home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!mcsun!news.funet.fi!hydra!klaava!torvalds
- From: torvalds@klaava.Helsinki.FI (Linus Benedict Torvalds)
- Newsgroups: alt.hackers
- Subject: Re: My First Post.
- Message-ID: <1992Aug25.214056.11488@klaava.Helsinki.FI>
- Date: 25 Aug 92 21:40:56 GMT
- References: <BtJzvF.6Bx@mentor.cc.purdue.edu>
- Organization: University of Helsinki
- Lines: 26
- Approved: most emphatically
-
- In article <BtJzvF.6Bx@mentor.cc.purdue.edu> devros@mentor.cc.purdue.edu (Devros) writes:
- >Well, Here goes....
-
- And on two /long/ lines..
-
- >ObHack: On the local SUN net, format a disk to SunOS and fdmount it.
- >Copy a shell to it and change permissions to setUID, setGID, and world
- >executable. Put the disk in a computer with a disk editor that you can
- >name the media type without a boot sector and find your UID on the disk,
- >changing it to a root uid. Mount it back onto the SUN and execute. I
- >haven't tried it, but theoretically you would get a setuid-root shell.
-
- "Theoretically" - if it indeed works, your sun setup isn't exactly
- secure. It's pretty easy to disallow suid programs, and I'd assume
- fdmount does exactly that, leaving you as badly off as you were from the
- beginning. Now if you had made the disk by doing a "cat > /dev/fd0",
- /that/ would be a hack (but it should still not work).
-
- Obhack: while debugging some weird VM bugs in the linux code, I used
-
- od -x /dev/kmem | less
-
- to find the problem. It worked too: it was a stray NULL pointer that
- corrupted the page directory. Who needs a kernel debugger anyway?
-
- Linus
-