Vulnerabilities / Threats
5/11/2011
03:13 PM
Connect Directly
Google+
Twitter
RSS
E-Mail
50%
50%

Google, VUPEN Spar Over Chrome Hack

If bypass of Chrome's sandbox indeed used a new Flash vulnerability in the browser, then it's both a Flash bug and a Chrome hack, says security researcher Dan Kaminksy

The Google Chrome browser hack revealed this week could come down to a dispute over third-party software: A Google security expert says a zero-day flaw discovered by VUPEN Security researchers to bypass the browser's sandbox actually lies in Adobe Flash. But VUPEN isn't budging on its claims that its researchers successfully "owned" the browser software itself.

But for now, only VUPEN and its "three-letter" government agency customers know the details about the two zero-day vulnerabilities that VUPEN says it exploited and successfully used to bypass Chrome's sandbox and other security features.

"Nobody knows how we bypassed Google Chrome's sandbox except us and our customers, and any claim is a pure speculation," says Chaouki Bekrar, CEO and head of research at VUPEN. "We will not reveal the details on how we achieved the full compromise of Chrome and its sandbox. We can, however, say that we did use the default and built-in attack surface of the browser without using a Windows kernel vulnerability. All users of Chrome should be aware now that this browser can be fully hacked despite its famous sandbox and despite all the marketing that Google has been doing around its security."

Bekrar would neither confirm nor deny whether his firm found the bug in Chrome's Flash plug-in implementation, which comes with the browser. But he did say this when asked about it: "Google Chrome sandboxes all its default plug-ins; thus, exploiting a plug-in vulnerability is not enough to bypass the sandbox."

Google, meanwhile, is challenging VUPEN's claims of a bug in Chrome. Google security researcher Tavis Ormandy said in a Twitter post that "VUPEN misunderstood how sandboxing worked in chrome, and only had a flash bug."

A Google spokesperson said the Chrome team is still investigating VUPEN's claim that it hacked Chrome's famed sandbox feature. Chrome's sandbox, along with its use of Microsoft's DEP and ASLR technologies, thus far has made the browser relatively impenetrable to hackers.

Aside from a video and blog posting on the attack, VUPEN has kept details of the flaws and exploits closely held: As of this posting, Google hadn't received them. Bekrar did reveal yesterday in an interview that VUPEN employed one zero-day that's exploited inside the sandbox, and another that's executed outside of it. "The first one results from a memory corruption leading to the execution of the first payload at low-integrity level, inside the sandbox," he says. "A second payload is then used to exploit another vulnerability, which allows the bypass of the sandbox and execution of the final payload with medium-integrity level, outside the sandbox."

But tweets today from Google's Ormandy appear to point to Flash as containing a vulnerability, not Chrome.

So what gives?

Both VUPEN and Google are technically right, says security expert Dan Kaminsky, who has been analyzing VUPEN's public disclosure and the debate that unfolded on Twitter between the two companies today. "VUPEN didn't break the primary Chrome sandbox. It might have broken a secondary Flash sandbox. But at the end of the day, the [Chrome] browser got owned," Kaminsky says. "This is sort of a rare, nuanced situation where everyone is right."

Kaminsky applauds Google's efforts to lock down Chrome. He says the issue with the VUPEN research is that it goes back to the issue of using and keeping third-party software safe. "Everyone has been scratching their heads for years about what to do about insecure plug-ins," he says. Google took Flash code "in-house," for example and wrote a PDF parser for PDF plug-ins.

Google basically assumed the responsibility of keeping its browser users as safe as possible with Flash, Kaminsky says. "Chrome has to push an update of Flash ... because [of] the code they shipped that's buggy," he says. "If you ship it in your browser and it's on by default, it's your responsibility. Sometimes you're going to get burnt because of it."

Kaminsky says that because Flash "breaks" when placed in the browser's primary sandbox, it's not protected by it. "[Google is] still working on a version of Flash that doesn't break when put in the sandbox," he says.

Have a comment on this story? Please click "Add Your Comment" below. If you'd like to contact Dark Reading's editors directly, send us a message.

Kelly Jackson Higgins is Senior Editor at DarkReading.com. She is an award-winning veteran technology and business journalist with more than two decades of experience in reporting and editing for various publications, including Network Computing, Secure Enterprise Magazine, ... View Full Bio

Comment  | 
Print  | 
More Insights
Register for Dark Reading Newsletters
White Papers
Flash Poll
Current Issue
Cartoon
Video
Slideshows
Twitter Feed
Dark Reading - Bug Report
Bug Report
Enterprise Vulnerabilities
From DHS/US-CERT's National Vulnerability Database
CVE-2012-4988
Published: 2014-07-09
Heap-based buffer overflow in the xjpegls.dll (aka JLS, JPEG-LS, or JPEG lossless) format plugin in XnView 1.99 and 1.99.1 allows remote attackers to execute arbitrary code via a crafted JLS image file.

CVE-2014-0207
Published: 2014-07-09
The cdf_read_short_sector function in cdf.c in file before 5.19, as used in the Fileinfo component in PHP before 5.4.30 and 5.5.x before 5.5.14, allows remote attackers to cause a denial of service (assertion failure and application exit) via a crafted CDF file.

CVE-2014-0537
Published: 2014-07-09
Adobe Flash Player before 13.0.0.231 and 14.x before 14.0.0.145 on Windows and OS X and before 11.2.202.394 on Linux, Adobe AIR before 14.0.0.137 on Android, Adobe AIR SDK before 14.0.0.137, and Adobe AIR SDK & Compiler before 14.0.0.137 allow attackers to bypass intended access restrictions via uns...

CVE-2014-0539
Published: 2014-07-09
Adobe Flash Player before 13.0.0.231 and 14.x before 14.0.0.145 on Windows and OS X and before 11.2.202.394 on Linux, Adobe AIR before 14.0.0.137 on Android, Adobe AIR SDK before 14.0.0.137, and Adobe AIR SDK & Compiler before 14.0.0.137 allow attackers to bypass intended access restrictions via uns...

CVE-2014-3309
Published: 2014-07-09
The NTP implementation in Cisco IOS and IOS XE does not properly support use of the access-group command for a "deny all" configuration, which allows remote attackers to bypass intended restrictions on time synchronization via a standard query, aka Bug ID CSCuj66318.

Best of the Web
Dark Reading Radio
Archived Dark Reading Radio
Marilyn Cohodas and her guests look at the evolving nature of the relationship between CIO and CSO.