13:16 <JeromedeBretagne[m]> Progress report on the Surface Pro 9 5G (sc8280xp-based), I've been able to get the built-in display backlight working, reusing the similar nodes from the Lenovo Thinkpad X13s, thanks jhovold: !

13:17 <JeromedeBretagne[m]> I've shared it here: https://github.com/JeromeDeBretagne/linux_surface_pro_9_5G/commits/surface/sp9-5G/v6.15-rc4/

13:17 <JeromedeBretagne[m]> and here is the actual commit : https://github.com/JeromeDeBretagne/linux_surface_pro_9_5G/commit/8eb986f4d99fbe4310feecd83a43aac3f0a6b74d

13:22 <JeromedeBretagne[m]> However, I still haven't found how to get the built-in display detected, despite quite a few attempts (testing mdss0_dp3 and mdss0_dp2)

13:30 <JeromedeBretagne[m]> @konradybcio: Any clue about this, since you've done the bring-up of the Surface Laptop 7 family (if I'm not mistaken)? How did you find the correct gpio for the vreg_edp_3p3 regulator and how to choose the right mdds node?

13:45 <JeromedeBretagne[m]> Here is what I've tried to enable mdss0_dp3 for instance:

13:45 <JeromedeBretagne[m]> https://github.com/JeromeDeBretagne/linux_surface_pro_9_5G/commit/3428f3cf01bf926328d2b3987d66bad7b7102f2a

13:45 <JeromedeBretagne[m]> is something obvious missing?

13:54 <JeromedeBretagne[m]> Here is the corresponding dmesg output:

13:54 <JeromedeBretagne[m]> https://gist.github.com/JeromeDeBretagne/2e4e243e771cd0baf8704fc243ec508c

13:55 <JeromedeBretagne[m]> I can see these 2 lines that don't look very good:

13:55 <JeromedeBretagne[m]> [ 12.105304] panel-simple-dp-aux aux-aea0000.displayport-controller: Prepare timeout after 5 tries

13:55 <JeromedeBretagne[m]> [ 12.107089] panel-simple-dp-aux aux-aea0000.displayport-controller: Couldn't power on panel to ID it; using conservative timings: -110

13:56 <JeromedeBretagne[m]> do I need to load some specific modules, in addition to panel_edp ?

13:58 <JeromedeBretagne[m]> Jens Glathe: do you have some suggestions maybe? I've also tried mdss_dp2 for your info, with a similar patch, I'm a bit out of ideas

15:32 <JensGlathe[m]> can I see your dt somewhere

15:34 <JeromedeBretagne[m]> The full devicetree? Sure, here you go:

15:34 <JeromedeBretagne[m]> https://github.com/JeromeDeBretagne/linux_surface_pro_9_5G/blob/3428f3cf01bf926328d2b3987d66bad7b7102f2a/arch/arm64/boot/dts/qcom/sc8280xp-microsoft-arcata.dts

16:02 <JensGlathe[m]> Try just 2 lanes? I don't know the res, but isn't 2 lanes enough?

16:02 <JensGlathe[m]> And I guess You need a power supply

16:03 <JeromedeBretagne[m]> Shouldn't I see some messages in dmesg if it was the link training that was failing, due to 4 lanes instead of 2 ?

16:03 <JensGlathe[m]> it says can't power up

16:03 <JeromedeBretagne[m]> the screen has this resolution : 2880 x 1920

16:04 <JensGlathe[m]> ok this could be 4 lanes

16:04 <JeromedeBretagne[m]> a missing power supply ? in which node ? I think I've stayed as close as possible to the nodes from the X13s dt

16:05 <JeromedeBretagne[m]> I have the vreg_edp_3p3 power supply in the panel node

16:06 <JeromedeBretagne[m]> but maybe it's the wrong gpio

16:07 <JeromedeBretagne[m]> I've tried both GPIO 25 (as on the X13s) and GPIO 70 (as used on the Surface Laptop 7 and the Surface Pro 11)

16:08 <JensGlathe[m]> X13s is the nearest (except for the Blackrock devkit)

21:17 <JensGlathe[m]> Huh. This fingerprint thing works

22:06 <kuruczgy[m]> How do kernel maintainers usually apply patch series with conflicts? I know about b4 and it seems to have this magic b4 am -3 flag to generate some fake base commits for 3way merges, but it doesn't seem be able to find the base commit for the whole patch series to begin with.

22:06 <kuruczgy[m]> Is getting git am to generate conflict markers without having the base commits for a 3way merge possible somehow? Otherwise it seems pretty hopeless to apply the whole series basically manually...

22:06 <kuruczgy[m]> (I am trying the MSM MST series btw (https://lore.kernel.org/all/20241205-dp_mst-v1-0-f8618d42a99a@quicinc.com/ ), it's from December so having quite a few conflicts is expected I guess.)

23:47 <robclark> depends on how bad the conflicts are.. I'll normally try to resolve them but if they are too bad I'll ask poster to rebase