feat(draw): Enable TSC Compression scheme for frame buffer in Nema acceleration

[MarkIoanAmat]

TSC Compression for FB

While using NemaGFX to accelerate LVGL, we have the option to use ThinkSilicon's proprietary TSC compression algorithms to encode/decode the frame buffer in favor of memory consumption.

Steps to Enable

• Enable TSC Frame buffer type. e.g lv_display_set_color_format(display, LV_COLOR_FORMAT_NEMA_TSC12A);
• Bypass lv_draw_sw_blend function with the use of the define below inside in lv_conf.h
#define LV_DRAW_BLEND_CUSTOM LV_DRAW_BLEND_CUSTOM_NEMA_GFX

Features not supported while using TSC compression for frame buffer

• LV_COLOR_FORMAT_ARGB8565
• LV_COLOR_FORMAT_RGB565A8
• Blending mode LV_BLEND_MODE_SUBTRACTIVE

[MarkIoanAmat]

@kisvegabor @liamHowatt

[uLipe]

First pass, I have a small comment.

[uLipe]

I like this initial checking, but I'd like to see an extra information to aid people to track when the things go wrong here, simply putting log-err or log-warning would be helpful here.

[kisvegabor]

Let's consider #7531 too, as it aims to solve the same issue.

[MarkIoanAmat]

@kisvegabor
I had a quick look at #7531 to my understanding these 2 solutions are complementary as they solve different problems. PR #7509 aims to by pass the whole lv_draw_sw_blend function and use a custom one. On the other one PR #7531 allow registering different blending handlers inside lv_draw_sw_blend. Also PR #7509 is adding a special feature which requires the presence of NemaGPU Plus the TSC Compression scheme for FB and it is working under specific constrictions. I can not think of a way merging those two solutions without polluting the common code, which is not preferred of course.
Waiting for your feedback. Thank you.