Core Security Technologies says it's discovered an IPv6-related buffer-overflow vulnerability in several versions of the OpenBSD operating system that would allow an attacker to knock a server offline or take control at the kernel level. The OpenBSD project, the group which makes the free operating system available, has issued a software patch to plug the hole.

According to Ivan Arce, CTO at Core Security, the firm discovered the IPv6-related vulnerability by chance as it was closely examining its own OpenBSD Servers.

"We were trying to replicate some odd behavior we saw with the servers crashing," Arce says. The vulnerability Core Security found in OpenBSD 4.1, 4.0, 3.9, 3.8, 3.6 and 3.1 -- other versions of OpenBSD that support the IPv6 protocol stack may also be vulnerable, the company adds -- would allow an attacker to add arbitrary code at the kernel level to subvert any security mechanisms.

"This allows the attacker to run whatever code he wants at the kernel level," Arce says, adding this may be the first buffer-overflow flaw discovered related to IPv6 in the OpenBSD operating system.

As a workaround to applying a patch, OpenBSD users with vulnerable systems could elect to block IPv6 packets using OpenBSD's firewall if they have no need to route IPv6 traffic, Core Security advises. The vast majority of IP-related traffic today is based on the older IPv4 protocol.

Arce says the U.S. government's mandate to support IPv6 traffic by mid-June 2008 is putting greater attention on IPv6-security-related issues. Core Security's penetration-testing tool, called Core Impact, is designed only for IPv4-based networks. However, Core has added this one detection for the IPv6-related flaw it discovered to Core Impact and will be looking at developing fuller IPv6-related products.