| Message ID | 8e812065f4a76325097c5f9c17f3386736d8c1d4.1574315190.git.amit.kucheria@linaro.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | drivers: thermal: step_wise: add support for hysteresis | expand |
On 11/21/2019 12:50 AM, Amit Kucheria wrote: > From: Ram Chandrasekar <rkumbako@codeaurora.org> > > Currently, step wise governor increases the mitigation when the > temperature goes above a threshold and decreases the mitigation when the > temperature goes below the threshold. If there is a case where the > temperature is wavering around the threshold, the mitigation will be > applied and removed every iteration, which is not very efficient. > > The use of hysteresis temperature could avoid this ping-pong of > mitigation by relaxing the mitigation to happen only when the > temperature goes below this lower hysteresis value. Hi Amit, Can this not lead to ping-pong around the hysteresis temperature? If the idea is to minimize ping-pong isn't average a better method? Warm Regards Thara > > Signed-off-by: Ram Chandrasekar <rkumbako@codeaurora.org> > Signed-off-by: Lina Iyer <ilina@codeaurora.org> > [Rebased patch from downstream] > Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org> > --- > drivers/thermal/step_wise.c | 35 ++++++++++++++++++++++++----------- > 1 file changed, 24 insertions(+), 11 deletions(-) > > diff --git a/drivers/thermal/step_wise.c b/drivers/thermal/step_wise.c > index 6e051cbd824ff..2c8a34a7cf959 100644 > --- a/drivers/thermal/step_wise.c > +++ b/drivers/thermal/step_wise.c > @@ -24,7 +24,7 @@ > * for this trip point > * d. if the trend is THERMAL_TREND_DROP_FULL, use lower limit > * for this trip point > - * If the temperature is lower than a trip point, > + * If the temperature is lower than a hysteresis temperature, > * 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 > @@ -115,30 +115,31 @@ static void update_passive_instance(struct thermal_zone_device *tz, > > static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) > { > - int trip_temp; > + int trip_temp, hyst_temp; > enum thermal_trip_type trip_type; > enum thermal_trend trend; > struct thermal_instance *instance; > - bool throttle = false; > + bool throttle; > int old_target; > > if (trip == THERMAL_TRIPS_NONE) { > - trip_temp = tz->forced_passive; > + hyst_temp = trip_temp = tz->forced_passive; > trip_type = THERMAL_TRIPS_NONE; > } else { > tz->ops->get_trip_temp(tz, trip, &trip_temp); > + hyst_temp = trip_temp; > + if (tz->ops->get_trip_hyst) { > + tz->ops->get_trip_hyst(tz, trip, &hyst_temp); > + hyst_temp = trip_temp - hyst_temp; > + } > tz->ops->get_trip_type(tz, trip, &trip_type); > } > > trend = get_tz_trend(tz, trip); > > - if (tz->temperature >= trip_temp) { > - throttle = true; > - trace_thermal_zone_trip(tz, trip, trip_type); > - } > - > - dev_dbg(&tz->device, "Trip%d[type=%d,temp=%d]:trend=%d,throttle=%d\n", > - trip, trip_type, trip_temp, trend, throttle); > + dev_dbg(&tz->device, > + "Trip%d[type=%d,temp=%d,hyst=%d]:trend=%d,throttle=%d\n", > + trip, trip_type, trip_temp, hyst_temp, trend, throttle); > > mutex_lock(&tz->lock); > > @@ -147,6 +148,18 @@ static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) > continue; > > old_target = instance->target; > + throttle = false; > + /* > + * Lower the mitigation only if the temperature > + * goes below the hysteresis temperature. > + */ > + if (tz->temperature >= trip_temp || > + (tz->temperature >= hyst_temp && > + old_target != THERMAL_NO_TARGET)) { > + throttle = true; > + trace_thermal_zone_trip(tz, trip, trip_type); > + } > + > instance->target = get_target_state(instance, trend, throttle); > dev_dbg(&instance->cdev->device, "old_target=%d, target=%d\n", > old_target, (int)instance->target); >
On Thu, Nov 21, 2019 at 7:40 PM Thara Gopinath <thara.gopinath@linaro.org> wrote: > > On 11/21/2019 12:50 AM, Amit Kucheria wrote: > > From: Ram Chandrasekar <rkumbako@codeaurora.org> > > > > Currently, step wise governor increases the mitigation when the > > temperature goes above a threshold and decreases the mitigation when the > > temperature goes below the threshold. If there is a case where the > > temperature is wavering around the threshold, the mitigation will be > > applied and removed every iteration, which is not very efficient. > > > > The use of hysteresis temperature could avoid this ping-pong of > > mitigation by relaxing the mitigation to happen only when the > > temperature goes below this lower hysteresis value. > Hi Amit, > > Can this not lead to ping-pong around the hysteresis temperature? That isn't how hysteresis is supposed to work if there is a sufficient delta between your trip point and your hysteresis value. e.g. if you have a trip at 80C and a hysteresis of 10C, it means that you will start throttling at 80C, but you won't STOP throttling until you cool down to 70C. At that point, you will wait again to get to 80C before throttling again. IOW, on the downward slope (80 -> 70) you still have throttling active and on the upward slope (70 -> 80), you have no[1] throttling, so different behaviour is the same temperature range depending on direction. If your hysteresis value was very low .e.g. 1C, it would certainly be useless. > If the idea is to minimize ping-pong isn't average a better method? We shouldn't ping-pong with the asymmetric behaviour described above. Regards, Amit [1] This is a simple example with a single trip point.
Hi Amit, On 11/21/2019 12:50 AM, Amit Kucheria wrote: > From: Ram Chandrasekar <rkumbako@codeaurora.org> > > Currently, step wise governor increases the mitigation when the > temperature goes above a threshold and decreases the mitigation when the > temperature goes below the threshold. If there is a case where the > temperature is wavering around the threshold, the mitigation will be > applied and removed every iteration, which is not very efficient. > > The use of hysteresis temperature could avoid this ping-pong of > mitigation by relaxing the mitigation to happen only when the > temperature goes below this lower hysteresis value. > > Signed-off-by: Ram Chandrasekar <rkumbako@codeaurora.org> > Signed-off-by: Lina Iyer <ilina@codeaurora.org> > [Rebased patch from downstream] > Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org> > --- > drivers/thermal/step_wise.c | 35 ++++++++++++++++++++++++----------- > 1 file changed, 24 insertions(+), 11 deletions(-) > > diff --git a/drivers/thermal/step_wise.c b/drivers/thermal/step_wise.c > index 6e051cbd824ff..2c8a34a7cf959 100644 > --- a/drivers/thermal/step_wise.c > +++ b/drivers/thermal/step_wise.c > @@ -24,7 +24,7 @@ > * for this trip point > * d. if the trend is THERMAL_TREND_DROP_FULL, use lower limit > * for this trip point > - * If the temperature is lower than a trip point, > + * If the temperature is lower than a hysteresis temperature, > * 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 > @@ -115,30 +115,31 @@ static void update_passive_instance(struct thermal_zone_device *tz, > > static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) > { > - int trip_temp; > + int trip_temp, hyst_temp; > enum thermal_trip_type trip_type; > enum thermal_trend trend; > struct thermal_instance *instance; > - bool throttle = false; > + bool throttle; There is no need to remove throttle = false here. You are setting it to false later down. > int old_target; > > if (trip == THERMAL_TRIPS_NONE) { > - trip_temp = tz->forced_passive; > + hyst_temp = trip_temp = tz->forced_passive; > trip_type = THERMAL_TRIPS_NONE; > } else { > tz->ops->get_trip_temp(tz, trip, &trip_temp); > + hyst_temp = trip_temp; > + if (tz->ops->get_trip_hyst) { > + tz->ops->get_trip_hyst(tz, trip, &hyst_temp); > + hyst_temp = trip_temp - hyst_temp; > + } > tz->ops->get_trip_type(tz, trip, &trip_type); > } > > trend = get_tz_trend(tz, trip); > > - if (tz->temperature >= trip_temp) { > - throttle = true; > - trace_thermal_zone_trip(tz, trip, trip_type); > - } > - > - dev_dbg(&tz->device, "Trip%d[type=%d,temp=%d]:trend=%d,throttle=%d\n", > - trip, trip_type, trip_temp, trend, throttle); > + dev_dbg(&tz->device, > + "Trip%d[type=%d,temp=%d,hyst=%d]:trend=%d,throttle=%d\n", > + trip, trip_type, trip_temp, hyst_temp, trend, throttle); throttle value is not final here. Why is the debug print and the setting of throttle reversed ? Idea is to print the final value of throttle. > > mutex_lock(&tz->lock); > > @@ -147,6 +148,18 @@ static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) > continue; > > old_target = instance->target; > + throttle = false; > + /* > + * Lower the mitigation only if the temperature > + * goes below the hysteresis temperature. > + */ I think this requires more comment here on why there is a check for old_target != THERMAL_NO_TARGET. Basically to ensure that the hysteresis is considered only when temperature is dropping. > + if (tz->temperature >= trip_temp || > + (tz->temperature >= hyst_temp && > + old_target != THERMAL_NO_TARGET)) { > + throttle = true; > + trace_thermal_zone_trip(tz, trip, trip_type); > + } > + > instance->target = get_target_state(instance, trend, throttle); > dev_dbg(&instance->cdev->device, "old_target=%d, target=%d\n", > old_target, (int)instance->target); >
On 11/21/2019 09:38 AM, Amit Kucheria wrote: > On Thu, Nov 21, 2019 at 7:40 PM Thara Gopinath > <thara.gopinath@linaro.org> wrote: >> >> On 11/21/2019 12:50 AM, Amit Kucheria wrote: >>> From: Ram Chandrasekar <rkumbako@codeaurora.org> >>> >>> Currently, step wise governor increases the mitigation when the >>> temperature goes above a threshold and decreases the mitigation when the >>> temperature goes below the threshold. If there is a case where the >>> temperature is wavering around the threshold, the mitigation will be >>> applied and removed every iteration, which is not very efficient. >>> >>> The use of hysteresis temperature could avoid this ping-pong of >>> mitigation by relaxing the mitigation to happen only when the >>> temperature goes below this lower hysteresis value. >> Hi Amit, >> >> Can this not lead to ping-pong around the hysteresis temperature? > > That isn't how hysteresis is supposed to work if there is a sufficient > delta between your trip point and your hysteresis value. > > e.g. if you have a trip at 80C and a hysteresis of 10C, it means that > you will start throttling at 80C, but you won't STOP throttling until > you cool down to 70C. At that point, you will wait again to get to 80C > before throttling again. > IOW, on the downward slope (80 -> 70) you still have throttling active > and on the upward slope (70 -> 80), you have no[1] throttling, so > different behaviour is the same temperature range depending on > direction. > > If your hysteresis value was very low .e.g. 1C, it would certainly be useless. Thanks for the explanation. I think averaging can still give a smoother experience/transition. But then it has to be implemented and tested against this solution. Other reason for this solution is hysteresis can be a higher value if needed where as averaging might not give that flexibility. I have some other comments on the patch which I have posted separately. > >> If the idea is to minimize ping-pong isn't average a better method? > > We shouldn't ping-pong with the asymmetric behaviour described above. > > Regards, > Amit > [1] This is a simple example with a single trip point. >
On Thu, Nov 21, 2019 at 11:21 AM Amit Kucheria <amit.kucheria@linaro.org> wrote: > > From: Ram Chandrasekar <rkumbako@codeaurora.org> > > Currently, step wise governor increases the mitigation when the > temperature goes above a threshold and decreases the mitigation when the > temperature goes below the threshold. If there is a case where the > temperature is wavering around the threshold, the mitigation will be > applied and removed every iteration, which is not very efficient. > > The use of hysteresis temperature could avoid this ping-pong of > mitigation by relaxing the mitigation to happen only when the > temperature goes below this lower hysteresis value. > > Signed-off-by: Ram Chandrasekar <rkumbako@codeaurora.org> > Signed-off-by: Lina Iyer <ilina@codeaurora.org> > [Rebased patch from downstream] > Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org> Daniel, Rui: ping. Keerthy: This works for you, right? > --- > drivers/thermal/step_wise.c | 35 ++++++++++++++++++++++++----------- > 1 file changed, 24 insertions(+), 11 deletions(-) > > diff --git a/drivers/thermal/step_wise.c b/drivers/thermal/step_wise.c > index 6e051cbd824ff..2c8a34a7cf959 100644 > --- a/drivers/thermal/step_wise.c > +++ b/drivers/thermal/step_wise.c > @@ -24,7 +24,7 @@ > * for this trip point > * d. if the trend is THERMAL_TREND_DROP_FULL, use lower limit > * for this trip point > - * If the temperature is lower than a trip point, > + * If the temperature is lower than a hysteresis temperature, > * 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 > @@ -115,30 +115,31 @@ static void update_passive_instance(struct thermal_zone_device *tz, > > static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) > { > - int trip_temp; > + int trip_temp, hyst_temp; > enum thermal_trip_type trip_type; > enum thermal_trend trend; > struct thermal_instance *instance; > - bool throttle = false; > + bool throttle; > int old_target; > > if (trip == THERMAL_TRIPS_NONE) { > - trip_temp = tz->forced_passive; > + hyst_temp = trip_temp = tz->forced_passive; > trip_type = THERMAL_TRIPS_NONE; > } else { > tz->ops->get_trip_temp(tz, trip, &trip_temp); > + hyst_temp = trip_temp; > + if (tz->ops->get_trip_hyst) { > + tz->ops->get_trip_hyst(tz, trip, &hyst_temp); > + hyst_temp = trip_temp - hyst_temp; > + } > tz->ops->get_trip_type(tz, trip, &trip_type); > } > > trend = get_tz_trend(tz, trip); > > - if (tz->temperature >= trip_temp) { > - throttle = true; > - trace_thermal_zone_trip(tz, trip, trip_type); > - } > - > - dev_dbg(&tz->device, "Trip%d[type=%d,temp=%d]:trend=%d,throttle=%d\n", > - trip, trip_type, trip_temp, trend, throttle); > + dev_dbg(&tz->device, > + "Trip%d[type=%d,temp=%d,hyst=%d]:trend=%d,throttle=%d\n", > + trip, trip_type, trip_temp, hyst_temp, trend, throttle); > > mutex_lock(&tz->lock); > > @@ -147,6 +148,18 @@ static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) > continue; > > old_target = instance->target; > + throttle = false; > + /* > + * Lower the mitigation only if the temperature > + * goes below the hysteresis temperature. > + */ > + if (tz->temperature >= trip_temp || > + (tz->temperature >= hyst_temp && > + old_target != THERMAL_NO_TARGET)) { > + throttle = true; > + trace_thermal_zone_trip(tz, trip, trip_type); > + } > + > instance->target = get_target_state(instance, trend, throttle); > dev_dbg(&instance->cdev->device, "old_target=%d, target=%d\n", > old_target, (int)instance->target); > -- > 2.17.1 >
On 21/11/2019 06:50, Amit Kucheria wrote: > From: Ram Chandrasekar <rkumbako@codeaurora.org> > > Currently, step wise governor increases the mitigation when the > temperature goes above a threshold and decreases the mitigation when the > temperature goes below the threshold. > > If there is a case where the > temperature is wavering around the threshold, the mitigation will be > applied and removed every iteration, which is not very efficient. > > The use of hysteresis temperature could avoid this ping-pong of > mitigation by relaxing the mitigation to happen only when the > temperature goes below this lower hysteresis value. What I'm worried about is how the hysteresis is used in the current code, where the destination of this data is to set the value in the sensor hardware if it is supported. Using the hysteresis in the governor seems like abusing the initial purpose of this information. Moreover, the hysteresis creates a gray area where the above algorithm (DROPPING && !throttle) => state-- or (RAISING && throttle) => state++ may drop the performances because we will continue mitigating even below the threshold. As the governor is an open-loop controller, I'm not sure if we can do something except adding some kind of low pass filter to prevent mitigation bounces. > Signed-off-by: Ram Chandrasekar <rkumbako@codeaurora.org> > Signed-off-by: Lina Iyer <ilina@codeaurora.org> > [Rebased patch from downstream] > Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org> > --- > drivers/thermal/step_wise.c | 35 ++++++++++++++++++++++++----------- > 1 file changed, 24 insertions(+), 11 deletions(-) > > diff --git a/drivers/thermal/step_wise.c b/drivers/thermal/step_wise.c > index 6e051cbd824ff..2c8a34a7cf959 100644 > --- a/drivers/thermal/step_wise.c > +++ b/drivers/thermal/step_wise.c > @@ -24,7 +24,7 @@ > * for this trip point > * d. if the trend is THERMAL_TREND_DROP_FULL, use lower limit > * for this trip point > - * If the temperature is lower than a trip point, > + * If the temperature is lower than a hysteresis temperature, > * 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 > @@ -115,30 +115,31 @@ static void update_passive_instance(struct thermal_zone_device *tz, > > static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) > { > - int trip_temp; > + int trip_temp, hyst_temp; > enum thermal_trip_type trip_type; > enum thermal_trend trend; > struct thermal_instance *instance; > - bool throttle = false; > + bool throttle; > int old_target; > > if (trip == THERMAL_TRIPS_NONE) { > - trip_temp = tz->forced_passive; > + hyst_temp = trip_temp = tz->forced_passive; > trip_type = THERMAL_TRIPS_NONE; > } else { > tz->ops->get_trip_temp(tz, trip, &trip_temp); > + hyst_temp = trip_temp; > + if (tz->ops->get_trip_hyst) { > + tz->ops->get_trip_hyst(tz, trip, &hyst_temp); > + hyst_temp = trip_temp - hyst_temp; > + } > tz->ops->get_trip_type(tz, trip, &trip_type); > } > > trend = get_tz_trend(tz, trip); > > - if (tz->temperature >= trip_temp) { > - throttle = true; > - trace_thermal_zone_trip(tz, trip, trip_type); > - } > - > - dev_dbg(&tz->device, "Trip%d[type=%d,temp=%d]:trend=%d,throttle=%d\n", > - trip, trip_type, trip_temp, trend, throttle); > + dev_dbg(&tz->device, > + "Trip%d[type=%d,temp=%d,hyst=%d]:trend=%d,throttle=%d\n", > + trip, trip_type, trip_temp, hyst_temp, trend, throttle); > > mutex_lock(&tz->lock); > > @@ -147,6 +148,18 @@ static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) > continue; > > old_target = instance->target; > + throttle = false; > + /* > + * Lower the mitigation only if the temperature > + * goes below the hysteresis temperature. > + */ > + if (tz->temperature >= trip_temp || > + (tz->temperature >= hyst_temp && > + old_target != THERMAL_NO_TARGET)) { > + throttle = true; > + trace_thermal_zone_trip(tz, trip, trip_type); > + } > + > instance->target = get_target_state(instance, trend, throttle); > dev_dbg(&instance->cdev->device, "old_target=%d, target=%d\n", > old_target, (int)instance->target); >
On 12/11/2019 6:35 AM, Daniel Lezcano wrote: > On 21/11/2019 06:50, Amit Kucheria wrote: >> From: Ram Chandrasekar <rkumbako@codeaurora.org> >> >> Currently, step wise governor increases the mitigation when the >> temperature goes above a threshold and decreases the mitigation when the >> temperature goes below the threshold. >> >> If there is a case where the >> temperature is wavering around the threshold, the mitigation will be >> applied and removed every iteration, which is not very efficient. >> >> The use of hysteresis temperature could avoid this ping-pong of >> mitigation by relaxing the mitigation to happen only when the >> temperature goes below this lower hysteresis value. > > What I'm worried about is how the hysteresis is used in the current > code, where the destination of this data is to set the value in the > sensor hardware if it is supported. > > Using the hysteresis in the governor seems like abusing the initial > purpose of this information. > > Moreover, the hysteresis creates a gray area where the above algorithm > (DROPPING && !throttle) => state-- or (RAISING && throttle) => state++ > may drop the performances because we will continue mitigating even below > the threshold. > > As the governor is an open-loop controller, I'm not sure if we can do > something except adding some kind of low pass filter to prevent > mitigation bounces. > We have two different use cases for the step wise algorithm, and the hysteresis makes sense only in one. For example, say we are controlling CPU junction temperature at 95C. When using step wise, mitigation is applied iteratively and there is a possibility that temperature can shoot up before the algorithm can reach an optimal mitigation level to keep the temperature near threshold. In order to help this state machine, we use a second back stop rule in the same thermal zone at a higher temperature (say 105C) with a hysteresis of 10C to mitigate CPU to a fixed value, by specifying upper/lower limit to be the same. The idea is that the second rule will place a hard hammer to bring the temperature down close to 95C and then it will remove the mitigation. Once mitigation is removed, the junction temperature rule state machine will re-adjust from that point to an optimal mitigation level. The junction temperature rule doesn’t use hysteresis. Another example is skin temperature mitigation for mobile devices, where the step wise algorithm with hysteresis just reduces the operating max frequency to a fixed value, when the threshold is reached. And the junction temperature rule starts mitigating from this operating max. That is the reason we have not generalized or mandated the hysteresis usage in this patch. The idea is to use it selectively based on use case. >> Signed-off-by: Ram Chandrasekar <rkumbako@codeaurora.org> >> Signed-off-by: Lina Iyer <ilina@codeaurora.org> >> [Rebased patch from downstream] >> Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org> >> --- >> drivers/thermal/step_wise.c | 35 ++++++++++++++++++++++++----------- >> 1 file changed, 24 insertions(+), 11 deletions(-) >> >> diff --git a/drivers/thermal/step_wise.c b/drivers/thermal/step_wise.c >> index 6e051cbd824ff..2c8a34a7cf959 100644 >> --- a/drivers/thermal/step_wise.c >> +++ b/drivers/thermal/step_wise.c >> @@ -24,7 +24,7 @@ >> * for this trip point >> * d. if the trend is THERMAL_TREND_DROP_FULL, use lower limit >> * for this trip point >> - * If the temperature is lower than a trip point, >> + * If the temperature is lower than a hysteresis temperature, >> * 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 >> @@ -115,30 +115,31 @@ static void update_passive_instance(struct thermal_zone_device *tz, >> >> static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) >> { >> - int trip_temp; >> + int trip_temp, hyst_temp; >> enum thermal_trip_type trip_type; >> enum thermal_trend trend; >> struct thermal_instance *instance; >> - bool throttle = false; >> + bool throttle; >> int old_target; >> >> if (trip == THERMAL_TRIPS_NONE) { >> - trip_temp = tz->forced_passive; >> + hyst_temp = trip_temp = tz->forced_passive; >> trip_type = THERMAL_TRIPS_NONE; >> } else { >> tz->ops->get_trip_temp(tz, trip, &trip_temp); >> + hyst_temp = trip_temp; >> + if (tz->ops->get_trip_hyst) { >> + tz->ops->get_trip_hyst(tz, trip, &hyst_temp); >> + hyst_temp = trip_temp - hyst_temp; >> + } >> tz->ops->get_trip_type(tz, trip, &trip_type); >> } >> >> trend = get_tz_trend(tz, trip); >> >> - if (tz->temperature >= trip_temp) { >> - throttle = true; >> - trace_thermal_zone_trip(tz, trip, trip_type); >> - } >> - >> - dev_dbg(&tz->device, "Trip%d[type=%d,temp=%d]:trend=%d,throttle=%d\n", >> - trip, trip_type, trip_temp, trend, throttle); >> + dev_dbg(&tz->device, >> + "Trip%d[type=%d,temp=%d,hyst=%d]:trend=%d,throttle=%d\n", >> + trip, trip_type, trip_temp, hyst_temp, trend, throttle); >> >> mutex_lock(&tz->lock); >> >> @@ -147,6 +148,18 @@ static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) >> continue; >> >> old_target = instance->target; >> + throttle = false; >> + /* >> + * Lower the mitigation only if the temperature >> + * goes below the hysteresis temperature. >> + */ >> + if (tz->temperature >= trip_temp || >> + (tz->temperature >= hyst_temp && >> + old_target != THERMAL_NO_TARGET)) { >> + throttle = true; >> + trace_thermal_zone_trip(tz, trip, trip_type); >> + } >> + >> instance->target = get_target_state(instance, trend, throttle); >> dev_dbg(&instance->cdev->device, "old_target=%d, target=%d\n", >> old_target, (int)instance->target); >> > >
On 08/01/2020 01:31, Ram Chandrasekar wrote: > > > On 12/11/2019 6:35 AM, Daniel Lezcano wrote: >> On 21/11/2019 06:50, Amit Kucheria wrote: >>> From: Ram Chandrasekar <rkumbako@codeaurora.org> >>> >>> Currently, step wise governor increases the mitigation when the >>> temperature goes above a threshold and decreases the mitigation when the >>> temperature goes below the threshold. >>> >>> If there is a case where the >>> temperature is wavering around the threshold, the mitigation will be >>> applied and removed every iteration, which is not very efficient. >>> >>> The use of hysteresis temperature could avoid this ping-pong of >>> mitigation by relaxing the mitigation to happen only when the >>> temperature goes below this lower hysteresis value. >> >> What I'm worried about is how the hysteresis is used in the current >> code, where the destination of this data is to set the value in the >> sensor hardware if it is supported. >> >> Using the hysteresis in the governor seems like abusing the initial >> purpose of this information. >> >> Moreover, the hysteresis creates a gray area where the above algorithm >> (DROPPING && !throttle) => state-- or (RAISING && throttle) => state++ >> may drop the performances because we will continue mitigating even below >> the threshold. >> >> As the governor is an open-loop controller, I'm not sure if we can do >> something except adding some kind of low pass filter to prevent >> mitigation bounces. >> > > We have two different use cases for the step wise algorithm, and the > hysteresis makes sense only in one. > > For example, say we are controlling CPU junction temperature at 95C. > When using step wise, mitigation is applied iteratively and there is a > possibility that temperature can shoot up before the algorithm can reach > an optimal mitigation level to keep the temperature near threshold. > > In order to help this state machine, we use a second back stop rule in > the same thermal zone at a higher temperature (say 105C) with a > hysteresis of 10C to mitigate CPU to a fixed value, by specifying > upper/lower limit to be the same. The idea is that the second rule will > place a hard hammer to bring the temperature down close to 95C and then > it will remove the mitigation. Once mitigation is removed, the junction > temperature rule state machine will re-adjust from that point to an > optimal mitigation level. The junction temperature rule doesn’t use > hysteresis. > > Another example is skin temperature mitigation for mobile devices, where > the step wise algorithm with hysteresis just reduces the operating max > frequency to a fixed value, when the threshold is reached. And the > junction temperature rule starts mitigating from this operating max. > > That is the reason we have not generalized or mandated the hysteresis > usage in this patch. The idea is to use it selectively based on use case. Did you ever try the IPA governor?
On 1/9/2020 3:46 PM, Daniel Lezcano wrote: > On 08/01/2020 01:31, Ram Chandrasekar wrote: >> >> >> On 12/11/2019 6:35 AM, Daniel Lezcano wrote: >>> On 21/11/2019 06:50, Amit Kucheria wrote: >>>> From: Ram Chandrasekar <rkumbako@codeaurora.org> >>>> >>>> Currently, step wise governor increases the mitigation when the >>>> temperature goes above a threshold and decreases the mitigation when the >>>> temperature goes below the threshold. >>>> >>>> If there is a case where the >>>> temperature is wavering around the threshold, the mitigation will be >>>> applied and removed every iteration, which is not very efficient. >>>> >>>> The use of hysteresis temperature could avoid this ping-pong of >>>> mitigation by relaxing the mitigation to happen only when the >>>> temperature goes below this lower hysteresis value. >>> >>> What I'm worried about is how the hysteresis is used in the current >>> code, where the destination of this data is to set the value in the >>> sensor hardware if it is supported. >>> >>> Using the hysteresis in the governor seems like abusing the initial >>> purpose of this information. >>> >>> Moreover, the hysteresis creates a gray area where the above algorithm >>> (DROPPING && !throttle) => state-- or (RAISING && throttle) => state++ >>> may drop the performances because we will continue mitigating even below >>> the threshold. >>> >>> As the governor is an open-loop controller, I'm not sure if we can do >>> something except adding some kind of low pass filter to prevent >>> mitigation bounces. >>> >> >> We have two different use cases for the step wise algorithm, and the >> hysteresis makes sense only in one. >> >> For example, say we are controlling CPU junction temperature at 95C. >> When using step wise, mitigation is applied iteratively and there is a >> possibility that temperature can shoot up before the algorithm can reach >> an optimal mitigation level to keep the temperature near threshold. >> >> In order to help this state machine, we use a second back stop rule in >> the same thermal zone at a higher temperature (say 105C) with a >> hysteresis of 10C to mitigate CPU to a fixed value, by specifying >> upper/lower limit to be the same. The idea is that the second rule will >> place a hard hammer to bring the temperature down close to 95C and then >> it will remove the mitigation. Once mitigation is removed, the junction >> temperature rule state machine will re-adjust from that point to an >> optimal mitigation level. The junction temperature rule doesn’t use >> hysteresis. >> >> Another example is skin temperature mitigation for mobile devices, where >> the step wise algorithm with hysteresis just reduces the operating max >> frequency to a fixed value, when the threshold is reached. And the >> junction temperature rule starts mitigating from this operating max. >> >> That is the reason we have not generalized or mandated the hysteresis >> usage in this patch. The idea is to use it selectively based on use case. > > Did you ever try the IPA governor? > > Yes. We understand IPA could bring down the power/frequency cap multiple levels compared to step wise which is doing iteratively. With cooling device weights this makes IPA good for skin temperature control. But when we evaluated IPA, we found step wise is giving a better performance compared to IPA for localized junction temperature rules. Additionally, step wise for the junction temperature rule is easier to tune with fewer parameters. With these additional enhancements, step wise turned out to be a better choice for local junction temperature rules.
diff --git a/drivers/thermal/step_wise.c b/drivers/thermal/step_wise.c index 6e051cbd824ff..2c8a34a7cf959 100644 --- a/drivers/thermal/step_wise.c +++ b/drivers/thermal/step_wise.c @@ -24,7 +24,7 @@ * for this trip point * d. if the trend is THERMAL_TREND_DROP_FULL, use lower limit * for this trip point - * If the temperature is lower than a trip point, + * If the temperature is lower than a hysteresis temperature, * 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 @@ -115,30 +115,31 @@ static void update_passive_instance(struct thermal_zone_device *tz, static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) { - int trip_temp; + int trip_temp, hyst_temp; enum thermal_trip_type trip_type; enum thermal_trend trend; struct thermal_instance *instance; - bool throttle = false; + bool throttle; int old_target; if (trip == THERMAL_TRIPS_NONE) { - trip_temp = tz->forced_passive; + hyst_temp = trip_temp = tz->forced_passive; trip_type = THERMAL_TRIPS_NONE; } else { tz->ops->get_trip_temp(tz, trip, &trip_temp); + hyst_temp = trip_temp; + if (tz->ops->get_trip_hyst) { + tz->ops->get_trip_hyst(tz, trip, &hyst_temp); + hyst_temp = trip_temp - hyst_temp; + } tz->ops->get_trip_type(tz, trip, &trip_type); } trend = get_tz_trend(tz, trip); - if (tz->temperature >= trip_temp) { - throttle = true; - trace_thermal_zone_trip(tz, trip, trip_type); - } - - dev_dbg(&tz->device, "Trip%d[type=%d,temp=%d]:trend=%d,throttle=%d\n", - trip, trip_type, trip_temp, trend, throttle); + dev_dbg(&tz->device, + "Trip%d[type=%d,temp=%d,hyst=%d]:trend=%d,throttle=%d\n", + trip, trip_type, trip_temp, hyst_temp, trend, throttle); mutex_lock(&tz->lock); @@ -147,6 +148,18 @@ static void thermal_zone_trip_update(struct thermal_zone_device *tz, int trip) continue; old_target = instance->target; + throttle = false; + /* + * Lower the mitigation only if the temperature + * goes below the hysteresis temperature. + */ + if (tz->temperature >= trip_temp || + (tz->temperature >= hyst_temp && + old_target != THERMAL_NO_TARGET)) { + throttle = true; + trace_thermal_zone_trip(tz, trip, trip_type); + } + instance->target = get_target_state(instance, trend, throttle); dev_dbg(&instance->cdev->device, "old_target=%d, target=%d\n", old_target, (int)instance->target);