Re: [PATCH] media: davinci: vpif: add check for NULL handler

[Sekhar Nori <nsekhar@ti.com>]

On 8/17/2012 2:45 PM, Laurent Pinchart wrote:
> Hi Prabhakar,
> 
> On Friday 17 August 2012 14:02:19 Prabhakar Lad wrote:
>> On Friday 17 August 2012 11:07 AM, Sekhar Nori wrote:
>>> On 8/17/2012 10:51 AM, Prabhakar Lad wrote:
>>>> On Thursday 16 August 2012 08:43 PM, Laurent Pinchart wrote:
>>>>> On Thursday 16 August 2012 19:32:00 Prabhakar Lad wrote:
>>>>>> From: Lad, Prabhakar <prabhakar.lad@ti.com>
>>>>>>
>>>>>> Signed-off-by: Lad, Prabhakar <prabhakar.lad@ti.com>
>>>>>> Signed-off-by: Manjunath Hadli <manjunath.hadli@ti.com>
>>>>>> Cc: Hans Verkuil <hans.verkuil@cisco.com>
>>>>>> ---
>>>>>>
>>>>>>  drivers/media/video/davinci/vpif_capture.c |   12 +++++++-----
>>>>>>  drivers/media/video/davinci/vpif_display.c |   14 ++++++++------
>>>>>>  2 files changed, 15 insertions(+), 11 deletions(-)
>>>>>>
>>>>>> diff --git a/drivers/media/video/davinci/vpif_capture.c
>>>>>> b/drivers/media/video/davinci/vpif_capture.c index 266025e..a87b7a5
>>>>>> 100644
>>>>>> --- a/drivers/media/video/davinci/vpif_capture.c
>>>>>> +++ b/drivers/media/video/davinci/vpif_capture.c
>>>>>> @@ -311,12 +311,14 @@ static int vpif_start_streaming(struct vb2_queue
>>>>>> *vq,
>>>>>> unsigned int count) }
>>>>>>
>>>>>>  	/* configure 1 or 2 channel mode */
>>>>>>
>>>>>> -	ret = vpif_config_data->setup_input_channel_mode
>>>>>> -					(vpif->std_info.ycmux_mode);
>>>>>> +	if (vpif_config_data->setup_input_channel_mode) {
>>>>>> +		ret = vpif_config_data->setup_input_channel_mode
>>>>>> +						(vpif->std_info.ycmux_mode);
>>>>>>
>>>>>> -	if (ret < 0) {
>>>>>> -		vpif_dbg(1, debug, "can't set vpif channel mode\n");
>>>>>> -		return ret;
>>>>>> +		if (ret < 0) {
>>>>>> +			vpif_dbg(1, debug, "can't set vpif channel mode\n");
>>>>>> +			return ret;
>>>>>> +		}
>>>>>
>>>>> This change looks good to me. However, note that you will need to get
>>>>> rid of board code callbacks at some point to implement device tree
>>>>> support. It would be worth thinking about how to do so now.
>>>>
>>>> Currently VPIF driver is only used by dm646x, and the handlers for this
>>>> in the the board code are not null. This patch is intended for da850
>>>> where this handlers will be null.
>>>>
>>>>>>  	}
>>>>>>  	
>>>>>>  	/* Call vpif_set_params function to set the parameters and addresses
>>>>>>  	*/
>>>>>>
>>>>>> diff --git a/drivers/media/video/davinci/vpif_display.c
>>>>>> b/drivers/media/video/davinci/vpif_display.c index e129c98..1e35f92
>>>>>> 100644
>>>>>> --- a/drivers/media/video/davinci/vpif_display.c
>>>>>> +++ b/drivers/media/video/davinci/vpif_display.c
>>>>>> @@ -280,12 +280,14 @@ static int vpif_start_streaming(struct vb2_queue
>>>>>> *vq,
>>>>>> unsigned int count) }
>>>>>>
>>>>>>  	/* clock settings */
>>>>>>
>>>>>> -	ret =
>>>>>> -	    vpif_config_data->set_clock(ch->vpifparams.std_info.ycmux_mode,
>>>>>> -					ch->vpifparams.std_info.hd_sd);
>>>>>> -	if (ret < 0) {
>>>>>> -		vpif_err("can't set clock\n");
>>>>>> -		return ret;
>>>>>> +	if (vpif_config_data->set_clock) {
>>>>>
>>>>> Does the DaVinci platform use the common clock framework ? If so, a
>>>>> better fix for this would be to pass a clock name through platform data
>>>>> instead of using a callback function.
>>>>
>>>> Currently DaVinci is not using the common clock framework.
>>>>
>>>> Can you ACK this patch?
>>>
>>> Yes, DaVinci has not migrated to common clock framework (yet). However,
>>> even without that it should be possible to use clock API in driver code.
>>> Using a callback to enable clocks or even passing the clock name from
>>> platform data would be bypassing an existing framework. Clock name
>>> should be IP specific, so it should be possible to use that in driver.
>>
>> The callback is not actually dealing with PSC clock's but with system
>> module registers.
> 
> Good to know. Then you'll have to create an API to expose the system module 
> registers to drivers.

Yes something like that is needed. Konstantin Baydarov is working on a
MFD driver for a similar module that exists in OMAP world called
"control module".

I guess this patch can go in though since it does not add any new
accesses but merely lets the existing function be passed as NULL.

Thanks,
Sekhar

[1] https://patchwork.kernel.org/patch/1236491/