KDBUS & Other Features You Won't FInd In The Linux 4.0 Kernel

[phoronix]

Phoronix: KDBUS & Other Features You Won't FInd In The Linux 4.0 Kernel

While Linux 4.0 is the next major kernel release and it does present a new lot of new functionality, there's still a number of high profile features not mainlined...

http://www.phoronix.com/scan.php?pag...ssing-Features

[GreatEmerald]

I read a bit of the comments on the last KDBUS RFC, and it's quite hilarious. It's like a battle of devil's advocates, where the systemd folks are adamant about loose coupling and making it a module, and the kernel developers are requesting tight coupling into a core part of the kernel due to all the benefits it gives.

[renox]

I read a bit of the comments on the last KDBUS RFC, and it's quite hilarious. It's like a battle of devil's advocates, where the systemd folks are adamant about loose coupling and making it a module, and the kernel developers are requesting tight coupling into a core part of the kernel due to all the benefits it gives.

You're sure that you didn't invert the advocates names? Because I don't see why systemd advocates would be against making KDBUS a tight coupling in the kernel: it would be a benefit for them..

[doom_Oo7]

You're sure that you didn't invert the advocates names? Because I don't see why systemd advocates would be against making KDBUS a tight coupling in the kernel: it would be a benefit for them..

No, it's correct : http://lkml.iu.edu/hypermail/linux/k...1.2/02699.html

[valeriodean]

No, it's correct : http://lkml.iu.edu/hypermail/linux/k...1.2/02699.html

I don't know (I'm not an expert in this field) but, if you use a special file system as base, then isn't logic to use the ioctls instead of syscalls?
In any case, there are other parts hilarious as well, something like:

kernel_dev: why you choose to do that in this way?
kdbus_dev: Because we want kdbus to be a optional driver...
kernel_dev: Why? If it will be adopted by systemd, it will be everywhere in a couple of years, so it will be a lot less optional than what you think

LOL.
When they choose the tight coupling way, people complain, when they choose the optional way, people complain...
It looks like is hard to be a systemd devs!
Ahahahah

[rdnetto]

You're sure that you didn't invert the advocates names? Because I don't see why systemd advocates would be against making KDBUS a tight coupling in the kernel: it would be a benefit for them..

Here's my take on it, having read the mailing list thread:
The kernel devs regard kdbus as a core kernel feature, and as such want it to use syscalls. (They also generally dislike ioctls due to the lack of type safety for user space.) (src)
It's also worth noting that Greg Kroah-hartman is one of the kdbus devs, which says a lot about how central this is to the kernel. (For the uninitiated, many regard him as Linus' second in command.)

Meanwhile, some of the kdbus devs regard it as an 'IPC driver' instead of a core feature. I suspect part of the reason for this is that core features are held to a higher standard in terms of documentation and API, which means it'll take longer to get the code mainlined. Having such a tight coupling with the kernel probably won't help with their PR either, given how many flame wars there have been on the subject.

[Maxim Levitsky]

Well

As they say, systemd is coming to crash an kernel near you

[MoonMoon]

As they say, systemd is coming to crash an kernel near you

You are too stupid to be an effective troll, try again.

[Maxim Levitsky]

You are too stupid to be an effective troll, try again.

If someone is stupid, its you
Because I was not trolling but just joking about the situation.

[curaga]

Damn, if that became mandatory it would push back Linux-tiny a few years' worth of work.