fix(db): correct unsafe RLock order (fixes #9906) (#9910)

Marshal() was called with the read lock held, and in turn called
Created() which also takes the read lock. This is fine by itself, but
there is a risk of deadlock if another call to lock the mutex happens
concurrently, as the lock call will block the inner rlock and the outer
rlock can never become unlocked.

It's an easy fix as marshalling is guaranteed to be called with a read
lock and does not need to call any methods that read lock themselves.

lib/db/meta.go

@@ -76,11 +76,12 @@ func (m *metadataTracker) Unmarshal(bs []byte) error {
 	return nil
 }
 
-// Marshal returns the protobuf representation of the metadataTracker
-func (m *metadataTracker) Marshal() ([]byte, error) {
+// protoMarshal returns the protobuf representation of the metadataTracker.
+// Must be called with the read lock held.
+func (m *metadataTracker) protoMarshal() ([]byte, error) {
 	dbc := &dbproto.CountsSet{
 		Counts:  make([]*dbproto.Counts, len(m.counts.Counts)),
-		Created: m.Created().UnixNano(),
+		Created: m.counts.Created,
 	}
 	for i, c := range m.counts.Counts {
 		dbc.Counts[i] = c.toWire()
@@ -109,7 +110,7 @@ func (m *metadataTracker) toDB(t backend.WriteTransaction, folder []byte) error
 		return nil
 	}
 
-	bs, err := m.Marshal()
+	bs, err := m.protoMarshal()
 	if err != nil {
 		return err
 	}