16:44 <abhinav___> bamse: I had a question about the eDP panel you had used for sc8180x as this came up during my discussion with our india team while they were developing the panel driver changes for sc7280 eDP panel. Today, for all panels not using "edp-panel" compatible, the panel-edp.c panel driver does not vote for the panel regulators during probe. If that doesnt happen EDID cannot be read from the panel to get the list of modes supported.

16:44 <abhinav___> So did you hard-code the modes for your panel like some other panels are doing . posting relevant snippet here

16:44 <abhinav___> https://www.irccloud.com/pastebin/Xrq02vAg/

16:47 <abhinav___> the second question was that, even if you did hard-code the mode, with our current eDP + DP driver, the link training is done in the hpd handler which shall again happen before panel power on. link training needs some DPCD reads which would fail if the panel is not powered up. So i wanted to understand how link training was succeeding for your panel with the current flow

16:47 <abhinav___> https://www.irccloud.com/pastebin/tljOc8dc/

16:48 <bamse> abhinav___: my panel is powered up by the bootloader, so the edid read does work for me...i haven't tried to power down the panel and see that the sequence of panel ops would allow it to be powered up and identified

16:49 <abhinav___> bamse: so your panel regulators are voted "always-on" ?

16:49 <bamse> abhinav___: i need to go back, but i think it runs on some external regulator and i don't touch the enable gpio

16:50 <abhinav___> bamse: got it, thanks !

16:50 <bamse> abhinav___: which i bet will fail spectacularly if i enter a proper suspend/resume cycle ;)

16:51 <abhinav___> bamse: yes we also expect it to :) just wanted to unravel the mystery of how it was working