We should always assume the filesystem is volatile and check each IO operation. While we're at it reorganize checks so that early-out errors are near one another.
We should always assume the filesystem is volatile and check each IO operation. While we're at it reorganize checks so that early-out errors are near one another.