Re: Example idea for how to solve the clock/cnf problem.

[Magnus Damm <magnus.damm@gmail.com>]

Hi Ian,

On Thu, Aug 6, 2009 at 7:43 AM, Ian Molton<ian@mnementh.co.uk> wrote:
> Ooops. Patch attached this time.
>
> Please comment on this. (its a WIP remember, so trivial stuff is to be
> ignored - I wanna get the functionality right first.)
>
> Note in particular the change in the suspend/resume paths - we no longer
> (ab)use the enable/disable hooks, which may break some users of the driver.
>
> I havent decided how I'm going to map the conf area in the MFD drivers in a
> nice way yet. As there are no known devices with TWO of these chips in them
> yet, I may do a static one-off mapping for now. At least this will leave
> tmio-mmc.{c,h} pure and free from all CNF area code.

Thanks for your work on this. There is a number of resources check
that is still fixed at 3 instead of 2, but apart from that all is ok
from the SuperH side of things.

In case you missed it, please see my earlier comments and measurements
in this email:

http://lkml.org/lkml/2009/8/18/67

Do you have any updates queued up for this patch, or do you want me to
submit my fix on top of this one?

Cheers,

/ magnus