On Wednesday, 20 of February 2008, Alan Stern wrote:
quoted text
> On Wed, 20 Feb 2008, Rafael J. Wysocki wrote:
> 
> > Well, below is an uncompiled and untested but illustrating the idea that
> > might allow people not to bother with device_pm_schedule_removal()
> > explicitly and can fix the issue at hand.
> > 
> > [There are some cases that need handling and are not covered here.]
> > 
> > Please have a look.
> > 
> > Thanks,
> > Rafael
> 
> > +static struct task_struct *suspending_task;
> > +static DEFINE_MUTEX(suspending_task_mtx);
> 
> I suspect you don't really need this mutex.
> 
> > +bool in_suspend_context(void)
> > +{
> > +	bool result;
> > +
> > +	mutex_lock(&suspending_task_mtx);
> > +	result = (suspending_task == current);
> > +	mutex_unlock(&suspending_task_mtx);
> > +	return result;
> > +}
> 
> If suspending_task == current then you are guaranteed to be serialized, 
> because everything a single task does is serial.

But in principle there could be a concurrent thread removind the device and
that should block on dev->sem held by us.

Right now that's not very likely to happen thanks to the freezer, but we're
doing all this stuff, because we want to get rid of the freezer eventually. :-)

quoted text
> > @@ -1162,7 +1162,10 @@ void device_destroy(struct class *class,
> >  	dev = class_find_device(class, &devt, __match_devt);
> >  	if (dev) {
> >  		put_device(dev);
> > -		device_unregister(dev);
> > +		if (in_suspend_context())
> > +			device_pm_schedule_removal(dev);
> > +		else
> > +			device_unregister(dev);
> >  	}
> >  }
> 
> But what about device_del()?  Can a similar change be made there?

I believe so.

I'm working on a more complete patch right now.

Thanks,
Rafael
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/