Renesas GLCDC dirver

[becseya]

Add display driver to support Renesas GLCDC peripheral.

[becseya]

tests/src/lv_test_conf_full.h was not extended with the new config macro, as build would fail, due to missing FSP libraries. I'm assuming this is preferred over mocking everything. @kisvegabor

[kisvegabor]

I'm assuming this is preferred over mocking everything.

Correct, normally we don't compile the hardware specific code.

[kisvegabor]

It should be called in line 204 for the frame buffer.

SCB_CleanInvalidateDCache_by_Addr(fb, DISPLAY_HSIZE_INPUT0 * DISPLAY_VSIZE_INPUT0 * 2);

[becseya]

replacing 191 or keep both? and why call for the whole buffer in partial mode?

[kisvegabor]

Only at line 191.
It would be more optimal to invalidate on the changed area. It can be added now or in a later PR as an optimization.

[becseya]

The suggested line does not work, as fb is being incremented in the loop, and pointing to a bad location by line 204.
This works* at the end of function:

SCB_CleanInvalidateDCache_by_Addr(fb_background[1], sizeof(fb_background[1]));

Or this in the loop:

SCB_CleanInvalidateDCache_by_Addr(fb, w * 2);

However it only works for the artifacts seen at switch transition. Pressing the "Log out" button also creates artifacts and they are not resolved by cache invalidation. It might be a different issue, as the stripes are inside the button, with white color. At no point the animation should be that, so it's not a "slow transition" related to cache emptying out.

This is not present in direct mode, however a scroll bar appears after pressing the button once. The bar does not appear in the PC emulator with matching screen resolution.

[kisvegabor]

Looks good in general, just a few comments.

Approved by accident

[kisvegabor]

Looks good! We can fix the issue of the partial rendering in a separate PR.

The failing CI is unrelated and being discussed here.

[becseya]

Please merge without squashing @kisvegabor

[kisvegabor]

Merged, thank you Akos!