Standardize how `display_set_orientation` functions

[danieldegrasse]

Is your enhancement proposal related to a problem? Please describe.
display_set_orientation is currently somewhat arbitrary- some drivers swap the x/y resolution they report when the display orientation is set, and other drivers do not.

For a list, see here: #73360 (comment)

Describe the solution you'd like
We should standardize whose responsibility it is to handle a display rotation event- should the driver change the resolution it reports, or should the consumer of the display API swap its x/y coordinates? Personally I think the consumer should handle this. Initial work to implement this for LVGL was done in #73556.

Describe alternatives you've considered
We can continue having display rotation be a bit arbitrary.

[IAmAnkur]

My 2 cents on this: (If i understood the enhancement request correctly)

I agree with the author's approach,
Since driver reports all required (3) parameters(x_res, y_res & current orientation) in get_capability call, i feel the standard way should be to offload swapping x_res, y_res to consumer so that unnecessary work in driver is avoided.
Consumer of driver API can check for orientation and can take a call accordingly (to swap or not to swap).