[ISN] New Linux Security Hole Found

[InfoSec News]

http://www.eweek.com/article2/0,1759,1612480,00.asp

By Steven J. Vaughan-Nichols 
June 14, 2004    

A Linux bug was recently uncovered by a young Norwegian programmer
that, when exploited by a simple C program, could crash most Linux 2.4
or 2.6 distributions running on an x86 architecture.

"Using this exploit to crash Linux systems requires the (ab)user to
have shell access or other means of uploading and running the program
- like cgi-bin and FTP access," reports the discoverer, Øyvind Sæther.

"The program works on any normal user account, and root access is not
required," Sæther reported. "This exploit has been reported used to
take down several 'lame free-shell providers' servers. [Running code
you know will damage a system intentionally and hacking in general] is
illegal in most parts of the world and strongly discouraged."

Along with the code needed to use the exploit, Sæther also posted
several patches to 2.4 and 2.6 kernels that will keep the exploit from
crashing systems.

Several security problems have been uncovered in Linux over the past
year. The most serious was uncovered in February by the Polish
security nonprofit organization iSEC Security Research.

The biggest of these security holes, called "Linux kernel do_mremap
VMA limit local privilege escalation vulnerability" by iSEC, could
have enabled a cracker to achieve full super-user and full
administration privileges. In each case, fixes were quickly delivered
by the Linux open-source community.

This latest security hole, however, can be used to crash a system, but
it doesn't give an attacker any other control of a Linux system.

Technically, the problem exists because the Linux kernel's signal
handler isn't handling floating-point (FP) exceptions correctly.  
Linux's creator, Linus Torvalds, said, "There's a path into the kernel
where if there is a pending FP error, the kernel will end up taking an
FP exception, and it will continue to take the FP exception forever.  
Duh."

Torvalds already has the problem well in hand, he said. "I fixed it in
my [source code] tree a few days ago, so it's in the current
snapshots, and if I wasn't in the middle of a move [to Portland, Ore.]
I'd have released a 2.6.7 already. As it is, I'll hopefully have it
done by tomorrow [June 15].

Eric Raymond, president of the Open Source Initiative, added, "It
isn't a big deal. This one can be trivially fixed. This fixable kernel
crasher doesn't cause any new problems."