Re: [PATCH/RFC] MMC: remove unbalanced pm_runtime_suspend()

From: "Rafael J. Wysocki" <rjw@sisk.pl>
To: Alan Stern <stern@rowland.harvard.edu>
Cc: Guennadi Liakhovetski <g.liakhovetski@gmx.de>,
	Ohad Ben-Cohen <ohad@wizery.com>,
	linux-mmc@vger.kernel.org, Magnus Damm <damm@opensource.se>,
	Simon Horman <horms@verge.net.au>,
	Linux PM mailing list <linux-pm@lists.linux-foundation.org>
Subject: Re: [PATCH/RFC] MMC: remove unbalanced pm_runtime_suspend()
Date: Thu, 21 Apr 2011 20:00:29 +0200	[thread overview]
Message-ID: <201104212000.29754.rjw@sisk.pl> (raw)
In-Reply-To: <Pine.LNX.4.44L0.1104210945470.1939-100000@iolanthe.rowland.org>

On Thursday, April 21, 2011, Alan Stern wrote:
> On Wed, 20 Apr 2011, Rafael J. Wysocki wrote:
> 
> > On Wednesday, April 20, 2011, Alan Stern wrote:
> > > On Wed, 20 Apr 2011, Rafael J. Wysocki wrote:
> > > 
> > > > On Wednesday, April 20, 2011, Alan Stern wrote:
> > > > > On Wed, 20 Apr 2011, Guennadi Liakhovetski wrote:
> > > > ...
> > > > > Ah, now I see the problem.  It looks like we did not give sufficient
> > > > > thought to the case where a device starts off (and therefore should
> > > > > finish up) in a powered-down state.  Calling pm_runtime_put_sync()
> > > > > after unbinding the device driver seems a little futile -- with no
> > > > > driver, the subsystem may not be able to power-down the device!
> > > > > 
> > > > > Rafael, how do you think we should handle this?  Get rid of the 
> > > > > pm_runtime_get_no_resume() and pm_runtime_put_sync() calls in 
> > > > > dd.c:__device_release_driver()?
> > > > 
> > > > I think we need pm_runtime_barrier() in there.  Otherwise we risk
> > > > removing the driver while there's a runtime PM request pending.
> > > > 
> > > > But we can move the pm_runtime_put_sync() before driver_sysfs_remove().
> > > 
> > > What happens if another runtime PM request is queued between the
> > > put_sync() and the remove callback?  We may need a safe way to prevent
> > > async runtime PM requests while still allowing synchronous requests.
> > 
> > What about making a rule that it is invalid to schedule a future suspend
> > or queue a resume request of a device whose driver is being removed?
> > 
> > Arguably, we can't prevent people from shooting themselves in the foot this
> > way or another and I'm not sure if this particular case is worth additional
> > handling.
> 
> After thinking about this, I tend to agree.  The synchronization 
> issues, combined with the unknown needs of the driver, make this very 
> difficult to handle in the PM core.
> 
> Here's another possible approach: If a driver wants to leave its device 
> in a powered-down state after unbinding then it can invoke its own 
> runtime_suspend callback directly, in the following way:
> 
> 	... unregister all child devices below dev ...
> 	pm_runtime_disable(dev);
> 	if (dev->power.runtime_status != RPM_SUSPENDED) {
> 		pm_set_suspended(dev);
> 		my_runtime_suspend_callback(dev);
> 	}

I think this would work too, but then possibly many drivers would have to
do the same thing in their "remove" routines.

> There may be issues regarding coordination with the subsystem or the
> power domain; at the moment it's not clear what should be done.  Maybe
> the runtime-PM core should include an API for directly invoking the
> appropriate callbacks.

If we choose this approach, then yes, we should provide a suitable API, but
I'm still thinking it would be simpler to move the pm_runtime_put_sync() before driver_sysfs_remove() and make the rule as I said previously. :-)

Rafael