@kernellogger@fosstodon.org
@kernellogger@fosstodon.org avatar

kernellogger

@kernellogger@fosstodon.org

Mainly tooting about #Linux the #kernel and things related to the #LinuxKernel – e.g. #bootloader, #compiler, #git, #glibc, #mesa, #qemu, #xorg, #X11, #wayland, and other stuff in the 'plumbing' layer.

Opinions are my own.

Topic account. Other accounts of mine:

https://social.linux.pizza/@knurd42 (EN): #FLOSS, #Fedora as well as Life, the Universe and Everything
https://norden.social/@thleemhuis (DE): Das Leben, das Universum und der ganze Rest
https://social.tchncs.de/@thleemhuisfoss (DE): #FLOSS

searchable

This profile is from a federated server and may be incomplete. Browse more on the original instance.

kernellogger, to linux
@kernellogger@fosstodon.org avatar

Jeremy Allison writes:

'" The data shows that “frozen” vendor kernels, created by branching off a release point and then using a team of engineers to select specific patches to back-port to that branch, are buggier than the upstream “stable” Linux created by Greg Kroah-Hartman. '"

https://ciq.com/blog/why-a-frozen-linux-kernel-isnt-the-safest-choice-for-security/

kernellogger, to linux
@kernellogger@fosstodon.org avatar

Annoyed by having to put in front on [1]?

Then use this instead[2]:

$ journalctl -k

It should work if the user executing this is a member of the groups "systemd-journal", "adm", or "wheel".

[1] which is the case if CONFIG_SECURITY_DMESG_RESTRICT is turned on in your 's .config – which recently switched on, something many other distros did already a while ago.

[2] works for the common case, for some fancier stuff you might still need dmesg

kernellogger, to linux
@kernellogger@fosstodon.org avatar

"[…] It is clear that the current process is based on the learnings, and frustrations, the [ 's CVE] team has faced in the past. […] By taking this position, this effort is now duplicated across thousands of engineering teams ad infinitum, […]"

Well, yeah, but guess what: maybe then the companies behind those engineering teams will join up and invest money to handle the problem "[…] at the source, in a central, efficient and reliable manner. […]". 😬

https://amanitasecurity.com/posts/dear-linux-kernel-cna-what-have-you-done/

kernellogger, (edited ) to linux
@kernellogger@fosstodon.org avatar

I think the 's praised "no regression" rule it massively subverted by one aspect:

Developers that caused a regression in a mainline release (say today's 6.4) are only obliged to fix it in mainline (e.g. for the next release, 6.5 currently).

There is nothing that ensures the fix will be backported to affected stable (e.g. 6.4.y) kernels (ideally quickly, unless the fix is risky).

1/ That being said: the…

kernellogger, to linux
@kernellogger@fosstodon.org avatar

As community, process, and quality problems are discussed due to critical toots from @marcan a quick reminder:

Many problems currently discussed are well known for years; below screenshot from the slide deck mentioned at the start of https://lwn.net/Articles/799134/ shows that.

Things didn't improve much since then. The main reason for that IMHO is not "resistance to change among the elders" – it's lack of funding for people that work out and establish better workflows.

kernellogger, to linux
@kernellogger@fosstodon.org avatar

A question for experts on #git bisecting the #Linux #Kernel:

Assume someone runs into a regression when updating from 6.1.90[1] to 6.6.30 that needs bisecting. What do you suggest:

  • Check manually which mainline release (e.g. 6.2, 6.3, ...) introduced the problem and afterwards bisect between that and the previous release.

  • Bisect straight between 6.1 and 6.6.30.

1/ I guess I would definitely go for…

[1] let's assume that 6.1 was fine for this scenario to keep things simpler

kernellogger, to linux
@kernellogger@fosstodon.org avatar

made it and was finally merged for 6.7: https://git.kernel.org/torvalds/c/9e87705289667a6c5185c619ea32f3d39314eb1b

223 files changed, 95037 insertions, 56 deletions

For a feature overview and other details about this new filesystem see: https://bcachefs.org/

kernellogger, (edited ) to linux
@kernellogger@fosstodon.org avatar

6.6 just became a (aka ) .

The 's stable team (Sasha Levin and @gregkh) just made this official by adding it to https://www.kernel.org/category/releases.html with the following change: https://git.kernel.org/pub/scm/docs/kernel/website.git/commit/?id=d3c85f300d9214949efae275e519f30cce155cca

Projected EOL is December 2026.

kernellogger, (edited ) to linux
@kernellogger@fosstodon.org avatar

Linus re-sorted the 's MAINTAINERS file again and had this to say:


The answer is "No. No we cannot".

I suggest that all kernel developers will need weekly training sessions, involving a lot of Big Bird and Sesame Street. And at the yearly maintainer summit, we will all sing the alphabet song together.```

😂

<https://git.kernel.org/torvalds/c/c192ac7357683f78c2e6d6e75adfcc29deb8c4ae> #LinuxKernel
kernellogger, to rust
@kernellogger@fosstodon.org avatar

Correlation vs. Causation

Reminder: the support for programming the [ using aka is not used for anything at all yet within the – and most likely disabled in this user's kernel image]

kernellogger, to linux
@kernellogger@fosstodon.org avatar

I know why Canonical/Ubuntu is doing it like that, but it nevertheless every time feels kinda odd when they switch to a series that upstream gave up and tagged "end of life" two-and-a-half months ago already[1] – something the article doesn't even mention.

https://www.omgubuntu.co.uk/2023/08/ubuntu-22-04-linux-kernel-6-2

[1] https://lore.kernel.org/all/2023051744-drainable-footwear-49bd@gregkh/

kernellogger, to linux
@kernellogger@fosstodon.org avatar

Hmmm, immutable distros are currently ignored by the "How to quickly build a trimmed " text I recently added to the 's documentation[1].

Is that fine for now? Or should I add a sentence or two about those?

For @fedora at al. it seems using "ostree admin unlock --hotfix" might be the best solution when say doing a bisection.

But what's the best way for @opensuse MicroOS?

[1]https://docs.kernel.org/next/admin-guide/quickly-build-trimmed-linux.html

kernellogger, to rust
@kernellogger@fosstodon.org avatar

Andrea Righi[1] wrote a scheduler in / using sched-ext[2]; he claims he was "'"pretty shocked to see that it doesn't just work, but it can even outperform the default scheduler (EEVDF) with certain workloads (i.e., gaming):"'"

He shared it on the : https://twitter.com/arighi/status/1746938387968254371

Github page: https://github.com/sched-ext/scx/

Video: https://www.youtube.com/watch?v=oCfVbz9jvVQ

[1] engineer @ Canonical
[2] https://lwn.net/Articles/922405/

kernellogger, (edited ) to linux
@kernellogger@fosstodon.org avatar

[EDIT: better ignore this, the example is obsolete; see 3/]

One of the things I find really annoying about development:

Things that are right when contributing to one part of the are wrong when contributing to another.

Sometimes even rules that I'd consider to be on a higher level are nullified on lower levels. The netdev-FAQ[1] for example tells contributors to not tag changes for stable backporting as explained in stable-kernel-rules.rst[2].

1/ I can even…

kernellogger, (edited ) to linux
@kernellogger@fosstodon.org avatar

A employee reported some out-of-tree code broke after something internal was renamed recently in 5.15.y – and as expected was told this is no regression at all, as the does not have a binary kernel interface, nor does it have a stable kernel interface:

https://lore.kernel.org/all/924449dc-9b1f-4943-afe3-a68c03aedbb5@canonical.com/

Christoph Hellwig in a reply also wrote: "given that Canonical ignores our licensing rules and tries to get away with it I'm not going to offer any help to Canonical at all."

kernellogger, to linux
@kernellogger@fosstodon.org avatar

Reminder: if you encounter a regression, try to find the change causing it using git bisect; then report the issue through the proper upstream channels, as with a bit of luck it then will be fixed quickly.

A good example is the "Dell systems hang at shutdown" problem with 6.5. It's now fixed in mainline, 22 days after the initial report, 3 days after being bisected. The fix is now en route for the next 6.5.y release (expected within 3 or 5 days).

1/ For details…

kernellogger, to linux
@kernellogger@fosstodon.org avatar

1.1.7 is now available to Windows Insiders and among others bumps the to the 6.1 (a longterm aka LTS )

https://github.com/microsoft/WSL/releases/tag/1.1.7

@chrissicool for mentioning it in a reply]

kernellogger, to linux
@kernellogger@fosstodon.org avatar

I'm working on a text for the docs describing how to perform a bisection when facing a .

I just published a first draft: https://www.leemhuis.info/files/misc/How%20to%20bisect%20a%20Linux%20kernel%20regression%20%e2%80%94%20The%20Linux%20Kernel%20documentation.html

Would be great if a few people could take a closer look and provide feedback. Either here, by mail, or via changes & notes in this document: https://docs.google.com/document/d/1Im7SPK0j6PUGQTSGZyCTSQv8h3S51EYsZuRRdyhfzto/edit?usp=sharing

Would be even better if a few people could play this through – but I guess chances are slim anyone seeing this currently has a need for that.

kernellogger, (edited ) to linux
@kernellogger@fosstodon.org avatar

Getting closer and closer to the point where I'll start a git tree[1] with fixes and reverts for regressions in the latest stable series, as from here it seems quite a few of the known problems could quickly be solved by a revert or applying fixes already queued[2]/still under review[3].

[1] ideally in collaboration with the package maintainers from distros like @archlinux, @fedora, and @opensuse

[2] https://lore.kernel.org/all/a810561a-14f3-412e-9903-acaba7a36160@leemhuis.info/

[3] https://lore.kernel.org/all/ded3e7ae-6a7d-48b2-8acc-c125874ee09f@leemhuis.info/

kernellogger, to linux
@kernellogger@fosstodon.org avatar

This week I learned: be a snitcher more often.

What happened? There is a bug in 6.4 causing memory corruption. The culprit is known since ~Mon. A workaround was found and supposed to head to Linus on Wed or Thu; that didn't happen till Sat, which finally made me CC Linus[1]. Only 12h later the problem was solved for real in mainline with a CC: <stable… tag.

If I had CCed Linus on Mon, I guess something similar had happened back then already.

[1] https://lore.kernel.org/lkml/df1d7d39-56f3-699c-0d0f-fcc8774f182e@leemhuis.info/

kernellogger, (edited ) to linux
@kernellogger@fosstodon.org avatar

The 's team just published their thousandth CVE[1]. 🥳 🙃

This happened 78 days after the effort was announced[2].

Note, 26 of the 1003 CVE entries published so far were later rejected. For details check https://git.kernel.org/pub/scm/linux/security/vulns.git/ or https://lore.kernel.org/linux-cve-announce/

[1] https://git.kernel.org/pub/scm/linux/security/vulns.git/commit/?id=55441d0dd1f40c5762cd7cf8c9ca312ed0964c4a

[2] http://www.kroah.com/log/blog/2024/02/13/linux-is-a-cna/

kernellogger, to linux
@kernellogger@fosstodon.org avatar

: I have to perform a bisection on my Thinkpad T14s G1 (AMD) with Fedora's full-blown distro config, as the problem does not occur with a config trimmed by localmodconfig. 😟

[1] s2idle resume broke with 6.6-rc

kernellogger, to linux
@kernellogger@fosstodon.org avatar

Sooner or later the afaics will need a group of "first level bug responders" to reduce workload and noise for maintainers and developers.

E.g. a group of people that triages incoming bug reports[1] and normally only lets those through to the real developers that meet some very basic requirements.

[1] point to dupes, check if the report is sane, check if all basic information is there, ...

kernellogger, (edited ) to rust
@kernellogger@fosstodon.org avatar

Congrats to the and -for- community: the now contains the first useful thing built using Rust[1]! 🥳 👏

It's a network driver for Asix PHYs. It's provided as an alternative to an existing driver written in C. The features are equivalent.

For more details see https://git.kernel.org/torvalds/c/d6beb085e8ff3d9547df8a5a55f15ccc7552c5d0, https://git.kernel.org/torvalds/c/cbe0e415089636170aa6eb540ca4af5dc9842a60, and https://lwn.net/Articles/949270/

[1] reminder, until now the 's Rust support was not used for anything practical upstream: https://lwn.net/Articles/952029/

kernellogger, to linux
@kernellogger@fosstodon.org avatar

If commands in a how-to do not work and no one reports it, does it have an impact?

The 'cloning #Linux from a #Git bundle' instructions[1] on #kernel.org were kinda broken for years: they until a few hours ago[2] contained a step to verify the bundle, which only worked if your working directory was part of a git repository.

Does that mean that nobody followed that how-to? Or that those who encountered the problem did not report it? 🤨 🧐

[1] https://www.kernel.org/cloning-linux-from-a-bundle.html

[2] https://git.kernel.org/pub/scm/docs/kernel/website.git/commit/?id=ae230f928042c79fe52c8769276cdac0816284fd

  • All
  • Subscribed
  • Moderated
  • Favorites
  • •
  • JUstTest
  • kavyap
  • DreamBathrooms
  • thenastyranch
  • magazineikmin
  • osvaldo12
  • khanakhh
  • Youngstown
  • mdbf
  • slotface
  • rosin
  • everett
  • ngwrru68w68
  • Durango
  • anitta
  • InstantRegret
  • GTA5RPClips
  • cubers
  • ethstaker
  • normalnudes
  • tacticalgear
  • cisconetworking
  • tester
  • Leos
  • modclub
  • megavids
  • provamag3
  • lostlight
  • All magazines
    •