Periodic 17.4 (32 bit) crash with kernel 5.8.16, not 4.9.240

[Member DaveW]

With Antix 17.4 (32 bit) on Dell D620, and kernel 4.9.240, everything is super stable. I can’t remember a crash of any kind.

For recent experiments of wireguard vpn, I installed kernel 5.8.16 (also tried the previous 5.8 kernel). Boot up time is slightly longer with the 5.8 kernels, but everything is stable, and wireguard works fine. Except…

With the 5.8 kernel, the boot process usually hangs, after the third re-boot (sometimes fourth, always by fifth). So, after three event free sessions, the fourth boot hangs up midway… If it does not hang, then shortly after the desktop appears, the computer freezes (with varying symptoms: usually frozen mouse and keyboard, sometimes with display corruption).

In either failure mode (boot hang or subsequent freeze), the cure is to reboot with the 4.9 kernel. After that, the 5.8 kernel will reboot and run for the next three sessions.

I don’t want to waste troubleshooting time on finding a fix (because a reboot with 4.9 kernel always fixes it).

So, I’m just wondering if this makes sense to anyone?

[Forum Admin rokytnji]

Helpful

Up

0

::

Makes perfect sense to me.

My panasonic C F-48 from early 2000 is picky on kernel boot up.

Regression is just a fact. I be just glad to get to a desktop.

My 32 bit atom netbooks never need need the latest and greatest kernel to run well also. Got me wondering now if those old custom netbook kernels are still floating around in the internet. Hmmmmm.

In fact. I have read threads over the years where something had to be remastered to get hardware support for dropped hardware in a newer kernel install.

So. Be glad you have a backup. I’d uninstall 5.8. Play around with othewr kernel choices since I prefer 2 boot able kernels.
One is backup for the other. Like you just went througho

Sometimes I drive a crooked road to get my mind straight.
Not all who Wander are Lost.
I'm not outa place. I'm from outer space.

Linux Registered User # 475019
How to Search for AntiX solutions to your problems

[Member andyprough]

Helpful

Up

0

::

I had a system recently that was creating log file entries related to my wifi card at a furious pace when I would use a 5.8 kernel. Eventually the system would freeze up as it seemingly created hundreds of log entries per minute. I finally tracked down a little power utility called tlp by which I could control this behavior. Don’t recall the commands I had to give it, but tlp had a website with extensive documentation that I used to figure it out. If I recall correctly, the system was constantly checking USB power to the wifi card, if that makes sense, when it did not need to. I think I told tlp to only check the power state to that device at shutdown, and that solved it.

[Member DaveW]

Helpful

Up

0

::

rokytnji – Well… That’s basically where I expected this would lead… and also the way I’ve been dealing with it. My only reason for using the 5.8 kernel, is that wireguard is already part of the kernel. So, when I need that, that’s the go to kernel. Otherwise, for now, 4.9 is best. I may try some other kernels, too. Thanks for sharing your experience.

andyprough – Thank you also for sharing your experience. It looks like tlp is already installed (cli version, no GUI) and apparently running in the default mode. However, I suppose, it is not active during the boot process (where the problem usually shows up). Perhaps it might help prevent a crash, after the system is up. I’ll keep that in mind. Thanks!

[Member andyprough]

Helpful

Up

0

::

You can configure tlp by editing /etc/tlp.conf. It’s a good idea to read the tlp documentation prior to messing with it: https://linrunner.de/tlp/

I’m looking through my tlp.conf now, looks like tlp has an “autosuspend usb” function that was creating havoc with my wifi card when using the 5.8 kernel, and I located the device ID and blacklisted it from tlp’s autosuspend. The tlp documentation does say that usb autosuspend can cause these types of problems. The usb blacklist option is under the “USB” section in tlp.conf, and the documentation is online at https://linrunner.de/tlp/settings/usb.html. Look in /var/log/boot.log, and if you find you have a device that is misbehaving in this way, look up the device ID with the ‘tlp-stat -u’ command (or with ‘lsusb’) and then try blacklisting it from usb autosuspend in tlpui.

There is also a gui called tlpui you can install from github to configure tlp: https://github.com/d4nj1/TLPUI
It’s a python3 package and has very simple build and install instructions.

• This reply was modified 2 years, 4 months ago by andyprough.

[Member DaveW]

Helpful

Up

0

::

andyprough,
It looks like your suggested cure has done the job. After disabling usb autosuspend (in the tlp.conf file), I have rebooted a dozens times, with 5.8 kernel, without a problem. Boot-up, with 5.8 kernel, is also quicker than it was previously.

By the way, on my system, the .conf file is /etc/default/tlp (without the .conf). The configuration file is shown near the top of the page when you run “tlp stat” from the command line.

For my purposes, tlp can be managed from the .conf file and/or cli. So, I’ll forego the GUI for the time being.

Thank you for sharing this!

• This reply was modified 2 years, 4 months ago by DaveW. Reason: added another line of information

[Author]

Posts