Mailing List Archive

Re: PUBLIC CHALLENGE: (was RE: devfs again, (was RE: USB devicealloc
On 8 Oct 1999, david parsons wrote:
> In article <linux.kernel.Pine.GSO.4.10.9910080409500.10704-100000@weyl.math.psu.edu>,
> Alexander Viro <viro@math.psu.edu> wrote:
>
> >> You don't need ownership, except as root, on a filesystem to boot
> >> a Linux+devfs system. It might not resemble a ``standard'' Unix
> >> system, but one of the spiffy things about a Unix kernel is that
> >> it can be adapted to a wide variety of environment.
> >
> > Ho-hum... /tmp without sticky bit is an interesting animal,
>
> The sticky bit hasn't always been there, and when I was at university
> 20 years ago we had a bunch of machines running BSD 2.7 with /tmp
> being 777.
Umhm... Remember the fun with rm /tmp/12345.s; cp my_code.s /tmp/12345.s;
laugh watching as the sucker goes on with compile ending up with a troyan,
don't you?
> >not to
> >mention the nice stuff in /var (sparse files),
>
> What sparse files live in /var? The big spaceeaters in /var on my
> machines are X11, spool (==mail), and adm (==log), for the machines
> that don't spit syslog off to a logserver.
utmp on a serious multi-user system. lastlog.
> >/usr/bin (suid/sgid not to
> >root),
>
> Why not? /usr/bin survives swimmingly even if it's owned by root.
That is, you've done audit of pine, lpc, screen, wall, write, mail, xterm,
etc. installed SGID-root, right? Would you mind sharing it? Ditto for
suid-root man, if you please. Ditto for the method of storing s[u|g]id in
the first place (doable, but you'll have to screw around with the
attributes).
Hey, wait-a-bloody-minute... What about the ownership of $HOME? OK, you
can mount it with uid=... Ouch. Speak about the binary security model...
> plus fat takes a little less room than ext2fs on a floppy, so that
> plus devfs may be enough so I can shoehorn the bloated elephant that
> will be 2.6 onto my install floppy and still have enough room to fit
> the installer and pcmcia utilities.
BTW, do you realize that 2.0.28 is very likely to have racey FAT? The fix
I did for 2.2 will not fly - it seriously depends on dcache, so you'll
have to roll your own. And I don't envy you - it's a messy work. Really. I
will not volunteer for _that_. Even with dcache it was PITA.
-
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/
Re: PUBLIC CHALLENGE: (was RE: devfs again, (was RE: USB devicealloc [ In reply to ]
> That is, you've done audit of pine, lpc, screen, wall, write, mail, xterm,
> etc. installed SGID-root, right? Would you mind sharing it? Ditto for
> suid-root man, if you please. Ditto for the method of storing s[u|g]id in
> the first place (doable, but you'll have to screw around with the
> attributes).
A lot of that work has already been done actually. With the utmper tool and
with glibc pty helpers screen and xterm are no longer required to be setuid
if set up that way. Pine can be run with IMAP access only and not setgid
root/mail.
-
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/
Re: PUBLIC CHALLENGE: (was RE: devfs again, (was RE: USB devicealloc [ In reply to ]
On Sat, 9 Oct 1999, Alan Cox wrote:
> > That is, you've done audit of pine, lpc, screen, wall, write, mail, xterm,
> > etc. installed SGID-root, right? Would you mind sharing it? Ditto for
> > suid-root man, if you please. Ditto for the method of storing s[u|g]id in
> > the first place (doable, but you'll have to screw around with the
> > attributes).
>
> A lot of that work has already been done actually. With the utmper tool and
> with glibc pty helpers screen and xterm are no longer required to be setuid
> if set up that way. Pine can be run with IMAP access only and not setgid
> root/mail.
David had switched to glibc??? It's sure sign that the end is near ;-)
-
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/
Re: PUBLIC CHALLENGE: (was RE: devfs again, (was RE: USB devicealloc [ In reply to ]
In article <linux.kernel.Pine.GSO.4.10.9910090607360.14121-100000@weyl.math.psu.edu>,
Alexander Viro <viro@math.psu.edu> wrote:
>David had switched to glibc??? It's sure sign that the end is near ;-)
Nope, I've got the sources for libc4, and I'm continuing to develop it.
____
david parsons \bi/ At least until I replace it with the BSD libc.
\/
-
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/