Re: [PATCH v2 19/20] media: i2c: ov4689: Refactor ov4689_s_stream

[Mikhail Rudenko <mike.rudenko@gmail.com>]

Hi Laurent,

and thanks for the review!

On 2024-02-23 at 13:48 +02, Laurent Pinchart <laurent.pinchart@ideasonboard.com> wrote:

> Hi Mikhail,
>
> Thank you for the patch.
>
> On Mon, Dec 18, 2023 at 08:40:40PM +0300, Mikhail Rudenko wrote:
>> Remove repetitive pm_runtime_put calls in ov4689_s_stream function,
>> and call pm_runtime_put once at the end of the "on" branch if any
>> error occurred.
>>
>> Signed-off-by: Mikhail Rudenko <mike.rudenko@gmail.com>
>> ---
>>  drivers/media/i2c/ov4689.c | 29 ++++++++++-------------------
>>  1 file changed, 10 insertions(+), 19 deletions(-)
>>
>> diff --git a/drivers/media/i2c/ov4689.c b/drivers/media/i2c/ov4689.c
>> index e997c3231e85..884761d02119 100644
>> --- a/drivers/media/i2c/ov4689.c
>> +++ b/drivers/media/i2c/ov4689.c
>> @@ -555,35 +555,26 @@ static int ov4689_s_stream(struct v4l2_subdev *sd, int on)
>>  					  ov4689_common_regs,
>>  					  ARRAY_SIZE(ov4689_common_regs),
>>  					  NULL);
>> -		if (ret) {
>> -			pm_runtime_put_sync(dev);
>> -			goto unlock_and_return;
>> -		}
>> +		if (ret)
>> +			goto cleanup_pm;
>>
>>  		ret = ov4689_setup_timings(ov4689, sd_state);
>> -		if (ret) {
>> -			pm_runtime_put(dev);
>> -			goto unlock_and_return;
>> -		}
>> +		if (ret)
>> +			goto cleanup_pm;
>>
>>  		ret = ov4689_setup_blc_anchors(ov4689, sd_state);
>> -		if (ret) {
>> -			pm_runtime_put(dev);
>> -			goto unlock_and_return;
>> -		}
>> +		if (ret)
>> +			goto cleanup_pm;
>>
>>  		ret = __v4l2_ctrl_handler_setup(&ov4689->ctrl_handler);
>> -		if (ret) {
>> -			pm_runtime_put_sync(dev);
>> -			goto unlock_and_return;
>> -		}
>> +		if (ret)
>> +			goto cleanup_pm;
>>
>>  		ret = cci_write(ov4689->regmap, OV4689_REG_CTRL_MODE,
>>  				OV4689_MODE_STREAMING, NULL);
>> -		if (ret) {
>> +cleanup_pm:
>
> A label within an if branch isn't great, readability-wise :-S Could we
> maybe split the ov4687_s_stream() function in two (streamon and
> streamoff, or similar names) ? You would then have a single
> pm_runtime_put_sync() call in ov4689_s_stream(), in the error handling
> path for the streamon function call.

Okay, will split it in v3.

>> +		if (ret)
>>  			pm_runtime_put_sync(dev);
>> -			goto unlock_and_return;
>> -		}
>>  	} else {
>>  		cci_write(ov4689->regmap, OV4689_REG_CTRL_MODE,
>>  			  OV4689_MODE_SW_STANDBY, NULL);


--
Best regards,
Mikhail Rudenko