Lenovo Ideapad Slim 5(x) 14Q8X9 (83HL)

[richardnpaul]

Start of discussion, brought over from launchpad

[richardnpaul]

As I've added to the launchpad, I wrote with rufus the img file (extracted from the xz file) and tried to boot. I did this twice once for each type A port in this laptop.

I've added a screen grab of the boot process at the point that it stops after a period of time it automatically reboots, much like the 25.04 image does, it just gets further than a blinking cursor.

[jglathe]

Hmm. Could you just try the reportedly "working" image for the ThinkBook 16? They have almost the same mainboard, only real difference is the type/address for the touchscreen (which wouldn't work). Writing the stick several times should confirm its working. Since its stalling so early... interesting.
For me this looks as if it behaves like without the patches for the Purwa (x1p42100) SoC. But they are definitely there. Have you updated the device firmware from Windows?

[richardnpaul]

Okay, that gets much further. Now I get to a segfault.

[richardnpaul]

This is fully patched firmware wise via Windows

[jglathe]

Well thats something, confirmation its not the ideapad image per se. That display panel is a Samsung ATNA40, could be that it needs a different driver. Let's see what the other tester says, maybe we get a better picture (he also has the Samsung panel).
It could very well be that they have some different chips on this for the USB repeaters, which may sound odd, but without schematics you can't know.

[jglathe]

I looked again at the boot try you did with the ThinkBook 16 image, that's odd. It really went way further than the Ideapad images. Trying to think up a reason for this.

[richardnpaul]

I looked again at the boot try you did with the ThinkBook 16 image, that's odd. It really went way further than the Ideapad images. Trying to think up a reason for this.

That was my impression too, that the Thinkbook image was loading for quite some time before failing. Sorry if I didn't mention it previously.

[jglathe]

Update from the other tester colleague: He brought it up, with keyboard, touchpad, touchscreen 😁 One hurdle is the not-really enabled bitlocker on windows, which needs to be once enabled (ugh) and then explicitly disabled. I vaguely remember this with my tb16, though. Bitlocker needs to be explicitly off.
He was using a type-A USB stick, so maybe its somewhere in the writing method...

[jglathe]

I'm a bit out of ideas, this sounds odd. Especially since Windows seems to run on it fine. You tried 2 images, both fail fairly hard. It could be something with uefi , but that doesn't help either. Thanks for the data, though. Thinking about this.

[richardnpaul]

Final updates as I'm out of ideas too. I've written the ideapad 5 image from a Windows Intel x86_64 machine using 4.8 version of rufus and I downloaded again just to be certain. That didn't work, booting from one of the type A ports. Then I updated to the latest balena etcher and tried writing the image again with that on the x86_64 system, and again the same issue. Also, I installed mesa-utils under WSL2 and ran that through gWSL for a few minutes and didn't see any issues there either (to try to nail down any obvious hardware issue, but I've not had any instability under Windows at all, it just greedily eats all the RAM leaving me little for doing anything productive on the machine 😁.)

[richardnpaul]

Ahhh, kaktus178 says that he's using a 2 in 1 which means that it's probably a 83GH based model and probably explains some of the differences.

[jglathe]

Still baffled. Yes, his is 83GH IdeaPad 5 2-in-1 14Q8X9 BIOS Q1CN22WW. I'll look for differences.

[richardnpaul]

It's annoying that there's no way that I can tell to roll back the BIOS as there's no list of previous versions or way to download those versions to roll back for testing purposes

[jglathe]

Hmm I thought you provided the ACPI tables and hwinfo64 log, but you... didn't? Could you please, since they are different models apparently
What a naming mess

[richardnpaul]

aarch64-laptops/build#141

This was my PR which was merged in, I think that this was what you looked at.

[jglathe]

Sorry, too much going on, did look at the launchpad bug. Thank you for the pointer. But well then I will have another look, need to separate these two things out.

[jglathe]

Oh interesting the log shows another PCIe bus. If that's essential in any way (I could bet) this would explain that stall. So I guess I have a handle for a new try.

[richardnpaul]

I'm game if you are. Just make sure that you're okay, if you have too much going on take a little time, I'm not pressed for time to get this up and, semi, working.

[jglathe]

It is pcie3, not-too-often seen on these devices. Probably for the card reader (Microsoft does this for the SP7 iirc), we'll find out.

[jglathe]

Ok I created and test-booted (with pcie3 enabled) on the tb16, worth a try. At the usual place, Ubuntu_Desktop_24.10_Lenovo_Ideapad_Slim_5x.img.xz.
The pcie3 part gave errors on thetb16, but I guess this is to be expected since its not enabled there.

[jglathe]

Oof. Actually its on line less than previous, doesn't sound like progress. My next step would be to go all-in on the pcie3 definition and also add the port and PMIC switches in the hope that it at least throws a non-fatal error. Will change this and upload again.

[richardnpaul]

Oof. Actually its on line less than previous, doesn't sound like progress. My next step would be to go all-in on the pcie3 definition and also add the port and PMIC switches in the hope that it at least throws a non-fatal error. Will change this and upload again.

Okay that sounds good to me. I'm more than happy to test anything you manage to put together 😊

[jglathe]

Currently preparing the next one. Getting the full pcie3 port definition in (with 6.15.0-jg-9) in the hope it pushes the box over the bump to actually boot. Since it was an interesting week, we're also having working X1-45 GPU support now, this will also be on the image. The first of its kind, so its a bit more fiddling. Should drop in the next few hours.

[richardnpaul]

Oh wow, that's great news I'll be happy to give it a spin and see where we get to. Am I right in recalling that there was a 6.15.1 kernel with patches for overheating GPUs due to throttling not occurring? That might have been the x1e but I don't know if if would also apply to the X1p.
Anyway, hopefully I can test and feedback. I'm well aware this must all be a pain without the ability to get hands on the hardware.

[jglathe]

x1p42 is still so new that you won't find it in mainline. I guess I will do a first release without GPU enabled, will be fun enough when it comes up. I have X1-45 GPU running on my TB16, but its one more moving part - or three.

[jglathe]

Alright new image is uploading to Google Drive now. It says ~20mins until up, still Ubuntu_Desktop_24.10_Lenovo_Ideapad_Slim_5x.img.xz (the old one was a dud, so...)

[jglathe]

Hmm I've seen some small detail on the Microsoft laptops, also regarding pcie3. Overwriting the image again.

[richardnpaul]

I'll test when I can when I get home

[jglathe]

Ok the new one is up.

[richardnpaul]

Okay, moment of truth, reboot about to commence...

[richardnpaul]

Right, that didn't work failing exactly the same way, but I'm just checking the sha sum to ensure validity of the files (you never know when you're using files of the same name over and over.)

Edit: no such luck I'm afraid, checksum of fc244bde3bcf6632911926a360328dd1cc1ef8fb7fb0a5e14638afe18e6cb43e checks out locally.

[jglathe]

Hmm sorry to hear that. debating with my self if I want to spend €1000 to get that sob of a laptop up 😼

[jglathe]

There are actually technical / performance reasons for not using slots. DDR6 8448MT is something, I enjoy it very much.

[richardnpaul]

The LPDDR5X spec is available on LPCAMM2 which is one of the benefits of it as it supports both standards including the higher speeds, and the clk chips have already shown that we can achieve high speeds on slotted memory, that's also on sodimm - just look for 6400 DDR5 to find it, just for your information. There are other reasons for not using slots, speed just isn't one of them. I don't expect people to follow hardware as closely as I do and so to know this 😁.
Similarly, I had to tell a guy who works at Nvidia that I know that Nvidia GPU with ARM CPU chips for laptops were potentially coming as early as later this year, because he thought the possibility of that was a way off yet; he works on FOSS there so not expected to know what hardware is going to come down the pipe until it does.

[jglathe]

Good q if any vendor actually puts it into hardware you can and want to buy. I actually bought an Orion O6 board with 64GB soldered on because it is 64GB (and supposedly open - compared to qcom it is). This one will be a kvm server eventually, so it is crucial to have enough RAM in the system. Aside from that I hold it with Douglas Adams (?): Every technical invention beyond the age of 30 is black magic.

[richardnpaul]

I liked the look of the o6 when I saw that they were going to release it, but I have enough mini-pcs already (4) running proxmox (HA) and it's, the o6, just quite expensive, especially if you want the 64GB of RAM (3 of my proxmox instances have 64GB of RAM, the other just a spare 16GB stick).

Mostly I liked that it was uefi and acpi because DTB just does not look to be scalable to me (it doesn't help that you have the likes of qcom doing half a job, from what I understand, wrt acpi).

[jglathe]

O6 is getting an AMI BIOS if I read correctly, so not everything is bad. It was a change of user experience just installing the debian image and everything apparently worked ootb.

[HarryDo289]

Hi,
Lenovo Ideapad Slim 5(x) 14Q8X9 (83HL) can instal Ubuntu ARM64?
I try to install this image but not support
https://launchpad.net/~ubuntu-concept/+archive/ubuntu/x1e

[jglathe]

Okay, if this works, could you please give me the HWInfo64 log (from Windows, I thought I had yours) and the dmesg output to check for errors? Also, a list what doesn't work? Also, output of lsusb would be nice.

[richardnpaul]

Yes, you (the arch repo) have the hwinfo log, that's from my machine. I'll try to sort out dmesg and lsusb, this installer is very slow and unresponsive to ctrl+alt+t

[richardnpaul]

dmesg.log
lsusb.txt
Here we go, dmesg and lsusb from the installer (I've not tried to install yet so as for the what works/doesn't work yet, I don't know.)

https://github.com/aarch64-laptops/build/blob/master/misc/lenovo-ideapad-5-slim-14Q8X9/hwinfo-14Q8X9.LOG is the hwinfo log that I submitted.

[richardnpaul]

Hmmm, the trackpad can't be changed. Seems to be stuck in "natural" scrolling regardless of the setting chosen. I set up fingerprint unlock, closed the lid, opened the lid then I just get continuous authentication errors and can't shut the laptop down.

[richardnpaul]

Also the console was getting spammed with message about not being able to write to the journal by systemd. In the end, after trying to hit ctrl+alt+del to force a shutdown, it got stuck with systemd and also unattended upgrades not stopping, so I had to REISUB to get the laptop to reboot.

I don't see the card reader, so I expect that those code changes you made may be helpful in this, yes backlight doesn't work but I don't think that we were expecting it to and I've not heard any audio out of it (system sounds and the like) so that probably doesn't work, but that's the same as noted elsewhere. I have dumped the dmesg and verbose lsusb, just need to get signed into Github on it later to upload them.

[sandiecao-dc]

The reboot should be caused by copy _firmware.service. There maybe something wrong with the firmware, so It comes to black screen again. We rewrite a new Image and remove the copy_firmware.service, then it works.

…

From: "Richard ***@***.***> Date:  2025年6月30日 (周一) 19:45 Subject:  Re: [jglathe/linux_ms_dev_kit] Lenovo Ideapad Slim 5(x) 14Q8X9 (83HL) (Discussion #36) To: ***@***.***> Cc: ***@***.***>, ***@***.***> Okay, I got a loading to black screen, followed by a reboot, then loading to a black screen again. So we have progress, however, there's no point where the screen comes back yet 😬  I'll check out external monitors shortly. — Reply to this email directly, view it on GitHub<#36 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BLTBRN6T2A4S5JPBKF3I6PD3GEPM7AVCNFSM6AAAAAB6Z7K57WVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNRRGY4TQOA>. You are receiving this because you were mentioned.[image]Message ID: ***@***.***>

[jglathe]

This sounds a lot like the Windows partition still being encrypted with Bitlocker, an unholy issue (because its saying its not, probably). But curious, after disabling the copy_firmware.service, did it boot to display? With todays image?
To actually disable Bitlocker:

• re-enable secure boot in UEFI
• boot to Windows, enable Bitlocker
• reboot to Windows, disable Bitlocker, explicitly let it decrypt the disk
• disable secure boot in UEFI

After that, you can either boot to Windows (should have secure boot disabled and no Bitlocker - can be checked with HWInfo64) or boot the Image after it was newly written to a stick/SSD.

[sandiecao-dc]

We already disabled Bitlocker, but we still need to disabled copy_firmware.service.

[sandiecao-dc]

Yes,  It did boot to desktop after we manually installed firmware. And it displayed on EDP.

…

From: "Jens ***@***.***> Date:  2025年6月30日 (周一) 20:17 Subject:  Re: [jglathe/linux_ms_dev_kit] Lenovo Ideapad Slim 5(x) 14Q8X9 (83HL) (Discussion #36) To: ***@***.***> Cc: ***@***.***>, ***@***.***> This sounds a lot like the Windows partition still being encrypted with Bitlocker, an unholy issue (because its saying its not, probably). But curious, after disabling the copy_firmware.service, did it boot to display? With todays image?  To actually disable Bitlocker: • re-enable secure boot in UEFI • boot to Windows, enable Bitlocker • reboot to Windows, disable Bitlocker, explicitly let it decrypt the disk • disable secure boot in UEFI After that, you can either boot to Windows (should have secure boot disabled and no Bitlocker - can be checked with HWInfo64) or boot the Image after it was newly written to a stick/SSD. — Reply to this email directly, view it on GitHub<#36 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BLTBRN6PFU2AIWWWKXSYPE33GETD3AVCNFSM6AAAAAB6Z7K57WVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNRRG4ZTMNA>. You are receiving this because you were mentioned.[image]Message ID: ***@***.***>

[jglathe]

Hmm good to know. So in your case, the Samsung driver works for the panel. Does Backlight brightness control work?

[jglathe]

Also, could you please send me the log file you can generate from HWInfo64? This contains the list of devices.

[jglathe]

output of lsusb and dmesg would also be great.

[sandiecao-dc]

Please note: we use the Image of 28 June with the newest dtb we built from GitHub source. It should be same with today’s Image.

From: "Jens ***@***.***> Date: 2025年6月30日 (周一) 20:17 Subject: Re: [jglathe/linux_ms_dev_kit] Lenovo Ideapad Slim 5(x) 14Q8X9 (83HL) (Discussion #36) To: ***@***.***> Cc: ***@***.***>, ***@***.***> This sounds a lot like the Windows partition still being encrypted with Bitlocker, an unholy issue (because its saying its not, probably). But curious, after disabling the copy_firmware.service, did it boot to display? With todays image?

To actually disable Bitlocker:

…

• re-enable secure boot in UEFI • boot to Windows, enable Bitlocker • reboot to Windows, disable Bitlocker, explicitly let it decrypt the disk • disable secure boot in UEFI After that, you can either boot to Windows (should have secure boot disabled and no Bitlocker - can be checked with HWInfo64) or boot the Image after it was newly written to a stick/SSD. — Reply to this email directly, view it on GitHub<#36 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BLTBRN6PFU2AIWWWKXSYPE33GETD3AVCNFSM6AAAAAB6Z7K57WVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNRRG4ZTMNA>. You are receiving this because you were mentioned.[image]Message ID: ***@***.***>

[jglathe]

Hmm okay. This should be the same. The HWInfo log would be nice, though. And a list of what doesn't work :)

[sandiecao-dc]

We can only provide it tomorrow, At least After 12-13 hours.

…

From: "Jens ***@***.***> Date: 2025年6月30日 (周一) 20:52 Subject: Re: [jglathe/linux_ms_dev_kit] Lenovo Ideapad Slim 5(x) 14Q8X9 (83HL) (Discussion #36) To: ***@***.***> Cc: ***@***.***>, ***@***.***> Hmm okay. This should be the same. The HWInfo log would be nice, though. And a list of what doesn't work :) — Reply to this email directly, view it on GitHub<#36 (reply in thread)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/BLTBRNZSRED2EOKV4T75T6L3GEXH5AVCNFSM6AAAAAB6Z7K57WVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTGNRRG43TOMY>. You are receiving this because you were mentioned.[image]Message ID: ***@***.***>

[sandiecao-dc]

Here is the Hwinfo log.
XPLUS Hwinfo.zip

[sandiecao-dc]

The Backlight brightness control doesn't work.
The audio doesn't work.
The type-c can't work for booting.

dmesg.log
lsusb.log

[HarryDo28998]

Do it work to install in Lenovo Ideapad Slim 5(x) 14Q8X9 (83HL)

[jglathe]

@richardnpaul @sandiecao-dc thank you for the logs. Two different display panels, nice :) Also a deviating device name. Anyway, Both logs show a kernel WARN, though, that needs to be investigated. This will also affect things like suspend/resume since the clock tree is not in the desired state, leading to the strangest errors.

Screen backlight brightness: This will be some experimenting to find out.
Spammed dmesg log with systemd messages: probably lost access to the USB stick (broken file system, re-enumerated device). Could be a side effect of the clock WARN we're getting.
No sound: yes, a bit risky to activate that one (no speaker protection). But possible.
Type-C unusable for boot: known issue, unfixable, only by not using it.

If the system is stable enough we should consider installing on internal nvme with dual boot. That's how I drive my devices, works well enough. And upstreaming of the trees ofc.

[richardnpaul]

Okay, so some things to fix up. So on the two different panels front, that's to be expected, there are at least two panels that are available for this model, an IPS LCD and an OLED. My laptop has the OLED panel; one of the two things that drew me to it, the other it being ARM based.

Upstreaming sounds good, as does getting this installed side by side on the NVMe.

[jglathe]

That clock splat might be a side effect of pcie3 (the card reader) not being in the dts. Will add it back.

[richardnpaul]

Feel free to send them over as patch files or whole files (it's been a minute since I've used a patch file but I have used/generated/applied them in the past) rather than rebuild and reuploads.

For the NVMe install it seems that we'll need something different as this img file didn't give you most/nearly all of the options in the installer that you usually get, including partitioning and location to install.

[jglathe]

Well, look at that

It says server, but that doesn't stop you from installing ubuntu-desktop afterwards
I guess we could do with this, will try

[jglathe]

New version of dtb to copy over:
x1p42100-lenovo-ideapad-slim5x.7z.txt

What's changed:

• re-enabled PCIE3, should enable the card reader
• added PWM backlight config of T14s, let's see if that one works

If it comes up, please send the dmesg and function reports.

[jglathe]

Having (at least similar) hardware improves the chances of discovering bugs a lot 😁 I managed to install on my 2in1, but it was very non-straightforward. However, I guess this is doable. Main hurdle is to have efivars, this ensures grub-install can set the bootloader entry . I'll write it up later as a wiki entry.

[davplsm]

Update: tested the ernstp/mesaaco PPA and it works 🚀 Also found efivars, RTC misconfigs (working now). Working on sound.

I've been trying to use kernel 6.16.0-12.12 from the Ubuntu Concept repository for a few days with same dtb, but the panel doesn't work.

[ 1.371286] /soc@0/gpu@3d00000: Unknown GPU revision: 67.3.12.0
[ 1.373492] /soc@0/display-subsystem@ae00000/displayport-controller@aea0000: Fixed dependency cycle(s) with /soc@0/display-subsystem@ae00000/displayport-co
ntroller@aea0000/aux-bus/panel
[ 1.374921] /soc@0/display-subsystem@ae00000/displayport-controller@aea0000/aux-bus/panel: Fixed dependency cycle(s) with /soc@0/display-subsystem@ae00000/displayport-controller@aea0000

[ 1.501228] ------------[ cut here ]------------
[ 1.501230] WARNING: CPU: 2 PID: 278 at drivers/gpu/drm/panel/panel-edp.c:814 generic_edp_panel_probe+0x140/0x290

Do you think it could be related to what you mentioned?
Thanks,

[jglathe]

Huh. Unknown GPU revision sounds interesting. Do you have the complete panel-edp splat? It is probably a different panel. Just got sound to work, will do a new image (I guess ) soon

[davplsm]

dmesg.txt
Attached dmesg using kernel 6.16.0-12-qcom-x1e and the same dtb that works for kernel 6.16.0-rc3-jg-2-qcom-x1e.
Thanks again for your help.

[richardnpaul]

I'll be testing out with the new firmware which was released (v01.50) when I get a bit of time.

I'd like to get an installation up and running at some point but I'll need directions.

This worked fine with the new EC and firmware

[davplsm]

Hi @jglathe,
With this device tree, could the device be added to dtbloader, or is it better to wait until the changes you mentioned have been applied?

[jglathe]

Hi @davplsm,
yes, please do. Since I don't have one ("only" the 2in1), I can't generate the required data. With this the chances drastically increase that the Ubuntu Concept ISO (and other distros) can "just boot" and install.
The current tip of development is sg/ubuntu-x1e80100-6.16-el2. It also contains the changes for RTC and stuff. I'm still fighting a bit with the ATNA40CT01-0 panel, when I feel masochistic 🤪

[Tokor0]

Hello,
I just today got a Slim 5x, and the ISO posted currently in the Google Drive is the closest I've gotten to installing Linux on this machine. However, it did not work. It seems to get very close, but then runs into the error "EFI variables are not supported on this system." I have attached the log from the installation.
log.txt
Thanks a lot!

[jglathe]

Nice laptop, though - I'm impressed with my 2in1

[Tokor0]

Thanks! It works wonderfully now! Only problems are the battery indicator and sound, but I already saw there was some talk about that in that forum.

[Tokor0]

Oh... The fix recommended here broke it. Do you have any working fix for now?

[jglathe]

😬 be careful. This here should do the trick. Put into /usr/local/bin, give execute rights, execute with sudo.

[baozich]

😬 be careful. This here should do the trick. Put into /usr/local/bin, give execute rights, execute with sudo.

Hello! Thanks for all the work. I have Ubuntu installed on my Lenovo Xiaoxin 14s (which is the Chinese model name of Ideapad Slim 5x I think). The only problems for me are also the battery indicator and sound mentioned in this thread. However I have remove my Window partition where the firmware blobs are stored... Is there any other place that I can find those firmware? Thanks.

[janantos]

HI, first I would very much thank you all for you effort, I was waiting for linux support since Feb when I bought Slim 5x. I have yesterday installed ubuntu using last version iso from gdrive, coppied dtb files, edited grub device tree, run script for copying firmware all looks good. However Sound is not detected , I can see in dmesg this

snd-x1e80100 sound: probe with driver snd-x1e80100 failed with error -22

I have read that we have to some degree working sound, however I can't find any docs how to get audio working

on the other note I also see in dmesg following error

[   16.311124] msm_dpu ae01000.display-controller: Direct firmware load for qcom/gen71500_sqe.fw failed with error -2
[   16.311127] msm_dpu ae01000.display-controller: [drm:adreno_request_fw [msm]] *ERROR* failed to load gen71500_sqe.fw

display is working but I cant set backlight, is this the reason? Am I missing some fw files?

Additionaly I have another two errors in dmesg (some usb devices?)

qcom_pmic_glink pmic-glink: Failed to create device link (0x180) with supplier a600000.usb for /pmic-glink/connector@0

and

qcom_pmic_glink pmic-glink: Failed to create device link (0x180) with supplier a800000.usb for /pmic-glink/connector@1

I am running kernel 6.17.0-8-qcom-x1e with this device tree x1p42100-lenovo-ideapad-slim5x.dtb

[jglathe]

Hi, first, congratulations! I am thoroughly impressed by the Ideapad series, and happy with the 2in1 that I have. I got the sound activated and it is working well, though. For setting up sound, there is a wiki entry that can be used as guidance. Since this is an Ubuntu Concept kernel, you need to add the kernel parameter to enable onboard sound. Also, the audioreach topology file is a symlink with the name X1E80100-LENOVO-Ideapad-5-tplg.bin.zst to the Thinkbook 16 variant. The alsa ucm branch jg/master should be good to go, all patches (that I use) are already merged in. All of this is also contained in the KERNEL_PKG partition of the ISO. Beware of the potential damage risk to the speakers, though.

[janantos]

BTW surprisingly fingerprint reader works out of the box

[olav-st]

@janantos Did you have to do something to enable the fingerprint reader? I am not seeing it on my Slim 5x (OLED version)

$ fprintd-enroll 
Impossible to enroll: GDBus.Error:net.reactivated.Fprint.Error.NoSuchDevice: No devices available

[Tokor0]

Hello!

I have been using your Ubuntu for about a month now on my Slim 5x, and it has been fantastic! Thank you very much for sharing your work with us! I am curious though, will the patches you have made to get Ubuntu to work be upstreamed, so as to let other distros also take advantage of them? Or perhaps, is some kind of documentation planned, which could be used to then adapt other distros to work as well? Although Ubuntu is great, it is not my distro of preference, hence my interest.

Many thanks!

[Tokor0]

Fantastic, I got it working by copying the firmware files from my Ubuntu partition. I also posted the firmware files to a repo for future convenience. Thank you!

[Tokor0]

New problems: what kernel modules and/or parameters are needed to get the following working?

1. Audio from the headphone jack (I do not much care about the speakers)
2. Display brightness control.

Especially the brightness control may just be a lack of appropriate control software, but the audio so far seems weird. I cannot find any journal entries related to it (except those related to the speaker topology missing) and I have configured PipeWire seemingly appropriately. Despite connecting headphones, pactl can only find the familiar Dummy Output sink and no audio is output.

[jglathe]

brightness: depends on the panel you have. Additionally, the ATNA40CT01 OLED panel is a b*tch. brightness doesn't work after suspend / resume, still no real idea why - or useful data to change that. Oh and its really bright, so...
For the slim5x, I have created two dtb variants, -oled and -dtb. They are in 6.18.3-jg-0. Maybe worth a try?
As for sound: As I already wrote, bleeding edge. T14s tplg has patches for dp sound that can f things up (like headphone), dmic config is still a mystery, alsa-ucm-conf is a problem in itself with all of this. And I don't understand it, a bit at a loss. So I got speakers working on ideapad 5 2in1 and was happy.

[Tokor0]

Yeah, noticed the OLED brightness thing while using Ubuntu. There the headphone jack worked out of the box though. I will try the kernel in the jg/ubuntu-qcom-x1e-6.18.y branch with the new dtb.

[valpackett]

dp sound that can f things up (like headphone)

wireplumber (or whoever's responsible) doesn't seem to autoswitch the use cases based on things getting plugged in, I got dp sound working on the latitude and I have to manually switch the profiles in pavucontrol…

dmic config is still a mystery

huuuuh?

[davplsm]

Hello,
After configuring the audio, following the wiki guide and the Thinkbook 16's audioreach topology, although the audio works, I notice some distortion, which is noticeable when someone speaks in the videos and sounds like a hiss at the end of the high frequencies of some words.
Has anyone else noticed the same problem?
It sounds fine in Windows, so it's not a speaker problem.

[jglathe]

Hi @davplsm , this is a known issue but @valpackett has worked on this issue and solved it IMO. You could try to use my jg/master branch from my alsa ucm repo, that’s pretty much less distortion. If you want to know more, search for Val and HPH_MODE.

[valpackett]

Specifically, alsa-project/alsa-ucm-conf#595 is the discussion in the ucm repo

[davplsm]

Thank you very much for your help.
Better and more usable now, although there is still some distortion.

[janantos]

Nope, just enabled fingerprint via Ubuntu settings. No kernel modules needed.

…

On Thu, 20. Nov 2025 at 12:19, Olav S. Thoresen ***@***.***> wrote: @janantos <https://github.com/janantos> Did you have to do something to enable the fingerprint reader? I am not seeing it on my Slim 5x (OLED version) $ fprintd-enroll Impossible to enroll: GDBus.Error:net.reactivated.Fprint.Error.NoSuchDevice: No devices available — Reply to this email directly, view it on GitHub <#36 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AETNTFVPAZBBGWIKFZPMKLL35WIRPAVCNFSM6AAAAAB6Z7K57WVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTKMBSGQYDCOA> . You are receiving this because you were mentioned.Message ID: ***@***.*** com>

[jglathe]

6.18 codebase is still a bit of a challenge, Had a few interesting days getting anything x1p up with display. Also suspend/resume behaves differently as far as I can tell for now. I will go into the Ideapad 5 corner in a few days and update what I found out and test if its working there, too. HDMI is the target :) And maybe de-buggering that fp reader.

[jglathe]

FP reader is sort of an oddity though - there is a hardware quirk which needs to be addressed in the qmpphy driver... since fp reader and hdmi bridge are on the same USB controller block, there are side efects which shouldn't be there. fp reader alone works, hot plug of HDMI works (but may kill the fp reader), HDMI cold plug kills the reader for sure. So, it may be partially useful, but for those that need HDMI it is good that it actually always works.

[richardnpaul]

FP reader is sort of an oddity though - there is a hardware quirk which needs to be addressed in the qmpphy driver... since fp reader and hdmi bridge are on the same USB controller block, there are side efects which shouldn't be there. fp reader alone works, hot plug of HDMI works (but may kill the fp reader), HDMI cold plug kills the reader for sure. So, it may be partially useful, but for those that need HDMI it is good that it actually always works.

Wow, the hardware decisions make for crazy work on the software side 😕 is this particularly egregious?

[jglathe]

I don't know, but it is an issue that has not yet been resolved. The whole ip block should have one driver, not 5 (or so), and they are dependent, especially re clock gating and power domain. In this case the hsphy and qmpphy are used independently, but the hsphy seems to require a clock from the qmpphy when both operate, not when hsphy is used alone (like, fp reader only). And the qmpphy driver doesn't do power management by default, for reasons*

[1-Me-Abhi]

6.18.1 kernel is still exhibiting the same behaviour for me where sleep doesn't work and the power light on the side of the chassis stays lit, rather than pulsing when it's asleep, the Wifi appears intermittently (some boots it's working others it isn't; I think you had a sleep somewhere, so I guess that it has a marginal setting where sometimes the sleep is long enough and other times it isn't is my guess if that memory is correct) and the fingerprint scanner doesn't seem to be available. (I saw the 6.18.0 you uploaded the other night and that wasn't in a good state, I'm glad that this one is 😁)

I also think sleep does not work correctly; the keyboard backlight stays on after sleep.

[richardnpaul]

So I've been installing the 25.04 image to an SSD locally. I've used encrypted LUKS which has made things a little "fun" to sort out. So far what I found that I had to do is:

• unlock and mount the encrypted partition and also mount the /boot partition.
  • I copied from the cdrom area of the usb image the x1p42100-lenovo-ideapad-slim5x.dtb to the /boot partition.
• Then I had to edit the /etc/grub.d/10_linux file on the encrypted drive to add a devicetree entry so that grub would find the devicetree in the /boot partition and boot up.
  • I added some extra lines like this:

     261:		devicetree ${rel_dirname}/x1p42100-lenovo-ideapad-slim5x.dtb
     273:		devicetree ${rel_dirname}/x1p42100-lenovo-ideapad-slim5x.dtb
     288:	devicetree ${rel_dirname}/x1p42100-lenovo-ideapad-slim5x.dtb
    

  • I also did this from the usb drive and had to go through mounting with --bind the dev, proc and sys paths so that I could run update-grub, however, a cuter way of avoiding this would have been to edit the grub entry with a devicetree /x1p42100-lenovo-ideapad-slim5x.dtb line and once I decrypted and got into the Ubuntu install I could have just done all the editing and updating of grub from there without all the research of how to generate grub config from another installation with access to the drive, but need to decrypt it first.
• Then I needed to install the qcom-firmware-extract package and then I need to edit that to get it to read the files off of my Windows partition. Luckily the utility is just a pretty flakey shell script which I then modified.
  • First I needed to get the device name which is what the script checks by running tr -d '\0' </proc/device-tree/model which should return Lenovo IdeaPad Slim 5x 14Q8X9.
  • With this I added this to the device_mode case check

                "Lenovo IdeaPad Slim 5x 14Q8X9")
              device_path="x1p42100/LENOVO/83HL"
              ;;
    

  • Then I modified the partition variable because my windows partition was connected via a USB-C port so was an sdx device
    part=$(lsblk -l -o NAME,FSTYPE | grep -E -m 1 "(^(nvme[0-9]n[0-9]p[0-9]{1,2}|sd[a-z][0-9]{1,2})\s+ntfs$)" | cut -d" " -f1)
  • Then I kicked off the extraction and that mostly seemed to work as I got working Wifi afterwards (I think anyway).
    sudo qcom-firmware-extract

So, my outstanding things to address are the sound not working and the battery still showing up with a red explanation mark. Edit: oh, no, I've previously mentioned it but the most egregious issue is that the trackpad is locked to the "Natural scrolling" mode so is the reverse of what I want but selecting either mode does not change the actual scrolling behaviour

[1-Me-Abhi]

this is the output:
WARNING: ... panel-edp.c:814 generic_edp_panel_probe
panel-simple-dp-aux ... Unknown panel CMN 0x1476
panel-simple-dp-aux ... DP AUX backlight is not supported

[jglathe]

Well this could be fun. I need to prepare a different dtb for it.

[1-Me-Abhi]

how much time will it take

[jglathe]

Since its Chimei Innolux, I already had a spat with these 😁 could be that its the normal PWM setup for that type of mainboard. Will prepare one to try

[jglathe]

@1-Me-Abhi you could give the output of cat /sys/devices/platform/soc@0/ae00000.display-subsystem/ae01000.display-controller/drm/card1/card1-eDP-1/edid|edid-decode, though. That would help :)

[1-Me-Abhi]

@jglathe this is the output :
➜ ~ cat /sys/devices/platform/soc@0/ae00000.display-subsystem/ae01000.display-controller/drm/card1/card1-eDP-1/edid | edid-decode

edid-decode (hex):

00 ff ff ff ff ff ff 00 0d ae 76 14 00 00 00 00
1c 21 01 04 a5 1e 13 78 03 ee 95 a3 54 4c 99 26
0f 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 42 3c 80 a0 70 b0 24 40 30 20
a6 00 2d bc 10 00 00 18 00 00 00 fd 00 28 3c 4a
4a 10 01 0a 20 20 20 20 20 20 00 00 00 fe 00 43
4d 4e 0a 20 20 20 20 20 20 20 20 20 00 00 00 fe
00 4e 31 34 30 4a 43 45 2d 45 4c 31 0a 20 00 1b

---

Block 0, Base EDID:
EDID Structure Version & Revision: 1.4
Vendor & Product Identification:
Manufacturer: CMN
Model: 5238
Made in: week 28 of 2023
Basic Display Parameters & Features:
Digital display
Bits per primary color channel: 8
DisplayPort interface
Maximum image size: 30 cm x 19 cm
Gamma: 2.20
Supported color formats: RGB 4:4:4
First detailed timing includes the native pixel format and preferred refresh rate
Display is continuous frequency
Color Characteristics:
Red : 0.6396, 0.3300
Green: 0.2998, 0.5996
Blue : 0.1503, 0.0595
White: 0.3134, 0.3291
Established Timings I & II: none
Standard Timings: none
Detailed Timing Descriptors:
DTD 1: 1920x1200 60.002801 Hz 8:5 74.163 kHz 154.260000 MHz (301 mm x 188 mm)
Hfront 48 Hsync 32 Hback 80 Hpol N
Vfront 10 Vsync 6 Vback 20 Vpol N
Display Range Limits:
Monitor ranges (Bare Limits): 40-60 Hz V, 74-74 kHz H, max dotclock 160 MHz
Alphanumeric Data String: 'CMN'
Alphanumeric Data String: 'N140JCE-EL1'
Checksum: 0x1b

[1-Me-Abhi]

Why the el2 dtb, that sounds odd. do you use slbounce and qebspil? Oterwise, it should be x1p42100-lenovo-ideapad-slim5x-lcd.dtb. flash-kernel should use different paths, though:

jglathe@t14s-jg:/boot$ ll
total 1291988
drwxr-xr-x  5 root root      4096 Jan  8 23:06 ./
drwxr-xr-x 20 root root      4096 Jul 17 09:44 ../
-rw-r--r--  1 root root    357352 Oct  4 15:02 config-6.16-el2-jg-7-qcom-x1e
-rw-r--r--  1 root root    360186 Nov  2 21:40 config-6.17.0-jg-2-qcom-x1e
-rw-r--r--  1 root root    360186 Nov 17 07:23 config-6.17.0-jg-3-qcom-x1e
-rw-r--r--  1 root root    360186 Dec 17 21:53 config-6.17.0-jg-5-qcom-x1e
-rw-r--r--  1 root root    312696 Jan  6 22:57 config-6.18.3-jg-0-qcom-x1e
lrwxrwxrwx  1 root root        73 Jan  8 23:06 dtb -> dtbs/6.18.3-jg-0-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled-el2.dtb
lrwxrwxrwx  1 root root        67 Sep 28 06:46 dtb-6.16.0-12-qcom-x1e -> dtbs/6.16.0-12-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled.dtb
lrwxrwxrwx  1 root root        71 Sep 20 07:19 dtb-6.16-el2-jg-5-qcom-x1e -> dtbs/6.16-el2-jg-5-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled.dtb
lrwxrwxrwx  1 root root        75 Sep 13 23:23 dtb-6.16-el2-jg-5-qcom-x1e-el2 -> dtbs/6.16-el2-jg-5-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled-el2.dtb
lrwxrwxrwx  1 root root        71 Oct  5 19:05 dtb-6.16-el2-jg-7-qcom-x1e -> dtbs/6.16-el2-jg-7-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled.dtb
lrwxrwxrwx  1 root root        75 Oct  5 19:05 dtb-6.16-el2-jg-7-qcom-x1e-el2 -> dtbs/6.16-el2-jg-7-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled-el2.dtb
lrwxrwxrwx  1 root root        60 Oct 14 19:32 dtb-6.17.0-jg-0-qcom-x1e -> dtbs/6.17.0-jg-0/qcom/x1e78100-lenovo-thinkpad-t14s-oled.dtb
lrwxrwxrwx  1 root root        64 Oct 14 19:32 dtb-6.17.0-jg-0-qcom-x1e-el2 -> dtbs/6.17.0-jg-0/qcom/x1e78100-lenovo-thinkpad-t14s-oled-el2.dtb
lrwxrwxrwx  1 root root        73 Nov  3 20:17 dtb-6.17.0-jg-2-qcom-x1e-el2 -> dtbs/6.17.0-jg-2-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled-el2.dtb
lrwxrwxrwx  1 root root        73 Nov 29 17:52 dtb-6.17.0-jg-3-qcom-x1e-el2 -> dtbs/6.17.0-jg-3-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled-el2.dtb
lrwxrwxrwx  1 root root        69 Jan  8 20:38 dtb-6.17.0-jg-5-qcom-x1e -> dtbs/6.17.0-jg-5-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled.dtb
lrwxrwxrwx  1 root root        73 Dec 18 08:04 dtb-6.17.0-jg-5-qcom-x1e-el2 -> dtbs/6.17.0-jg-5-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled-el2.dtb
lrwxrwxrwx  1 root root        73 Jan  8 23:06 dtb-6.18.3-jg-0-qcom-x1e -> dtbs/6.18.3-jg-0-qcom-x1e/qcom/x1e78100-lenovo-thinkpad-t14s-oled-el2.dtb
drwxr-xr-x 11 root root      4096 Jan  8 23:05 dtbs/
drwxr-xr-x  7 root root      4096 Jan  1  1970 efi/
drwxr-xr-x  5 root root      4096 Jan  8 23:09 grub/
lrwxrwxrwx  1 root root        31 Jan  8 23:05 initrd.img -> initrd.img-6.18.3-jg-0-qcom-x1e
-rw-r--r--  1 root root 261718074 Dec 26 23:38 initrd.img-6.16-el2-jg-7-qcom-x1e
-rw-r--r--  1 root root 261776746 Dec 26 23:38 initrd.img-6.17.0-jg-2-qcom-x1e
-rw-r--r--  1 root root 261774152 Dec 26 23:37 initrd.img-6.17.0-jg-3-qcom-x1e
-rw-r--r--  1 root root 261887492 Jan  8 20:38 initrd.img-6.17.0-jg-5-qcom-x1e
-rw-r--r--  1 root root 147175850 Jan  8 23:05 initrd.img-6.18.3-jg-0-qcom-x1e
lrwxrwxrwx  1 root root        31 Jan  8 23:05 initrd.img.old -> initrd.img-6.17.0-jg-5-qcom-x1e
-rw-------  1 root root   7965492 Oct  4 15:02 System.map-6.16-el2-jg-7-qcom-x1e
-rw-------  1 root root   8033918 Nov  2 21:40 System.map-6.17.0-jg-2-qcom-x1e
-rw-------  1 root root   8033910 Nov 17 07:23 System.map-6.17.0-jg-3-qcom-x1e
-rw-------  1 root root   8033987 Dec 17 21:53 System.map-6.17.0-jg-5-qcom-x1e
-rw-------  1 root root   6335781 Jan  6 22:57 System.map-6.18.3-jg-0-qcom-x1e
lrwxrwxrwx  1 root root        28 Jan  8 23:05 vmlinuz -> vmlinuz-6.18.3-jg-0-qcom-x1e
-rw-------  1 root root  13967872 Oct  4 15:02 vmlinuz-6.16-el2-jg-7-qcom-x1e
-rw-------  1 root root  20633088 Nov  2 21:40 vmlinuz-6.17.0-jg-2-qcom-x1e
-rw-------  1 root root  20633600 Nov 17 07:23 vmlinuz-6.17.0-jg-3-qcom-x1e
-rw-------  1 root root  20635648 Dec 17 21:53 vmlinuz-6.17.0-jg-5-qcom-x1e
-rw-------  1 root root  12509696 Jan  6 22:57 vmlinuz-6.18.3-jg-0-qcom-x1e
lrwxrwxrwx  1 root root        28 Jan  8 23:05 vmlinuz.old -> vmlinuz-6.17.0-jg-5-qcom-x1e

As you can see, it does symlinks from /boot to /boot/dtbs//qcom/ to reference the dtb. In this case on my T14s, I actually use slbounce and qebspil, and have a mix of normal and -el2 device trees.

Did you also install the -nostubble variant of the kernel packages?

Yes I installed nostubble variant of kernel

[1-Me-Abhi]

You were right — I’m not using slbounce or qebspil.

/boot/dtb was pointing to x1p42100-lenovo-ideapad-slim5x-lcd-el2.dtb.
I switched it to the non-EL2 DTB:

x1p42100-lenovo-ideapad-slim5x-lcd.dtb

After reboot, the system now reports:

Lenovo IdeaPad Slim 5x 14Q8X9 (LCD)

Panel detection and EDID match the LCD panel (CMN N140JCE-EL1), and everything works as expected.

Thanks — this resolved the issue.

[1-Me-Abhi]

@jglathe Confirmed: with non-EL2 LCD DTB loaded, panel detection is correct (CMN N140JCE-EL1) and system reports (LCD).

/sys/class/backlight only exposes a generic backlight device:

max_brightness = 4095
brightness changes have no effect on panel

This looks like the Slim 5x LCD DTB is missing or not binding a PWM backlight node.

[1-Me-Abhi]

Brightness to change using keys and sysfs

What happens:
Value changes in sysfs but screen brightness does not

Steps to reproduce:

1. echo 100 > /sys/class/backlight/.../brightness
2. No visible change

Logs:
(dmesg / journalctl output)
sudo dmesg grep Ei backlight.txt
sudo journalctl b grep.txt

[1-Me-Abhi]

@jglathe
Hii any update regarding the brightness

[richardnpaul]

So, here are two dmesg outputs, one is when wifi works and the other is when it doesn't work. Looking at these two files side by side in meld makes it look like something strange is happening, maybe in the dtb loading (remember that I don't really know what I'm talking about 😁) as the efi line has a difference for the initrd and the Kernel command line is missing vt.handoff=7 from the end of it. There may be other changes which are apparent to someone more familiar. This is one of the more annoying issues because it will randomly either be not working or working between reboots.
Just a question @jglathe can you (or anyone else, but I suspect that you may be able to fix it or know where to report it as a bug) try out a zoom call and let me know if the camera works for you; in that it should be mirroring back to you? I get a black screen though when I select the camera I can see that the feed is working from the camera (you can make a test call from here click on join and then click on the "Join from your browser" link rather than the blue call to action button). The reason that I know about this is that I decided to have a Zoom call on this laptop and couldn't see the other person or myself 😂
dmesg-no-wifi.txt
dmesg-wifi.txt

[jglathe]

What I can see is that pwrseq initializes before the pcie slots in the "no wifi" case - which is bad, should be the other way round. So, pwrseq needs to defer its probe until the pcie slot is there/up. On the list, sounds interesting and annoying at the same time.

Also there is this:

[ 30.690931] pcieport 0004:00:00.0: AER: Correctable error message received from 0004:00:00.0
[ 30.690948] pcieport 0004:00:00.0: PCIe Bus Error: severity=Correctable, type=Physical Layer, (Receiver ID)
[ 30.690955] pcieport 0004:00:00.0: device [17cb:0111] error status/mask=00000001/0000e000
[ 30.690963] pcieport 0004:00:00.0: [ 0] RxErr (First)
[ 30.743282] pcieport 0004:00:00.0: AER: Uncorrectable (Fatal) error message received from 0004:00:00.0
[ 30.743310] pcieport 0004:00:00.0: PCIe Bus Error: severity=Uncorrectable (Fatal), type=Transaction Layer, (Receiver ID)
[ 30.743318] pcieport 0004:00:00.0: device [17cb:0111] error status/mask=00000020/00400000
[ 30.743325] pcieport 0004:00:00.0: [ 5] SDES (First)
[ 30.743340] pci 0004:01:00.0: AER: can't recover (no error_detected callback)
[ 31.769194] pcieport 0004:00:00.0: broken device, retraining non-functional downstream link at 2.5GT/s
[ 31.876174] pcieport 0004:00:00.0: AER: Root Port link has been reset (0)
[ 31.876313] pcieport 0004:00:00.0: AER: device recovery failed

pcie4 not doing what it should. I took a look into this pwrseq driver a while back, it is sort of important because it handles activation quirks for this hardware.

On another note, adding cma=128M in your boot parameters should resolve those cma errors - which result from ath12k trying to get continuous physical memory ranges for its dma transfers.

[richardnpaul]

What I can see is that pwrseq initializes before the pcie slots in the "no wifi" case - which is bad, should be the other way round. So, pwrseq needs to defer its probe until the pcie slot is there/up. On the list, sounds interesting and annoying at the same time.

So this does fit in with my feeling that it was a race condition.

Edit:

On another note, adding cma=128M in your boot parameters should resolve those cma errors

done, thanks, and they are gone. Should this be a default on this hardware?

[jglathe]

Yes. But with the wcn7850 being standard equipment it may be that this is a reason to put it into the device tree. Had a discussion re cma on the ThinkBook 16, and didn't know that ath12k requires it. Thought it was only some camera stuff that needs it... UVC cameras don't.

[jglathe]

I switched my T14s to 6.18.3-jg-0, and I had the "no WiFi" error for the first time on first boot. So there is a change that makes it more sensitive to this race.

[janantos]

Just be aware, I was stupid enough to confirm in Ubuntu to make upgrade to latest version, unfortunately this made my Ubuntu not booting and ending in black screen and boot reset loop :-( Never drink Glühwein and make system changes 🙃

[jglathe]

Trying to re-activate sound on my 2in1, also kind of purgatory.

[janantos]

Kernel is vmlinuz-6.17.0-8-qcom-x1e but seem it disappeared dtb settings from grub :-)

[janantos]

@jglathe do we have somewhere tutorial for ideapad slim 5x audio activation? Last time I was somehow scared? So living without audio

[jglathe]

Bleeding Edge. I got sound activated on Ideapad 5, but no microphone yet. And lets say this is a clusterfuck. The culprit is - ofcourse - progress. Activating soundwire on video outs seems to be a breaking change.

[jglathe]

USB and Bluetooth audio works, though

[richardnpaul]

Ah, it's annoying to get this far through the build process only to hit a hard coded path issue (also I don't have the stubble hwids):

FileNotFoundError: [Errno 2] No such file or directory: '/home/jglathe/src/stubble/hwids'

Is this something that you could put into another repo?

[richardnpaul]

In the end I found the files locally in /usr/share/stubble/hwids because I'm building from my slim 5x. The other thing that tripped me up was your hardcoded email address passed as the gpg signing key, I think. And zfs wouldn't build, removing that line managed to get things moving further along. Finally, I added a new changelong to, hopefully, get a different kernel build Name so that I can install with "my"* first dts changes.

* I prompted an LLM to resolve a warning that I was seeing in the logs. Hopefully I can work through the errors and warnings until there are far fewer of them.

[richardnpaul]

Hi Jens, can you tell me what options that you use to build your kernels that you put in Google Drive?
I see my builds black screen and then reboot after 30 odd seconds. If I remember correctly this might be what you encounter if you don't have the windows partition available to load the firmware from (I don't because I didn't bring it with me to Brazil), though I should already have the firmware because I've already installed and just want to update the kernel. I don't see this behaviour with your builds so I'm expecting that you pass an option to the build to skip this or some other behaviour?

[jglathe]

Hi @richardnpaul ,

the config I use is in my kernels: zcat /proc/config.gz. This doesn't help with missing patches, though. My tree has the one required for x1-45 GPU. Without, black screen. https://lore.kernel.org/all/20251117-topic-sm8x50-fix-a6xx-non-ifpc-v1-1-e4473cbf5903@linaro.org/

My tree is sort-of current with 6.18.y or 6.18.2-jg-2.

with best regards

Jens

[richardnpaul]

Ah, nice, I've just seen the zfs-2.4 module build which will help with building the 6.18.x kernel. Thanks Jens (Although now I've seen that it has a conflict with the initiramfs package so maybe this isn't ready for primetime)

[jglathe]

I had no real issue here. depending on what is installed, some packages fail. sudo apt install -f removes the unconfigured packages. I have zfs2.40 on some boxes and it works well. What doesn't work is building the kernel with zfs support. The corresponding ubuntu packages are not defined yet. (?)

[janantos]

Good to know.

…

On Sun, 28. Dec 2025 at 21:21, Jens Glathe ***@***.***> wrote: USB and Bluetooth audio works, though — Reply to this email directly, view it on GitHub <#36 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AETNTFQSSUSHUXJACKSO2YT4EAUS7AVCNFSM6AAAAAB6Z7K57WVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTKMZVHAYDGNI> . You are receiving this because you were mentioned.Message ID: ***@***.*** com>

[richardnpaul]

Hmm since installing the 6.18.3 kernel I've suddenly got an issue with the screen in that the logo shows up at expected brightness then dims to what seems like the lowest non-off setting. The back light controls (fn keys) don't allow brightness control and eventually the screen fades to blackness. This includes trying to enter the bios. Booting an old kernel didn't help neither did uninstalling the old kernel and neither again did adding the old windows ssd back in and booting that so I'm guessing some corruption has happened somewhere unrelated to the drive itself but maybe the display firmware? It has unfortunately made the laptop somewhat of a brick where only using hdmi gives it any use.

Edit: Accidentally posted to a thread rather to the main discussion

[richardnpaul]

Hmm your dmesg says it didn't recognize the WiFi card... that's the pcie slot throwing errors. My first analysis still stands: a timing issue with powerseq. I have rebooted quite a few other ath11k and ath12k devices (both with powerseq), but I had only 2 occasions I remember where that happened. Only fix for now is "try again" or "use older kernel".

I went all the way back to booting with 6.17.0 I think, it still showed the same errors. I wondered if the firmware update has introduced a breaking change.

[jglathe]

Actually not impossible. Did you also update the firmware files from the Linux POV by extracting them again, and updating initramfs if they're in there?

[richardnpaul]

I just went through finding out what the dtb name was again as my 6.17.0-jg-3 kernel I had installed didn't work without providing the devicetree (actually this kernel was fine it was the previous kernel from when updated to 25.10 that needed the dtb specifying; no stubble for this hardware). Anyway, I can't update because I can't boot as I hit this issue, to be able to update the files from the windows partition and I'm tired so I can't think how I'd sideload this and regenerate the initramfs for an Arm64 system right now.

Edit:

As it stands this is early enough in the boot process that I don't think that the task to update the firmware files will be run because I've not decrypted the partition yet.

I guess that the approach is to grab the files from the windows drive, then mount and decrypt the partition from the Linux install, copy the firmware across and then do an update-initramfs for the offline drive

[jglathe]

Sounds like a plan. Boot from an installer ISO. But its a PITA with the chroot part to get a valid initramfs. I woul probably try several times to boot the existing system until it succeeds once, then do the update stuff from inside the system.

[richardnpaul]

Unfortunately I have tried this, trying to get it to boot, like some 10 or 20+ times with no progress. I've tried with pci=noaer but still not got any further in the process.

[digidude512]

nostubble kernel images seems to stop me from booting.
Should I be using the normal ones instead or have I just got something missing/misconfigured?

[richardnpaul]

Nostubble means that you have manually define the dtb devicetree file that you want to load in grub so normally you don't want to manually have to configure that and that's what stubble manages for you

[digidude512]

Ah I see, How can I configure that? I want to see if 6.18.4-jg-0 fixes a new bug I've got.

[richardnpaul]

When I was doing it I couldn't get it working any other way other than when grub loads, edit, navigate to the bottom and add a devicetree /linux-lenovo-ideapad-slim5x.dtb line, or something like that. I put the file directly into /boot and found that I didn't have to include /boot in the location (in fact I think that it didn't work with /boot/.. as the path but that might just be a hazy memory). There is a way to set this without having to do it on every boot but I didn't figure out how to get that working.
(Sorry about the delay in responding)

[jglathe]

6.18.4-jg-0 should boot without any additional devicetree statement. It should have a working stubble/hwid/dtb combination built-in.