Re: [PATCH 2/4] ARM: SAMSUNG: Add API to set platform data for s5p-tv driver

[Tushar Behera <tushar.behera@linaro.org>]

On 07/20/2012 04:26 AM, Kukjin Kim wrote:
> 
>> -----Original Message-----
>> From: Tushar Behera [mailto:tushar.behera@linaro.org]
>> Sent: Tuesday, July 17, 2012 12:33 PM
>> To: Kukjin Kim
>> Cc: 'Sachin Kamat'; linux-samsung-soc@vger.kernel.org; patches@linaro.org;
>> a.sim@samsung.com; jiun.yu@samsung.com
>> Subject: Re: [PATCH 2/4] ARM: SAMSUNG: Add API to set platform data for
>> s5p-tv driver
>>
>> On 07/17/2012 08:11 AM, Kukjin Kim wrote:
>>> Sachin Kamat wrote:
>>>>
>>>> From: Tushar Behera <tushar.behera@linaro.org>
>>>>
>>>> Commit 350f2f4dad64 ("[media] v4l: s5p-tv: hdmi: add support for
>>>> platform data") makes the presence of platform data mandatory for s5p-
>> tv
>>>> driver. Adding an API to plat-samsung for this purpose.
>>>>
>>>> Signed-off-by: Tushar Behera <tushar.behera@linaro.org>
>>>> Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
>>>> ---
>>>>  arch/arm/plat-samsung/devs.c              |   26
>>> ++++++++++++++++++++++++++
>>>>  arch/arm/plat-samsung/include/plat/hdmi.h |   16 ++++++++++++++++
>>>>  2 files changed, 42 insertions(+), 0 deletions(-)
>>>>  create mode 100644 arch/arm/plat-samsung/include/plat/hdmi.h
>>>>
>>>> diff --git a/arch/arm/plat-samsung/devs.c b/arch/arm/plat-
>> samsung/devs.c
>>>> index 07b5ac6..921ed88 100644
>>>> --- a/arch/arm/plat-samsung/devs.c
>>>> +++ b/arch/arm/plat-samsung/devs.c
>>>> @@ -32,6 +32,8 @@
>>>>  #include <linux/platform_data/s3c-hsudc.h>
>>>>  #include <linux/platform_data/s3c-hsotg.h>
>>>>
>>>> +#include <media/s5p_hdmi.h>
>>>> +
>>>>  #include <asm/irq.h>
>>>>  #include <asm/pmu.h>
>>>>  #include <asm/mach/arch.h>
>>>> @@ -759,6 +761,30 @@ void __init s5p_i2c_hdmiphy_set_platdata(struct
>>>> s3c2410_platform_i2c *pd)
>>>>  	npd = s3c_set_platdata(pd, sizeof(struct s3c2410_platform_i2c),
>>>>  			       &s5p_device_i2c_hdmiphy);
>>>>  }
>>>> +
>>>> +struct s5p_hdmi_platform_data s5p_hdmi_def_platdata;
>>>> +
>>>> +void __init s5p_hdmi_set_platdata(struct i2c_board_info *hdmiphy_info,
>>>> +				struct i2c_board_info *mhl_info, int
>>> mhl_bus)
>>>> +{
>>>> +	struct s5p_hdmi_platform_data *pd = &s5p_hdmi_def_platdata;
>>>> +
>>>> +	if (soc_is_exynos4210() || soc_is_exynos4212() ||
>>>> +						soc_is_exynos4412())
>>>> +		pd->hdmiphy_bus = 8;
>>>> +	else if (soc_is_s5pv210())
>>>> +		pd->hdmiphy_bus = 3;
>>>> +	else
>>>> +		pd->hdmiphy_bus = 0;
>>>> +
>>>> +	pd->hdmiphy_info = hdmiphy_info;
>>>> +	pd->mhl_info = mhl_info;
>>>> +	pd->mhl_bus = mhl_bus;
>>>> +
>>>> +	s3c_set_platdata(pd, sizeof(struct s5p_hdmi_platform_data),
>>>> +				&s5p_device_hdmi);
>>>> +}
>>>
>>> (Cc'ed Ayoung Sim and Jiun Yu in my team)
>>>
>>> Totally I wonder, do we really need s5p_i2c_hdmiphy_set_platdata and
>> this
>>> adding together? You updated s5p_i2c_hdmiphy_set_platdata in your 1/4
>> patch
>>> so I don't see why we should keep the bus number in duplicated...
>>>
>>
>> s5p_i2c_hdmiphy_set_platdata sets platform data for
>> s5p_device_i2c_hdmiphy. 'bus_num' assigned in that structure is not
>> accessible in s5p-tv driver. But s5p-tv driver needs the bus number to
>> get appropriate i2c adapter. Hence we need to update the bus number in
>> both the places.
>>
> Yeah, I know the situation, but I don't know why we should having the same
> i2c channel is included in different platform_data. I need to talk about
> this approach to hdmi guys in my team, I'll let you know as soon as
> possible.
> 

While configuring all the i2c channels, we need to pass the channel
number. Since the channel number is not fixed for hdmiphy, it was
updated as per respective CPUs.

The same would be done for other I2C channels also, we pass channel
number in platform data (s3c_i2c0_set_platdata) and then specify the bus
number as the first parameter in i2c_register_board_info(...) while
adding i2c devices.

It would be great if we can find a resolution on this before the merge
window. HDMI support in mainline kernel is broken since 3.4-rc1 and we
would love to get this fixed during 3.6.

> Thanks.
> 
> Best regards,
> Kgene.
> --
> Kukjin Kim <kgene.kim@samsung.com>, Senior Engineer,
> SW Solution Development Team, Samsung Electronics Co., Ltd.
> 


-- 
Tushar Behera