R70/R71 on debug v3 board ?

Werner Almesberger werner at openmoko.org
Mon Feb 11 13:09:35 CET 2008


Andy Green wrote:
> How about if it doesn't ship with flash_unlock?

There may still be some installer that does it. Or some other code may
unlock it by accident. (Unlikely, but then you never know.) People may
suggest - in a honest attempt to help - upgrade procedures that involve
re-adding the unlock. Or even some malicious code could do it.

The bottom line is that users should never have to update the NOR. It's
there to fix the system when the NAND is screwed up, but that's all.
What's important is that it's reliably there in the hour of need. The
less we have to think about ways in which a software protection could
be bypassed, and how to thwart them, the better ;-)

Now, we will continue to make improvements, but even without them, a
system will be perfectly recoverable with what we'll put there in
production. The code involved is something we use all the time and,
for most of it, have done so for many months, so I think the risk of
a crippling bug that requires in-the-field upgrades sneaking in is
pretty low.

Hard-core developers who also want to play with NOR will probably get
the debug board anyway. And even if they don't, they could just tie
H-TP4711 to a positive supply (Vpp is 12V-tolerant).

> I can see the logic, but I can also see it puts us in a unique position
> with it for the next couple of weeks until RTM, we're not going to be
> able to generally update whatever it is we put in there subsequently.

Hmm, I didn't understand that one. Who wouldn't be able to update
the NOR, I suppose, when ? (BTW, my question about R70/71 wasn't
rethorical - all my debug v3 boards are some R&D prototypes, so I
really don't know what's on the production units.)

- Werner




More information about the openmoko-kernel mailing list