On Fri, Dec 12, 2025 at 10:04:47AM +0530, Dutta, Anurag wrote:
> On 09-12-2025 16:00, Mark Brown wrote:

Quoting seems messed up here - your whole mail is quoted.

> > I think we want this, possibly using pm_runtime_force_resume() instead
> > (not 100% sure on that one, glancing at the documentation there might be
> > issues though it feels like the intent of what we're doing here).  Can
> > you send a patch please?

> > Hi Mark
> > If we use pm_runtime_force_resume(), then for CONFIG_PM_SLEEP, it will work.
> > But, in case of !CONFIG_PM_SLEEP, pm_runtime_force_resume() returns -ENXIO,
> > because of which :

Huh, hadn't noticed an interaction with PM_SLEEP.  My thinking was
partly that this is convenient since it lets you know if runtime PM
isn't enabled and you need to unwind the clocks separately.  But I don't
super mind so long as we've got something that works.

> > Let me send a patch with pm_runtime_get_sync(). If changes are required,
> > please let me know.

Yes, please.