[coreboot] Acer C710-2856 coreboot+seabios blue screen?

John Lewis jlewis at johnlewis.ie
Fri Feb 21 13:08:51 CET 2014


On 21/02/14 02:02, Justin Roylance wrote:
> Ok, so I have downloaded and compiled flashrom from source and it sees
> my Winbond flash chip "W25Q64.V".  When I first installed coreboot I
> backed up my original firmware, but I have two files.  The first is from
> the command "flashrom -r stock-firmware.rom", and the second is from the
> command "chromeos-firmwareupdate --sb_extract ."  Both files are the
> same size, but diff tells me that they are different.  I understand that
> the first one is the stock firmware read straight from the flash chip,
> and the second command gets you a copy of the parts of the current
> firmware which can't be read by flashrom if running on the target
> machine.  So, which one of these two files do I flash back into my C7?

They should be identical save for a missing/invalid serial number from 
the (S)hell(Ball) created ROM. Flash back the backup ROM.

> I read at http://chromebooklinux.blogspot.com/2013_09_01_archive.html
> that you can check the validity of a firmware backup by looking at
> address 0x00001000 of the firmware file, if that byte contains 0xFF the
> backup is invalid, but neither of my two original backup files have 0xFF
> at that address.

Basically, if you backed the ROM up when the write-protect screw/jumper 
was still on, you would have a Management Engine binary which is 
garbage. Even that is okay though, as it won't completely brick your 
Chromebook, it will just switch it off every 30 mins.

Since I haven't completely bricked my chromebook I
> want to make sure I don't flash a bad firmware backup in as this might
> be my last chance to save it without requiring an external programmer.
>
>
I'm still not sure what went wrong with the original flash, it sounds as 
though the PCI id's in the ROM for the graphics card are wrong, except 
that normally you'd get a completely blank screen, not a blue one. I'll 
have to keep an eye out if anyone else reports that issue.

John.



More information about the coreboot mailing list