is there a calculation sheet which would allow seeing the correlations? TI i.e. has them for some of there DSI products (either an excel sheet and/or some GUI based SW).I think the minimum HFP requirement is that when in burst mode, there needs to be sufficient time to drop to LP11 state and re-enter HS.
Otherwise, the DSI Host has to remain in HS mode for the whole vertical active period; and for mysterious reasons, it seems to make the timings go awry on RP1, causing the "runt line". (The latter may be fixable by not setting the burst mode flag in hardware. However, staying in HS for the whole frame is unreliable at best.)
The minimum time for HS->LP->HS depends on the MIPI D-PHY clock rate, inter alia (the DSI spec gives some calculations but they are confusing; we have tables of magic numbers from Synopsys that the RP1 driver uses).
Now VC4 and RP1 will use different clock rates. RP1 will try to fit the MIPI rate exactly to the video mode (put another way, it expects the requested video mode to match the desired DSI timings; which was perhaps a naive interpretation, but the only really well-defined one).
Often RP1 will use a lower data rate than VC4, which means it needs longer porches.
Knowing which parameter to change would be simpler than running a try&error type of test.
Statistics: Posted by aBUGSworstnightmare — Thu Aug 01, 2024 10:34 am
