System reboots ~10 seconds after I try to suspend/hibernate since 6.16

alba4k · 2026-03-15 12:33:06

Sorry for bothering, but would it suffice to start with https://gitlab.archlinux.org/archlinux/ … 886aa4650d (which is the commit containing the config you linked) and building by simply running

makepkg -efi

repeatedly, alternated to the git bisect steps?

alba4k · 2026-03-15 12:59:55

Nevermind, using the PKGBUILD from linux-git and changing the config will probably be better

Last edited by alba4k (2026-03-15 13:17:58)

seth · 2026-03-15 13:37:29

Because of your first question it is important to understand that you must *not* bisect the AUR package but the kernel tree it packages.
linux-git is the mainline kernel w/o patches Greg or the archlinux fork might have applied, it will not have any versioned tags and the commits I posted will likely not work.

It is not any kernel you've been using so far but if you can isolate an offending commit and test to revert that from https://github.com/archlinux/linux to confirm it's the culprit that's a perfectly serviceble approach.

alba4k · 2026-03-15 14:07:29

I'm only using the AUR package because of the PKGBUILD, which I've slightly modified, it's just to build. I am trying to build https://github.com/archlinux/linux (using the modified config and AUR PKGBUILD). That's why I mentioned using `makepkg -e`, my plan was to just have the PKGBUILD clone the archlinux fork and the run the bisect in linux-git/src/linux

So now all'ive done is let `makepkg` clone, go into the cloned dir, start the bisect

I'm a bit worried about the changes done by make olddefconfig every time while building

Last edited by alba4k (2026-03-15 14:32:34)

alba4k · 2026-03-17 08:41:24

I just tried booting a recent arch iso (I have no clue why I didn't before, I was sure I had tried that) and it was able to maintain suspension just fine. That seems to indicate that the kernel alone is probably not the issue.

While a bisection might still be useful, do you think there's something more time effective I could try? especially since I seem to be unable to build the kernel this time ?

libbpf.c: In function ‘kallsyms_cb’:
libbpf.c:8184:13: error: assignment discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
 8184 |         res = strstr(sym_name, ".llvm.");
      |             ^
libbpf.c: In function ‘avail_kallsyms_cb’:
libbpf.c:11487:31: error: assignment discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
11487 |                 if (!(sym_sfx = strstr(sym_name, ".llvm.")))
      |                               ^
libbpf.c: In function ‘resolve_full_path’:
libbpf.c:12075:35: error: assignment discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
12075 |                         next_path = strchr(s, ':');
      |                                   ^
cc1: all warnings being treated as errors

seth · 2026-03-17 14:05:31

Do you have an updated journal (previous ones turned 404)

For userspace triggers your GUI session, anything nvidia (?) and also NetworkManager (try to stop that before the sleep) might be concerns.
But you're also not relying on the same root device (nvme? => https://wiki.archlinux.org/title/Solid_ … ST_support )

alba4k · 2026-03-17 21:05:51

Here you go: https://0x0.st/P9Z4.txt

You can see the system reboot after going into suspension at "mar 17 21:54:35"

And that was without NetworkManager, cups, all gui sessions not running. bluetooth wasn't stopping but disabling that made no difference either.

I only have an intel integrated gpu (i7-1260P), so no nvidia involved.

But you're also not relying on the same root device

What do you mean? I do have an nvme drive, but the reccommended kernel args do not seem to help.

Last edited by alba4k (2026-03-17 22:31:28)

seth · 2026-03-17 22:11:08

What do you mean?

The life system runs off a usb key, the installed system relies on the nvme, luks and apparently TPM?

Another thing is

mar 17 09:30:38 dell-xps sudo[20743]:   alba4k : TTY=pts/0 ; PWD=/home/alba4k ; USER=root ; COMMAND=/usr/bin/efibootmgr --create --disk /dev/nvme0n1 --part 3 --label 'Windows Boot Manager' --loader \\EFI\\Microsoft\\Boot\\bootmgfw.efi

3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.
A hibernating windows triggering the reboot is way more likely than anything else here - the kernel version might just be a fluke/red herring and the ACPI will typically distinguish between boot devices (ie. when you boot from usb the running hibernation and scheduled acpi signals are ignored)

alba4k · 2026-03-17 22:30:26

Windows is completely unrelated. I installed it a few weeks ago because of Safe Exam Browser (long after this issue started) and this laptop never had Windows on it before that. That command was just because I noticed that the Windows boot entry had disappeared (and interestingly didn't come back with that, I just added it back from the bios)

The life system runs off a usb key, the installed system relies on the nvme, luks and apparently TPM?

Oh, I see. Yeah exactly, I have arch installed on btrfs, inside Luks (unlocked by tpm, secure boot is on). Do you think any of those might interfere with suspension?

but the kernel version is most definitely not just a fluke, as going back to anything older than 6.16 still fixes the issue. Of course, it can't be the only factor as the february ISO (runninf 6.18 iirc) suspends just fine

Related to the timing the reboot: while on suspension it seems quite random (~10s), when trying to hibernate it happens to be almost exactly a split second after the system appears to have completely shut down (i.e. I see the touchpad backlight turning off)

Last edited by alba4k (2026-03-17 22:34:09)

alba4k · 2026-03-18 10:04:04

Small update.. I sort of f-ed up

I started wondering if the TPM was maybe not liking something (as we mentioned it here) so I tried to blacklist those modules (thinking that worst-case I could just lift that blacklist

Turns out, my system is now refusing to mount that partition, even after removing it, for some reason

https://bbs.archlinux.org/viewtopic.php?id=312695

edit: solved

Last edited by alba4k (2026-04-01 21:59:40)

alba4k · 2026-04-01 22:00:38

probably not too important, but I noticed that if I wake up the system just before it was going to reboot, it sort of wakes up (i.e. the screen turns on) but then reboots shortly after

alba4k · 2026-04-05 19:25:55

I guess I'll try to stick with linux-lts612 for the time being...

seth · 2026-04-05 19:35:00

alba4k wrote:

going back to anything older than 6.16 still fixes the issue. Of course, it can't be the only factor as the february ISO (running 6.18 iirc) suspends just fine

alba4k wrote:

a bisection might still be useful

Compare

pacman -Qs firmware
lsmod

between iso and booted system - both systems are entering S3? You're not comparing that to s2idle or so?

alba4k · 2026-04-05 19:55:52

I only have s2idle avaiable (on the booted system at least). from what I can find online, the i7 1260P I have doesn't even seem to support S3 sleep

anyway, I uploaded the initial outputs here: https://github.com/alba4k/temporary-throwaway

I will now go get a live usb and update this comment once I've added those and the diffs

edit: done. they both look quite different.

Last edited by alba4k (2026-04-05 20:05:24)

Arch Linux

#26 2026-03-15 12:33:06

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#27 2026-03-15 12:59:55

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#28 2026-03-15 13:37:29

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#29 2026-03-15 14:07:29

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#30 2026-03-17 08:41:24

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#31 2026-03-17 14:05:31

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#32 2026-03-17 21:05:51

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#33 2026-03-17 22:11:08

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#34 2026-03-17 22:30:26

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#35 2026-03-18 10:04:04

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#36 2026-04-01 22:00:38

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#37 2026-04-05 19:25:55

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#38 2026-04-05 19:35:00

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

#39 2026-04-05 19:55:52

Re: System reboots ~10 seconds after I try to suspend/hibernate since 6.16

Board footer