Skip to main content


```articles that recommend to disable # extensions, such as […] # […] provides background on what these extensions do […] and why it is normally a bad idea to turn them off. ```

(written by # kernel # co-maintaner Florian Westphal)
TCP window scaling, timestamps and SACK



# 5.8.1 is out(¹). That means: prepare for headlines like ```5.8 Gets Its First Point Release, It's Now Ready for Mass Adoption``` or ```5.8 is stable now``` that totally miss the point.

(¹) and a few other stable and longterm # versions, too




[bookmark=https://twitter.com/kernellogger/status/1127099676493328384]# # 5.1.1 is out.

IOW: Prepare for headlines like ```5.1 Gets Its First Point Release, It's Now Ready for Mass Adoption``` or ```5.1 is stable now```, which IMHO give quite a misleading impression of how the Linux development process and its versions numbers work.

[/bookmark]



[bookmark=https://twitter.com/fagiolinux/status/1293114449478508544]QEMU is going to switch to Meson! A fully bisectable switch to a new build system, in 147 easy steps patchew.org/QEMU/159707934…

... almost, there's a patch in there that I forgot to squash :) [/bookmark]

@Iron Bug Sounds interesting! To be honest, i'm a funtoo user anyways, so python is at the heart of my distribution anyways. ;)

What is your model? Are there any docs for your approach? ;)

there's no any docs. I just build my system from sources. it's based on Void Linux with musl, but it's gone too far from the original distributive, so it's my own compilation. I don't publish it, because don't want to turn into maintainer: I have no time for this.
I deleted python from production, but still need to delete it from build system. one variant is to rewrite meson in C or make converter of meson to normal autotools. if I have time for this I will do it.
and there're some troubles with python in mesa and some gnome libraries. I don't like gnome, but many programs use its libraries. so it's one of the tasks I haven't done yet. but I don't hurry up. I have a plenty of years ahead, I will kill it everywhere :)



[bookmark=https://twitter.com/wipawel/status/1291050377958379520]I am very happy, because my pet project goes public today! It is a simple # # # that I use for my low-level security research. I hope it can be useful for others too! [/bookmark]




[bookmark=https://twitter.com/linuxplumbers/status/1292877241823756290]We're trying our best to have the virtual version of # be as close to the in-person # spririt as we can.

That specifically means that it should foster the same intense technical discussions, keep its focus on solving problems, and - above all -[/bookmark]



[bookmark=https://twitter.com/linuxplumbers/status/1292877237075804161]Here's the promised # registration update:
We're currently sold out but we will be adding at least another 100 tickets and_maybe_ more. 😊[/bookmark]



[bookmark=https://twitter.com/hughsient/status/1292783799525429249]A large hardware vendor wants to join the LVFS, but only on the agreement that every user has to agree to a English-only EULA text when deploying their firmware updates. This is the first vendor that's required this condition, and breaks all kinds of automated deployment. Do we:[/bookmark]



[bookmark=https://twitter.com/perfsummit1/status/1289872001755017216]Recent versions of Android come with eBPF support. Learn more about ways one of its earliest adopters, Facebook, is leveraging it for low-overhead performance and reliability investigations. [/bookmark]



From the # merge for # # 5.9: "[…] made inode flushing fully asynchronous, and we no longer block memory reclaim on this […] the reflink control loops have been redesigned[…] behavior of DAX between ext4 and xfs has been stabilized[…]"

git.kernel.org/torvalds/c/563…



[bookmark=https://twitter.com/knurd42/status/1292489640482570240]In case you are already looking for a Christmas gift for yourself: # # [/bookmark]




[bookmark=https://twitter.com/linuxplumbers/status/1292148382337847298]We really want to make sure as many people as we can support can attend! 🥰[/bookmark]



[bookmark=https://twitter.com/read_rust/status/1292044550488199168]Why QEMU should move from C to Rust by Stefan Hajnoczi: blog.vmsplice.net/2020/08/why-qe… # [/bookmark]
#os



[bookmark=https://twitter.com/linuxplumbers/status/1292130750423281666]We finished announcing all of our microconferences for # 2020. And we have 16 microconferences this year! That's not too bad. Let's take a quick look what we have:[/bookmark]



If you wonder what the new # # syscall close_range() does: "It allows to efficiently close a range of file descriptors up to all file descriptors of a calling task."

For more details see this merge git.kernel.org/torvalds/c/4f3… and this article: lwn.net/Articles/78902…
[bookmark=https://twitter.com/haerwu/status/1292023176923422720]5.9 merge window in progress. Unicore32 architecture is no longer supported, there is new close_range() system call.

So I updated my system calls table. [/bookmark]



[bookmark=https://twitter.com/haerwu/status/1292023176923422720]5.9 merge window in progress. Unicore32 architecture is no longer supported, there is new close_range() system call.

So I updated my system calls table. [/bookmark]



Merged for # # 5.9: code to block all # writes from userspace apart from a few registers on a allowlist (there is only one on it for now).

The plan is to make this the default mode of operation, but for now it only logs all writes.

git.kernel.org/torvalds/c/a7e…



[bookmark=https://twitter.com/brau_ner/status/1291731342326145024]I bet the seccomp notifier can be used to use bpf() in unprivileged containers... Let's try 👇😬
people.kernel.org/brauner/the-se…[/bookmark]



[bookmark=https://twitter.com/brau_ner/status/1291740147516678148]I used the seccomp notifier, pidfd_getfd(), and the new seccomp fd injection work to use # in unprivileged containers. 🙃 [/bookmark]
#BPF