A weekly roundup of Linux Kernel Mailing List threads but with a lot more sarcasm.

kernel shit weekly 2017/04/08

[PATCH 00/24] Kernel lockdown

This wins the “cool” award for this week. This patch set would provide a mechanism to prevent user-space from modifying the running kernel image.

[PATCH 0/2] Create an initial user-space API manual

As someone who is used to finding the specific commit messages in the kernel to use as the “docs”, this patch is YUGEEE.


Blast from the past! We used to use LEGO Mindstorms in college, this patch is for the driver for the battery.

[PATCH net-next 0/6] bpf: program testing framework

Introduces a BPF testing framework, which solves numerous problems with skb and xdp programs being difficult to test.

[PATCH v1 14] syscalls: Restore address limit after a syscall

Prevents a syscall from modifying the address limit of the caller. Patch could be used to mitigate

kernel shit weekly 2017/04/01

[PATCH] MAINTAINERS: Drain the swamp

Let’s be honest, everyone knows who the real members of the swamp in the kernel are…

[PATCH net-next v6 00/11] Landlock LSM: Toward unprivileged sandboxing

HOLY CRIKEY BATMAN even though this is the 6th patch set sent out for this feature I am still giddy like a school girl.

Re: syscall_geterror() && TS checks

Welcome to Naming Things 101.

kernel shit weekly 2017/03/23

Bcachefs - encryption, fsck, and more

There was a new Bcachefs release and we are all reminded that is a thing that exists and has about as many production users as attendees at an OpenBSD meetup.

[PATCH 0/7] LLVM: make x86_64 kernel build with clang.

All I have to say is


RFC to add a new flag to mitigate some path transversal bugs, also has some obvious sandboxing use cases.

kernel shit weekly 2017/03/16

Re: [scsi] scsi: ufs: don’t check unsigned type for a negative value

Human error when a reviewer spells their own email wrong. Jokes are made but all in good fun.

Re: [RFC PATCH 00/13] Introduce first class virtual address spaces

Someone tries to reinvent the wheel, realizes most of the functionality can already be achieved and sadly that their patch name is too “marketing”.

Re: [git pull] drm for v4.11 - main pull request

The drm subsystem is on the shit list this week for making Linus go all caps mode over nested defines & gcc warnings.

Re: [PATCH net-next RFC v1 00/27] afnetns: new namespace type for separation on protocol level

A new namespace is proposed to benefit container networking performance. This just neat imho and there is no drama yet which is the actual shocker.

Compiling kernels faster (was Re: v4.10: kernel stack frame pointer .. has bad value (null))

A thread that started as a bug, turned into a compiler version debate, morphed into a thread focusing on compiling kernels faster. Unsure if this can even be considered a bike shed or if it needs a new name.