[PATCH 2/2] i2c-mux-gpio: eliminate i2c channel order assumptions

[PATCH 2/2] i2c-mux-gpio: eliminate i2c channel order assumptions

[Ionut Nicu]

The i2c-mux driver uses the chan_id parameter provided
in i2c_add_mux_adapter as a parameter to the select
and deselect callbacks while the i2c-mux-gpio driver
uses the chan_id as an index in the mux->data.values
array.

A simple example of where this doesn't work is when we
have a device tree like this:

i2cmux {
	i2c@1 {
		reg = <1>;
		...
	};

	i2c@0 {
		reg = <0>;
		...
	};
};

The mux->data.values array will be { 1, 0 }, but when
the i2-mux driver will try to select channel 0, the
i2c-mux-gpio driver will actually use values[0], hence 1
as the gpio selection value.

Signed-off-by: Ionut Nicu <ioan.nicu.ext@nsn.com>
---
 drivers/i2c/muxes/i2c-mux-gpio.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/i2c/muxes/i2c-mux-gpio.c b/drivers/i2c/muxes/i2c-mux-gpio.c
index b5f17ef..3505d0e 100644
--- a/drivers/i2c/muxes/i2c-mux-gpio.c
+++ b/drivers/i2c/muxes/i2c-mux-gpio.c
@@ -43,7 +43,7 @@ static int i2c_mux_gpio_select(struct i2c_adapter *adap, void *data, u32 chan)
 {
 	struct gpiomux *mux = data;
 
-	i2c_mux_gpio_set(mux, mux->data.values[chan]);
+	i2c_mux_gpio_set(mux, chan);
 
 	return 0;
 }
@@ -233,7 +233,7 @@ static int i2c_mux_gpio_probe(struct platform_device *pdev)
 		unsigned int class = mux->data.classes ? mux->data.classes[i] : 0;
 
 		mux->adap[i] = i2c_add_mux_adapter(parent, &pdev->dev, mux, nr,
-						   i, class,
+						   mux->data.values[i], class,
 						   i2c_mux_gpio_select, deselect);
 		if (!mux->adap[i]) {
 			ret = -ENODEV;
-- 
1.7.1

Re: [PATCH 2/2] i2c-mux-gpio: eliminate i2c channel order assumptions

[Alexander Sverdlin]

Hi!

On 10/10/2013 10:39 AM, Ionut Nicu wrote:
> The i2c-mux driver uses the chan_id parameter provided
> in i2c_add_mux_adapter as a parameter to the select
> and deselect callbacks while the i2c-mux-gpio driver
> uses the chan_id as an index in the mux->data.values
> array.
> 
> A simple example of where this doesn't work is when we
> have a device tree like this:
> 
> i2cmux {
> 	i2c@1 {
> 		reg = <1>;
> 		...
> 	};
> 
> 	i2c@0 {
> 		reg = <0>;
> 		...
> 	};
> };
> 
> The mux->data.values array will be { 1, 0 }, but when
> the i2-mux driver will try to select channel 0, the
> i2c-mux-gpio driver will actually use values[0], hence 1
> as the gpio selection value.

The patch itself is correct, but the description is not precise,
I suppose... i2c-mux-gpio is consistent inside itself, it will
receive for every child adapter the value it has configured.
The problem happens inside i2c-mux.c, i2c_add_mux_adapter():

               	for_each_child_of_node(mux_dev->of_node, child) {
                       	ret = of_property_read_u32(child, "reg", &reg);
                       	if (ret)
                               	continue;
                       	if (chan_id == reg) {
                               	priv->adap.dev.of_node = child;

Which means, i2c-mux-gpio MUST pass reg, not its logical index inside
array. Otherwise node will not be correctly assigned and i2c-mux will
have problems selecting right adapter for the multiplexed devices.

> Signed-off-by: Ionut Nicu <ioan.nicu.ext-OYasijW0DpE@public.gmane.org>

So, for the code itself

Acked-by: Alexander Sverdlin <alexander.sverdlin-OYasijW0DpE@public.gmane.org>

> ---
>  drivers/i2c/muxes/i2c-mux-gpio.c |    4 ++--
>  1 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/i2c/muxes/i2c-mux-gpio.c b/drivers/i2c/muxes/i2c-mux-gpio.c
> index b5f17ef..3505d0e 100644
> --- a/drivers/i2c/muxes/i2c-mux-gpio.c
> +++ b/drivers/i2c/muxes/i2c-mux-gpio.c
> @@ -43,7 +43,7 @@ static int i2c_mux_gpio_select(struct i2c_adapter *adap, void *data, u32 chan)
>  {
>  	struct gpiomux *mux = data;
>  
> -	i2c_mux_gpio_set(mux, mux->data.values[chan]);
> +	i2c_mux_gpio_set(mux, chan);
>  
>  	return 0;
>  }
> @@ -233,7 +233,7 @@ static int i2c_mux_gpio_probe(struct platform_device *pdev)
>  		unsigned int class = mux->data.classes ? mux->data.classes[i] : 0;
>  
>  		mux->adap[i] = i2c_add_mux_adapter(parent, &pdev->dev, mux, nr,
> -						   i, class,
> +						   mux->data.values[i], class,
>  						   i2c_mux_gpio_select, deselect);
>  		if (!mux->adap[i]) {
>  			ret = -ENODEV;
> 

-- 
Best regards,
Alexander Sverdlin.

Re: [PATCH 2/2] i2c-mux-gpio: eliminate i2c channel order assumptions

[Ionut Nicu]

Hi,

On 10.10.2013 12:34, Alexander Sverdlin wrote:
> Hi!
> 
> On 10/10/2013 10:39 AM, Ionut Nicu wrote:
>> The i2c-mux driver uses the chan_id parameter provided
>> in i2c_add_mux_adapter as a parameter to the select
>> and deselect callbacks while the i2c-mux-gpio driver
>> uses the chan_id as an index in the mux->data.values
>> array.
>>
>> A simple example of where this doesn't work is when we
>> have a device tree like this:
>>
>> i2cmux {
>> 	i2c@1 {
>> 		reg = <1>;
>> 		...
>> 	};
>>
>> 	i2c@0 {
>> 		reg = <0>;
>> 		...
>> 	};
>> };
>>
>> The mux->data.values array will be { 1, 0 }, but when
>> the i2-mux driver will try to select channel 0, the
>> i2c-mux-gpio driver will actually use values[0], hence 1
>> as the gpio selection value.
> 
> The patch itself is correct, but the description is not precise,
> I suppose... i2c-mux-gpio is consistent inside itself, it will
> receive for every child adapter the value it has configured.
> The problem happens inside i2c-mux.c, i2c_add_mux_adapter():
> 
>                	for_each_child_of_node(mux_dev->of_node, child) {
>                        	ret = of_property_read_u32(child, "reg", &reg);
>                        	if (ret)
>                                	continue;
>                        	if (chan_id == reg) {
>                                	priv->adap.dev.of_node = child;
> 
> Which means, i2c-mux-gpio MUST pass reg, not its logical index inside
> array. Otherwise node will not be correctly assigned and i2c-mux will
> have problems selecting right adapter for the multiplexed devices.
> 
>> Signed-off-by: Ionut Nicu <ioan.nicu.ext-OYasijW0DpE@public.gmane.org>
> 
> So, for the code itself
> 
> Acked-by: Alexander Sverdlin <alexander.sverdlin-OYasijW0DpE@public.gmane.org>
> 

You are right, the patch description is not so good. I will try to change
it so it's clearer for everyone what I'm trying to fix here and after that
I will re-submit the series.

>> ---
>>  drivers/i2c/muxes/i2c-mux-gpio.c |    4 ++--
>>  1 files changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/i2c/muxes/i2c-mux-gpio.c b/drivers/i2c/muxes/i2c-mux-gpio.c
>> index b5f17ef..3505d0e 100644
>> --- a/drivers/i2c/muxes/i2c-mux-gpio.c
>> +++ b/drivers/i2c/muxes/i2c-mux-gpio.c
>> @@ -43,7 +43,7 @@ static int i2c_mux_gpio_select(struct i2c_adapter *adap, void *data, u32 chan)
>>  {
>>  	struct gpiomux *mux = data;
>>  
>> -	i2c_mux_gpio_set(mux, mux->data.values[chan]);
>> +	i2c_mux_gpio_set(mux, chan);
>>  
>>  	return 0;
>>  }
>> @@ -233,7 +233,7 @@ static int i2c_mux_gpio_probe(struct platform_device *pdev)
>>  		unsigned int class = mux->data.classes ? mux->data.classes[i] : 0;
>>  
>>  		mux->adap[i] = i2c_add_mux_adapter(parent, &pdev->dev, mux, nr,
>> -						   i, class,
>> +						   mux->data.values[i], class,
>>  						   i2c_mux_gpio_select, deselect);
>>  		if (!mux->adap[i]) {
>>  			ret = -ENODEV;
>>
>