linux/drivers/thermal/gov_step_wise.c
Rafael J. Wysocki b684682698 thermal: gov_step_wise: Restore passive polling management
Consider a thermal zone with one passive trip point, a cooling device
with 3 states (0, 1, 2) bound to it, passive polling enabled (nonzero
passive_delay_jiffies) and no regular polling (polling_delay_jiffies
equal to 0) that is managed by the Step-Wise governor.  Suppose that
the initial state of the cooling device is 0 and the zone temperature
is below the trip point to start with.

When the trip point is crossed, tz->passive is incremented by the
thermal core and the governor's .manage() callback is invoked.  It
sets 'throttle' to 'true' for the trip in question and
get_target_state() returns 1 for the instance corresponding to the
cooling device (say that 'upper' and 'lower' are set to 2 and 0 for
it, respectively), so its state changes to 1.

Passive polling is still active for the zone, so next time the
temperature is updated, the governor's .manage() callback will be
invoked again.  If the temperature is still rising, it will change
the state of the cooling device to 2.

Now suppose that next time the zone temperature is updated, it falls
below the trip point, so tz->passive is decremented for the zone (say
it becomes 0 then) and the governor's .manage() callbacks runs.

It finds that the temperature trend for the zone is 'falling' and
'throttle' will be set to 'false' for the trip in question, so the
cooling device's state will be changed to 1.  However, because
tz->polling is 0 for the zone, the governor's .manage() callback
may not be invoked again for a long time and the cooling device's
state will not be reset back to 0.

This can happen because commit 042a3d80f1 ("thermal: core: Move
passive polling management to the core") removed passive polling
management from the Step-Wise governor.

Before that change, thermal_zone_trip_update() would bump up
tz->passive when changing the target state for a thermal instance
from "no target" to a specific value and it would drop tz->passive
when changing it back to "no target" which would cause passive
polling to be active for the zone until the governor has reset the
states of all cooling devices.  In particular, in the example above
tz->passive would be incremented when changing the state of the
cooling device from 0 to 1 and then it would be still nonzero when
the state of the cooling device was changed from 2 to 1.

To prevent this problem from occurring, restore the passive polling
management in the Step-Wise governor by partially reverting the
commit in question and update the comment in the restored code
to explain its role more clearly.

Fixes: 042a3d80f1 ("thermal: core: Move passive polling management to the core")
Closes: https://lore.kernel.org/linux-pm/ZmVfcEOxmjUHZTSX@hovoldconsulting.com
Reported-by: Johan Hovold <johan+linaro@kernel.org>
Tested-by: Johan Hovold <johan+linaro@kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2024-06-11 21:00:44 +02:00

155 lines
4.6 KiB
C

// SPDX-License-Identifier: GPL-2.0-only
/*
* step_wise.c - A step-by-step Thermal throttling governor
*
* Copyright (C) 2012 Intel Corp
* Copyright (C) 2012 Durgadoss R <durgadoss.r@intel.com>
*
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*/
#include <linux/thermal.h>
#include <linux/minmax.h>
#include "thermal_trace.h"
#include "thermal_core.h"
/*
* If the temperature is higher than a trip point,
* a. if the trend is THERMAL_TREND_RAISING, use higher cooling
* state for this trip point
* b. if the trend is THERMAL_TREND_DROPPING, do nothing
* If the temperature is lower than a trip point,
* a. if the trend is THERMAL_TREND_RAISING, do nothing
* b. if the trend is THERMAL_TREND_DROPPING, use lower cooling
* state for this trip point, if the cooling state already
* equals lower limit, deactivate the thermal instance
*/
static unsigned long get_target_state(struct thermal_instance *instance,
enum thermal_trend trend, bool throttle)
{
struct thermal_cooling_device *cdev = instance->cdev;
unsigned long cur_state;
/*
* We keep this instance the way it is by default.
* Otherwise, we use the current state of the
* cdev in use to determine the next_target.
*/
cdev->ops->get_cur_state(cdev, &cur_state);
dev_dbg(&cdev->device, "cur_state=%ld\n", cur_state);
if (!instance->initialized) {
if (throttle)
return clamp(cur_state + 1, instance->lower, instance->upper);
return THERMAL_NO_TARGET;
}
if (throttle) {
if (trend == THERMAL_TREND_RAISING)
return clamp(cur_state + 1, instance->lower, instance->upper);
} else if (trend == THERMAL_TREND_DROPPING) {
if (cur_state <= instance->lower)
return THERMAL_NO_TARGET;
return clamp(cur_state - 1, instance->lower, instance->upper);
}
return instance->target;
}
static void thermal_zone_trip_update(struct thermal_zone_device *tz,
const struct thermal_trip *trip,
int trip_threshold)
{
enum thermal_trend trend = get_tz_trend(tz, trip);
int trip_id = thermal_zone_trip_id(tz, trip);
struct thermal_instance *instance;
bool throttle = false;
if (tz->temperature >= trip_threshold) {
throttle = true;
trace_thermal_zone_trip(tz, trip_id, trip->type);
}
dev_dbg(&tz->device, "Trip%d[type=%d,temp=%d]:trend=%d,throttle=%d\n",
trip_id, trip->type, trip_threshold, trend, throttle);
list_for_each_entry(instance, &tz->thermal_instances, tz_node) {
int old_target;
if (instance->trip != trip)
continue;
old_target = instance->target;
instance->target = get_target_state(instance, trend, throttle);
dev_dbg(&instance->cdev->device, "old_target=%d, target=%ld\n",
old_target, instance->target);
if (instance->initialized && old_target == instance->target)
continue;
if (trip->type == THERMAL_TRIP_PASSIVE) {
/*
* If the target state for this thermal instance
* changes from THERMAL_NO_TARGET to something else,
* ensure that the zone temperature will be updated
* (assuming enabled passive cooling) until it becomes
* THERMAL_NO_TARGET again, or the cooling device may
* not be reset to its initial state.
*/
if (old_target == THERMAL_NO_TARGET &&
instance->target != THERMAL_NO_TARGET)
tz->passive++;
else if (old_target != THERMAL_NO_TARGET &&
instance->target == THERMAL_NO_TARGET)
tz->passive--;
}
instance->initialized = true;
mutex_lock(&instance->cdev->lock);
instance->cdev->updated = false; /* cdev needs update */
mutex_unlock(&instance->cdev->lock);
}
}
static void step_wise_manage(struct thermal_zone_device *tz)
{
const struct thermal_trip_desc *td;
struct thermal_instance *instance;
lockdep_assert_held(&tz->lock);
/*
* Throttling Logic: Use the trend of the thermal zone to throttle.
* If the thermal zone is 'heating up', throttle all of the cooling
* devices associated with each trip point by one step. If the zone
* is 'cooling down', it brings back the performance of the devices
* by one step.
*/
for_each_trip_desc(tz, td) {
const struct thermal_trip *trip = &td->trip;
if (trip->temperature == THERMAL_TEMP_INVALID ||
trip->type == THERMAL_TRIP_CRITICAL ||
trip->type == THERMAL_TRIP_HOT)
continue;
thermal_zone_trip_update(tz, trip, td->threshold);
}
list_for_each_entry(instance, &tz->thermal_instances, tz_node)
thermal_cdev_update(instance->cdev);
}
static struct thermal_governor thermal_gov_step_wise = {
.name = "step_wise",
.manage = step_wise_manage,
};
THERMAL_GOVERNOR_DECLARE(thermal_gov_step_wise);