Please add marshalling support for the Mutex class:
$ irb
Marshal.dump(Mutex.new)
TypeError: no marshal_dump is defined for class Mutex
    from (irb):1:in dump'     from (irb):1     from /usr/bin/irb:12:in '
Because of, Almost all OSs don't have a way to restore mutex locking state.
I have no idea how implement it. ;-)
Here is the workaround I am currently using:
class Mutex
 def marshal_dump
  []
 end
 def marshal_load array
  # do nothing
 end
end
Thanks for your consideration.
Hmmm??
It seems locking state lost. Is this acceptable?