OK I can reproduce the problem now.
Preliminary observations: It looks like DSI is running slow: horizontal line period is something like 17.25us rather than 16.96us, and I guess that causes a growing mismatch (FIFO fill-up) between DPI and DSI which culminates in the short line. It is no doubt aggravated by remaining in HS mode for the whole frame.
That shouldn't happen as all clock rates are exactly achievable and related. We have tables of timeouts for HS/LP transitions but since it's not doing any, I don't think they are to blame -- and while it remains in HS, I don't see any opportunities for e.g. the PHY to exert back-pressure.
It vaguely points to some timing misconfiguration in the DSI Host.
Preliminary observations: It looks like DSI is running slow: horizontal line period is something like 17.25us rather than 16.96us, and I guess that causes a growing mismatch (FIFO fill-up) between DPI and DSI which culminates in the short line. It is no doubt aggravated by remaining in HS mode for the whole frame.
That shouldn't happen as all clock rates are exactly achievable and related. We have tables of timeouts for HS/LP transitions but since it's not doing any, I don't think they are to blame -- and while it remains in HS, I don't see any opportunities for e.g. the PHY to exert back-pressure.
It vaguely points to some timing misconfiguration in the DSI Host.
Statistics: Posted by njh — Tue Jul 30, 2024 4:30 pm