Re: more evil firmwares found

[Nathanael Nerode <neroden@twcny.rr.com>]

Adam Majer wrote:
> But for other things, you can write things like for the programmable DAC
> converters (or ADC for that matter). It is actually *easier* to write
> stuff like that in hex than attempting to write/modify another
> assembler. Of course, first you have to write a program for something
> like the Cypress EZ-USB FX chip so that it loads the program onto the
> DAC or ADC - ie. firmware loads more firmware. But that can be done in
> assembly or even C.
> 
> Some chips only have a handful of op codes (like 4 or whatever). The
> entire programs can be only a few to a few dozen of bytes.

I don't think anyone's been complaining about really short hex blobs, under
the "give them the benefit of the doubt" rule.  What we've been looking at
are long (and patternless) enough that it's really hard to believe they
were edited in hex.  Furthermore, the lack of comments is especially
suspicious.  (And it's even more suspicious when the maintainers of the
driver can't explain what the microcode actually does.)

Note for contrast that there is some 'hand-assembled' x86 16-bit machine
code in the kernel sources, but it has comments explaining what the
corresponding assembly is, and what it does.  :-)

-- 
There are none so blind as those who will not see.