Re: [PATCH 5/6] media: staging: rkisp1: params: add '__must_hold' to rkisp1_params_apply_params_cfg

[Helen Koike <helen.koike@xxxxxxxxxxxxx>]

Hi Dafna,

On 10/2/20 3:42 PM, Dafna Hirschfeld wrote:
> The function rkisp1_params_apply_params_cfg must be under the
> the lock of params->config_lock. Add the __must_hold annotation
> to indicate it.
> 
> Signed-off-by: Dafna Hirschfeld <dafna.hirschfeld@xxxxxxxxxxxxx>
> ---
>  drivers/staging/media/rkisp1/rkisp1-params.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/staging/media/rkisp1/rkisp1-params.c b/drivers/staging/media/rkisp1/rkisp1-params.c
> index 3afbc24ca05e..aa18a113245b 100644
> --- a/drivers/staging/media/rkisp1/rkisp1-params.c
> +++ b/drivers/staging/media/rkisp1/rkisp1-params.c
> @@ -1187,6 +1187,7 @@ static void rkisp1_isp_isr_meas_config(struct rkisp1_params *params,
>  
>  static void rkisp1_params_apply_params_cfg(struct rkisp1_params *params,
>  					   unsigned int frame_sequence)
> +	__must_hold(&params->config_lock)
>  {
>  	struct rkisp1_params_cfg *new_params;
>  	struct rkisp1_buffer *cur_buf = NULL;
> 

Instead of adding __must_hold, why no to lock inside this function?
It seems there are only two places that call it, and they don't do anything else
in the critical section.

Regards,
Helen