While reading devfs guide on the gentoo site I've bumped into really weird statement: For instance, with devfs, you don't have to worry about major/minor pairs. It is still supported (for backwards compatibility), but isn't needed. This makes it possible for Linux to support even more devices, since there are no limits anymore (numbers always have boundaries :) To the best of my knowledge devfs doesn't get rid of major/minor device numbers in any way and therefore doesn't break any boundaries. devfs dynamic creation of filsystem nodes doesn't provide any namespace-based access and the access to the devices still relies on the major/minor numbers provided. Reproducible: Always Steps to Reproduce: 1. 2. 3.
yes but devfs generates nodes dynamically and can assign major/minor pairs dynamically the limit used to be 8bits (256) but since things are done dynamically, we arent limited to 8bits anymore
It would be strange to suppose there's nothing behind the statement. But it says "there are no limits anymore" and puts the devfs allocation strategy as opposite to numbers. I'd say that the wording is inaccurate to say the least. If accuracy of the documentation matters...
DevFS uses it for backwards compatibility although it isn't set in stone. If we want to we can remove this bc setting and use dynamic numbering (so instead of giving each type of hardware it's own maj/min, we can just increase the maj/min every time a new hardware device is found). But I have not seen a working implementation of this as too many tools relie on the static numbering scheme. It's in the guide to inform the users about DevFS' possibilities, not how Gentoo uses it.
I think we all know what goes on and what is stated in the documentation. No matter if we speak on devfs or devfs in gentoo, the question is how accurate we describe the concept. I don't feel like convincing you the's something sheerly wrong in the documentation or god forbid software. All I wanted is, assuming you know the concepts behind devfs, is to draw your attention to the paragraph quoted.
