How to correctly add support for iMX6 GPU?
Some calls from stm32_dma2d are in the source files
\lv_core\lv_refr.c
\lv_draw\lv_draw_blend.c
\lv_draw\lv_draw_img.c
the documentation only requires callback implementation for gpu_fill_cbfill and gpu_blend_cb
I’d start with the higher-level callbacks (gpu_fill_cb and gpu_blend_cb). They’re designed to provide an average increase in speed, be more generic across hardware types, and be implemented externally by the user.
If you want even more speed, we added internal, built-in support for the STM32 in the specific cases where its GPU provided the best speed. You would have to do your own experimenting with the code to find where the iMX6 GPU provides the most speed benefits, and then you could add the iMX6 GPU code directly into the core rendering logic (like we did for STM32).