Re: [PATCH] Futex Asynchronous Interface

[kaih@khms.westfalen.de (Kai Henningsen)]

torvalds@transmeta.com (Linus Torvalds)  wrote on 08.06.02 in <Pine.LNX.4.44.0206081523410.11630-100000@home.transmeta.com>:

> On Fri, 7 Jun 2002, Peter Waechtler wrote:
> >
> > What about /proc/futex then?
>
> Why?
>
> Tell me _one_ advantage from having the thing exposed as a filename?

None, of course - the shell can't do the other futex ops, either. Futex  
file handles mean you can implement select() on them, but that's about all  
they have in common with files - there is certainly no read() or write()  
operation here!

> > Give it an entry in the namespace, why not with sockets (unix and ip)
> > also?
>
> Perhaps because you cannot enumerate sockets and pipes? They don't _have_
> names before they are created. Same as futexes, btw.

Now *there* I disagree, at least for sockets. First of all, there's  
absolutely no need to be able to enumerate unopened sockets to justify  
putting them into the namespace - you can create them (in a special fs, of  
course) at the moment they are opened. (As in fact is done, in /proc/<pid>/ 
fd/.) And second, you *can* read() and write() there.

However, I don't think that's all that important. What I'd rather see is  
making the network devices into namespace nodes. The situation of eth0 and  
friends, from a Unix perspective, is utterly unnatural.

MfG Kai