[coreboot] Who has experience with the Intel RVP7 (or RVP15) CRB?

Nico Huber nico.h at gmx.de
Sat Nov 4 20:59:54 CET 2017


On 04.11.2017 20:31, Jay Talbott wrote:
>> Yes, tell us what you do about the other (non-host) firmware components
>> in your flash chip (e.g. Firmware Descriptor, ME Firmware). And the
>> tool (command line if you use flashrom) you use to flash. Often the
>> culprit is a broken descriptor.
>>
> 
> I left the factory descriptor, ME, etc. that came with the original UEFI
> BIOS intact. I just built a 2MB coreboot image, not a complete IFWI. I
> then just wrote my coreboot binary to the end of the BIOS region (after
> first having saved the whole IFWI from the flash as a backup). This
> works on other CRBs (and always has since the inception of the FSP), at
> least to initially get coreboot up and running without having to worry
> about the other firmware components.

Hmm, sounds pretty much like my own workflow.

> 
>> If you are sure that everything in the flash chip is perfectly fine (to
>> be sure, ME firmware should be a fresh image that wasn't ever altered
>> during runtime (ME does that), i.e. not a dump of an ever run
>> firmware), I'd suggest to remove all power, take a break, eat a Pizza
>> or watch a movie and try again later.
>>
> 
> Before changing to XIP, doing the above worked fine to get coreboot to
> begin booting, but then it was hanging due to the signature mis-match.
> But once the signature mis-match was fixed by enabling XIP, and
> execution proceeded into Fsp_M, that's when the board then became
> non-functional. Restoring the original IFWI with the UEFI BIOS has
> always worked to bring a CRB back to life, essentially restoring things
> to where they were before introducing coreboot, but not in this case...

I'm not sure what IFWI refers to. I thought it's only about the host
firmware (i.e. the BIOS region of the flash). Make sure the other
regions are restored as well. If you have an image of the factory
firmware (not a backup), I'd try to flash that.

The FW can theoretically damage the hardware, but it's unlikely. If
you haven't tried that yet, I'd remove the CMOS battery, too. Some
registers are powered by it and could still hold a value set by core-
boot/FSP. A warning, though, I've encountered one board where this
made things worse (but the board survived).

Nico



More information about the coreboot mailing list