The second Windows Vista vulnerability patched in the last week may be very difficult to exploit, a security researcher said Wednesday, because of the new operating system's additional defenses.
"This is a 'double-free' vulnerability, which is significantly harder to exploit in Windows Vista due to [its] heap protection mechanisms," said Minoo Hamilton, senior security researcher at nCircle Network Security.
Double-free vulnerabilities are those where an address pointer is freed twice; before the introduction of Windows XP SP2, double-free bugs were relatively easy to exploit, according to researchers like Symantec's Matt Conover. But various protections introduced in Vista or carried over from XP SP2 -- including randomized heap base -- should make it much tougher for attackers to craft a workable exploit, at least for now.
Hamilton explained: "The actual research into how these protections work hasn't been done, so they're not fully understood yet," he said. "A lot of research has to happen before reliable exploits against double-free vulnerabilities can be created."
The bottom line, he said, is: "There is a low likelihood of a Vista exploit for the 'MsgBox (CSRSS) Remote Code Execution Vulnerability' in MS07-021."
Something that hasn't changed, however, is Hamilton's opinion of Vista. "I thought I was going out on a limb when I talked about Vista and vulnerabilities, but now I feel different," he said. "I feel kind of vindicated, what with two Vista vulnerabilities in just one week."
Still, the caveat holds, he said. "But it may take time for researchers to understand Vista and how its vulnerabilities can be exploited."
Tuesday's Vista vulnerability was also notable because it was present in Beta 1 and Beta 2, which were heavily distributed and downloaded in 2006, as well as in the final editions rolled out in the beginning in November. However, it may not have been in the release candidate (RC) editions between the betas and final. (In tests, Computerworld discovered that machines running final Vista code got the MS07-021 update, while systems running RC2 (Release Candidate 2) did not.)
"I don't think we can confirm that with a high degree of certainty," said Vince Hwang, group product manager of Symantec's security response team. "But this was initially disclosed Dec. 15, 2006, so it's quite plausible that it had always been there through all the betas."
When asked to explain how the vulnerability escaped notice -- unlike the animated cursor (ANI) bug patched last week, the MsgBox flaw went public and was even accompanied by proof-of-concept code four months ago -- Microsoft said only that it hadn't considered the flaw important enough to fix post haste. "In this case, Microsoft was not aware of any active attacks and did not feel the level of malicious activity around this vulnerability prompted an out-of-cycle release," a company spokeswoman said in an e-mail exchange today.
During the follow-on webcast that the Microsoft Security Response Center (MSRC) regularly conducts the day after updates debut, MSRC program manager Christopher Budd offered a few details of the MS07-021 timeline. "Originally, the public discussion was that it was only a local elevation of privilege. That's how it was reported and publicly discussed," said Budd. "But an internal investigation did find the possibility of code execution."
Ironically, the MsgBox bug was acknowledged by Microsoft on Dec. 22 2006, just two days after Determina researcher Alexander Sotirov reported the ANI issue to Microsoft.