[PATCH] i2c-bfin-twi: Register adapter with a specific bus number

[PATCH] i2c-bfin-twi: Register adapter with a specific bus number

[Kalle Pokki]

All the users of this driver explicitly specify the I2C bus numbers
to be used in their platform data. Make the driver respect that.

Signed-off-by: Kalle Pokki <kalle.pokki-loK5ZF7nR+A@public.gmane.org>
---
 drivers/i2c/busses/i2c-bfin-twi.c |    3 ++-
 1 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/i2c/busses/i2c-bfin-twi.c b/drivers/i2c/busses/i2c-bfin-twi.c
index 67224a4..7dbdaeb 100644
--- a/drivers/i2c/busses/i2c-bfin-twi.c
+++ b/drivers/i2c/busses/i2c-bfin-twi.c
@@ -550,6 +550,7 @@ static int i2c_bfin_twi_probe(struct platform_device *dev)
 
 	p_adap = &iface->adap;
 	p_adap->id = I2C_HW_BLACKFIN;
+	p_adap->nr = dev->id;
 	strlcpy(p_adap->name, dev->name, sizeof(p_adap->name));
 	p_adap->algo = &bfin_twi_algorithm;
 	p_adap->algo_data = iface;
@@ -576,7 +577,7 @@ static int i2c_bfin_twi_probe(struct platform_device *dev)
 	bfin_write_TWI_CONTROL(bfin_read_TWI_CONTROL() | TWI_ENA);
 	SSYNC();
 
-	rc = i2c_add_adapter(p_adap);
+	rc = i2c_add_numbered_adapter(p_adap);
 	if (rc < 0)
 		free_irq(iface->irq, iface);
 	else
-- 
1.4.4.2


_______________________________________________
i2c mailing list
i2c-GZX6beZjE8VD60Wz+7aTrA@public.gmane.org
http://lists.lm-sensors.org/mailman/listinfo/i2c

Re: [PATCH] i2c-bfin-twi: Register adapter with a specific bus number

[Bryan Wu]

On Tue, 2008-01-22 at 23:40 +0200, Kalle Pokki wrote:
> All the users of this driver explicitly specify the I2C bus numbers
> to be used in their platform data. Make the driver respect that.
> 
> Signed-off-by: Kalle Pokki <kalle.pokki-loK5ZF7nR+A@public.gmane.org>
> ---
>  drivers/i2c/busses/i2c-bfin-twi.c |    3 ++-
>  1 files changed, 2 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-bfin-twi.c b/drivers/i2c/busses/i2c-bfin-twi.c
> index 67224a4..7dbdaeb 100644
> --- a/drivers/i2c/busses/i2c-bfin-twi.c
> +++ b/drivers/i2c/busses/i2c-bfin-twi.c
> @@ -550,6 +550,7 @@ static int i2c_bfin_twi_probe(struct platform_device *dev)
>  
>  	p_adap = &iface->adap;
>  	p_adap->id = I2C_HW_BLACKFIN;
> +	p_adap->nr = dev->id;

How do you think of following code, Kalle?

	 /*
         * If "dev->id" is negative we consider it as zero.
         * The reason to do so is to avoid sysfs names that only make
         * sense when there are multiple adapters.
         */
        p_adap->nr = dev->id >= 0 ? dev->id : 0;

I found some i2c-gpio, i2c-ibm_iic.c and i2c-pxa.c using this code,
while others use your idea.

>  	strlcpy(p_adap->name, dev->name, sizeof(p_adap->name));
>  	p_adap->algo = &bfin_twi_algorithm;
>  	p_adap->algo_data = iface;
> @@ -576,7 +577,7 @@ static int i2c_bfin_twi_probe(struct platform_device *dev)
>  	bfin_write_TWI_CONTROL(bfin_read_TWI_CONTROL() | TWI_ENA);
>  	SSYNC();
>  
> -	rc = i2c_add_adapter(p_adap);
> +	rc = i2c_add_numbered_adapter(p_adap);
>  	if (rc < 0)
>  		free_irq(iface->irq, iface);
>  	else

Thanks a lot
-Bryan Wu

_______________________________________________
i2c mailing list
i2c-GZX6beZjE8VD60Wz+7aTrA@public.gmane.org
http://lists.lm-sensors.org/mailman/listinfo/i2c

Re: [PATCH] i2c-bfin-twi: Register adapter with a specific bus number

[Kalle Pokki]

On Wed, 23 Jan 2008, Bryan Wu wrote:

> How do you think of following code, Kalle?
> 
> 	 /*
>          * If "dev->id" is negative we consider it as zero.
>          * The reason to do so is to avoid sysfs names that only make
>          * sense when there are multiple adapters.
>          */
>         p_adap->nr = dev->id >= 0 ? dev->id : 0;
> 
> I found some i2c-gpio, i2c-ibm_iic.c and i2c-pxa.c using this code,
> while others use your idea.

I don't see the point in ever defining dev->id as -1. None of the current 
users of i2c-bfin-twi does that. Another option could be to use dynamic 
bus numbers with dev->id == -1, but I think the only sane way to enumerate 
this kind of SOC busses is to fix the numbers in the machine 
initialization time.

_______________________________________________
i2c mailing list
i2c-GZX6beZjE8VD60Wz+7aTrA@public.gmane.org
http://lists.lm-sensors.org/mailman/listinfo/i2c

Re: [PATCH] i2c-bfin-twi: Register adapter with a specific bus number

[Jean Delvare]

Hi Kalle,

On Wed, 23 Jan 2008 10:54:45 +0200 (EET), Kalle Pokki wrote:
> On Wed, 23 Jan 2008, Bryan Wu wrote:
> 
> > How do you think of following code, Kalle?
> > 
> > 	 /*
> >          * If "dev->id" is negative we consider it as zero.
> >          * The reason to do so is to avoid sysfs names that only make
> >          * sense when there are multiple adapters.
> >          */
> >         p_adap->nr = dev->id >= 0 ? dev->id : 0;
> > 
> > I found some i2c-gpio, i2c-ibm_iic.c and i2c-pxa.c using this code,
> > while others use your idea.
> 
> I don't see the point in ever defining dev->id as -1.

The point is spelled out just above:

"The reason to do so is to avoid sysfs names that only make sense when
there are multiple adapters."

I.e., some people prefer to have their platform device named just foo
instead of foo.0. Me, I wouldn't have bothered making an exception for
this, but the fact is that the platform core code supports it.

>                                                         None of the current 
> users of i2c-bfin-twi does that. Another option could be to use dynamic 
> bus numbers with dev->id == -1, but I think the only sane way to enumerate 
> this kind of SOC busses is to fix the numbers in the machine 
> initialization time.

I don't really care myself. I'll apply your patch as is, but if you
want to update it, that's equally fine.

-- 
Jean Delvare

_______________________________________________
i2c mailing list
i2c-GZX6beZjE8VD60Wz+7aTrA@public.gmane.org
http://lists.lm-sensors.org/mailman/listinfo/i2c

Re: [PATCH] i2c-bfin-twi: Register adapter with a specific bus number

[Kalle Pokki]

On Wed, 23 Jan 2008, Jean Delvare wrote:

> > I don't see the point in ever defining dev->id as -1.
> 
> The point is spelled out just above:
> 
> "The reason to do so is to avoid sysfs names that only make sense when
> there are multiple adapters."
> 
> I.e., some people prefer to have their platform device named just foo
> instead of foo.0. Me, I wouldn't have bothered making an exception for
> this, but the fact is that the platform core code supports it.

OK, now I get it. I didn't even consider that the platform core handled 
this as a special case, too.

Each in-tree user of the driver already sets dev->id = 0, even if there 
is only one device, so adding the extra check wouldn't make much 
difference now.

_______________________________________________
i2c mailing list
i2c-GZX6beZjE8VD60Wz+7aTrA@public.gmane.org
http://lists.lm-sensors.org/mailman/listinfo/i2c