Kernel 5.16.1 resume from suspend fails; 5.15.x works fine

Support Kernel
21

What I mean is that whatever broke suspend in 5.16 will likely eventually be ported to 5.15.

1 Like
22

That would suck.

23

Another experiment: new partition, same drive, fresh install of manjaro w/ 5.16 kernel. suspend and resume work fine. So there’s some gremlin in my main work partition. So, I’ll eventually migrate my apps to the fresh install if I have to.

24

Such a thing did in fact happen. It’s resolved in 5.17-rc7, and hopefully the fix will again be backported to 5.15 and 5.16. Currently, the latest 5.15 and 5.16 kernels (via Manjaro’s stable updates) have inherited this nasty backported bug.

Those on a fresh install of Manjaro (from ISO) or who haven’t yet updated, are spared from the s2idle / suspend bug.

1 Like
25

Migrated and done with this bug. I’m marking it answered, as it now seems clear that for me, at least, the problem was due to something in my old installation that couldn’t cope with 5.16. The fresh install suspends and resumes fine on both 5.15 and 5.16. My old work partition was showing other problems too. I’m keeping it around in case I want to bang my head against that wall some more.

26

Fresh install plus updates?

What is the current kernel you’re running on precisely?

uname -r
27

Fresh install all stable updates, kernel 5.16.14-1-MANJARO

28

Looks like you’re on the newer kernels then (which have the fix.) :+1:

The Stable Updates were released recently.

:white_check_mark: Fresh install? You’re safe.

:white_check_mark: Fully updated with latest kernels? You’re safe. (5.15.28 and 5.16.14)

:warning: Those affected were “caught in between”.

They were on updated packages, which pulled in kernels that had the backported “fix” (really, “bug”) that broke suspend for some laptop users. However, they weren’t updated enough, since the Manjaro Stable Updates hadn’t yet been released which contain the fixed kernels. (5.15.28 and 5.16.14 )

29

well, the work machine was on unstable and had the problem with every 5.16 kernel , and the new machine didn’t have the problem with any 5.16 kernel I used (11,12, now 14). I don’t believe the broken machine will be fixed by the new kernel, but will test it when I get time.

30

That would be interesting to see. For the other machine, update it so that it uses 5.15.28+ or 5.16.14+, and then reboot. (Or even 5.17-rc7)

If not, then you might have been facing yet another different bug. :hushed:

31

all 5.15 kernels work fine on the “broken” machine including 5.15.28.

32

So the broken machine isn’t really “broken” anymore, right? :slightly_smiling_face:

33

The physical machine is fine. The ‘work’ system on a separate partition was initially broken only in that suspend/resume didn’t work with kernel 5.16.x (but worked fine with 5.15.x). Then other issues started appearing beyond the 5.16.x suspend problem, e.g., oxygen window decorations broke shade/unshade, then Firefox started showing blank pages on some web sites. I don’t know what I did to break it (I installed several applications from source code, like the python3 operable zenmap), and decided to give up trying to fix it. The ‘broken’ ‘work’ system is now on a partition that I only mount when I need to get some old data.

I installed a fresh version of Manjaro on another partition of the same nvme drive, reinstalled all the apps I need and everything works fine. I haven’t tested the ‘broken’ system since I finished setting up the new one.

34

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.