00:01 <steev> okay, replaced the ssd in my t14s, so that's the first step to switching it over to linux :D

00:01 <steev> i did try out tobhe_ 's ubuntu image and it was actually quite pleasant aside from the really slow wifi speeds

06:03 <krzk> steev: does usb-c orientation and altmode works for you on the ubuntu concept image?

06:04 <steev> krzk: i need to check; i dunno about altmode as i don't think my only usb-c display adapter does that, but i can check that too i think

06:04 <krzk> I am using that one and neither on ubuntu kernel nor on johan's tree I cannot get it working - there is simply no PMIC_RTR_ADSP_APPS rpmsg device which is responsible for altmode/orientation

06:05 <krzk> steev: just plug any USB-hub or USB adapter after system boots... or re-plug it with different orientation.

06:06 <krzk> I already added all possible modules to initrd and tweaked the order and still no help. The funny part is that looks like this is problem of my ubuntu as I did not see many reports (and it works in other folsk setups)

06:06 <krzk> steev: USB-C to be precise

06:06 <steev> that's for orientation though not altmode, right?

06:09 <krzk> steev: both

06:10 <krzk> that's the rpmsg device which will spawn pmic_glink device handling entire usb-c magic

06:21 <steev> krzk: able to use the Apple usb-c digital av multiport adapter in both orientations in the usb-c port closest to the front - about to test back https://usercontent.irccloud-cdn.com/file/7VPvGjfh/1743661173.JPG

06:21 <steev> Port closest to the display as well yeah

06:24 <steev> i don't know if i have the latest version of the image, but it's the #30 kernel build here

06:24 <steev> this is the oled no touch screen variant

06:25 <krzk> steev: thanks for checking... so maybe the moment of Ubuntu concept installation mattered? I wonder why my setup has it broken. :/

06:26 <krzk> btw, are using encrypted rootfs?

06:26 <steev> this is the livecd not an install

06:26 <steev> for personal reasons, i will never install ubuntu

06:27 <krzk> ah, ok

06:38 <steev> maybe try the live iso and see if it works there for you? i don't know what is in build 30 of their kernel, but it worked in both orientations of both ports here

06:45 <krzk> steev: my livecd also works, because it is something with my installation of ubuntu, that's why I was wondering if you installed it

06:45 <steev> ah, got ya, sadly, no, sorry

09:54 <tobhe> krzk: the livecd is probably on a slightly older kernel than your installed system, so that would be one possible difference

09:55 <tobhe> i also don't remember anyone else reporting this though

10:19 <krzk> tobhe: I also booted same kernel as livecd (I think -5) and tried others as well

10:20 <krzk> tobhe: I already spent few days debugging this and have no clue. The difference I have comparing to many folks is that I use full disk encryption and mayybe this changes something how initramfs is handled and how modules are loaded

10:21 <krzk> tobhe: for example on my system - on the livecd and ubuntu (your) kernels, I do not have sound, because not all modules are loaded. I need to modprobe lpass and WCD codecs manually.

10:21 <krzk> Or manually add them to /etc/initramfs-tools/modules - then I have sound. FW is already in initramfs and rootfs, btw.

10:22 <krzk> So I suspect there is something similar happening for rpmsg/glink/pmic_glink/pmic_altmode

10:22 <krzk> For now what I spotted is some race condition in glink (narrowing it now), but I don't think this is the cause

10:39 <tobhe> maybe a timing issue?

10:40 <tobhe> maybe you hit some timeout during the password prompt

14:37 <kuruczgy[m]> Were the slim7x DP patches dropped from the ubuntu-concept kernel branch? Can't find it in either qcom-x1e or unstable-x1e. (Not sure which one is supposed to be the most up-to-date.)

14:47 <tobhe> possibly, I did reduce the number of out-of-tree patches over the last month

14:54 <tobhe> yep looks like it

14:57 <adrian_> JensGlathe[m]: it's me, hi. :-P

14:57 <adrian_> (I'm the problem, it's me)

14:58 <JensGlathe[m]> Did I miss something?

15:02 <JensGlathe[m]> Okay the gold bar dropped. Yeah 😆

15:02 <tobhe> an engineer with a screwdriver (and a song)

16:11 <SpieringsAE> adrian_: same lol, working at a small company, I both program the systems and assemble them

16:51 <SpieringsAE> JensGlathe: alexandrs vinarskis managed to get the hdmi working on the zenbook and with those changes applied it also works on the vivobook

16:51 <SpieringsAE> its very borked though

16:51 <SpieringsAE> but I managed to get display out of it, it completely blew up when I unplugged the hdmi cable though lol

16:52 <SpieringsAE> It also only seems to work when the hdmi cable is plugged in on boot time, not inserted later

16:53 <JensGlathe[m]> Well thats progress

17:32 <krzk> tobhe: abelvesa: JosDehaes[m]: so the answer for this my damn issue - PMIC_RTR_ADSP_APPS was never sent from FW via glink to kernel. Never. It turns out that except proper firmware in qcom/x1e/LENOVO/21N1/ I also had some ADSP files with JSN files in qcom/x1e/ (so top level) and they were silently messing things up!

17:34 <krzk> After removing these top level ADSP files - PMIC_RTR_ADSP_APPS appears and type-c orientation works. I don't know how did these files appear. Maybe from linux-firmare? Or maybe I put them there after installation? Who knows? But I do not have them in my installation log (and I follow/document steps) thus I suspect linux-firmware or some other Ubuntu package.

17:38 <krzk> Uh, I pinged wrong person, should be jhovold above

17:45 <tobhe> dpkg -S /file/path should tell you if it was part of a package

17:46 <tobhe> like $ dpkg -S /lib/firmware/qcom/x1e80100/adsps.jsn

17:46 <tobhe> linux-firmware: /lib/firmware/qcom/x1e80100/adsps.jsn

17:47 <tobhe> and https://web.git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/qcom/x1e80100

17:52 <tobhe> why does this only happen on your machine though?

17:59 <krzk> tobhe: no clue! Maybe depends on the actual FW? Before I removed windows, so like 5 months ago, I upgraded the Lenovo FW to newest available

18:04 <sgerhold> krzk: the jsn files shouldn't be doing anything anymore, since we've had the in-kernel pd-mapper enabled since 6.12... if they really do make a difference, make sure you remove the userspace service from your distro install

18:04 <krzk> tobhe: ah, maybe this was zst from linux-firmware? So the /lib/firmware/qcom/x1e80100/LENOVO/21N1/qcadsp8380.mbn.zst? I see I removed it as well.

18:05 <krzk> sgerhold: pdmapper is disabled for this platform, I think sinvce forever... and never got enabled

18:05 <sgerhold> krzk: so removing what exactly did make a difference here?

18:05 <krzk> tobhe: I copied the firmware from lenovo EXE package, which had unpacked files, but linux-firmware brought duplicated and maybe they conflicted?

18:06 <krzk> sgerhold: unfortunately I did two things - dropped top-level adsp files and the packaged files from linux-firmware

18:06 <sgerhold> I don't see how removing some top-level .mbn files could make a difference

18:06 <sgerhold> Having two different adsp versions with/without .zst seems more likely then

18:07 <sgerhold> The upstream linux-firmware should work fine for T14s, maybe try a clean /lib/firmware from upstream and remove everything custom?

18:10 <krzk> sgerhold: yeah, I'll try. Initially I needed the self-unpacked firmware, but it's not the case anymore

18:19 <krzk> sgerhold: tobhe: so clean linux-firmware works, but when I added firmware from Lenovo package - where the binaries differ - I gave both packed zst (from linux-firmware) and unpacked mbn (from lenovo) and this breaks things.

18:19 <krzk> So it was still me who broke my setup :/

18:20 <krzk> s/I gave both/I have both/

18:30 <tobhe> best possible outcome :) at least not everything is broken

20:32 <dgilmore> steev: I should make a fedora image for the t14s

20:35 <steev> dgilmore: you should :D

20:35 <steev> i'd test it out

21:54 <dgilmore> Excellent