bcachefs: Btree write buffer

This adds a new method of doing btree updates - a straight write buffer,
implemented as a flat fixed size array.

This is only useful when we don't need to read from the btree in order
to do the update, and when reading is infrequent - perfect for the LRU
btree.

This will make LRU btree updates fast enough that we'll be able to use
it for persistently indexing buckets by fragmentation, which will be a
massive boost to copygc performance.

Changes:
 - A new btree_insert_type enum, for btree_insert_entries. Specifies
   btree, btree key cache, or btree write buffer.

 - bch2_trans_update_buffered(): updates via the btree write buffer
   don't need a btree path, so we need a new update path.

 - Transaction commit path changes:
   The update to the btree write buffer both mutates global, and can
   fail if there isn't currently room. Therefore we do all write buffer
   updates in the transaction all at once, and also if it fails we have
   to revert filesystem usage counter changes.

   If there isn't room we flush the write buffer in the transaction
   commit error path and retry.

 - A new persistent option, for specifying the number of entries in the
   write buffer.

Signed-off-by: Kent Overstreet <kent.overstreet@linux.dev>

1 files changed, 41 insertions, 0 deletions

diff --git a/fs/bcachefs/buckets.c b/fs/bcachefs/buckets.c
index 153987376b89..86f48f5762dd 100644
--- a/fs/bcachefs/buckets.c
+++ b/fs/bcachefs/buckets.c
@@ -1278,6 +1278,47 @@ int bch2_mark_reflink_p(struct btree_trans *trans,
 	return ret;
 }
 
+void bch2_trans_fs_usage_revert(struct btree_trans *trans,
+				struct replicas_delta_list *deltas)
+{
+	struct bch_fs *c = trans->c;
+	struct bch_fs_usage *dst;
+	struct replicas_delta *d, *top = (void *) deltas->d + deltas->used;
+	s64 added = 0;
+	unsigned i;
+
+	percpu_down_read(&c->mark_lock);
+	preempt_disable();
+	dst = fs_usage_ptr(c, trans->journal_res.seq, false);
+
+	/* revert changes: */
+	for (d = deltas->d; d != top; d = replicas_delta_next(d)) {
+		switch (d->r.data_type) {
+		case BCH_DATA_btree:
+		case BCH_DATA_user:
+		case BCH_DATA_parity:
+			added += d->delta;
+		}
+		BUG_ON(__update_replicas(c, dst, &d->r, -d->delta));
+	}
+
+	dst->nr_inodes -= deltas->nr_inodes;
+
+	for (i = 0; i < BCH_REPLICAS_MAX; i++) {
+		added				-= deltas->persistent_reserved[i];
+		dst->reserved			-= deltas->persistent_reserved[i];
+		dst->persistent_reserved[i]	-= deltas->persistent_reserved[i];
+	}
+
+	if (added > 0) {
+		trans->disk_res->sectors += added;
+		this_cpu_add(*c->online_reserved, added);
+	}
+
+	preempt_enable();
+	percpu_up_read(&c->mark_lock);
+}
+
 int bch2_trans_fs_usage_apply(struct btree_trans *trans,
 			      struct replicas_delta_list *deltas)
 {