+/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _BCACHEFS_H
#define _BCACHEFS_H
*
* BTREE NODES:
*
- * Our unit of allocation is a bucket, and we we can't arbitrarily allocate and
+ * Our unit of allocation is a bucket, and we can't arbitrarily allocate and
* free smaller than a bucket - so, that's how big our btree nodes are.
*
* (If buckets are really big we'll only use part of the bucket for a btree node
*/
#undef pr_fmt
+#ifdef __KERNEL__
#define pr_fmt(fmt) "bcachefs: %s() " fmt "\n", __func__
+#else
+#define pr_fmt(fmt) "%s() " fmt "\n", __func__
+#endif
+#include <linux/backing-dev-defs.h>
#include <linux/bug.h>
#include <linux/bio.h>
#include <linux/closure.h>
#include <linux/kobject.h>
-#include <linux/lglock.h>
#include <linux/list.h>
+#include <linux/math64.h>
#include <linux/mutex.h>
#include <linux/percpu-refcount.h>
-#include <linux/radix-tree.h>
-#include <linux/rbtree.h>
+#include <linux/percpu-rwsem.h>
+#include <linux/refcount.h>
#include <linux/rhashtable.h>
#include <linux/rwsem.h>
+#include <linux/semaphore.h>
#include <linux/seqlock.h>
#include <linux/shrinker.h>
+#include <linux/srcu.h>
#include <linux/types.h>
#include <linux/workqueue.h>
+#include <linux/zstd.h>
#include "bcachefs_format.h"
-#include "bset.h"
+#include "errcode.h"
#include "fifo.h"
+#include "nocow_locking_types.h"
#include "opts.h"
+#include "recovery_types.h"
+#include "sb-errors_types.h"
+#include "seqmutex.h"
#include "util.h"
-#include <linux/dynamic_fault.h>
+#ifdef CONFIG_BCACHEFS_DEBUG
+#define BCH_WRITE_REF_DEBUG
+#endif
+
+#ifndef dynamic_fault
+#define dynamic_fault(...) 0
+#endif
+
+#define race_fault(...) dynamic_fault("bcachefs:race")
+
+#define count_event(_c, _name) this_cpu_inc((_c)->counters[BCH_COUNTER_##_name])
-#define bch2_fs_init_fault(name) \
+#define trace_and_count(_c, _name, ...) \
+do { \
+ count_event(_c, _name); \
+ trace_##_name(__VA_ARGS__); \
+} while (0)
+
+#define bch2_fs_init_fault(name) \
dynamic_fault("bcachefs:bch_fs_init:" name)
#define bch2_meta_read_fault(name) \
dynamic_fault("bcachefs:meta:read:" name)
dynamic_fault("bcachefs:meta:write:" name)
#ifdef __KERNEL__
-#define bch2_fmt(_c, fmt) "bcachefs (%s): " fmt "\n", ((_c)->name)
+#define BCACHEFS_LOG_PREFIX
+#endif
+
+#ifdef BCACHEFS_LOG_PREFIX
+
+#define bch2_log_msg(_c, fmt) "bcachefs (%s): " fmt, ((_c)->name)
+#define bch2_fmt_dev(_ca, fmt) "bcachefs (%s): " fmt "\n", ((_ca)->name)
+#define bch2_fmt_dev_offset(_ca, _offset, fmt) "bcachefs (%s sector %llu): " fmt "\n", ((_ca)->name), (_offset)
+#define bch2_fmt_inum(_c, _inum, fmt) "bcachefs (%s inum %llu): " fmt "\n", ((_c)->name), (_inum)
+#define bch2_fmt_inum_offset(_c, _inum, _offset, fmt) \
+ "bcachefs (%s inum %llu offset %llu): " fmt "\n", ((_c)->name), (_inum), (_offset)
+
#else
-#define bch2_fmt(_c, fmt) fmt "\n"
+
+#define bch2_log_msg(_c, fmt) fmt
+#define bch2_fmt_dev(_ca, fmt) "%s: " fmt "\n", ((_ca)->name)
+#define bch2_fmt_dev_offset(_ca, _offset, fmt) "%s sector %llu: " fmt "\n", ((_ca)->name), (_offset)
+#define bch2_fmt_inum(_c, _inum, fmt) "inum %llu: " fmt "\n", (_inum)
+#define bch2_fmt_inum_offset(_c, _inum, _offset, fmt) \
+ "inum %llu offset %llu: " fmt "\n", (_inum), (_offset)
+
#endif
+#define bch2_fmt(_c, fmt) bch2_log_msg(_c, fmt "\n")
+
+__printf(2, 3)
+void __bch2_print(struct bch_fs *c, const char *fmt, ...);
+
+#define maybe_dev_to_fs(_c) _Generic((_c), \
+ struct bch_dev *: ((struct bch_dev *) (_c))->fs, \
+ struct bch_fs *: (_c))
+
+#define bch2_print(_c, ...) __bch2_print(maybe_dev_to_fs(_c), __VA_ARGS__)
+
+#define bch2_print_ratelimited(_c, ...) \
+do { \
+ static DEFINE_RATELIMIT_STATE(_rs, \
+ DEFAULT_RATELIMIT_INTERVAL, \
+ DEFAULT_RATELIMIT_BURST); \
+ \
+ if (__ratelimit(&_rs)) \
+ bch2_print(_c, __VA_ARGS__); \
+} while (0)
+
#define bch_info(c, fmt, ...) \
- printk(KERN_INFO bch2_fmt(c, fmt), ##__VA_ARGS__)
+ bch2_print(c, KERN_INFO bch2_fmt(c, fmt), ##__VA_ARGS__)
#define bch_notice(c, fmt, ...) \
- printk(KERN_NOTICE bch2_fmt(c, fmt), ##__VA_ARGS__)
+ bch2_print(c, KERN_NOTICE bch2_fmt(c, fmt), ##__VA_ARGS__)
#define bch_warn(c, fmt, ...) \
- printk(KERN_WARNING bch2_fmt(c, fmt), ##__VA_ARGS__)
+ bch2_print(c, KERN_WARNING bch2_fmt(c, fmt), ##__VA_ARGS__)
+#define bch_warn_ratelimited(c, fmt, ...) \
+ bch2_print_ratelimited(c, KERN_WARNING bch2_fmt(c, fmt), ##__VA_ARGS__)
+
#define bch_err(c, fmt, ...) \
- printk(KERN_ERR bch2_fmt(c, fmt), ##__VA_ARGS__)
+ bch2_print(c, KERN_ERR bch2_fmt(c, fmt), ##__VA_ARGS__)
+#define bch_err_dev(ca, fmt, ...) \
+ bch2_print(c, KERN_ERR bch2_fmt_dev(ca, fmt), ##__VA_ARGS__)
+#define bch_err_dev_offset(ca, _offset, fmt, ...) \
+ bch2_print(c, KERN_ERR bch2_fmt_dev_offset(ca, _offset, fmt), ##__VA_ARGS__)
+#define bch_err_inum(c, _inum, fmt, ...) \
+ bch2_print(c, KERN_ERR bch2_fmt_inum(c, _inum, fmt), ##__VA_ARGS__)
+#define bch_err_inum_offset(c, _inum, _offset, fmt, ...) \
+ bch2_print(c, KERN_ERR bch2_fmt_inum_offset(c, _inum, _offset, fmt), ##__VA_ARGS__)
+
+#define bch_err_ratelimited(c, fmt, ...) \
+ bch2_print_ratelimited(c, KERN_ERR bch2_fmt(c, fmt), ##__VA_ARGS__)
+#define bch_err_dev_ratelimited(ca, fmt, ...) \
+ bch2_print_ratelimited(ca, KERN_ERR bch2_fmt_dev(ca, fmt), ##__VA_ARGS__)
+#define bch_err_dev_offset_ratelimited(ca, _offset, fmt, ...) \
+ bch2_print_ratelimited(ca, KERN_ERR bch2_fmt_dev_offset(ca, _offset, fmt), ##__VA_ARGS__)
+#define bch_err_inum_ratelimited(c, _inum, fmt, ...) \
+ bch2_print_ratelimited(c, KERN_ERR bch2_fmt_inum(c, _inum, fmt), ##__VA_ARGS__)
+#define bch_err_inum_offset_ratelimited(c, _inum, _offset, fmt, ...) \
+ bch2_print_ratelimited(c, KERN_ERR bch2_fmt_inum_offset(c, _inum, _offset, fmt), ##__VA_ARGS__)
+
+#define bch_err_fn(_c, _ret) \
+do { \
+ if (_ret && !bch2_err_matches(_ret, BCH_ERR_transaction_restart))\
+ bch_err(_c, "%s(): error %s", __func__, bch2_err_str(_ret));\
+} while (0)
+
+#define bch_err_msg(_c, _ret, _msg, ...) \
+do { \
+ if (_ret && !bch2_err_matches(_ret, BCH_ERR_transaction_restart))\
+ bch_err(_c, "%s(): error " _msg " %s", __func__, \
+ ##__VA_ARGS__, bch2_err_str(_ret)); \
+} while (0)
#define bch_verbose(c, fmt, ...) \
do { \
- if ((c)->opts.verbose_recovery) \
+ if ((c)->opts.verbose) \
bch_info(c, fmt, ##__VA_ARGS__); \
} while (0)
+#define pr_verbose_init(opts, fmt, ...) \
+do { \
+ if (opt_get(opts, verbose)) \
+ pr_info(fmt, ##__VA_ARGS__); \
+} while (0)
+
/* Parameters that are useful for debugging, but should always be compiled in: */
#define BCH_DEBUG_PARAMS_ALWAYS() \
BCH_DEBUG_PARAM(key_merging_disabled, \
"btree node it traverses") \
BCH_DEBUG_PARAM(btree_gc_rewrite_disabled, \
"Disables rewriting of btree nodes during mark and sweep")\
- BCH_DEBUG_PARAM(btree_gc_coalesce_disabled, \
- "Disables coalescing of btree nodes") \
BCH_DEBUG_PARAM(btree_shrinker_disabled, \
- "Disables the shrinker callback for the btree node cache")
-
-/* Parameters that should only be compiled in in debug mode: */
-#define BCH_DEBUG_PARAMS_DEBUG() \
- BCH_DEBUG_PARAM(expensive_debug_checks, \
- "Enables various runtime debugging checks that " \
- "significantly affect performance") \
- BCH_DEBUG_PARAM(debug_check_bkeys, \
- "Run bkey_debugcheck (primarily checking GC/allocation "\
- "information) when iterating over keys") \
- BCH_DEBUG_PARAM(version_stress_test, \
- "Assigns random version numbers to newly written " \
- "extents, to test overlapping extent cases") \
+ "Disables the shrinker callback for the btree node cache")\
BCH_DEBUG_PARAM(verify_btree_ondisk, \
"Reread btree nodes at various points to verify the " \
"mergesort in the read path against modifications " \
"done in memory") \
+ BCH_DEBUG_PARAM(verify_all_btree_replicas, \
+ "When reading btree nodes, read all replicas and " \
+ "compare them") \
+ BCH_DEBUG_PARAM(backpointers_no_use_write_buffer, \
+ "Don't use the write buffer for backpointers, enabling "\
+ "extra runtime checks")
+
+/* Parameters that should only be compiled in debug mode: */
+#define BCH_DEBUG_PARAMS_DEBUG() \
+ BCH_DEBUG_PARAM(expensive_debug_checks, \
+ "Enables various runtime debugging checks that " \
+ "significantly affect performance") \
+ BCH_DEBUG_PARAM(debug_check_iterators, \
+ "Enables extra verification for btree iterators") \
+ BCH_DEBUG_PARAM(debug_check_btree_accounting, \
+ "Verify btree accounting for keys within a node") \
+ BCH_DEBUG_PARAM(journal_seq_verify, \
+ "Store the journal sequence number in the version " \
+ "number of every btree key, and verify that btree " \
+ "update ordering is preserved during recovery") \
+ BCH_DEBUG_PARAM(inject_invalid_keys, \
+ "Store the journal sequence number in the version " \
+ "number of every btree key, and verify that btree " \
+ "update ordering is preserved during recovery") \
+ BCH_DEBUG_PARAM(test_alloc_startup, \
+ "Force allocator startup to use the slowpath where it" \
+ "can't find enough free buckets without invalidating" \
+ "cached data") \
+ BCH_DEBUG_PARAM(force_reconstruct_read, \
+ "Force reads to use the reconstruct path, when reading" \
+ "from erasure coded extents") \
+ BCH_DEBUG_PARAM(test_restart_gc, \
+ "Test restarting mark and sweep gc when bucket gens change")
#define BCH_DEBUG_PARAMS_ALL() BCH_DEBUG_PARAMS_ALWAYS() BCH_DEBUG_PARAMS_DEBUG()
#define BCH_DEBUG_PARAMS() BCH_DEBUG_PARAMS_ALWAYS()
#endif
-/* name, frequency_units, duration_units */
-#define BCH_TIME_STATS() \
- BCH_TIME_STAT(btree_node_mem_alloc, sec, us) \
- BCH_TIME_STAT(btree_gc, sec, ms) \
- BCH_TIME_STAT(btree_coalesce, sec, ms) \
- BCH_TIME_STAT(btree_split, sec, us) \
- BCH_TIME_STAT(btree_sort, ms, us) \
- BCH_TIME_STAT(btree_read, ms, us) \
- BCH_TIME_STAT(journal_write, us, us) \
- BCH_TIME_STAT(journal_delay, ms, us) \
- BCH_TIME_STAT(journal_blocked, sec, ms) \
- BCH_TIME_STAT(journal_flush_seq, us, us)
+#define BCH_DEBUG_PARAM(name, description) extern bool bch2_##name;
+BCH_DEBUG_PARAMS()
+#undef BCH_DEBUG_PARAM
+
+#ifndef CONFIG_BCACHEFS_DEBUG
+#define BCH_DEBUG_PARAM(name, description) static const __maybe_unused bool bch2_##name;
+BCH_DEBUG_PARAMS_DEBUG()
+#undef BCH_DEBUG_PARAM
+#endif
+
+#define BCH_TIME_STATS() \
+ x(btree_node_mem_alloc) \
+ x(btree_node_split) \
+ x(btree_node_compact) \
+ x(btree_node_merge) \
+ x(btree_node_sort) \
+ x(btree_node_read) \
+ x(btree_interior_update_foreground) \
+ x(btree_interior_update_total) \
+ x(btree_gc) \
+ x(data_write) \
+ x(data_read) \
+ x(data_promote) \
+ x(journal_flush_write) \
+ x(journal_noflush_write) \
+ x(journal_flush_seq) \
+ x(blocked_journal_low_on_space) \
+ x(blocked_journal_low_on_pin) \
+ x(blocked_journal_max_in_flight) \
+ x(blocked_allocate) \
+ x(blocked_allocate_open_bucket) \
+ x(blocked_write_buffer_full) \
+ x(nocow_lock_contended)
+
+enum bch_time_stats {
+#define x(name) BCH_TIME_##name,
+ BCH_TIME_STATS()
+#undef x
+ BCH_TIME_STAT_NR
+};
#include "alloc_types.h"
+#include "btree_types.h"
+#include "btree_write_buffer_types.h"
#include "buckets_types.h"
+#include "buckets_waiting_for_journal_types.h"
#include "clock_types.h"
-#include "io_types.h"
+#include "disk_groups_types.h"
+#include "ec_types.h"
#include "journal_types.h"
#include "keylist_types.h"
-#include "move_types.h"
+#include "quota_types.h"
+#include "rebalance_types.h"
+#include "replicas_types.h"
+#include "subvolume_types.h"
#include "super_types.h"
-/* 256k, in sectors */
-#define BTREE_NODE_SIZE_MAX 512
-
-/*
- * Number of nodes we might have to allocate in a worst case btree split
- * operation - we split all the way up to the root, then allocate a new root.
- */
-#define btree_reserve_required_nodes(depth) (((depth) + 1) * 2 + 1)
-
/* Number of nodes btree coalesce will try to coalesce at once */
#define GC_MERGE_NODES 4U
/* Maximum number of nodes we might need to allocate atomically: */
-#define BTREE_RESERVE_MAX \
- (btree_reserve_required_nodes(BTREE_MAX_DEPTH) + GC_MERGE_NODES)
+#define BTREE_RESERVE_MAX (BTREE_MAX_DEPTH + (BTREE_MAX_DEPTH - 1))
/* Size of the freelist we allocate btree nodes from: */
-#define BTREE_NODE_RESERVE (BTREE_RESERVE_MAX * 2)
+#define BTREE_NODE_RESERVE (BTREE_RESERVE_MAX * 4)
+
+#define BTREE_NODE_OPEN_BUCKET_RESERVE (BTREE_RESERVE_MAX * BCH_REPLICAS_MAX)
struct btree;
-struct crypto_blkcipher;
-struct crypto_ahash;
+
+struct log_output {
+ spinlock_t lock;
+ wait_queue_head_t wait;
+ struct printbuf buf;
+};
enum gc_phase {
- GC_PHASE_SB_METADATA = BTREE_ID_NR + 1,
+ GC_PHASE_NOT_RUNNING,
+ GC_PHASE_START,
+ GC_PHASE_SB,
+
+ GC_PHASE_BTREE_stripes,
+ GC_PHASE_BTREE_extents,
+ GC_PHASE_BTREE_inodes,
+ GC_PHASE_BTREE_dirents,
+ GC_PHASE_BTREE_xattrs,
+ GC_PHASE_BTREE_alloc,
+ GC_PHASE_BTREE_quotas,
+ GC_PHASE_BTREE_reflink,
+ GC_PHASE_BTREE_subvolumes,
+ GC_PHASE_BTREE_snapshots,
+ GC_PHASE_BTREE_lru,
+ GC_PHASE_BTREE_freespace,
+ GC_PHASE_BTREE_need_discard,
+ GC_PHASE_BTREE_backpointers,
+ GC_PHASE_BTREE_bucket_gens,
+ GC_PHASE_BTREE_snapshot_trees,
+ GC_PHASE_BTREE_deleted_inodes,
+ GC_PHASE_BTREE_logged_ops,
+ GC_PHASE_BTREE_rebalance_work,
+
GC_PHASE_PENDING_DELETE,
- GC_PHASE_DONE
};
struct gc_pos {
unsigned level;
};
-struct bch_member_cpu {
- u64 nbuckets; /* device size */
- u16 first_bucket; /* index of first bucket used */
- u16 bucket_size; /* sectors */
- u8 state;
- u8 tier;
- u8 has_metadata;
- u8 has_data;
- u8 replacement;
- u8 discard;
- u8 valid;
+struct reflink_gc {
+ u64 offset;
+ u32 size;
+ u32 refcount;
+};
+
+typedef GENRADIX(struct reflink_gc) reflink_gc_table;
+
+struct io_count {
+ u64 sectors[2][BCH_DATA_NR];
};
struct bch_dev {
struct kobject kobj;
struct percpu_ref ref;
+ struct completion ref_completion;
struct percpu_ref io_ref;
- struct completion stop_complete;
- struct completion offline_complete;
+ struct completion io_ref_completion;
struct bch_fs *fs;
* Committed by bch2_write_super() -> bch_fs_mi_update()
*/
struct bch_member_cpu mi;
- uuid_le uuid;
+ atomic64_t errors[BCH_MEMBER_ERROR_NR];
+
+ __uuid_t uuid;
char name[BDEVNAME_SIZE];
- struct bcache_superblock disk_sb;
+ struct bch_sb_handle disk_sb;
+ struct bch_sb *sb_read_scratch;
+ int sb_write_error;
+ dev_t dev;
+ atomic_t flush_seq;
- struct dev_group self;
+ struct bch_devs_mask self;
- /* biosets used in cloned bios for replicas and moving_gc */
+ /* biosets used in cloned bios for writing multiple replicas */
struct bio_set replica_set;
- struct task_struct *alloc_thread;
-
- struct prio_set *disk_buckets;
-
/*
- * When allocating new buckets, prio_write() gets first dibs - since we
- * may not be allocate at all without writing priorities and gens.
- * prio_last_buckets[] contains the last buckets we wrote priorities to
- * (so gc can mark them as metadata).
+ * Buckets:
+ * Per-bucket arrays are protected by c->mark_lock, bucket_lock and
+ * gc_lock, for device resize - holding any is sufficient for access:
+ * Or rcu_read_lock(), but only for ptr_stale():
*/
- u64 *prio_buckets;
- u64 *prio_last_buckets;
- spinlock_t prio_buckets_lock;
- struct bio *bio_prio;
- bool prio_read_done;
- bool need_prio_write;
- struct mutex prio_write_lock;
+ struct bucket_array __rcu *buckets_gc;
+ struct bucket_gens __rcu *bucket_gens;
+ u8 *oldest_gen;
+ unsigned long *buckets_nouse;
+ struct rw_semaphore bucket_lock;
- /*
- * free: Buckets that are ready to be used
- *
- * free_inc: Incoming buckets - these are buckets that currently have
- * cached data in them, and we can't reuse them until after we write
- * their new gen to disk. After prio_write() finishes writing the new
- * gens/prios, they'll be moved to the free list (and possibly discarded
- * in the process)
- */
- DECLARE_FIFO(long, free)[RESERVE_NR];
- DECLARE_FIFO(long, free_inc);
- spinlock_t freelist_lock;
+ struct bch_dev_usage *usage_base;
+ struct bch_dev_usage __percpu *usage[JOURNAL_BUF_NR];
+ struct bch_dev_usage __percpu *usage_gc;
- size_t fifo_last_bucket;
+ /* Allocator: */
+ u64 new_fs_bucket_idx;
+ u64 alloc_cursor;
- /* Allocation stuff: */
+ unsigned nr_open_buckets;
+ unsigned nr_btree_reserve;
- /* most out of date gen in the btree */
- u8 *oldest_gens;
- struct bucket *buckets;
- unsigned short bucket_bits; /* ilog2(bucket_size) */
-
- /* last calculated minimum prio */
- u16 min_prio[2];
-
- /*
- * Bucket book keeping. The first element is updated by GC, the
- * second contains a saved copy of the stats from the beginning
- * of GC.
- */
- struct bch_dev_usage __percpu *usage_percpu;
- struct bch_dev_usage usage_cached;
-
- atomic_long_t saturated_count;
size_t inc_gen_needs_gc;
+ size_t inc_gen_really_needs_gc;
+ size_t buckets_waiting_on_journal;
- bucket_heap alloc_heap;
- bucket_heap copygc_heap;
-
- /* Moving GC: */
- struct task_struct *moving_gc_read;
-
- struct bch_pd_controller moving_gc_pd;
-
- /* Tiering: */
- struct write_point tiering_write_point;
-
- struct write_point copygc_write_point;
+ atomic64_t rebalance_work;
struct journal_device journal;
+ u64 prev_journal_sector;
struct work_struct io_error_work;
/* The rest of this all shows up in sysfs */
- atomic64_t meta_sectors_written;
- atomic64_t btree_sectors_written;
- u64 __percpu *sectors_written;
+ atomic64_t cur_latency[2];
+ struct bch2_time_stats io_latency[2];
+
+#define CONGESTED_MAX 1024
+ atomic_t congested;
+ u64 congested_last;
+
+ struct io_count __percpu *io_done;
};
/*
- * Flag bits for what phase of startup/shutdown the cache set is at, how we're
- * shutting down, etc.:
+ * fsck_done - kill?
*
- * BCH_FS_UNREGISTERING means we're not just shutting down, we're detaching
- * all the backing devices first (their cached data gets invalidated, and they
- * won't automatically reattach).
+ * replace with something more general from enumated fsck passes/errors:
+ * initial_gc_unfixed
+ * error
+ * topology error
*/
-enum {
- BCH_FS_INITIAL_GC_DONE,
- BCH_FS_EMERGENCY_RO,
- BCH_FS_WRITE_DISABLE_COMPLETE,
- BCH_FS_GC_STOPPING,
- BCH_FS_GC_FAILURE,
- BCH_FS_BDEV_MOUNTED,
- BCH_FS_ERROR,
- BCH_FS_FSCK_FIXED_ERRORS,
- BCH_FS_FSCK_DONE,
- BCH_FS_FIXED_GENS,
+
+#define BCH_FS_FLAGS() \
+ x(started) \
+ x(may_go_rw) \
+ x(rw) \
+ x(was_rw) \
+ x(stopping) \
+ x(emergency_ro) \
+ x(going_ro) \
+ x(write_disable_complete) \
+ x(clean_shutdown) \
+ x(fsck_done) \
+ x(initial_gc_unfixed) \
+ x(need_another_gc) \
+ x(need_delete_dead_snapshots) \
+ x(error) \
+ x(topology_error) \
+ x(errors_fixed) \
+ x(errors_not_fixed)
+
+enum bch_fs_flags {
+#define x(n) BCH_FS_##n,
+ BCH_FS_FLAGS()
+#undef x
};
struct btree_debug {
unsigned id;
- struct dentry *btree;
- struct dentry *btree_format;
- struct dentry *failed;
};
-struct bch_tier {
- unsigned idx;
- struct task_struct *migrate;
- struct bch_pd_controller pd;
+#define BCH_TRANSACTIONS_NR 128
- struct dev_group devs;
+struct btree_transaction_stats {
+ struct bch2_time_stats lock_hold_times;
+ struct mutex lock;
+ unsigned nr_max_paths;
+ unsigned journal_entries_size;
+ unsigned max_mem;
+ char *max_paths_text;
};
-enum bch_fs_state {
- BCH_FS_STARTING = 0,
- BCH_FS_STOPPING,
- BCH_FS_RO,
- BCH_FS_RW,
+struct bch_fs_pcpu {
+ u64 sectors_available;
+};
+
+struct journal_seq_blacklist_table {
+ size_t nr;
+ struct journal_seq_blacklist_table_entry {
+ u64 start;
+ u64 end;
+ bool dirty;
+ } entries[];
+};
+
+struct journal_keys {
+ struct journal_key {
+ u64 journal_seq;
+ u32 journal_offset;
+ enum btree_id btree_id:8;
+ unsigned level:8;
+ bool allocated;
+ bool overwritten;
+ struct bkey_i *k;
+ } *d;
+ /*
+ * Gap buffer: instead of all the empty space in the array being at the
+ * end of the buffer - from @nr to @size - the empty space is at @gap.
+ * This means that sequential insertions are O(n) instead of O(n^2).
+ */
+ size_t gap;
+ size_t nr;
+ size_t size;
+ atomic_t ref;
+ bool initial_ref_held;
+};
+
+struct btree_trans_buf {
+ struct btree_trans *trans;
+};
+
+#define REPLICAS_DELTA_LIST_MAX (1U << 16)
+
+#define BCACHEFS_ROOT_SUBVOL_INUM \
+ ((subvol_inum) { BCACHEFS_ROOT_SUBVOL, BCACHEFS_ROOT_INO })
+
+#define BCH_WRITE_REFS() \
+ x(trans) \
+ x(write) \
+ x(promote) \
+ x(node_rewrite) \
+ x(stripe_create) \
+ x(stripe_delete) \
+ x(reflink) \
+ x(fallocate) \
+ x(discard) \
+ x(invalidate) \
+ x(delete_dead_snapshots) \
+ x(snapshot_delete_pagecache) \
+ x(sysfs) \
+ x(btree_write_buffer)
+
+enum bch_write_ref {
+#define x(n) BCH_WRITE_REF_##n,
+ BCH_WRITE_REFS()
+#undef x
+ BCH_WRITE_REF_NR,
};
struct bch_fs {
struct list_head list;
struct kobject kobj;
+ struct kobject counters_kobj;
struct kobject internal;
struct kobject opts_dir;
struct kobject time_stats;
int minor;
struct device *chardev;
struct super_block *vfs_sb;
+ dev_t dev;
char name[40];
+ struct log_output *output;
+ struct task_struct *output_filter;
- /* ro/rw, add/remove devices: */
- struct mutex state_lock;
- enum bch_fs_state state;
+ /* ro/rw, add/remove/resize devices: */
+ struct rw_semaphore state_lock;
/* Counts outstanding writes, for clean transition to read-only */
+#ifdef BCH_WRITE_REF_DEBUG
+ atomic_long_t writes[BCH_WRITE_REF_NR];
+#else
struct percpu_ref writes;
+#endif
+ /*
+ * Analagous to c->writes, for asynchronous ops that don't necessarily
+ * need fs to be read-write
+ */
+ refcount_t ro_ref;
+ wait_queue_head_t ro_ref_wait;
+
struct work_struct read_only_work;
struct bch_dev __rcu *devs[BCH_SB_MEMBERS_MAX];
+ struct bch_replicas_cpu replicas;
+ struct bch_replicas_cpu replicas_gc;
+ struct mutex replicas_gc_lock;
+ mempool_t replicas_delta_pool;
+
+ struct journal_entry_res btree_root_journal_res;
+ struct journal_entry_res replicas_journal_res;
+ struct journal_entry_res clock_journal_res;
+ struct journal_entry_res dev_usage_journal_res;
+
+ struct bch_disk_groups_cpu __rcu *disk_groups;
+
struct bch_opts opts;
/* Updated by bch2_sb_update():*/
struct {
- uuid_le uuid;
- uuid_le user_uuid;
+ __uuid_t uuid;
+ __uuid_t user_uuid;
- u16 block_size;
- u16 btree_node_size;
+ u16 version;
+ u16 version_min;
+ u16 version_upgrade_complete;
u8 nr_devices;
u8 clean;
- u8 meta_replicas_have;
- u8 data_replicas_have;
-
- u8 str_hash_type;
u8 encryption_type;
u64 time_base_lo;
u32 time_base_hi;
- u32 time_precision;
+ unsigned time_units_per_sec;
+ unsigned nsec_per_time_unit;
+ u64 features;
+ u64 compat;
} sb;
- struct bch_sb *disk_sb;
- unsigned disk_sb_order;
+
+ struct bch_sb_handle disk_sb;
unsigned short block_bits; /* ilog2(block_size) */
+ u16 btree_foreground_merge_threshold;
+
struct closure sb_write;
struct mutex sb_lock;
- struct backing_dev_info bdi;
+ /* snapshot.c: */
+ struct snapshot_table __rcu *snapshots;
+ size_t snapshot_table_size;
+ struct mutex snapshot_table_lock;
+ struct rw_semaphore snapshot_create_lock;
+
+ struct work_struct snapshot_delete_work;
+ struct work_struct snapshot_wait_for_pagecache_and_delete_work;
+ snapshot_id_list snapshots_unlinked;
+ struct mutex snapshots_unlinked_lock;
/* BTREE CACHE */
- struct bio_set btree_read_bio;
+ struct bio_set btree_bio;
+ struct workqueue_struct *io_complete_wq;
- struct btree_root btree_roots[BTREE_ID_NR];
+ struct btree_root btree_roots_known[BTREE_ID_NR];
+ DARRAY(struct btree_root) btree_roots_extra;
struct mutex btree_root_lock;
- bool btree_cache_table_init_done;
- struct rhashtable btree_cache_table;
-
- /*
- * We never free a struct btree, except on shutdown - we just put it on
- * the btree_cache_freed list and reuse it later. This simplifies the
- * code, and it doesn't cost us much memory as the memory usage is
- * dominated by buffers that hold the actual btree node data and those
- * can be freed - and the number of struct btrees allocated is
- * effectively bounded.
- *
- * btree_cache_freeable effectively is a small cache - we use it because
- * high order page allocations can be rather expensive, and it's quite
- * common to delete and allocate btree nodes in quick succession. It
- * should never grow past ~2-3 nodes in practice.
- */
- struct mutex btree_cache_lock;
- struct list_head btree_cache;
- struct list_head btree_cache_freeable;
- struct list_head btree_cache_freed;
-
- /* Number of elements in btree_cache + btree_cache_freeable lists */
- unsigned btree_cache_used;
- unsigned btree_cache_reserve;
- struct shrinker btree_cache_shrink;
-
- /*
- * If we need to allocate memory for a new btree node and that
- * allocation fails, we can cannibalize another node in the btree cache
- * to satisfy the allocation - lock to guarantee only one thread does
- * this at a time:
- */
- struct closure_waitlist mca_wait;
- struct task_struct *btree_cache_alloc_lock;
-
- mempool_t btree_reserve_pool;
+ struct btree_cache btree_cache;
/*
* Cache of allocated btree nodes - if we allocate a btree node and
* when allocating btree reserves fail halfway through) - instead, we
* can stick them here:
*/
- struct btree_alloc {
- struct open_bucket *ob;
- BKEY_PADDED(k);
- } btree_reserve_cache[BTREE_NODE_RESERVE * 2];
+ struct btree_alloc btree_reserve_cache[BTREE_NODE_RESERVE * 2];
unsigned btree_reserve_cache_nr;
struct mutex btree_reserve_cache_lock;
mempool_t btree_interior_update_pool;
struct list_head btree_interior_update_list;
+ struct list_head btree_interior_updates_unwritten;
struct mutex btree_interior_update_lock;
+ struct closure_waitlist btree_interior_update_wait;
- struct workqueue_struct *wq;
- /* copygc needs its own workqueue for index updates.. */
- struct workqueue_struct *copygc_wq;
+ struct workqueue_struct *btree_interior_update_worker;
+ struct work_struct btree_interior_update_work;
- /* ALLOCATION */
- struct bch_pd_controller foreground_write_pd;
- struct delayed_work pd_controllers_update;
- unsigned pd_controllers_update_seconds;
- spinlock_t foreground_write_pd_lock;
- struct bch_write_op *write_wait_head;
- struct bch_write_op *write_wait_tail;
+ struct list_head pending_node_rewrites;
+ struct mutex pending_node_rewrites_lock;
+
+ /* btree_io.c: */
+ spinlock_t btree_write_error_lock;
+ struct btree_write_stats {
+ atomic64_t nr;
+ atomic64_t bytes;
+ } btree_write_stats[BTREE_WRITE_TYPE_NR];
+
+ /* btree_iter.c: */
+ struct seqmutex btree_trans_lock;
+ struct list_head btree_trans_list;
+ mempool_t btree_trans_pool;
+ mempool_t btree_trans_mem_pool;
+ struct btree_trans_buf __percpu *btree_trans_bufs;
+
+ struct srcu_struct btree_trans_barrier;
+ bool btree_trans_barrier_initialized;
- struct timer_list foreground_write_wakeup;
+ struct btree_key_cache btree_key_cache;
+ unsigned btree_key_cache_btrees;
+ struct btree_write_buffer btree_write_buffer;
+
+ struct workqueue_struct *btree_update_wq;
+ struct workqueue_struct *btree_io_complete_wq;
+ /* copygc needs its own workqueue for index updates.. */
+ struct workqueue_struct *copygc_wq;
/*
- * These contain all r/w devices - i.e. devices we can currently
- * allocate from:
+ * Use a dedicated wq for write ref holder tasks. Required to avoid
+ * dependency problems with other wq tasks that can block on ref
+ * draining, such as read-only transition.
*/
- struct dev_group all_devs;
- struct bch_tier tiers[BCH_TIER_MAX];
- /* NULL if we only have devices in one tier: */
- struct bch_tier *fastest_tier;
+ struct workqueue_struct *write_ref_wq;
+
+ /* ALLOCATION */
+ struct bch_devs_mask rw_devs[BCH_DATA_NR];
u64 capacity; /* sectors */
* and forces them to be revalidated
*/
u32 capacity_gen;
+ unsigned bucket_size_max;
atomic64_t sectors_available;
+ struct mutex sectors_available_lock;
- struct bch_fs_usage __percpu *usage_percpu;
- struct bch_fs_usage usage_cached;
- struct lglock usage_lock;
+ struct bch_fs_pcpu __percpu *pcpu;
- struct mutex bucket_lock;
+ struct percpu_rw_semaphore mark_lock;
- struct closure_waitlist freelist_wait;
+ seqcount_t usage_lock;
+ struct bch_fs_usage *usage_base;
+ struct bch_fs_usage __percpu *usage[JOURNAL_BUF_NR];
+ struct bch_fs_usage __percpu *usage_gc;
+ u64 __percpu *online_reserved;
- /*
- * When we invalidate buckets, we use both the priority and the amount
- * of good data to determine which buckets to reuse first - to weight
- * those together consistently we keep track of the smallest nonzero
- * priority of any bucket.
- */
- struct prio_clock prio_clock[2];
+ /* single element mempool: */
+ struct mutex usage_scratch_lock;
+ struct bch_fs_usage_online *usage_scratch;
struct io_clock io_clock[2];
- /* SECTOR ALLOCATOR */
- struct list_head open_buckets_open;
- struct list_head open_buckets_free;
- unsigned open_buckets_nr_free;
+ /* JOURNAL SEQ BLACKLIST */
+ struct journal_seq_blacklist_table *
+ journal_seq_blacklist_table;
+ struct work_struct journal_seq_blacklist_gc_work;
+
+ /* ALLOCATOR */
+ spinlock_t freelist_lock;
+ struct closure_waitlist freelist_wait;
+ u64 blocked_allocate;
+ u64 blocked_allocate_open_bucket;
+
+ open_bucket_idx_t open_buckets_freelist;
+ open_bucket_idx_t open_buckets_nr_free;
struct closure_waitlist open_buckets_wait;
- spinlock_t open_buckets_lock;
struct open_bucket open_buckets[OPEN_BUCKETS_COUNT];
+ open_bucket_idx_t open_buckets_hash[OPEN_BUCKETS_COUNT];
+
+ open_bucket_idx_t open_buckets_partial[OPEN_BUCKETS_COUNT];
+ open_bucket_idx_t open_buckets_partial_nr;
struct write_point btree_write_point;
+ struct write_point rebalance_write_point;
- struct write_point write_points[WRITE_POINT_COUNT];
- struct write_point promote_write_point;
+ struct write_point write_points[WRITE_POINT_MAX];
+ struct hlist_head write_points_hash[WRITE_POINT_HASH_NR];
+ struct mutex write_points_hash_lock;
+ unsigned write_points_nr;
- /*
- * This write point is used for migrating data off a device
- * and can point to any other device.
- * We can't use the normal write points because those will
- * gang up n replicas, and for migration we want only one new
- * replica.
- */
- struct write_point migration_write_point;
+ struct buckets_waiting_for_journal buckets_waiting_for_journal;
+ struct work_struct discard_work;
+ struct work_struct invalidate_work;
/* GARBAGE COLLECTION */
struct task_struct *gc_thread;
atomic_t kick_gc;
+ unsigned long gc_count;
+
+ enum btree_id gc_gens_btree;
+ struct bpos gc_gens_pos;
/*
* Tracks GC's progress - everything in the range [ZERO_KEY..gc_cur_pos]
* has been marked by GC.
*
- * gc_cur_phase is a superset of btree_ids (BTREE_ID_EXTENTS etc.)
- *
- * gc_cur_phase == GC_PHASE_DONE indicates that gc is finished/not
- * currently running, and gc marks are currently valid
+ * gc_cur_phase is a superset of btree_ids (BTREE_ID_extents etc.)
*
* Protected by gc_pos_lock. Only written to by GC thread, so GC thread
* can read without a lock.
* it's not while a gc is in progress.
*/
struct rw_semaphore gc_lock;
+ struct mutex gc_gens_lock;
/* IO PATH */
+ struct semaphore io_in_flight;
struct bio_set bio_read;
struct bio_set bio_read_split;
struct bio_set bio_write;
struct mutex bio_bounce_pages_lock;
mempool_t bio_bounce_pages;
+ struct bucket_nocow_lock_table
+ nocow_locks;
+ struct rhashtable promote_table;
- mempool_t lz4_workspace_pool;
- void *zlib_workspace;
- struct mutex zlib_workspace_lock;
mempool_t compression_bounce[2];
+ mempool_t compress_workspace[BCH_COMPRESSION_TYPE_NR];
+ mempool_t decompress_workspace;
+ size_t zstd_workspace_size;
struct crypto_shash *sha256;
- struct crypto_blkcipher *chacha20;
+ struct crypto_sync_skcipher *chacha20;
struct crypto_shash *poly1305;
atomic64_t key_version;
- struct bio_list read_retry_list;
- struct work_struct read_retry_work;
- spinlock_t read_retry_lock;
+ mempool_t large_bkey_pool;
- /* ERRORS */
- struct list_head fsck_errors;
- struct mutex fsck_error_lock;
- bool fsck_alloc_err;
+ /* MOVE.C */
+ struct list_head moving_context_list;
+ struct mutex moving_context_lock;
+
+ /* REBALANCE */
+ struct bch_fs_rebalance rebalance;
+
+ /* COPYGC */
+ struct task_struct *copygc_thread;
+ struct write_point copygc_write_point;
+ s64 copygc_wait_at;
+ s64 copygc_wait;
+ bool copygc_running;
+ wait_queue_head_t copygc_running_wq;
+
+ /* STRIPES: */
+ GENRADIX(struct stripe) stripes;
+ GENRADIX(struct gc_stripe) gc_stripes;
+
+ struct hlist_head ec_stripes_new[32];
+ spinlock_t ec_stripes_new_lock;
+
+ ec_stripes_heap ec_stripes_heap;
+ struct mutex ec_stripes_heap_lock;
+
+ /* ERASURE CODING */
+ struct list_head ec_stripe_head_list;
+ struct mutex ec_stripe_head_lock;
+
+ struct list_head ec_stripe_new_list;
+ struct mutex ec_stripe_new_lock;
+ wait_queue_head_t ec_stripe_new_wait;
+
+ struct work_struct ec_stripe_create_work;
+ u64 ec_stripe_hint;
+
+ struct work_struct ec_stripe_delete_work;
+
+ struct bio_set ec_bioset;
- /* FILESYSTEM */
- wait_queue_head_t writeback_wait;
- atomic_t writeback_pages;
- unsigned writeback_pages_max;
- atomic_long_t nr_inodes;
+ /* REFLINK */
+ reflink_gc_table reflink_gc_table;
+ size_t reflink_gc_nr;
+
+ /* fs.c */
+ struct list_head vfs_inodes_list;
+ struct mutex vfs_inodes_lock;
+
+ /* VFS IO PATH - fs-io.c */
+ struct bio_set writepage_bioset;
+ struct bio_set dio_write_bioset;
+ struct bio_set dio_read_bioset;
+ struct bio_set nocow_flush_bioset;
+
+ /* QUOTAS */
+ struct bch_memquota_type quotas[QTYP_NR];
+
+ /* RECOVERY */
+ u64 journal_replay_seq_start;
+ u64 journal_replay_seq_end;
+ /*
+ * Two different uses:
+ * "Has this fsck pass?" - i.e. should this type of error be an
+ * emergency read-only
+ * And, in certain situations fsck will rewind to an earlier pass: used
+ * for signaling to the toplevel code which pass we want to run now.
+ */
+ enum bch_recovery_pass curr_recovery_pass;
+ /* bitmap of explicitly enabled recovery passes: */
+ u64 recovery_passes_explicit;
+ /* bitmask of recovery passes that we actually ran */
+ u64 recovery_passes_complete;
+ /* never rewinds version of curr_recovery_pass */
+ enum bch_recovery_pass recovery_pass_done;
+ struct semaphore online_fsck_mutex;
/* DEBUG JUNK */
- struct dentry *debug;
+ struct dentry *fs_debug_dir;
+ struct dentry *btree_debug_dir;
struct btree_debug btree_debug[BTREE_ID_NR];
-#ifdef CONFIG_BCACHEFS_DEBUG
struct btree *verify_data;
struct btree_node *verify_ondisk;
struct mutex verify_lock;
-#endif
- u64 unused_inode_hint;
+ u64 *unused_inode_hints;
+ unsigned inode_shard_bits;
/*
* A btree node on disk could have too many bsets for an iterator to fit
mempool_t btree_bounce_pool;
struct journal journal;
+ GENRADIX(struct journal_replay *) journal_entries;
+ u64 journal_entries_base_seq;
+ struct journal_keys journal_keys;
+ struct list_head journal_iters;
- unsigned bucket_journal_seq;
+ u64 last_bucket_seq_cleanup;
- /* The rest of this all shows up in sysfs */
- atomic_long_t read_realloc_races;
+ u64 counters_on_mount[BCH_COUNTER_NR];
+ u64 __percpu *counters;
- unsigned foreground_write_ratelimit_enabled:1;
+ unsigned btree_gc_periodic:1;
unsigned copy_gc_enabled:1;
- unsigned tiering_enabled:1;
- unsigned tiering_percent;
+ bool promote_whole_extents;
- /*
- * foreground writes will be throttled when the number of free
- * buckets is below this percentage
- */
- unsigned foreground_target_percent;
+ struct bch2_time_stats times[BCH_TIME_STAT_NR];
-#define BCH_DEBUG_PARAM(name, description) bool name;
- BCH_DEBUG_PARAMS_ALL()
-#undef BCH_DEBUG_PARAM
+ struct btree_transaction_stats btree_transaction_stats[BCH_TRANSACTIONS_NR];
-#define BCH_TIME_STAT(name, frequency_units, duration_units) \
- struct time_stats name##_time;
- BCH_TIME_STATS()
-#undef BCH_TIME_STAT
+ /* ERRORS */
+ struct list_head fsck_error_msgs;
+ struct mutex fsck_error_msgs_lock;
+ bool fsck_alloc_msgs_err;
+
+ bch_sb_errors_cpu fsck_error_counts;
+ struct mutex fsck_error_counts_lock;
};
-static inline bool bch2_fs_running(struct bch_fs *c)
+extern struct wait_queue_head bch2_read_only_wait;
+
+static inline void bch2_write_ref_get(struct bch_fs *c, enum bch_write_ref ref)
+{
+#ifdef BCH_WRITE_REF_DEBUG
+ atomic_long_inc(&c->writes[ref]);
+#else
+ percpu_ref_get(&c->writes);
+#endif
+}
+
+static inline bool __bch2_write_ref_tryget(struct bch_fs *c, enum bch_write_ref ref)
{
- return c->state == BCH_FS_RO || c->state == BCH_FS_RW;
+#ifdef BCH_WRITE_REF_DEBUG
+ return !test_bit(BCH_FS_going_ro, &c->flags) &&
+ atomic_long_inc_not_zero(&c->writes[ref]);
+#else
+ return percpu_ref_tryget(&c->writes);
+#endif
}
-static inline unsigned bucket_pages(const struct bch_dev *ca)
+static inline bool bch2_write_ref_tryget(struct bch_fs *c, enum bch_write_ref ref)
{
- return ca->mi.bucket_size / PAGE_SECTORS;
+#ifdef BCH_WRITE_REF_DEBUG
+ return !test_bit(BCH_FS_going_ro, &c->flags) &&
+ atomic_long_inc_not_zero(&c->writes[ref]);
+#else
+ return percpu_ref_tryget_live(&c->writes);
+#endif
+}
+
+static inline void bch2_write_ref_put(struct bch_fs *c, enum bch_write_ref ref)
+{
+#ifdef BCH_WRITE_REF_DEBUG
+ long v = atomic_long_dec_return(&c->writes[ref]);
+
+ BUG_ON(v < 0);
+ if (v)
+ return;
+ for (unsigned i = 0; i < BCH_WRITE_REF_NR; i++)
+ if (atomic_long_read(&c->writes[i]))
+ return;
+
+ set_bit(BCH_FS_write_disable_complete, &c->flags);
+ wake_up(&bch2_read_only_wait);
+#else
+ percpu_ref_put(&c->writes);
+#endif
+}
+
+static inline bool bch2_ro_ref_tryget(struct bch_fs *c)
+{
+ if (test_bit(BCH_FS_stopping, &c->flags))
+ return false;
+
+ return refcount_inc_not_zero(&c->ro_ref);
+}
+
+static inline void bch2_ro_ref_put(struct bch_fs *c)
+{
+ if (refcount_dec_and_test(&c->ro_ref))
+ wake_up(&c->ro_ref_wait);
+}
+
+static inline void bch2_set_ra_pages(struct bch_fs *c, unsigned ra_pages)
+{
+#ifndef NO_BCACHEFS_FS
+ if (c->vfs_sb)
+ c->vfs_sb->s_bdi->ra_pages = ra_pages;
+#endif
}
static inline unsigned bucket_bytes(const struct bch_dev *ca)
static inline unsigned block_bytes(const struct bch_fs *c)
{
- return c->sb.block_size << 9;
+ return c->opts.block_size;
+}
+
+static inline unsigned block_sectors(const struct bch_fs *c)
+{
+ return c->opts.block_size >> 9;
+}
+
+static inline size_t btree_sectors(const struct bch_fs *c)
+{
+ return c->opts.btree_node_size >> 9;
+}
+
+static inline bool btree_id_cached(const struct bch_fs *c, enum btree_id btree)
+{
+ return c->btree_key_cache_btrees & (1U << btree);
+}
+
+static inline struct timespec64 bch2_time_to_timespec(const struct bch_fs *c, s64 time)
+{
+ struct timespec64 t;
+ s32 rem;
+
+ time += c->sb.time_base_lo;
+
+ t.tv_sec = div_s64_rem(time, c->sb.time_units_per_sec, &rem);
+ t.tv_nsec = rem * c->sb.nsec_per_time_unit;
+ return t;
}
+static inline s64 timespec_to_bch2_time(const struct bch_fs *c, struct timespec64 ts)
+{
+ return (ts.tv_sec * c->sb.time_units_per_sec +
+ (int) ts.tv_nsec / c->sb.nsec_per_time_unit) - c->sb.time_base_lo;
+}
+
+static inline s64 bch2_current_time(const struct bch_fs *c)
+{
+ struct timespec64 now;
+
+ ktime_get_coarse_real_ts64(&now);
+ return timespec_to_bch2_time(c, now);
+}
+
+static inline bool bch2_dev_exists2(const struct bch_fs *c, unsigned dev)
+{
+ return dev < c->sb.nr_devices && c->devs[dev];
+}
+
+#define BKEY_PADDED_ONSTACK(key, pad) \
+ struct { struct bkey_i key; __u64 key ## _pad[pad]; }
+
#endif /* _BCACHEFS_H */