Re: Suggested dual human/binary interface for proc/devfs

From: Alexander Viro (viro@math.psu.edu)
Date: Thu Apr 13 2000 - 09:36:41 EST


On Thu, 13 Apr 2000, Richard Gooch wrote:

> But that means I can't abolish those butt-ugly device numbers from my
> kernel.

Yes, it does. Too fscking bad, but you are _not_ going to get rid of them.
Period. mknod(2) is there to stay. Deal. If killing device numbers is your
design goal it looks like you lose. They are ugly, but there is no way in
hell to remove them - you are _NOT_ getting devfs mandatory. Forget it.

[snip]
> That's fine. All I'm saying is that to get both ways, I can't rip out
> gobs of code from devfs and go all the way with your scheme.

Let's see: devfs contains
        registry for block devices
        registry for character devices
        automounter on strange drugs
        half-assed template for virtual filesystems
        half-assed union-mount (very special case of it)
        half-assed multiple-mounts (racey)
Remember "... should do one thing and do it well"? Most of that stuff
needs to be done in VFS and done right. Instead of doing it there you've
lumped kinda-sorta implementations together, made the result look
superficially similar to fs driver and strongly oppose to taking stuff out
of this animal? Mind boggleth...

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Sat Apr 15 2000 - 21:00:21 EST