Intel chip flaw gets double exposure
Security researchers converge on cache vuln.
Security researchers are due to publish research on how an Intel chip flaw might be used for potentially malign purposes on Thursday.
Both Joanna Rutkowska, of Blue Pill renown, and Loic Duflot have separately announced plans to release a research paper and proof of concept exploit code on how an Intel CPU caching vulnerability might be misused. The flaw creates a potential means to attack the SMM (System Management Mode) of systems. Duflot will present his work at CanSEcWest in Vancover while Rutkowska is due to publish a paper on the same subject, also on Thursday.
"The attack allows for privilege escalation from Ring 0 to the SMM on many recent motherboards with Intel CPUs," Rutkowska, chief exec of of Invisible Things Lab explains.
The Intel caching hack is all the more serious because it burrows into SMM space, which is more privileged than a hypervisor area. As such it is a possible mechanism for creating "incredibly stealthy" rootkits, security blogger James Heary adds.
Invisible Things Lab's paper touches on the long-running debate about security disclosure. According to Rutkowska, Intel's own employees first wrote about how this class of CPU caching vulnerability might be exploited back in early 2005.
Duflot came across the problem and was the first to write a proof-of-concept exploit in October 2008. Rutkowska independently discovered the same issue a couple of months later.
"If there is a bug somewhere and if it stays unpatched for enough time, it is almost guaranteed that various people will (re)discover and exploit it, sooner or later," Rutkowska writes. "So, don't blame researchers that they find and publish information about bugs — they actually do a favor to our society." ®
Why oh why do *nix fanbois' go on about bloatware being an MS problem?
last time i installed lenny it was over a gig in size, installed 12 different text editors, 3 different GUI front ends, 4 different sound engines, 4 different programmers kits for everything from C to Java, and god knows what else.
bloatware is everywhere and linux is just as guilty as windows.
yes, you can prune linux down, but you can trim windows down too. as a standard install lenny installed more useless crap on my system than vista did. At least vista only gave me one of each application instead of 4 or 5.
Coat on back, runnin' like hell ;)
That's my reading of it too - as the researchers say you essentially require admin access to a vulnerable machine to exploit this. However if this payload (in the form of a rootkit) was piggy-backed on some plain vanilla malware it could be very nasty.
The ability to conceal itself no-holds-barred is precisely the point: that after the regular malware is removed and the vulnerability patched the machine is still pWn3d.
Presumably if a fix is possible it will be a BIOS update for mainboards like the Intel DQ35 and, as this board will be the basis of many thousands of OEM machines, it is rather shocking that Intel haven't moved faster to fix it.
But what about mainboards based on the 965/945/915 chipsets? There's still plenty of those around - are they vulnerable too? And third-party chipsets?
There's more at http://isc.sans.org/diary.html?storyid=6046 and even more on a PDF that the diary entry links to.
To summarise, the SMM code lives in a protected area of memory, but if you are already running at ring 0 you can trick the processor into thinking that this area of memory is cached and you can then prime the cache with your exploit prior to triggering an SMM interrupt.
It's cute, but if your system already has malware running at ring 0 then the ability of that malware to further conceal itself is the least of your worries.