[RFC PATCH for 6.13 v1 05/20] thermal: core: Fix race between zone registration and system suspend

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>

If the registration of a thermal zone takes place at the same time when
system suspend is started, thermal_pm_notify() can run before the new
thermal zone is added to thermal_tz_list and its "suspended" flag will
not be set.  Consequently, if __thermal_zone_device_update() is called
for that thermal zone, it will not return early as expected which may
cause some destructive interference with the system suspend or resume
flow to occur.

To avoid that, make thermal_zone_init_complete() introduced previously
set the "suspended" flag for new thermal zones if it runs during system
suspend or resume.

Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@xxxxxxxxx>
---
 drivers/thermal/thermal_core.c |   18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)

Index: linux-pm/drivers/thermal/thermal_core.c
===================================================================
--- linux-pm.orig/drivers/thermal/thermal_core.c
+++ linux-pm/drivers/thermal/thermal_core.c
@@ -39,6 +39,8 @@ static DEFINE_MUTEX(thermal_governor_loc
 
 static struct thermal_governor *def_governor;
 
+static bool thermal_pm_suspended;
+
 /*
  * Governor section: set of functions to handle thermal governors
  *
@@ -1323,6 +1325,14 @@ static void thermal_zone_init_complete(s
 	guard(thermal_zone)(tz);
 
 	tz->state &= ~TZ_STATE_FLAG_INIT;
+	/*
+	 * If system suspend or resume is in progress at this point, the
+	 * new thermal zone needs to be marked as suspended because
+	 * thermal_pm_notify() has run already.
+	 */
+	if (thermal_pm_suspended)
+		tz->state |= TZ_STATE_FLAG_SUSPENDED;
+
 	__thermal_zone_device_update(tz, THERMAL_EVENT_UNSPECIFIED);
 }
 
@@ -1494,10 +1504,10 @@ thermal_zone_device_register_with_trips(
 	list_for_each_entry(cdev, &thermal_cdev_list, node)
 		thermal_zone_cdev_bind(tz, cdev);
 
-	mutex_unlock(&thermal_list_lock);
-
 	thermal_zone_init_complete(tz);
 
+	mutex_unlock(&thermal_list_lock);
+
 	thermal_notify_tz_create(tz);
 
 	thermal_debug_tz_add(tz);
@@ -1718,6 +1728,8 @@ static int thermal_pm_notify(struct noti
 	case PM_SUSPEND_PREPARE:
 		mutex_lock(&thermal_list_lock);
 
+		thermal_pm_suspended = true;
+
 		list_for_each_entry(tz, &thermal_tz_list, node)
 			thermal_zone_pm_prepare(tz);
 
@@ -1728,6 +1740,8 @@ static int thermal_pm_notify(struct noti
 	case PM_POST_SUSPEND:
 		mutex_lock(&thermal_list_lock);
 
+		thermal_pm_suspended = false;
+
 		list_for_each_entry(tz, &thermal_tz_list, node)
 			thermal_zone_pm_complete(tz);