Re: [PATCH 1/5] i2c: Convert i2c-octeon.c to use device tree.

[David Daney <david.daney-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org>]

On 03/26/2012 06:20 PM, Rob Herring wrote:
[...]
>> ---
>>
>> Should probably go via Ralf's linux-mips.org tree.
>>
>>   arch/mips/cavium-octeon/octeon-irq.c      |    2 -
>>   arch/mips/cavium-octeon/octeon-platform.c |   84 -------------------------
>>   arch/mips/include/asm/octeon/octeon.h     |    5 --
>>   drivers/i2c/busses/i2c-octeon.c           |   94 ++++++++++++++++-------------
>>   4 files changed, 52 insertions(+), 133 deletions(-)
>
> snip
>
>>
>> -	if (i2c_data == NULL) {
>> -		dev_err(i2c->dev, "no I2C frequency data\n");
>> +	/*
>> +	 * "clock-rate" is a legacy binding, the official binding is
>> +	 * "clock-frequency".  Try the official one first and then
>> +	 * fall back if it doesn't exist.
>> +	 */
>> +	data = of_get_property(pdev->dev.of_node, "clock-frequency",&len);
>> +	if (!data || len != sizeof(*data))
>> +		data = of_get_property(pdev->dev.of_node, "clock-rate",&len);
>> +	if (data&&  len == sizeof(*data)) {
>> +		i2c->twsi_freq = be32_to_cpup(data);
>
> Can't you use of_property_read_u32?

I will investigate, and use it if possible.

>
> Does the legacy binding really exist as DT support is new?
>

Yes, unfortunately some bootloaders escaped into the wild with the 
"clock-rate" binding.

> Otherwise,
>
> Acked-by: Rob Herring<rob.herring-bsGFqQB8/DxBDgjK7y7TUQ@public.gmane.org>
>

Thanks,