5

According to this LWN article kdbus is supposed to replace D-Bus. Can this be confirmed somehow?

I wonder if after all this will be an easy task as I would imagine that both have a distinct API? To what I understand it will be necessary to redo some work/update to shift the programms from using D-Bus to kdbus? Hence if as mentioned in the LWN article it is the goal to replace D-Bus then, I think it would incur some update work,right?

Or maybe will it be that there will be some time in which both systems will work in parallel?

Improve this question

1 Answer 1

Reset to default
3

I'll give you my take on the article.

kdbus is supposed to replace D-Bus. Can this be confirmed somehow?

That is explicitly the intent, but WRT "confirming" it, there's no central authority that can say, "Yes, here is our timeline for the future of GNU/Linux" -- beyond the kernel, that's a heterogeneous and de-centralized realm.

Of course, seeing as how it is the kernel, a lot of decision makers in that de-centralized realm will probably be interested in cooperating. It sounds like a good thing.

will be an easy task

I don't see any indication that they can't both be used at the same time -- which would be by far the sanest form of transition. So "easy" depends on your context...

imagine that both have a distinct API?

The Greg K-H announcement linked at the beginning refers to a userland compatibility layer, which is also very sane with regard to transitioning; initially some distros could make both available, others move directly to the compatibility layer, etc.

Sometimes it is good to move forward in a way that sacrifices backward compatibility. Consider perl 5 vs. perl 4 or python 3 vs. 2; within the major version, improvements (indicated by new minor versions, 5.8, 5.9, 5.10, etc.) are made with backward compatibility a priority, and meanwhile work on the work on the next major version (which is incompatible, but presumably much improved based on experience with the current version) may be progressing.

Distros that have versions are similar in the sense that while the current version is being maintained and updated work on a new version is also going on. This makes the incorporation of fundamental changes like kdbus easier. I guess we'll see what happens.

Improve this answer
2
  • thank you for your insight!, Can you (and do you dare to) give a time horizon/estimation as regards to when, or if the kdbus based on the insides? Having read the article I am not even sure if kdbus might not even eventually be rejected still? (which coming from a security appreciative perspective I would not mind very much :)
    – humanityANDpeace
    Nov 16, 2014 at 17:30
  • If I had to make a total guess based on, e.g., the time required for SysV init to be phased out, you are looking at years. Although it doesn't stand to make much difference to my life, I hope it happens; I like kernel-userspace interfaces. Ideally, the development of new technology leads to more security, again because of lessons learned from older things.
    – goldilocks
    Nov 16, 2014 at 18:18

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .