The Heated KDBUS Debate For The Linux Kernel Has Fizzled Out

[gens]

Pretty much, yeah. Like I said in my previous post - dbus is *already* the defacto standard IPC on Linux, because that's what 99% of the IPC-using software on Linux uses. There's no value in creating hypothetical better systems, unless someone is prepared to do the work to implement them, and to migrate hundreds or thousands of packages to use the new system. A "better" system that nobody uses doesn't really count as "better" in my book...

i hear java is popular
we should put it in the kernel so we could get 100x speedups

on a less sarcastic note
no, by far the most widely used IPC on linux are UNIX domain sockets
i won't even start on how inefficient of a protocol dbus is

[gens]

kdbus is that specialized ipc protocol

specialized for what ?
"we must have a generic IPC in the kernel"

i told you a long time ago to stop insulting people
(at least i think it was you)

[timtas]

First: There is no _formal_ API spec of dbus, it is merely ad-hoc
Here is a link: http://gentooexperimental.org/~patri...s/2014-11.html
Second: As there are better IPCs available (at a cost) why implement the bloated dbus?
It is a crappy piece of software, unfortunately I had to dig into it, and its's awful. (and systemd id dbus-based)

Great read, and very well summed up with: "If this is the future I'll stay firmly stuck in the past ..."

[gens]

i have to put this here

https://www.youtube.com/watch?v=eFTLKWw542g

[stalkerg]

Do proper optimizations of the actual real costs before starting to
work on kernel stuff. It's *stupid* to add a kernel driver to get 2x
improvement, when there's a 10x bloat in user space.

You right Linus!

PS http://lkml.iu.edu/hypermail/linux/k...4.3/02886.html