6.17.10 broke my standby (Lenovo Yoga 7, s2idle)

I have a Lenovo Yoga 7 2-in-1 16APK10 AMD convertible. It only supports s2idle for standby (you can set S3 in the BIOS).

On kernel 6.17.9 resuming from standby works great, on 6.17.10 it doesn’t resume.

I just got my X1 2-in-1 Gen 10 today and noticed that I only have s2idle as well. Kind of a big problem, especially since I read reviews on this unit and sleep was said to be working and that’s why I bought it.

Same for me. Going back to 6.17.9 (+ Pluton Security Processor being enabled) makes the resume from standby working at least (but it takes about 10s most of the times to resume).

Really hope there will be some fixes in the nearby future (since I don’t expect Lenovo to update the BIOS & enable an option for S3).

Same issue on 6.17.11.

It’s probably even suspending issue. Usually my powerbutton led blinks when the device is in standby, but it doesn’t on 6.17.10 & 6.17.11.

In case this is usefull:

journalctl -b | grep suspend

returns:

Low-power S0 idle used by default for system suspend
nvme 0000:01:00.0: platform quirk: setting simple suspend

Turns out in my case I just know nothing about owning modern computers. s2idle is the only mode supported by my laptop, but now that I’ve had a chance to properly test it, it looses less than 1%/hr while asleep. The light blinks as expected while sleeping.

Which kernel version do you use (uname -r)?

For me only 6.17.9 is working, on 6.17.10+ s2idle standby is broken.

I have a Yoga 16AKP10. I have found that the system takes a while to wake from sleep when on battery. I’ve seen something about “deep sleep”. Have you let it go for a while to see if it wakes? I’m finding it takes 45-60seconds sometimes to wake.

Also, it didn’t used to take this long, but started recently. Maybe with Kernel 6.17? Not sure exactly when it happened, but it’s only been the last month or so?

Just wanted to add another data point: HP G1a (AMD Strix Halo APU) - the same problem on 6.17.10, the laptop won’t wake up. This laptop also supports only s2idle afaik.

Yes, it really takes long to resume, even on 6.17.9 for me, when the hardware is in the “deepest sleep state”. Tbf it saves a lot of power (it only consumes 1% of battery in ~5h on my device), but I would prefer faster resume over powersaving (for my usecase at least).

On 6.17.10+ my 16AKP10 isn’t going to standby properly at all. Usually the led on my powerbutton blinks when the device entered the standby, but on 6.17.10+ it stays on permanentely.

I’m currently trying to figure out which changes could have caused the issue: Making sure you're not a bot!

It’s really unfortunate that there is nothing in the logs on my system. Do you get any informations from your logs?

Here are a couple of differences in the kernel logs (journalctl -k) between 6.17.10 and 6.17.1 that caught my eye:

  1. logs for kernel 6.17.10 end with fedora kernel: PM: suspend entry (s2idle) and then goes nothing. I opened the lid of my laptop and stared at the reflection of my mug for quite some time, so there should have been something logged, but there was nothing, the system got frozen with no details.
  2. kernel 6.17.10 reports the following lines that 6.17.1 doesn’t:
kernel: ACPI BIOS Error (bug): Could not resolve symbol [^PMF.ATST], AE_NOT_FOUND (20250404/psargs-332)
fedora kernel: ACPI Error: Aborting method \_SB.SCBI due to previous error (AE_NOT_FOUND) (20250404/psparse-529)
fedora kernel: amdgpu 0000:c3:00.0: amdgpu: Register(0) [regVPEC_QUEUE_RESET_REQ_6_1_1] failed to reach value 0x00000000 != 0x00000001n
fedora kernel: amdgpu 0000:c3:00.0: amdgpu: VPE queue reset failed
fedora kernel: amdgpu 0000:c3:00.0: [drm:amdgpu_ib_ring_tests [amdgpu]] *ERROR* IB test failed on vpe (-110).
fedora kernel: amdgpu 0000:c3:00.0: amdgpu: ib ring test failed (-110).

Plmk if there is anything of interest or whether I could look deeper. I’m not very proficient with Linux, though.

1 Like

You’re right, this does sound very similar to my issue: Kernel issue?: Fresh install of Fedora 43 hard locks when Thunderbolt 4 Dock is (dis)connected during sleep

As time goes by, I think it’s sleep in general, and not necessarily the dock, though the dock definitely exacerbates the issue.

I’ll try x.1 and report back.

Interessting, thanks for sharing!

I opened a bug report for the kernel: Making sure you're not a bot!

Maybe you could add your logs there, please?

I’ll try to follow this comment and run bisection tomorrow. Not sure if just attaching logs is what people there expect (never posted anything in the kernel space, so I might be wrong :slight_smile: ).

Btw, have you tried newer kernels, like 6.18.1 or 6.17.12?

1 Like

FYI: just tried 6.18.1 - still broken, my laptop doesn’t wake up.

That’s sad to hear, thanks for testing!

Can I manually keep my working kernel 6.17.9 installed somehow? I can’t update at the moment because of it.

I guess it should be possible to update all software, except for the kernel. But I’m too lazy, I just let it install everything, then removed the new 6.17.12 kernel that doesn’t work for me anyway. Fedora keeps 3 kernels by default, but you can extend that limit. I just removed unused kernels before the update (I assume you don’t want 6.17.10 and 11 hanging around). List installed kernel packages with sudo dnf list --installed | grep kernel, then remove unused ones (e.g. sudo dnf remove kernel-core-6.17.12).

Btw, couldn’t get to bisecting today, so if you have time for that, please do.

The currently running kernel will never be removed.

So if you are running 6.17.9, and you also have 6.17.10 and 6.17.11 installed - then the update will remove 6.17.10 and install 6.17.12, leaving 6.17.9 untouched.

Stumbled upon a discussion of the same problem in the Framework community: Wake from Suspend? - #21 by Jean-Marc_Le_Roux - Framework Desktop - Framework Community . Looks like we just need to wait until this kernel patch lands, sometime in 6.18.

1 Like