On 19.01.2017 22:12, Chris Packham wrote:

On 14/01/17 20:50, Chris Packham wrote:

quoted

On 13/01/17 22:54, Sebastian Hesselbarth wrote:

quoted

On 13.01.2017 10:12, Chris Packham wrote:

quoted

From: Kalyan Kinthada <redacted>

This pinctrl driver supports the 98DX3236, 98DX3336 and 98DX4251 SoCs
from Marvell.

Signed-off-by: Kalyan Kinthada <redacted>
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Acked-by: Rob Herring <robh@kernel.org>
Acked-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
---

Notes:
    Changes in v2:
    - include sdio support for the 98DX4251
    Changes in v3:
    - None
    Changes in v4:
    - Correct some discrepencies between binding and driver.

Well, unfortunately I still see differences between the "gpio" in
the binding and "gpo" in the driver.

Please go back to that list I sent you yesterday and fix them all.

I think you may have missed my initial reply [1]. Or I have missed your
response to it. Long story short "gpo" is intentional because some of
those pins can't be used as inputs. But if you still want me to change
it I will.

[1] - https://lkml.org/lkml/2017/1/12/117

Did you get a chance to consider this. Do you still want me to change 
gpo -> gpio given the information above?

Chris,

sorry if I wasn't clear enough. I don't want you to change every gpo
into gpio. All I was referring to is the _difference_ between driver
implementation and device tree binding - and soley resolve that.

So, for the gpo's I see that the binding doc still says "gpio" for the
available functions where the driver expects "gpo".

e.g. the binding has this:

mpp6          6        gpio, sd0(clk), dev(a2)

if you change it to

mpp6          6        gpo, sd0(clk), dev(a2)

both binding and driver are the same, right?

I do understand that the hardware is gp-output only and you correctly
reflected that in the pinctrl driver - but the binding doc does not
reflect that for those mpps in the list.

Did I make it clearer now or am I still missing the point?

Sebastian