kdbus

So did you see the fallout on the LKML?

http://lkml.iu.edu/hypermail/linux/kernel/1504.1/index.html#03936

SystemDisease developers try again to infect the kernel with their nasty kernel-dbus, and Linus and several other senior maintainers trash it for being a terrible, insecure layer of crap. SystemDisease developers just smiley-face (seriously) and try to act innocent, and claim that if people don't want it, they can just disable it in the kernel, that it won't become a hard dependency. Meanwhile, the same people over on their SystemDisease blogs state that as soon as it's in the kernel, it will become a hard dependency by SystemDisease and several other projects they control.

Senior maintainers discover this and call them out on their shenanigans. SystemDisease developers then claim Linux is full of mysognists and racists, trying to use identity politics to brow-beat Linus and friends into submission. Linus tells to fuck off and that he'll never accept any pull requests from them.

Please disregard duplicate post.

>>2
no

yes I saw this

i don't like it

I have no power over it

Desktop linux is dead and I don't see how this is relevant to mobile or embedded.

kdubs

i wonder if greg is taking government money to push this crap or if he actually has a mental issue that makes him unable to comprehend any criticism

Debating with Kroah-Hartmann is like talking to a wall. The prime example of this is the “It's not mandatory, don't use it if you don't like it!”/“Of course it became mandatory, did you expect people to disregard extra features?” spiel he pulls around four times in that thread alone.

I hope he's taking money for this shit like >>7-san is saying, because the alternative is a retarded kernel developer with way too much power. LKML is way too polite to him.

>>8
I assume some well heeled corporate sponsors of the Linux Foundation are leaning on people to get this merged. This wouldn't be a problem if the leaned upon would at least explain what use cases kdbus is supposed to enable - right now it seems to exist in this strange limbo between high performance ugly things (DRM, shmem) and lower performance simple things (pipes).

Most desktop applications using dbus today do not have high performance requirements, and the few that do seem to do well enough with the existing IPC primitives. I can't help but imagine some cabal of proprietary middleware developers at Intel or Samsung are trying to do really insane things with dbus and not telling anyone why, because they don't really know themselves.

>>1

and Linus and several other senior maintainers trash it for being a terrible, insecure layer of crap.

I don't see Linus trashing anyone there, nerd. I like it how you misogynistic nerds have to go deeper and deeper into your autistic fantasies as the world moves on. And I really enjoy seeing you flail your appendages in bouts of impotent rage whenever the reality comes through to you and you realize (but not really) that you're becoming obsolete.

>>7
What criticism?
I'm about 20 messages into the thread and haven't seen a single technical argument against kdbus, only inane rants about 'no choice' and 'systemd being forced into my anii'

>>10
Leave this place.

>>12
Make me, nerd ;-)

>>11
There are valid complaints about the implementation as well (large ABI, unclear locking rules, unclear use of many existing kernel interfaces like capabilities that most assume have nothing to do with an IPC mechanism...)

To be frank I think the distrust of certain freedesktop developers is totally justified at this point. That bunch has a history of unilaterally changing things in user space that kernel developers are known to rely on (udev firmware loading, kernel log buffer abuse) and totally refusing to own up to the breakage it causes. There's no reason to expect this pattern of behavior to stop once they have a foothold in kernel space.

>>11
Since when did mandatory SystemDisease become an okay thing and objections to it "inane"?

>>15
They won the init war, deal with it. Install BSD.

>>16
No, they didn't. My Gentoo is free of systemDick.

>>17
For now... wait til the depenancy creep sets in.

https://lwn.net/Articles/641275/#Comments

>>16
https://uglyman.kremlin.cc/gitweb/gitweb.cgi?p=systembsd.git

>>5
GNU/Linux is a server os

Check my dbus.

>>22
Checking dbus...
note: udevd[225] exited with preempt_count 2
Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: c0147299

I still can't see what's the point of dbus, there are much better IPC mechanisms and most of them are implemented as POSIX functions.

Remember this thread? https://dis.4chan.org/read/prog/1219766316/

>>21
My phone isn't a server!

>>25
Your phone probably does not run any GNU program.

>>26
Sure it does, compiled some of them myself.

I don't know much about 9P, but it seems like it would be a cleaner interface for desktop IPC where you don't want to use sockets/pipes. Are there any drawbacks to 9P?

>>27
People who regularly use phones are sub-human plebians.

DBUS incoming.

>>28
It's too slow and unreliable for network comms, even 9P2000 does not improve on this much.
If somebody that was not a cat-v shill worked on it we might have something good.

>>31
What would the benefit of high-performance networked 9P filesystem bring to IPC? Would it be necessary for a functioning alternative desktop system?

DBUS leverages privilege separation and world-class Unix domain socket IPC stack to provide programmers worldwide with robust, scalable, modern turnkey implementations of flexible, modular, cutting edge message-based run-time system of of algorithmic architectures that accelerate response to theoretic and real-world integration demands and reliably adapt to evolving software needs, seamlessly and efficiently integrating and synchronizing with their existing legacy codebases, enhancing the information sharing capabilities of their code production environments across the enterprise while giving them a critical competitive advantage and taking them to the next level.

http://dbus.freedesktop.org/doc/system-activation.txt

>>34
WTF does "activation" even mean? Don't tell me it's just another buzzword for "run the program"...

>>35
Get the level of bloated idiocy in this:

Launching programs using dbus has been a topic of interest for many months.
This would allow simple systems to only start services that are needed,
and that are automatically started only when first requested.
This removes the need for an init system, and means that we can trivially
startup services in parallel.

Turns out that starting only services that are needed is somehow a great achievement, lol.

🆒🆓

>>36

Turns out that starting only services that are needed is somehow a great achievement, lol.

Well. Are you saying that not doing that is an even greater achievement, being true to the ways of our forefathers probably?

>>38
No, I'm not.