Re: ASoC updates for v4.2

[Takashi Iwai <tiwai@suse.de>]

At Mon, 22 Jun 2015 11:30:34 +0100,
Mark Brown wrote:
> 
> On Mon, Jun 22, 2015 at 11:58:24AM +0200, Takashi Iwai wrote:
> 
> > In addition to rsnd-card.c, I get a compile warning as below:
> 
> > sound/soc/mediatek/mtk-afe-pcm.c:1035:12: warning: 'mtk_afe_runtime_suspend' defined but not used [-Wunused-function]
> 
> > And, looking at the code, it seems calling runtime suspend in the
> > following way at probe:
> 
> >         pm_runtime_enable(&pdev->dev);
> >         if (!pm_runtime_enabled(&pdev->dev)) {
> >                 ret = mtk_afe_runtime_resume(&pdev->dev);
> >                 if (ret)
> >                         goto err_pm_disable;
> >         }
> 
> I'm confused, where's the call to runtime suspend?

It's in
static const struct dev_pm_ops mtk_afe_pm_ops = {
	SET_RUNTIME_PM_OPS(mtk_afe_runtime_suspend, mtk_afe_runtime_resume,
			   NULL)
};

But my concern above isn't about the warning itself.  I just stumbled
on the code invoking runtime resume while looking at this warning, and
wondered the behavior with CONFIG_PM=n.

Usually this kind of warning could be simply fixed by adding a proper
ifdef.  But, this driver calls runtime resume in the probe manually.

> > I'm not sure whether this really behaves correctly, especially when a
> > kernel is built without CONFIG_PM.
> 
> Could you be more specific about the problem you're seeing?  If runtime
> PM is disabled pm_runtime_enabled() will return false and we'll run
> through the resume path during probe() instead, otherwise we'll runtime
> resume whenever we need to use the hardware.

The runtime suspend is never called when CONFIG_PM=n.  OTOH, we call
runtime resume *always* at probe when CONFIG_PM=n.  This looks
inconsistent to me.

If it's a part of the mandatory initialization, it should be named
explicitly so, and make the runtime resume callback just calls it
instead.


Takashi