[coreboot] call on AMD to release src+specs+datasheets for ryzen

Taiidan at gmx.com Taiidan at gmx.com
Sat Mar 4 17:54:25 CET 2017

On 03/04/2017 06:39 AM, Nico Huber wrote:

> On 04.03.2017 02:57, Taiidan at gmx.com wrote:
>> Of course they also must release the signing keys as well afaik, or we
>> would be stuck at a tivo style not really open source impasse.
>> Nobody has mentioned this fact in that thread.
> Please don't ask for that.
> If somebody put a signature verification for his firmware in place, you
> should first discuss the reasons and alternatives (for the particular
> design in question). Sure there are alternatives to signature verifi-
> cations to put some trust in hardware (like ROMs or the RO partitions
> in cros devices). But removing the security checks from hardware who's
> trust is designed around these checks? You'd likely end up with a sys-
> tem where you have to check the flash contents with external hardware
> before every boot (if it can be tampered with from the running system).
> Of course you can ask for alternatives in new designs.
> For yet released platforms, however, it's more feasible to ask for docu-
> mentation, reproducible binaries and signatures (e.g. for fixes / reim-
> plementations).
> Nico
I am simply stating that source code is pointless without the ability to 
flash it and have the hardware execute it.

The issue isn't that there are signing keys in the first place (which 
are common sense to prevent rogue BIOS updates) it is that the hardware 
enforces them for manual external flashes.

Your idea isn't a free platform, it isn't owner controlled because you 
can't modify it - you can only be on the outside looking in.

More information about the coreboot mailing list