kernel/linux.git/drivers/thermal/gov_power_allocator.c, branch v6.19.11

thermal: gov_power_allocator: Update total_weight on bind and cdev updates

2025-02-25T11:30:45+00:00

params->total_weight is not initialized during bind and not updated when the bound cdev changes. The cooling device weight will not be used due to the uninitialized total_weight, until an update via sysfs is triggered. The bound cdevs are updated during thermal zone registration, where each cooling device will be bound to the thermal zone one by one, but power_allocator_bind() can be called without an additional cdev update when manually changing the policy of a thermal zone via sysfs. Add a new function to handle weight update logic, including updating total_weight, and call it when bind, weight changes, and cdev updates to ensure total_weight is always correct. Fixes: a3cd6db4cc2e ("thermal: gov_power_allocator: Support new update callback of weights") Signed-off-by: Yu-Che Cheng Link: https://patch.msgid.link/20250222-fix-power-allocator-weight-v2-1-a94de86b685a@chromium.org [ rjw: Changelog edits ] Signed-off-by: Rafael J. Wysocki

thermal: gov_power_allocator: Fix incorrect calculation in divvy_up_power()

2025-02-21T17:13:10+00:00

divvy_up_power() should use weighted_req_power instead of req_power to calculate granted_power. Otherwise, granted_power may be unexpected as the denominator total_req_power is a weighted sum. This is a mistake made during the previous refactor. Replace req_power with weighted_req_power in divvy_up_power() calculation. Fixes: 912e97c67cc3 ("thermal: gov_power_allocator: Move memory allocation out of throttle()") Signed-off-by: Yu-Che Cheng Reviewed-by: Lukasz Luba Link: https://patch.msgid.link/20250219-fix-power-allocator-calc-v1-1-48b860291919@chromium.org [ rjw: Subject and changelog edits ] Signed-off-by: Rafael J. Wysocki

thermal: gov_power_allocator: Add missing NULL pointer check

2024-11-26T12:27:56+00:00

Commit 0dc23567c206 ("thermal: core: Move lists of thermal instances to trip descriptors") overlooked the case in which the Power Allocator governor attempts to bind to a tripless thermal zone and params->trip_max is NULL in check_power_actors(). No power actors can be found in that case, so check_power_actors() needs to be made return 0 then to restore its previous behavior. Fixes: 0dc23567c206 ("thermal: core: Move lists of thermal instances to trip descriptors") Closes: https://lore.kernel.org/linux-pm/Z0NeGF4ryCe_b5rr@sashalap/ Reported-by: Sasha Levin Signed-off-by: Rafael J. Wysocki Reviewed-by: Lukasz Luba Link: https://patch.msgid.link/2761105.mvXUDI8C0e@rjwysocki.net

thermal: gov_power_allocator: Granted power set to max when nobody request power

2024-10-24T15:24:03+00:00

When total_req_power is 0, divvy_up_power() will set granted_power to 0, and cdev will be limited to the lowest performance. If our polling delay is set to 200ms, it means that cdev cannot perform better within 200ms even if cdev has a sudden load. This will affect the performance of cdev and is not as expected. For this reason, if nobody requests power, then set the granted power to the max_power. Signed-off-by: ZhengShaobo Reviewed-by: Lukasz Luba Link: https://patch.msgid.link/20241021121138.422-1-zhengshaobo1@xiaomi.com [ rjw: Fixed up tags ] Signed-off-by: Rafael J. Wysocki

thermal: core: Add and use cooling device guard

2024-10-24T12:48:23+00:00

Add and use a special guard for cooling devices. This allows quite a few error code paths to be simplified among other things and brings in code size reduction for a good measure. No intentional functional impact. Signed-off-by: Rafael J. Wysocki Link: https://patch.msgid.link/5837621.DvuYhMxLoT@rjwysocki.net Reviewed-by: Lukasz Luba

thermal: core: Introduce thermal_cdev_update_nocheck()

2024-10-23T09:57:11+00:00

Three thermal governors call __thermal_cdev_update() under the cdev lock without doing any checks, so in order to reduce the related code duplication, introduce a new helper function called thermal_cdev_update_nocheck() for them and make them use it. No intentional functional impact. Signed-off-by: Rafael J. Wysocki Link: https://patch.msgid.link/1835097.VLH7GnMWUR@rjwysocki.net Reviewed-by: Lukasz Luba

thermal: core: Move lists of thermal instances to trip descriptors

2024-10-22T10:08:03+00:00

In almost all places where a thermal zone's list of thermal instances is walked, there is a check to match a specific trip point and it is walked in vain whenever there are no cooling devices associated with the given trip. To address this, store the lists of thermal instances in trip point descriptors instead of storing them in thermal zones and adjust all code using those lists accordingly. No intentional functional impact. Signed-off-by: Rafael J. Wysocki Link: https://patch.msgid.link/5522726.Sb9uPGUboI@rjwysocki.net Reviewed-by: Lukasz Luba

thermal: gov_power_allocator: Return early in manage if trip_max is NULL

2024-07-04T11:35:50+00:00

Commit da781936e7c3 ("thermal: gov_power_allocator: Allow binding without trip points") allowed the governor to bind even when trip_max is NULL. This allows a NULL pointer dereference to happen in the manage callback. Add an early return to prevent it, since the governor is expected to not do anything in this case. Fixes: da781936e7c3 ("thermal: gov_power_allocator: Allow binding without trip points") Signed-off-by: Nícolas F. R. A. Prado Link: https://patch.msgid.link/20240702-power-allocator-null-trip-max-v1-1-47a60dc55414@collabora.com Cc: All applicable Signed-off-by: Rafael J. Wysocki

thermal: core: Move passive polling management to the core

2024-04-30T19:16:13+00:00

Passive polling is enabled by setting the 'passive' field in struct thermal_zone_device to a positive value so long as the 'passive_delay_jiffies' field is greater than zero. It causes the thermal core to actively check the thermal zone temperature periodically which in theory should be done after crossing a passive trip point on the way up in order to allow governors to react more rapidly to temperature changes and adjust mitigation more precisely. However, the 'passive' field in struct thermal_zone_device is currently managed by governors which is quite problematic. First of all, only two governors, Step-Wise and Power Allocator, update that field at all, so the other governors do not benefit from passive polling, although in principle they should. Moreover, if the zone governor is changed from, say, Step-Wise to Fair-Share after 'passive' has been incremented by the former, it is not going to be reset back to zero by the latter even if the zone temperature falls down below all passive trip points. For this reason, make handle_thermal_trip() increment 'passive' to enable passive polling for the given thermal zone whenever a passive trip point is crossed on the way up and decrement it whenever a passive trip point is crossed on the way down. Also remove the 'passive' field updates from governors and additionally clear it in thermal_zone_device_init() to prevent passive polling from being enabled after a system resume just beacuse it was enabled before suspending the system. Signed-off-by: Rafael J. Wysocki Reviewed-by: Lukasz Luba Tested-by: Lukasz Luba

thermal: gov_power_allocator: Eliminate a redundant variable

2024-04-23T18:39:50+00:00

Notice that the passive field in struct thermal_zone_device is not used by the Power Allocator governor itself and so the ordering of its updates with respect to allow_maximum_power() or allocate_power() does not matter. Accordingly, make power_allocator_manage() update that field right before returning, which allows the current value of it to be passed directly to allow_maximum_power() without using the additional update variable that can be dropped. Signed-off-by: Rafael J. Wysocki Reviewed-by: Lukasz Luba