09:19 <rawoul> lumag: you mean this patch ? https://github.com/rawoul/linux/commit/927d68a85ce12a13e76711627e9fcd3fc7fdaaf1

09:20 <rawoul> I don't think it's actually correct, since the frame size can vary between 4 and 6

09:21 <rawoul> although you have to be careful not to allocate a static buffer with HDMI_INFOFRAME_SIZE(VENDOR), which might be too small. The STI driver defines HDMI_INFOFRAME_MAX_SIZE 6 for this case

10:24 <lumag> rawoul, and also the actual handling and packing of the vendor infoframe

10:31 <lumag> rawoul, I think the define should be removed completely, as not to be misleading. Or renamed to _MIN_SIZE

11:19 <lumag> rawoul, just to clarify: I'd like to ensure that we pass proper infoframes, AVI, audio, VS, SPD, etc, It looks like this becomes more important over the time, as monitors start to require them.

11:31 <rawoul> I agree. I remember now why I have this patch in my branch, I had also added vendor infoframe somewhere else and was hit by the exact problem I described: https://pastebin.com/UbVM6hiL

11:33 <lumag> yeah. that's one of the reasons why I wrote that we should remove that define

11:37 <lumag> And another problem with that define is that the HF-VSIF, vendor frame from HDMI 2.0, can take up to 32 +3 bytes

20:53 <konradybcio> intel just killed their risc-v development accelerator program

20:53 <konradybcio> wrong chat.. sort of..