Spoke with Steve,
He says:
What is described there does not appear to be a problem with the FAX
protocol, even though the text says it is. They refer to problems
handling DHT (define huffman table) and COM (comment) sections when they
are used incorrectly. Those are tags in JPEG images. It looks like that
are talking about an issue with JPEG decoders that don't allow for
certain types of corrupt image, and the exploit would be specific to
particular models of FAX machines. The problem would be in the JPEG
decoder itself. Spandsp calls the system's JPEG library to do that
decoding. Any flaw that might exist would be in that system's JPEG library.
Steve and Brian,
thanks.
Yes, that's it:
While their research was undoubtedly eye-opening, similar headlines and statements claiming that the “T.30 protocol itself has been compromised”
are categorically false; a bit of drama designed to make headlines.
More specifically, the researches were able to demonstrate the Faxploit by exposing
a flaw in software developed internally at HP, rendering it vulnerable when handling color/JPEG faxes.