On 2023-12-11 at 20:19 +02, Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx> wrote:
> Hi Mikhail,
>
> Thank you for the patch.
>
> On Mon, Dec 11, 2023 at 08:50:11PM +0300, Mikhail Rudenko wrote:
>> As the sensor may be accessible right after its async sub-device is
>> registered, enable runtime PM before doing so.
>
> While at it, could you also switch to runtime PM autosuspend, possibly
> in a separate patch ? See for instance the imx290 driver.
I actually had a plan to do this in this series, but it had grown a
little bigger then expected and I had to drop a few non-essential
patches, including PM autosuspend. But if you suggest, I'll bring in
back in v2.
>> Signed-off-by: Mikhail Rudenko <mike.rudenko@xxxxxxxxx>
>> ---
>> drivers/media/i2c/ov4689.c | 8 ++++----
>> 1 file changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/media/i2c/ov4689.c b/drivers/media/i2c/ov4689.c
>> index 2eef64cd0070..ba33b0ced532 100644
>> --- a/drivers/media/i2c/ov4689.c
>> +++ b/drivers/media/i2c/ov4689.c
>> @@ -874,16 +874,16 @@ static int ov4689_probe(struct i2c_client *client)
>> goto err_clean_entity;
>> }
>>
>> + pm_runtime_set_active(dev);
>> + pm_runtime_enable(dev);
>> + pm_runtime_idle(dev);
>> +
>> ret = v4l2_async_register_subdev_sensor(sd);
>> if (ret) {
>> dev_err(dev, "v4l2 async register subdev failed\n");
>> goto err_clean_subdev;
>
> Don't you need to disable runtime PM in the error path ?
Ack, will improve error in v2.
>> }
>>
>> - pm_runtime_set_active(dev);
>> - pm_runtime_enable(dev);
>> - pm_runtime_idle(dev);
>> -
>> return 0;
>>
>> err_clean_subdev:
--
Best regards,
Mikhail Rudenko
