Hi,
On Wed, 16 Feb 2022, Takashi Iwai wrote:
> > static void silent_stream_enable(struct hda_codec *codec,
> > struct hdmi_spec_per_pin *per_pin)
> > {
(snip)
> > unlock_out:
> > mutex_unlock(&per_pin->lock);
> > +
> > + if (err || !keep_power)
> > + snd_hda_power_down_pm(codec);
>
> So this may leave the power up. But where is the corresponding part
> that turns it down? The newly added snd_hda_power_up_pm() in
> silent_stream_disable() is paired with the one at the tail of that
> function, so it looks like a refcount unbalance. Kai?
indeed it is. I'll send a V2 with this fixed.
Br, Kai
