Add ability to set a user defined DMA transfer callback

[steppers]

• This allows the user to avoid unnecessary spin locks while waiting for a DMA transfer to complete.

• This has the same limitations as the dmaWait() function in terms of supported devices.

• This has been tested on PyGamer M4 to signal to the application that a framebuffer copy has completed and the framebuffer can be modified before initiating the next DMA transfer. In essence this allows us to mimic a vblank interrupt (though a lack of broken out sync pin from the display module prevents us avoiding tearing altogether).

• If you require a direct test example, do let me know and I’ll throw something together.