Hi, On Wed, Jan 15, 2025 at 1:45 AM Andy Yan <andyshrk@xxxxxxx> wrote: > > > Hi Doug, > > 在 2025-01-15 00:44:41，"Doug Anderson" <dianders@xxxxxxxxxxxx> 写道： > >Hi, > > > >On Tue, Jan 14, 2025 at 1:05 AM Andy Yan <andyshrk@xxxxxxx> wrote: > >> > >> > >> Hi All, > >> > >> At 2025-01-13 18:17:38, "Andy Yan" <andyshrk@xxxxxxx> wrote: > >> > > >> >Sorry, please don't merge this patch. after further testing, > >> >I found that there are still some changce, it can't read edid. > >> > >> It turns out that we need set hpd-reliable-delay-ms = 120 in dts to ensure > >> the right time to access edid. > > > >That seems awfully high and feels likely to be a problem with your > >board design and not the panel. Are you sure HPD is even hooked up > >properly on your board? Maybe you're missing a pullup/pulldown config > >somewhere? Would it be better to just specify "no-hpd" and get the > >full "HPD absent" delay? > > > > Yes, you are right, after checking the schematic, I found that the HDP indeed does not > hooked up on the board. > > I do more tests with hpd-reliable-delay-ms to a short value, From the current tests, this > value can be set below 10, even lower, but I need to do more tests to confirm how low > it can actually be set。 If HPD isn't hooked up on your board then I would suggest using the "no-hpd" + "hpd-absent-delay-ms" instead of "hpd-reliable-delay-ms". That more accurately describes your hardware. I'd personally pick "200 ms" for your "hpd-absent-delay-ms" unless you know for sure that all the panels you might hook up have less or more. 200 ms is very common for the max HPD time. -Doug
