torvalds@transmeta.com (Linus Torvalds) writes:
> In article <20000110145913.01335@colin.muc.de>, Andi Kleen <ak@muc.de> wrote:
> >
> >I planed to map the Unix Sockets abstract name space to a file system
> >for some time now. Because it would be silly to write another file
> >system just for that rather obscure feature, would it be possible
> >to use a subdirectory in your new shm filesystem?
>
> I'd much rather _not_ do it like that.
>
> Remembe: do one thing, and do it well. Mixing up seveal conceptually
> unrelated things just ends up a maintenance nightmare, and implies tying
> the two together more than necessary.
>
> What if somebody wants shared memory but not UNIX domain sockets? Or
> vice versa? Issues like that add extra and unnecessary complexity if you
> want to tie the functionality together.
But if we think of the shm fs as an in memory fs we could mount an
instance somewhere simply to get the directory handling for Unix
Domain sockets. The Unix domain sockets then only needed the path to
their instance to put in front of the sun_path like SYSV shm needs it
right now.
For this we would need to decouple the shm fs from SYSV shm and also
add features like multiple mounts, directories and special files.
But I think it's worth a thought.
Greetings
Christoph
-
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 Jan 15 2000 - 21:00:18 EST