[flashrom] add Numonyx/Micron N25Q128

[Stefan Tauner]

On Wed, 11 Jul 2012 16:10:49 +0200
frederic.temporelli at bull.net wrote:

Hello Frederic,

> the main idea is to unlock the regions from the ME
> Document with datails about ME messages for X79/C600 are Intel 
> Restricted... :-(

that is/was our main idea too, but since we could not obtain the needed
docs and my REing skills are not as good as needed there was little
progress in that direction. i have documented most of what i could find
out in our documentation, see
http://flashrom.org/trac/flashrom/browser/trunk/Documentation/mysteries_intel.txt#L17
if you can spot any errors or can provide more detail i am happy to
refine (and use) it.

> But seems that ME messages (HMRFPO and others) used in Maho Bay and Chief 
> River platform are the same,
> and "BIOS Writers Guide" for these systems (document 490124) is only 
> "Intel Confidential" (and HMRFPO is described in this document...)
> I hope that Intel will soon set all these documents available to 
> everybody...

hope does not really help... :) do you have any clear indications that
they might open them up?

> note: there are lot of work to do... first, there's only one IOCTL 
> developed in mei driver (mei driver may be the best way to interact with 
> ME...)

i have an almost complete (but unpublished) user-space reimplementation
of the parts of the linux kernel MEI driver that is needed to send the
HMRFPO message from flashrom... if i would know the GUIDs and payloads
to send (and receive on success), there could be ME unlocking support
in a matter of days i think.

> Other topic (may be more related to layout feature): currently, when using 
> flashrom for BIOS upgrade, Serial Numbers and some other information in 
> DMI are overwritten...
> (seen on Supermicro motherboards)
> I don't have details, but it may be nice to investigate this issue...

yes this happens sometimes but flashrom is not the right place to fix
it. flashrom has the goal to read and write images to flash chips. if
the images are not suited for the system the flash chip is used
afterwards then it is per definition not our problem.
one could of course use (lib)flashrom as a backend and create something
that downloads and prepares correct images for all possible systems...
but the person implementing this is not in sight yet ;)

so the most realistic approach for such problems for now is that you
prepare such images by hand or some customized tools/binary diffs,
sorry that there is nothing better than that yet.
-- 
Kind regards/Mit freundlichen Grüßen, Stefan Tauner