Re: Putting together a computer from old components

Spiros Bousbouras <[email protected]> writes:
> Richard Kettlewell <[email protected]> wrote:
>> Spiros Bousbouras <[email protected]> writes:
>>> SH <[email protected]> wrote:
>>>> what is it about the intel management engine that creeps you out?
>>>
>>> That there is a part of the processor running secret code which has
>>> access to everything on the computer (memory , storage media ,
>>> internet communications) and nothing in the software that you choose
>>> to run on your computer can affect this.
>> 
>> You could say much the same about the CPU microcode or the platform
>> firmware (e.g. UEFI, or BIOS if you can find something old enough).
>
> Sorry , I forgot to reply to that part.
>
> If you mean that CPU microcode potentially has access to the same
> things then yes. But if you mean that it actually does then there's no
> reason to think so. If for example I learned that , microcode of some
> CPU , which microcode ostensibly exists to compute the sine fucntion ,
> tries for access to the network , I would be worried.

It’s not close to things like network interfaces, and it’s not very
large, and it’s not well documented outside CPU vendors, but it does
control the semantics of many machine instructions, so in practice it’s
pretty powerful.

> Regarding firmware , similar considerations apply but I only have a
> vague idea what firmware duties are. But one central criterion is the
> same : do the accesses follow from the nature of its functions or are
> they arbitrary ? If it's the latter , I'd rather avoid the extra risk.

It can interrupt the OS at any time (into SMM) and do anything it
likes. Powerful and, I suspect, a lot more flexible than microcode

-- 
https://www.greenend.org.uk/rjk/