RE: [PATCH 3/3] Thermal: do thermal zone update after a cooling device registered

[Chen, Yu C]

(resend for broken display)

Hi,
> -----Original Message-----
> From: Javi Merino [mailto:javi.merino@xxxxxxx]
> Sent: Thursday, October 15, 2015 10:05 PM
> To: Chen, Yu C
> Cc: linux-pm@xxxxxxxxxxxxxxx; edubezval@xxxxxxxxx; Zhang, Rui; linux- 
> kernel@xxxxxxxxxxxxxxx; stable@xxxxxxxxxxxxxxx; Pandruvada, Srinivas
> Subject: Re: [PATCH 3/3] Thermal: do thermal zone update after a 
> cooling device registered
> 
> On Wed, Oct 14, 2015 at 07:23:55PM +0000, Chen, Yu C wrote:
> > > -----Original Message-----
> > > From: Javi Merino [mailto:javi.merino@xxxxxxx]
> > > Sent: Thursday, October 15, 2015 1:08 AM
> > > To: Chen, Yu C
> > > Cc: linux-pm@xxxxxxxxxxxxxxx; edubezval@xxxxxxxxx; Zhang, Rui;
> > > linux- kernel@xxxxxxxxxxxxxxx; stable@xxxxxxxxxxxxxxx; Pandruvada, 
> > > Srinivas
> > > Subject: Re: [PATCH 3/3] Thermal: do thermal zone update after a 
> > > cooling device registered
> > >
> > > On Mon, Oct 12, 2015 at 09:23:28AM +0000, Chen, Yu C wrote:
> > > > Hi, Javi
> > > > Sorry for my late response,
> > > >
> > > > > -----Original Message-----
> > > > > From: Javi Merino [mailto:javi.merino@xxxxxxx]
> > > > > Sent: Wednesday, September 30, 2015 12:02 AM
> > > > > To: Chen, Yu C
> > > > > Cc: linux-pm@xxxxxxxxxxxxxxx; edubezval@xxxxxxxxx; Zhang, Rui;
> > > > > linux- kernel@xxxxxxxxxxxxxxx; stable@xxxxxxxxxxxxxxx
> > > > > Subject: Re: [PATCH 3/3] Thermal: do thermal zone update after 
> > > > > a cooling device registered
> > > > >
> > > > > Hi Yu,
> > > > >
> > > > > On Mon, Sep 28, 2015 at 06:52:00PM +0100, Chen, Yu C wrote:
> > > > > > Hi, Javi,
> > > > > >
> > > > > > > -----Original Message-----
> > > > > > > From: Javi Merino [mailto:javi.merino@xxxxxxx]
> > > > > > > Sent: Monday, September 28, 2015 10:29 PM
> > > > > > > To: Chen, Yu C
> > > > > > > Cc: linux-pm@xxxxxxxxxxxxxxx; edubezval@xxxxxxxxx; Zhang, 
> > > > > > > Rui;
> > > > > > > linux- kernel@xxxxxxxxxxxxxxx; stable@xxxxxxxxxxxxxxx
> > > > > > > Subject: Re: [PATCH 3/3] Thermal: do thermal zone update 
> > > > > > > after a cooling device registered
> > > > > > >
> > > > > > > On Sun, Sep 27, 2015 at 06:48:44AM +0100, Chen Yu wrote:
> > > > > > > > From: Zhang Rui <rui.zhang@xxxxxxxxx>
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > > > I think you need to hold cdev->lock here, to make sure 
> > > > > > > that no thermal zone is added or removed from
> > > > > > > cdev->thermal_instances while
> > > > > you are looping.
> > > > > > >
> > > > > > Ah right, will add. If I add the cdev ->lock here, will 
> > > > > > there be a AB-BA lock with thermal_zone_unbind_cooling_device?
> > > > >
> > > > > You're right, it could lead to a deadlock.  The locks can't be 
> > > > > swapped because that won't work in step_wise.
> > > > >
> > > > > The best way that I can think of accessing thermal_instances 
> > > > > atomically is by making it RCU protected instead of with mutexes.
> > > > > What do you think?
> > > > >
> > > > RCU would need extra spinlocks to protect the list, and need to 
> > > > sync_rcu after we delete one instance from thermal_instance 
> > > > list, I think it is too complicated for me to rewrite: ( How 
> > > > about using
> > > thermal_list_lock instead of cdev ->lock?
> > > > This guy should be big enough to protect the 
> > > > device.thermal_instance
> list.
> > >
> > > thermal_list_lock protects thermal_tz_list and thermal_cdev_list, 
> > > but it doesn't protect the thermal_instances list.  For example,
> > > thermal_zone_bind_cooling_device() adds a cooling device to the
> > > cdev->thermal_instances list without taking thermal_tz_list.
> > >
> > Before thermal_zone_bind_cooling_device is invoked, the 
> > thermal_list_lock will be firstly gripped:
> >
> > static void bind_cdev(struct thermal_cooling_device *cdev) { 
> > mutex_lock(&thermal_list_lock);
> > either tz->ops->bind    :   thermal_zone_bind_cooling_device
> > or __bind()  :   thermal_zone_bind_cooling_device
> > mutex_unlock(&thermal_list_lock);
> > }
> >
> > And it is the same as in  passive_store.
> > So when code is trying to add/delete thermal_instance of cdev, he 
> > has already hold thermal_list_lock IMO. Or do I miss anything?
> 
> thermal_zone_bind_cooling_device() is exported, so you can't really 
> rely on the static thermal_list_lock being acquired in every single call.
> 
> thermal_list_lock and protects the lists thermal_tz_list and thermal_cdev_list.
> Making it implicitly protect the cooling device's and thermal zone 
> device's instances list because no sensible code would call
> thermal_zone_bind_cooling_device() outside of a bind function is just 
> asking for trouble.
> 
Yes, from this point of view,it is true. 

> Locking is hard to understand and easy to get wrong so let's keep it simple.
> 
How about the following 2 methods:
1. avoid accessing device's thermal_instance,
but access all thermal_zone_device directly,
 although there might be some redundancy,
 some thermal zones do not need to be updated, 
but we can avoid gripping dev->lock:

	mutex_lock(&thermal_list_lock);
	list_for_each_entry(pos, &thermal_tz_list, node)
		thermal_zone_device_update(tz);
	mutex_unlock(&thermal_list_lock); 

or,
2. Once we bind the new device with the thermal_zone_device,
 we can record that thermal_zone_device, 
and update that thermal_zone_device alone,the the code would be:
	
	mutex_lock(&thermal_list_lock);
	list_for_each_entry(pos, &thermal_tz_list, node){
		if (tz->need_update)
			thermal_zone_device_update(tz);
	}
	mutex_unlock(&thermal_list_lock);


BTW, since thermal_zone_device_update is not atomic, 
we might need another patch to make it into atomic or 
something like that, but for now, I think these three patches 
are just for fixing the regressions.


Thanks

Best Regards,
Yu