On Mon, Aug 11, 2014 at 11:49 PM, Henrik Nordström henrik@henriknordstrom.net wrote:
mån 2014-08-11 klockan 15:46 +0100 skrev Luke Kenneth Casson Leighton:
c) part of that area involves the device-tree "merging" code (that will need to be written), so that the CPU Cards can dynamically detect what they're plugged into, load the Device-Tree fragment out of the EEPROM and go from there.
Beaglebone have something like that in their kernels with device tree overlays loaded runtime based on eeprom contents of connected "shields".
oh gooood, whew.
publishing "buttons" via a USB-HID, and publishing the analog joystick again as a USB mouse, in theeeoorrryyy "Random OS Mk VII" *might* actually work.
I might even say SHOULD work, that's the point of having USB standards. But not all buttons might be mapped to meaningful events out of the box.
true
And joystick do have it's own USB-HID profile.
oh, duh :) thank you for pointing that out.
l.