Bringing D-Bus Into The Linux Kernel

[Svartalf]

With plasma fire you mean. And all the devs who had this idea. Then spread the dust through space and lets never talk about it again!

If you're using plasma fire, I'm thinking that Schlock just eats it.

I'm thinking that we just take off and nuke it from orbit- it's the only way to be sure. >:-D

[fabiank22]

Let's move Firefox into the kernel too. I can post benchmarks then showing how faster it is

Well guess what? It ?sn't. At least not by much. The only reason for the massive speedup in D-Bus is because D-Bus is a userspace daemon that has to do A LOT of switching to Kernel Mode whenever it's used. Firefox needs to do nothing in kernel mode.

Or to make a better analogy: Paying a bus driver to leave his bus and stand around outside the exact moment nobody needs to driven and have him haste back into the bus whenever customers are around makes no sense. Yet with D-Bus you advocate that.

And once again: The kernel is modular - you don't want D-Bus don't build it.

[Svartalf]

A commercial company already did that with a proprietary kernel module:
http://www.microxwin.com/

What they won't tell you is that it's a proprietary module- shipping something with it to anyone other than in-house clients is a GPL violation against the Linux kernel itself.

[DeepDayze]

Oh yeah, what exactly is "ugly" and "Overengineered"? Do you have any idea how D-Bus works or have developed with it at all? D-Bus was created to adress real problems(e.g. DCOP being old and not up to the task anymore), and also has been around for a while now. There's a freedesktop spec and all.

Secondly those are not "bad design decisions", D-Bus was created with KDE/The Desktop in mind, not with embedded systems. My normal CPU can handle the (small) overhead, but ARM/Snapdragon-Phones and devices?

Also Power Consumption is a good point, Kernel Space = less wakeup calls due to context switching. This is Computer Science 101.

DCOP is a KDE thing not a general userspace message bus and definitely old and has reached its limits, thus dbus was born. Also Gnome just started using dbus as well so it isn't just a KDE thing now any longer and it has gotten the same benefits as KDE did.

[DeepDayze]

Ewww. D-Bus itself is completely unnecessary, keep that thing out of the kernel.

Agreed, that would bring a lot of bugs plus more work for the kernel devs to keep up with it. The more crap you bring into the kernel the more bloated it will be

[RealNC]

Let's move Firefox into the kernel too. I can post benchmarks then showing how faster it is. While we're at it, move everything into the kernel. A distro then consists of a 20GB vmlinuz kernel image. It will be very fast, I guarantee it, since everything is in the kernel.

Of course, in order to do updates, you have to update the whole kernel. But who gives a fsck? They did it with drivers. Let's do it with everything else too.

[fabiank22]

Step 1: design Yet Another Ugly and Overengineered IPC System (YAUOIS)
Step 2: move it to the kernel to overcome the problems caused by bad design decisions.

Oh yeah, what exactly is "ugly" and "Overengineered"? Do you have any idea how D-Bus works or have developed with it at all? D-Bus was created to adress real problems(e.g. DCOP being old and not up to the task anymore), and also has been around for a while now. There's a freedesktop spec and all.

Secondly those are not "bad design decisions", D-Bus was created with KDE/The Desktop in mind, not with embedded systems. My normal CPU can handle the (small) overhead, but ARM/Snapdragon-Phones and devices?

Also Power Consumption is a good point, Kernel Space = less wakeup calls due to context switching. This is Computer Science 101.

[BlackStar]

A 1.12x increase in performance is nothing to laugh at. Linux sucks in power consumption compared to Windows or Mac OS X and every bit that helps bridge that gap is a good thing in my book.

[curaga]

@diegocg

Definitely. You forgot step 1.5: Make it Yet Another Daemon.

So that people who just want to pass messages can get more resource usage, always there for you, up and running.

[diegocg]

Step 1: design Yet Another Ugly and Overengineered IPC System (YAUOIS)
Step 2: move it to the kernel to overcome the problems caused by bad design decisions.

Seriously, I doubt this will be accepted. Android is already trying to push in their own (YAUOIS)...how many of the do we need? Can't D-Bus use the Android crap?