đŸ’Ÿ Archived View for republic.circumlunar.space â€ș users â€ș flexibeast â€ș gemlog â€ș 2020-09-18.gmi captured on 2023-03-20 at 18:44:33. Gemini links have been rewritten to link to archived content

View Raw

More Information

âŹ…ïž Previous capture (2022-07-16)

âžĄïž Next capture (2023-05-24)

-=-=-=-=-=-=-

Recent Linux kernel issues

i've been using Linux since 1998, when i started out on Red Hat 5.2 (apparently codenamed ‘Apollo’). That came with version 2.0.36 of the Linux kernel. Since then, i've found the Linux kernel to be pretty stable - in general, there haven't been many problems i've encountered which were (or which seemed to be) an issue with the kernel specifically, rather than with something higher up the stack (X, PEBKAC).

That seems like it might be changing.

One kernel in the recent 5.7 series resulted in a panic at boot. i honestly don't remember the last time i saw a kernel panic. The error was:

VFS: Unable to mount root fs on unknown-block(0,0)

One user had found this to be caused by a full ‘/boot’; my own ‘/boot’, however, was nowhere near full. Booting from a previous version in the 5.7 series worked without problems.

In the last couple of days i started facing odd wifi issues - the connection would suddenly stop working, even though my laptop was still associated with the AP. i tried restarting wpa_supplicant and dhcpcd - nope, after a few hours, the connection would drop again. Okay, well, there was an issue reported on the OpenBSD ports-changes list about issues after upgrading to dhcpcd 9.2.0, so i tried downgrading to 9.1.4 - nope, again, after a few hours, the connection would drop. Each time, the output of ‘ip a’ suggested that all was in order, but trying to do things like ‘ip link set wlp2s0 down’ would result in “RTNETLINK answers: Input/output error”.

(This is a sterling example of why a successful kernel boot is not an indication that it's ‘working’ - problems might take hours or days to appear.)

Last night, over 12 hours ago, i booted from 5.8.8 rather than 5.8.9. Lo and behold: no more connectivity issues since.

The thing is, i consciously make a choice to _not_ use the latest hardware. i generally don't need high-powered CPUs or GPUs. i don't tend to manually compile software written in languages that are CPU-intensive (e.g. C++, Rust); in most cases, i use precompiled binaries. i'm basically not a gamer; i mainly play roguelikes, which i love, so i have no need for increasingly powerful graphics. i don't have the income to run on the constant-upgrade treadmill; and even if i did, i wouldn't be enthusiastic about continually moving to new hardware, for environmental/ecological reasons.

So i buy consumer hardware that's discounted due to no longer being the latest line. Not only is it cheaper, but it's more likely that its hardware is common enough, and has been around long enough, that it will be sufficiently well supported by the Linux kernel. i've had my current laptop for a few years now; the wifi adaptor in it is a QCA9377. Until now, i've not had any problems with it. (Though more generally, the laptop _does_ continually re-block the WLAN after each reboot, such that i need to have ‘rfkill unblock wlan’ in ‘/etc/rc.local’.)

i've seen a few comments suggesting that it feels like the kernel is becoming more flaky. The sample size is obviously way too small, but given my overall experiences compared to my recent experiences, i'm starting to wonder ....

--

đŸ· ict

Glossary

Gemlog Home