linux/fs/bcachefs/errcode.h
Kent Overstreet 33bd5d0686 bcachefs: Deadlock cycle detector
We've outgrown our own deadlock avoidance strategy.

The btree iterator API provides an interface where the user doesn't need
to concern themselves with lock ordering - different btree iterators can
be traversed in any order. Without special care, this will lead to
deadlocks.

Our previous strategy was to define a lock ordering internally, and
whenever we attempt to take a lock and trylock() fails, we'd check if
the current btree transaction is holding any locks that cause a lock
ordering violation. If so, we'd issue a transaction restart, and then
bch2_trans_begin() would re-traverse all previously used iterators, but
in the correct order.

That approach had some issues, though.
 - Sometimes we'd issue transaction restarts unnecessarily, when no
   deadlock would have actually occured. Lock ordering restarts have
   become our primary cause of transaction restarts, on some workloads
   totally 20% of actual transaction commits.

 - To avoid deadlock or livelock, we'd often have to take intent locks
   when we only wanted a read lock: with the lock ordering approach, it
   is actually illegal to hold _any_ read lock while blocking on an intent
   lock, and this has been causing us unnecessary lock contention.

 - It was getting fragile - the various lock ordering rules are not
   trivial, and we'd been seeing occasional livelock issues related to
   this machinery.

So, since bcachefs is already a relational database masquerading as a
filesystem, we're stealing the next traditional database technique and
switching to a cycle detector for avoiding deadlocks.

When we block taking a btree lock, after adding ourself to the waitlist
but before sleeping, we do a DFS of btree transactions waiting on other
btree transactions, starting with the current transaction and walking
our held locks, and transactions blocking on our held locks.

If we find a cycle, we emit a transaction restart. Occasionally (e.g.
the btree split path) we can not allow the lock() operation to fail, so
if necessary we'll tell another transaction that it has to fail.

Result: trans_restart_would_deadlock events are reduced by a factor of
10 to 100, and we'll be able to delete a whole bunch of grotty, fragile
code.

Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
2023-10-22 17:09:41 -04:00

97 lines
3.7 KiB
C

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _BCACHEFS_ERRCODE_H
#define _BCACHEFS_ERRCODE_H
#define BCH_ERRCODES() \
x(ENOSPC, ENOSPC_disk_reservation) \
x(ENOSPC, ENOSPC_bucket_alloc) \
x(ENOSPC, ENOSPC_disk_label_add) \
x(ENOSPC, ENOSPC_stripe_create) \
x(ENOSPC, ENOSPC_stripe_reuse) \
x(ENOSPC, ENOSPC_inode_create) \
x(ENOSPC, ENOSPC_str_hash_create) \
x(ENOSPC, ENOSPC_snapshot_create) \
x(ENOSPC, ENOSPC_subvolume_create) \
x(ENOSPC, ENOSPC_sb) \
x(ENOSPC, ENOSPC_sb_journal) \
x(ENOSPC, ENOSPC_sb_quota) \
x(ENOSPC, ENOSPC_sb_replicas) \
x(ENOSPC, ENOSPC_sb_members) \
x(0, open_buckets_empty) \
x(0, freelist_empty) \
x(BCH_ERR_freelist_empty, no_buckets_found) \
x(0, insufficient_devices) \
x(0, transaction_restart) \
x(BCH_ERR_transaction_restart, transaction_restart_fault_inject) \
x(BCH_ERR_transaction_restart, transaction_restart_relock) \
x(BCH_ERR_transaction_restart, transaction_restart_relock_path) \
x(BCH_ERR_transaction_restart, transaction_restart_relock_path_intent) \
x(BCH_ERR_transaction_restart, transaction_restart_relock_after_fill) \
x(BCH_ERR_transaction_restart, transaction_restart_too_many_iters) \
x(BCH_ERR_transaction_restart, transaction_restart_lock_node_reused) \
x(BCH_ERR_transaction_restart, transaction_restart_fill_relock) \
x(BCH_ERR_transaction_restart, transaction_restart_fill_mem_alloc_fail)\
x(BCH_ERR_transaction_restart, transaction_restart_mem_realloced) \
x(BCH_ERR_transaction_restart, transaction_restart_in_traverse_all) \
x(BCH_ERR_transaction_restart, transaction_restart_would_deadlock) \
x(BCH_ERR_transaction_restart, transaction_restart_would_deadlock_write)\
x(BCH_ERR_transaction_restart, transaction_restart_deadlock_recursion_limit)\
x(BCH_ERR_transaction_restart, transaction_restart_upgrade) \
x(BCH_ERR_transaction_restart, transaction_restart_key_cache_upgrade) \
x(BCH_ERR_transaction_restart, transaction_restart_key_cache_fill) \
x(BCH_ERR_transaction_restart, transaction_restart_key_cache_raced) \
x(BCH_ERR_transaction_restart, transaction_restart_key_cache_realloced)\
x(BCH_ERR_transaction_restart, transaction_restart_journal_preres_get) \
x(BCH_ERR_transaction_restart, transaction_restart_nested) \
x(0, no_btree_node) \
x(BCH_ERR_no_btree_node, no_btree_node_relock) \
x(BCH_ERR_no_btree_node, no_btree_node_upgrade) \
x(BCH_ERR_no_btree_node, no_btree_node_drop) \
x(BCH_ERR_no_btree_node, no_btree_node_lock_root) \
x(BCH_ERR_no_btree_node, no_btree_node_up) \
x(BCH_ERR_no_btree_node, no_btree_node_down) \
x(BCH_ERR_no_btree_node, no_btree_node_init) \
x(BCH_ERR_no_btree_node, no_btree_node_cached) \
x(0, lock_fail_node_reused) \
x(0, lock_fail_root_changed) \
x(0, journal_reclaim_would_deadlock) \
x(0, fsck) \
x(BCH_ERR_fsck, fsck_fix) \
x(BCH_ERR_fsck, fsck_ignore) \
x(BCH_ERR_fsck, fsck_errors_not_fixed) \
x(BCH_ERR_fsck, fsck_repair_unimplemented) \
x(BCH_ERR_fsck, fsck_repair_impossible) \
x(0, need_snapshot_cleanup) \
x(0, need_topology_repair)
enum bch_errcode {
BCH_ERR_START = 2048,
#define x(class, err) BCH_ERR_##err,
BCH_ERRCODES()
#undef x
BCH_ERR_MAX
};
const char *bch2_err_str(int);
bool __bch2_err_matches(int, int);
static inline bool _bch2_err_matches(int err, int class)
{
return err && __bch2_err_matches(err, class);
}
#define bch2_err_matches(_err, _class) \
({ \
BUILD_BUG_ON(!__builtin_constant_p(_class)); \
_bch2_err_matches(_err, _class); \
})
int __bch2_err_class(int);
static inline long bch2_err_class(long err)
{
return err < 0 ? __bch2_err_class(err) : err;
}
#endif /* _BCACHFES_ERRCODE_H */