06:47 <Tooniis[m]> lumag_: im using mdp5. i dont think icc is too important, i just moved it as well while moving iommus since i saw there was also a similar change in reading it from dts in that patch

06:47 <Tooniis[m]> ill try that series

07:01 <Tooniis[m]> lumag_: i see what you mean now. i confused the changes between dpu1 and mdp5. mdp5 still tries to get iommus from the mdp node.

07:02 <Tooniis[m]> im not sure what is causing this then

07:02 <Tooniis[m]> ill try only moving interconnects and see if that does anything. i didnt suspect it because icc wasnt a blocker for display even before it was supported

07:36 <Tooniis[m]> lumag_: you were right, it was only caused by the interconnect change.

07:37 <Tooniis[m]> ill send a new version of the icc dts patch, and probably convert mdp5.txt to yaml and document interconnects to make it clear where they must be put in

07:37 <Tooniis[m]> thanks for helping

15:42 <aka_[m]> Marijn: any chance you can give any tip how to debug downstream display subsystem?

15:42 <aka_[m]> I get hwccomposer dead while trying to use LE.UM.2.X 4.9 fork on 8976.

17:33 <aka_[m]> ioctls seems to be same like in partial log and throw invalid argument somewhere in data passed to one of these ioctls... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/mRIlruYozhMndpRoxeSNVmwI)

17:56 <aka_[m]> ok downstream feel disabled here

17:56 <aka_[m]> disabled dsi pll1 and tux is no longer on screen

17:56 <aka_[m]> so something has changed

21:35 <Marijn[m]> aka_: Sorry I have no idea, I don't use hwcomposer on downstream :/

21:36 <Marijn[m]> aka_: Also I think you asked something about drivers we modified for 8976. Angelo has a bunch of hfpll changes that I don't remember if they are already sent upstream. I have a gcc patch that makes some clocks shared_ops (`enable_safe_config` on """newer""" downstream), but it is incompatible with your branch. I did not have to touch pinctrl (yet).

21:36 <Marijn[m]> aka_: Anything you're working on?