Re: Linux's future: //posix/ipc, //root and so on ?

From: Khimenko Victor (khim@dell.sch57.msk.ru)
Date: Tue Feb 29 2000 - 16:56:21 EST


On Tue, 29 Feb 2000, Alexander Viro wrote:

>
>
> On Wed, 1 Mar 2000, Khimenko Victor wrote:
>
> > AV> Ewwwwwwwwwwwwwwwwwww... We don't need to implement Weirdnix - Dave Cutler
> > AV> had already done that.
> >
> > Hmm. I've heard that Unix also existed as well long before Linux was
> > implemented...
>
> Search on the net for "weirdnix", OK? It's amusing, honest.
>
Hmm. AFAIK "weirdnix" is just alias for "not quite straight Unix
implementation", no ? And my point was that why we need Linux is we want
JUST unix ??? There are already exists "just unix". Even free one (*BSD).
Of course we should not stuff Linux with additions without thinking.

> > AV> The whole Missed'em'V IPC thing is a completely fucked up API (OK, as the
> > AV> whole Missed'em'V API is) and we'ld better add a filesystem, and mount it
> > AV> during the boot like every honest UNIX should do with filesystems.
> >
> > Linux HAS SVR4's IPC. What it does not have is POSIX's IPC. And what you want
> > to do with chroot, BTW ? YOU were one who complained about problems with devfs
> > in chroot'ed environment AFAIK. And YOU want still more such problems ??
>
> For a filesystem with flat namespace that doesn't try to do all this
> foo-on-demand mess multiple mounts _are_ easy.

There are limited number of mountable filesystem AFAIK. If you have
500-1000 users you just can not mount ipcfs, procfs, devfs for each...
You CAN provide sane environment with hardlinks, though...

> So... not a valid reason to stuff lookup code with even more
> ad-hackery. If you want namespaces - say so. When we'll have SMP-safe
> dcache - fine, they'll go. Until then...
> Thanks, but No Thanks.
>
> > AV> But doing it via magic names? Yuk. _Please_, don't do that.
> >
> > Ok. What's alternative ? BTW I can not do this (and it was Linus's idea, not
> > mine after all).
>
> man 8 mount
>
Ok. It IS alternative to some extent. Just not sure if it's BETTER
alternative...

-
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 : Tue Feb 29 2000 - 21:00:23 EST