00:31 <akawolf> steev: did I miss anything?

00:33 <steev> i don't know?

04:01 <bamse> https://lore.kernel.org/linux-arm-msm/20240125-x13s-touchscreen-v1-0-ab8c882def9c@quicinc.com/ <- anyone interested in testing touchscreen on x13s?

04:01 <bamse> steev: ^^ not sure if you have one of those?

06:02 <steev> afaik, everyone does

06:02 <steev> just that it hasn't been working, i'll definitely kick the tires on that though (and then disable it :P )

06:09 <steev> bamse: definitely works here

06:10 <steev> sorta?

06:10 <steev> probably not a you thing though, definitely can touch the screen to launch stuff, but gnome isn't letting me touch the titlebar and drag things around

06:18 <steev> oh, it's just not letting me drag alacritty, gnome terminal is fine

06:21 <steev> and verified i wasn't just using the udev rule to enable it

08:34 <jhovold> steev: no, not all x13s come with a touchscreen. Mine does not (fortunately).

08:37 <Jasper[m]> Mine neither

09:45 <jhovold> Here's an updated wip branch for the X13s:

09:45 <jhovold> - https://github.com/jhovold/linux/tree/wip/sc8280xp-v6.8-rc1

09:45 <jhovold> Changes include:

09:45 <jhovold> - fix battery driver crash on boot (6.8 regression)

09:45 <jhovold> - fix lowest level of headphone volume control

09:45 <jhovold> - add missing delays to enable touchscreen

09:45 <jhovold> - drop arm64.nopauth hack

09:45 <jhovold> - drop efi=noruntime hack

09:45 <jhovold> - drop sc8280xp-crd wifi calibration hack

09:45 <jhovold> - johan_defconfig: enable camera clock controller to not block icc sync_state

09:45 <jhovold> IMPORTANT:

09:45 <jhovold> - 'arm64.nopauth' now needs to be provided in the command line (as with mainline)

09:45 <jhovold> - 'efi=noruntime' may need to be provided depending on boot firmware version and config ("Linux Boot" option) or reboot crashes (as with mainline)

09:45 <jhovold> I've also pushed on more alsa-ucm-conf change to lower the default headphones volume, which was still set too high. Anyone packaging the ucm config may want to include that one as well (e.g. clover).

09:45 <jhovold> My 6.7 branch as also been updated with the new headphone volume fix for anyone that prefers to stay on 6.7 for a while still.

10:23 <craftyguy> jhovold: nice, thank you :D

10:31 <_[m]123> no touchscreen either here, do they even come in matte ?

10:33 <_[m]123> the nopauth and noruntime hacks, does it mean I didn't have to put them before?

10:33 <_[m]123> because I did

10:35 <danielt> I've got a touchscreen in my X13s. It's been working for a long time but it has needed the userspace workaround. Does the testing above involve turning that off (I haven't read the commit messages yet)?

10:44 <danielt> IIRC the "standard" spec for X13s as a 300nit non-touchscreen and when I got mine there were optional upgrades to 400nit non-touch or 300nit touch. However the screen options have disappeared from the configurator in the UK store (and I believe that the "Build your own" options have disappearing entirely from the Lenovo store in some of the geographies).

10:59 <jglathe> touch screen here. will try this weekend

11:04 <konradybcio> does anyone have an x13s with a "privacy screen" option?

11:09 <_[m]123> I was not aware that was possible even !

11:09 <_[m]123> > Display Panel - 13,3" WUXGA (1920 x 1200), IPS, ontspiegeld, geen touchscreen, 100% sRGB, 400 nits, 60 Hz

11:09 <_[m]123> it says "demirrored/unmirrored"

11:09 <_[m]123> clearly pretty motivated to get the best one I could lol no wonder it was $$$

12:00 <jhovold> _[m]123: you didn't need to use the nopauth and efi noruntime parameters with my wip branches (and by extension steev's branches), they've always been needed with mainline

12:01 <jhovold> danielt: yes, i'm sure bamse didn't have a udev rule to reprobe the driver

12:05 <_[m]123> but now we'll be needing them? I think the confusion comes from installing your distro with mainline

12:10 <jhovold> right, I'm aligning my wip branches with mainline so that people can more easily switch between the two

12:11 <jhovold> hopefully there is no confusion, you need those two kernel parameters with mainline, the wip branches used to have a workaround that effectively set them for you

12:11 <jhovold> the hope was that the underlying firmware bugs would be fixed by now, but we seem to be stuck with those parameters for a while still

14:25 <ema> I've added a few words about the firmware file needed for audio to work to https://wiki.debian.org/InstallingDebianOn/Thinkpad/X13s#Audio - I think a few people were asking about that here in the past

14:26 <ema> hopefully soon https://bugs.debian.org/1057040 will be fixed and we'll be able to remove the words :-)

15:03 <ema> quite beautiful that a few seconds ago there was only silence here, now there's Master of Puppets

15:09 <ema> also kernel 6.7.1 currently in experimental includes QCOM_QSEECOM/QCOM_QSEECOM_UEFISECAPP=y https://tracker.debian.org/news/1496976/accepted-linux-671-1exp1-source-into-experimental/ which means EFI variables work fine

15:10 <ema> as in efibootmgr and friends

16:07 <bamse> jhovold: CONFIG_FW_DEVLINK_SYNC_STATE_TIMEOUT=y seems cause sync_state to happen in the desired fashion i talked about on Wednesday...

17:29 <steev> jhovold: ah, mine was supposed to *not* come with it, but it did anyway

17:31 <travmurav[m]> inb4 only disabled in software

17:36 <steev> travmurav[m]: heh, that's how i discovered it; i'd run the echo command to see what it did... and suddenly my touchscreen was working

17:38 <steev> jhovold: have you looked over krzk's patch for the microphone?

17:39 <steev> srini says it breaks it but i don't see how

17:39 <steev> https://lore.kernel.org/all/20240125154531.417098-1-krzysztof.kozlowski@linaro.org/

17:51 <craftyguy> so for booting jhovold's 6.8_rc1 branch, only the two cmdline params are required now? efi=noruntime and efi=noruntime, and nothing else ?

17:52 <Jasper[m]> swap one of those for arm64.nopauth

17:52 <craftyguy> oops, copy/paste error :P

17:52 <Jasper[m]> Yeah the rest is fixed upstream afaik

17:52 <craftyguy> nice, thanks :)

17:52 <Jasper[m]> Don't have anything else on fedora either

18:24 <dgilmore> bamse: touchscreen works here with that patch and without a udev rule

18:31 <jhovold> craftyguy, Jasper[m]: no, you still need the ignore-unused ones

18:32 <jhovold> the command line section from the wiki no applies to both wip and mainline (and distro) kernels:

18:32 <jhovold> https://github.com/jhovold/linux/wiki/X13s#kernel-command-line

18:32 <craftyguy> oh ok, that wasn't clear to me since it said "to boot mainline"

18:33 <craftyguy> should probably say "to boot mainline OR my branch" or something :P

18:33 <craftyguy> (or "and", whatever. I read that wiki section that thought it was only for mainline and not your branch)

18:33 <jhovold> yeah, that word was there since before you only needed the two ignore-unused ones with wip

18:33 <craftyguy> so to clarify, this is what I need with your 6.8_rc1 branch?__"clk_ignore_unused pd_ignore_unused arm64.nopauth efi=noruntime"_

18:34 <jhovold> correct

18:34 <jhovold> that is, same as with mainline and distro kernels

18:34 <craftyguy> gotcha, thank you :)

18:36 <Jasper[m]> jhovold: Ah wait yeah, my bad

18:36 <jhovold> I've dropped "mainline" from that section in the wiki now (I considered doing so earlier today, but left it in in the end)

18:39 <craftyguy> jhovold: ya that's a lot more clear now :)

18:50 <clover[m]> jhovold: interesting updates and thanks fo rheads up

18:51 <clover[m]> for heads up*

18:52 <clover[m]> as for the guy who wants kernel config changes, yes i take PRs

20:43 <steev> i plan to go over them for 6.8 changes

21:04 <derzahl> hey guys, I have a question. I've been playing with an aarch64 port of proxmox on my opi5plus(rk3588) and surprised to find that 1) the iso img boots and installs without a hitch, and 2) the resulting install also boots without any dtb being specified(albeit, after a 5-10 minute hang) and seems to function decently

21:05 <derzahl> i take it the kernel is using ACPI instead of the devicetree to boot...?

21:06 <derzahl> if so, im wondering what, if any, disadvantages that may bring

21:06 <derzahl> sorry for being slightly off-topic and possibly a dumb question

21:11 <derzahl> i forgot to mention, the thing that struck me as the strangest is that the generic 6.1 kernel even showed a text login prompt on the HDMI out. Graphic support has only been present in the old 5.10 BSP kernel up until very recently(mainline 6.7 i believe)

21:30 <steev> derzahl: likely it's just generic stuffs; but are you sure u-boot doesn't provide a dtb at all? i thought most did

21:34 <robclark> check /proc/device-tree/ .. or dmesg should print the machine name from dtb

21:40 <derzahl> theres no uboot, i think the intention of the port is to run on arm64 actual server, no SOCs

21:41 <derzahl> i recall reading the UEFI firmware i flashed might provide a BSP devicetree

21:41 <derzahl> but im also booting in acpi-only mode

21:42 <derzahl> robclark: will check, i compiled a 6.7.1 kernel which also boots in ACPI-only mode, but now tried to boot in devicetree mode while passing the proper dtb and its hanging

21:43 <derzahl> putting it back to working config right now

21:52 <derzahl> ok, its back up. no /proc/device-tree exists. so the UEFI firmware must be supplying configuration info through ACPI?

22:03 <robclark> it's possible, the kernel does support ACPI on arm.. which works for devices that are simple enough

22:04 <robclark> what the kernel doesn't support is the PEP stuff (which is kinda the way MS/qcom punted when it came to things too hard to do properly in ACPI)

22:11 <derzahl> Platform extension plugin? what kind of device(s) would you expect to need them to function properly?

22:14 <robclark> yeah, platform extension plugin. It basically encapsulates the various clk controller, genpd/gdsc, and interconnect drivers (and maybe regulators and that stuff too?)

22:15 <robclark> I'm not entirely handles situations where client driver wants to directly and explicitly set bandwidth, like display controller where required bandwidth is a function of resolution

22:15 <konradybcio> power*

22:16 <konradybcio> actually, power engine* IIRC

22:16 <robclark> is that a windows API?

22:18 <konradybcio> sorta kinda

22:18 <konradybcio> saying "windows APIs" is very liberal, as everybody just NIHs stuff in their own hundreds-of-megabytes-big drivers

22:18 <konradybcio> see https://learn.microsoft.com/en-us/windows-hardware/drivers/bringup/device-power-management and https://learn.microsoft.com/en-us/windows-hardware/drivers/kernel/overview-of-the-power-management-framework

22:24 <robclark> hmm, did they backronym "platform extension plugin" to "power engine plugin"?

22:24 <derzahl> bonkers...i cant even get this kernel to boot when i pass it a dtb. i know DTB support is compiled in but I based my config on the initial proxmox aarch64 generic kernel...must be something enabled that conflicts with device tree usage

22:24 <robclark> https://learn.microsoft.com/en-us/windows-hardware/drivers/kernel/using-peps-for-acpi-services <-- older doc that refers to it as "platform extension plugin"

22:30 <steev> oh, you're using the opi5plus uefi firmware

22:45 <craftyguy> hmm I got 2 lockups in a row with jhovold's 6.8_rc1 on the x13s. no clue what caused it, whatever was in the kernel log when it happened wasn't written to disk

22:54 <robclark> /o\

22:54 <robclark> hard enough lockup that you can't even ping the laptop?

23:03 <steev> i've not had a chance to play yet