[mdlug] Big brother gives M$ a 10 out of 10

[LAP]

On Fri, 24 May 2024 07:34:30 -0400
LAP <mail1 at lapiet.info> wrote:

> 
> http://0pointer.de/public/gnomeasia2014.pdf
> 
> The goal of systemd is to become the one-and-only interface
> between user space and the kernel.
>

I forgot to mention that systemd goes beyond user space
and extends into the kernel itself, or at least has been
attempting to do so.

>From the beginning, the systemd folks have been trying
to get their brand of inter-process communication, or IPC,
into the kernel itself.

The current IPC used by Linux today is dbus, but this is far
too limited for the goals of systemd.  They first tried to
push their kdbus into the kernel tree but, fortunately, it
was rejected by the kernel developers.  Since then they
are developing bus1 as the new in-kernel IPC.  When, or if,
bus1 becomes part of the Linux kernel only time will tell,
but if it does then systemd will be able to dictate how most
other software can behave.

Even at present, systemd is probably the only user-space
software that demands a certain kernel configuration.
Kernel cgroups must be enabled or systemd cannot function
and a system will not boot.  This reliance on a purely
kernel-space capability further indicates that
systemd intends to be on par with the kernel itself.

Remember also the fiasco when Kay Sievers, a systemd developer,
began to infiltrate the kernel to output boot log messages
for systemd.  Linus Torvalds had to literally cut off his
access privileges until he stopped the practice.  Obviously,
the belief was that systemd is equal to the kernel itself.

Poettering himself has also stated that POSIX principles
should be ignored because they are too limiting.

All of this adds up to systemd being a major threat to
long established Unix methods and hence GNU/Linux freedom
and diversity.