Re: [PATCH v3 19/20] media: i2c: ov4689: Refactor ov4689_s_stream

From: Mikhail Rudenko
Date: Mon Mar 04 2024 - 16:50:04 EST



Hi Laurent,

and thanks for the review.

On 2024-03-04 at 23:31 +02, Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx> wrote:

> Hi Mikhail,
>
> Thank you for the patch.
>
> On Thu, Feb 29, 2024 at 07:53:32PM +0300, Mikhail Rudenko wrote:
>> Split ov4689_s_stream into __ov4689_stream_on and __ov4689_stream_off
>> functions. Also remove repetitive pm_runtime_put calls and call
>> pm_runtime_put once at the end of the __ov4689_stream_off function if
>> any error occurred.
>>
>> Signed-off-by: Mikhail Rudenko <mike.rudenko@xxxxxxxxx>
>> ---
>> drivers/media/i2c/ov4689.c | 100 ++++++++++++++++++++-----------------
>> 1 file changed, 53 insertions(+), 47 deletions(-)
>>
>> diff --git a/drivers/media/i2c/ov4689.c b/drivers/media/i2c/ov4689.c
>> index 2496067b90a0..5cea9b5ba201 100644
>> --- a/drivers/media/i2c/ov4689.c
>> +++ b/drivers/media/i2c/ov4689.c
>> @@ -537,61 +537,67 @@ static int ov4689_setup_blc_anchors(struct ov4689 *ov4689,
>> return ret;
>> }
>>
>> +static int __ov4689_stream_on(struct ov4689 *ov4689,
>
> No need for the __ prefix. Same for __ov4689_stream_off().

Will remove the prefix in v4.

>> + struct v4l2_subdev_state *sd_state)
>> +{
>> + int ret;
>> +
>> + ret = pm_runtime_resume_and_get(ov4689->dev);
>> + if (ret < 0)
>> + return ret;
>> +
>> + ret = cci_multi_reg_write(ov4689->regmap, ov4689_common_regs,
>> + ARRAY_SIZE(ov4689_common_regs), NULL);
>> + if (ret)
>> + goto cleanup_pm;
>> +
>> + ret = ov4689_setup_timings(ov4689, sd_state);
>> + if (ret)
>> + goto cleanup_pm;
>> +
>> + ret = ov4689_setup_blc_anchors(ov4689, sd_state);
>> + if (ret)
>> + goto cleanup_pm;
>> +
>> + ret = __v4l2_ctrl_handler_setup(&ov4689->ctrl_handler);
>> + if (ret)
>> + goto cleanup_pm;
>> +
>> + ret = cci_write(ov4689->regmap, OV4689_REG_CTRL_MODE,
>> + OV4689_MODE_STREAMING, NULL);
>> + if (ret)
>> + goto cleanup_pm;
>> +
>> + return 0;
>> +
>> + cleanup_pm:
>
> No space before the label. I would also name it just "error".

Thanks for the suggestion, will do so in v4.

> With those small changes,
>
> Reviewed-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
>
>> + pm_runtime_put(ov4689->dev);
>> + return ret;
>> +}
>> +
>> +static int __ov4689_stream_off(struct ov4689 *ov4689)
>> +{
>> + cci_write(ov4689->regmap, OV4689_REG_CTRL_MODE, OV4689_MODE_SW_STANDBY,
>> + NULL);
>> + pm_runtime_mark_last_busy(ov4689->dev);
>> + pm_runtime_put_autosuspend(ov4689->dev);
>> +
>> + return 0;
>> +}
>> +
>> static int ov4689_s_stream(struct v4l2_subdev *sd, int on)
>> {
>> struct ov4689 *ov4689 = to_ov4689(sd);
>> struct v4l2_subdev_state *sd_state;
>> - struct device *dev = ov4689->dev;
>> - int ret = 0;
>> + int ret;
>>
>> sd_state = v4l2_subdev_lock_and_get_active_state(&ov4689->subdev);
>>
>> - if (on) {
>> - ret = pm_runtime_resume_and_get(dev);
>> - if (ret < 0)
>> - goto unlock_and_return;
>> -
>> - ret = cci_multi_reg_write(ov4689->regmap,
>> - ov4689_common_regs,
>> - ARRAY_SIZE(ov4689_common_regs),
>> - NULL);
>> - if (ret) {
>> - pm_runtime_put(dev);
>> - goto unlock_and_return;
>> - }
>> -
>> - ret = ov4689_setup_timings(ov4689, sd_state);
>> - if (ret) {
>> - pm_runtime_put(dev);
>> - goto unlock_and_return;
>> - }
>> -
>> - ret = ov4689_setup_blc_anchors(ov4689, sd_state);
>> - if (ret) {
>> - pm_runtime_put(dev);
>> - goto unlock_and_return;
>> - }
>> -
>> - ret = __v4l2_ctrl_handler_setup(&ov4689->ctrl_handler);
>> - if (ret) {
>> - pm_runtime_put(dev);
>> - goto unlock_and_return;
>> - }
>> -
>> - ret = cci_write(ov4689->regmap, OV4689_REG_CTRL_MODE,
>> - OV4689_MODE_STREAMING, NULL);
>> - if (ret) {
>> - pm_runtime_put(dev);
>> - goto unlock_and_return;
>> - }
>> - } else {
>> - cci_write(ov4689->regmap, OV4689_REG_CTRL_MODE,
>> - OV4689_MODE_SW_STANDBY, NULL);
>> - pm_runtime_mark_last_busy(dev);
>> - pm_runtime_put_autosuspend(dev);
>> - }
>> + if (on)
>> + ret = __ov4689_stream_on(ov4689, sd_state);
>> + else
>> + ret = __ov4689_stream_off(ov4689);
>>
>> -unlock_and_return:
>> v4l2_subdev_unlock_state(sd_state);
>>
>> return ret;


--
Best regards,
Mikhail Rudenko