Not sure if I've missed something crashingly obvious here. I've adapted the BF526 USB mass storage sample to use the PID ramdisk device with the following results:
The 'disk' is present with a working file system as far as the BF526 firmware is concerned.
The device also appears as a USB drive when plugged into my Windows 7x64 development machine.
The firmware detects USB connections and disconnections and correctly lists files on the embedded file system
What is not happening (and I'm not sure if it should!) is that files written to the USB device from Windows actually make it to the embedded file system. I suspect they are just being buffered by Windows.
In the mass storage example the code simply spins whilst connected, and then spins once again whilst disconnected.
So here is the $64K question: Is any additional application level code required to ensure USB packets actually get written to the embedded file system?
If so does anyone have an example?
If not, what/where do I need to hook into the stack?