[Unstable Update] 2021-11-11 - Gnome 41.1, Xorg-Server 21.1, LxQt 1.0, Nvidia 470.86, Plasma 5.23.3

Linux 5.16-rc1 Released With Intel AMX, FUTEX2, Folios and I really cant wait to test it …

I know you are busy fixing other things , have a nice day :slight_smile:

1 Like

@wolfyrion
I think the target thread for the 5.16rc1 kernel request is the Update requests for packages built by Manjaro developers
Please post there and remove here.

I realized awful behavior of the

linux516
5.16rc1.211115.gfa55b7d-1

It present in safely removal of USB stick drives: may “PC was hanged” 2 times while that operation: operation did not complete (first time ejecting via GUI menu, second time with udisksctl power-off (after successful sudo eject sda). All other components, GUI works fine, but in order to get safely remove I call PC reboot via GUI and reboot procedure it hangs for several minutes. Hard reset was then.
Second time while udisksctl power-off hangs with no answer I switch to tty and systemctl reboot and it semi-hangs for 20 minutes: kernel periodicaly shows messages about hanged component.
Both times I need to hard reboot PC.

After first hard reset I lost history file of zsh: I have cleaned history now.

Like logs? ok: here is `journalctl`

First hang

~/Desktop ❯ journalctl -b-5 -p3 | grep -i "dumped core"
Nov 16 22:54:33 pc systemd-coredump[24975]: Process 805 (Xorg) of user 0 dumped core.
Nov 16 22:54:35 pc systemd-coredump[25082]: Process 909 (kwin_x11) of user 1000 dumped core.
Nov 16 22:54:36 pc systemd-coredump[25105]: Process 1421 (opera) of user 1000 dumped core.
~/Desktop ❯      

Second hang:

Nov 17 11:05:06 pc kernel: INFO: task pool-udisksd:4995 blocked for more than 860 seconds.
Nov 17 11:05:06 pc kernel:       Not tainted 5.16.0-1-MANJARO #1
Nov 17 11:05:06 pc kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Nov 17 11:07:09 pc kernel: INFO: task pool-udisksd:4995 blocked for more than 983 seconds.
Nov 17 11:07:09 pc kernel:       Not tainted 5.16.0-1-MANJARO #1
Nov 17 11:07:09 pc kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Nov 17 11:09:11 pc kernel: INFO: task pool-udisksd:4995 blocked for more than 1105 seconds.
Nov 17 11:09:11 pc kernel:       Not tainted 5.16.0-1-MANJARO #1
Nov 17 11:09:11 pc kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
Nov 17 11:11:14 pc kernel: INFO: task pool-udisksd:4995 blocked for more than 1228 seconds.
Nov 17 11:11:14 pc kernel:       Not tainted 5.16.0-1-MANJARO #1
Nov 17 11:11:14 pc kernel: "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.

The purpose of that post:

  1. Use the 5.16rc1 kernel version only of you need it. Wait for the rc2 at least.
  2. Do not propagate the rc1 into testing and stable “downstream” branches.

Thanks for the warning.

I can’t help it, but this comes to mind:

  1. Nobody needs an rc kernel
  2. IMHO it does not matter if this goes to testing - it is an rc and not a stable kernel. People should not use those unless they want to find and report bugs to https://bugzilla.kernel.org/.
3 Likes

I use RCs since about 5.13rcX. Besides that I remember some unstable behavior only once in rc5, but that was so nasty, that I skipped rc6 and switch back only when rc7 released.

I can submit them, but currently I have 5+ bug report to submit: I constantly postponed them previously, so I did not went deep to figure out what exactly the problem in rc1, but all 2 times it was for USB stick removal. Before the 5.16rc1 and now on 5.15th gen everything works fine, but I reinstalled all packages (that was about 2.7 GiB to download) in order to fix possibly damaged executable files while I did that hard reset.
As I said I lost at least zsh history file, still do not know what else config files and my user file I lost/damage. Moreover it could be much more dangerous for me: I run so-called self-healed BTRFS but with LUKS. Any LUKS header damage and my recent data lost. Probably now I will start to backup LUKS header also, but the header is required to be in super-private place only.

We need LUKS2 with it’s ability to backup the LUKS header copy deeper inside of disk/partition - as second / backup copy.

Hi, sparky!
Welcome to community!
Thank you for your feedback, I use PC without any sleep/suspending/hibernation (do not stop disks, network, etc…) for years except:

to turn off monitor

but may be I have some deeper settings I am not aware of.

Has anybody had issues recently with kernel 5.14.19-1? :thinking:

My system won’t be able to boot the above kernel. It stops booting right after reaching Switch root target.

I’m using systemd-based initramfs. I’m on systemd version 249.6-3
My rootfs is LVM-based.

On the other hand, my system boots fine with kernels 5.15.2-2 and 5.10.80-1

This is not big deal for me as I always use the latest stable kernel and LTS as fallback, but it might be for someone who decides to stick with 5.14 for now.

I’m not dumping kernel 5.14 yet, just let me know if you guys would like to troubleshoot this and need more information from me.

2 Likes

I haven’t faced this probably because yesterday I noticed that 5.14.20 had already been published on kernel.org, but Manjaro’s repos didn’t get it in time so I just built and installed 5.14.20 myself.
Last time I checked it still wasn’t available in repos…
I also have LVM + underlying LUKS2 so I usually pay extra attention to everything as low level as kernel / systemd / so on. Kinda exhausting but fun when nothing breaks too hard :wink:

1 Like

On second thought, 5.14 will become EOL in a near future (as soon as 5.16 gets officially released).

The kernel maintainers are treating 5.15 as LTS with an anticipated EOL in 2023. (See releases section at kernel.org).

So, at least for my use case, I don’t see the point in keeping 5.14 around. It just adds maintenance complexity for everyone.

3 Likes

5.14 is a stable kernel and for some hardware 5.15 is currently not working (well) - it still has it’s use.

1 Like

Thing is, 5.14 hasn’t been dropped by upstream yet, and it has gone a nice round of bug fixes which 5.15 has only just begun receiving. It will be around for a while until 5.15 is deemed stable enough, that’s how things are usually done by Greg KH. My bet for EOLing 5.14 is when 5.15 will get 5th patch set or so.
PS: this is not that type of complexity that is overwhelming. It’s of the one that offers more flexibility. But of course the easiest way is to simply run one and only 5.10, although it’s too boring.

2 Likes

Or you can have just latest stable and LTS like our upstream folks do :wink:

Anyway, I don’t intend to start a controversy here.

I’d be glad to see if 5.14.20 gets booted up on my system, but at the end of the day, we all will have to make the move to 5.15 once 5.14 is declared EOL.

1 Like

I just installed it. I can confirm my system boots up this kernel flawlessly. :slightly_smiling_face:

Time to move on! :laughing:

I have the Transcend JetFlash 8GB USB 2.0 stick drive (it’s release date was 10+ years ago)

2 posts were merged into an existing topic: [Stable Update] 2021-11-19 - Kernels, Gnome 41.1, Plasma 5.23.3, Frameworks 5.88, LxQt 1.0, Xorg-Server 21.1, Mesa