CPUd
Member
- Joined
- May 12, 2012
- Messages
- 22,978
Thanks.
So malware affecting your OS can be detected and fixed by anti-virus software. Malware affecting your drivers should be able to be detected and fixed by anti-virus software (and if not, easily downloaded from the vendor's website). But anti-virus software doesn't look at firmware I guess. And that's where the problem is? And if true, what would be the disadvantage of using nonprogrammable eproms for their firmware as a solution to stop malware from infecting firmware?
They often need to be able to save and modify variables locally, especially diagnostic data. Also, the manufacturer needs a way to make future updates if necessary.
Antivirus programs could look at firmware, but with something like the stuffs described in the article, it would not find anything wrong if you're running it from the OS. With a special utility, you can test if the firmware checksum matches the checksum provided by the manufacturer, and you can flash a new one, so it's kinda all or nothing at this level. This is because it can't otherwise look directly at hardware like that, it is like if you message someone (assuming no encryption) and your messsaging app is compromised. A 3rd party could be watching both ends of the conversation. The 3rd party could suppress the other person's message and insert their own, if they do it carefully, you would never know you were talking to a 3rd party. There are 2 problems here:
1.) finding out you are not talking to the person you think you are
2.) stopping it
If your 3rd party is clever enough, it is nearly impossible for you to detect something is wrong. It gets easier if you are aware beforehand that this has been happening to others. To address #2, you reinstall with a known clean version of the messaging app.
I don't understand. tangent4ronpaul wrote that: The malicious firmware created a secret storage vault [on your harddrive]. cpu'd wrote that: the secret storage vault would be incorrectly labelled a bad sector by the software because it was unable to read or write to it.
So if the firmware creates what the OS thinks is a bad sector, and chkdsk checks for bad sectors, why wouldn't a successful running of chkdsk (that shows you have no bad sectors) prove your firmware is not infected by this malware?
I was just using the bad sector thing as an example, there are other ways, like moving it around on unallocated space.
If the drive is reporting no bad sectors, then no, there is no secret storage vault masquerading as a bad sector. Drives older than a year are probably going to have bad sectors.
The way it was described in the article, this is a sophisticated group of programs that can act like a Swiss army knife; if one attack vector is not feasible, it can use another. An example: your AV program might be able to detect/remove part of it in the regular file system, or running in memory but when you reboot, before the OS even loads the drive firmware could copy stuff back to the file system, including infected system files that are made to appear OK to the OS and AV programs. You could reflash the firmware, but if you miss something in the OS, it could reflash again with their firmware.
It is a vicious cycle with a lot of moving parts that aren't fully understood yet. The researchers are still working out exactly what is happening, and might take a few weeks before they start releasing tools for removal.
Last edited: