On Tue Mar 21, 2023 at 5:50 PM CET, Roman Beranek wrote: > > Also, how was it tested/confirmed? > > By counting Vblank interrupts (GIC 118). Sorry, that was perhaps too abbreviated. To test this change, I set up an A64 board running kmscube on DSI-1 and verified that the rate of Vblank IRQs tracked with a video mode set on DSI-1, once with a 2-lane panel and once with a 4-lane panel. Roman
