09:17 <srinik> DavidHeidelberg[m]: qcom,apr-domain is marked as deprecated to discourage new users. apr driver still has the code to support it. But if you have a patch to cleanup the exisiting ones please send.

09:19 <DavidHeidelberg[m]> srinik: I'll do

18:15 <Marijn[m]> I've got commandmode panels working properly on sm6125 🎉, and am now working towards cleaning all the patches for submission. Hitting the first snag though, we're on `MSM_DSI_6G_VER_MINOR_V2_3_0` and that selects an `sdm845_dsi_cfg` with mismatching `io_start` (wrong address, and we only have one DSI PHY). Is there a desired way to fix this? Are these addresses supposed to be hardcoded in the driver anyway or should we just provide the

18:15 <Marijn[m]> PHY id/index in DT?

19:00 <minecrell> Marijn[m]: I think we had this discussion here before

19:02 <Marijn[m]> minecrell: October first apparently, not that long ago despite me completely forgetting. Doesn't seem like we came up with any conclusive solution though, so it's worth asking again

19:03 <Marijn[m]> (Matrix still has the history, in case someone needs it to re-read the problem and proposed solutions)

19:03 <minecrell> yeah iirc everyone was sharing a bit of frustration and then moved on :p

19:04 <Marijn[m]> Yeah, discussion moved on to other frustrating problems :( - we should really just start tackling and fixing these now

19:38 <robclark> abhinav___, lumag: So, it seems like the "blank screen" issue I was hitting the other day (which I *think* started after the dsi host vs bridge probe order changes) is somehow some sort of race condition.. I can semi-easily reproduce it if I disable serial console and turn down the loglevel.. but the only thing I'm seeing in disp snapshots that is consistently the same between multiple good and multiple bad snapshots is dspp_0

19:38 <robclark> starting at offset 0xc00 .. ie. in the snapshots of bad state they are all zero's, but in good snapshots they are all non-zeros. I don't *think* we are writing anything in that range (as best I can tell from looking at the code) but no idea what is at those offsets.. bridge register dumps match between good/bad.. and I can turn on bridge colorbar pattern w/ i2cset and that works, so I think the bridge itself is in an ok state

19:38 <robclark> fwiw https://people.freedesktop.org/~robclark/tmp/disp-good.txt and https://people.freedesktop.org/~robclark/tmp/disp-bad.txt

19:39 <robclark> (hmm, maybe the display snapshots should add something so we can tell what hw it is from.. but I guess that is a separate topic)

21:25 <robclark> also, it seems unlikely that we are programming *any* dspp regs.. maybe what we are dumping at offset 0xc00 is actually something else?