On Tue, 23 Dec 2014 00:55:47 -0800
Matt Taggart <matt(a)lackof.org> wrote:
Replying to just you to keep noise off the list.
Please don't. The topic is quite generic and posting it on the list may
save me some time in the future and discussing such stuff is exactly
what the mailing list is for.
Stefan Tauner writes:
That's not unusual for various reasons (e.g.
the firmware uses the
flash as scratchpad). The rom.bin on your website looks sane to me and
since I don't expect any problems regarding reading the chip with
flashrom, I will mark probing and reading as tested, thanks.
I would like to write test this board and I wanted to run my idea for a
testing procedure by you.
I managed to find an older vendor BIOS version and can use the vendor DOS
flash tools to switch between them. I have used flashrom to read an image
of each. So how about trying this:
1) run 'flashrom -p internal -r rom1.bin' to get a dump of the current
2) run 'flashrom -p internal -w rom1.bin' to write the same version back,
and hopefully get back a report that they are identical
3) run 'flashrom -p internal -w rom2.bin` to write the vendor's other
version, and hopefully the write succeeds
4) reboot and confirm running the rom2.bin version
5) run 'flashrom -p internal -w rom1.bin' to put the original back
6) reboot and confirm running the rom1.bin version
7) if anything went wrong, buy an external clip and use it to write one of
the images back. Since we think the reading is working, should the images I
have be the same as what I'd want to write with an external? (or would I
need an image originally read with the external?)
Usually boards either allow to access the chip or they don't. So in
theory one could simply do
- flashrom -r backup
- flashrom -E
- flashrom -w backup
Practically erase + write is slightly more risky than flashing another
good image directly because flashrom writes in the smallest block size
possible and that sometimes saves one from a brick.
If you plan to hack on coreboot you should get a programmer anyway (I
recommend a BBB because it is very versatile for coreboot development
(SPI programmer (even easily convertible to 1.8V), USB debug).
Images read with flashrom from a chip should always be identical no
matter which programmer is used. The only reasons for differences are
bugs and firmware that writes to the flash while booting (introducing
discrepancies between the flash contents while powered on and off,
which are usually benign... because the firmware must be prepared for
crashing OSes ;)
Hope that answers your questions.
Kind regards/Mit freundlichen Grüßen, Stefan Tauner