The mitigation episodes are recorded. A mitigation episode happens
when the first trip point is crossed the way up and then the way
down. During this episode other trip points can be crossed also and
are accounted for this mitigation episode. The interesting information
is the average temperature at the trip point, the undershot and the
overshot. The standard deviation of the mitigated temperature will be
added later.
The thermal debugfs directory structure tries to stay consistent with
the sysfs one but in a very simplified way:
thermal/
`-- thermal_zones
|-- 0
| `-- mitigations
`-- 1
`-- mitigations
The content of the mitigations file has the following format:
,-Mitigation at 349988258us, duration=130136ms
| trip | type | temp(°mC) | hyst(°mC) | duration | avg(°mC) | min(°mC) | max(°mC) |
| 0 | passive | 65000 | 2000 | 130136 | 68227 | 62500 | 75625 |
| 1 | passive | 75000 | 2000 | 104209 | 74857 | 71666 | 77500 |
,-Mitigation at 272451637us, duration=75000ms
| trip | type | temp(°mC) | hyst(°mC) | duration | avg(°mC) | min(°mC) | max(°mC) |
| 0 | passive | 65000 | 2000 | 75000 | 68561 | 62500 | 75000 |
| 1 | passive | 75000 | 2000 | 60714 | 74820 | 70555 | 77500 |
,-Mitigation at 238184119us, duration=27316ms
| trip | type | temp(°mC) | hyst(°mC) | duration | avg(°mC) | min(°mC) | max(°mC) |
| 0 | passive | 65000 | 2000 | 27316 | 73377 | 62500 | 75000 |
| 1 | passive | 75000 | 2000 | 19468 | 75284 | 69444 | 77500 |
,-Mitigation at 39863713us, duration=136196ms
| trip | type | temp(°mC) | hyst(°mC) | duration | avg(°mC) | min(°mC) | max(°mC) |
| 0 | passive | 65000 | 2000 | 136196 | 73922 | 62500 | 75000 |
| 1 | passive | 75000 | 2000 | 91721 | 74386 | 69444 | 78125 |
More information for a better understanding of the thermal behavior
will be added after. The idea is to give detailed statistics
information about the undershots and overshots, the temperature speed,
etc... As all the information in a single file is too much, the idea
would be to create a directory named with the mitigation timestamp
where all data could be added.
Please note this code is immune against trip ordering but not against
a trip temperature change while a mitigation is happening. However,
this situation should be extremely rare, perhaps not happening and we
might question ourselves if something should be done in the core
framework for other components first.
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
[ rjw: White space fixups, rebase ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The thermal framework does not have any debug information except a
sysfs stat which is a bit controversial. This one allocates big chunks
of memory for every cooling devices with a high number of states and
could represent on some systems in production several megabytes of
memory for just a portion of it. As the sysfs is limited to a page
size, the output is not exploitable with large data array and gets
truncated.
The patch provides the same information than sysfs except the
transitions are dynamically allocated, thus they won't show more
events than the ones which actually occurred. There is no longer a
size limitation and it opens the field for more debugging information
where the debugfs is designed for, not sysfs.
The thermal debugfs directory structure tries to stay consistent with
the sysfs one but in a very simplified way:
thermal/
-- cooling_devices
|-- 0
| |-- clear
| |-- time_in_state_ms
| |-- total_trans
| `-- trans_table
|-- 1
| |-- clear
| |-- time_in_state_ms
| |-- total_trans
| `-- trans_table
|-- 2
| |-- clear
| |-- time_in_state_ms
| |-- total_trans
| `-- trans_table
|-- 3
| |-- clear
| |-- time_in_state_ms
| |-- total_trans
| `-- trans_table
`-- 4
|-- clear
|-- time_in_state_ms
|-- total_trans
`-- trans_table
The content of the files in the cooling devices directory is the same
as the sysfs one except for the trans_table which has the following
format:
Transition Hits
1->0 246
0->1 246
2->1 632
1->2 632
3->2 98
2->3 98
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
[ rjw: White space fixups, rebase ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
There are several rountines in the thermal netlink API that take a
thermal zone ID or a thermal zone type as their arguments, but from
their callers perspective it would be more convenient to pass a thermal
zone pointer to them and let them extract the necessary data from the
given thermal zone object by themselves.
Modify the code accordingly.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Because thermal_notify_tz_trip_add/delete() are never used, drop them
entirely along with the related code.
The addition or removal of trip points is not supported by the thermal
core and is unlikely to be supported in the future, so it is also
unlikely that these functions will ever be needed.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Instead of requiring the callers of thermal_notify_tz_trip_up/down() to
provide specific values needed to populate struct param in them, make
them extract those values from objects passed by the callers via const
pointers.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Reviewed-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Instead of requiring the caller of thermal_notify_tz_trip_change() to
provide specific values needed to populate struct param in it, make it
extract those values from objects passed to it by the caller via const
pointers.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Reviewed-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Because thermal_zone_trip_id() does not update the thermal zone object
passed to it, its pointer argument representing the thermal zone can be
const, so adjust its definition accordingly.
No functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
- Fixed DT bindings issue on Loongson (Binbin Zhou)
- Fixed returning NULL instead of -ENODEV on Loogsoo (Binbin Zhou)
- Added the DT binding for the tsens on SM8650 platform (Neil Armstrong)
- Added a reboot on critical option feature (Fabio Estevam)
- Made usage of DEFINE_SIMPLE_DEV_PM_OPS on AmLogic (Uwe Kleine-König)
- Added the D1/T113s THS controller support on Sun8i (Maxim Kiselev)
- Fixed example in the DT binding for QCom SPMI (Johan Hovold)
- Fixed compilation warning for the tmon utility (Florian Eckert)
- Added interrupt based configuration on Exynos along with a set of
related cleanups (Mateusz Majewski)
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEGn3N4YVz0WNVyHskqDIjiipP6E8FAmWT0yUACgkQqDIjiipP
6E+YjggAsrsQNrsUgiW0/M0i75kfcEBgLfXscyPFsUYwpByb+haWzAJSLrvCBko8
zPFvNE0or6KTQJCtseWWQDNQKWilAKARurEg7vuahfWo5LmfauPGMxsw+iHM9TgW
8Ptkc1biy3TNr1zVCpQrCZK9GdLGsG7JRgxHi4Hfr/Hb/FZN9Mm0Yk1pRpFU+pn8
Ff5UScMcU7NWhhxlQavLOoMmyAmh2k/jvfCSlmXGj7kxRrX8YIC02dTDjFm5cpyP
Toy2POWkcZr4Xr+kWOh8pxojZVwKU2pN7cYyLJn8+OO+rpUAf0ol4PW0mly7uMgH
1HCZ0x0hiNJQ8N6SwN+Eptq9TpgCBw==
=W2kc
-----END PGP SIGNATURE-----
Merge tag 'thermal-v6.8-rc1' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/thermal/linux into thermal
Merge thermal control material for 6.8-rc1 from Daniel Lezcano:
"- Converted Mediatek Thermal to the json-schema (Rafał Miłecki)
- Fixed DT bindings issue on Loongson (Binbin Zhou)
- Fixed returning NULL instead of -ENODEV on Loogsoo (Binbin Zhou)
- Added the DT binding for the tsens on SM8650 platform (Neil Armstrong)
- Added a reboot on critical option feature (Fabio Estevam)
- Made usage of DEFINE_SIMPLE_DEV_PM_OPS on AmLogic (Uwe Kleine-König)
- Added the D1/T113s THS controller support on Sun8i (Maxim Kiselev)
- Fixed example in the DT binding for QCom SPMI (Johan Hovold)
- Fixed compilation warning for the tmon utility (Florian Eckert)
- Added interrupt based configuration on Exynos along with a set of
related cleanups (Mateusz Majewski)"
* tag 'thermal-v6.8-rc1' of ssh://gitolite.kernel.org/pub/scm/linux/kernel/git/thermal/linux: (24 commits)
thermal/drivers/exynos: Use set_trips ops
thermal/drivers/exynos: Use BIT wherever possible
thermal/drivers/exynos: Split initialization of TMU and the thermal zone
thermal/drivers/exynos: Stop using the threshold mechanism on Exynos 4210
thermal/drivers/exynos: Simplify regulator (de)initialization
thermal/drivers/exynos: Handle devm_regulator_get_optional return value correctly
thermal/drivers/exynos: Wwitch from workqueue-driven interrupt handling to threaded interrupts
thermal/drivers/exynos: Drop id field
thermal/drivers/exynos: Remove an unnecessary field description
tools/thermal/tmon: Fix compilation warning for wrong format
dt-bindings: thermal: qcom-spmi-adc-tm5/hc: Clean up examples
dt-bindings: thermal: qcom-spmi-adc-tm5/hc: Fix example node names
thermal/drivers/sun8i: Add D1/T113s THS controller support
dt-bindings: thermal: sun8i: Add binding for D1/T113s THS controller
thermal: amlogic: Use DEFINE_SIMPLE_DEV_PM_OPS for PM functions
thermal: amlogic: Make amlogic_thermal_disable() return void
thermal/thermal_of: Allow rebooting after critical temp
reboot: Introduce thermal_zone_device_critical_reboot()
thermal/core: Prepare for introduction of thermal reboot
dt-bindings: thermal-zones: Document critical-action
...
Currently, each trip point defined in the device tree corresponds to a
single hardware interrupt. This commit instead switches to using two
hardware interrupts, whose values are set dynamically using the
set_trips callback. Additionally, the critical temperature threshold is
handled specifically.
Setting interrupts in this way also fixes a long-standing lockdep
warning, which was caused by calling thermal_zone_get_trips with our
lock being held. Do note that this requires TMU initialization to be
split into two parts, as done by the parent commit: parts of the
initialization call into the thermal_zone_device structure and so must
be done after its registration, but the initialization is also
responsible for setting up calibration, which must be done before
thermal_zone_device registration, which will call set_trips for the
first time; if the calibration is not done in time, the interrupt values
will be silently wrong!
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231201095625.301884-10-m.majewski2@samsung.com
The original driver did not use that macro and it allows us to make our
intentions slightly clearer.
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231201095625.301884-9-m.majewski2@samsung.com
This will be needed in the future, as the thermal zone subsystem might
call our callbacks right after devm_thermal_of_zone_register. Currently
we just make get_temp return EAGAIN in such case, but this will not be
possible with state-modifying callbacks, for instance set_trips.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231201095625.301884-8-m.majewski2@samsung.com
Exynos 4210 supports setting a base threshold value, which is added to
all trip points. This might be useful, but is not really necessary in
our usecase, so we always set it to 0 to simplify the code a bit.
Additionally, this change makes it so that we convert the value to the
calibrated one in a slightly different place. This is more correct
morally, though it does not make any change when single-point
calibration is being used (which is the case currently).
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231201095625.301884-7-m.majewski2@samsung.com
We rewrite the initialization to enable the regulator as part of devm,
which allows us to not handle the struct instance manually.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org> Link:
https://lore.kernel.org/r/20231201095625.301884-6-m.majewski2@samsung.com
Currently, if regulator is required in the SoC, but
devm_regulator_get_optional fails for whatever reason, the execution
will proceed without propagating the error. Meanwhile there is no
reason to output the error in case of -ENODEV.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231201095625.301884-5-m.majewski2@samsung.com
The workqueue boilerplate is mostly one-to-one what the threaded
interrupts do.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231201095625.301884-4-m.majewski2@samsung.com
We do not use the value, and only Exynos 7 defines this alias anyway.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231201095625.301884-3-m.majewski2@samsung.com
It seems that the field has been removed in one of the previous commits,
but the description has been forgotten.
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Mateusz Majewski <m.majewski2@samsung.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231201095625.301884-2-m.majewski2@samsung.com
This patch adds a thermal sensor controller support for the D1/T113s,
which is similar to the one on H6, but with only one sensor and
different scale and offset values.
Signed-off-by: Maxim Kiselev <bigunclemax@gmail.com>
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231217210629.131486-3-bigunclemax@gmail.com
This macro has the advantage over SIMPLE_DEV_PM_OPS that we don't have to
care about when the functions are actually used, so the corresponding
__maybe_unused can be dropped.
Also make use of pm_ptr() to discard all PM related stuff if CONFIG_PM
isn't enabled.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231116112633.668826-3-u.kleine-koenig@pengutronix.de
amlogic_thermal_disable() returned zero unconditionally and
amlogic_thermal_remove() already ignores the return value.
Make it return no value and modify amlogic_thermal_suspend to not check
the value.
This patch introduces no semantic changes, but makes it more obvious for
a human reader that amlogic_thermal_suspend() cannot fail.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231116112633.668826-2-u.kleine-koenig@pengutronix.de
Currently, the default mechanism is to trigger a shutdown after the
critical temperature is reached.
In some embedded cases, such behavior does not suit well, as the board may
be unattended in the field and rebooting may be a better approach.
The bootloader may also check the temperature and only allow the boot to
proceed when the temperature is below a certain threshold.
Introduce support for allowing a reboot to be triggered after the
critical temperature is reached.
If the "critical-action" devicetree property is not found, fall back to
the shutdown action to preserve the existing default behavior.
If a custom ops->critical exists, then it takes preference over
critical-actions.
Tested on a i.MX8MM board with the following devicetree changes:
thermal-zones {
cpu-thermal {
critical-action = "reboot";
};
};
Signed-off-by: Fabio Estevam <festevam@denx.de>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231129124330.519423-4-festevam@gmail.com
Introduce thermal_zone_device_critical_reboot() to trigger an
emergency reboot.
It is a counterpart of thermal_zone_device_critical() with the
difference that it will force a reboot instead of shutdown.
The motivation for doing this is to allow the thermal subystem
to trigger a reboot when the temperature reaches the critical
temperature.
Signed-off-by: Fabio Estevam <festevam@denx.de>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231129124330.519423-3-festevam@gmail.com
Add some helper functions to make it easier introducing the support
for thermal reboot.
No functional change.
Signed-off-by: Fabio Estevam <festevam@denx.de>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20231129124330.519423-2-festevam@gmail.com
When the thermal instance's weight is updated from the sysfs the governor
update_tz() callback is triggered. Implement proper reaction to this
event in the IPA, which would save CPU cycles spent in throttle().
This will speed-up the main throttle() IPA function and clean it up
a bit.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Support governors update when the thermal instance's weight has changed.
This allows to adjust internal state for the governor.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
[ rjw: Add two empty code lines aroung the locking ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
User space can change the weight of a thermal instance via sysfs while the
.throttle() callback is running for a governor, because weight_store()
does not use the zone lock.
The IPA governor uses instance weight values for power calculations and
caches the sum of them as total_weight, so it gets confused when one of
them changes while its .throttle() callback is running.
To prevent that from happening, use thermal zone locking in
weight_store().
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
There is a need to check if the cooling device in the thermal zone
supports IPA callback and is set for control trip point.
Refactor the code which validates the power actor capabilities and
make it more consistent in all places.
No intentional functional impact.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The new thermal callback allows to react to the change of cooling
instances in the thermal zone. Move the memory allocation to that new
callback and save CPU cycles in the throttle() code path.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Change trace event trace_thermal_power_allocator() to not use dynamic
array for requested power and granted power for all power actors.
Instead, simplify the trace event and print other simple values.
Add new trace event to print power actor information of requested power
and granted power. That trace event would be called in a loop for each
power actor. The trace data would be easier to parse comparing to the
dynamic array implementation.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Simplify the code and remove one extra 'if' block.
No intentional functional impact.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
In preparation for a subsequent change, rearrange check_power_actors().
No intentional functional impact.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Add a new callback to the struct thermal_governor. It can be used for
updating governors when there is a change in the thermal zone internals,
e.g. thermal cooling device is bind to the thermal zone.
That makes possible to move some heavy operations like memory allocations
related to the number of cooling instances out of the throttle() callback.
Both callback code paths (throttle() and update_tz()) are protected with
the same thermal zone lock, which guaranties the consistency.
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Add a helper function to check if there are listeners for
thermal_gnl_family multicast groups.
For now use it to avoid unnecessary allocations and sending
thermal genl messages when there are no recipients.
In the future, in conjunction with (not yet implemented) notification
of change in the netlink socket group membership, this helper can be
used to open/close hardware interfaces based on the presence of
user space subscribers.
Signed-off-by: Stanislaw Gruszka <stanislaw.gruszka@linux.intel.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Use enum instead of hard-coded numbers for indexing multicast groups.
Signed-off-by: Stanislaw Gruszka <stanislaw.gruszka@linux.intel.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The resume of thermal zones in thermal_pm_notify() is carried out
sequentially, which may be a problem if __thermal_zone_device_update()
takes a significant time to run for some thermal zones, because some
other thermal zones may need to wait for them to resume then and if
any other PM notifiers are going to be invoked after the thermal one,
they will need to wait for it either.
To address this, make thermal_pm_notify() switch the poll_queue delayed
work over to a one-shot thermal_zone_device_resume() work function that
will restore the original one during the thermal zone resume and queue
up poll_queue without a delay for each thermal zone.
Link: https://lore.kernel.org/linux-pm/20231120234015.3273143-1-radusolea@google.com/
Reported-by: Radu Solea <radusolea@google.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
In preparation for a subsequent change, move the initialization of the
poll_queue delayed work from thermal_zone_device_register_with_trips()
to thermal_zone_device_init() which is called by the former.
However, because thermal_zone_device_init() is also called by
thermal_pm_notify(), make the latter call cancel_delayed_work() on
poll_queue before invoking the former, so as to allow the work
item to be re-initialized safely.
Also move thermal_zone_device_check() which needs to be defined
before thermal_zone_device_init(), so the latter can pass it to the
INIT_DELAYED_WORK() macro.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
There are 3 synchronization issues with thermal zone suspend-resume
during system-wide transitions:
1. The resume code runs in a PM notifier which is invoked after user
space has been thawed, so it can run concurrently with user space
which can trigger a thermal zone device removal. If that happens,
the thermal zone resume code may use a stale pointer to the next
list element and crash, because it does not hold thermal_list_lock
while walking thermal_tz_list.
2. The thermal zone resume code calls thermal_zone_device_init()
outside the zone lock, so user space or an update triggered by
the platform firmware may see an inconsistent state of a
thermal zone leading to unexpected behavior.
3. Clearing the in_suspend global variable in thermal_pm_notify()
allows __thermal_zone_device_update() to continue for all thermal
zones and it may as well run before the thermal_tz_list walk (or
at any point during the list walk for that matter) and attempt to
operate on a thermal zone that has not been resumed yet. It may
also race destructively with thermal_zone_device_init().
To address these issues, add thermal_list_lock locking to
thermal_pm_notify(), especially arount the thermal_tz_list,
make it call thermal_zone_device_init() back-to-back with
__thermal_zone_device_update() under the zone lock and replace
in_suspend with per-zone bool "suspend" indicators set and unset
under the given zone's lock.
Link: https://lore.kernel.org/linux-pm/20231218162348.69101-1-bo.ye@mediatek.com/
Reported-by: Bo Ye <bo.ye@mediatek.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Correct one misuse of kernel-doc notation and one spelling error as
reported by codespell.
cpuidle_cooling.c:152: warning: cannot understand function prototype: 'struct thermal_cooling_device_ops cpuidle_cooling_ops = '
For the kernel-doc warning, don't use "/**" for a comment on data.
kernel-doc can be used for structure declarations but not definitions.
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
If device_register() in thermal_zone_device_register_with_trips()
returns an error, the tz variable is set to NULL and subsequently
dereferenced in kfree(tz->tzp).
Commit adc8749b15 ("thermal/drivers/core: Use put_device() if
device_register() fails") added the tz = NULL assignment in question to
avoid a possible double-free after dropping the reference to the zone
device. However, after commit 4649620d94 ("thermal: core: Make
thermal_zone_device_unregister() return after freeing the zone"), that
assignment has become redundant, because dropping the reference to the
zone device does not cause the zone object to be freed any more.
Drop it to address the NULL pointer dereference.
Fixes: 3d439b1a2a ("thermal/core: Alloc-copy-free the thermal zone parameters structure")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Initially the check against the get_temp ops in the
thermal_zone_device_update() was put in there in order to catch
drivers not providing this method.
Instead of checking again and again the function if the ops exists in
the update function, let's do the check at registration time, so it is
checked one time and for all.
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The _store callbacks of the trip point temperature and hysteresis sysfs
attributes invoke thermal_notify_tz_trip_change() to send a notification
regarding the trip point change, but when trip points are updated by the
platform firmware, trip point change notifications are not sent.
To make the behavior after a trip point change more consistent,
modify all of the 3 places where trip point temperature is updated
to use a new function called thermal_zone_set_trip_temp() for this
purpose and make that function call thermal_notify_tz_trip_change().
Note that trip point hysteresis can only be updated via sysfs and
trip_point_hyst_store() calls thermal_notify_tz_trip_change() already,
so this code path need not be changed.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Make thermal_genl_cmd_tz_get_trip() use for_each_trip() instead of an open-
coded loop over trip indices.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Make __thermal_zone_get_temp() use for_each_trip() instead of an open-
coded loop over trip indices.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Make __thermal_zone_set_trips() use for_each_trip() instead of an open-
coded loop over trip indices.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
The __thermal_zone_get_trip() header in drivers/thermal/thermal_core.h
is redundant, because there is one already in thermal.h, so drop it.
No functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
In order to avoid running __thermal_zone_device_update() for thermal
zones going away, the thermal zone lock is held around device_del()
in thermal_zone_device_unregister() and thermal_zone_device_update()
passes the given thermal zone device to device_is_registered().
This allows thermal_zone_device_update() to skip the
__thermal_zone_device_update() if device_del() has already run for
the thermal zone at hand.
However, instead of looking at driver core internals, the thermal
subsystem may as well rely on its own data structures for this
purpose. Namely, if the thermal zone is not present in
thermal_tz_list, it can be regarded as unavailable, which in fact is
already the case in thermal_zone_device_unregister(). Accordingly,
the device_is_registered() check in thermal_zone_device_update() can
be replaced with checking whether or not the node list_head in struct
thermal_zone_device is empty, in which case it is not there in
thermal_tz_list.
To make this work, though, it is necessary to initialize tz->node
in thermal_zone_device_register_with_trips() before registering the
thermal zone device and it needs to be added to thermal_tz_list and
deleted from it under its zone lock.
After the above modifications, the zone lock does not need to be
held around device_del() in thermal_zone_device_unregister() any more.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-and-tested-by: Lukasz Luba <lukasz.luba@arm.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Multiple places in the thermal subsystem (most importantly, sysfs
attribute callback functions) check if the given thermal zone device is
still registered in order to return early in case the device_del() in
thermal_zone_device_unregister() has run already.
However, after thermal_zone_device_unregister() has been made wait for
all of the zone-related activity to complete before returning, it is
not necessary to do that any more, because all of the code holding a
reference to the thermal zone device object will be waited for even if
it does not do anything special to enforce this.
Accordingly, drop all of the device_is_registered() checks that are now
redundant and get rid of the zone locking that is not necessary any more
after dropping them.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-and-tested-by: Lukasz Luba <lukasz.luba@arm.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Make thermal_zone_device_unregister() wait until all of the references
to the given thermal zone object have been dropped and free it before
returning.
This guarantees that when thermal_zone_device_unregister() returns,
there is no leftover activity regarding the thermal zone in question
which is required by some of its callers (for instance, modular driver
code that wants to know when it is safe to let the module go away).
Subsequently, this will allow some confusing device_is_registered()
checks to be dropped from the thermal sysfs and core code.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-and-tested-by: Lukasz Luba <lukasz.luba@arm.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Rework the _show() callback functions for the trip point temperature,
hysteresis and type attributes to avoid copying the values of struct
thermal_trip fields that they do not use and make them carry out the
same validation checks as the corresponding _store() callback functions.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Daniel Lezcano <daniel.lezcano@linaro.org>
