[Arm-netbook] EOMA-68-A20 Card, RTC?

[luke.leighton]

On Fri, Aug 16, 2013 at 4:30 PM, Scott Sullivan <scott at ss.org> wrote:
> Does the first A20 card have a built in RTC?

 yes but no battery, so you'll have to set it each time.  i didn't
want to put a huge battery on-board (space, cost), but we've done a
bit of math and i believe a 100uF capacitor would provide a
significant amount of time (hours) at the 30mA current required by the
A20.

> Follow up questions:
>
> If no to the above, does the Flying Squirrel have an RTC to compensate?

 yes.  STM32F.

> Do future cards currently on the drawing boards have an RTC?
>
> What is the policy on where the RTC should live?

 i didn't really think about this, and it really does have to be part
of the spec.

> The I/O board or the card?
> Is it acceptable that neither have one?
> Which wins if both have an RTC? (luck of kernel driver order?)

 exactly.  chaos rules.  you cannot ever and i do mean ever have
"options" in standards, for *exactly* this reason. if it's optional in
either then it's not optional *at all*, and *all* CPU Cards _and_ all
I/O boards would need to have an RTC.

> I/O Board for all of the above, with no RTC at all being acceptable (think
> router). I favour the I/O board as usually has the luxury of space to
> include it. It's also a part of any design that won't change much.

 well the question we asked ourselves was: is it acceptable to let an
unpowered CPU Card just lose its time setting after say e.g. 24 hours?

l.