Bug Fixes in Multimedia
Daring logic fix (or so I hope). There is a potential (read: actual)
threat from having needData and enoughData called from potentially
(read: actually) 2 different threads, as needData relies on KIO's resume
function, which only queues data dispatching. If KIO is suspended before
that queued call arrives no data will be sent. We hope that moving it
here should fix that though.