Google Discloses Another Unpatched Windows Vulnerability

Megalith

24-bit/48kHz
Joined
Aug 20, 2006
Messages
13,000
Google’s Project Zero team has disclosed a vulnerability concerning the Windows GDI (Graphics Device Interface) (gdi32.dll) that allows an attacker to use EMF files to read the content of a user’s memory. The issue was originally reported back in November 16 but was never patched properly, so its details are now out in the wild. If Microsoft didn’t botch their security update this month, this may have been avoided, assuming the fix was part of that package. Another exploit that still needs to be addressed is the malicious SMB one that was unveiled a couple of weeks ago.

Google gives companies 90 days after disclosure of vulnerabilities to fix the issue. If the time period elapses without a patch that is made available to the public, the vulnerability is disclosed to the public. Jurczyk reported the issue to Microsoft on November 16, 2016. Microsoft did not release a patch in time, which is why the system revealed the issue and the example exploit code. Good news for Windows users is that the issue should not be of major concern as it requires access to the machine to exploit the issue. Woody notes that an attacker would have to log on to the machine to execute a specially prepared EMF file to exploit the issue.
 
I think Google is doing the right thing by notifying Microsoft first then letting it be made public later. You can argue whether or not the two months or whatever they give them is truly enough time to fix it and test all possible fallouts of that fix. But I suppose that is something Microsoft and Google will have to decide on behind the scenes. Ultimately it's Microsoft deciding if it's worth fixing or not not Google.
 
I think Google is doing the right thing by notifying Microsoft first then letting it be made public later. You can argue whether or not the two months or whatever they give them is truly enough time to fix it and test all possible fallouts of that fix. But I suppose that is something Microsoft and Google will have to decide on behind the scenes. Ultimately it's Microsoft deciding if it's worth fixing or not not Google.
From what I read over at slashdot, google disclosed it (March 2016), and Microsoft *tried* to patch it once (June 2016), only partially succeeding. So Google let them know again (November 2016), and Microsoft didn't fix it within the 90 days that Google waits before disclosing the bug publicly.
 
From what I read over at slashdot, google disclosed it (March 2016), and Microsoft *tried* to patch it once (June 2016), only partially succeeding. So Google let them know again (November 2016), and Microsoft didn't fix it within the 90 days that Google waits before disclosing the bug publicly.
That's a good example of my point about is that enough time? Windows is a huge complex OS. Changing 1 thing can screw up a lot of other things. Is Google giving them enough time to truly code and test a fix?
 
That SMB issues is so far off the radar for anyone it's nuts to even mention it.
1. Who is going to go through the trouble of setting up a bad SMB share on someone's local network to cause trouble like this? How would they get a targeted person to try to attach to it? It seems like a whole lot of effort for very little effect.
2. Who would be stupid enough to attach to an SMB share over the internet to an untrusted source? Even to an assumed trusted source? A person would have to be awfully stupid to even try it.
3. It only temporarily crashes the machine that attaches to it. It doesn't infect the victim machine, and doesn't cause any permanent damage unless someone had an edited but unsaved document open. So, at worst, it is an annoyance. At best, barely a minor annoyance.
Why should Microsoft even put any effort into fixing it?
 
It's not even a large security issue, it is another one that requires physical access.... an attacker would have to log on to the machine to execute a specially prepared EMF file to exploit the issue. Simple fix, GPO a policy that denying the execution of EMF files and make sure your printers are configured to print directly with out spooling first. It sounds to me more likely this is a series of bugs that exists between two legacy systems, one for graphics and another for printing. Fixing those with out killing thousands of printers in the process is going to be a tough one.
 
Sounds like it needs a re-write then if its so "touchy" to enhance and maintain!

It's a complex system. The question is how long does it take to test and validate. Even if it isn't touchy, you should still be testing and validating even if the answer is "works as expected, nothing broken" 999 times out of 1000 or better.
 
Back
Top