diff options
| author | Kent Overstreet <kent.overstreet@linux.dev> | 2024-04-05 23:27:27 -0400 |
|---|---|---|
| committer | Kent Overstreet <kent.overstreet@linux.dev> | 2024-04-06 13:50:26 -0400 |
| commit | 6088234ce83acec4aaf56ecc0e9525bac18b4295 (patch) | |
| tree | b72e80255b37eaba6b7574b56f5391277b5635cd /fs/bcachefs/btree_key_cache.c | |
| parent | 05801b6526156aefe55c0440fab877109c9a89c5 (diff) | |
bcachefs: JOURNAL_SPACE_LOW
"bcachefs; Fix deadlock in bch2_btree_update_start()" was a significant
performance regression (nearly 50%) on multithreaded random writes with
fio.
The reason is that the journal watermark checks multiple things,
including the state of the btree write buffer, and on multithreaded
update heavy workloads we're bottleneked on write buffer flushing - we
don't want kicknig off btree updates to depend on the state of the write
buffer.
This isn't strictly correct; the interior btree update path does do
write buffer updates, but it's a tiny fraction of total accounting
updates and we're more concerned with space in the journal itself.
Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>
Diffstat (limited to 'fs/bcachefs/btree_key_cache.c')
| -rw-r--r-- | fs/bcachefs/btree_key_cache.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/fs/bcachefs/btree_key_cache.c b/fs/bcachefs/btree_key_cache.c index 581edcb0911b..4ae7890f07c6 100644 --- a/fs/bcachefs/btree_key_cache.c +++ b/fs/bcachefs/btree_key_cache.c @@ -659,7 +659,7 @@ static int btree_key_cache_flush_pos(struct btree_trans *trans, commit_flags |= BCH_WATERMARK_reclaim; if (ck->journal.seq != journal_last_seq(j) || - j->watermark == BCH_WATERMARK_stripe) + !test_bit(JOURNAL_SPACE_LOW, &c->journal.flags)) commit_flags |= BCH_TRANS_COMMIT_no_journal_res; ret = bch2_btree_iter_traverse(&b_iter) ?: |