diff mbox

[2/3] Thermal: handle thermal zone device properly during system sleep

Message ID 1427174490-7974-3-git-send-email-rui.zhang@intel.com (mailing list archive)
State Superseded, archived
Delegated to: Zhang Rui
Headers show

Commit Message

Zhang Rui March 24, 2015, 5:21 a.m. UTC
Current thermal code does not handle system sleep well because
1. the cooling device cooling state may be changed during suspend
2. the previous temperature reading becomes invalid after resumed because
   it is got before system sleep
3. updating thermal zone device during suspending/resuming
   is wrong because some devices may have already been suspended
   or may have not been resumed.

Thus, the proper way to do this is to cancel all thermal zone
device update requirements during suspend/resume, and after all
the devices have been resumed, reset and update every registered
thermal zone devices.

This also fixes a regression introduced by
commit 19593a1fb1f6718406afca5b867dab184289d406
Author: Aaron Lu <aaron.lu@intel.com>
Date:   Tue Nov 19 16:59:20 2013 +0800

    ACPI / fan: convert to platform driver

    Convert ACPI fan driver to a platform driver for the purpose of phasing
    out ACPI bus.

    Signed-off-by: Aaron Lu <aaron.lu@intel.com>
    Signed-off-by: Zhang Rui <rui.zhang@intel.com>

Because, with the commit applied, all the fan devices are attached
to the acpi_general_pm_domain, and they are turned on by the pm_domain
automatically after resume, without the awareness of thermal core.

CC: <stable@vger.kernel.org> #3.18+
Reference: https://bugzilla.kernel.org/show_bug.cgi?id=78201
Reference: https://bugzilla.kernel.org/show_bug.cgi?id=91411
Tested-by: Manuel Krause <manuelkrause@netscape.net>
Tested-by: szegad <szegadlo@poczta.onet.pl>
Tested-by: prash <prash.n.rao@gmail.com>
Tested-by: amish <ammdispose-arch@yahoo.com>
Tested-by: Matthias <morpheusxyz123@yahoo.de>
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
---
 drivers/thermal/thermal_core.c | 37 +++++++++++++++++++++++++++++++++++++
 1 file changed, 37 insertions(+)

Comments

Eduardo Valentin March 24, 2015, 3:06 p.m. UTC | #1
Hey Rui

On Tue, Mar 24, 2015 at 01:21:29PM +0800, Zhang Rui wrote:
> Current thermal code does not handle system sleep well because
> 1. the cooling device cooling state may be changed during suspend
> 2. the previous temperature reading becomes invalid after resumed because
>    it is got before system sleep
> 3. updating thermal zone device during suspending/resuming
>    is wrong because some devices may have already been suspended
>    or may have not been resumed.
> 
> Thus, the proper way to do this is to cancel all thermal zone
> device update requirements during suspend/resume, and after all
> the devices have been resumed, reset and update every registered
> thermal zone devices.
> 
> This also fixes a regression introduced by
> commit 19593a1fb1f6718406afca5b867dab184289d406
> Author: Aaron Lu <aaron.lu@intel.com>
> Date:   Tue Nov 19 16:59:20 2013 +0800
> 
>     ACPI / fan: convert to platform driver
> 
>     Convert ACPI fan driver to a platform driver for the purpose of phasing
>     out ACPI bus.
> 
>     Signed-off-by: Aaron Lu <aaron.lu@intel.com>
>     Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> 
> Because, with the commit applied, all the fan devices are attached
> to the acpi_general_pm_domain, and they are turned on by the pm_domain
> automatically after resume, without the awareness of thermal core.
> 
> CC: <stable@vger.kernel.org> #3.18+
> Reference: https://bugzilla.kernel.org/show_bug.cgi?id=78201
> Reference: https://bugzilla.kernel.org/show_bug.cgi?id=91411
> Tested-by: Manuel Krause <manuelkrause@netscape.net>
> Tested-by: szegad <szegadlo@poczta.onet.pl>
> Tested-by: prash <prash.n.rao@gmail.com>
> Tested-by: amish <ammdispose-arch@yahoo.com>
> Tested-by: Matthias <morpheusxyz123@yahoo.de>
> Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> ---
>  drivers/thermal/thermal_core.c | 37 +++++++++++++++++++++++++++++++++++++
>  1 file changed, 37 insertions(+)
> 
> diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
> index 9d6f71b..9c03561 100644
> --- a/drivers/thermal/thermal_core.c
> +++ b/drivers/thermal/thermal_core.c
> @@ -37,6 +37,7 @@
>  #include <linux/of.h>
>  #include <net/netlink.h>
>  #include <net/genetlink.h>
> +#include <linux/suspend.h>
>  
>  #define CREATE_TRACE_POINTS
>  #include <trace/events/thermal.h>
> @@ -59,6 +60,9 @@ static LIST_HEAD(thermal_governor_list);
>  static DEFINE_MUTEX(thermal_list_lock);
>  static DEFINE_MUTEX(thermal_governor_lock);
>  
> +static struct notifier_block thermal_pm_nb;
> +static bool no_thermal_update;

Should this variable be considered to be accessed using a lock?

> +
>  static struct thermal_governor *def_governor;
>  
>  static struct thermal_governor *__find_governor(const char *name)
> @@ -491,6 +495,9 @@ void thermal_zone_device_update(struct thermal_zone_device *tz)
>  {
>  	int count;
>  
> +	if (no_thermal_update)
> +		return;
> +
>  	if (!tz->ops->get_temp)
>  		return;
>  
> @@ -1823,6 +1830,33 @@ static void thermal_unregister_governors(void)
>  	thermal_gov_user_space_unregister();
>  }
>  
> +static int thermal_notify(struct notifier_block *nb,
> +				unsigned long mode, void *_unused)

I believe thermal_pm_notify sounds a better naming for this case.

> +{
> +	struct thermal_zone_device *tz;
> +
> +	switch (mode) {
> +	case PM_HIBERNATION_PREPARE:
> +	case PM_RESTORE_PREPARE:
> +	case PM_SUSPEND_PREPARE:
> +		no_thermal_update = true;
> +		break;
> +	case PM_POST_HIBERNATION:
> +	case PM_POST_RESTORE:
> +	case PM_POST_SUSPEND:
> +		no_thermal_update = false;
> +		list_for_each_entry(tz, &thermal_tz_list, node) {
> +			thermal_zone_device_reset(tz);
> +			thermal_zone_device_update(tz);
> +		}
> +		break;
> +	default:
> +		break;
> +	}
> +	return 0;
> +}
> +
> +
>  static int __init thermal_init(void)
>  {
>  	int result;
> @@ -1843,6 +1877,9 @@ static int __init thermal_init(void)
>  	if (result)
>  		goto exit_netlink;
>  
> +	thermal_pm_nb.notifier_call = thermal_notify;

I believe you can declare thermal_pm_nb already with the callback
initialized:



static struct notifier_block thermal_pm_nb = {
	.notifier_call = thermal_notify,
};


just put it after the thermal_notify function.

> +	register_pm_notifier(&thermal_pm_nb);
> +
>  	return 0;
>  
>  exit_netlink:
> -- 
> 1.9.1
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-pm" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
Javi Merino March 24, 2015, 4:39 p.m. UTC | #2
One minor nit

On Tue, Mar 24, 2015 at 05:21:29AM +0000, Zhang Rui wrote:
> Current thermal code does not handle system sleep well because
> 1. the cooling device cooling state may be changed during suspend
> 2. the previous temperature reading becomes invalid after resumed because
>    it is got before system sleep
> 3. updating thermal zone device during suspending/resuming
>    is wrong because some devices may have already been suspended
>    or may have not been resumed.
> 
> Thus, the proper way to do this is to cancel all thermal zone
> device update requirements during suspend/resume, and after all
> the devices have been resumed, reset and update every registered
> thermal zone devices.
> 
> This also fixes a regression introduced by
> commit 19593a1fb1f6718406afca5b867dab184289d406
> Author: Aaron Lu <aaron.lu@intel.com>
> Date:   Tue Nov 19 16:59:20 2013 +0800
> 
>     ACPI / fan: convert to platform driver
> 
>     Convert ACPI fan driver to a platform driver for the purpose of phasing
>     out ACPI bus.
> 
>     Signed-off-by: Aaron Lu <aaron.lu@intel.com>
>     Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> 
> Because, with the commit applied, all the fan devices are attached
> to the acpi_general_pm_domain, and they are turned on by the pm_domain
> automatically after resume, without the awareness of thermal core.
> 
> CC: <stable@vger.kernel.org> #3.18+
> Reference: https://bugzilla.kernel.org/show_bug.cgi?id=78201
> Reference: https://bugzilla.kernel.org/show_bug.cgi?id=91411
> Tested-by: Manuel Krause <manuelkrause@netscape.net>
> Tested-by: szegad <szegadlo@poczta.onet.pl>
> Tested-by: prash <prash.n.rao@gmail.com>
> Tested-by: amish <ammdispose-arch@yahoo.com>
> Tested-by: Matthias <morpheusxyz123@yahoo.de>
> Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> ---
>  drivers/thermal/thermal_core.c | 37 +++++++++++++++++++++++++++++++++++++
>  1 file changed, 37 insertions(+)
> 
> diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
> index 9d6f71b..9c03561 100644
> --- a/drivers/thermal/thermal_core.c
> +++ b/drivers/thermal/thermal_core.c
> @@ -37,6 +37,7 @@
>  #include <linux/of.h>
>  #include <net/netlink.h>
>  #include <net/genetlink.h>
> +#include <linux/suspend.h>
>  
>  #define CREATE_TRACE_POINTS
>  #include <trace/events/thermal.h>
> @@ -59,6 +60,9 @@ static LIST_HEAD(thermal_governor_list);
>  static DEFINE_MUTEX(thermal_list_lock);
>  static DEFINE_MUTEX(thermal_governor_lock);
>  
> +static struct notifier_block thermal_pm_nb;
> +static bool no_thermal_update;

Can this have a name without a negative?  It's a bit hard to read
the double-negative in "no_thermal_update = false".  Maybe
"in_suspend" is better?

Cheers,
Javi

> +
>  static struct thermal_governor *def_governor;
>  
>  static struct thermal_governor *__find_governor(const char *name)
> @@ -491,6 +495,9 @@ void thermal_zone_device_update(struct thermal_zone_device *tz)
>  {
>  	int count;
>  
> +	if (no_thermal_update)
> +		return;
> +
>  	if (!tz->ops->get_temp)
>  		return;
>  
> @@ -1823,6 +1830,33 @@ static void thermal_unregister_governors(void)
>  	thermal_gov_user_space_unregister();
>  }
>  
> +static int thermal_notify(struct notifier_block *nb,
> +				unsigned long mode, void *_unused)
> +{
> +	struct thermal_zone_device *tz;
> +
> +	switch (mode) {
> +	case PM_HIBERNATION_PREPARE:
> +	case PM_RESTORE_PREPARE:
> +	case PM_SUSPEND_PREPARE:
> +		no_thermal_update = true;
> +		break;
> +	case PM_POST_HIBERNATION:
> +	case PM_POST_RESTORE:
> +	case PM_POST_SUSPEND:
> +		no_thermal_update = false;
> +		list_for_each_entry(tz, &thermal_tz_list, node) {
> +			thermal_zone_device_reset(tz);
> +			thermal_zone_device_update(tz);
> +		}
> +		break;
> +	default:
> +		break;
> +	}
> +	return 0;
> +}
> +
> +
>  static int __init thermal_init(void)
>  {
>  	int result;

--
To unsubscribe from this list: send the line "unsubscribe linux-pm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Zhang Rui March 25, 2015, 2:25 a.m. UTC | #3
> -----Original Message-----
> From: linux-pm-owner@vger.kernel.org [mailto:linux-pm-
> owner@vger.kernel.org] On Behalf Of Eduardo Valentin
> Sent: Tuesday, March 24, 2015 11:07 PM
> To: Zhang, Rui
> Cc: linux-pm@vger.kernel.org; stable@vger.kernel.org
> Subject: Re: [PATCH 2/3] Thermal: handle thermal zone device properly during
> system sleep
> Importance: High
> 
> Hey Rui
> 
> On Tue, Mar 24, 2015 at 01:21:29PM +0800, Zhang Rui wrote:
> > Current thermal code does not handle system sleep well because 1. the
> > cooling device cooling state may be changed during suspend 2. the
> > previous temperature reading becomes invalid after resumed because
> >    it is got before system sleep
> > 3. updating thermal zone device during suspending/resuming
> >    is wrong because some devices may have already been suspended
> >    or may have not been resumed.
> >
> > Thus, the proper way to do this is to cancel all thermal zone device
> > update requirements during suspend/resume, and after all the devices
> > have been resumed, reset and update every registered thermal zone
> > devices.
> >
> > This also fixes a regression introduced by commit
> > 19593a1fb1f6718406afca5b867dab184289d406
> > Author: Aaron Lu <aaron.lu@intel.com>
> > Date:   Tue Nov 19 16:59:20 2013 +0800
> >
> >     ACPI / fan: convert to platform driver
> >
> >     Convert ACPI fan driver to a platform driver for the purpose of phasing
> >     out ACPI bus.
> >
> >     Signed-off-by: Aaron Lu <aaron.lu@intel.com>
> >     Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> >
> > Because, with the commit applied, all the fan devices are attached to
> > the acpi_general_pm_domain, and they are turned on by the pm_domain
> > automatically after resume, without the awareness of thermal core.
> >
> > CC: <stable@vger.kernel.org> #3.18+
> > Reference: https://bugzilla.kernel.org/show_bug.cgi?id=78201
> > Reference: https://bugzilla.kernel.org/show_bug.cgi?id=91411
> > Tested-by: Manuel Krause <manuelkrause@netscape.net>
> > Tested-by: szegad <szegadlo@poczta.onet.pl>
> > Tested-by: prash <prash.n.rao@gmail.com>
> > Tested-by: amish <ammdispose-arch@yahoo.com>
> > Tested-by: Matthias <morpheusxyz123@yahoo.de>
> > Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> > ---
> >  drivers/thermal/thermal_core.c | 37
> > +++++++++++++++++++++++++++++++++++++
> >  1 file changed, 37 insertions(+)
> >
> > diff --git a/drivers/thermal/thermal_core.c
> > b/drivers/thermal/thermal_core.c index 9d6f71b..9c03561 100644
> > --- a/drivers/thermal/thermal_core.c
> > +++ b/drivers/thermal/thermal_core.c
> > @@ -37,6 +37,7 @@
> >  #include <linux/of.h>
> >  #include <net/netlink.h>
> >  #include <net/genetlink.h>
> > +#include <linux/suspend.h>
> >
> >  #define CREATE_TRACE_POINTS
> >  #include <trace/events/thermal.h>
> > @@ -59,6 +60,9 @@ static LIST_HEAD(thermal_governor_list);  static
> > DEFINE_MUTEX(thermal_list_lock);  static
> > DEFINE_MUTEX(thermal_governor_lock);
> >
> > +static struct notifier_block thermal_pm_nb; static bool
> > +no_thermal_update;
> 
> Should this variable be considered to be accessed using a lock?
> 
Hmmm, why?
It is set once when entering suspend, and cleared once when resuming,
and this whole process is protected by the pm_mutex lock, right?

> > +
> >  static struct thermal_governor *def_governor;
> >
> >  static struct thermal_governor *__find_governor(const char *name) @@
> > -491,6 +495,9 @@ void thermal_zone_device_update(struct
> > thermal_zone_device *tz)  {
> >  	int count;
> >
> > +	if (no_thermal_update)
> > +		return;
> > +
> >  	if (!tz->ops->get_temp)
> >  		return;
> >
> > @@ -1823,6 +1830,33 @@ static void thermal_unregister_governors(void)
> >  	thermal_gov_user_space_unregister();
> >  }
> >
> > +static int thermal_notify(struct notifier_block *nb,
> > +				unsigned long mode, void *_unused)
> 
> I believe thermal_pm_notify sounds a better naming for this case.
> 
Okay, will change it to thermal_pm_notify in next version.

> > +{
> > +	struct thermal_zone_device *tz;
> > +
> > +	switch (mode) {
> > +	case PM_HIBERNATION_PREPARE:
> > +	case PM_RESTORE_PREPARE:
> > +	case PM_SUSPEND_PREPARE:
> > +		no_thermal_update = true;
> > +		break;
> > +	case PM_POST_HIBERNATION:
> > +	case PM_POST_RESTORE:
> > +	case PM_POST_SUSPEND:
> > +		no_thermal_update = false;
> > +		list_for_each_entry(tz, &thermal_tz_list, node) {
> > +			thermal_zone_device_reset(tz);
> > +			thermal_zone_device_update(tz);
> > +		}
> > +		break;
> > +	default:
> > +		break;
> > +	}
> > +	return 0;
> > +}
> > +
> > +
> >  static int __init thermal_init(void)
> >  {
> >  	int result;
> > @@ -1843,6 +1877,9 @@ static int __init thermal_init(void)
> >  	if (result)
> >  		goto exit_netlink;
> >
> > +	thermal_pm_nb.notifier_call = thermal_notify;
> 
> I believe you can declare thermal_pm_nb already with the callback
> initialized:
> 
> 
> 
> static struct notifier_block thermal_pm_nb = {
> 	.notifier_call = thermal_notify,
> };
> 
Yes, will do this.

Thanks,
rui
> 
> just put it after the thermal_notify function.
> 
> > +	register_pm_notifier(&thermal_pm_nb);
> > +
> >  	return 0;
> >
> >  exit_netlink:
> > --
> > 1.9.1
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-pm" in
> > the body of a message to majordomo@vger.kernel.org More majordomo info
> > at  http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-pm" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Zhang Rui March 25, 2015, 2:28 a.m. UTC | #4
> -----Original Message-----

> From: Javi Merino [mailto:javi.merino@arm.com]

> Sent: Wednesday, March 25, 2015 12:39 AM

> To: Zhang, Rui

> Cc: linux-pm@vger.kernel.org; stable@vger.kernel.org

> Subject: Re: [PATCH 2/3] Thermal: handle thermal zone device properly during

> system sleep

> Importance: High

> 

> One minor nit

> 

> On Tue, Mar 24, 2015 at 05:21:29AM +0000, Zhang Rui wrote:

> > Current thermal code does not handle system sleep well because 1. the

> > cooling device cooling state may be changed during suspend 2. the

> > previous temperature reading becomes invalid after resumed because

> >    it is got before system sleep

> > 3. updating thermal zone device during suspending/resuming

> >    is wrong because some devices may have already been suspended

> >    or may have not been resumed.

> >

> > Thus, the proper way to do this is to cancel all thermal zone device

> > update requirements during suspend/resume, and after all the devices

> > have been resumed, reset and update every registered thermal zone

> > devices.

> >

> > This also fixes a regression introduced by commit

> > 19593a1fb1f6718406afca5b867dab184289d406

> > Author: Aaron Lu <aaron.lu@intel.com>

> > Date:   Tue Nov 19 16:59:20 2013 +0800

> >

> >     ACPI / fan: convert to platform driver

> >

> >     Convert ACPI fan driver to a platform driver for the purpose of phasing

> >     out ACPI bus.

> >

> >     Signed-off-by: Aaron Lu <aaron.lu@intel.com>

> >     Signed-off-by: Zhang Rui <rui.zhang@intel.com>

> >

> > Because, with the commit applied, all the fan devices are attached to

> > the acpi_general_pm_domain, and they are turned on by the pm_domain

> > automatically after resume, without the awareness of thermal core.

> >

> > CC: <stable@vger.kernel.org> #3.18+

> > Reference: https://bugzilla.kernel.org/show_bug.cgi?id=78201

> > Reference: https://bugzilla.kernel.org/show_bug.cgi?id=91411

> > Tested-by: Manuel Krause <manuelkrause@netscape.net>

> > Tested-by: szegad <szegadlo@poczta.onet.pl>

> > Tested-by: prash <prash.n.rao@gmail.com>

> > Tested-by: amish <ammdispose-arch@yahoo.com>

> > Tested-by: Matthias <morpheusxyz123@yahoo.de>

> > Signed-off-by: Zhang Rui <rui.zhang@intel.com>

> > ---

> >  drivers/thermal/thermal_core.c | 37

> > +++++++++++++++++++++++++++++++++++++

> >  1 file changed, 37 insertions(+)

> >

> > diff --git a/drivers/thermal/thermal_core.c

> > b/drivers/thermal/thermal_core.c index 9d6f71b..9c03561 100644

> > --- a/drivers/thermal/thermal_core.c

> > +++ b/drivers/thermal/thermal_core.c

> > @@ -37,6 +37,7 @@

> >  #include <linux/of.h>

> >  #include <net/netlink.h>

> >  #include <net/genetlink.h>

> > +#include <linux/suspend.h>

> >

> >  #define CREATE_TRACE_POINTS

> >  #include <trace/events/thermal.h>

> > @@ -59,6 +60,9 @@ static LIST_HEAD(thermal_governor_list);  static

> > DEFINE_MUTEX(thermal_list_lock);  static

> > DEFINE_MUTEX(thermal_governor_lock);

> >

> > +static struct notifier_block thermal_pm_nb; static bool

> > +no_thermal_update;

> 

> Can this have a name without a negative?  It's a bit hard to read the double-

> negative in "no_thermal_update = false".  Maybe "in_suspend" is better?

> 

Sounds reasonable, will do it in next version.

Thanks,
Rui

> Cheers,

> Javi

> 

> > +

> >  static struct thermal_governor *def_governor;

> >

> >  static struct thermal_governor *__find_governor(const char *name) @@

> > -491,6 +495,9 @@ void thermal_zone_device_update(struct

> > thermal_zone_device *tz)  {

> >  	int count;

> >

> > +	if (no_thermal_update)

> > +		return;

> > +

> >  	if (!tz->ops->get_temp)

> >  		return;

> >

> > @@ -1823,6 +1830,33 @@ static void thermal_unregister_governors(void)

> >  	thermal_gov_user_space_unregister();

> >  }

> >

> > +static int thermal_notify(struct notifier_block *nb,

> > +				unsigned long mode, void *_unused) {

> > +	struct thermal_zone_device *tz;

> > +

> > +	switch (mode) {

> > +	case PM_HIBERNATION_PREPARE:

> > +	case PM_RESTORE_PREPARE:

> > +	case PM_SUSPEND_PREPARE:

> > +		no_thermal_update = true;

> > +		break;

> > +	case PM_POST_HIBERNATION:

> > +	case PM_POST_RESTORE:

> > +	case PM_POST_SUSPEND:

> > +		no_thermal_update = false;

> > +		list_for_each_entry(tz, &thermal_tz_list, node) {

> > +			thermal_zone_device_reset(tz);

> > +			thermal_zone_device_update(tz);

> > +		}

> > +		break;

> > +	default:

> > +		break;

> > +	}

> > +	return 0;

> > +}

> > +

> > +

> >  static int __init thermal_init(void)

> >  {

> >  	int result;
Eduardo Valentin March 25, 2015, 2:40 p.m. UTC | #5
On Wed, Mar 25, 2015 at 02:25:06AM +0000, Zhang, Rui wrote:
> 
> 
> > -----Original Message-----
> > From: linux-pm-owner@vger.kernel.org [mailto:linux-pm-
> > owner@vger.kernel.org] On Behalf Of Eduardo Valentin
> > Sent: Tuesday, March 24, 2015 11:07 PM
> > To: Zhang, Rui
> > Cc: linux-pm@vger.kernel.org; stable@vger.kernel.org
> > Subject: Re: [PATCH 2/3] Thermal: handle thermal zone device properly during
> > system sleep
> > Importance: High
> > 
> > Hey Rui
> > 
> > On Tue, Mar 24, 2015 at 01:21:29PM +0800, Zhang Rui wrote:
> > > Current thermal code does not handle system sleep well because 1. the
> > > cooling device cooling state may be changed during suspend 2. the
> > > previous temperature reading becomes invalid after resumed because
> > >    it is got before system sleep
> > > 3. updating thermal zone device during suspending/resuming
> > >    is wrong because some devices may have already been suspended
> > >    or may have not been resumed.
> > >
> > > Thus, the proper way to do this is to cancel all thermal zone device
> > > update requirements during suspend/resume, and after all the devices
> > > have been resumed, reset and update every registered thermal zone
> > > devices.
> > >
> > > This also fixes a regression introduced by commit
> > > 19593a1fb1f6718406afca5b867dab184289d406
> > > Author: Aaron Lu <aaron.lu@intel.com>
> > > Date:   Tue Nov 19 16:59:20 2013 +0800
> > >
> > >     ACPI / fan: convert to platform driver
> > >
> > >     Convert ACPI fan driver to a platform driver for the purpose of phasing
> > >     out ACPI bus.
> > >
> > >     Signed-off-by: Aaron Lu <aaron.lu@intel.com>
> > >     Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> > >
> > > Because, with the commit applied, all the fan devices are attached to
> > > the acpi_general_pm_domain, and they are turned on by the pm_domain
> > > automatically after resume, without the awareness of thermal core.
> > >
> > > CC: <stable@vger.kernel.org> #3.18+
> > > Reference: https://bugzilla.kernel.org/show_bug.cgi?id=78201
> > > Reference: https://bugzilla.kernel.org/show_bug.cgi?id=91411
> > > Tested-by: Manuel Krause <manuelkrause@netscape.net>
> > > Tested-by: szegad <szegadlo@poczta.onet.pl>
> > > Tested-by: prash <prash.n.rao@gmail.com>
> > > Tested-by: amish <ammdispose-arch@yahoo.com>
> > > Tested-by: Matthias <morpheusxyz123@yahoo.de>
> > > Signed-off-by: Zhang Rui <rui.zhang@intel.com>
> > > ---
> > >  drivers/thermal/thermal_core.c | 37
> > > +++++++++++++++++++++++++++++++++++++
> > >  1 file changed, 37 insertions(+)
> > >
> > > diff --git a/drivers/thermal/thermal_core.c
> > > b/drivers/thermal/thermal_core.c index 9d6f71b..9c03561 100644
> > > --- a/drivers/thermal/thermal_core.c
> > > +++ b/drivers/thermal/thermal_core.c
> > > @@ -37,6 +37,7 @@
> > >  #include <linux/of.h>
> > >  #include <net/netlink.h>
> > >  #include <net/genetlink.h>
> > > +#include <linux/suspend.h>
> > >
> > >  #define CREATE_TRACE_POINTS
> > >  #include <trace/events/thermal.h>
> > > @@ -59,6 +60,9 @@ static LIST_HEAD(thermal_governor_list);  static
> > > DEFINE_MUTEX(thermal_list_lock);  static
> > > DEFINE_MUTEX(thermal_governor_lock);
> > >
> > > +static struct notifier_block thermal_pm_nb; static bool
> > > +no_thermal_update;
> > 
> > Should this variable be considered to be accessed using a lock?
> > 
> Hmmm, why?

Because you access the variable out of the suspend path.

> It is set once when entering suspend, and cleared once when resuming,
> and this whole process is protected by the pm_mutex lock, right?
> 

yeah, if you would be accessing it only inside the suspend path, but you
have an extra reader...

> > > +
> > >  static struct thermal_governor *def_governor;
> > >
> > >  static struct thermal_governor *__find_governor(const char *name) @@
> > > -491,6 +495,9 @@ void thermal_zone_device_update(struct
> > > thermal_zone_device *tz)  {
> > >  	int count;
> > >
> > > +	if (no_thermal_update)
> > > +		return;
> > > +

.. right here.

> > >  	if (!tz->ops->get_temp)
> > >  		return;
> > >
> > > @@ -1823,6 +1830,33 @@ static void thermal_unregister_governors(void)
> > >  	thermal_gov_user_space_unregister();
> > >  }
> > >
> > > +static int thermal_notify(struct notifier_block *nb,
> > > +				unsigned long mode, void *_unused)
> > 
> > I believe thermal_pm_notify sounds a better naming for this case.
> > 
> Okay, will change it to thermal_pm_notify in next version.
> 
> > > +{
> > > +	struct thermal_zone_device *tz;
> > > +
> > > +	switch (mode) {
> > > +	case PM_HIBERNATION_PREPARE:
> > > +	case PM_RESTORE_PREPARE:
> > > +	case PM_SUSPEND_PREPARE:
> > > +		no_thermal_update = true;
> > > +		break;
> > > +	case PM_POST_HIBERNATION:
> > > +	case PM_POST_RESTORE:
> > > +	case PM_POST_SUSPEND:
> > > +		no_thermal_update = false;
> > > +		list_for_each_entry(tz, &thermal_tz_list, node) {
> > > +			thermal_zone_device_reset(tz);
> > > +			thermal_zone_device_update(tz);
> > > +		}
> > > +		break;
> > > +	default:
> > > +		break;
> > > +	}
> > > +	return 0;
> > > +}
> > > +
> > > +
> > >  static int __init thermal_init(void)
> > >  {
> > >  	int result;
> > > @@ -1843,6 +1877,9 @@ static int __init thermal_init(void)
> > >  	if (result)
> > >  		goto exit_netlink;
> > >
> > > +	thermal_pm_nb.notifier_call = thermal_notify;
> > 
> > I believe you can declare thermal_pm_nb already with the callback
> > initialized:
> > 
> > 
> > 
> > static struct notifier_block thermal_pm_nb = {
> > 	.notifier_call = thermal_notify,
> > };
> > 
> Yes, will do this.
> 
> Thanks,
> rui
> > 
> > just put it after the thermal_notify function.
> > 
> > > +	register_pm_notifier(&thermal_pm_nb);
> > > +
> > >  	return 0;
> > >
> > >  exit_netlink:
> > > --
> > > 1.9.1
> > >
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe linux-pm" in
> > > the body of a message to majordomo@vger.kernel.org More majordomo info
> > > at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
index 9d6f71b..9c03561 100644
--- a/drivers/thermal/thermal_core.c
+++ b/drivers/thermal/thermal_core.c
@@ -37,6 +37,7 @@ 
 #include <linux/of.h>
 #include <net/netlink.h>
 #include <net/genetlink.h>
+#include <linux/suspend.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/thermal.h>
@@ -59,6 +60,9 @@  static LIST_HEAD(thermal_governor_list);
 static DEFINE_MUTEX(thermal_list_lock);
 static DEFINE_MUTEX(thermal_governor_lock);
 
+static struct notifier_block thermal_pm_nb;
+static bool no_thermal_update;
+
 static struct thermal_governor *def_governor;
 
 static struct thermal_governor *__find_governor(const char *name)
@@ -491,6 +495,9 @@  void thermal_zone_device_update(struct thermal_zone_device *tz)
 {
 	int count;
 
+	if (no_thermal_update)
+		return;
+
 	if (!tz->ops->get_temp)
 		return;
 
@@ -1823,6 +1830,33 @@  static void thermal_unregister_governors(void)
 	thermal_gov_user_space_unregister();
 }
 
+static int thermal_notify(struct notifier_block *nb,
+				unsigned long mode, void *_unused)
+{
+	struct thermal_zone_device *tz;
+
+	switch (mode) {
+	case PM_HIBERNATION_PREPARE:
+	case PM_RESTORE_PREPARE:
+	case PM_SUSPEND_PREPARE:
+		no_thermal_update = true;
+		break;
+	case PM_POST_HIBERNATION:
+	case PM_POST_RESTORE:
+	case PM_POST_SUSPEND:
+		no_thermal_update = false;
+		list_for_each_entry(tz, &thermal_tz_list, node) {
+			thermal_zone_device_reset(tz);
+			thermal_zone_device_update(tz);
+		}
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+
 static int __init thermal_init(void)
 {
 	int result;
@@ -1843,6 +1877,9 @@  static int __init thermal_init(void)
 	if (result)
 		goto exit_netlink;
 
+	thermal_pm_nb.notifier_call = thermal_notify;
+	register_pm_notifier(&thermal_pm_nb);
+
 	return 0;
 
 exit_netlink: