linux-kernel - [PATCH 2/5] thermal/core: Rework the monitoring a bit

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20220805153834.2510142-2-daniel.lezcano@linaro.org>
Date:   Fri,  5 Aug 2022 17:38:31 +0200
From:   Daniel Lezcano <daniel.lezcano@...aro.org>
To:     daniel.lezcano@...aro.org, rafael@...nel.org
Cc:     rui.zhang@...el.com, linux-pm@...r.kernel.org,
        linux-kernel@...r.kernel.org, Amit Kucheria <amitk@...nel.org>
Subject: [PATCH 2/5] thermal/core: Rework the monitoring a bit

The should_stop_polling() function wraps the function
thermal_zone_device_is_enabled().

The monitor_thermal_zone() function checks if the thermal zone is
enabled via the should_stop_polling() function.

However, the instant after checking the thermal zone is enabled, this
one can be disabled, so even if that reduces the race window, it does
not prevent that and the monitoring can be set again with the thermal
zone disabled.

For this reason, the function should_stop_polling() is replaced by a
direct check of the thermal zone mode with the mutex locks held, that
prevents the situation described above.

As the semantic is clear with the thermal_zone_is_enabled() function,
we can remove the should_stop_polling() function and replace the check
with the former function.

While at it, reorder the checks to improve the readability of the
monitor_thermal_zone() function.

In the future, the thermal_zone_device_disable() and the
thermal_zone_device_enable() functions should unset / set the polling
timer directly instead of relying on the next
thermal_zone_device_update() call to do that. That will make a
synchronous thermal zone mode change but the locking scheme should be
double checked for that which out of the scope of this change.

Signed-off-by: Daniel Lezcano <daniel.lezcano@...aro.org>
---
 drivers/thermal/thermal_core.c | 19 +++++--------------
 1 file changed, 5 insertions(+), 14 deletions(-)

diff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c
index 4e1a83987b99..d7029fd1c112 100644
--- a/drivers/thermal/thermal_core.c
+++ b/drivers/thermal/thermal_core.c
@@ -295,25 +295,16 @@ static void thermal_zone_device_set_polling(struct thermal_zone_device *tz,
 		cancel_delayed_work(&tz->poll_queue);
 }

-static inline bool should_stop_polling(struct thermal_zone_device *tz)
-{
-	return !thermal_zone_device_is_enabled(tz);
-}
-
 static void monitor_thermal_zone(struct thermal_zone_device *tz)
 {
-	bool stop;
-
-	stop = should_stop_polling(tz);
-
 	mutex_lock(&tz->lock);

-	if (!stop && tz->passive)
+	if (tz->mode != THERMAL_DEVICE_ENABLED)
+		thermal_zone_device_set_polling(tz, 0);
+	else if (tz->passive)
 		thermal_zone_device_set_polling(tz, tz->passive_delay_jiffies);
-	else if (!stop && tz->polling_delay_jiffies)
+	else if (tz->polling_delay_jiffies)
 		thermal_zone_device_set_polling(tz, tz->polling_delay_jiffies);
-	else
-		thermal_zone_device_set_polling(tz, 0);

 	mutex_unlock(&tz->lock);
 }
@@ -480,7 +471,7 @@ void thermal_zone_device_update(struct thermal_zone_device *tz,
 {
 	int count;

-	if (should_stop_polling(tz))
+	if (!thermal_zone_device_is_enabled(tz))
 		return;

 	if (atomic_read(&in_suspend))
-- 
2.25.1