08:53 <mszyprow> javier__: ping

08:55 <mszyprow> javier__: which kernel version was the last one you have tested on peach-pi? I'm surprised that peach-pit (exynos5420 one) fails to boot since v4.9-rc1... I was convinced that peach-pi and peach-pit are so similar, that both should work with latest kernel...

09:11 <javier__> mszyprow: it was v4.15-rc3, that's when I 976a9b35d77a ("ARM: dts: exynos: Enable Mixer node for Exynos5800 Peach Pi machine")

09:11 <javier__> mszyprow: did you try booting with EXYNOS_DRM disabled?

09:11 <javier__> the biggest diff between pi and pit I think is that the latter has a eDP/LVDS bridge

09:11 <mszyprow> javier__: yes, without DRM it boots fine

09:12 <javier__> mszyprow: that's what I thought, I guess the bridge code path hasn't been exercised for a while since kernelci and myself only have a pi

09:12 <mszyprow> javier__: so it is indeed related to exynos_dp and bridge handling

09:14 <mszyprow> javier__: bisecting with power-domain code disabled pointed me to my commit f0a8b49c03d2 "drm/bridge: analogix dp: Fix runtime PM state on driver bind", what is a bit strange

09:14 <mszyprow> javier__: but I suspect that this is related to lack of proper eDP/LVDS setup

09:16 <javier__> yeah, power domains, clocks and the exynos drm driver caused a lot of issues in the past

09:17 <javier__> things got better since Inki provided his dev tree in -next since kernelci could catch issues earlier

09:17 <javier__> mszyprow: on a related note, I remember watching a talk from one of your Samsung poland colleagues about setting a LAVA instance and sending report back to kernelci

09:17 <javier__> do you know what happened to that effort?

09:18 <mszyprow> hmmm, and pit uses different eDP/LVDS (parade,ps8625) than snow (nxp,ptn3460)

09:18 <javier__> mszyprow: that's correct. Although I thought some RK board used the same bridge?

09:18 <mszyprow> javier__: probably lack of time on Inki's side

09:19 <mszyprow> javier__: during last few releases he hardly managed to prepare his tree on time for merge...

09:20 <javier__> mszyprow: oh, it seems I didn't explain myself clearly. Inki does now provide a branch that's pulled by -next (and so tested by kernelci)

09:21 <javier__> I meant if you knew what happened to the effort from your poland colleague about having your own lava instance and report to kernelci

09:21 <javier__> that would help a lot since you have HW that no one else has

09:22 <javier__> mszyprow: this is the talk I'm referring btw https://www.youtube.com/watch?v=g4Hw5I2sk3A

09:22 <mszyprow> javier__: that still on todo list with very low priority... :/

09:22 <mszyprow> javier__: well, Pawel is from different group

09:23 <mszyprow> javier__: he is indeed working on local farm of boards

09:23 <mszyprow> javier__: and service for managing them

09:23 <javier__> ah sorry, I read Samsung R&D Poland and (wrongly) assumed you were from the same group

09:24 <mszyprow> javier__: but we are waiting with adding our boards until they finish and have it ready

09:24 <javier__> mszyprow: I see

09:25 <mszyprow> javier__: it might be also a lot of bureaucracy needed to have it available to the public

09:26 <javier__> yeah, good point