08:22 <qschulz> alikates: we have DSI working on 6.6 on a PX30 board

08:22 <qschulz> https://git.theobroma-systems.com/ringneck-linux.git/tree/arch/arm64/boot/dts/rockchip/px30-ringneck-haikou-video-demo.dts

08:22 <qschulz> I vaguely remember that there's some trickery downstream with dsi1 which hasn't being worked on upstream yet, so it's possibly just not working

08:34 <mmind00> alikates: "copy pasted" sounds "dangerous" ... i.e. the vendor-kernel's init-sequences encode the actual mipi commands, so need modification for a panel driver ... is your panel-driver somewhere on git{hub,lab,*} ?

08:59 <alikates> mmind00: by copypaste i mean that i used them as is, but removing the command, delay and length parts and left only the actual payload

09:00 <alikates> qschulz: thanks, i will check it for inspiration. correct me if im wrong, but iirc rk3326 has only one dsi interface, so I don't think thats the problem

09:02 <alikates> ooh, maybe given that px30 has 2 and 3326 ony one, the missing one may be dsi0

09:03 <qschulz> alikates: mmmm maybe I mixed stuff and this dsi stuff is rather for RK3399

09:03 <qschulz> actually pretty sure that's the case

09:03 <qschulz> don't see any dsi1 or whatever in Rockchip's kernel for PX30 or RK3326

09:04 <qschulz> though the RK3326 has only VOPB and not VOPL so that's a note-worthy difference

09:04 <qschulz> sorry for the mislead

09:05 <qschulz> alikates: depending on the boot order and HW design, there may be an issue with the DSI PHY not being supplied power in time as well, but IIRC when this happens the driver should not have probed

09:06 <qschulz> that we had an issue with, which resolved itself because we now have a GPIO expander that defer the probe of everything behind, including the DSI panel...

09:18 <alikates> from what i can read of the decompiled DTB, both the phy ans controller are in the same power domain. Maybe I need to load one driver after another...

09:18 <alikates> anyways the backlight turns on quite late in the boot process, just before switching to usermode

09:29 <alikates> there was some quirk config for changing the probe order of panel phy and controller right??