Add canvas mode support to support pixel based rendering instead of character based rendering.

[hydra]

From BF discord:

belrik: Would you be open to helping FPV.WTF create a canvas-mode renderer for vector-based OSD? That would increase the number of users able to use canvas-mode.

hydra: i fully agree that it would, and if you can do per-pixel drawing on the goggles frame buffer then yes, do it.
hydra: I personally do not have the available time to embark on such a project.
hydra: I would suggest the implementation just uses the existing FC code for the FrSky UART Canvas Mode driver, and then there's nothing else to do on the FC side of things. Then on the goggle side of things just start with the clear screen, and displaywrite implementations of which you already have existing code in the form of the MSP implementation, but obviously ditching the grid and use the pixel-based offsets instead.
hydra: obviously you need the uart protocol code too. but it doesn't seem like that big of a job to get uart protocol, basic fc/osd detection/capabilities handshake and then screen clear + font rendering going.
hydra: at least, for someone already familiar with the DJI goggle programming environment and system libraries/api that is, which also isn't me