Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1610492 > unrolled thread

Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Linus Walleij <linus.walleij@linaro.org> - 2017-03-28 11:50 +0200
    Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc jacopo@jmondi.org - 2017-03-28 16:50 +0200
      Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Linus Walleij <linus.walleij@linaro.org> - 2017-03-29 04:40 +0200
        Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Geert Uytterhoeven <geert@linux-m68k.org> - 2017-03-29 09:40 +0200
          Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Linus Walleij <linus.walleij@linaro.org> - 2017-03-29 12:20 +0200
            Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Geert Uytterhoeven <geert@linux-m68k.org> - 2017-03-29 13:30 +0200
              Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc jacopo <jacopo@jmondi.org> - 2017-03-29 14:10 +0200
                RE: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Chris Brandt <Chris.Brandt@renesas.com> - 2017-03-29 14:40 +0200
                  Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Linus Walleij <linus.walleij@linaro.org> - 2017-03-29 15:20 +0200
                    RE: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Chris Brandt <Chris.Brandt@renesas.com> - 2017-03-29 16:20 +0200
              Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc Linus Walleij <linus.walleij@linaro.org> - 2017-03-29 15:10 +0200

#1610492 — Re: [PATCH v2 2/7] dt-bindings: pinctrl: Add RZ/A1 bindings doc

On Thu, Mar 23, 2017 at 5:02 PM, jacopo <jacopo@jmondi.org> wrote:

>> > +  Required properties:
>> > +    - renesas,pins
>>
>> Just "pins"?
>>
>
> You know, I've been thinking about this, bu the "pins" property
> definition in pinctrl-bidings is the following one:
>
> Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt
> ---
> - pins takes a list of pin names or IDs as a required argument. The
>   specific binding for the hardware defines:
>       - Whether the entries are integers or strings, and their
>         meaning.
> ---
>
> And all examples there assume one "pin name" or "ID" per pin.
>
> Now, we use 2 values per each pin (the pin ID and the alternate
> function number), so to me this is different from what the generic
> binding describes.
> Also, pinctrl-single, and pinctrl-imx which have and ABI similar to
> the one this driver define, use "pinctrl-single,pins" and "fsl,pins"
> respectively as property names.
> So either they have to be updated yet, or we should keep using
> "renesas,pins" for our own defined ABI.
>
> Maybe Linus or other pinctrl people can give some suggestion here.

To me as subsystem maintainer any "necessarily different" bindings
are just a big confusion for the head.

Since you're adding a new driver, try to stick to the generic bindings
even if it deviates from what you are used to for Renesas, because
even if it may be more work for you guys or make you annoyed that
now a certain Renesas is different from all other Renesas platforms,
for the community this makes things easier to maintain because
we can look at the driver and its bindings and say "ah I know this".

The fact that historically all the early adopters of pinctrl in device tree
have these funky custom bindings is unfortunate but just something
that we need to live with.

Yours,
Linus Walleij

[toc] | [next] | [standalone]

#1611033

Hi Linus,

On 2017-03-28 11:46, Linus Walleij wrote:
> On Thu, Mar 23, 2017 at 5:02 PM, jacopo <jacopo@jmondi.org> wrote:
> 
>>> > +  Required properties:
>>> > +    - renesas,pins
>>> 
>>> Just "pins"?
>>> 
>> 
>> You know, I've been thinking about this, bu the "pins" property
>> definition in pinctrl-bidings is the following one:
>> 
>> Documentation/devicetree/bindings/pinctrl/pinctrl-bindings.txt
>> ---
>> - pins takes a list of pin names or IDs as a required argument. The
>>   specific binding for the hardware defines:
>>       - Whether the entries are integers or strings, and their
>>         meaning.
>> ---
>> 
>> And all examples there assume one "pin name" or "ID" per pin.
>> 
>> Now, we use 2 values per each pin (the pin ID and the alternate
>> function number), so to me this is different from what the generic
>> binding describes.
>> Also, pinctrl-single, and pinctrl-imx which have and ABI similar to
>> the one this driver define, use "pinctrl-single,pins" and "fsl,pins"
>> respectively as property names.
>> So either they have to be updated yet, or we should keep using
>> "renesas,pins" for our own defined ABI.
>> 
>> Maybe Linus or other pinctrl people can give some suggestion here.
> 
> To me as subsystem maintainer any "necessarily different" bindings
> are just a big confusion for the head.
> 

Understandable :)

> Since you're adding a new driver, try to stick to the generic bindings
> even if it deviates from what you are used to for Renesas, because
> even if it may be more work for you guys or make you annoyed that
> now a certain Renesas is different from all other Renesas platforms,
> for the community this makes things easier to maintain because
> we can look at the driver and its bindings and say "ah I know this".
> 
> The fact that historically all the early adopters of pinctrl in device 
> tree
> have these funky custom bindings is unfortunate but just something
> that we need to live with.
> 

To avoid any confusion, please bear with me and clarify this once and 
for all,
since I'm not certain I fully got you here.

Are you suggesting:

1) Use "pins" property with the currently implemented ABI (which 
slightly differs
    from the standard documented one as explained above. Not sure it is 
fine overriding
    it or not)

2) Use "pins" property and change our ABI to match the documented one: 
one ID (integer
    or string) per pin, not 2 as we're doing now.

Both solutions are easily implementable; 2) requires some more work to 
make pin id, function number
and pin configuration fit in one single integer, but is achievable for 
sure.

Thanks
    j

> Yours,
> Linus Walleij

[toc] | [prev] | [next] | [standalone]

#1611520

On Tue, Mar 28, 2017 at 4:38 PM,  <jacopo@jmondi.org> wrote:

>> The fact that historically all the early adopters of pinctrl in device
>> tree
>> have these funky custom bindings is unfortunate but just something
>> that we need to live with.
>>
>
> To avoid any confusion, please bear with me and clarify this once and for
> all,
> since I'm not certain I fully got you here.
>
> Are you suggesting:
>
> 1) Use "pins" property with the currently implemented ABI (which slightly
> differs
>    from the standard documented one as explained above. Not sure it is fine
> overriding
>    it or not)

Correction: you should be using the property "pinmux", because you
are setting group and function at the same time.

See for example:
include/dt-bindings/pinctrl/mt65xx.h

And how that is used in:
arch/arm/boot/dts/mt2701-pinfunc.h
arch/arm/boot/dts/mt2701-evb.dts

The docs are here:
Documentation/devicetree/bindings/pinctrl/pinctrl-mt65xx.txt

I'm sorry that "pinmux" is not part of the generic documentation, it'd be
great if you would like to add it with a patch.

Yours,
Linus Walleij

[toc] | [prev] | [next] | [standalone]

#1611649

Hi Linus,

On Wed, Mar 29, 2017 at 4:30 AM, Linus Walleij <linus.walleij@linaro.org> wrote:
> On Tue, Mar 28, 2017 at 4:38 PM,  <jacopo@jmondi.org> wrote:
>>> The fact that historically all the early adopters of pinctrl in device
>>> tree
>>> have these funky custom bindings is unfortunate but just something
>>> that we need to live with.
>>
>> To avoid any confusion, please bear with me and clarify this once and for
>> all,
>> since I'm not certain I fully got you here.
>>
>> Are you suggesting:
>>
>> 1) Use "pins" property with the currently implemented ABI (which slightly
>> differs
>>    from the standard documented one as explained above. Not sure it is fine
>> overriding
>>    it or not)
>
> Correction: you should be using the property "pinmux", because you
> are setting group and function at the same time.

OK.

> See for example:
> include/dt-bindings/pinctrl/mt65xx.h
>
> And how that is used in:
> arch/arm/boot/dts/mt2701-pinfunc.h
> arch/arm/boot/dts/mt2701-evb.dts
>
> The docs are here:
> Documentation/devicetree/bindings/pinctrl/pinctrl-mt65xx.txt

All of the above pack the information for a pin into a single 32-bit integer.
Is that what you want us to use, too?
Currently we use two integers: 1) pin index, and 2) function/flags combo.

> I'm sorry that "pinmux" is not part of the generic documentation, it'd be
> great if you would like to add it with a patch.

That would mean pinmux could be an array of single values, or tuples.
Is that OK?

Thanks!

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

[toc] | [prev] | [next] | [standalone]

#1611784

On Wed, Mar 29, 2017 at 9:35 AM, Geert Uytterhoeven
<geert@linux-m68k.org> wrote:
> Me:

>> See for example:
>> include/dt-bindings/pinctrl/mt65xx.h
>>
>> And how that is used in:
>> arch/arm/boot/dts/mt2701-pinfunc.h
>> arch/arm/boot/dts/mt2701-evb.dts
>>
>> The docs are here:
>> Documentation/devicetree/bindings/pinctrl/pinctrl-mt65xx.txt
>
> All of the above pack the information for a pin into a single 32-bit integer.
> Is that what you want us to use, too?
> Currently we use two integers: 1) pin index, and 2) function/flags combo.

I don't really know what you need, sorry. But some kind of figure, yes.
I would say whatever makes sense. 16+16 bits makes sense in most
combinatorial spaces does it not? If you split 32 bits  in 16 bits for
pin and 16 bits for function, do you have more than 2^16 pins or 2^16
functions?

If you really do we may need to go for u64 but ... really? Is there
a rational reason for that other than "we did it like this first"?

I do not understand the notion of "flags" here. I hope that is not referring
to pin config, because I expect that to use the standard pin config
bindings outside of the pinmux value which should just define the
pin+function combo:

node {
    pinmux = <PIN_NUMBER_PINMUX>;
    GENERIC_PINCONFIG;
};

Example from Mediatek:

i2c1_pins_a: i2c1@0 {
    pins {
        pinmux = <MT8135_PIN_195_SDA1__FUNC_SDA1>,
                        <MT8135_PIN_196_SCL1__FUNC_SCL1>;
        bias-pull-up = <55>;
    };
};

So this allows for a combine pin+function number but pull ups,
bias etc are not baked into the thing, they have to be added on
separately with the generic bindings, which is nice and very readable.

Yours,
Linus Walleij

[toc] | [prev] | [next] | [standalone]

#1611845

Hi Linus,

On Wed, Mar 29, 2017 at 12:15 PM, Linus Walleij
<linus.walleij@linaro.org> wrote:
> On Wed, Mar 29, 2017 at 9:35 AM, Geert Uytterhoeven
> <geert@linux-m68k.org> wrote:
>>> See for example:
>>> include/dt-bindings/pinctrl/mt65xx.h
>>>
>>> And how that is used in:
>>> arch/arm/boot/dts/mt2701-pinfunc.h
>>> arch/arm/boot/dts/mt2701-evb.dts
>>>
>>> The docs are here:
>>> Documentation/devicetree/bindings/pinctrl/pinctrl-mt65xx.txt
>>
>> All of the above pack the information for a pin into a single 32-bit integer.
>> Is that what you want us to use, too?
>> Currently we use two integers: 1) pin index, and 2) function/flags combo.
>
> I don't really know what you need, sorry. But some kind of figure, yes.
> I would say whatever makes sense. 16+16 bits makes sense in most
> combinatorial spaces does it not? If you split 32 bits  in 16 bits for
> pin and 16 bits for function, do you have more than 2^16 pins or 2^16
> functions?
>
> If you really do we may need to go for u64 but ... really? Is there
> a rational reason for that other than "we did it like this first"?
>
> I do not understand the notion of "flags" here. I hope that is not referring

Flags refers to BI_DIR, SWIO_IN, and SWIO_OUT, from
https://patchwork.kernel.org/patch/9643047/

32-bit should be enough to cover pins, function, and flags.

> to pin config, because I expect that to use the standard pin config
> bindings outside of the pinmux value which should just define the
> pin+function combo:
>
> node {
>     pinmux = <PIN_NUMBER_PINMUX>;
>     GENERIC_PINCONFIG;
> };
>
> Example from Mediatek:
>
> i2c1_pins_a: i2c1@0 {
>     pins {
>         pinmux = <MT8135_PIN_195_SDA1__FUNC_SDA1>,
>                         <MT8135_PIN_196_SCL1__FUNC_SCL1>;

If we follow this example, then we can list all combinations in
include/dt-bindings/pinctrl/r7s72100-pinctrl.h, instead of creating the value
by combining the bits using a macro where we need it in the DTS.

It's gonna be a long list, though...

>         bias-pull-up = <55>;
>     };
> };
>
> So this allows for a combine pin+function number but pull ups,
> bias etc are not baked into the thing, they have to be added on
> separately with the generic bindings, which is nice and very readable.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

[toc] | [prev] | [next] | [standalone]

#1611876

Hi Geert, Linus

On Wed, Mar 29, 2017 at 01:20:39PM +0200, Geert Uytterhoeven wrote:
> Hi Linus,
> 
> On Wed, Mar 29, 2017 at 12:15 PM, Linus Walleij
> <linus.walleij@linaro.org> wrote:
> > On Wed, Mar 29, 2017 at 9:35 AM, Geert Uytterhoeven
> > <geert@linux-m68k.org> wrote:
> >>> See for example:
> >>> include/dt-bindings/pinctrl/mt65xx.h
> >>>
> >>> And how that is used in:
> >>> arch/arm/boot/dts/mt2701-pinfunc.h
> >>> arch/arm/boot/dts/mt2701-evb.dts
> >>>
> >>> The docs are here:
> >>> Documentation/devicetree/bindings/pinctrl/pinctrl-mt65xx.txt
> >>
> >> All of the above pack the information for a pin into a single 32-bit integer.
> >> Is that what you want us to use, too?
> >> Currently we use two integers: 1) pin index, and 2) function/flags combo.
> >
> > I don't really know what you need, sorry. But some kind of figure, yes.
> > I would say whatever makes sense. 16+16 bits makes sense in most
> > combinatorial spaces does it not? If you split 32 bits  in 16 bits for
> > pin and 16 bits for function, do you have more than 2^16 pins or 2^16
> > functions?
> >
> > If you really do we may need to go for u64 but ... really? Is there
> > a rational reason for that other than "we did it like this first"?
> >
> > I do not understand the notion of "flags" here. I hope that is not referring
> 
> Flags refers to BI_DIR, SWIO_IN, and SWIO_OUT, from
> https://patchwork.kernel.org/patch/9643047/
> 
> 32-bit should be enough to cover pins, function, and flags.
> 

Geert already replied, but to avoid any confusion I'll try to remove
from driver the use of "pin config" when referring to this three
flags, which are just additional informations the pin controller needs
to perform pin muxing properly. They're not related the standard pin
config properties (pull-up/down, bias etc.. actually our hardware does
not even support these natively)

> > to pin config, because I expect that to use the standard pin config
> > bindings outside of the pinmux value which should just define the
> > pin+function combo:
> >
> > node {
> >     pinmux = <PIN_NUMBER_PINMUX>;
> >     GENERIC_PINCONFIG;
> > };
> >
> > Example from Mediatek:
> >
> > i2c1_pins_a: i2c1@0 {
> >     pins {
> >         pinmux = <MT8135_PIN_195_SDA1__FUNC_SDA1>,
> >                         <MT8135_PIN_196_SCL1__FUNC_SCL1>;
> 
> If we follow this example, then we can list all combinations in
> include/dt-bindings/pinctrl/r7s72100-pinctrl.h, instead of creating the value
> by combining the bits using a macro where we need it in the DTS.
> 
> It's gonna be a long list, though...
> 

I'm strongly in favour of something like
    pinmux = <PIN(1, 4) | FUNC# | FLAGS>, .... ;

opposed to
    pinmux = <PIN1_4_FUNC#_FLAGS>, ... ;

Not only because it will save use from having a loong list(*) of macros that has
to be kept up to date when/if new RZ hardware will arrive, but also
because of readability and simplicity for down-stream and BSP users.
Speaking of which, I would like to know what does Chris think of this.

Is there any strong preference between the two for you Linus?

Thanks
   j

(*) 12 ports, 16 pins, 8 functions, 3 flags: you can do the math
yourselves and see this is going to be hard to maintain.

> >         bias-pull-up = <55>;
> >     };
> > };
> >
> > So this allows for a combine pin+function number but pull ups,
> > bias etc are not baked into the thing, they have to be added on
> > separately with the generic bindings, which is nice and very readable.
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
> 
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                 -- Linus Torvalds

[toc] | [prev] | [next] | [standalone]

#1611896

Adding my 2 cents here:

On Wednesday, March 29, 2017, jacopo wrote:
> > > If you really do we may need to go for u64 but ... really? Is there
> > > a rational reason for that other than "we did it like this first"?
> > >
> > > I do not understand the notion of "flags" here. I hope that is not
> > > referring
> >
> > Flags refers to BI_DIR, SWIO_IN, and SWIO_OUT, from
> > https://patchwork.kernel.org/patch/9643047/
> >
> > 32-bit should be enough to cover pins, function, and flags.
> >
> 
> Geert already replied, but to avoid any confusion I'll try to remove from
> driver the use of "pin config" when referring to this three flags, which
> are just additional informations the pin controller needs to perform pin
> muxing properly. They're not related the standard pin config properties
> (pull-up/down, bias etc.. actually our hardware does not even support
> these natively)
> 
> > > to pin config, because I expect that to use the standard pin config
> > > bindings outside of the pinmux value which should just define the
> > > pin+function combo:
> > >
> > > node {
> > >     pinmux = <PIN_NUMBER_PINMUX>;
> > >     GENERIC_PINCONFIG;
> > > };
> > >
> > > Example from Mediatek:
> > >
> > > i2c1_pins_a: i2c1@0 {
> > >     pins {
> > >         pinmux = <MT8135_PIN_195_SDA1__FUNC_SDA1>,
> > >                         <MT8135_PIN_196_SCL1__FUNC_SCL1>;
> >
> > If we follow this example, then we can list all combinations in
> > include/dt-bindings/pinctrl/r7s72100-pinctrl.h, instead of creating
> > the value by combining the bits using a macro where we need it in the
> DTS.
> >
> > It's gonna be a long list, though...
> >
> 
> I'm strongly in favour of something like
>     pinmux = <PIN(1, 4) | FUNC# | FLAGS>, .... ;
> 
> opposed to
>     pinmux = <PIN1_4_FUNC#_FLAGS>, ... ;


I agree. I like "<PIN(1, 4) | FUNC# | FLAGS>".



> Not only because it will save use from having a loong list(*) of macros
> that has to be kept up to date when/if new RZ hardware will arrive, but
> also because of readability and simplicity for down-stream and BSP users.
> Speaking of which, I would like to know what does Chris think of this.

The list of macros would be very long, especially against the different
packaging version of the RZ/A1 series. 11 ports, 16-pins for each port, 8 different
function options for each pin....2 different package/pin variations.

And at the end of the day....there is no benefit for the user over just using
a macro.


A little about "this controller" for the RZ/A1: In my opinion it's a one-shot usage.
I don't foresee future RZ/A SoCs using this exact pin controller.
The reason for the "FLAGS" is to work around a quirky hardware design (in my opinion).

However, future RZ/A SoCs will use a 'similar' type controller where each pin has 8 function
options (but the FLAGs won't be needed anymore...as far as I can see...).

So I foresee the DT interface staying more or less the same, but the underlying register
setup in the driver Will change (become more simple).


Now...if we can only convince the R-Car guys to move to a more simple pin-mux controller...


Chris

[toc] | [prev] | [next] | [standalone]

#1611925

On Wed, Mar 29, 2017 at 2:38 PM, Chris Brandt <Chris.Brandt@renesas.com> wrote:

>> I'm strongly in favour of something like
>>     pinmux = <PIN(1, 4) | FUNC# | FLAGS>, .... ;
>>
>> opposed to
>>     pinmux = <PIN1_4_FUNC#_FLAGS>, ... ;
>
>
> I agree. I like "<PIN(1, 4) | FUNC# | FLAGS>".

I don't know if you have seen the Mediatek things really.
include/dt-bindings/pinctrl/mt6397-pinfunc.h

Example:
#define MT6397_PIN_2_SRCLKEN_PERI__FUNC_GPIO2 (MTK_PIN_NO(2) | 0)

If you prefer to use preprocessor macros or whatever to make the
bitmasks or how you want to organize the constants in your
include files is not my concern, do whatever you seem fit, just
pack it into a 32bit thing somehow which makes sense from a
maintenance point of view.

>> Not only because it will save use from having a loong list(*) of macros
>> that has to be kept up to date when/if new RZ hardware will arrive, but
>> also because of readability and simplicity for down-stream and BSP users.
>> Speaking of which, I would like to know what does Chris think of this.
>
> The list of macros would be very long, especially against the different
> packaging version of the RZ/A1 series. 11 ports, 16-pins for each port, 8 different
> function options for each pin....2 different package/pin variations.
>
> And at the end of the day....there is no benefit for the user over just using
> a macro.

I don't know who has this idea that you could not use macros, certainly
not me. Some misunderstanding must be going on. For what I'm concerned
you can write hex numbers in the pinmux = <0x12345678>;

> The reason for the "FLAGS" is to work around a quirky hardware design (in my opinion).

The flags I don't like at all, and think they should be converted to generic
pin config because they have nothing to do with muxing.

But I will point that out in the specific patch adding them.

Yours,
Linus Walleij

[toc] | [prev] | [next] | [standalone]

#1611975

On Wednesday, March 29, 2017, Linus Walleij:
> If you prefer to use preprocessor macros or whatever to make the bitmasks
> or how you want to organize the constants in your include files is not my
> concern, do whatever you seem fit, just pack it into a 32bit thing somehow
> which makes sense from a maintenance point of view.

OK, I think everyone agrees that a single 32-bit value is fine because macros
will also for good readability and maintenance.


> >> Not only because it will save use from having a loong list(*) of
> >> macros that has to be kept up to date when/if new RZ hardware will
> >> arrive, but also because of readability and simplicity for down-stream
> and BSP users.
> >> Speaking of which, I would like to know what does Chris think of this.
> >
> > The list of macros would be very long, especially against the
> > different packaging version of the RZ/A1 series. 11 ports, 16-pins for
> > each port, 8 different function options for each pin....2 different
> package/pin variations.
> >
> > And at the end of the day....there is no benefit for the user over
> > just using a macro.
> 
> I don't know who has this idea that you could not use macros, certainly
> not me. Some misunderstanding must be going on. For what I'm concerned you
> can write hex numbers in the pinmux = <0x12345678>;
> 
> > The reason for the "FLAGS" is to work around a quirky hardware design
> (in my opinion).
> 
> The flags I don't like at all, and think they should be converted to
> generic pin config because they have nothing to do with muxing.
> 
> But I will point that out in the specific patch adding them.

OK, I think I understand your issue a little better of mixing user-defined config
with generic pinmux.

From our perspective, the FLAGS (BI_DIR, SWIO_IN, and SWIO_OUT ) were not really
optional when selecting what you want the pin to do....so we considered it part
of the pin-mux.

In the hardware manual, there are tables that say that for 'some' certain
pins/functions "just setting the pin to a function is not enough...you need to make
another register setting (that may or may not make sense), otherwise it's not going
to work".

So, trying to get the driver to be that smart across all the different pin/package
variations seems to be way too ugly (and a maintenance nightmare). Simply putting
that in the DT binding was much more cleaner.


As for how to pass this HW info into the driver, I'll move over to the other email
thread where you started to give some suggestions...


Chris

[toc] | [prev] | [next] | [standalone]

#1611915

On Wed, Mar 29, 2017 at 1:20 PM, Geert Uytterhoeven
<geert@linux-m68k.org> wrote:

>> I do not understand the notion of "flags" here. I hope that is not referring
>
> Flags refers to BI_DIR, SWIO_IN, and SWIO_OUT, from
> https://patchwork.kernel.org/patch/9643047/

Aha I will go in and review that closer because it doesn't seem right.

Sorry for missing it.

>> i2c1_pins_a: i2c1@0 {
>>     pins {
>>         pinmux = <MT8135_PIN_195_SDA1__FUNC_SDA1>,
>>                         <MT8135_PIN_196_SCL1__FUNC_SCL1>;
>
> If we follow this example, then we can list all combinations in
> include/dt-bindings/pinctrl/r7s72100-pinctrl.h, instead of creating the value
> by combining the bits using a macro where we need it in the DTS.
>
> It's gonna be a long list, though...

Size is not the issue, readability is the issue. I don't see why you would
need to list "all combinations" since the trees go through the C preprocessor
so you can use macros and bit | OR to build them?

Yours,
Linus Walleij

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web