Hi Jiada
> > 2nd, can we keep usrcnt setup as-is ?
> > I guess we can just avoid rsnd_ssi_master_clk_start() if ssi->rate was not
> > 0 ?
>
> I don't fully understand your 2nd question,
> in case of rsnd_ssi_master_clk_stop(), if avoid
> rsnd_ssi_master_clk_stop() when ssi->rate is 0 by apply following
> change
>
> static void rsnd_ssi_master_clk_stop(struct rsnd_mod *mod,
> struct rsnd_dai_stream *io)
> {
> ...
> - if (ssi->usrcnt > 1)
> + if (ssi->rate == 0)
> return;
> ...
> }
>
> then when any IO stream with same SSI calls .hw_free,
> the other IO stream's clock will be stopped too.
I think we can find more simple solution if we can find good ideas.
For example, how about to add new counter for hw_params/hw_free ?
Anyway, [3/3] patch is too much over-kill to me.
And, please don't exchange usrcnt inc/dec position at [2/3].
It is for open/close.
Thank you for your help !!
Best regards
---
Kuninori Morimoto
