bug: udev does not create /dev/fd pseudo directory
Dennis Veatch
dveatch at woh.rr.com
Fri Jan 20 01:21:13 UTC 2006
On Thursday 19 January 2006 19:05, sofar wrote:
> On Thu, 19 Jan 2006 19:00:34 -0500, Dennis Veatch <dveatch at woh.rr.com>
wrote:
> > On Thursday 19 January 2006 18:43, sofar wrote:
> >> On Thu, 19 Jan 2006 15:34:27 -0500, Dennis Veatch <dveatch at woh.rr.com>
> >
> > wrote:
> >> > On Thursday 19 January 2006 15:03, Dennis Veatch wrote:
> >> >> On Thursday 19 January 2006 14:19, Moritz Heiber wrote:
> >> >> > On Thu, 19 Jan 2006 18:01:46 +0000
> >> >> >
> >> >> > sofar <sofar at foo-projects.org> wrote:
> >> >> > > Can this symlink be added to udev somehow?
> >> >> >
> >> >> > udev.rules one-liner.
> >> >> >
> >> >> > Regards,
> >> >> >
> >> >> > Moritz
> >> >>
> >> >> What should the MODE and GROUP be?
> >> >
> >> > Would something like this be right? :
> >> >
> >> > KERNEL=="self/fd[0-2]*", NAME="proc/%n", SYMLINK+="%k", MODE="0600"
> >>
> >> no
> >>
> >> you should *ONLY* symlink /proc/self/fd to /dev/fd
> >>
> >> don't symlink anything else, the kernel (procfs) takes care of
> >
> > everything
> >
> >> under /proc/fd.
> >>
> >> no clue about udev syntax - I need help for that.
> >>
> >> Auke
> >
> > Well then if I understand right, this should do the trick (I think);
> >
> > KERNEL=="/proc/self/fd[0-2]*", NAME="/dev/%n", SYMLINK+="%k",
> > MODE="1777"
>
> ????
>
> /proc/self/fd is a DIRECTORY that always exists, there is no fd0, fd1, fd2
> etc. ONLY fd
>
> > So to me my own lab rat. Poked that line in my udev rules (ouch!) and
> > reboot. .
> > .
> > .
> >
> > Hey I have a "self" in /proc now!
>
> that already existed before you dumbass! it is a reference to $$ or the
> /proc/nnn directory of the "current" process!
>
> Auke
>
Hey asshole, cut me some slack here, these be uncharted waters fer me and just
tryin' ta help.
I don't see a way to do that with udev rules.... but that's just me.
Run with florin's suggestion. Which is how Lunar *used* to do it.
--
You can tuna piano but you can't tune a fish.
More information about the Lunar-dev
mailing list