Re: [RESEND PATCH v4 1/5] thermal: Add support for hardware-tracked trip points
From: Javi Merino <hidden>
Date: 2016-06-02 17:30:21
Also in:
linux-rockchip, lkml
Hi Caesar, On Fri, May 27, 2016 at 04:36:44PM +0800, Caesar Wang wrote:
quoted hunk ↗ jump to hunk
From: Sascha Hauer <s.hauer@pengutronix.de> This adds support for hardware-tracked trip points to the device tree thermal sensor framework. The framework supports an arbitrary number of trip points. Whenever the current temperature is updated, the trip points immediately below and above the current temperature are found. A .set_trips callback is then called with the temperatures. If there is no trip point above or below the current temperature, the passed trip temperature will be -INT_MAX or INT_MAX respectively. In this callback, the driver should program the hardware such that it is notified when either of these trip points are triggered. When a trip point is triggered, the driver should call `thermal_zone_device_update' for the respective thermal zone. This will cause the trip points to be updated again. If .set_trips is not implemented, the framework behaves as before. This patch is based on an earlier version from Mikko Perttunen [off-list ref] Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de> Signed-off-by: Caesar Wang <redacted> Cc: Zhang Rui <rui.zhang@intel.com> Cc: Eduardo Valentin <edubezval@gmail.com> Cc: linux-pm@vger.kernel.org --- Changes in v4: - Missing the lock added in v3. Changes in v3: - as Javi comments on https://patchwork.kernel.org/patch/9001281/. - add the lock for preventing the called from multi placce - add the note for pre_low/high_trip. Changes in v2: - update the sysfs-api.txt for set_trips. Documentation/thermal/sysfs-api.txt | 7 +++++ drivers/thermal/thermal_core.c | 56 +++++++++++++++++++++++++++++++++++++ include/linux/thermal.h | 7 +++++ 3 files changed, 70 insertions(+)diff --git a/Documentation/thermal/sysfs-api.txt b/Documentation/thermal/sysfs-api.txt index efc3f3d..75d8838 100644 --- a/Documentation/thermal/sysfs-api.txt +++ b/Documentation/thermal/sysfs-api.txt@@ -49,6 +49,9 @@ temperature) and throttle appropriate devices. .bind: bind the thermal zone device with a thermal cooling device. .unbind: unbind the thermal zone device with a thermal cooling device. .get_temp: get the current temperature of the thermal zone. + .set_trips: set the trip points window. Whenever the current temperature + is updated, the trip points immediately below and above the + current temperature are found. .get_mode: get the current mode (enabled/disabled) of the thermal zone. - "enabled" means the kernel thermal management is enabled. - "disabled" will prevent kernel thermal driver action upon trip points@@ -95,6 +98,10 @@ temperature) and throttle appropriate devices. get_temp: a pointer to a function that reads the sensor temperature. This is mandatory callback provided by sensor driver. + set_trips: a pointer to a function that sets a + temperature window. When this window is + left the driver must inform the thermal + core via thermal_zone_device_update. get_trend: a pointer to a function that reads the sensor temperature trend. set_emul_temp: a pointer to a function that setsdiff --git a/drivers/thermal/thermal_core.c b/drivers/thermal/thermal_core.c index 5133cd1..0591438 100644 --- a/drivers/thermal/thermal_core.c +++ b/drivers/thermal/thermal_core.c@@ -520,6 +520,55 @@ exit: } EXPORT_SYMBOL_GPL(thermal_zone_get_temp); +static void thermal_zone_set_trips(struct thermal_zone_device *tz) +{ + int low = -INT_MAX; + int high = INT_MAX; + int trip_temp, hysteresis; + int temp = tz->temperature; + int i, ret; + + if (!tz->ops->set_trips) + return; + + for (i = 0; i < tz->trips; i++) { + int trip_low; + + tz->ops->get_trip_temp(tz, i, &trip_temp); + tz->ops->get_trip_hyst(tz, i, &hysteresis); + + trip_low = trip_temp - hysteresis;
You only use the hysteresis for the low trip point and not for the high trip point. Shouldn't you also add: trip_high = trip_temp + hysteresis; ...
+ + if (trip_low < temp && trip_low > low) + low = trip_low; + + if (trip_temp > temp && trip_temp < high) + high = trip_temp;
... and here: if (trip_high > temp && trip_high < high) high = trip_high;
+ } + + /* No need to change trip points */ + if (tz->prev_low_trip == low && tz->prev_high_trip == high) + return; + + mutex_lock(&tz->lock);
I think you should get the lock for the whole function, including the read of tz->temperature . I know I said that the lock should be acquired before the if condition before but thinking about it, I think there may be a race condition if userspace updates the hysteresis while thermal_zone_device_update() is updating the temperature: you could have the cpu that is updating the hysteresis seeing the old temperature while the cpu that is updating the temperature would see the old hysteresis. I think it's safer if this whole function is protected from the lock, from the read of tz->temperature until the end. Cheers, Javi
quoted hunk ↗ jump to hunk
+ + tz->prev_low_trip = low; + tz->prev_high_trip = high; + + dev_dbg(&tz->device, "new temperature boundaries: %d < x < %d\n", + low, high); + + /* + * Set a temperature window. When this window is left the driver + * must inform the thermal core via thermal_zone_device_update. + */ + ret = tz->ops->set_trips(tz, low, high); + if (ret) + dev_err(&tz->device, "Failed to set trips: %d\n", ret); + + mutex_unlock(&tz->lock); +} + static void update_temperature(struct thermal_zone_device *tz) { int temp, ret;@@ -569,6 +618,8 @@ void thermal_zone_device_update(struct thermal_zone_device *tz) update_temperature(tz); + thermal_zone_set_trips(tz); + for (count = 0; count < tz->trips; count++) handle_thermal_trip(tz, count); }@@ -754,6 +805,9 @@ trip_point_hyst_store(struct device *dev, struct device_attribute *attr, */ ret = tz->ops->set_trip_hyst(tz, trip, temperature); + if (!ret) + thermal_zone_set_trips(tz); + return ret ? ret : count; }