It Doesn't Look Like KDBUS Will Make It For Linux 4.1

[uid313]

Plan 9 from Bell Labs

Why not just use Plumber with 9P from Plan 9 from Bell Labs?

It is IPC invented for Plan 9 (aka Unix 2.0) by the founders of Unix.

[You-]

systemd FTW! kdbus is just a matter of time. It must go in. It has to. Sad but true.

Why is it sad?

[You-]

Why not just use Plumber with 9P from Plan 9 from Bell Labs?

It is IPC invented for Plan 9 (aka Unix 2.0) by the founders of Unix.

Because kernel developers have been unable to agree on a generic IPC interface and the developers have also decided that dbus has been tested in the real world for a decade now, so consider it to be acceptable for their use cases.

kdbus is very close to merging. Linus wanted a couple of concepts removing from it as per security review (he more or less pointed out which "reviews" to ignore and which to respond to), but otherwise he seemed to want to get it in too.

[Awesomeness]

Linus wanted a couple of concepts removing from it as per security review (he more or less pointed out which "reviews" to ignore and which to respond to)

I don't think someone with a mere masters degree is qualified to make such calls.

[erendorn]

I don't think someone with a mere masters degree is qualified to make such calls.

Cannot tell if you're serious (and stupid) or if it's sarcasm.

[cb88]

Cannot tell if you're serious (and stupid) or if it's sarcasm.

Trollin'


Also, Greg Kroah-Hartman is often arrogant sure but he's made alot of good contributions. Unfortunately his work has been poisoned by Lennart Pottering when he basically absorbed udev into systemd. Sure you can still sort of build and use udev separately but now it is maintained by those guys and they could care less about portability across all platforms. At least the old udev lives on in eudev and people that don't *really* need udev can use the simpler mdev and friends etc...

[cjcox]

Why is it sad?

Sad because it will soon be a requirement and it was done so poorly that it couldn't make the kernel. Remember we're betting the farm on the systemd folks.

[You-]

Sad because it will soon be a requirement and it was done so poorly that it couldn't make the kernel.

You are mixing tenses here making the sentiment pointless.

WHEN it is merged, it will be considered to be of high enough quality to be merged. That is why there are reviews. Some of them may be political, but not all of them are and people who are experts in different areas are chiming in with their reviews and suggestions.

Before merging details of how kdbus will be implemented have changed a number of times in response to previous requests for comments. As there were no outstanding non-political issues, a pull request was made, which forced other people to review the patches.

This shows that the project is working.

Remember we're betting the farm on the systemd folks.

systemd and kdbus are different projects.

kdbus is not written by the systemd guys (though given the low leve nature of the projects, there may be some cross over). Poettering and Sievers are not involved in writing kdbus, though they see the see the benefit in utilising it when it is merged.

[Awesomeness]

Cannot tell if you're serious (and stupid) or if it's sarcasm.

Torvalds doesn't really work on Linux these days. He mostly rants and insults other people. He's also not an expert in IT security.

[gens]

Sad because it will soon be a requirement and it was done so poorly that it couldn't make the kernel. Remember we're betting the farm on the systemd folks.

thank god that the kernel folk don't care about what is popular in user-space

the only real reason for dbus in the kernel is to transport multimedia, and that is a really really stupid reason
(dbus it too heavy a protocol to transport mass data)

EDIT: to be clear, it won't make the computer boot faster