Re: [PATCH V2] dmaengine: tegra-apb: Really fix runtime-pm usage

[Vinod Koul <vinod.koul-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>]

On Tue, Jun 27, 2017 at 12:44:58PM +0100, Jon Hunter wrote:

> >>  static const struct dev_pm_ops tegra_dma_dev_pm_ops = {
> >>  	SET_RUNTIME_PM_OPS(tegra_dma_runtime_suspend, tegra_dma_runtime_resume,
> >>  			   NULL)
> >> -	SET_SYSTEM_SLEEP_PM_OPS(tegra_dma_pm_suspend, tegra_dma_pm_resume)
> >> +	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
> >> +				pm_runtime_force_resume)
> > 
> > Is that even necessary? I thought runtime PM was going to be triggered
> > for system sleep anyway, but it looks like there are other examples of
> > this usage, so maybe I'm mistaken.
> 
> Yes this is necessary. No RPM is not automatically trigger by system
> suspend AFAICT.

Yes I was earlier under the same impression but later did realize that the
behaviour seems to be arch specific and we don't have guarantee on this

-- 
~Vinod