[U-Boot] [PATCH] arm: socfpga: Enable DM_I2C

[Marek Vasut <marex@denx.de>]

On 03/18/2016 01:16 PM, Stefan Roese wrote:
> On 18.03.2016 12:20, Marek Vasut wrote:
>> On 03/18/2016 08:55 AM, Stefan Roese wrote:
>>> The designware I2C driver now supports DM. So lets use it and enable
>>> DM_I2C for this platform per default.
>>>
>>> Signed-off-by: Stefan Roese <sr@denx.de>
>>> Cc: Marek Vasut <marex@denx.de>
>>> ---
>>>   arch/arm/Kconfig                 |  1 +
>>>   arch/arm/dts/socfpga.dtsi        |  4 ++++
>>>   include/configs/socfpga_common.h | 16 ----------------
>>>   3 files changed, 5 insertions(+), 16 deletions(-)
>>>
>>> diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
>>> index e5f57ef..98c1f10 100644
>>> --- a/arch/arm/Kconfig
>>> +++ b/arch/arm/Kconfig
>>> @@ -516,6 +516,7 @@ config ARCH_SOCFPGA
>>>       select DM
>>>       select DM_SPI_FLASH
>>>       select DM_SPI
>>> +    select DM_I2C
>>>
>>>   config TARGET_CM_T43
>>>       bool "Support cm_t43"
>>> diff --git a/arch/arm/dts/socfpga.dtsi b/arch/arm/dts/socfpga.dtsi
>>> index 8588221..fe55722 100644
>>> --- a/arch/arm/dts/socfpga.dtsi
>>> +++ b/arch/arm/dts/socfpga.dtsi
>>> @@ -23,6 +23,10 @@
>>>           spi0 = &qspi;
>>>           spi1 = &spi0;
>>>           spi2 = &spi1;
>>> +        i2c0 = &i2c0;
>>> +        i2c1 = &i2c1;
>>> +        i2c2 = &i2c2;
>>> +        i2c3 = &i2c3;
>>
>> Something tells me that you should be super-careful here, because some
>> socfpga boards actually do use i2c. But just adding the aliases here
>> without actually enabling the i2c via DT now will cause breakage.
> 
> Right. But if the boards are also used in Linux and the dts is synced
> with the Linux one, then the I2C DT node should be enabled, right?

In theory, yes. I hope practice is in-line with theory here.

>> One way to solve it is to enable all four buses in socfpga-cyclone5.dtsi
>> but that's not something I find very appealing.
> 
> I don't like it. In the Linux socpfga.dtsi the I2C DT nodes are also
> disabled per default. We should stay in sync here.

I agree with this.

>> The other is to keep the
>> patch this way and wait until someone complains his i2c is not working.
>>
>> What do you think ?
> 
> Or we could move the aliases into the board specific dts files? I
> could start with the SR1500 board.
> 
> What do you think?

The aliases are fine. It's the "status = "okay"" bit which I suspect
might be missing for some boards. But then, let's see if something
breaks, we should be able to find it before 2016.05 is out anyway.

-- 
Best regards,
Marek Vasut