On Sat, 26 May 2012 09:58:06 -0700, Dong Aisheng [off-list ref] wrote:

On Fri, May 25, 2012 at 5:29 PM, Grant Likely [off-list ref] wrote:

quoted

On Fri, 25 May 2012 21:36:20 +0800, Dong Aisheng [off-list ref] wrote:

quoted

From: Dong Aisheng <redacted>

This patch implements a standard common binding for pinctrl gpio ranges.
Each SoC can add gpio ranges through device tree by adding a gpio-maps property
under their pinctrl devices node with the format:
<&gpio $gpio-specifier $pin_offset $count>
while the gpio phandle and gpio-specifier are the standard approach
to represent a gpio in device tree.
Then we can cooperate it with the gpio xlate function to get the gpio number
from device tree to set up the gpio ranges map.

Then the pinctrl driver can call pinctrl_dt_add_gpio_ranges(pctldev, node)
to parse and register the gpio ranges from device tree.

Signed-off-by: Dong Aisheng <redacted>
---
Personally i'm not very satisfied with current solution due to a few reasons:
1) i can not user standard gpio api to get gpio number
2) i need to reinvent a new api of_parse_phandles_with_args_ext which i'm not
sure if it can be accepted by DT maintainer.

Right, as mentioned in my other email, doing it this way completely
breaks the way the phandle-with-args pattern works. ??That pattern
depends on the phandle node to have a #-cells property telling it how
many cells to process for the binding. ??Adding additional data cells
means the kernel is no longer able to parse multiple entries in the
gpios property.

Hmm, it can still parse multiple entries in the gpios property except
that it adds two args although it's not related to gpio, but it is useful
for users for special case like pinctrl gpio ranges map.

Really?  How exactly does it know that each record is longer than
#gpio-cells specifies (I'm speaking from the binding level; not having
custom code that just "knows" the the records have additional
padding).

I have no interest in creating exceptions to the phandle-with-args
pattern since it adds yet more implicit knowledge about how to parse.
For example, the common gpio code can no longer parse a gpios property
that is padded out because the common code doesn't know anything about
padding.

g.

quoted

Hmmm.... I need more information about this gpio-maps property. ??How
is it arranged? ??What kind of data is in it. ??Can you give some
specific examples of how hardware would be described with a gpio-maps
property?

For exampe:
MX6Q_PAD_SD2_DAT2__GPIO_1_13 means MX6Q_PAD_SD2_DAT2 can be used as GPIO_1_13,
For reference gpio1,13, we usually do: xx-gpios = <gpio1 13 0> in device tree.
Here we want to create a pin map of gpio1,13 to MX6Q_PAD_SD2_DAT2 for
pinctrl gpio ranges map,
the format should be <GPIO_NUMBER PIN_ID NPINS>, then the pinctrl core
can automatically mux
the PIN_ID to gpio function by refer to this map.
For GPIO_NUMBER, we want to use the standard gpio dt represent way
since the gpio base may
be dynamically.
Assume MX6Q_PAD_SD2_DAT2 pin id is 1 and only one pin starting from it
can be used as gpio.
Then the gpio-maps for MX6Q_PAD_SD2_DAT2 can be:
gpio-maps = <gpio1 13 0 1 1>

We may have several pins can be used as gpio on mx6q.
Then the gpio-maps may becomes:
gpio-maps = <gpio1 13 0 1 1>,
                   <gpio1 14 0 5 1>,
                   <gpio2 0 0 20 1>,
                   ................

Since the format is a little different from the standard gpio
represent way, so i can not use the standard gpio
api to parse the gpio number. That's why i need to invent
of_parse_phandle_args_ext function for this special
format.

we still did not find any better way to do that.
Do you have any suggestion for this special case?

Oh, I see....  Does this gpio-maps property sit beside a normal
"gpios" property?  Or is it in a completely separate node?  If it sits
beside a normal "gpios" property and lines up with the gpio properties
there, then I would just make it a tuple for each gpio.  Ie:

gpios = <&gpio1 13 0>, <&gpio1 14 0>, <&gpio2 0 0>;
gpio-pinmux = <1 1>,   <5 1>,         <20 1>;

g.