05:37 <pstef> update-initramfs on my Debian Trixie points out that I'm missing firmware: a640_gmu.bin a630_zap.mbn a619_gmu.bin a615_zap.mbn a540_gpmu.fw2 a530_zap.b02 a530_zap.b01 a530_zap.b00

05:38 <pstef> not sure what to think about it because it looks like it's not missing anything

05:47 <JensGlathe[m]> those are older ones, should be no issue

05:52 <pstef> thanks!

06:35 <kuruczgy[m]> I am having issues with sleep on my yoga, it wakes up immediately (according to the logs), but the screen fails to resume:... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/SDybmgdVOGlcNAlHzeODtOOA>)

06:36 <JosDehaes[m]> Yes, I have the same, I think it's just not working yet

06:48 <pundir> JensGlathe[m]: how do I get started with your Ubuntu_Desktop_24.04_x1e_6.11rc_v5.img.xz build on my Yoga Slim 7x? I have not dealt with Windows machines since early 2000s, so I'd need a bit of hand holding.

06:49 <kuruczgy[m]> pundir: Like, you need help with booting the image?

06:52 <pundir> kuruczgy[m]: yes please. does it boot like an usual x86 live disk image (or we are not there yet?)

06:52 <pundir> i'm trying to follow this wiki https://gist.github.com/joske/52be3f1e5d0239706cd5a4252606644b

06:53 <pundir> but no idea which tools to use for bitlocker/resizing/disabling-secure-boot

06:53 <pundir> again absolutely noob on Windows platform

06:54 <kuruczgy[m]> Then I would link you to my guide, it has all the Windows specific stuff described in detail: https://github.com/kuruczgy/x1e-nixos-config#setup-guide

06:54 <kuruczgy[m]> (You can ignore all the stuff about NixOS)

06:55 <kuruczgy[m]> But basically: you need to turn off bitlocker, disable secureboot, then just boot the USB by pressing F2 during boot and selecting it

06:55 <kuruczgy[m]> Feel free to ask questions if something is unclear

06:55 <pundir> nice! that would definitely help. I'll try that today

06:57 <kuruczgy[m]> (In theory you could also just wipe Windows and not care about it, but I would not recommend that any time in the near future, until Linux support is super stable.)

07:40 <JensGlathe[m]> @pundir @kuruczgy I would add further instructions on the [wiki page](https://github.com/jglathe/linux_ms_dev_kit/wiki/Bootable-Image-for-multiple-Snapdragon-(SC8280XP)-and-Snapdragon-X-Elite-(X1E80100)-laptops) as they come in and I find time to structure /write it. My recommendation would be to *not* wipe the windows install, better go for a new SSD / gparted magic if you do want a permanent install. The image is suited for

07:40 <JensGlathe[m]> this. Occasionally, you do need some bits from the Windows side, there is even a script on the image (fetch_sc8280xp_fw.sh) that can do this. Should be a good template for your own machine specific cariant, or (Jehova!) a generic version.

07:46 <JensGlathe[m]> Oh and the image itself is bootable from USB. It has its own grub/efi bootloader and a few dtbs wied up, no need to even touch the local disk. Disabling secure boot / bitlocker is a sensible decision, though. You avoid a lot of pain this way.

07:49 <JensGlathe[m]> Hopefully I can add the HP Omnibook X14 to the mix soon

08:05 <JensGlathe[m]> s/wied/wired/

08:43 <pundir> JensGlathe[m] @kuruczgy[m] I plan to resize the Window's [C:] partition and install Ubuntu_Desktop_24.04_x1e_6.11rc_v5.img.xz on the new partition from the live disk. What are the caveats (if any)?

08:50 <JensGlathe[m]> The image is already 2 partitions. One of it is also ESP. The SSD will also have an ESP, and ther should only be one used. So, it boils down to only copying over the rootfs partition, and doing some rewiring for the already on SSD ESP. Including re-installing grub. I would recommend to do the boot from USB (SSD) first. If that works well enough, you can run sudo update-grub and it will add the local Windows install to the boot

08:50 <JensGlathe[m]> menu (and kill some text maybe, but that's nothing functional). I'm sure I've described ways to do the install on SSD with commands in the discussions already, need to fo a Wiki section for it.

08:52 <JensGlathe[m]> As a precaution I would recommend to have a backup /recovery image / separate SSD to try these things with some fallback/restore option.

08:55 <kuruczgy[m]> (I am not using the Ubuntu image so my setup might be less relevant for you, but I am still adding it as a data point.)

08:55 <kuruczgy[m]> Here is my partition layout:

08:55 <kuruczgy[m]> I have two ESPs and I moved around some of the windows partitions with gparted

08:56 <kuruczgy[m]> I also managed to mess up Windows once and had to download Lenovo's recovery image (needs windows, talk about a chicken-and-egg problem), it's a pain but it managed to reinstall a working Windows.

08:58 <JensGlathe[m]> Nice. As long as you don't change UUIDs of the partition or filesystem Windows may still be able to boot. My experience is to rather use the ESP from Windows than risking Windows not booting (as long as you require it)

09:01 <travmurav[m]> fwiw I also added second esp + manually added a boot entry for grub/sd-boot on it via efi shell bcfg

09:02 <travmurav[m]> So I get both windows boot manager or sd-boot in uefi boot menu, with linux being the default in my case

09:03 <travmurav[m]> But if firmware uses esp for efi capsules, it can break with two esp on the same disk

09:03 <travmurav[m]> As in firmware updates via windows can break

09:51 <JensGlathe[m]> Hmm sound is odd since 6.11

10:02 <JensGlathe[m]> oh its now broken on 6.10 too, I would say alsa-ucm has a problem

12:25 <pundir> JensGlathe[m]: Ubuntu_Desktop_24.04_x1e_6.11rc_v5.img.xz failed to boot for me on Yoga Slim 7x from a USB disk https://usercontent.irccloud-cdn.com/file/4ZJAh2AW/PXL_20240827_122011635.MP.jpg

13:07 <kuruczgy[m]> <pundir> "Jens Glathe: Ubuntu_Desktop_24.0..." <- Why does it say "Hardware name: Windows Dev Kit ..."? It should say "Lenovo Yoga Slim 7x ..." or "LENOVO 83ED ...". To me it seems like it doesn't have the right device tree.

13:14 <FarchordSteveCossette[m]> <JensGlathe[m]> "@pundir @kuruczgy I would add..." <- Love this btw. The USB-C issue on that last paragraph is my BANE!

13:54 <anonymix007[m]> Is something wrong with system-boot entry? https://katb.in/fonezisuvol

14:04 <kuruczgy[m]> <anonymix007[m]> "Is something wrong with system-..." <- Hm are you sure the kernel image is called "vmlinux"? On arm64 it's usually called "Image".

14:11 <anonymix007[m]> Yes, I'm sure: https://katb.in/neyecapexaz

14:15 <anonymix007[m]> Aren't these "EFI stub" messages from the kernel itself already?

14:25 <kuruczgy[m]> <anonymix007[m]> "Aren't these "EFI stub" messages..." <- Yeah, good point actually, they are. Then I have no idea why the EFI stub is failing, haven't seen this yet...

14:42 <jhovold> kuruczgy[m]: please avoid using matrix quoting when replying here

15:03 <JensGlathe[m]> @pundir ouch something with the GIC. But yeah, you need to choose the right boot entry otherwise bad things happen. Will check the wiring again.

15:03 <JensGlathe[m]> I can vaguely remember that there was something with gic, but that early...

15:04 <travmurav[m]> I mean you will obviously get a system error if you try to talk to gic using 8c3 address while running on x1e

15:04 <travmurav[m]> <del>if only there was a way for the image to pick correct dtb for you automatically</del>

15:04 <TSIN[m]> wrong kernel？

15:06 <TSIN[m]> * wrong kernel/dtb？

15:06 <JensGlathe[m]> travmurav: was focused on the bsod issue, didn't want to upset things more. Yes dtbloader should be used. That's what it's there for. To be fair you would also need an initramfs specific to the dtb

15:06 <travmurav[m]> can you not just include everything for the installer though?

15:07 <JensGlathe[m]> you mean, like on the ISO?

15:08 <travmurav[m]> in the said initramfs

15:08 <travmurav[m]> as in

15:08 <travmurav[m]> include a superset of "needed for x1e" + "needed for 8c3"

15:09 <JensGlathe[m]> yes, you would need 2 initramfs as it seems. the modules lists need to be different to not get into trouble later

15:10 <travmurav[m]> well something is already horribly wrong if making some extra unused modules available causes trouble I'd say...

15:11 <steev> JensGlathe[m]: audio is broken on 6.10/6.11 on what? my x13s is doing just fine?

15:12 <JensGlathe[m]> x13s. sounds odd, is good on Windows.

15:12 <steev> too fast/too slow. tinny...

15:13 <JensGlathe[m]> tinny yes, like pieces ofthe waveform missing

15:13 <JensGlathe[m]> travmurav: I agree.

15:13 <steev> that's not new, we don't have the full range

15:14 <JensGlathe[m]> it sounds horrible

15:14 <steev> patches welcome :)

15:23 <steev> from memory, we artificially limit the power so that the speakers can't be broken. maybe we need to write some profiles for speakersafetyd

16:09 <anonymix007[m]> I have no idea why, but it just booted with UKI. Literally everything is the same, just packaged into one .efi file. This must be systemd-boot issue

16:30 <jhovold> JensGlathe[m]: have you recently switched to pipewire? audio works just fine on the x13s with pulseaudio

16:30 <jhovold> there's a workaround for playback, but capture is really broken

16:31 <jhovold> should be the same problem with all qcom socs

16:32 <jhovold> https://github.com/jhovold/linux/wiki/X13s#audio

16:43 <anonymix007[m]> Is brightness control supposed to work on x1e already?

16:45 <robclark> works on my yoga 7x

16:46 <konradybcio> anonymix007: such things are per device matters

16:46 <konradybcio> although they tend to be identical on 95% of the devices..

16:47 <anonymix007[m]> Well, it doesn't work on T14s

16:47 <robclark> I've not looked closely but don't see anything in 7x dts for backlight so I guess it is probably controlled via dp-aux

16:50 <anonymix007[m]> panel-simple-dp-aux complains that the BOE 0x0b66 panel is unknown. Could this be related?

16:59 <robclark> I don't think so.. if it doesn't know the panel it just uses conservative (long) delays in the powerup sequence

16:59 <robclark> fixing that might shave a few ms off of boot time

17:13 <HdkR> If backlight controls don't work on the t14s when I get mine then I'll be complaining, I don't want to burn in that OLED :P

17:23 <\[m]> you bought one?

17:23 <\[m]> which cpu?

17:32 <FarchordSteveCossette[m]> The oled on those x1es are goooorgeous

17:33 <HdkR> \[m]: The only CPU available, X1E-78-100

17:35 <FarchordSteveCossette[m]> tbf there are others. But not nearly as much choice in laptops though

17:35 <HdkR> If the Snapdragon dev kit ever becomes a real product then I'll use that for most testing anyway

17:41 <kuruczgy[m]> > please avoid using matrix quoting when replying here

17:41 <kuruczgy[m]> (It might be nice to document this in the channel topic, instead of having to explain it to every new member. Maybe if a generic "matrix bridge etiquette" document exists somewhere linking to that would be enough.)

17:41 <kuruczgy[m]> jhovold: Sorry, I thought that they just did not render on IRC, but now I checked and apparently the bridge is appending the quote to the beginning of the message, I can see how that makes the reading experience worse...

17:41 <kuruczgy[m]> * > please avoid using matrix quoting when replying here

17:41 <kuruczgy[m]> jhovold: Sorry, I thought that they just did not render on IRC, but now I checked and apparently the bridge is appending the quote to the beginning of the message, I can see how that makes the reading experience worse...

17:41 <kuruczgy[m]> (It might be nice to document this in the channel topic, instead of having to explain it to every new member. Maybe if a generic "matrix bridge etiquette" document exists somewhere linking to that would be enough.)

17:41 <kuruczgy[m]> And now the matrix editing feature apparently duplicated my message on IRC... okay, I definitely have to pay more attention to how stuff renders on IRC

17:47 <robclark> HdkR: did you get the t14s with OLED? Would be nice to get ACPI dump from that. My suspicion is that we need different dtb for OLED vs non-OLED t14s. (Would like an edid-decode)

17:48 <robclark> if the OLED is using the same or similar samsung panel as the 7x it shouldn't be so hard to get working (and I guess backlight controls will "just work")

17:48 <anonymix007[m]> Resuming is also broken. After closing/opening the lid screen doesn't come up and this is in dmesg:... (full message at <https://matrix.org/_matrix/media/v3/download/matrix.org/WKAdJwAAihpeEPUoWzxuPOLC>)

17:48 <HdkR> robclark: Yea, got the one with the 120hz oled

17:49 <HdkR> Ships sometime next week

17:49 <robclark> ohh, mine is only a measly 90hz

17:49 <robclark> still, quite nice looking

17:50 <HdkR> The 90hz on the 7x is pretty nice as well

17:54 <kuruczgy[m]> anonymix007: I have the same issue, please definitely post about it if you make progress.

17:54 <kuruczgy[m]> (Also mine isn't sleeping properly, according to the kernel logs at least, you might also want to check whether yours really goes to sleep and it's not just the screen going black.)

17:58 <anonymix007[m]> Well, my ssh session appears to be suspended when the lid is closed. Here's dmesg (after resume): https://katb.in/yabeteronef

18:00 <anonymix007[m]> Funnily enough, the screen comes up (for kernel logs) back while being rebooted.

18:04 <kuruczgy[m]> <anonymix007[m]> "Well, my ssh session appears..." <- Hm, but in the log timings there is only a 1ms gap between the CPUs going to sleep and then waking.

18:04 <kuruczgy[m]> But if your SSH session gets suspended until you open the lid, that must mean that the system IS sleeping, but the clock does not run during sleep?

18:04 <kuruczgy[m]> I have never seen something like that before. But this could mean that sleeping IS working after all for me too, I need to check using SSH too I guess.

18:04 <robclark> there are I think still several reasons why it isn't hitting lowest power states.. and depending on dts the panel regulator might not be switching off (see `arm64: dts: qcom: x1e80100-yoga: Update panel bindings` .. if vreg_edp_3p3 is regulator-always-on for your laptop then panel is probably not switching off completely

18:05 <kuruczgy[m]> (Also FYI I can bring the screen back just by switching to another VT and back.)

18:05 <anonymix007[m]> Isn't this clock thing universal? AFAIR it is the same on x86 as well

18:09 <kuruczgy[m]> Just checked on my x86 laptop, the (kernel log) clock does seem to advance during sleep

18:14 <JensGlathe[m]> jhovold: No haven't changed anything re audio

18:15 <jhovold> JensGlathe[m]: but are you using pipewire? What you describe sounds like the pipewire issue

18:16 <JensGlathe[m]> it is pipewire as per default of Ubuntu

18:16 <jhovold> perhaps some setting changed when it was updated

18:18 <JensGlathe[m]> re sc8280xp vs x1e driver clash: sounds like it may be better to build separate images

18:18 <robclark> pipewire is also what fedora uses.. not really sure what the pipewire vs audio issue is, but it certainly would be nice to see it fixed

18:19 <robclark> what is the driver clash?

18:19 <jhovold> playback is broken for some users, depends on quantum settings

18:19 <jhovold> capture is broken for everyone, sounds like a robot with hiccups

18:20 <robclark> hmm, on x13s iirc I was still seeing "Dummy Output" in gnome sound settings (although alsactl was showing something sensible)

18:20 <jhovold> there is no known workaround for the capture issue, so this would need to be fixed before we can do video conferencing

18:21 <jhovold> (video in firefox depends on pipewire)

18:21 <robclark> hmm, ok, playback is what I was more interested (only for selfish reasons, VC is all on corp devices)

18:21 <jhovold> camera I meant

18:22 <jhovold> not sure if distro's use different default settings, iirc debian even documented that you may need to fiddle with the quantum settings if you want to use pipewire

18:22 <jhovold> the capture issue appears to be qcom specific though

18:23 <robclark> hmm, "fiddle with settings I'm not familiar with" isn't a great UX

18:23 <jhovold> indeed

18:24 <jhovold> https://wiki.debian.org/PipeWire#choppy_audio_on_systems_with_high_load

18:25 <jhovold> not the same description of the issue, but the "fix" was

18:31 <steev> maybe we need something like 90219f1bd273 "ASoC: SOF: intel: hda: Clean up link DMA for IPC3 during stop" - no real idea, just did a search for other robotic stuff

18:39 <steev> but the good news is, the robotic stuff isn't only us on pipewire

18:50 <JensGlathe[m]> The driver clash is sc8280xp drivers and a x1e driver in /etc/initramfs-tools/modules. On Volterra it loads tcsrcc_x1e80100 - leading to my aforementioned bsod in my tests, despite camcc_sc8280xp being blacklisted. Although pundir might have chosen the first entry (which is for Windows Dev Kit) it would be an issue even with dtbload.

18:53 <robclark> hmm, even if you modprobe the drivers, they shouldn't be binding since wrong compatible

18:56 <JensGlathe[m]> Despite my esoteric bsod issue on Volterra (who has that hw anyway) I decided to give it a go. as anybody booted it succesfully?

20:06 <steev> booted what

20:07 <JensGlathe[m]> on x1e hardware. I tested it on Volterra and X13s

20:10 <robclark> I think we kinda want to figure out why x1e modules cause problems on sc8280xp.. that is _supposed_ to work (and I don't see any obvious reason that it shouldn't work)

20:10 <JensGlathe[m]> that, too

20:12 <steev> unless it's somehow matching on a generic tcsr

20:13 <JensGlathe[m]> I'm afraid this may be the reason. Need to check

20:13 <robclark> https://www.irccloud.com/pastebin/FcyGYU46/

20:13 <robclark> looks ok to me

20:14 <JensGlathe[m]> hmm

20:22 <JensGlathe[m]> But its loaded due to the modules list in initramfs. So it will do some init code, like probing

20:23 <robclark> it shouldn't probe, just register a driver (which doesn't match against any device)

20:23 <JensGlathe[m]> okay