= ddi_get_lbolt();
5681 * If this buffer is here because of a prefetch, then either:
5682 * - clear the flag if this is a "referencing" read
5683 * (any subsequent access will bump this into the MFU state).
5685 * - move the buffer to the head of the list if this is
5686 * another prefetch (to make it less likely to be evicted).
5688 if (HDR_PREFETCH(hdr
) || HDR_PRESCIENT_PREFETCH(hdr
)) {
5689 if (zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
) == 0) {
5690 /* link protected by hash lock */
5691 ASSERT(multilist_link_active(
5692 &hdr
->b_l1hdr
.b_arc_node
));
5694 arc_hdr_clear_flags(hdr
,
5696 ARC_FLAG_PRESCIENT_PREFETCH
);
5697 atomic_inc_32(&hdr
->b_l1hdr
.b_mru_hits
);
5698 ARCSTAT_BUMP(arcstat_mru_hits
);
5700 hdr
->b_l1hdr
.b_arc_access
= now
;
5705 * This buffer has been "accessed" only once so far,
5706 * but it is still in the cache. Move it to the MFU
5709 if (ddi_time_after(now
, hdr
->b_l1hdr
.b_arc_access
+
5712 * More than 125ms have passed since we
5713 * instantiated this buffer. Move it to the
5714 * most frequently used state.
5716 hdr
->b_l1hdr
.b_arc_access
= now
;
5717 DTRACE_PROBE1(new_state__mfu
, arc_buf_hdr_t
*, hdr
);
5718 arc_change_state(arc_mfu
, hdr
, hash_lock
);
5720 atomic_inc_32(&hdr
->b_l1hdr
.b_mru_hits
);
5721 ARCSTAT_BUMP(arcstat_mru_hits
);
5722 } else if (hdr
->b_l1hdr
.b_state
== arc_mru_ghost
) {
5723 arc_state_t
*new_state
;
5725 * This buffer has been "accessed" recently, but
5726 * was evicted from the cache. Move it to the
5730 if (HDR_PREFETCH(hdr
) || HDR_PRESCIENT_PREFETCH(hdr
)) {
5731 new_state
= arc_mru
;
5732 if (zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
) > 0) {
5733 arc_hdr_clear_flags(hdr
,
5735 ARC_FLAG_PRESCIENT_PREFETCH
);
5737 DTRACE_PROBE1(new_state__mru
, arc_buf_hdr_t
*, hdr
);
5739 new_state
= arc_mfu
;
5740 DTRACE_PROBE1(new_state__mfu
, arc_buf_hdr_t
*, hdr
);
5743 hdr
->b_l1hdr
.b_arc_access
= ddi_get_lbolt();
5744 arc_change_state(new_state
, hdr
, hash_lock
);
5746 atomic_inc_32(&hdr
->b_l1hdr
.b_mru_ghost_hits
);
5747 ARCSTAT_BUMP(arcstat_mru_ghost_hits
);
5748 } else if (hdr
->b_l1hdr
.b_state
== arc_mfu
) {
5750 * This buffer has been accessed more than once and is
5751 * still in the cache. Keep it in the MFU state.
5753 * NOTE: an add_reference() that occurred when we did
5754 * the arc_read() will have kicked this off the list.
5755 * If it was a prefetch, we will explicitly move it to
5756 * the head of the list now.
5759 atomic_inc_32(&hdr
->b_l1hdr
.b_mfu_hits
);
5760 ARCSTAT_BUMP(arcstat_mfu_hits
);
5761 hdr
->b_l1hdr
.b_arc_access
= ddi_get_lbolt();
5762 } else if (hdr
->b_l1hdr
.b_state
== arc_mfu_ghost
) {
5763 arc_state_t
*new_state
= arc_mfu
;
5765 * This buffer has been accessed more than once but has
5766 * been evicted from the cache. Move it back to the
5770 if (HDR_PREFETCH(hdr
) || HDR_PRESCIENT_PREFETCH(hdr
)) {
5772 * This is a prefetch access...
5773 * move this block back to the MRU state.
5775 new_state
= arc_mru
;
5778 hdr
->b_l1hdr
.b_arc_access
= ddi_get_lbolt();
5779 DTRACE_PROBE1(new_state__mfu
, arc_buf_hdr_t
*, hdr
);
5780 arc_change_state(new_state
, hdr
, hash_lock
);
5782 atomic_inc_32(&hdr
->b_l1hdr
.b_mfu_ghost_hits
);
5783 ARCSTAT_BUMP(arcstat_mfu_ghost_hits
);
5784 } else if (hdr
->b_l1hdr
.b_state
== arc_l2c_only
) {
5786 * This buffer is on the 2nd Level ARC.
5789 hdr
->b_l1hdr
.b_arc_access
= ddi_get_lbolt();
5790 DTRACE_PROBE1(new_state__mfu
, arc_buf_hdr_t
*, hdr
);
5791 arc_change_state(arc_mfu
, hdr
, hash_lock
);
5793 cmn_err(CE_PANIC
, "invalid arc state 0x%p",
5794 hdr
->b_l1hdr
.b_state
);
5799 * This routine is called by dbuf_hold() to update the arc_access() state
5800 * which otherwise would be skipped for entries in the dbuf cache.
5803 arc_buf_access(arc_buf_t
*buf
)
5805 mutex_enter(&buf
->b_evict_lock
);
5806 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
5809 * Avoid taking the hash_lock when possible as an optimization.
5810 * The header must be checked again under the hash_lock in order
5811 * to handle the case where it is concurrently being released.
5813 if (hdr
->b_l1hdr
.b_state
== arc_anon
|| HDR_EMPTY(hdr
)) {
5814 mutex_exit(&buf
->b_evict_lock
);
5818 kmutex_t
*hash_lock
= HDR_LOCK(hdr
);
5819 mutex_enter(hash_lock
);
5821 if (hdr
->b_l1hdr
.b_state
== arc_anon
|| HDR_EMPTY(hdr
)) {
5822 mutex_exit(hash_lock
);
5823 mutex_exit(&buf
->b_evict_lock
);
5824 ARCSTAT_BUMP(arcstat_access_skip
);
5828 mutex_exit(&buf
->b_evict_lock
);
5830 ASSERT(hdr
->b_l1hdr
.b_state
== arc_mru
||
5831 hdr
->b_l1hdr
.b_state
== arc_mfu
);
5833 DTRACE_PROBE1(arc__hit
, arc_buf_hdr_t
*, hdr
);
5834 arc_access(hdr
, hash_lock
);
5835 mutex_exit(hash_lock
);
5837 ARCSTAT_BUMP(arcstat_hits
);
5838 ARCSTAT_CONDSTAT(!HDR_PREFETCH(hdr
) && !HDR_PRESCIENT_PREFETCH(hdr
),
5839 demand
, prefetch
, !HDR_ISTYPE_METADATA(hdr
), data
, metadata
, hits
);
5842 /* a generic arc_read_done_func_t which you can use */
5845 arc_bcopy_func(zio_t
*zio
, const zbookmark_phys_t
*zb
, const blkptr_t
*bp
,
5846 arc_buf_t
*buf
, void *arg
)
5851 bcopy(buf
->b_data
, arg
, arc_buf_size(buf
));
5852 arc_buf_destroy(buf
, arg
);
5855 /* a generic arc_read_done_func_t */
5858 arc_getbuf_func(zio_t
*zio
, const zbookmark_phys_t
*zb
, const blkptr_t
*bp
,
5859 arc_buf_t
*buf
, void *arg
)
5861 arc_buf_t
**bufp
= arg
;
5864 ASSERT(zio
== NULL
|| zio
->io_error
!= 0);
5867 ASSERT(zio
== NULL
|| zio
->io_error
== 0);
5869 ASSERT(buf
->b_data
!= NULL
);
5874 arc_hdr_verify(arc_buf_hdr_t
*hdr
, blkptr_t
*bp
)
5876 if (BP_IS_HOLE(bp
) || BP_IS_EMBEDDED(bp
)) {
5877 ASSERT3U(HDR_GET_PSIZE(hdr
), ==, 0);
5878 ASSERT3U(arc_hdr_get_compress(hdr
), ==, ZIO_COMPRESS_OFF
);
5880 if (HDR_COMPRESSION_ENABLED(hdr
)) {
5881 ASSERT3U(arc_hdr_get_compress(hdr
), ==,
5882 BP_GET_COMPRESS(bp
));
5884 ASSERT3U(HDR_GET_LSIZE(hdr
), ==, BP_GET_LSIZE(bp
));
5885 ASSERT3U(HDR_GET_PSIZE(hdr
), ==, BP_GET_PSIZE(bp
));
5886 ASSERT3U(!!HDR_PROTECTED(hdr
), ==, BP_IS_PROTECTED(bp
));
5891 arc_read_done(zio_t
*zio
)
5893 blkptr_t
*bp
= zio
->io_bp
;
5894 arc_buf_hdr_t
*hdr
= zio
->io_private
;
5895 kmutex_t
*hash_lock
= NULL
;
5896 arc_callback_t
*callback_list
;
5897 arc_callback_t
*acb
;
5898 boolean_t freeable
= B_FALSE
;
5901 * The hdr was inserted into hash-table and removed from lists
5902 * prior to starting I/O. We should find this header, since
5903 * it's in the hash table, and it should be legit since it's
5904 * not possible to evict it during the I/O. The only possible
5905 * reason for it not to be found is if we were freed during the
5908 if (HDR_IN_HASH_TABLE(hdr
)) {
5909 arc_buf_hdr_t
*found
;
5911 ASSERT3U(hdr
->b_birth
, ==, BP_PHYSICAL_BIRTH(zio
->io_bp
));
5912 ASSERT3U(hdr
->b_dva
.dva_word
[0], ==,
5913 BP_IDENTITY(zio
->io_bp
)->dva_word
[0]);
5914 ASSERT3U(hdr
->b_dva
.dva_word
[1], ==,
5915 BP_IDENTITY(zio
->io_bp
)->dva_word
[1]);
5917 found
= buf_hash_find(hdr
->b_spa
, zio
->io_bp
, &hash_lock
);
5919 ASSERT((found
== hdr
&&
5920 DVA_EQUAL(&hdr
->b_dva
, BP_IDENTITY(zio
->io_bp
))) ||
5921 (found
== hdr
&& HDR_L2_READING(hdr
)));
5922 ASSERT3P(hash_lock
, !=, NULL
);
5925 if (BP_IS_PROTECTED(bp
)) {
5926 hdr
->b_crypt_hdr
.b_ot
= BP_GET_TYPE(bp
);
5927 hdr
->b_crypt_hdr
.b_dsobj
= zio
->io_bookmark
.zb_objset
;
5928 zio_crypt_decode_params_bp(bp
, hdr
->b_crypt_hdr
.b_salt
,
5929 hdr
->b_crypt_hdr
.b_iv
);
5931 if (BP_GET_TYPE(bp
) == DMU_OT_INTENT_LOG
) {
5934 tmpbuf
= abd_borrow_buf_copy(zio
->io_abd
,
5935 sizeof (zil_chain_t
));
5936 zio_crypt_decode_mac_zil(tmpbuf
,
5937 hdr
->b_crypt_hdr
.b_mac
);
5938 abd_return_buf(zio
->io_abd
, tmpbuf
,
5939 sizeof (zil_chain_t
));
5941 zio_crypt_decode_mac_bp(bp
, hdr
->b_crypt_hdr
.b_mac
);
5945 if (zio
->io_error
== 0) {
5946 /* byteswap if necessary */
5947 if (BP_SHOULD_BYTESWAP(zio
->io_bp
)) {
5948 if (BP_GET_LEVEL(zio
->io_bp
) > 0) {
5949 hdr
->b_l1hdr
.b_byteswap
= DMU_BSWAP_UINT64
;
5951 hdr
->b_l1hdr
.b_byteswap
=
5952 DMU_OT_BYTESWAP(BP_GET_TYPE(zio
->io_bp
));
5955 hdr
->b_l1hdr
.b_byteswap
= DMU_BSWAP_NUMFUNCS
;
5959 arc_hdr_clear_flags(hdr
, ARC_FLAG_L2_EVICTED
);
5960 if (l2arc_noprefetch
&& HDR_PREFETCH(hdr
))
5961 arc_hdr_clear_flags(hdr
, ARC_FLAG_L2CACHE
);
5963 callback_list
= hdr
->b_l1hdr
.b_acb
;
5964 ASSERT3P(callback_list
, !=, NULL
);
5966 if (hash_lock
&& zio
->io_error
== 0 &&
5967 hdr
->b_l1hdr
.b_state
== arc_anon
) {
5969 * Only call arc_access on anonymous buffers. This is because
5970 * if we've issued an I/O for an evicted buffer, we've already
5971 * called arc_access (to prevent any simultaneous readers from
5972 * getting confused).
5974 arc_access(hdr
, hash_lock
);
5978 * If a read request has a callback (i.e. acb_done is not NULL), then we
5979 * make a buf containing the data according to the parameters which were
5980 * passed in. The implementation of arc_buf_alloc_impl() ensures that we
5981 * aren't needlessly decompressing the data multiple times.
5983 int callback_cnt
= 0;
5984 for (acb
= callback_list
; acb
!= NULL
; acb
= acb
->acb_next
) {
5990 if (zio
->io_error
!= 0)
5993 int error
= arc_buf_alloc_impl(hdr
, zio
->io_spa
,
5994 &acb
->acb_zb
, acb
->acb_private
, acb
->acb_encrypted
,
5995 acb
->acb_compressed
, acb
->acb_noauth
, B_TRUE
,
5999 * Assert non-speculative zios didn't fail because an
6000 * encryption key wasn't loaded
6002 ASSERT((zio
->io_flags
& ZIO_FLAG_SPECULATIVE
) ||
6006 * If we failed to decrypt, report an error now (as the zio
6007 * layer would have done if it had done the transforms).
6009 if (error
== ECKSUM
) {
6010 ASSERT(BP_IS_PROTECTED(bp
));
6011 error
= SET_ERROR(EIO
);
6012 if ((zio
->io_flags
& ZIO_FLAG_SPECULATIVE
) == 0) {
6013 spa_log_error(zio
->io_spa
, &acb
->acb_zb
);
6014 zfs_ereport_post(FM_EREPORT_ZFS_AUTHENTICATION
,
6015 zio
->io_spa
, NULL
, &acb
->acb_zb
, zio
, 0, 0);
6021 * Decompression or decryption failed. Set
6022 * io_error so that when we call acb_done
6023 * (below), we will indicate that the read
6024 * failed. Note that in the unusual case
6025 * where one callback is compressed and another
6026 * uncompressed, we will mark all of them
6027 * as failed, even though the uncompressed
6028 * one can't actually fail. In this case,
6029 * the hdr will not be anonymous, because
6030 * if there are multiple callbacks, it's
6031 * because multiple threads found the same
6032 * arc buf in the hash table.
6034 zio
->io_error
= error
;
6039 * If there are multiple callbacks, we must have the hash lock,
6040 * because the only way for multiple threads to find this hdr is
6041 * in the hash table. This ensures that if there are multiple
6042 * callbacks, the hdr is not anonymous. If it were anonymous,
6043 * we couldn't use arc_buf_destroy() in the error case below.
6045 ASSERT(callback_cnt
< 2 || hash_lock
!= NULL
);
6047 hdr
->b_l1hdr
.b_acb
= NULL
;
6048 arc_hdr_clear_flags(hdr
, ARC_FLAG_IO_IN_PROGRESS
);
6049 if (callback_cnt
== 0)
6050 ASSERT(hdr
->b_l1hdr
.b_pabd
!= NULL
|| HDR_HAS_RABD(hdr
));
6052 ASSERT(zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
) ||
6053 callback_list
!= NULL
);
6055 if (zio
->io_error
== 0) {
6056 arc_hdr_verify(hdr
, zio
->io_bp
);
6058 arc_hdr_set_flags(hdr
, ARC_FLAG_IO_ERROR
);
6059 if (hdr
->b_l1hdr
.b_state
!= arc_anon
)
6060 arc_change_state(arc_anon
, hdr
, hash_lock
);
6061 if (HDR_IN_HASH_TABLE(hdr
))
6062 buf_hash_remove(hdr
);
6063 freeable
= zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
);
6067 * Broadcast before we drop the hash_lock to avoid the possibility
6068 * that the hdr (and hence the cv) might be freed before we get to
6069 * the cv_broadcast().
6071 cv_broadcast(&hdr
->b_l1hdr
.b_cv
);
6073 if (hash_lock
!= NULL
) {
6074 mutex_exit(hash_lock
);
6077 * This block was freed while we waited for the read to
6078 * complete. It has been removed from the hash table and
6079 * moved to the anonymous state (so that it won't show up
6082 ASSERT3P(hdr
->b_l1hdr
.b_state
, ==, arc_anon
);
6083 freeable
= zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
);
6086 /* execute each callback and free its structure */
6087 while ((acb
= callback_list
) != NULL
) {
6088 if (acb
->acb_done
!= NULL
) {
6089 if (zio
->io_error
!= 0 && acb
->acb_buf
!= NULL
) {
6091 * If arc_buf_alloc_impl() fails during
6092 * decompression, the buf will still be
6093 * allocated, and needs to be freed here.
6095 arc_buf_destroy(acb
->acb_buf
,
6097 acb
->acb_buf
= NULL
;
6099 acb
->acb_done(zio
, &zio
->io_bookmark
, zio
->io_bp
,
6100 acb
->acb_buf
, acb
->acb_private
);
6103 if (acb
->acb_zio_dummy
!= NULL
) {
6104 acb
->acb_zio_dummy
->io_error
= zio
->io_error
;
6105 zio_nowait(acb
->acb_zio_dummy
);
6108 callback_list
= acb
->acb_next
;
6109 kmem_free(acb
, sizeof (arc_callback_t
));
6113 arc_hdr_destroy(hdr
);
6117 * "Read" the block at the specified DVA (in bp) via the
6118 * cache. If the block is found in the cache, invoke the provided
6119 * callback immediately and return. Note that the `zio' parameter
6120 * in the callback will be NULL in this case, since no IO was
6121 * required. If the block is not in the cache pass the read request
6122 * on to the spa with a substitute callback function, so that the
6123 * requested block will be added to the cache.
6125 * If a read request arrives for a block that has a read in-progress,
6126 * either wait for the in-progress read to complete (and return the
6127 * results); or, if this is a read with a "done" func, add a record
6128 * to the read to invoke the "done" func when the read completes,
6129 * and return; or just return.
6131 * arc_read_done() will invoke all the requested "done" functions
6132 * for readers of this block.
6135 arc_read(zio_t
*pio
, spa_t
*spa
, const blkptr_t
*bp
,
6136 arc_read_done_func_t
*done
, void *private, zio_priority_t priority
,
6137 int zio_flags
, arc_flags_t
*arc_flags
, const zbookmark_phys_t
*zb
)
6139 arc_buf_hdr_t
*hdr
= NULL
;
6140 kmutex_t
*hash_lock
= NULL
;
6142 uint64_t guid
= spa_load_guid(spa
);
6143 boolean_t compressed_read
= (zio_flags
& ZIO_FLAG_RAW_COMPRESS
) != 0;
6144 boolean_t encrypted_read
= BP_IS_ENCRYPTED(bp
) &&
6145 (zio_flags
& ZIO_FLAG_RAW_ENCRYPT
) != 0;
6146 boolean_t noauth_read
= BP_IS_AUTHENTICATED(bp
) &&
6147 (zio_flags
& ZIO_FLAG_RAW_ENCRYPT
) != 0;
6150 ASSERT(!BP_IS_EMBEDDED(bp
) ||
6151 BPE_GET_ETYPE(bp
) == BP_EMBEDDED_TYPE_DATA
);
6154 if (!BP_IS_EMBEDDED(bp
)) {
6156 * Embedded BP's have no DVA and require no I/O to "read".
6157 * Create an anonymous arc buf to back it.
6159 hdr
= buf_hash_find(guid
, bp
, &hash_lock
);
6163 * Determine if we have an L1 cache hit or a cache miss. For simplicity
6164 * we maintain encrypted data seperately from compressed / uncompressed
6165 * data. If the user is requesting raw encrypted data and we don't have
6166 * that in the header we will read from disk to guarantee that we can
6167 * get it even if the encryption keys aren't loaded.
6169 if (hdr
!= NULL
&& HDR_HAS_L1HDR(hdr
) && (HDR_HAS_RABD(hdr
) ||
6170 (hdr
->b_l1hdr
.b_pabd
!= NULL
&& !encrypted_read
))) {
6171 arc_buf_t
*buf
= NULL
;
6172 *arc_flags
|= ARC_FLAG_CACHED
;
6174 if (HDR_IO_IN_PROGRESS(hdr
)) {
6175 zio_t
*head_zio
= hdr
->b_l1hdr
.b_acb
->acb_zio_head
;
6177 ASSERT3P(head_zio
, !=, NULL
);
6178 if ((hdr
->b_flags
& ARC_FLAG_PRIO_ASYNC_READ
) &&
6179 priority
== ZIO_PRIORITY_SYNC_READ
) {
6181 * This is a sync read that needs to wait for
6182 * an in-flight async read. Request that the
6183 * zio have its priority upgraded.
6185 zio_change_priority(head_zio
, priority
);
6186 DTRACE_PROBE1(arc__async__upgrade__sync
,
6187 arc_buf_hdr_t
*, hdr
);
6188 ARCSTAT_BUMP(arcstat_async_upgrade_sync
);
6190 if (hdr
->b_flags
& ARC_FLAG_PREDICTIVE_PREFETCH
) {
6191 arc_hdr_clear_flags(hdr
,
6192 ARC_FLAG_PREDICTIVE_PREFETCH
);
6195 if (*arc_flags
& ARC_FLAG_WAIT
) {
6196 cv_wait(&hdr
->b_l1hdr
.b_cv
, hash_lock
);
6197 mutex_exit(hash_lock
);
6200 ASSERT(*arc_flags
& ARC_FLAG_NOWAIT
);
6203 arc_callback_t
*acb
= NULL
;
6205 acb
= kmem_zalloc(sizeof (arc_callback_t
),
6207 acb
->acb_done
= done
;
6208 acb
->acb_private
= private;
6209 acb
->acb_compressed
= compressed_read
;
6210 acb
->acb_encrypted
= encrypted_read
;
6211 acb
->acb_noauth
= noauth_read
;
6214 acb
->acb_zio_dummy
= zio_null(pio
,
6215 spa
, NULL
, NULL
, NULL
, zio_flags
);
6217 ASSERT3P(acb
->acb_done
, !=, NULL
);
6218 acb
->acb_zio_head
= head_zio
;
6219 acb
->acb_next
= hdr
->b_l1hdr
.b_acb
;
6220 hdr
->b_l1hdr
.b_acb
= acb
;
6221 mutex_exit(hash_lock
);
6224 mutex_exit(hash_lock
);
6228 ASSERT(hdr
->b_l1hdr
.b_state
== arc_mru
||
6229 hdr
->b_l1hdr
.b_state
== arc_mfu
);
6232 if (hdr
->b_flags
& ARC_FLAG_PREDICTIVE_PREFETCH
) {
6234 * This is a demand read which does not have to
6235 * wait for i/o because we did a predictive
6236 * prefetch i/o for it, which has completed.
6239 arc__demand__hit__predictive__prefetch
,
6240 arc_buf_hdr_t
*, hdr
);
6242 arcstat_demand_hit_predictive_prefetch
);
6243 arc_hdr_clear_flags(hdr
,
6244 ARC_FLAG_PREDICTIVE_PREFETCH
);
6247 if (hdr
->b_flags
& ARC_FLAG_PRESCIENT_PREFETCH
) {
6249 arcstat_demand_hit_prescient_prefetch
);
6250 arc_hdr_clear_flags(hdr
,
6251 ARC_FLAG_PRESCIENT_PREFETCH
);
6254 ASSERT(!BP_IS_EMBEDDED(bp
) || !BP_IS_HOLE(bp
));
6256 /* Get a buf with the desired data in it. */
6257 rc
= arc_buf_alloc_impl(hdr
, spa
, zb
, private,
6258 encrypted_read
, compressed_read
, noauth_read
,
6262 * Convert authentication and decryption errors
6263 * to EIO (and generate an ereport if needed)
6264 * before leaving the ARC.
6266 rc
= SET_ERROR(EIO
);
6267 if ((zio_flags
& ZIO_FLAG_SPECULATIVE
) == 0) {
6268 spa_log_error(spa
, zb
);
6270 FM_EREPORT_ZFS_AUTHENTICATION
,
6271 spa
, NULL
, zb
, NULL
, 0, 0);
6275 (void) remove_reference(hdr
, hash_lock
,
6277 arc_buf_destroy_impl(buf
);
6281 /* assert any errors weren't due to unloaded keys */
6282 ASSERT((zio_flags
& ZIO_FLAG_SPECULATIVE
) ||
6284 } else if (*arc_flags
& ARC_FLAG_PREFETCH
&&
6285 zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
) == 0) {
6286 arc_hdr_set_flags(hdr
, ARC_FLAG_PREFETCH
);
6288 DTRACE_PROBE1(arc__hit
, arc_buf_hdr_t
*, hdr
);
6289 arc_access(hdr
, hash_lock
);
6290 if (*arc_flags
& ARC_FLAG_PRESCIENT_PREFETCH
)
6291 arc_hdr_set_flags(hdr
, ARC_FLAG_PRESCIENT_PREFETCH
);
6292 if (*arc_flags
& ARC_FLAG_L2CACHE
)
6293 arc_hdr_set_flags(hdr
, ARC_FLAG_L2CACHE
);
6294 mutex_exit(hash_lock
);
6295 ARCSTAT_BUMP(arcstat_hits
);
6296 ARCSTAT_CONDSTAT(!HDR_PREFETCH(hdr
),
6297 demand
, prefetch
, !HDR_ISTYPE_METADATA(hdr
),
6298 data
, metadata
, hits
);
6301 done(NULL
, zb
, bp
, buf
, private);
6303 uint64_t lsize
= BP_GET_LSIZE(bp
);
6304 uint64_t psize
= BP_GET_PSIZE(bp
);
6305 arc_callback_t
*acb
;
6308 boolean_t devw
= B_FALSE
;
6313 * Gracefully handle a damaged logical block size as a
6316 if (lsize
> spa_maxblocksize(spa
)) {
6317 rc
= SET_ERROR(ECKSUM
);
6322 /* this block is not in the cache */
6323 arc_buf_hdr_t
*exists
= NULL
;
6324 arc_buf_contents_t type
= BP_GET_BUFC_TYPE(bp
);
6325 hdr
= arc_hdr_alloc(spa_load_guid(spa
), psize
, lsize
,
6326 BP_IS_PROTECTED(bp
), BP_GET_COMPRESS(bp
), type
,
6329 if (!BP_IS_EMBEDDED(bp
)) {
6330 hdr
->b_dva
= *BP_IDENTITY(bp
);
6331 hdr
->b_birth
= BP_PHYSICAL_BIRTH(bp
);
6332 exists
= buf_hash_insert(hdr
, &hash_lock
);
6334 if (exists
!= NULL
) {
6335 /* somebody beat us to the hash insert */
6336 mutex_exit(hash_lock
);
6337 buf_discard_identity(hdr
);
6338 arc_hdr_destroy(hdr
);
6339 goto top
; /* restart the IO request */
6343 * This block is in the ghost cache or encrypted data
6344 * was requested and we didn't have it. If it was
6345 * L2-only (and thus didn't have an L1 hdr),
6346 * we realloc the header to add an L1 hdr.
6348 if (!HDR_HAS_L1HDR(hdr
)) {
6349 hdr
= arc_hdr_realloc(hdr
, hdr_l2only_cache
,
6353 if (GHOST_STATE(hdr
->b_l1hdr
.b_state
)) {
6354 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
6355 ASSERT(!HDR_HAS_RABD(hdr
));
6356 ASSERT(!HDR_IO_IN_PROGRESS(hdr
));
6357 ASSERT0(zfs_refcount_count(
6358 &hdr
->b_l1hdr
.b_refcnt
));
6359 ASSERT3P(hdr
->b_l1hdr
.b_buf
, ==, NULL
);
6360 ASSERT3P(hdr
->b_l1hdr
.b_freeze_cksum
, ==, NULL
);
6361 } else if (HDR_IO_IN_PROGRESS(hdr
)) {
6363 * If this header already had an IO in progress
6364 * and we are performing another IO to fetch
6365 * encrypted data we must wait until the first
6366 * IO completes so as not to confuse
6367 * arc_read_done(). This should be very rare
6368 * and so the performance impact shouldn't
6371 cv_wait(&hdr
->b_l1hdr
.b_cv
, hash_lock
);
6372 mutex_exit(hash_lock
);
6377 * This is a delicate dance that we play here.
6378 * This hdr might be in the ghost list so we access
6379 * it to move it out of the ghost list before we
6380 * initiate the read. If it's a prefetch then
6381 * it won't have a callback so we'll remove the
6382 * reference that arc_buf_alloc_impl() created. We
6383 * do this after we've called arc_access() to
6384 * avoid hitting an assert in remove_reference().
6386 arc_access(hdr
, hash_lock
);
6387 arc_hdr_alloc_abd(hdr
, encrypted_read
);
6390 if (encrypted_read
) {
6391 ASSERT(HDR_HAS_RABD(hdr
));
6392 size
= HDR_GET_PSIZE(hdr
);
6393 hdr_abd
= hdr
->b_crypt_hdr
.b_rabd
;
6394 zio_flags
|= ZIO_FLAG_RAW
;
6396 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
6397 size
= arc_hdr_size(hdr
);
6398 hdr_abd
= hdr
->b_l1hdr
.b_pabd
;
6400 if (arc_hdr_get_compress(hdr
) != ZIO_COMPRESS_OFF
) {
6401 zio_flags
|= ZIO_FLAG_RAW_COMPRESS
;
6405 * For authenticated bp's, we do not ask the ZIO layer
6406 * to authenticate them since this will cause the entire
6407 * IO to fail if the key isn't loaded. Instead, we
6408 * defer authentication until arc_buf_fill(), which will
6409 * verify the data when the key is available.
6411 if (BP_IS_AUTHENTICATED(bp
))
6412 zio_flags
|= ZIO_FLAG_RAW_ENCRYPT
;
6415 if (*arc_flags
& ARC_FLAG_PREFETCH
&&
6416 zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
))
6417 arc_hdr_set_flags(hdr
, ARC_FLAG_PREFETCH
);
6418 if (*arc_flags
& ARC_FLAG_PRESCIENT_PREFETCH
)
6419 arc_hdr_set_flags(hdr
, ARC_FLAG_PRESCIENT_PREFETCH
);
6420 if (*arc_flags
& ARC_FLAG_L2CACHE
)
6421 arc_hdr_set_flags(hdr
, ARC_FLAG_L2CACHE
);
6422 if (BP_IS_AUTHENTICATED(bp
))
6423 arc_hdr_set_flags(hdr
, ARC_FLAG_NOAUTH
);
6424 if (BP_GET_LEVEL(bp
) > 0)
6425 arc_hdr_set_flags(hdr
, ARC_FLAG_INDIRECT
);
6426 if (*arc_flags
& ARC_FLAG_PREDICTIVE_PREFETCH
)
6427 arc_hdr_set_flags(hdr
, ARC_FLAG_PREDICTIVE_PREFETCH
);
6428 ASSERT(!GHOST_STATE(hdr
->b_l1hdr
.b_state
));
6430 acb
= kmem_zalloc(sizeof (arc_callback_t
), KM_SLEEP
);
6431 acb
->acb_done
= done
;
6432 acb
->acb_private
= private;
6433 acb
->acb_compressed
= compressed_read
;
6434 acb
->acb_encrypted
= encrypted_read
;
6435 acb
->acb_noauth
= noauth_read
;
6438 ASSERT3P(hdr
->b_l1hdr
.b_acb
, ==, NULL
);
6439 hdr
->b_l1hdr
.b_acb
= acb
;
6440 arc_hdr_set_flags(hdr
, ARC_FLAG_IO_IN_PROGRESS
);
6442 if (HDR_HAS_L2HDR(hdr
) &&
6443 (vd
= hdr
->b_l2hdr
.b_dev
->l2ad_vdev
) != NULL
) {
6444 devw
= hdr
->b_l2hdr
.b_dev
->l2ad_writing
;
6445 addr
= hdr
->b_l2hdr
.b_daddr
;
6447 * Lock out L2ARC device removal.
6449 if (vdev_is_dead(vd
) ||
6450 !spa_config_tryenter(spa
, SCL_L2ARC
, vd
, RW_READER
))
6455 * We count both async reads and scrub IOs as asynchronous so
6456 * that both can be upgraded in the event of a cache hit while
6457 * the read IO is still in-flight.
6459 if (priority
== ZIO_PRIORITY_ASYNC_READ
||
6460 priority
== ZIO_PRIORITY_SCRUB
)
6461 arc_hdr_set_flags(hdr
, ARC_FLAG_PRIO_ASYNC_READ
);
6463 arc_hdr_clear_flags(hdr
, ARC_FLAG_PRIO_ASYNC_READ
);
6466 * At this point, we have a level 1 cache miss. Try again in
6467 * L2ARC if possible.
6469 ASSERT3U(HDR_GET_LSIZE(hdr
), ==, lsize
);
6471 DTRACE_PROBE4(arc__miss
, arc_buf_hdr_t
*, hdr
, blkptr_t
*, bp
,
6472 uint64_t, lsize
, zbookmark_phys_t
*, zb
);
6473 ARCSTAT_BUMP(arcstat_misses
);
6474 ARCSTAT_CONDSTAT(!HDR_PREFETCH(hdr
),
6475 demand
, prefetch
, !HDR_ISTYPE_METADATA(hdr
),
6476 data
, metadata
, misses
);
6478 if (vd
!= NULL
&& l2arc_ndev
!= 0 && !(l2arc_norw
&& devw
)) {
6480 * Read from the L2ARC if the following are true:
6481 * 1. The L2ARC vdev was previously cached.
6482 * 2. This buffer still has L2ARC metadata.
6483 * 3. This buffer isn't currently writing to the L2ARC.
6484 * 4. The L2ARC entry wasn't evicted, which may
6485 * also have invalidated the vdev.
6486 * 5. This isn't prefetch and l2arc_noprefetch is set.
6488 if (HDR_HAS_L2HDR(hdr
) &&
6489 !HDR_L2_WRITING(hdr
) && !HDR_L2_EVICTED(hdr
) &&
6490 !(l2arc_noprefetch
&& HDR_PREFETCH(hdr
))) {
6491 l2arc_read_callback_t
*cb
;
6495 DTRACE_PROBE1(l2arc__hit
, arc_buf_hdr_t
*, hdr
);
6496 ARCSTAT_BUMP(arcstat_l2_hits
);
6497 atomic_inc_32(&hdr
->b_l2hdr
.b_hits
);
6499 cb
= kmem_zalloc(sizeof (l2arc_read_callback_t
),
6501 cb
->l2rcb_hdr
= hdr
;
6504 cb
->l2rcb_flags
= zio_flags
;
6506 asize
= vdev_psize_to_asize(vd
, size
);
6507 if (asize
!= size
) {
6508 abd
= abd_alloc_for_io(asize
,
6509 HDR_ISTYPE_METADATA(hdr
));
6510 cb
->l2rcb_abd
= abd
;
6515 ASSERT(addr
>= VDEV_LABEL_START_SIZE
&&
6516 addr
+ asize
<= vd
->vdev_psize
-
6517 VDEV_LABEL_END_SIZE
);
6520 * l2arc read. The SCL_L2ARC lock will be
6521 * released by l2arc_read_done().
6522 * Issue a null zio if the underlying buffer
6523 * was squashed to zero size by compression.
6525 ASSERT3U(arc_hdr_get_compress(hdr
), !=,
6526 ZIO_COMPRESS_EMPTY
);
6527 rzio
= zio_read_phys(pio
, vd
, addr
,
6530 l2arc_read_done
, cb
, priority
,
6531 zio_flags
| ZIO_FLAG_DONT_CACHE
|
6533 ZIO_FLAG_DONT_PROPAGATE
|
6534 ZIO_FLAG_DONT_RETRY
, B_FALSE
);
6535 acb
->acb_zio_head
= rzio
;
6537 if (hash_lock
!= NULL
)
6538 mutex_exit(hash_lock
);
6540 DTRACE_PROBE2(l2arc__read
, vdev_t
*, vd
,
6542 ARCSTAT_INCR(arcstat_l2_read_bytes
,
6543 HDR_GET_PSIZE(hdr
));
6545 if (*arc_flags
& ARC_FLAG_NOWAIT
) {
6550 ASSERT(*arc_flags
& ARC_FLAG_WAIT
);
6551 if (zio_wait(rzio
) == 0)
6554 /* l2arc read error; goto zio_read() */
6555 if (hash_lock
!= NULL
)
6556 mutex_enter(hash_lock
);
6558 DTRACE_PROBE1(l2arc__miss
,
6559 arc_buf_hdr_t
*, hdr
);
6560 ARCSTAT_BUMP(arcstat_l2_misses
);
6561 if (HDR_L2_WRITING(hdr
))
6562 ARCSTAT_BUMP(arcstat_l2_rw_clash
);
6563 spa_config_exit(spa
, SCL_L2ARC
, vd
);
6567 spa_config_exit(spa
, SCL_L2ARC
, vd
);
6568 if (l2arc_ndev
!= 0) {
6569 DTRACE_PROBE1(l2arc__miss
,
6570 arc_buf_hdr_t
*, hdr
);
6571 ARCSTAT_BUMP(arcstat_l2_misses
);
6575 rzio
= zio_read(pio
, spa
, bp
, hdr_abd
, size
,
6576 arc_read_done
, hdr
, priority
, zio_flags
, zb
);
6577 acb
->acb_zio_head
= rzio
;
6579 if (hash_lock
!= NULL
)
6580 mutex_exit(hash_lock
);
6582 if (*arc_flags
& ARC_FLAG_WAIT
) {
6583 rc
= zio_wait(rzio
);
6587 ASSERT(*arc_flags
& ARC_FLAG_NOWAIT
);
6592 /* embedded bps don't actually go to disk */
6593 if (!BP_IS_EMBEDDED(bp
))
6594 spa_read_history_add(spa
, zb
, *arc_flags
);
6599 arc_add_prune_callback(arc_prune_func_t
*func
, void *private)
6603 p
= kmem_alloc(sizeof (*p
), KM_SLEEP
);
6605 p
->p_private
= private;
6606 list_link_init(&p
->p_node
);
6607 zfs_refcount_create(&p
->p_refcnt
);
6609 mutex_enter(&arc_prune_mtx
);
6610 zfs_refcount_add(&p
->p_refcnt
, &arc_prune_list
);
6611 list_insert_head(&arc_prune_list
, p
);
6612 mutex_exit(&arc_prune_mtx
);
6618 arc_remove_prune_callback(arc_prune_t
*p
)
6620 boolean_t wait
= B_FALSE
;
6621 mutex_enter(&arc_prune_mtx
);
6622 list_remove(&arc_prune_list
, p
);
6623 if (zfs_refcount_remove(&p
->p_refcnt
, &arc_prune_list
) > 0)
6625 mutex_exit(&arc_prune_mtx
);
6627 /* wait for arc_prune_task to finish */
6629 taskq_wait_outstanding(arc_prune_taskq
, 0);
6630 ASSERT0(zfs_refcount_count(&p
->p_refcnt
));
6631 zfs_refcount_destroy(&p
->p_refcnt
);
6632 kmem_free(p
, sizeof (*p
));
6636 * Notify the arc that a block was freed, and thus will never be used again.
6639 arc_freed(spa_t
*spa
, const blkptr_t
*bp
)
6642 kmutex_t
*hash_lock
;
6643 uint64_t guid
= spa_load_guid(spa
);
6645 ASSERT(!BP_IS_EMBEDDED(bp
));
6647 hdr
= buf_hash_find(guid
, bp
, &hash_lock
);
6652 * We might be trying to free a block that is still doing I/O
6653 * (i.e. prefetch) or has a reference (i.e. a dedup-ed,
6654 * dmu_sync-ed block). If this block is being prefetched, then it
6655 * would still have the ARC_FLAG_IO_IN_PROGRESS flag set on the hdr
6656 * until the I/O completes. A block may also have a reference if it is
6657 * part of a dedup-ed, dmu_synced write. The dmu_sync() function would
6658 * have written the new block to its final resting place on disk but
6659 * without the dedup flag set. This would have left the hdr in the MRU
6660 * state and discoverable. When the txg finally syncs it detects that
6661 * the block was overridden in open context and issues an override I/O.
6662 * Since this is a dedup block, the override I/O will determine if the
6663 * block is already in the DDT. If so, then it will replace the io_bp
6664 * with the bp from the DDT and allow the I/O to finish. When the I/O
6665 * reaches the done callback, dbuf_write_override_done, it will
6666 * check to see if the io_bp and io_bp_override are identical.
6667 * If they are not, then it indicates that the bp was replaced with
6668 * the bp in the DDT and the override bp is freed. This allows
6669 * us to arrive here with a reference on a block that is being
6670 * freed. So if we have an I/O in progress, or a reference to
6671 * this hdr, then we don't destroy the hdr.
6673 if (!HDR_HAS_L1HDR(hdr
) || (!HDR_IO_IN_PROGRESS(hdr
) &&
6674 zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
))) {
6675 arc_change_state(arc_anon
, hdr
, hash_lock
);
6676 arc_hdr_destroy(hdr
);
6677 mutex_exit(hash_lock
);
6679 mutex_exit(hash_lock
);
6685 * Release this buffer from the cache, making it an anonymous buffer. This
6686 * must be done after a read and prior to modifying the buffer contents.
6687 * If the buffer has more than one reference, we must make
6688 * a new hdr for the buffer.
6691 arc_release(arc_buf_t
*buf
, void *tag
)
6693 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
6696 * It would be nice to assert that if its DMU metadata (level >
6697 * 0 || it's the dnode file), then it must be syncing context.
6698 * But we don't know that information at this level.
6701 mutex_enter(&buf
->b_evict_lock
);
6703 ASSERT(HDR_HAS_L1HDR(hdr
));
6706 * We don't grab the hash lock prior to this check, because if
6707 * the buffer's header is in the arc_anon state, it won't be
6708 * linked into the hash table.
6710 if (hdr
->b_l1hdr
.b_state
== arc_anon
) {
6711 mutex_exit(&buf
->b_evict_lock
);
6712 ASSERT(!HDR_IO_IN_PROGRESS(hdr
));
6713 ASSERT(!HDR_IN_HASH_TABLE(hdr
));
6714 ASSERT(!HDR_HAS_L2HDR(hdr
));
6715 ASSERT(HDR_EMPTY(hdr
));
6717 ASSERT3U(hdr
->b_l1hdr
.b_bufcnt
, ==, 1);
6718 ASSERT3S(zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
), ==, 1);
6719 ASSERT(!list_link_active(&hdr
->b_l1hdr
.b_arc_node
));
6721 hdr
->b_l1hdr
.b_arc_access
= 0;
6724 * If the buf is being overridden then it may already
6725 * have a hdr that is not empty.
6727 buf_discard_identity(hdr
);
6733 kmutex_t
*hash_lock
= HDR_LOCK(hdr
);
6734 mutex_enter(hash_lock
);
6737 * This assignment is only valid as long as the hash_lock is
6738 * held, we must be careful not to reference state or the
6739 * b_state field after dropping the lock.
6741 arc_state_t
*state
= hdr
->b_l1hdr
.b_state
;
6742 ASSERT3P(hash_lock
, ==, HDR_LOCK(hdr
));
6743 ASSERT3P(state
, !=, arc_anon
);
6745 /* this buffer is not on any list */
6746 ASSERT3S(zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
), >, 0);
6748 if (HDR_HAS_L2HDR(hdr
)) {
6749 mutex_enter(&hdr
->b_l2hdr
.b_dev
->l2ad_mtx
);
6752 * We have to recheck this conditional again now that
6753 * we're holding the l2ad_mtx to prevent a race with
6754 * another thread which might be concurrently calling
6755 * l2arc_evict(). In that case, l2arc_evict() might have
6756 * destroyed the header's L2 portion as we were waiting
6757 * to acquire the l2ad_mtx.
6759 if (HDR_HAS_L2HDR(hdr
))
6760 arc_hdr_l2hdr_destroy(hdr
);
6762 mutex_exit(&hdr
->b_l2hdr
.b_dev
->l2ad_mtx
);
6766 * Do we have more than one buf?
6768 if (hdr
->b_l1hdr
.b_bufcnt
> 1) {
6769 arc_buf_hdr_t
*nhdr
;
6770 uint64_t spa
= hdr
->b_spa
;
6771 uint64_t psize
= HDR_GET_PSIZE(hdr
);
6772 uint64_t lsize
= HDR_GET_LSIZE(hdr
);
6773 boolean_t
protected = HDR_PROTECTED(hdr
);
6774 enum zio_compress compress
= arc_hdr_get_compress(hdr
);
6775 arc_buf_contents_t type
= arc_buf_type(hdr
);
6776 VERIFY3U(hdr
->b_type
, ==, type
);
6778 ASSERT(hdr
->b_l1hdr
.b_buf
!= buf
|| buf
->b_next
!= NULL
);
6779 (void) remove_reference(hdr
, hash_lock
, tag
);
6781 if (arc_buf_is_shared(buf
) && !ARC_BUF_COMPRESSED(buf
)) {
6782 ASSERT3P(hdr
->b_l1hdr
.b_buf
, !=, buf
);
6783 ASSERT(ARC_BUF_LAST(buf
));
6787 * Pull the data off of this hdr and attach it to
6788 * a new anonymous hdr. Also find the last buffer
6789 * in the hdr's buffer list.
6791 arc_buf_t
*lastbuf
= arc_buf_remove(hdr
, buf
);
6792 ASSERT3P(lastbuf
, !=, NULL
);
6795 * If the current arc_buf_t and the hdr are sharing their data
6796 * buffer, then we must stop sharing that block.
6798 if (arc_buf_is_shared(buf
)) {
6799 ASSERT3P(hdr
->b_l1hdr
.b_buf
, !=, buf
);
6800 VERIFY(!arc_buf_is_shared(lastbuf
));
6803 * First, sever the block sharing relationship between
6804 * buf and the arc_buf_hdr_t.
6806 arc_unshare_buf(hdr
, buf
);
6809 * Now we need to recreate the hdr's b_pabd. Since we
6810 * have lastbuf handy, we try to share with it, but if
6811 * we can't then we allocate a new b_pabd and copy the
6812 * data from buf into it.
6814 if (arc_can_share(hdr
, lastbuf
)) {
6815 arc_share_buf(hdr
, lastbuf
);
6817 arc_hdr_alloc_abd(hdr
, B_FALSE
);
6818 abd_copy_from_buf(hdr
->b_l1hdr
.b_pabd
,
6819 buf
->b_data
, psize
);
6821 VERIFY3P(lastbuf
->b_data
, !=, NULL
);
6822 } else if (HDR_SHARED_DATA(hdr
)) {
6824 * Uncompressed shared buffers are always at the end
6825 * of the list. Compressed buffers don't have the
6826 * same requirements. This makes it hard to
6827 * simply assert that the lastbuf is shared so
6828 * we rely on the hdr's compression flags to determine
6829 * if we have a compressed, shared buffer.
6831 ASSERT(arc_buf_is_shared(lastbuf
) ||
6832 arc_hdr_get_compress(hdr
) != ZIO_COMPRESS_OFF
);
6833 ASSERT(!ARC_BUF_SHARED(buf
));
6836 ASSERT(hdr
->b_l1hdr
.b_pabd
!= NULL
|| HDR_HAS_RABD(hdr
));
6837 ASSERT3P(state
, !=, arc_l2c_only
);
6839 (void) zfs_refcount_remove_many(&state
->arcs_size
,
6840 arc_buf_size(buf
), buf
);
6842 if (zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
)) {
6843 ASSERT3P(state
, !=, arc_l2c_only
);
6844 (void) zfs_refcount_remove_many(
6845 &state
->arcs_esize
[type
],
6846 arc_buf_size(buf
), buf
);
6849 hdr
->b_l1hdr
.b_bufcnt
-= 1;
6850 if (ARC_BUF_ENCRYPTED(buf
))
6851 hdr
->b_crypt_hdr
.b_ebufcnt
-= 1;
6853 arc_cksum_verify(buf
);
6854 arc_buf_unwatch(buf
);
6856 /* if this is the last uncompressed buf free the checksum */
6857 if (!arc_hdr_has_uncompressed_buf(hdr
))
6858 arc_cksum_free(hdr
);
6860 mutex_exit(hash_lock
);
6863 * Allocate a new hdr. The new hdr will contain a b_pabd
6864 * buffer which will be freed in arc_write().
6866 nhdr
= arc_hdr_alloc(spa
, psize
, lsize
, protected,
6867 compress
, type
, HDR_HAS_RABD(hdr
));
6868 ASSERT3P(nhdr
->b_l1hdr
.b_buf
, ==, NULL
);
6869 ASSERT0(nhdr
->b_l1hdr
.b_bufcnt
);
6870 ASSERT0(zfs_refcount_count(&nhdr
->b_l1hdr
.b_refcnt
));
6871 VERIFY3U(nhdr
->b_type
, ==, type
);
6872 ASSERT(!HDR_SHARED_DATA(nhdr
));
6874 nhdr
->b_l1hdr
.b_buf
= buf
;
6875 nhdr
->b_l1hdr
.b_bufcnt
= 1;
6876 if (ARC_BUF_ENCRYPTED(buf
))
6877 nhdr
->b_crypt_hdr
.b_ebufcnt
= 1;
6878 nhdr
->b_l1hdr
.b_mru_hits
= 0;
6879 nhdr
->b_l1hdr
.b_mru_ghost_hits
= 0;
6880 nhdr
->b_l1hdr
.b_mfu_hits
= 0;
6881 nhdr
->b_l1hdr
.b_mfu_ghost_hits
= 0;
6882 nhdr
->b_l1hdr
.b_l2_hits
= 0;
6883 (void) zfs_refcount_add(&nhdr
->b_l1hdr
.b_refcnt
, tag
);
6886 mutex_exit(&buf
->b_evict_lock
);
6887 (void) zfs_refcount_add_many(&arc_anon
->arcs_size
,
6888 arc_buf_size(buf
), buf
);
6890 mutex_exit(&buf
->b_evict_lock
);
6891 ASSERT(zfs_refcount_count(&hdr
->b_l1hdr
.b_refcnt
) == 1);
6892 /* protected by hash lock, or hdr is on arc_anon */
6893 ASSERT(!multilist_link_active(&hdr
->b_l1hdr
.b_arc_node
));
6894 ASSERT(!HDR_IO_IN_PROGRESS(hdr
));
6895 hdr
->b_l1hdr
.b_mru_hits
= 0;
6896 hdr
->b_l1hdr
.b_mru_ghost_hits
= 0;
6897 hdr
->b_l1hdr
.b_mfu_hits
= 0;
6898 hdr
->b_l1hdr
.b_mfu_ghost_hits
= 0;
6899 hdr
->b_l1hdr
.b_l2_hits
= 0;
6900 arc_change_state(arc_anon
, hdr
, hash_lock
);
6901 hdr
->b_l1hdr
.b_arc_access
= 0;
6903 mutex_exit(hash_lock
);
6904 buf_discard_identity(hdr
);
6910 arc_released(arc_buf_t
*buf
)
6914 mutex_enter(&buf
->b_evict_lock
);
6915 released
= (buf
->b_data
!= NULL
&&
6916 buf
->b_hdr
->b_l1hdr
.b_state
== arc_anon
);
6917 mutex_exit(&buf
->b_evict_lock
);
6923 arc_referenced(arc_buf_t
*buf
)
6927 mutex_enter(&buf
->b_evict_lock
);
6928 referenced
= (zfs_refcount_count(&buf
->b_hdr
->b_l1hdr
.b_refcnt
));
6929 mutex_exit(&buf
->b_evict_lock
);
6930 return (referenced
);
6935 arc_write_ready(zio_t
*zio
)
6937 arc_write_callback_t
*callback
= zio
->io_private
;
6938 arc_buf_t
*buf
= callback
->awcb_buf
;
6939 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
6940 blkptr_t
*bp
= zio
->io_bp
;
6941 uint64_t psize
= BP_IS_HOLE(bp
) ? 0 : BP_GET_PSIZE(bp
);
6942 fstrans_cookie_t cookie
= spl_fstrans_mark();
6944 ASSERT(HDR_HAS_L1HDR(hdr
));
6945 ASSERT(!zfs_refcount_is_zero(&buf
->b_hdr
->b_l1hdr
.b_refcnt
));
6946 ASSERT(hdr
->b_l1hdr
.b_bufcnt
> 0);
6949 * If we're reexecuting this zio because the pool suspended, then
6950 * cleanup any state that was previously set the first time the
6951 * callback was invoked.
6953 if (zio
->io_flags
& ZIO_FLAG_REEXECUTED
) {
6954 arc_cksum_free(hdr
);
6955 arc_buf_unwatch(buf
);
6956 if (hdr
->b_l1hdr
.b_pabd
!= NULL
) {
6957 if (arc_buf_is_shared(buf
)) {
6958 arc_unshare_buf(hdr
, buf
);
6960 arc_hdr_free_abd(hdr
, B_FALSE
);
6964 if (HDR_HAS_RABD(hdr
))
6965 arc_hdr_free_abd(hdr
, B_TRUE
);
6967 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
6968 ASSERT(!HDR_HAS_RABD(hdr
));
6969 ASSERT(!HDR_SHARED_DATA(hdr
));
6970 ASSERT(!arc_buf_is_shared(buf
));
6972 callback
->awcb_ready(zio
, buf
, callback
->awcb_private
);
6974 if (HDR_IO_IN_PROGRESS(hdr
))
6975 ASSERT(zio
->io_flags
& ZIO_FLAG_REEXECUTED
);
6977 arc_hdr_set_flags(hdr
, ARC_FLAG_IO_IN_PROGRESS
);
6979 if (BP_IS_PROTECTED(bp
) != !!HDR_PROTECTED(hdr
))
6980 hdr
= arc_hdr_realloc_crypt(hdr
, BP_IS_PROTECTED(bp
));
6982 if (BP_IS_PROTECTED(bp
)) {
6983 /* ZIL blocks are written through zio_rewrite */
6984 ASSERT3U(BP_GET_TYPE(bp
), !=, DMU_OT_INTENT_LOG
);
6985 ASSERT(HDR_PROTECTED(hdr
));
6987 if (BP_SHOULD_BYTESWAP(bp
)) {
6988 if (BP_GET_LEVEL(bp
) > 0) {
6989 hdr
->b_l1hdr
.b_byteswap
= DMU_BSWAP_UINT64
;
6991 hdr
->b_l1hdr
.b_byteswap
=
6992 DMU_OT_BYTESWAP(BP_GET_TYPE(bp
));
6995 hdr
->b_l1hdr
.b_byteswap
= DMU_BSWAP_NUMFUNCS
;
6998 hdr
->b_crypt_hdr
.b_ot
= BP_GET_TYPE(bp
);
6999 hdr
->b_crypt_hdr
.b_dsobj
= zio
->io_bookmark
.zb_objset
;
7000 zio_crypt_decode_params_bp(bp
, hdr
->b_crypt_hdr
.b_salt
,
7001 hdr
->b_crypt_hdr
.b_iv
);
7002 zio_crypt_decode_mac_bp(bp
, hdr
->b_crypt_hdr
.b_mac
);
7006 * If this block was written for raw encryption but the zio layer
7007 * ended up only authenticating it, adjust the buffer flags now.
7009 if (BP_IS_AUTHENTICATED(bp
) && ARC_BUF_ENCRYPTED(buf
)) {
7010 arc_hdr_set_flags(hdr
, ARC_FLAG_NOAUTH
);
7011 buf
->b_flags
&= ~ARC_BUF_FLAG_ENCRYPTED
;
7012 if (BP_GET_COMPRESS(bp
) == ZIO_COMPRESS_OFF
)
7013 buf
->b_flags
&= ~ARC_BUF_FLAG_COMPRESSED
;
7014 } else if (BP_IS_HOLE(bp
) && ARC_BUF_ENCRYPTED(buf
)) {
7015 buf
->b_flags
&= ~ARC_BUF_FLAG_ENCRYPTED
;
7016 buf
->b_flags
&= ~ARC_BUF_FLAG_COMPRESSED
;
7019 /* this must be done after the buffer flags are adjusted */
7020 arc_cksum_compute(buf
);
7022 enum zio_compress compress
;
7023 if (BP_IS_HOLE(bp
) || BP_IS_EMBEDDED(bp
)) {
7024 compress
= ZIO_COMPRESS_OFF
;
7026 ASSERT3U(HDR_GET_LSIZE(hdr
), ==, BP_GET_LSIZE(bp
));
7027 compress
= BP_GET_COMPRESS(bp
);
7029 HDR_SET_PSIZE(hdr
, psize
);
7030 arc_hdr_set_compress(hdr
, compress
);
7032 if (zio
->io_error
!= 0 || psize
== 0)
7036 * Fill the hdr with data. If the buffer is encrypted we have no choice
7037 * but to copy the data into b_radb. If the hdr is compressed, the data
7038 * we want is available from the zio, otherwise we can take it from
7041 * We might be able to share the buf's data with the hdr here. However,
7042 * doing so would cause the ARC to be full of linear ABDs if we write a
7043 * lot of shareable data. As a compromise, we check whether scattered
7044 * ABDs are allowed, and assume that if they are then the user wants
7045 * the ARC to be primarily filled with them regardless of the data being
7046 * written. Therefore, if they're allowed then we allocate one and copy
7047 * the data into it; otherwise, we share the data directly if we can.
7049 if (ARC_BUF_ENCRYPTED(buf
)) {
7050 ASSERT3U(psize
, >, 0);
7051 ASSERT(ARC_BUF_COMPRESSED(buf
));
7052 arc_hdr_alloc_abd(hdr
, B_TRUE
);
7053 abd_copy(hdr
->b_crypt_hdr
.b_rabd
, zio
->io_abd
, psize
);
7054 } else if (zfs_abd_scatter_enabled
|| !arc_can_share(hdr
, buf
)) {
7056 * Ideally, we would always copy the io_abd into b_pabd, but the
7057 * user may have disabled compressed ARC, thus we must check the
7058 * hdr's compression setting rather than the io_bp's.
7060 if (BP_IS_ENCRYPTED(bp
)) {
7061 ASSERT3U(psize
, >, 0);
7062 arc_hdr_alloc_abd(hdr
, B_TRUE
);
7063 abd_copy(hdr
->b_crypt_hdr
.b_rabd
, zio
->io_abd
, psize
);
7064 } else if (arc_hdr_get_compress(hdr
) != ZIO_COMPRESS_OFF
&&
7065 !ARC_BUF_COMPRESSED(buf
)) {
7066 ASSERT3U(psize
, >, 0);
7067 arc_hdr_alloc_abd(hdr
, B_FALSE
);
7068 abd_copy(hdr
->b_l1hdr
.b_pabd
, zio
->io_abd
, psize
);
7070 ASSERT3U(zio
->io_orig_size
, ==, arc_hdr_size(hdr
));
7071 arc_hdr_alloc_abd(hdr
, B_FALSE
);
7072 abd_copy_from_buf(hdr
->b_l1hdr
.b_pabd
, buf
->b_data
,
7076 ASSERT3P(buf
->b_data
, ==, abd_to_buf(zio
->io_orig_abd
));
7077 ASSERT3U(zio
->io_orig_size
, ==, arc_buf_size(buf
));
7078 ASSERT3U(hdr
->b_l1hdr
.b_bufcnt
, ==, 1);
7080 arc_share_buf(hdr
, buf
);
7084 arc_hdr_verify(hdr
, bp
);
7085 spl_fstrans_unmark(cookie
);
7089 arc_write_children_ready(zio_t
*zio
)
7091 arc_write_callback_t
*callback
= zio
->io_private
;
7092 arc_buf_t
*buf
= callback
->awcb_buf
;
7094 callback
->awcb_children_ready(zio
, buf
, callback
->awcb_private
);
7098 * The SPA calls this callback for each physical write that happens on behalf
7099 * of a logical write. See the comment in dbuf_write_physdone() for details.
7102 arc_write_physdone(zio_t
*zio
)
7104 arc_write_callback_t
*cb
= zio
->io_private
;
7105 if (cb
->awcb_physdone
!= NULL
)
7106 cb
->awcb_physdone(zio
, cb
->awcb_buf
, cb
->awcb_private
);
7110 arc_write_done(zio_t
*zio
)
7112 arc_write_callback_t
*callback
= zio
->io_private
;
7113 arc_buf_t
*buf
= callback
->awcb_buf
;
7114 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
7116 ASSERT3P(hdr
->b_l1hdr
.b_acb
, ==, NULL
);
7118 if (zio
->io_error
== 0) {
7119 arc_hdr_verify(hdr
, zio
->io_bp
);
7121 if (BP_IS_HOLE(zio
->io_bp
) || BP_IS_EMBEDDED(zio
->io_bp
)) {
7122 buf_discard_identity(hdr
);
7124 hdr
->b_dva
= *BP_IDENTITY(zio
->io_bp
);
7125 hdr
->b_birth
= BP_PHYSICAL_BIRTH(zio
->io_bp
);
7128 ASSERT(HDR_EMPTY(hdr
));
7132 * If the block to be written was all-zero or compressed enough to be
7133 * embedded in the BP, no write was performed so there will be no
7134 * dva/birth/checksum. The buffer must therefore remain anonymous
7137 if (!HDR_EMPTY(hdr
)) {
7138 arc_buf_hdr_t
*exists
;
7139 kmutex_t
*hash_lock
;
7141 ASSERT3U(zio
->io_error
, ==, 0);
7143 arc_cksum_verify(buf
);
7145 exists
= buf_hash_insert(hdr
, &hash_lock
);
7146 if (exists
!= NULL
) {
7148 * This can only happen if we overwrite for
7149 * sync-to-convergence, because we remove
7150 * buffers from the hash table when we arc_free().
7152 if (zio
->io_flags
& ZIO_FLAG_IO_REWRITE
) {
7153 if (!BP_EQUAL(&zio
->io_bp_orig
, zio
->io_bp
))
7154 panic("bad overwrite, hdr=%p exists=%p",
7155 (void *)hdr
, (void *)exists
);
7156 ASSERT(zfs_refcount_is_zero(
7157 &exists
->b_l1hdr
.b_refcnt
));
7158 arc_change_state(arc_anon
, exists
, hash_lock
);
7159 mutex_exit(hash_lock
);
7160 arc_hdr_destroy(exists
);
7161 exists
= buf_hash_insert(hdr
, &hash_lock
);
7162 ASSERT3P(exists
, ==, NULL
);
7163 } else if (zio
->io_flags
& ZIO_FLAG_NOPWRITE
) {
7165 ASSERT(zio
->io_prop
.zp_nopwrite
);
7166 if (!BP_EQUAL(&zio
->io_bp_orig
, zio
->io_bp
))
7167 panic("bad nopwrite, hdr=%p exists=%p",
7168 (void *)hdr
, (void *)exists
);
7171 ASSERT(hdr
->b_l1hdr
.b_bufcnt
== 1);
7172 ASSERT(hdr
->b_l1hdr
.b_state
== arc_anon
);
7173 ASSERT(BP_GET_DEDUP(zio
->io_bp
));
7174 ASSERT(BP_GET_LEVEL(zio
->io_bp
) == 0);
7177 arc_hdr_clear_flags(hdr
, ARC_FLAG_IO_IN_PROGRESS
);
7178 /* if it's not anon, we are doing a scrub */
7179 if (exists
== NULL
&& hdr
->b_l1hdr
.b_state
== arc_anon
)
7180 arc_access(hdr
, hash_lock
);
7181 mutex_exit(hash_lock
);
7183 arc_hdr_clear_flags(hdr
, ARC_FLAG_IO_IN_PROGRESS
);
7186 ASSERT(!zfs_refcount_is_zero(&hdr
->b_l1hdr
.b_refcnt
));
7187 callback
->awcb_done(zio
, buf
, callback
->awcb_private
);
7189 abd_put(zio
->io_abd
);
7190 kmem_free(callback
, sizeof (arc_write_callback_t
));
7194 arc_write(zio_t
*pio
, spa_t
*spa
, uint64_t txg
,
7195 blkptr_t
*bp
, arc_buf_t
*buf
, boolean_t l2arc
,
7196 const zio_prop_t
*zp
, arc_write_done_func_t
*ready
,
7197 arc_write_done_func_t
*children_ready
, arc_write_done_func_t
*physdone
,
7198 arc_write_done_func_t
*done
, void *private, zio_priority_t priority
,
7199 int zio_flags
, const zbookmark_phys_t
*zb
)
7201 arc_buf_hdr_t
*hdr
= buf
->b_hdr
;
7202 arc_write_callback_t
*callback
;
7204 zio_prop_t localprop
= *zp
;
7206 ASSERT3P(ready
, !=, NULL
);
7207 ASSERT3P(done
, !=, NULL
);
7208 ASSERT(!HDR_IO_ERROR(hdr
));
7209 ASSERT(!HDR_IO_IN_PROGRESS(hdr
));
7210 ASSERT3P(hdr
->b_l1hdr
.b_acb
, ==, NULL
);
7211 ASSERT3U(hdr
->b_l1hdr
.b_bufcnt
, >, 0);
7213 arc_hdr_set_flags(hdr
, ARC_FLAG_L2CACHE
);
7215 if (ARC_BUF_ENCRYPTED(buf
)) {
7216 ASSERT(ARC_BUF_COMPRESSED(buf
));
7217 localprop
.zp_encrypt
= B_TRUE
;
7218 localprop
.zp_compress
= HDR_GET_COMPRESS(hdr
);
7219 localprop
.zp_byteorder
=
7220 (hdr
->b_l1hdr
.b_byteswap
== DMU_BSWAP_NUMFUNCS
) ?
7221 ZFS_HOST_BYTEORDER
: !ZFS_HOST_BYTEORDER
;
7222 bcopy(hdr
->b_crypt_hdr
.b_salt
, localprop
.zp_salt
,
7224 bcopy(hdr
->b_crypt_hdr
.b_iv
, localprop
.zp_iv
,
7226 bcopy(hdr
->b_crypt_hdr
.b_mac
, localprop
.zp_mac
,
7228 if (DMU_OT_IS_ENCRYPTED(localprop
.zp_type
)) {
7229 localprop
.zp_nopwrite
= B_FALSE
;
7230 localprop
.zp_copies
=
7231 MIN(localprop
.zp_copies
, SPA_DVAS_PER_BP
- 1);
7233 zio_flags
|= ZIO_FLAG_RAW
;
7234 } else if (ARC_BUF_COMPRESSED(buf
)) {
7235 ASSERT3U(HDR_GET_LSIZE(hdr
), !=, arc_buf_size(buf
));
7236 localprop
.zp_compress
= HDR_GET_COMPRESS(hdr
);
7237 zio_flags
|= ZIO_FLAG_RAW_COMPRESS
;
7239 callback
= kmem_zalloc(sizeof (arc_write_callback_t
), KM_SLEEP
);
7240 callback
->awcb_ready
= ready
;
7241 callback
->awcb_children_ready
= children_ready
;
7242 callback
->awcb_physdone
= physdone
;
7243 callback
->awcb_done
= done
;
7244 callback
->awcb_private
= private;
7245 callback
->awcb_buf
= buf
;
7248 * The hdr's b_pabd is now stale, free it now. A new data block
7249 * will be allocated when the zio pipeline calls arc_write_ready().
7251 if (hdr
->b_l1hdr
.b_pabd
!= NULL
) {
7253 * If the buf is currently sharing the data block with
7254 * the hdr then we need to break that relationship here.
7255 * The hdr will remain with a NULL data pointer and the
7256 * buf will take sole ownership of the block.
7258 if (arc_buf_is_shared(buf
)) {
7259 arc_unshare_buf(hdr
, buf
);
7261 arc_hdr_free_abd(hdr
, B_FALSE
);
7263 VERIFY3P(buf
->b_data
, !=, NULL
);
7266 if (HDR_HAS_RABD(hdr
))
7267 arc_hdr_free_abd(hdr
, B_TRUE
);
7269 if (!(zio_flags
& ZIO_FLAG_RAW
))
7270 arc_hdr_set_compress(hdr
, ZIO_COMPRESS_OFF
);
7272 ASSERT(!arc_buf_is_shared(buf
));
7273 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, ==, NULL
);
7275 zio
= zio_write(pio
, spa
, txg
, bp
,
7276 abd_get_from_buf(buf
->b_data
, HDR_GET_LSIZE(hdr
)),
7277 HDR_GET_LSIZE(hdr
), arc_buf_size(buf
), &localprop
, arc_write_ready
,
7278 (children_ready
!= NULL
) ? arc_write_children_ready
: NULL
,
7279 arc_write_physdone
, arc_write_done
, callback
,
7280 priority
, zio_flags
, zb
);
7286 arc_memory_throttle(spa_t
*spa
, uint64_t reserve
, uint64_t txg
)
7289 uint64_t available_memory
= arc_free_memory();
7293 MIN(available_memory
, vmem_size(heap_arena
, VMEM_FREE
));
7296 if (available_memory
> arc_all_memory() * arc_lotsfree_percent
/ 100)
7299 if (txg
> spa
->spa_lowmem_last_txg
) {
7300 spa
->spa_lowmem_last_txg
= txg
;
7301 spa
->spa_lowmem_page_load
= 0;
7304 * If we are in pageout, we know that memory is already tight,
7305 * the arc is already going to be evicting, so we just want to
7306 * continue to let page writes occur as quickly as possible.
7308 if (current_is_kswapd()) {
7309 if (spa
->spa_lowmem_page_load
>
7310 MAX(arc_sys_free
/ 4, available_memory
) / 4) {
7311 DMU_TX_STAT_BUMP(dmu_tx_memory_reclaim
);
7312 return (SET_ERROR(ERESTART
));
7314 /* Note: reserve is inflated, so we deflate */
7315 atomic_add_64(&spa
->spa_lowmem_page_load
, reserve
/ 8);
7317 } else if (spa
->spa_lowmem_page_load
> 0 && arc_reclaim_needed()) {
7318 /* memory is low, delay before restarting */
7319 ARCSTAT_INCR(arcstat_memory_throttle_count
, 1);
7320 DMU_TX_STAT_BUMP(dmu_tx_memory_reclaim
);
7321 return (SET_ERROR(EAGAIN
));
7323 spa
->spa_lowmem_page_load
= 0;
7324 #endif /* _KERNEL */
7329 arc_tempreserve_clear(uint64_t reserve
)
7331 atomic_add_64(&arc_tempreserve
, -reserve
);
7332 ASSERT((int64_t)arc_tempreserve
>= 0);
7336 arc_tempreserve_space(spa_t
*spa
, uint64_t reserve
, uint64_t txg
)
7342 reserve
> arc_c
/4 &&
7343 reserve
* 4 > (2ULL << SPA_MAXBLOCKSHIFT
))
7344 arc_c
= MIN(arc_c_max
, reserve
* 4);
7347 * Throttle when the calculated memory footprint for the TXG
7348 * exceeds the target ARC size.
7350 if (reserve
> arc_c
) {
7351 DMU_TX_STAT_BUMP(dmu_tx_memory_reserve
);
7352 return (SET_ERROR(ERESTART
));
7356 * Don't count loaned bufs as in flight dirty data to prevent long
7357 * network delays from blocking transactions that are ready to be
7358 * assigned to a txg.
7361 /* assert that it has not wrapped around */
7362 ASSERT3S(atomic_add_64_nv(&arc_loaned_bytes
, 0), >=, 0);
7364 anon_size
= MAX((int64_t)(zfs_refcount_count(&arc_anon
->arcs_size
) -
7365 arc_loaned_bytes
), 0);
7368 * Writes will, almost always, require additional memory allocations
7369 * in order to compress/encrypt/etc the data. We therefore need to
7370 * make sure that there is sufficient available memory for this.
7372 error
= arc_memory_throttle(spa
, reserve
, txg
);
7377 * Throttle writes when the amount of dirty data in the cache
7378 * gets too large. We try to keep the cache less than half full
7379 * of dirty blocks so that our sync times don't grow too large.
7381 * In the case of one pool being built on another pool, we want
7382 * to make sure we don't end up throttling the lower (backing)
7383 * pool when the upper pool is the majority contributor to dirty
7384 * data. To insure we make forward progress during throttling, we
7385 * also check the current pool's net dirty data and only throttle
7386 * if it exceeds zfs_arc_pool_dirty_percent of the anonymous dirty
7387 * data in the cache.
7389 * Note: if two requests come in concurrently, we might let them
7390 * both succeed, when one of them should fail. Not a huge deal.
7392 uint64_t total_dirty
= reserve
+ arc_tempreserve
+ anon_size
;
7393 uint64_t spa_dirty_anon
= spa_dirty_data(spa
);
7395 if (total_dirty
> arc_c
* zfs_arc_dirty_limit_percent
/ 100 &&
7396 anon_size
> arc_c
* zfs_arc_anon_limit_percent
/ 100 &&
7397 spa_dirty_anon
> anon_size
* zfs_arc_pool_dirty_percent
/ 100) {
7399 uint64_t meta_esize
= zfs_refcount_count(
7400 &arc_anon
->arcs_esize
[ARC_BUFC_METADATA
]);
7401 uint64_t data_esize
=
7402 zfs_refcount_count(&arc_anon
->arcs_esize
[ARC_BUFC_DATA
]);
7403 dprintf("failing, arc_tempreserve=%lluK anon_meta=%lluK "
7404 "anon_data=%lluK tempreserve=%lluK arc_c=%lluK\n",
7405 arc_tempreserve
>> 10, meta_esize
>> 10,
7406 data_esize
>> 10, reserve
>> 10, arc_c
>> 10);
7408 DMU_TX_STAT_BUMP(dmu_tx_dirty_throttle
);
7409 return (SET_ERROR(ERESTART
));
7411 atomic_add_64(&arc_tempreserve
, reserve
);
7416 arc_kstat_update_state(arc_state_t
*state
, kstat_named_t
*size
,
7417 kstat_named_t
*evict_data
, kstat_named_t
*evict_metadata
)
7419 size
->value
.ui64
= zfs_refcount_count(&state
->arcs_size
);
7420 evict_data
->value
.ui64
=
7421 zfs_refcount_count(&state
->arcs_esize
[ARC_BUFC_DATA
]);
7422 evict_metadata
->value
.ui64
=
7423 zfs_refcount_count(&state
->arcs_esize
[ARC_BUFC_METADATA
]);
7427 arc_kstat_update(kstat_t
*ksp
, int rw
)
7429 arc_stats_t
*as
= ksp
->ks_data
;
7431 if (rw
== KSTAT_WRITE
) {
7432 return (SET_ERROR(EACCES
));
7434 arc_kstat_update_state(arc_anon
,
7435 &as
->arcstat_anon_size
,
7436 &as
->arcstat_anon_evictable_data
,
7437 &as
->arcstat_anon_evictable_metadata
);
7438 arc_kstat_update_state(arc_mru
,
7439 &as
->arcstat_mru_size
,
7440 &as
->arcstat_mru_evictable_data
,
7441 &as
->arcstat_mru_evictable_metadata
);
7442 arc_kstat_update_state(arc_mru_ghost
,
7443 &as
->arcstat_mru_ghost_size
,
7444 &as
->arcstat_mru_ghost_evictable_data
,
7445 &as
->arcstat_mru_ghost_evictable_metadata
);
7446 arc_kstat_update_state(arc_mfu
,
7447 &as
->arcstat_mfu_size
,
7448 &as
->arcstat_mfu_evictable_data
,
7449 &as
->arcstat_mfu_evictable_metadata
);
7450 arc_kstat_update_state(arc_mfu_ghost
,
7451 &as
->arcstat_mfu_ghost_size
,
7452 &as
->arcstat_mfu_ghost_evictable_data
,
7453 &as
->arcstat_mfu_ghost_evictable_metadata
);
7455 ARCSTAT(arcstat_size
) = aggsum_value(&arc_size
);
7456 ARCSTAT(arcstat_meta_used
) = aggsum_value(&arc_meta_used
);
7457 ARCSTAT(arcstat_data_size
) = aggsum_value(&astat_data_size
);
7458 ARCSTAT(arcstat_metadata_size
) =
7459 aggsum_value(&astat_metadata_size
);
7460 ARCSTAT(arcstat_hdr_size
) = aggsum_value(&astat_hdr_size
);
7461 ARCSTAT(arcstat_l2_hdr_size
) = aggsum_value(&astat_l2_hdr_size
);
7462 ARCSTAT(arcstat_dbuf_size
) = aggsum_value(&astat_dbuf_size
);
7463 ARCSTAT(arcstat_dnode_size
) = aggsum_value(&astat_dnode_size
);
7464 ARCSTAT(arcstat_bonus_size
) = aggsum_value(&astat_bonus_size
);
7466 as
->arcstat_memory_all_bytes
.value
.ui64
=
7468 as
->arcstat_memory_free_bytes
.value
.ui64
=
7470 as
->arcstat_memory_available_bytes
.value
.i64
=
7471 arc_available_memory();
7478 * This function *must* return indices evenly distributed between all
7479 * sublists of the multilist. This is needed due to how the ARC eviction
7480 * code is laid out; arc_evict_state() assumes ARC buffers are evenly
7481 * distributed between all sublists and uses this assumption when
7482 * deciding which sublist to evict from and how much to evict from it.
7485 arc_state_multilist_index_func(multilist_t
*ml
, void *obj
)
7487 arc_buf_hdr_t
*hdr
= obj
;
7490 * We rely on b_dva to generate evenly distributed index
7491 * numbers using buf_hash below. So, as an added precaution,
7492 * let's make sure we never add empty buffers to the arc lists.
7494 ASSERT(!HDR_EMPTY(hdr
));
7497 * The assumption here, is the hash value for a given
7498 * arc_buf_hdr_t will remain constant throughout its lifetime
7499 * (i.e. its b_spa, b_dva, and b_birth fields don't change).
7500 * Thus, we don't need to store the header's sublist index
7501 * on insertion, as this index can be recalculated on removal.
7503 * Also, the low order bits of the hash value are thought to be
7504 * distributed evenly. Otherwise, in the case that the multilist
7505 * has a power of two number of sublists, each sublists' usage
7506 * would not be evenly distributed.
7508 return (buf_hash(hdr
->b_spa
, &hdr
->b_dva
, hdr
->b_birth
) %
7509 multilist_get_num_sublists(ml
));
7513 * Called during module initialization and periodically thereafter to
7514 * apply reasonable changes to the exposed performance tunings. Non-zero
7515 * zfs_* values which differ from the currently set values will be applied.
7518 arc_tuning_update(void)
7520 uint64_t allmem
= arc_all_memory();
7521 unsigned long limit
;
7523 /* Valid range: 64M - <all physical memory> */
7524 if ((zfs_arc_max
) && (zfs_arc_max
!= arc_c_max
) &&
7525 (zfs_arc_max
>= 64 << 20) && (zfs_arc_max
< allmem
) &&
7526 (zfs_arc_max
> arc_c_min
)) {
7527 arc_c_max
= zfs_arc_max
;
7529 arc_p
= (arc_c
>> 1);
7530 if (arc_meta_limit
> arc_c_max
)
7531 arc_meta_limit
= arc_c_max
;
7532 if (arc_dnode_limit
> arc_meta_limit
)
7533 arc_dnode_limit
= arc_meta_limit
;
7536 /* Valid range: 32M - <arc_c_max> */
7537 if ((zfs_arc_min
) && (zfs_arc_min
!= arc_c_min
) &&
7538 (zfs_arc_min
>= 2ULL << SPA_MAXBLOCKSHIFT
) &&
7539 (zfs_arc_min
<= arc_c_max
)) {
7540 arc_c_min
= zfs_arc_min
;
7541 arc_c
= MAX(arc_c
, arc_c_min
);
7544 /* Valid range: 16M - <arc_c_max> */
7545 if ((zfs_arc_meta_min
) && (zfs_arc_meta_min
!= arc_meta_min
) &&
7546 (zfs_arc_meta_min
>= 1ULL << SPA_MAXBLOCKSHIFT
) &&
7547 (zfs_arc_meta_min
<= arc_c_max
)) {
7548 arc_meta_min
= zfs_arc_meta_min
;
7549 if (arc_meta_limit
< arc_meta_min
)
7550 arc_meta_limit
= arc_meta_min
;
7551 if (arc_dnode_limit
< arc_meta_min
)
7552 arc_dnode_limit
= arc_meta_min
;
7555 /* Valid range: <arc_meta_min> - <arc_c_max> */
7556 limit
= zfs_arc_meta_limit
? zfs_arc_meta_limit
:
7557 MIN(zfs_arc_meta_limit_percent
, 100) * arc_c_max
/ 100;
7558 if ((limit
!= arc_meta_limit
) &&
7559 (limit
>= arc_meta_min
) &&
7560 (limit
<= arc_c_max
))
7561 arc_meta_limit
= limit
;
7563 /* Valid range: <arc_meta_min> - <arc_meta_limit> */
7564 limit
= zfs_arc_dnode_limit
? zfs_arc_dnode_limit
:
7565 MIN(zfs_arc_dnode_limit_percent
, 100) * arc_meta_limit
/ 100;
7566 if ((limit
!= arc_dnode_limit
) &&
7567 (limit
>= arc_meta_min
) &&
7568 (limit
<= arc_meta_limit
))
7569 arc_dnode_limit
= limit
;
7571 /* Valid range: 1 - N */
7572 if (zfs_arc_grow_retry
)
7573 arc_grow_retry
= zfs_arc_grow_retry
;
7575 /* Valid range: 1 - N */
7576 if (zfs_arc_shrink_shift
) {
7577 arc_shrink_shift
= zfs_arc_shrink_shift
;
7578 arc_no_grow_shift
= MIN(arc_no_grow_shift
, arc_shrink_shift
-1);
7581 /* Valid range: 1 - N */
7582 if (zfs_arc_p_min_shift
)
7583 arc_p_min_shift
= zfs_arc_p_min_shift
;
7585 /* Valid range: 1 - N ms */
7586 if (zfs_arc_min_prefetch_ms
)
7587 arc_min_prefetch_ms
= zfs_arc_min_prefetch_ms
;
7589 /* Valid range: 1 - N ms */
7590 if (zfs_arc_min_prescient_prefetch_ms
) {
7591 arc_min_prescient_prefetch_ms
=
7592 zfs_arc_min_prescient_prefetch_ms
;
7595 /* Valid range: 0 - 100 */
7596 if ((zfs_arc_lotsfree_percent
>= 0) &&
7597 (zfs_arc_lotsfree_percent
<= 100))
7598 arc_lotsfree_percent
= zfs_arc_lotsfree_percent
;
7600 /* Valid range: 0 - <all physical memory> */
7601 if ((zfs_arc_sys_free
) && (zfs_arc_sys_free
!= arc_sys_free
))
7602 arc_sys_free
= MIN(MAX(zfs_arc_sys_free
, 0), allmem
);
7607 arc_state_init(void)
7609 arc_anon
= &ARC_anon
;
7611 arc_mru_ghost
= &ARC_mru_ghost
;
7613 arc_mfu_ghost
= &ARC_mfu_ghost
;
7614 arc_l2c_only
= &ARC_l2c_only
;
7616 arc_mru
->arcs_list
[ARC_BUFC_METADATA
] =
7617 multilist_create(sizeof (arc_buf_hdr_t
),
7618 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7619 arc_state_multilist_index_func
);
7620 arc_mru
->arcs_list
[ARC_BUFC_DATA
] =
7621 multilist_create(sizeof (arc_buf_hdr_t
),
7622 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7623 arc_state_multilist_index_func
);
7624 arc_mru_ghost
->arcs_list
[ARC_BUFC_METADATA
] =
7625 multilist_create(sizeof (arc_buf_hdr_t
),
7626 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7627 arc_state_multilist_index_func
);
7628 arc_mru_ghost
->arcs_list
[ARC_BUFC_DATA
] =
7629 multilist_create(sizeof (arc_buf_hdr_t
),
7630 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7631 arc_state_multilist_index_func
);
7632 arc_mfu
->arcs_list
[ARC_BUFC_METADATA
] =
7633 multilist_create(sizeof (arc_buf_hdr_t
),
7634 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7635 arc_state_multilist_index_func
);
7636 arc_mfu
->arcs_list
[ARC_BUFC_DATA
] =
7637 multilist_create(sizeof (arc_buf_hdr_t
),
7638 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7639 arc_state_multilist_index_func
);
7640 arc_mfu_ghost
->arcs_list
[ARC_BUFC_METADATA
] =
7641 multilist_create(sizeof (arc_buf_hdr_t
),
7642 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7643 arc_state_multilist_index_func
);
7644 arc_mfu_ghost
->arcs_list
[ARC_BUFC_DATA
] =
7645 multilist_create(sizeof (arc_buf_hdr_t
),
7646 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7647 arc_state_multilist_index_func
);
7648 arc_l2c_only
->arcs_list
[ARC_BUFC_METADATA
] =
7649 multilist_create(sizeof (arc_buf_hdr_t
),
7650 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7651 arc_state_multilist_index_func
);
7652 arc_l2c_only
->arcs_list
[ARC_BUFC_DATA
] =
7653 multilist_create(sizeof (arc_buf_hdr_t
),
7654 offsetof(arc_buf_hdr_t
, b_l1hdr
.b_arc_node
),
7655 arc_state_multilist_index_func
);
7657 zfs_refcount_create(&arc_anon
->arcs_esize
[ARC_BUFC_METADATA
]);
7658 zfs_refcount_create(&arc_anon
->arcs_esize
[ARC_BUFC_DATA
]);
7659 zfs_refcount_create(&arc_mru
->arcs_esize
[ARC_BUFC_METADATA
]);
7660 zfs_refcount_create(&arc_mru
->arcs_esize
[ARC_BUFC_DATA
]);
7661 zfs_refcount_create(&arc_mru_ghost
->arcs_esize
[ARC_BUFC_METADATA
]);
7662 zfs_refcount_create(&arc_mru_ghost
->arcs_esize
[ARC_BUFC_DATA
]);
7663 zfs_refcount_create(&arc_mfu
->arcs_esize
[ARC_BUFC_METADATA
]);
7664 zfs_refcount_create(&arc_mfu
->arcs_esize
[ARC_BUFC_DATA
]);
7665 zfs_refcount_create(&arc_mfu_ghost
->arcs_esize
[ARC_BUFC_METADATA
]);
7666 zfs_refcount_create(&arc_mfu_ghost
->arcs_esize
[ARC_BUFC_DATA
]);
7667 zfs_refcount_create(&arc_l2c_only
->arcs_esize
[ARC_BUFC_METADATA
]);
7668 zfs_refcount_create(&arc_l2c_only
->arcs_esize
[ARC_BUFC_DATA
]);
7670 zfs_refcount_create(&arc_anon
->arcs_size
);
7671 zfs_refcount_create(&arc_mru
->arcs_size
);
7672 zfs_refcount_create(&arc_mru_ghost
->arcs_size
);
7673 zfs_refcount_create(&arc_mfu
->arcs_size
);
7674 zfs_refcount_create(&arc_mfu_ghost
->arcs_size
);
7675 zfs_refcount_create(&arc_l2c_only
->arcs_size
);
7677 aggsum_init(&arc_meta_used
, 0);
7678 aggsum_init(&arc_size
, 0);
7679 aggsum_init(&astat_data_size
, 0);
7680 aggsum_init(&astat_metadata_size
, 0);
7681 aggsum_init(&astat_hdr_size
, 0);
7682 aggsum_init(&astat_l2_hdr_size
, 0);
7683 aggsum_init(&astat_bonus_size
, 0);
7684 aggsum_init(&astat_dnode_size
, 0);
7685 aggsum_init(&astat_dbuf_size
, 0);
7687 arc_anon
->arcs_state
= ARC_STATE_ANON
;
7688 arc_mru
->arcs_state
= ARC_STATE_MRU
;
7689 arc_mru_ghost
->arcs_state
= ARC_STATE_MRU_GHOST
;
7690 arc_mfu
->arcs_state
= ARC_STATE_MFU
;
7691 arc_mfu_ghost
->arcs_state
= ARC_STATE_MFU_GHOST
;
7692 arc_l2c_only
->arcs_state
= ARC_STATE_L2C_ONLY
;
7696 arc_state_fini(void)
7698 zfs_refcount_destroy(&arc_anon
->arcs_esize
[ARC_BUFC_METADATA
]);
7699 zfs_refcount_destroy(&arc_anon
->arcs_esize
[ARC_BUFC_DATA
]);
7700 zfs_refcount_destroy(&arc_mru
->arcs_esize
[ARC_BUFC_METADATA
]);
7701 zfs_refcount_destroy(&arc_mru
->arcs_esize
[ARC_BUFC_DATA
]);
7702 zfs_refcount_destroy(&arc_mru_ghost
->arcs_esize
[ARC_BUFC_METADATA
]);
7703 zfs_refcount_destroy(&arc_mru_ghost
->arcs_esize
[ARC_BUFC_DATA
]);
7704 zfs_refcount_destroy(&arc_mfu
->arcs_esize
[ARC_BUFC_METADATA
]);
7705 zfs_refcount_destroy(&arc_mfu
->arcs_esize
[ARC_BUFC_DATA
]);
7706 zfs_refcount_destroy(&arc_mfu_ghost
->arcs_esize
[ARC_BUFC_METADATA
]);
7707 zfs_refcount_destroy(&arc_mfu_ghost
->arcs_esize
[ARC_BUFC_DATA
]);
7708 zfs_refcount_destroy(&arc_l2c_only
->arcs_esize
[ARC_BUFC_METADATA
]);
7709 zfs_refcount_destroy(&arc_l2c_only
->arcs_esize
[ARC_BUFC_DATA
]);
7711 zfs_refcount_destroy(&arc_anon
->arcs_size
);
7712 zfs_refcount_destroy(&arc_mru
->arcs_size
);
7713 zfs_refcount_destroy(&arc_mru_ghost
->arcs_size
);
7714 zfs_refcount_destroy(&arc_mfu
->arcs_size
);
7715 zfs_refcount_destroy(&arc_mfu_ghost
->arcs_size
);
7716 zfs_refcount_destroy(&arc_l2c_only
->arcs_size
);
7718 multilist_destroy(arc_mru
->arcs_list
[ARC_BUFC_METADATA
]);
7719 multilist_destroy(arc_mru_ghost
->arcs_list
[ARC_BUFC_METADATA
]);
7720 multilist_destroy(arc_mfu
->arcs_list
[ARC_BUFC_METADATA
]);
7721 multilist_destroy(arc_mfu_ghost
->arcs_list
[ARC_BUFC_METADATA
]);
7722 multilist_destroy(arc_mru
->arcs_list
[ARC_BUFC_DATA
]);
7723 multilist_destroy(arc_mru_ghost
->arcs_list
[ARC_BUFC_DATA
]);
7724 multilist_destroy(arc_mfu
->arcs_list
[ARC_BUFC_DATA
]);
7725 multilist_destroy(arc_mfu_ghost
->arcs_list
[ARC_BUFC_DATA
]);
7726 multilist_destroy(arc_l2c_only
->arcs_list
[ARC_BUFC_METADATA
]);
7727 multilist_destroy(arc_l2c_only
->arcs_list
[ARC_BUFC_DATA
]);
7729 aggsum_fini(&arc_meta_used
);
7730 aggsum_fini(&arc_size
);
7731 aggsum_fini(&astat_data_size
);
7732 aggsum_fini(&astat_metadata_size
);
7733 aggsum_fini(&astat_hdr_size
);
7734 aggsum_fini(&astat_l2_hdr_size
);
7735 aggsum_fini(&astat_bonus_size
);
7736 aggsum_fini(&astat_dnode_size
);
7737 aggsum_fini(&astat_dbuf_size
);
7741 arc_target_bytes(void)
7749 uint64_t percent
, allmem
= arc_all_memory();
7750 mutex_init(&arc_adjust_lock
, NULL
, MUTEX_DEFAULT
, NULL
);
7751 cv_init(&arc_adjust_waiters_cv
, NULL
, CV_DEFAULT
, NULL
);
7753 arc_min_prefetch_ms
= 1000;
7754 arc_min_prescient_prefetch_ms
= 6000;
7758 * Register a shrinker to support synchronous (direct) memory
7759 * reclaim from the arc. This is done to prevent kswapd from
7760 * swapping out pages when it is preferable to shrink the arc.
7762 spl_register_shrinker(&arc_shrinker
);
7764 /* Set to 1/64 of all memory or a minimum of 512K */
7765 arc_sys_free
= MAX(allmem
/ 64, (512 * 1024));
7769 /* Set max to 1/2 of all memory */
7770 arc_c_max
= allmem
/ 2;
7773 /* Set min cache to 1/32 of all memory, or 32MB, whichever is more */
7774 arc_c_min
= MAX(allmem
/ 32, 2ULL << SPA_MAXBLOCKSHIFT
);
7777 * In userland, there's only the memory pressure that we artificially
7778 * create (see arc_available_memory()). Don't let arc_c get too
7779 * small, because it can cause transactions to be larger than
7780 * arc_c, causing arc_tempreserve_space() to fail.
7782 arc_c_min
= MAX(arc_c_max
/ 2, 2ULL << SPA_MAXBLOCKSHIFT
);
7786 arc_p
= (arc_c
>> 1);
7788 /* Set min to 1/2 of arc_c_min */
7789 arc_meta_min
= 1ULL << SPA_MAXBLOCKSHIFT
;
7790 /* Initialize maximum observed usage to zero */
7793 * Set arc_meta_limit to a percent of arc_c_max with a floor of
7794 * arc_meta_min, and a ceiling of arc_c_max.
7796 percent
= MIN(zfs_arc_meta_limit_percent
, 100);
7797 arc_meta_limit
= MAX(arc_meta_min
, (percent
* arc_c_max
) / 100);
7798 percent
= MIN(zfs_arc_dnode_limit_percent
, 100);
7799 arc_dnode_limit
= (percent
* arc_meta_limit
) / 100;
7801 /* Apply user specified tunings */
7802 arc_tuning_update();
7804 /* if kmem_flags are set, lets try to use less memory */
7805 if (kmem_debugging())
7807 if (arc_c
< arc_c_min
)
7813 * The arc must be "uninitialized", so that hdr_recl() (which is
7814 * registered by buf_init()) will not access arc_reap_zthr before
7817 ASSERT(!arc_initialized
);
7820 list_create(&arc_prune_list
, sizeof (arc_prune_t
),
7821 offsetof(arc_prune_t
, p_node
));
7822 mutex_init(&arc_prune_mtx
, NULL
, MUTEX_DEFAULT
, NULL
);
7824 arc_prune_taskq
= taskq_create("arc_prune", max_ncpus
, defclsyspri
,
7825 max_ncpus
, INT_MAX
, TASKQ_PREPOPULATE
| TASKQ_DYNAMIC
);
7827 arc_ksp
= kstat_create("zfs", 0, "arcstats", "misc", KSTAT_TYPE_NAMED
,
7828 sizeof (arc_stats
) / sizeof (kstat_named_t
), KSTAT_FLAG_VIRTUAL
);
7830 if (arc_ksp
!= NULL
) {
7831 arc_ksp
->ks_data
= &arc_stats
;
7832 arc_ksp
->ks_update
= arc_kstat_update
;
7833 kstat_install(arc_ksp
);
7836 arc_adjust_zthr
= zthr_create(arc_adjust_cb_check
,
7837 arc_adjust_cb
, NULL
);
7838 arc_reap_zthr
= zthr_create_timer(arc_reap_cb_check
,
7839 arc_reap_cb
, NULL
, SEC2NSEC(1));
7841 arc_initialized
= B_TRUE
;
7845 * Calculate maximum amount of dirty data per pool.
7847 * If it has been set by a module parameter, take that.
7848 * Otherwise, use a percentage of physical memory defined by
7849 * zfs_dirty_data_max_percent (default 10%) with a cap at
7850 * zfs_dirty_data_max_max (default 4G or 25% of physical memory).
7852 if (zfs_dirty_data_max_max
== 0)
7853 zfs_dirty_data_max_max
= MIN(4ULL * 1024 * 1024 * 1024,
7854 allmem
* zfs_dirty_data_max_max_percent
/ 100);
7856 if (zfs_dirty_data_max
== 0) {
7857 zfs_dirty_data_max
= allmem
*
7858 zfs_dirty_data_max_percent
/ 100;
7859 zfs_dirty_data_max
= MIN(zfs_dirty_data_max
,
7860 zfs_dirty_data_max_max
);
7870 spl_unregister_shrinker(&arc_shrinker
);
7871 #endif /* _KERNEL */
7873 /* Use B_TRUE to ensure *all* buffers are evicted */
7874 arc_flush(NULL
, B_TRUE
);
7876 arc_initialized
= B_FALSE
;
7878 if (arc_ksp
!= NULL
) {
7879 kstat_delete(arc_ksp
);
7883 taskq_wait(arc_prune_taskq
);
7884 taskq_destroy(arc_prune_taskq
);
7886 mutex_enter(&arc_prune_mtx
);
7887 while ((p
= list_head(&arc_prune_list
)) != NULL
) {
7888 list_remove(&arc_prune_list
, p
);
7889 zfs_refcount_remove(&p
->p_refcnt
, &arc_prune_list
);
7890 zfs_refcount_destroy(&p
->p_refcnt
);
7891 kmem_free(p
, sizeof (*p
));
7893 mutex_exit(&arc_prune_mtx
);
7895 list_destroy(&arc_prune_list
);
7896 mutex_destroy(&arc_prune_mtx
);
7897 (void) zthr_cancel(arc_adjust_zthr
);
7898 zthr_destroy(arc_adjust_zthr
);
7900 (void) zthr_cancel(arc_reap_zthr
);
7901 zthr_destroy(arc_reap_zthr
);
7903 mutex_destroy(&arc_adjust_lock
);
7904 cv_destroy(&arc_adjust_waiters_cv
);
7907 * buf_fini() must proceed arc_state_fini() because buf_fin() may
7908 * trigger the release of kmem magazines, which can callback to
7909 * arc_space_return() which accesses aggsums freed in act_state_fini().
7914 ASSERT0(arc_loaned_bytes
);
7920 * The level 2 ARC (L2ARC) is a cache layer in-between main memory and disk.
7921 * It uses dedicated storage devices to hold cached data, which are populated
7922 * using large infrequent writes. The main role of this cache is to boost
7923 * the performance of random read workloads. The intended L2ARC devices
7924 * include short-stroked disks, solid state disks, and other media with
7925 * substantially faster read latency than disk.
7927 * +-----------------------+
7929 * +-----------------------+
7932 * l2arc_feed_thread() arc_read()
7936 * +---------------+ |
7938 * +---------------+ |
7943 * +-------+ +-------+
7945 * | cache | | cache |
7946 * +-------+ +-------+
7947 * +=========+ .-----.
7948 * : L2ARC : |-_____-|
7949 * : devices : | Disks |
7950 * +=========+ `-_____-'
7952 * Read requests are satisfied from the following sources, in order:
7955 * 2) vdev cache of L2ARC devices
7957 * 4) vdev cache of disks
7960 * Some L2ARC device types exhibit extremely slow write performance.
7961 * To accommodate for this there are some significant differences between
7962 * the L2ARC and traditional cache design:
7964 * 1. There is no eviction path from the ARC to the L2ARC. Evictions from
7965 * the ARC behave as usual, freeing buffers and placing headers on ghost
7966 * lists. The ARC does not send buffers to the L2ARC during eviction as
7967 * this would add inflated write latencies for all ARC memory pressure.
7969 * 2. The L2ARC attempts to cache data from the ARC before it is evicted.
7970 * It does this by periodically scanning buffers from the eviction-end of
7971 * the MFU and MRU ARC lists, copying them to the L2ARC devices if they are
7972 * not already there. It scans until a headroom of buffers is satisfied,
7973 * which itself is a buffer for ARC eviction. If a compressible buffer is
7974 * found during scanning and selected for writing to an L2ARC device, we
7975 * temporarily boost scanning headroom during the next scan cycle to make
7976 * sure we adapt to compression effects (which might significantly reduce
7977 * the data volume we write to L2ARC). The thread that does this is
7978 * l2arc_feed_thread(), illustrated below; example sizes are included to
7979 * provide a better sense of ratio than this diagram:
7982 * +---------------------+----------+
7983 * ARC_mfu |:::::#:::::::::::::::|o#o###o###|-->. # already on L2ARC
7984 * +---------------------+----------+ | o L2ARC eligible
7985 * ARC_mru |:#:::::::::::::::::::|#o#ooo####|-->| : ARC buffer
7986 * +---------------------+----------+ |
7987 * 15.9 Gbytes ^ 32 Mbytes |
7989 * l2arc_feed_thread()
7991 * l2arc write hand <--[oooo]--'
7995 * +==============================+
7996 * L2ARC dev |####|#|###|###| |####| ... |
7997 * +==============================+
8000 * 3. If an ARC buffer is copied to the L2ARC but then hit instead of
8001 * evicted, then the L2ARC has cached a buffer much sooner than it probably
8002 * needed to, potentially wasting L2ARC device bandwidth and storage. It is
8003 * safe to say that this is an uncommon case, since buffers at the end of
8004 * the ARC lists have moved there due to inactivity.
8006 * 4. If the ARC evicts faster than the L2ARC can maintain a headroom,
8007 * then the L2ARC simply misses copying some buffers. This serves as a
8008 * pressure valve to prevent heavy read workloads from both stalling the ARC
8009 * with waits and clogging the L2ARC with writes. This also helps prevent
8010 * the potential for the L2ARC to churn if it attempts to cache content too
8011 * quickly, such as during backups of the entire pool.
8013 * 5. After system boot and before the ARC has filled main memory, there are
8014 * no evictions from the ARC and so the tails of the ARC_mfu and ARC_mru
8015 * lists can remain mostly static. Instead of searching from tail of these
8016 * lists as pictured, the l2arc_feed_thread() will search from the list heads
8017 * for eligible buffers, greatly increasing its chance of finding them.
8019 * The L2ARC device write speed is also boosted during this time so that
8020 * the L2ARC warms up faster. Since there have been no ARC evictions yet,
8021 * there are no L2ARC reads, and no fear of degrading read performance
8022 * through increased writes.
8024 * 6. Writes to the L2ARC devices are grouped and sent in-sequence, so that
8025 * the vdev queue can aggregate them into larger and fewer writes. Each
8026 * device is written to in a rotor fashion, sweeping writes through
8027 * available space then repeating.
8029 * 7. The L2ARC does not store dirty content. It never needs to flush
8030 * write buffers back to disk based storage.
8032 * 8. If an ARC buffer is written (and dirtied) which also exists in the
8033 * L2ARC, the now stale L2ARC buffer is immediately dropped.
8035 * The performance of the L2ARC can be tweaked by a number of tunables, which
8036 * may be necessary for different workloads:
8038 * l2arc_write_max max write bytes per interval
8039 * l2arc_write_boost extra write bytes during device warmup
8040 * l2arc_noprefetch skip caching prefetched buffers
8041 * l2arc_headroom number of max device writes to precache
8042 * l2arc_headroom_boost when we find compressed buffers during ARC
8043 * scanning, we multiply headroom by this
8044 * percentage factor for the next scan cycle,
8045 * since more compressed buffers are likely to
8047 * l2arc_feed_secs seconds between L2ARC writing
8049 * Tunables may be removed or added as future performance improvements are
8050 * integrated, and also may become zpool properties.
8052 * There are three key functions that control how the L2ARC warms up:
8054 * l2arc_write_eligible() check if a buffer is eligible to cache
8055 * l2arc_write_size() calculate how much to write
8056 * l2arc_write_interval() calculate sleep delay between writes
8058 * These three functions determine what to write, how much, and how quickly
8063 l2arc_write_eligible(uint64_t spa_guid
, arc_buf_hdr_t
*hdr
)
8066 * A buffer is *not* eligible for the L2ARC if it:
8067 * 1. belongs to a different spa.
8068 * 2. is already cached on the L2ARC.
8069 * 3. has an I/O in progress (it may be an incomplete read).
8070 * 4. is flagged not eligible (zfs property).
8072 if (hdr
->b_spa
!= spa_guid
|| HDR_HAS_L2HDR(hdr
) ||
8073 HDR_IO_IN_PROGRESS(hdr
) || !HDR_L2CACHE(hdr
))
8080 l2arc_write_size(void)
8085 * Make sure our globals have meaningful values in case the user
8088 size
= l2arc_write_max
;
8090 cmn_err(CE_NOTE
, "Bad value for l2arc_write_max, value must "
8091 "be greater than zero, resetting it to the default (%d)",
8093 size
= l2arc_write_max
= L2ARC_WRITE_SIZE
;
8096 if (arc_warm
== B_FALSE
)
8097 size
+= l2arc_write_boost
;
8104 l2arc_write_interval(clock_t began
, uint64_t wanted
, uint64_t wrote
)
8106 clock_t interval
, next
, now
;
8109 * If the ARC lists are busy, increase our write rate; if the
8110 * lists are stale, idle back. This is achieved by checking
8111 * how much we previously wrote - if it was more than half of
8112 * what we wanted, schedule the next write much sooner.
8114 if (l2arc_feed_again
&& wrote
> (wanted
/ 2))
8115 interval
= (hz
* l2arc_feed_min_ms
) / 1000;
8117 interval
= hz
* l2arc_feed_secs
;
8119 now
= ddi_get_lbolt();
8120 next
= MAX(now
, MIN(now
+ interval
, began
+ interval
));
8126 * Cycle through L2ARC devices. This is how L2ARC load balances.
8127 * If a device is returned, this also returns holding the spa config lock.
8129 static l2arc_dev_t
*
8130 l2arc_dev_get_next(void)
8132 l2arc_dev_t
*first
, *next
= NULL
;
8135 * Lock out the removal of spas (spa_namespace_lock), then removal
8136 * of cache devices (l2arc_dev_mtx). Once a device has been selected,
8137 * both locks will be dropped and a spa config lock held instead.
8139 mutex_enter(&spa_namespace_lock
);
8140 mutex_enter(&l2arc_dev_mtx
);
8142 /* if there are no vdevs, there is nothing to do */
8143 if (l2arc_ndev
== 0)
8147 next
= l2arc_dev_last
;
8149 /* loop around the list looking for a non-faulted vdev */
8151 next
= list_head(l2arc_dev_list
);
8153 next
= list_next(l2arc_dev_list
, next
);
8155 next
= list_head(l2arc_dev_list
);
8158 /* if we have come back to the start, bail out */
8161 else if (next
== first
)
8164 } while (vdev_is_dead(next
->l2ad_vdev
));
8166 /* if we were unable to find any usable vdevs, return NULL */
8167 if (vdev_is_dead(next
->l2ad_vdev
))
8170 l2arc_dev_last
= next
;
8173 mutex_exit(&l2arc_dev_mtx
);
8176 * Grab the config lock to prevent the 'next' device from being
8177 * removed while we are writing to it.
8180 spa_config_enter(next
->l2ad_spa
, SCL_L2ARC
, next
, RW_READER
);
8181 mutex_exit(&spa_namespace_lock
);
8187 * Free buffers that were tagged for destruction.
8190 l2arc_do_free_on_write(void)
8193 l2arc_data_free_t
*df
, *df_prev
;
8195 mutex_enter(&l2arc_free_on_write_mtx
);
8196 buflist
= l2arc_free_on_write
;
8198 for (df
= list_tail(buflist
); df
; df
= df_prev
) {
8199 df_prev
= list_prev(buflist
, df
);
8200 ASSERT3P(df
->l2df_abd
, !=, NULL
);
8201 abd_free(df
->l2df_abd
);
8202 list_remove(buflist
, df
);
8203 kmem_free(df
, sizeof (l2arc_data_free_t
));
8206 mutex_exit(&l2arc_free_on_write_mtx
);
8210 * A write to a cache device has completed. Update all headers to allow
8211 * reads from these buffers to begin.
8214 l2arc_write_done(zio_t
*zio
)
8216 l2arc_write_callback_t
*cb
;
8219 arc_buf_hdr_t
*head
, *hdr
, *hdr_prev
;
8220 kmutex_t
*hash_lock
;
8221 int64_t bytes_dropped
= 0;
8223 cb
= zio
->io_private
;
8224 ASSERT3P(cb
, !=, NULL
);
8225 dev
= cb
->l2wcb_dev
;
8226 ASSERT3P(dev
, !=, NULL
);
8227 head
= cb
->l2wcb_head
;
8228 ASSERT3P(head
, !=, NULL
);
8229 buflist
= &dev
->l2ad_buflist
;
8230 ASSERT3P(buflist
, !=, NULL
);
8231 DTRACE_PROBE2(l2arc__iodone
, zio_t
*, zio
,
8232 l2arc_write_callback_t
*, cb
);
8234 if (zio
->io_error
!= 0)
8235 ARCSTAT_BUMP(arcstat_l2_writes_error
);
8238 * All writes completed, or an error was hit.
8241 mutex_enter(&dev
->l2ad_mtx
);
8242 for (hdr
= list_prev(buflist
, head
); hdr
; hdr
= hdr_prev
) {
8243 hdr_prev
= list_prev(buflist
, hdr
);
8245 hash_lock
= HDR_LOCK(hdr
);
8248 * We cannot use mutex_enter or else we can deadlock
8249 * with l2arc_write_buffers (due to swapping the order
8250 * the hash lock and l2ad_mtx are taken).
8252 if (!mutex_tryenter(hash_lock
)) {
8254 * Missed the hash lock. We must retry so we
8255 * don't leave the ARC_FLAG_L2_WRITING bit set.
8257 ARCSTAT_BUMP(arcstat_l2_writes_lock_retry
);
8260 * We don't want to rescan the headers we've
8261 * already marked as having been written out, so
8262 * we reinsert the head node so we can pick up
8263 * where we left off.
8265 list_remove(buflist
, head
);
8266 list_insert_after(buflist
, hdr
, head
);
8268 mutex_exit(&dev
->l2ad_mtx
);
8271 * We wait for the hash lock to become available
8272 * to try and prevent busy waiting, and increase
8273 * the chance we'll be able to acquire the lock
8274 * the next time around.
8276 mutex_enter(hash_lock
);
8277 mutex_exit(hash_lock
);
8282 * We could not have been moved into the arc_l2c_only
8283 * state while in-flight due to our ARC_FLAG_L2_WRITING
8284 * bit being set. Let's just ensure that's being enforced.
8286 ASSERT(HDR_HAS_L1HDR(hdr
));
8289 * Skipped - drop L2ARC entry and mark the header as no
8290 * longer L2 eligibile.
8292 if (zio
->io_error
!= 0) {
8294 * Error - drop L2ARC entry.
8296 list_remove(buflist
, hdr
);
8297 arc_hdr_clear_flags(hdr
, ARC_FLAG_HAS_L2HDR
);
8299 ARCSTAT_INCR(arcstat_l2_psize
, -arc_hdr_size(hdr
));
8300 ARCSTAT_INCR(arcstat_l2_lsize
, -HDR_GET_LSIZE(hdr
));
8302 bytes_dropped
+= arc_hdr_size(hdr
);
8303 (void) zfs_refcount_remove_many(&dev
->l2ad_alloc
,
8304 arc_hdr_size(hdr
), hdr
);
8308 * Allow ARC to begin reads and ghost list evictions to
8311 arc_hdr_clear_flags(hdr
, ARC_FLAG_L2_WRITING
);
8313 mutex_exit(hash_lock
);
8316 atomic_inc_64(&l2arc_writes_done
);
8317 list_remove(buflist
, head
);
8318 ASSERT(!HDR_HAS_L1HDR(head
));
8319 kmem_cache_free(hdr_l2only_cache
, head
);
8320 mutex_exit(&dev
->l2ad_mtx
);
8322 vdev_space_update(dev
->l2ad_vdev
, -bytes_dropped
, 0, 0);
8324 l2arc_do_free_on_write();
8326 kmem_free(cb
, sizeof (l2arc_write_callback_t
));
8330 l2arc_untransform(zio_t
*zio
, l2arc_read_callback_t
*cb
)
8333 spa_t
*spa
= zio
->io_spa
;
8334 arc_buf_hdr_t
*hdr
= cb
->l2rcb_hdr
;
8335 blkptr_t
*bp
= zio
->io_bp
;
8336 uint8_t salt
[ZIO_DATA_SALT_LEN
];
8337 uint8_t iv
[ZIO_DATA_IV_LEN
];
8338 uint8_t mac
[ZIO_DATA_MAC_LEN
];
8339 boolean_t no_crypt
= B_FALSE
;
8342 * ZIL data is never be written to the L2ARC, so we don't need
8343 * special handling for its unique MAC storage.
8345 ASSERT3U(BP_GET_TYPE(bp
), !=, DMU_OT_INTENT_LOG
);
8346 ASSERT(MUTEX_HELD(HDR_LOCK(hdr
)));
8347 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
8350 * If the data was encrypted, decrypt it now. Note that
8351 * we must check the bp here and not the hdr, since the
8352 * hdr does not have its encryption parameters updated
8353 * until arc_read_done().
8355 if (BP_IS_ENCRYPTED(bp
)) {
8356 abd_t
*eabd
= arc_get_data_abd(hdr
, arc_hdr_size(hdr
), hdr
);
8358 zio_crypt_decode_params_bp(bp
, salt
, iv
);
8359 zio_crypt_decode_mac_bp(bp
, mac
);
8361 ret
= spa_do_crypt_abd(B_FALSE
, spa
, &cb
->l2rcb_zb
,
8362 BP_GET_TYPE(bp
), BP_GET_DEDUP(bp
), BP_SHOULD_BYTESWAP(bp
),
8363 salt
, iv
, mac
, HDR_GET_PSIZE(hdr
), eabd
,
8364 hdr
->b_l1hdr
.b_pabd
, &no_crypt
);
8366 arc_free_data_abd(hdr
, eabd
, arc_hdr_size(hdr
), hdr
);
8371 * If we actually performed decryption, replace b_pabd
8372 * with the decrypted data. Otherwise we can just throw
8373 * our decryption buffer away.
8376 arc_free_data_abd(hdr
, hdr
->b_l1hdr
.b_pabd
,
8377 arc_hdr_size(hdr
), hdr
);
8378 hdr
->b_l1hdr
.b_pabd
= eabd
;
8381 arc_free_data_abd(hdr
, eabd
, arc_hdr_size(hdr
), hdr
);
8386 * If the L2ARC block was compressed, but ARC compression
8387 * is disabled we decompress the data into a new buffer and
8388 * replace the existing data.
8390 if (HDR_GET_COMPRESS(hdr
) != ZIO_COMPRESS_OFF
&&
8391 !HDR_COMPRESSION_ENABLED(hdr
)) {
8392 abd_t
*cabd
= arc_get_data_abd(hdr
, arc_hdr_size(hdr
), hdr
);
8393 void *tmp
= abd_borrow_buf(cabd
, arc_hdr_size(hdr
));
8395 ret
= zio_decompress_data(HDR_GET_COMPRESS(hdr
),
8396 hdr
->b_l1hdr
.b_pabd
, tmp
, HDR_GET_PSIZE(hdr
),
8397 HDR_GET_LSIZE(hdr
));
8399 abd_return_buf_copy(cabd
, tmp
, arc_hdr_size(hdr
));
8400 arc_free_data_abd(hdr
, cabd
, arc_hdr_size(hdr
), hdr
);
8404 abd_return_buf_copy(cabd
, tmp
, arc_hdr_size(hdr
));
8405 arc_free_data_abd(hdr
, hdr
->b_l1hdr
.b_pabd
,
8406 arc_hdr_size(hdr
), hdr
);
8407 hdr
->b_l1hdr
.b_pabd
= cabd
;
8409 zio
->io_size
= HDR_GET_LSIZE(hdr
);
8420 * A read to a cache device completed. Validate buffer contents before
8421 * handing over to the regular ARC routines.
8424 l2arc_read_done(zio_t
*zio
)
8427 l2arc_read_callback_t
*cb
= zio
->io_private
;
8429 kmutex_t
*hash_lock
;
8430 boolean_t valid_cksum
;
8431 boolean_t using_rdata
= (BP_IS_ENCRYPTED(&cb
->l2rcb_bp
) &&
8432 (cb
->l2rcb_flags
& ZIO_FLAG_RAW_ENCRYPT
));
8434 ASSERT3P(zio
->io_vd
, !=, NULL
);
8435 ASSERT(zio
->io_flags
& ZIO_FLAG_DONT_PROPAGATE
);
8437 spa_config_exit(zio
->io_spa
, SCL_L2ARC
, zio
->io_vd
);
8439 ASSERT3P(cb
, !=, NULL
);
8440 hdr
= cb
->l2rcb_hdr
;
8441 ASSERT3P(hdr
, !=, NULL
);
8443 hash_lock
= HDR_LOCK(hdr
);
8444 mutex_enter(hash_lock
);
8445 ASSERT3P(hash_lock
, ==, HDR_LOCK(hdr
));
8448 * If the data was read into a temporary buffer,
8449 * move it and free the buffer.
8451 if (cb
->l2rcb_abd
!= NULL
) {
8452 ASSERT3U(arc_hdr_size(hdr
), <, zio
->io_size
);
8453 if (zio
->io_error
== 0) {
8455 abd_copy(hdr
->b_crypt_hdr
.b_rabd
,
8456 cb
->l2rcb_abd
, arc_hdr_size(hdr
));
8458 abd_copy(hdr
->b_l1hdr
.b_pabd
,
8459 cb
->l2rcb_abd
, arc_hdr_size(hdr
));
8464 * The following must be done regardless of whether
8465 * there was an error:
8466 * - free the temporary buffer
8467 * - point zio to the real ARC buffer
8468 * - set zio size accordingly
8469 * These are required because zio is either re-used for
8470 * an I/O of the block in the case of the error
8471 * or the zio is passed to arc_read_done() and it
8474 abd_free(cb
->l2rcb_abd
);
8475 zio
->io_size
= zio
->io_orig_size
= arc_hdr_size(hdr
);
8478 ASSERT(HDR_HAS_RABD(hdr
));
8479 zio
->io_abd
= zio
->io_orig_abd
=
8480 hdr
->b_crypt_hdr
.b_rabd
;
8482 ASSERT3P(hdr
->b_l1hdr
.b_pabd
, !=, NULL
);
8483 zio
->io_abd
= zio
->io_orig_abd
= hdr
->b_l1hdr
.b_pabd
;
8487 ASSERT3P(zio
->io_abd
, !=, NULL
);
8490 * Check this survived the L2ARC journey.
8492 ASSERT(zio
->io_abd
== hdr
->b_l1hdr
.b_pabd
||
8493 (HDR_HAS_RABD(hdr
) && zio
->io_abd
== hdr
->b_crypt_hdr
.b_rabd
));
8494 zio
->io_bp_copy
= cb
->l2rcb_bp
; /* XXX fix in L2ARC 2.0 */
8495 zio
->io_bp
= &zio
->io_bp_copy
; /* XXX fix in L2ARC 2.0 */
8497 valid_cksum
= arc_cksum_is_equal(hdr
, zio
);
8500 * b_rabd will always match the data as it exists on disk if it is
8501 * being used. Therefore if we are reading into b_rabd we do not
8502 * attempt to untransform the data.
8504 if (valid_cksum
&& !using_rdata
)
8505 tfm_error
= l2arc_untransform(zio
, cb
);
8507 if (valid_cksum
&& tfm_error
== 0 && zio
->io_error
== 0 &&
8508 !HDR_L2_EVICTED(hdr
)) {
8509 mutex_exit(hash_lock
);
8510 zio
->io_private
= hdr
;
8513 mutex_exit(hash_lock
);
8515 * Buffer didn't survive caching. Increment stats and
8516 * reissue to the original storage device.
8518 if (zio
->io_error
!= 0) {
8519 ARCSTAT_BUMP(arcstat_l2_io_error
);
8521 zio
->io_error
= SET_ERROR(EIO
);
8523 if (!valid_cksum
|| tfm_error
!= 0)
8524 ARCSTAT_BUMP(arcstat_l2_cksum_bad
);
8527 * If there's no waiter, issue an async i/o to the primary
8528 * storage now. If there *is* a waiter, the caller must
8529 * issue the i/o in a context where it's OK to block.
8531 if (zio
->io_waiter
== NULL
) {
8532 zio_t
*pio
= zio_unique_parent(zio
);
8533 void *abd
= (using_rdata
) ?
8534 hdr
->b_crypt_hdr
.b_rabd
: hdr
->b_l1hdr
.b_pabd
;
8536 ASSERT(!pio
|| pio
->io_child_type
== ZIO_CHILD_LOGICAL
);
8538 zio_nowait(zio_read(pio
, zio
->io_spa
, zio
->io_bp
,
8539 abd
, zio
->io_size
, arc_read_done
,
8540 hdr
, zio
->io_priority
, cb
->l2rcb_flags
,
8545 kmem_free(cb
, sizeof (l2arc_read_callback_t
));
8549 * This is the list priority from which the L2ARC will search for pages to
8550 * cache. This is used within loops (0..3) to cycle through lists in the
8551 * desired order. This order can have a significant effect on cache
8554 * Currently the metadata lists are hit first, MFU then MRU, followed by
8555 * the data lists. This function returns a locked list, and also returns
8558 static multilist_sublist_t
*
8559 l2arc_sublist_lock(int list_num
)
8561 multilist_t
*ml
= NULL
;
8564 ASSERT(list_num
>= 0 && list_num
< L2ARC_FEED_TYPES
);
8568 ml
= arc_mfu
->arcs_list
[ARC_BUFC_METADATA
];
8571 ml
= arc_mru
->arcs_list
[ARC_BUFC_METADATA
];
8574 ml
= arc_mfu
->arcs_list
[ARC_BUFC_DATA
];
8577 ml
= arc_mru
->arcs_list
[ARC_BUFC_DATA
];
8584 * Return a randomly-selected sublist. This is acceptable
8585 * because the caller feeds only a little bit of data for each
8586 * call (8MB). Subsequent calls will result in different
8587 * sublists being selected.
8589 idx
= multilist_get_random_index(ml
);
8590 return (multilist_sublist_lock(ml
, idx
));
8594 * Evict buffers from the device write hand to the distance specified in
8595 * bytes. This distance may span populated buffers, it may span nothing.
8596 * This is clearing a region on the L2ARC device ready for writing.
8597 * If the 'all' boolean is set, every buffer is evicted.
8600 l2arc_evict(l2arc_dev_t
*dev
, uint64_t distance
, boolean_t all
)
8603 arc_buf_hdr_t
*hdr
, *hdr_prev
;
8604 kmutex_t
*hash_lock
;
8607 buflist
= &dev
->l2ad_buflist
;
8609 if (!all
&& dev
->l2ad_first
) {
8611 * This is the first sweep through the device. There is
8617 if (dev
->l2ad_hand
>= (dev
->l2ad_end
- (2 * distance
))) {
8619 * When nearing the end of the device, evict to the end
8620 * before the device write hand jumps to the start.
8622 taddr
= dev
->l2ad_end
;
8624 taddr
= dev
->l2ad_hand
+ distance
;
8626 DTRACE_PROBE4(l2arc__evict
, l2arc_dev_t
*, dev
, list_t
*, buflist
,
8627 uint64_t, taddr
, boolean_t
, all
);
8630 mutex_enter(&dev
->l2ad_mtx
);
8631 for (hdr
= list_tail(buflist
); hdr
; hdr
= hdr_prev
) {
8632 hdr_prev
= list_prev(buflist
, hdr
);
8634 hash_lock
= HDR_LOCK(hdr
);
8637 * We cannot use mutex_enter or else we can deadlock
8638 * with l2arc_write_buffers (due to swapping the order
8639 * the hash lock and l2ad_mtx are taken).
8641 if (!mutex_tryenter(hash_lock
)) {
8643 * Missed the hash lock. Retry.
8645 ARCSTAT_BUMP(arcstat_l2_evict_lock_retry
);
8646 mutex_exit(&dev
->l2ad_mtx
);
8647 mutex_enter(hash_lock
);
8648 mutex_exit(hash_lock
);
8653 * A header can't be on this list if it doesn't have L2 header.
8655 ASSERT(HDR_HAS_L2HDR(hdr
));
8657 /* Ensure this header has finished being written. */
8658 ASSERT(!HDR_L2_WRITING(hdr
));
8659 ASSERT(!HDR_L2_WRITE_HEAD(hdr
));
8661 if (!all
&& (hdr
->b_l2hdr
.b_daddr
>= taddr
||
8662 hdr
->b_l2hdr
.b_daddr
< dev
->l2ad_hand
)) {
8664 * We've evicted to the target address,
8665 * or the end of the device.
8667 mutex_exit(hash_lock
);
8671 if (!HDR_HAS_L1HDR(hdr
)) {
8672 ASSERT(!HDR_L2_READING(hdr
));
8674 * This doesn't exist in the ARC. Destroy.
8675 * arc_hdr_destroy() will call list_remove()
8676 * and decrement arcstat_l2_lsize.
8678 arc_change_state(arc_anon
, hdr
, hash_lock
);
8679 arc_hdr_destroy(hdr
);
8681 ASSERT(hdr
->b_l1hdr
.b_state
!= arc_l2c_only
);
8682 ARCSTAT_BUMP(arcstat_l2_evict_l1cached
);
8684 * Invalidate issued or about to be issued
8685 * reads, since we may be about to write
8686 * over this location.
8688 if (HDR_L2_READING(hdr
)) {
8689 ARCSTAT_BUMP(arcstat_l2_evict_reading
);
8690 arc_hdr_set_flags(hdr
, ARC_FLAG_L2_EVICTED
);
8693 arc_hdr_l2hdr_destroy(hdr
);
8695 mutex_exit(hash_lock
);
8697 mutex_exit(&dev
->l2ad_mtx
);
8701 * Handle any abd transforms that might be required for writing to the L2ARC.
8702 * If successful, this function will always return an abd with the data
8703 * transformed as it is on disk in a new abd of asize bytes.
8706 l2arc_apply_transforms(spa_t
*spa
, arc_buf_hdr_t
*hdr
, uint64_t asize
,
8711 abd_t
*cabd
= NULL
, *eabd
= NULL
, *to_write
= hdr
->b_l1hdr
.b_pabd
;
8712 enum zio_compress compress
= HDR_GET_COMPRESS(hdr
);
8713 uint64_t psize
= HDR_GET_PSIZE(hdr
);
8714 uint64_t size
= arc_hdr_size(hdr
);
8715 boolean_t ismd
= HDR_ISTYPE_METADATA(hdr
);
8716 boolean_t bswap
= (hdr
->b_l1hdr
.b_byteswap
!= DMU_BSWAP_NUMFUNCS
);
8717 dsl_crypto_key_t
*dck
= NULL
;
8718 uint8_t mac
[ZIO_DATA_MAC_LEN
] = { 0 };
8719 boolean_t no_crypt
= B_FALSE
;
8721 ASSERT((HDR_GET_COMPRESS(hdr
) != ZIO_COMPRESS_OFF
&&
8722 !HDR_COMPRESSION_ENABLED(hdr
)) ||
8723 HDR_ENCRYPTED(hdr
) || HDR_SHARED_DATA(hdr
) || psize
!= asize
);
8724 ASSERT3U(psize
, <=, asize
);
8727 * If this data simply needs its own buffer, we simply allocate it
8728 * and copy the data. This may be done to elimiate a depedency on a
8729 * shared buffer or to reallocate the buffer to match asize.
8731 if (HDR_HAS_RABD(hdr
) && asize
!= psize
) {
8732 ASSERT3U(asize
, >=, psize
);
8733 to_write
= abd_alloc_for_io(asize
, ismd
);
8734 abd_copy(to_write
, hdr
->b_crypt_hdr
.b_rabd
, psize
);
8736 abd_zero_off(to_write
, psize
, asize
- psize
);
8740 if ((compress
== ZIO_COMPRESS_OFF
|| HDR_COMPRESSION_ENABLED(hdr
)) &&
8741 !HDR_ENCRYPTED(hdr
)) {
8742 ASSERT3U(size
, ==, psize
);
8743 to_write
= abd_alloc_for_io(asize
, ismd
);
8744 abd_copy(to_write
, hdr
->b_l1hdr
.b_pabd
, size
);
8746 abd_zero_off(to_write
, size
, asize
- size
);
8750 if (compress
!= ZIO_COMPRESS_OFF
&& !HDR_COMPRESSION_ENABLED(hdr
)) {
8751 cabd
= abd_alloc_for_io(asize
, ismd
);
8752 tmp
= abd_borrow_buf(cabd
, asize
);
8754 psize
= zio_compress_data(compress
, to_write
, tmp
, size
);
8755 ASSERT3U(psize
, <=, HDR_GET_PSIZE(hdr
));
8757 bzero((char *)tmp
+ psize
, asize
- psize
);
8758 psize
= HDR_GET_PSIZE(hdr
);
8759 abd_return_buf_copy(cabd
, tmp
, asize
);
8763 if (HDR_ENCRYPTED(hdr
)) {
8764 eabd
= abd_alloc_for_io(asize
, ismd
);
8767 * If the dataset was disowned before the buffer
8768 * made it to this point, the key to re-encrypt
8769 * it won't be available. In this case we simply
8770 * won't write the buffer to the L2ARC.
8772 ret
= spa_keystore_lookup_key(spa
, hdr
->b_crypt_hdr
.b_dsobj
,
8777 ret
= zio_do_crypt_abd(B_TRUE
, &dck
->dck_key
,
8778 hdr
->b_crypt_hdr
.b_ot
, bswap
, hdr
->b_crypt_hdr
.b_salt
,
8779 hdr
->b_crypt_hdr
.b_iv
, mac
, psize
, to_write
, eabd
,
8785 abd_copy(eabd
, to_write
, psize
);
8788 abd_zero_off(eabd
, psize
, asize
- psize
);
8790 /* assert that the MAC we got here matches the one we saved */
8791 ASSERT0(bcmp(mac
, hdr
->b_crypt_hdr
.b_mac
, ZIO_DATA_MAC_LEN
));
8792 spa_keystore_dsl_key_rele(spa
, dck
, FTAG
);
8794 if (to_write
== cabd
)
8801 ASSERT3P(to_write
, !=, hdr
->b_l1hdr
.b_pabd
);
8802 *abd_out
= to_write
;
8807 spa_keystore_dsl_key_rele(spa
, dck
, FTAG
);
8818 * Find and write ARC buffers to the L2ARC device.
8820 * An ARC_FLAG_L2_WRITING flag is set so that the L2ARC buffers are not valid
8821 * for reading until they have completed writing.
8822 * The headroom_boost is an in-out parameter used to maintain headroom boost
8823 * state between calls to this function.
8825 * Returns the number of bytes actually written (which may be smaller than
8826 * the delta by which the device hand has changed due to alignment).
8829 l2arc_write_buffers(spa_t
*spa
, l2arc_dev_t
*dev
, uint64_t target_sz
)
8831 arc_buf_hdr_t
*hdr
, *hdr_prev
, *head
;
8832 uint64_t write_asize
, write_psize
, write_lsize
, headroom
;
8834 l2arc_write_callback_t
*cb
;
8836 uint64_t guid
= spa_load_guid(spa
);
8838 ASSERT3P(dev
->l2ad_vdev
, !=, NULL
);
8841 write_lsize
= write_asize
= write_psize
= 0;
8843 head
= kmem_cache_alloc(hdr_l2only_cache
, KM_PUSHPAGE
);
8844 arc_hdr_set_flags(head
, ARC_FLAG_L2_WRITE_HEAD
| ARC_FLAG_HAS_L2HDR
);
8847 * Copy buffers for L2ARC writing.
8849 for (int try = 0; try < L2ARC_FEED_TYPES
; try++) {
8850 multilist_sublist_t
*mls
= l2arc_sublist_lock(try);
8851 uint64_t passed_sz
= 0;
8853 VERIFY3P(mls
, !=, NULL
);
8856 * L2ARC fast warmup.
8858 * Until the ARC is warm and starts to evict, read from the
8859 * head of the ARC lists rather than the tail.
8861 if (arc_warm
== B_FALSE
)
8862 hdr
= multilist_sublist_head(mls
);
8864 hdr
= multilist_sublist_tail(mls
);
8866 headroom
= target_sz
* l2arc_headroom
;
8867 if (zfs_compressed_arc_enabled
)
8868 headroom
= (headroom
* l2arc_headroom_boost
) / 100;
8870 for (; hdr
; hdr
= hdr_prev
) {
8871 kmutex_t
*hash_lock
;
8872 abd_t
*to_write
= NULL
;
8874 if (arc_warm
== B_FALSE
)
8875 hdr_prev
= multilist_sublist_next(mls
, hdr
);
8877 hdr_prev
= multilist_sublist_prev(mls
, hdr
);
8879 hash_lock
= HDR_LOCK(hdr
);
8880 if (!mutex_tryenter(hash_lock
)) {
8882 * Skip this buffer rather than waiting.
8887 passed_sz
+= HDR_GET_LSIZE(hdr
);
8888 if (passed_sz
> headroom
) {
8892 mutex_exit(hash_lock
);
8896 if (!l2arc_write_eligible(guid
, hdr
)) {
8897 mutex_exit(hash_lock
);
8902 * We rely on the L1 portion of the header below, so
8903 * it's invalid for this header to have been evicted out
8904 * of the ghost cache, prior to being written out. The
8905 * ARC_FLAG_L2_WRITING bit ensures this won't happen.
8907 ASSERT(HDR_HAS_L1HDR(hdr
));
8909 ASSERT3U(HDR_GET_PSIZE(hdr
), >, 0);
8910 ASSERT3U(arc_hdr_size(hdr
), >, 0);
8911 ASSERT(hdr
->b_l1hdr
.b_pabd
!= NULL
||
8913 uint64_t psize
= HDR_GET_PSIZE(hdr
);
8914 uint64_t asize
= vdev_psize_to_asize(dev
->l2ad_vdev
,
8917 if ((write_asize
+ asize
) > target_sz
) {
8919 mutex_exit(hash_lock
);
8924 * We rely on the L1 portion of the header below, so
8925 * it's invalid for this header to have been evicted out
8926 * of the ghost cache, prior to being written out. The
8927 * ARC_FLAG_L2_WRITING bit ensures this won't happen.
8929 arc_hdr_set_flags(hdr
, ARC_FLAG_L2_WRITING
);
8930 ASSERT(HDR_HAS_L1HDR(hdr
));
8932 ASSERT3U(HDR_GET_PSIZE(hdr
), >, 0);
8933 ASSERT(hdr
->b_l1hdr
.b_pabd
!= NULL
||
8935 ASSERT3U(arc_hdr_size(hdr
), >, 0);
8938 * If this header has b_rabd, we can use this since it
8939 * must always match the data exactly as it exists on
8940 * disk. Otherwise, the L2ARC can normally use the
8941 * hdr's data, but if we're sharing data between the
8942 * hdr and one of its bufs, L2ARC needs its own copy of
8943 * the data so that the ZIO below can't race with the
8944 * buf consumer. To ensure that this copy will be
8945 * available for the lifetime of the ZIO and be cleaned
8946 * up afterwards, we add it to the l2arc_free_on_write
8947 * queue. If we need to apply any transforms to the
8948 * data (compression, encryption) we will also need the
8951 if (HDR_HAS_RABD(hdr
) && psize
== asize
) {
8952 to_write
= hdr
->b_crypt_hdr
.b_rabd
;
8953 } else if ((HDR_COMPRESSION_ENABLED(hdr
) ||
8954 HDR_GET_COMPRESS(hdr
) == ZIO_COMPRESS_OFF
) &&
8955 !HDR_ENCRYPTED(hdr
) && !HDR_SHARED_DATA(hdr
) &&
8957 to_write
= hdr
->b_l1hdr
.b_pabd
;
8960 arc_buf_contents_t type
= arc_buf_type(hdr
);
8962 ret
= l2arc_apply_transforms(spa
, hdr
, asize
,
8965 arc_hdr_clear_flags(hdr
,
8966 ARC_FLAG_L2_WRITING
);
8967 mutex_exit(hash_lock
);
8971 l2arc_free_abd_on_write(to_write
, asize
, type
);
8976 * Insert a dummy header on the buflist so
8977 * l2arc_write_done() can find where the
8978 * write buffers begin without searching.
8980 mutex_enter(&dev
->l2ad_mtx
);
8981 list_insert_head(&dev
->l2ad_buflist
, head
);
8982 mutex_exit(&dev
->l2ad_mtx
);
8985 sizeof (l2arc_write_callback_t
), KM_SLEEP
);
8986 cb
->l2wcb_dev
= dev
;
8987 cb
->l2wcb_head
= head
;
8988 pio
= zio_root(spa
, l2arc_write_done
, cb
,
8992 hdr
->b_l2hdr
.b_dev
= dev
;
8993 hdr
->b_l2hdr
.b_hits
= 0;
8995 hdr
->b_l2hdr
.b_daddr
= dev
->l2ad_hand
;
8996 arc_hdr_set_flags(hdr
, ARC_FLAG_HAS_L2HDR
);
8998 mutex_enter(&dev
->l2ad_mtx
);
8999 list_insert_head(&dev
->l2ad_buflist
, hdr
);
9000 mutex_exit(&dev
->l2ad_mtx
);
9002 (void) zfs_refcount_add_many(&dev
->l2ad_alloc
,
9003 arc_hdr_size(hdr
), hdr
);
9005 wzio
= zio_write_phys(pio
, dev
->l2ad_vdev
,
9006 hdr
->b_l2hdr
.b_daddr
, asize
, to_write
,
9007 ZIO_CHECKSUM_OFF
, NULL
, hdr
,
9008 ZIO_PRIORITY_ASYNC_WRITE
,
9009 ZIO_FLAG_CANFAIL
, B_FALSE
);
9011 write_lsize
+= HDR_GET_LSIZE(hdr
);
9012 DTRACE_PROBE2(l2arc__write
, vdev_t
*, dev
->l2ad_vdev
,
9015 write_psize
+= psize
;
9016 write_asize
+= asize
;
9017 dev
->l2ad_hand
+= asize
;
9019 mutex_exit(hash_lock
);
9021 (void) zio_nowait(wzio
);
9024 multilist_sublist_unlock(mls
);
9030 /* No buffers selected for writing? */
9032 ASSERT0(write_lsize
);
9033 ASSERT(!HDR_HAS_L1HDR(head
));
9034 kmem_cache_free(hdr_l2only_cache
, head
);
9038 ASSERT3U(write_asize
, <=, target_sz
);
9039 ARCSTAT_BUMP(arcstat_l2_writes_sent
);
9040 ARCSTAT_INCR(arcstat_l2_write_bytes
, write_psize
);
9041 ARCSTAT_INCR(arcstat_l2_lsize
, write_lsize
);
9042 ARCSTAT_INCR(arcstat_l2_psize
, write_psize
);
9043 vdev_space_update(dev
->l2ad_vdev
, write_psize
, 0, 0);
9046 * Bump device hand to the device start if it is approaching the end.
9047 * l2arc_evict() will already have evicted ahead for this case.
9049 if (dev
->l2ad_hand
>= (dev
->l2ad_end
- target_sz
)) {
9050 dev
->l2ad_hand
= dev
->l2ad_start
;
9051 dev
->l2ad_first
= B_FALSE
;
9054 dev
->l2ad_writing
= B_TRUE
;
9055 (void) zio_wait(pio
);
9056 dev
->l2ad_writing
= B_FALSE
;
9058 return (write_asize
);
9062 * This thread feeds the L2ARC at regular intervals. This is the beating
9063 * heart of the L2ARC.
9067 l2arc_feed_thread(void *unused
)
9072 uint64_t size
, wrote
;
9073 clock_t begin
, next
= ddi_get_lbolt();
9074 fstrans_cookie_t cookie
;
9076 CALLB_CPR_INIT(&cpr
, &l2arc_feed_thr_lock
, callb_generic_cpr
, FTAG
);
9078 mutex_enter(&l2arc_feed_thr_lock
);
9080 cookie
= spl_fstrans_mark();
9081 while (l2arc_thread_exit
== 0) {
9082 CALLB_CPR_SAFE_BEGIN(&cpr
);
9083 (void) cv_timedwait_sig(&l2arc_feed_thr_cv
,
9084 &l2arc_feed_thr_lock
, next
);
9085 CALLB_CPR_SAFE_END(&cpr
, &l2arc_feed_thr_lock
);
9086 next
= ddi_get_lbolt() + hz
;
9089 * Quick check for L2ARC devices.
9091 mutex_enter(&l2arc_dev_mtx
);
9092 if (l2arc_ndev
== 0) {
9093 mutex_exit(&l2arc_dev_mtx
);
9096 mutex_exit(&l2arc_dev_mtx
);
9097 begin
= ddi_get_lbolt();
9100 * This selects the next l2arc device to write to, and in
9101 * doing so the next spa to feed from: dev->l2ad_spa. This
9102 * will return NULL if there are now no l2arc devices or if
9103 * they are all faulted.
9105 * If a device is returned, its spa's config lock is also
9106 * held to prevent device removal. l2arc_dev_get_next()
9107 * will grab and release l2arc_dev_mtx.
9109 if ((dev
= l2arc_dev_get_next()) == NULL
)
9112 spa
= dev
->l2ad_spa
;
9113 ASSERT3P(spa
, !=, NULL
);
9116 * If the pool is read-only then force the feed thread to
9117 * sleep a little longer.
9119 if (!spa_writeable(spa
)) {
9120 next
= ddi_get_lbolt() + 5 * l2arc_feed_secs
* hz
;
9121 spa_config_exit(spa
, SCL_L2ARC
, dev
);
9126 * Avoid contributing to memory pressure.
9128 if (arc_reclaim_needed()) {
9129 ARCSTAT_BUMP(arcstat_l2_abort_lowmem
);
9130 spa_config_exit(spa
, SCL_L2ARC
, dev
);
9134 ARCSTAT_BUMP(arcstat_l2_feeds
);
9136 size
= l2arc_write_size();
9139 * Evict L2ARC buffers that will be overwritten.
9141 l2arc_evict(dev
, size
, B_FALSE
);
9144 * Write ARC buffers.
9146 wrote
= l2arc_write_buffers(spa
, dev
, size
);
9149 * Calculate interval between writes.
9151 next
= l2arc_write_interval(begin
, size
, wrote
);
9152 spa_config_exit(spa
, SCL_L2ARC
, dev
);
9154 spl_fstrans_unmark(cookie
);
9156 l2arc_thread_exit
= 0;
9157 cv_broadcast(&l2arc_feed_thr_cv
);
9158 CALLB_CPR_EXIT(&cpr
); /* drops l2arc_feed_thr_lock */
9163 l2arc_vdev_present(vdev_t
*vd
)
9167 mutex_enter(&l2arc_dev_mtx
);
9168 for (dev
= list_head(l2arc_dev_list
); dev
!= NULL
;
9169 dev
= list_next(l2arc_dev_list
, dev
)) {
9170 if (dev
->l2ad_vdev
== vd
)
9173 mutex_exit(&l2arc_dev_mtx
);
9175 return (dev
!= NULL
);
9179 * Add a vdev for use by the L2ARC. By this point the spa has already
9180 * validated the vdev and opened it.
9183 l2arc_add_vdev(spa_t
*spa
, vdev_t
*vd
)
9185 l2arc_dev_t
*adddev
;
9187 ASSERT(!l2arc_vdev_present(vd
));
9190 * Create a new l2arc device entry.
9192 adddev
= kmem_zalloc(sizeof (l2arc_dev_t
), KM_SLEEP
);
9193 adddev
->l2ad_spa
= spa
;
9194 adddev
->l2ad_vdev
= vd
;
9195 adddev
->l2ad_start
= VDEV_LABEL_START_SIZE
;
9196 adddev
->l2ad_end
= VDEV_LABEL_START_SIZE
+ vdev_get_min_asize(vd
);
9197 adddev
->l2ad_hand
= adddev
->l2ad_start
;
9198 adddev
->l2ad_first
= B_TRUE
;
9199 adddev
->l2ad_writing
= B_FALSE
;
9200 list_link_init(&adddev
->l2ad_node
);
9202 mutex_init(&adddev
->l2ad_mtx
, NULL
, MUTEX_DEFAULT
, NULL
);
9204 * This is a list of all ARC buffers that are still valid on the
9207 list_create(&adddev
->l2ad_buflist
, sizeof (arc_buf_hdr_t
),
9208 offsetof(arc_buf_hdr_t
, b_l2hdr
.b_l2node
));
9210 vdev_space_update(vd
, 0, 0, adddev
->l2ad_end
- adddev
->l2ad_hand
);
9211 zfs_refcount_create(&adddev
->l2ad_alloc
);
9214 * Add device to global list
9216 mutex_enter(&l2arc_dev_mtx
);
9217 list_insert_head(l2arc_dev_list
, adddev
);
9218 atomic_inc_64(&l2arc_ndev
);
9219 mutex_exit(&l2arc_dev_mtx
);
9223 * Remove a vdev from the L2ARC.
9226 l2arc_remove_vdev(vdev_t
*vd
)
9228 l2arc_dev_t
*dev
, *nextdev
, *remdev
= NULL
;
9231 * Find the device by vdev
9233 mutex_enter(&l2arc_dev_mtx
);
9234 for (dev
= list_head(l2arc_dev_list
); dev
; dev
= nextdev
) {
9235 nextdev
= list_next(l2arc_dev_list
, dev
);
9236 if (vd
== dev
->l2ad_vdev
) {
9241 ASSERT3P(remdev
, !=, NULL
);
9244 * Remove device from global list
9246 list_remove(l2arc_dev_list
, remdev
);
9247 l2arc_dev_last
= NULL
; /* may have been invalidated */
9248 atomic_dec_64(&l2arc_ndev
);
9249 mutex_exit(&l2arc_dev_mtx
);
9252 * Clear all buflists and ARC references. L2ARC device flush.
9254 l2arc_evict(remdev
, 0, B_TRUE
);
9255 list_destroy(&remdev
->l2ad_buflist
);
9256 mutex_destroy(&remdev
->l2ad_mtx
);
9257 zfs_refcount_destroy(&remdev
->l2ad_alloc
);
9258 kmem_free(remdev
, sizeof (l2arc_dev_t
));
9264 l2arc_thread_exit
= 0;
9266 l2arc_writes_sent
= 0;
9267 l2arc_writes_done
= 0;
9269 mutex_init(&l2arc_feed_thr_lock
, NULL
, MUTEX_DEFAULT
, NULL
);
9270 cv_init(&l2arc_feed_thr_cv
, NULL
, CV_DEFAULT
, NULL
);
9271 mutex_init(&l2arc_dev_mtx
, NULL
, MUTEX_DEFAULT
, NULL
);
9272 mutex_init(&l2arc_free_on_write_mtx
, NULL
, MUTEX_DEFAULT
, NULL
);
9274 l2arc_dev_list
= &L2ARC_dev_list
;
9275 l2arc_free_on_write
= &L2ARC_free_on_write
;
9276 list_create(l2arc_dev_list
, sizeof (l2arc_dev_t
),
9277 offsetof(l2arc_dev_t
, l2ad_node
));
9278 list_create(l2arc_free_on_write
, sizeof (l2arc_data_free_t
),
9279 offsetof(l2arc_data_free_t
, l2df_list_node
));
9286 * This is called from dmu_fini(), which is called from spa_fini();
9287 * Because of this, we can assume that all l2arc devices have
9288 * already been removed when the pools themselves were removed.
9291 l2arc_do_free_on_write();
9293 mutex_destroy(&l2arc_feed_thr_lock
);
9294 cv_destroy(&l2arc_feed_thr_cv
);
9295 mutex_destroy(&l2arc_dev_mtx
);
9296 mutex_destroy(&l2arc_free_on_write_mtx
);
9298 list_destroy(l2arc_dev_list
);
9299 list_destroy(l2arc_free_on_write
);
9305 if (!(spa_mode_global
& FWRITE
))
9308 (void) thread_create(NULL
, 0, l2arc_feed_thread
, NULL
, 0, &p0
,
9309 TS_RUN
, defclsyspri
);
9315 if (!(spa_mode_global
& FWRITE
))
9318 mutex_enter(&l2arc_feed_thr_lock
);
9319 cv_signal(&l2arc_feed_thr_cv
); /* kick thread out of startup */
9320 l2arc_thread_exit
= 1;
9321 while (l2arc_thread_exit
!= 0)
9322 cv_wait(&l2arc_feed_thr_cv
, &l2arc_feed_thr_lock
);
9323 mutex_exit(&l2arc_feed_thr_lock
);
9326 #if defined(_KERNEL)
9327 EXPORT_SYMBOL(arc_buf_size
);
9328 EXPORT_SYMBOL(arc_write
);
9329 EXPORT_SYMBOL(arc_read
);
9330 EXPORT_SYMBOL(arc_buf_info
);
9331 EXPORT_SYMBOL(arc_getbuf_func
);
9332 EXPORT_SYMBOL(arc_add_prune_callback
);
9333 EXPORT_SYMBOL(arc_remove_prune_callback
);
9336 module_param(zfs_arc_min
, ulong
, 0644);
9337 MODULE_PARM_DESC(zfs_arc_min
, "Min arc size");
9339 module_param(zfs_arc_max
, ulong
, 0644);
9340 MODULE_PARM_DESC(zfs_arc_max
, "Max arc size");
9342 module_param(zfs_arc_meta_limit
, ulong
, 0644);
9343 MODULE_PARM_DESC(zfs_arc_meta_limit
, "Meta limit for arc size");
9345 module_param(zfs_arc_meta_limit_percent
, ulong
, 0644);
9346 MODULE_PARM_DESC(zfs_arc_meta_limit_percent
,
9347 "Percent of arc size for arc meta limit");
9349 module_param(zfs_arc_meta_min
, ulong
, 0644);
9350 MODULE_PARM_DESC(zfs_arc_meta_min
, "Min arc metadata");
9352 module_param(zfs_arc_meta_prune
, int, 0644);
9353 MODULE_PARM_DESC(zfs_arc_meta_prune
, "Meta objects to scan for prune");
9355 module_param(zfs_arc_meta_adjust_restarts
, int, 0644);
9356 MODULE_PARM_DESC(zfs_arc_meta_adjust_restarts
,
9357 "Limit number of restarts in arc_adjust_meta");
9359 module_param(zfs_arc_meta_strategy
, int, 0644);
9360 MODULE_PARM_DESC(zfs_arc_meta_strategy
, "Meta reclaim strategy");
9362 module_param(zfs_arc_grow_retry
, int, 0644);
9363 MODULE_PARM_DESC(zfs_arc_grow_retry
, "Seconds before growing arc size");
9365 module_param(zfs_arc_p_dampener_disable
, int, 0644);
9366 MODULE_PARM_DESC(zfs_arc_p_dampener_disable
, "disable arc_p adapt dampener");
9368 module_param(zfs_arc_shrink_shift
, int, 0644);
9369 MODULE_PARM_DESC(zfs_arc_shrink_shift
, "log2(fraction of arc to reclaim)");
9371 module_param(zfs_arc_pc_percent
, uint
, 0644);
9372 MODULE_PARM_DESC(zfs_arc_pc_percent
,
9373 "Percent of pagecache to reclaim arc to");
9375 module_param(zfs_arc_p_min_shift
, int, 0644);
9376 MODULE_PARM_DESC(zfs_arc_p_min_shift
, "arc_c shift to calc min/max arc_p");
9378 module_param(zfs_arc_average_blocksize
, int, 0444);
9379 MODULE_PARM_DESC(zfs_arc_average_blocksize
, "Target average block size");
9381 module_param(zfs_compressed_arc_enabled
, int, 0644);
9382 MODULE_PARM_DESC(zfs_compressed_arc_enabled
, "Disable compressed arc buffers");
9384 module_param(zfs_arc_min_prefetch_ms
, int, 0644);
9385 MODULE_PARM_DESC(zfs_arc_min_prefetch_ms
, "Min life of prefetch block in ms");
9387 module_param(zfs_arc_min_prescient_prefetch_ms
, int, 0644);
9388 MODULE_PARM_DESC(zfs_arc_min_prescient_prefetch_ms
,
9389 "Min life of prescient prefetched block in ms");
9391 module_param(l2arc_write_max
, ulong
, 0644);
9392 MODULE_PARM_DESC(l2arc_write_max
, "Max write bytes per interval");
9394 module_param(l2arc_write_boost
, ulong
, 0644);
9395 MODULE_PARM_DESC(l2arc_write_boost
, "Extra write bytes during device warmup");
9397 module_param(l2arc_headroom
, ulong
, 0644);
9398 MODULE_PARM_DESC(l2arc_headroom
, "Number of max device writes to precache");
9400 module_param(l2arc_headroom_boost
, ulong
, 0644);
9401 MODULE_PARM_DESC(l2arc_headroom_boost
, "Compressed l2arc_headroom multiplier");
9403 module_param(l2arc_feed_secs
, ulong
, 0644);
9404 MODULE_PARM_DESC(l2arc_feed_secs
, "Seconds between L2ARC writing");
9406 module_param(l2arc_feed_min_ms
, ulong
, 0644);
9407 MODULE_PARM_DESC(l2arc_feed_min_ms
, "Min feed interval in milliseconds");
9409 module_param(l2arc_noprefetch
, int, 0644);
9410 MODULE_PARM_DESC(l2arc_noprefetch
, "Skip caching prefetched buffers");
9412 module_param(l2arc_feed_again
, int, 0644);
9413 MODULE_PARM_DESC(l2arc_feed_again
, "Turbo L2ARC warmup");
9415 module_param(l2arc_norw
, int, 0644);
9416 MODULE_PARM_DESC(l2arc_norw
, "No reads during writes");
9418 module_param(zfs_arc_lotsfree_percent
, int, 0644);
9419 MODULE_PARM_DESC(zfs_arc_lotsfree_percent
,
9420 "System free memory I/O throttle in bytes");
9422 module_param(zfs_arc_sys_free
, ulong
, 0644);
9423 MODULE_PARM_DESC(zfs_arc_sys_free
, "System free memory target size in bytes");
9425 module_param(zfs_arc_dnode_limit
, ulong
, 0644);
9426 MODULE_PARM_DESC(zfs_arc_dnode_limit
, "Minimum bytes of dnodes in arc");
9428 module_param(zfs_arc_dnode_limit_percent
, ulong
, 0644);
9429 MODULE_PARM_DESC(zfs_arc_dnode_limit_percent
,
9430 "Percent of ARC meta buffers for dnodes");
9432 module_param(zfs_arc_dnode_reduce_percent
, ulong
, 0644);
9433 MODULE_PARM_DESC(zfs_arc_dnode_reduce_percent
,
9434 "Percentage of excess dnodes to try to unpin");