07:50 <narmstrong> xdarklight: for g12a they added RXSENSE and dynamic HDR, the big dev was done on GXBB and GXL/GXM to support div40 and TMDS scrambling

07:51 <narmstrong> but maybe the PHY arch was good enough to permit these

18:29 <xdarklight> narmstrong: I'm making the assumption that the color problem is due to the VIU registers (so between VIU_ADDR_START and VIU_ADDR_END). do you think that this makes sense? if so then I guess I'll start register diff'ing (again...). here's what I get on EC-100's Endless OS for these registers (CVBS is connected, HDMI not): https://pastebin.com/JMyGVYw4

18:34 <narmstrong> xdarklight: yes it could be

18:35 <xdarklight> narmstrong: honestly, I expected more differences than these 18 registers: https://pastebin.com/pU4mtSRM

18:39 <xdarklight> ah, later in the boot process it's more differences: https://pastebin.com/1p93nYeZ

18:39 <narmstrong> Interesting but I don’t have the ref numbers in my head

18:40 <narmstrong> Reg*

19:16 <xdarklight> weird

19:16 <xdarklight> if I use "OSD_BLK_MODE_32 | OSD_COLOR_MATRIX_32_ARGB" then I don't see the boot logo or framebuffer console but kmscube output looks fine

19:18 <xdarklight> however, if I use "OSD_BLK_MODE_32 | OSD_COLOR_MATRIX_32_RGBA" then I see the boot logo and framebuffer console but with a blue tint and the colors in kmscube are wrong