07:57 <narmstrong> Darkmatter66_: is it specific to v5.7 ?

10:01 <Darkmatter66> narmstrong, I think so, i didn't have that issue in 5.6 if i recall correctly

10:06 <Darkmatter66> narmstrong, I can git bisect the "drivers/gpu/drm/" path, but i don't know of a guaranteed way to reproduce the crah each time, but I might try it later anyway

10:06 <Darkmatter66> crash*

10:07 <narmstrong> Darkmatter66: you can do a few runs per bisect step, if you have 1 fail you consider it’s a bad commit

10:08 <Darkmatter66> narmstrong, yes that was my thinking, but it might take a bit of a long time

10:08 <narmstrong> I did this to bisect a panfrost crash, I automated 10 runs via our LAVA lab per bisect step to find the offending commit, it was long (2h) but worth it !

10:09 <Darkmatter66> narmstrong, should i bisect "drivers/gpu/drm/meson" or the whole "drivers/gpu/drm/"

10:11 <narmstrong> Better bisect the whole, you will have almost the same number of steeps anyway and it could be an issue out of drm

13:05 <Darkmatter66> narmstrong, It seems that the drm issue has existed for ever

13:06 <Darkmatter66> narmstrong, I tested 5.7 5.6 and 5.5 and they all had the issue

13:06 <Darkmatter66> i didn't bisect in-between though i was just trying to find a good commit to start with

13:08 <Darkmatter66> narmstrong, the crash can be reproduced at my end as follows

13:09 <Darkmatter66> 1- the board is booted while pluged-in to the screen and the screen is powered on

13:09 <Darkmatter66> 2- the screen is powered of from it's remote

13:10 <Darkmatter66> 3- power on the screen from the remote again, as soon as i press the power on button the crash happens,before even the screen backlight is on

13:10 <Darkmatter66> the crash at this point is without the "flip_done timed out" lines at the end

13:11 <Darkmatter66> 4- at this point the screen is fully powered on but shows no signal, and now the "flip_done timed out" starts appearing one by one

13:12 <Darkmatter66> 5- as soon as the last line "[PLANE:35:meson_primary_plane] flip_done timed out" appears in the dmesg the screen detects the signal and shows the hdmi output of the box

13:13 <Darkmatter66> narmstrong, I can test older kernels if it helps, but i think it's fair to say that the issue is old(maybe from the begining) but no testing has caught it

16:07 <narmstrong> Darkmatter66: strange since you're the first one reportig this

16:07 <narmstrong> maybe your screen is doing funky stuff with HPD

16:08 <narmstrong> can you try with another screen/monitor ?