00:12 <bamse> i've rewritten the dp stuff completely...

00:13 <bamse> some of it had made it upstream and got some new stuff on the list

00:13 <bamse> in particular there's still no agreement of how the tcpm should invoke the hpd interrupt on the dp driver...but it's being looked at

00:15 <qzed> okay, I'll need to start with making things not crash first though...

00:15 <qzed> last time I had the dispcc enabled things got weird, panel flickering and going dark, machine unresponsive to ssh

00:39 <bamse> seems consistent with an SError

11:08 <qzed> bamse: SError?

14:02 <HdkR> https://www.engadget.com/acers-chromebook-spin-714-specs-price-availability-134551833.html?src=rss Not necessarily a laptop, but a tablet+ChromeOS+keyboard case. Another 7cG2 device.

14:43 <robclark> hmm, I think that is quackingstick?

15:26 <bamse> qzed: the typical outcome when you do things such as touch unclocked registers

15:28 <qzed> ah, I guess that makes sense

15:29 <qzed> bamse: I'm struggling a bit with the backlight definition

15:29 <qzed> specifically the PWM control

15:29 <qzed> when I try to add the LPG controller it fails probing with -22

15:30 <qzed> qcom-spmi-lpg c440000.spmi:pmic@5:lpg: failed to add PWM chip: ret -22

15:30 <qzed> and I'm not sure if that's actually used to drive the backlight pwm signal

15:30 <qzed> there's some xml config in acpi with

15:30 <qzed> <BacklightPmicControlType>3</BacklightPmicControlType>

15:31 <qzed> according to some older doc I found, that is neither WLED nor LPG

15:31 <qzed> https://developer.qualcomm.com/qfile/34758/lm80-p0337-1_b_display_drivers_acpi_and_xml.pdf

15:33 <bamse> qzed: did you enable CONFIG_PWM?

15:33 <qzed> huh, no... I guess that's it

15:34 <qzed> thanks

15:34 <bamse> qzed: jhovold posted a fix to depend on that, because without it we'll just fail to register the pwm_chip and bail from probe

15:34 <qzed> yeah, I just went through the xmenu config and enabled all dependencies... didn't think to look for PWM specifically

15:36 <qzed> also just checked for the XML config thing in the Flex DSDT, same control type 3

15:37 <qzed> so I guess if it works there it should work on the pro x as well

15:45 <bamse> that xml is such a wonderful creation

15:50 <qzed> uhm are the pins and pingroup names offset by one in the pinctrl-spmi-gpio driver?

15:50 <qzed> group names start with gpio1

15:51 <qzed> do I reference that pin with <&pmc8180e_gpios 0 ...> or <&pmc8180e_gpios 1 ...> now?

15:55 <qzed> bamse: so according to https://elixir.bootlin.com/linux/latest/source/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c#L1053 pin with index 0 in the kernel has number 0 but group gpio1

15:56 <qzed> so I need to reference with <&pmc8180e_gpios 0 ...> but configure with pins = "gpio1"?

15:56 <qzed> or am I missing something?

15:56 <bamse> qzed: https://elixir.bootlin.com/linux/latest/source/drivers/pinctrl/qcom/pinctrl-spmi-gpio.c#L949

15:56 <bamse> i think we fixed that for you...

15:56 <qzed> ah

15:56 <qzed> thanks!

15:57 <qzed> ah, now I see it's also used in pmic_gpio_of_xlate

16:01 <bamse> ahh, so i got the wrong line for you after all

16:01 <bamse> yeah, but you're right...the of_xlate has the same offset

16:02 <bamse> qzed: we started off without that, and it was a complete mess :)

16:02 <qzed> I can imagine, was quite a bit confused just now

16:03 <bamse> that's just because you expected that it wouldn't be user friendly

16:03 <qzed> yeah xD

16:06 <qzed> neat, backlight control actually works!

16:06 <qzed> thanks for your help, yet again!

16:36 <bamse> nice!

16:41 <bamse> qzed: i'll take some time to get the sc8180x.dtsi on the list, and am looking forward to see some patches from you :)

16:46 <qzed> nice! I'll probably need to clean up my DT a bit before sending parts in

16:49 <bamse> no worries, i'm glad to see the progress

16:49 <bamse> and that i managed to land what i think was the last dependency

16:50 <bamse> i haven't checked though if georgi picked up the interconnect fixes, so perhaps that will show up in 5.20 instead

16:50 <bamse> but at least we should be able to run the primus & lenovo machine off mainline/linux-next shortly, with just a dt added

16:51 <bamse> and there are discussions related to usb/dp and the battery power-supply driver

17:21 <qzed> awesome!

17:23 <qzed> guess I should start sending in the refactoring changes I made to the surface specific stuff to make DT/of integration easier (also to enable Pro 8 HID support)... currently sitting in our kernel for testing

17:24 <qzed> the fun thing is that DT/OF integration for the Surface Aggregator isn't actually that much, it's more refactoring that module loading on platforms with ACPI works well

17:25 <qzed> we already use firmware nodes to describe devices that aren't described in ACPI (because MS decided to enumerate them in some other for some reason)

17:25 <qzed> *platforms without ACPI

18:40 <bamse> interesting, but that sounds promising