On Wed, Sep 19, 2012 at 06:04:48PM +0100, Pawel Moll wrote:

Hi Guenter,

Thanks for your quick response (and apologies about me being delayed), I
appreciate your time!

On Tue, 2012-09-18 at 16:24 +0100, Guenter Roeck wrote:

quoted

and a highly unusual way of, as much as I understand of it, bypass the hwmon
infrastructure as much as possible.

I can assure you that it wasn't my aim. Generally, my platform has a lot
of small "control" devices providing one sensor each. And as they are
separate from the device model point of view, I wanted them to be
logically grouped around a single hwmon device (I was actually looking
at the coretemp driver). But it's not a big deal, really.

I don't think the coretemp driver does what you are doing in your driver ...

quoted

I don't even understand what you are trying to do, much less why you don't
just use the existing infrastructure, and I don't have time to try to figure
it out. Maybe Jean has time to review this driver, but not me.

So, no, for my part I don't think it would be a good idea to rush this driver
into 3.7.

Really, I would suggest to submit a standard hwmon driver (there are lots of
examples out there).

Sure thing, I'll quickly spin a simplified version and post it for
review.

When you do that, please have a look at Documentation/hwmon/submitting-patches,
specifically the "New drivers" section.

Regarding your device model, yes, it is a bit odd that you have a separate
platform device for each sensor, especially if the vexpress is physically a
single chip. It would be much simpler to have a single vexpress-hwmon platform
device instead. But if you insist having such a complicated platform device
architecture, you should not try to clean it up with a complicated and
difficult to understand hwmon driver.

Thanks,
Guenter