A relatively rare concurrency bug existed in these methods due to them not correctly holding the partition lock during calls to _FAT_directory_entryFromPath. That method can trigger reads to the partition. Due to the lock not being held, another thread could still trigger a concurrent read, potentially corrupting the cache. The fix is simply to ensure the lock is held around the calls.
A relatively rare concurrency bug existed in these methods due to them not correctly holding the partition lock during calls to
_FAT_directory_entryFromPath. That method can trigger reads to the partition. Due to the lock not being held, another thread could still trigger a concurrent read, potentially corrupting the cache. The fix is simply to ensure the lock is held around the calls.