On Thu, Jan 31, 2019 at 12:59 PM Takashi Iwai tiwai@suse.de wrote:
On Thu, 31 Jan 2019 12:46:54 +0100, Rafael J. Wysocki wrote:
On Thu, Jan 31, 2019 at 12:21 PM Takashi Iwai tiwai@suse.de wrote:
On Thu, 31 Jan 2019 12:05:30 +0100, Thierry Reding wrote:
On Wed, Jan 30, 2019 at 05:40:42PM +0100, Takashi Iwai wrote:
[cut]
If I understand correctly the code, the pm domain is already activated at calling driver's probe callback.
As far as I can tell, the domain will also be powered off again after probe finished, unless the device grabs a runtime PM reference. This is what happens via the dev->pm_domain->sync() call after successful probe of a driver.
Ah, a good point. This can be a problem with a probe work like this case.
It seems to me like it's not a very well defined case what to do when a device needs to be powered up but runtime PM is not enabled.
Adding Rafael and linux-pm, maybe they can provide some guidance on what to do in these situations.
To summarize, what we're debating here is how to handle powering up a device if the pm_runtime infrastructure doesn't take care of it. Jon's proposal here was, and we use this elsewhere, to do something like this:
pm_runtime_enable(dev); if (!pm_runtime_enabled(dev)) { err = foo_runtime_resume(dev); if (err < 0) goto fail; }So basically when runtime PM is not available, we explicitly "resume" the device to power it up.
It seems to me like that's a fairly common problem, so I'm wondering if there's something that the runtime PM core could do to help with this. Or perhaps there's already a way to achieve this that we're all overlooking?
Rafael, any suggestions?
If any, a common helper would be appreciated, indeed.
I'm not sure that I understand the problem correctly, so let me restate it the way I understand it.
What we're talking about is a driver ->probe() callback. Runtime PM is disabled initially and the device is off. It needs to be powered up, but the way to do that depends on some configuration of the board etc., so ideally
pm_runtime_enable(dev); ret = pm_runtime_resume(dev);
should just work, but the question is what to do if runtime PM doesn't work as expected. That is, CONFIG_PM_RUNTIME is unset? Or something else?
Yes, the question is how to write the code for both with and without CONFIG_PM (or CONFIG_PM_RUNTIME).
This basically is about setup, because after that point all should just work in both cases.
Personally, I would do
if (IS_ENABLED(CONFIG_PM)) { do setup based on pm-runtime } else { do manual setup }
Right now, we have a code like below, pushing the initialization in an async work and let the probe returning quickly.
hda_tegra_probe() { ....
So why don't you do
if (!IS_ENABLED(CONFIG_PM)) { do manual clock setup }
here?
pm_runtime_enable(); schedule_work(); return; } hda_tegra_probe_work() { pm_runtime_get_sync(); .... pm_runtime_put_sync(); }Then it truned outhis code lacks of the clock initialization when runtime PM isn't enabled. Normally it's done via runtime resume
hda_tegra_runtime_resume() { hda_tegra_enable_clocks(); .... }And now the question is what is the standard idiom in such a case.
IMO, calling pm_runtime_resume() inside the probe function looks weird, and my preference was to initialize the clocks explicitly, then enable runtime PM. But if using pm_runtime_resume() in the proc should be seen as a standard procedure, I'm fine with that.
Well, people do pm_runtime_resume() in ->probe() too, but pm_runtime_resume() returns 1 for CONFIG_PM unset, so that won't give you what you want anyway. :-)
Cheers, Rafael