Commit 26984d9d58 ("PM / devfreq: passive: Keep cpufreq_policy for
possible cpus") reworked governor_passive.c, and accidently added a
tab in the first line, i.e., the SPDX-License-Identifier line.
The checkpatch script warns with the SPDX_LICENSE_TAG warning, and hence
pointed this issue out while investigating checkpatch warnings.
Revert this editing accident. No functional change.
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Remove cpufreq_passive_unregister_notifier from
cpufreq_passive_register_notifier in case of error as devfreq core
already call unregister on GOV_START fail.
This fix the kernel always printing a WARN on governor PROBE_DEFER as
cpufreq_passive_unregister_notifier is called two times and return
error on the second call as the cpufreq is already unregistered.
Fixes: a03dacb031 ("PM / devfreq: Add cpu based scaling support to passive governor")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
On a devfreq PROBE_DEFER, the freq_table in the driver profile struct,
is never reset and may be leaved in an undefined state.
This comes from the fact that we store the freq_table in the driver
profile struct that is commonly defined as static and not reset on
PROBE_DEFER.
We currently skip the reinit of the freq_table if we found
it's already defined since a driver may declare his own freq_table.
This logic is flawed in the case devfreq core generate a freq_table, set
it in the profile struct and then PROBE_DEFER, freeing the freq_table.
In this case devfreq will found a NOT NULL freq_table that has been
freed, skip the freq_table generation and probe the driver based on the
wrong table.
To fix this and correctly handle PROBE_DEFER, use a local freq_table and
max_state in the devfreq struct and never modify the freq_table present
in the profile struct if it does provide it.
Fixes: 0ec09ac2ce ("PM / devfreq: Set the freq_table of devfreq device")
Cc: stable@vger.kernel.org
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
of_get_child_by_name() returns a node pointer with refcount
incremented, we should use of_node_put() on it when done.
This function only calls of_node_put() in normal path,
missing it in error paths.
Add missing of_node_put() to avoid refcount leak.
Fixes: f262f28c14 ("PM / devfreq: event: Add devfreq_event class")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
HZ macros has been centralized in units.h since [1]. Use it to avoid
duplicated definition.
[1] commit e2c77032fc ("units: add the HZ macros")
Signed-off-by: Yicong Yang <yangyicong@hisilicon.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
With the passive governor, the cpu based scaling can PROBE_DEFER due to
the fact that CPU policy are not ready.
The cpufreq passive unregister notifier is called both from the
GOV_START errors and for the GOV_STOP and assume the notifier is
successfully registred every time. With GOV_START failing it's wrong to
loop over each possible CPU since the register path has failed for
some CPU policy not ready. Change the logic and unregister the notifer
based on the current allocated parent_cpu_data list to correctly handle
errors and the governor unregister path.
Fixes: a03dacb031 ("PM / devfreq: Add cpu based scaling support to passive governor")
Signed-off-by: Christian 'Ansuel' Marangi <ansuelsmth@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Don't print warning when a governor PROBE_DEFER as it's not a real
GOV_START fail.
Fixes: a03dacb031 ("PM / devfreq: Add cpu based scaling support to passive governor")
Signed-off-by: Christian 'Ansuel' Marangi <ansuelsmth@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
The cpufreq passive register notifier can PROBE_DEFER and the devfreq
struct is freed and then reallocaed on probe retry.
The current logic assume that the code can't PROBE_DEFER so the devfreq
struct in the this variable in devfreq_passive_data is assumed to be
(if already set) always correct.
This cause kernel panic as the code try to access the wrong address.
To correctly handle this, update the this variable in
devfreq_passive_data to the devfreq reallocated struct.
Fixes: a03dacb031 ("PM / devfreq: Add cpu based scaling support to passive governor")
Signed-off-by: Christian 'Ansuel' Marangi <ansuelsmth@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Each devfreq governor specifies the supported governor event
such as GOV_START and GOV_STOP. When not-supported event is required,
just return non-error. But, commit ce9a0d88d97a ("PM / devfreq: Add
cpu based scaling support to passive governor") returned the error
value. So that return non-error value when not-supported event is required.
Fixes: ce9a0d88d97a ("PM / devfreq: Add cpu based scaling support to passive governor")
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The passive governor requires the cpu data to get the next target frequency
of devfreq device if depending on cpu. In order to reduce the unnecessary
memory data, keep cpufreq_policy data for possible cpus instead of NR_CPU.
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Johnson Wang <johnson.wang@mediatek.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
In order to keep the consistent coding style between passive_devfreq
and passive_cpufreq, use common code for handling required opp property.
Also remove the unneed conditional statement and unify the comment
of both passive_devfreq and passive_cpufreq when getting the target frequency.
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Johnson Wang <johnson.wang@mediatek.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Many CPU architectures have caches that can scale independent of the
CPUs. Frequency scaling of the caches is necessary to make sure that the
cache is not a performance bottleneck that leads to poor performance and
power. The same idea applies for RAM/DDR.
To achieve this, this patch adds support for cpu based scaling to the
passive governor. This is accomplished by taking the current frequency
of each CPU frequency domain and then adjust the frequency of the cache
(or any devfreq device) based on the frequency of the CPUs. It listens
to CPU frequency transition notifiers to keep itself up to date on the
current CPU frequency.
To decide the frequency of the device, the governor does one of the
following:
* Derives the optimal devfreq device opp from required-opps property of
the parent cpu opp_table.
* Scales the device frequency in proportion to the CPU frequency. So, if
the CPUs are running at their max frequency, the device runs at its
max frequency. If the CPUs are running at their min frequency, the
device runs at its min frequency. It is interpolated for frequencies
in between.
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Johnson Wang <johnson.wang@mediatek.com>
Signed-off-by: Saravana Kannan <skannan@codeaurora.org>
[Sibi: Integrated cpu-freqmap governor into passive_governor]
Signed-off-by: Sibi Sankar <sibis@codeaurora.org>
[Chanwoo: Fix conflict with latest code and cleanup code]
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
In order to get frequency range within devfreq governors,
export devfreq_get_freq_range symbol within devfreq.
Reviewed-by: Matthias Kaehlcke <mka@chromium.org>
Tested-by: Chen-Yu Tsai <wenst@chromium.org>
Tested-by: Johnson Wang <johnson.wang@mediatek.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
See the previous patch ("soc: rockchip: power-domain: Manage resource
conflicts with firmware") for a thorough explanation of the conflicts.
While ARM Trusted Firmware may be modifying memory controller and
power-domain states, we need to block the kernel's power-domain driver.
If the power-domain driver is disabled, there is no resource conflict
and this becomes a no-op.
Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Acked-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
This static struct can get reused if the device gets removed/reprobed,
and that causes use-after-free in its ->freq_table.
Let's just move the struct to our dynamic allocation.
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
We want to keep the idle time fixed, so compute based on the current DDR
frequency.
The old properties were deprecated and never used, so we can safely drop
them from the driver.
This is a rewritten version of work by Lin Huang <hl@rock-chips.com>.
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Implement the newly-defined properties to allow disabling certain
power-saving-at-idle features at higher frequencies.
This is a rewritten version of work by Lin Huang <hl@rock-chips.com>.
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
We're going to add new usages, and it's cleaner to work with macros
instead of comments and magic numbers.
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
All of these properties are initialized by ARM Trusted Firmware, and
have been since the early days of this chip. It's redundant (and
possibly wrong) to do this here now. What's more, there seems to be some
confusion about the units and some of the definitions of this timing
struct: the DT docs say MHz for many of these, but downstream users were
in Hz (and therefore, the ATF interface was Hz). Also, the in-driver
usage for some of these (e.g., for comparing to target frequency) were
in Hz too. So doubly wrong.
We can avoid thinking about who got the right units by dropping the
unnecessary code and properties. They are marked deprecated in the
binding schema.
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
These properties are:
* undocumented
* directly representing software properties, not hardware properties
* unused (no in-tree users, yet; this IP block has so far only been used
in downstream kernels)
Let's just stick the values that downstream users have been using
directly in the driver and call it a day.
Signed-off-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Suppose devfreq_update_status() failure in devfreq_set_target() is not a
critical error, reduces the log severity.
Signed-off-by: Tzung-Bi Shih <tzungbi@google.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Compile-testing the driver without CONFIG_COMMON_CLK causes
a link failure:
arm-linux-gnueabi-ld: drivers/devfreq/sun8i-a33-mbus.o: in function `sun8i_a33_mbus_remove':
sun8i-a33-mbus.c:(.text+0x450): undefined reference to `clk_rate_exclusive_put'
arm-linux-gnueabi-ld: sun8i-a33-mbus.c:(.text+0x460): undefined reference to `clk_rate_exclusive_put'
arm-linux-gnueabi-ld: drivers/devfreq/sun8i-a33-mbus.o: in function `sun8i_a33_mbus_probe':
sun8i-a33-mbus.c:(.text+0x85c): undefined reference to `clk_rate_exclusive_get'
arm-linux-gnueabi-ld: sun8i-a33-mbus.c:(.text+0x878): undefined reference to `clk_rate_exclusive_get'
Fixes: 8bfd4858b4 ("PM / devfreq: Add a driver for the sun8i/sun50i MBUS")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
This driver works by adjusting the divider on the DRAM controller's
module clock. Thus there is no fixed set of OPPs, only "full speed" down
to "quarter speed" (or whatever the maximum divider is on that variant).
It makes use of the MDFS hardware in the MBUS, in "DFS" mode, which
takes care of updating registers during the critical section while DRAM
is inaccessible.
This driver should support several sunxi SoCs, starting with the A33,
which have a DesignWare DDR3 controller with merged PHY register space
and the matching MBUS register layout (so not A63 or later). However,
the driver has only been tested on the A64/H5, so those are the only
compatibles enabled for now.
Acked-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Since commit ea572f8160 ("PM / devfreq: Change return type of
devfreq_set_freq_table()"), all devfreq devices are expected to have a
valid freq_table. The devfreq core unconditionally dereferences
freq_table in the sysfs code and in get_freq_range().
Therefore, we need to ensure that freq_table is both non-null and
non-empty (length is > 0). If either check fails, replace the table
using set_freq_table() or return the error.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
When parsing devicetree, the function of_get_devfreq_events(), for each
device child node, iterates over array of possible events "ppmu_events"
till it finds one matching by node name. When match is found the
ppmu_events[i] points to element having both the name of the event and
the counters ID.
Each PPMU device child node might have an "event-name" property with the
name of the event, however due to the design of devfreq it must be the
same as the device node name. If it is not the same, the devfreq client
won't be able to use it via devfreq_event_get_edev_by_phandle().
Since PPMU device child node name must be equal to the "event-name"
property (event-name == ppmu_events[i].name), there is no need to find
the counters ID by the "event-name". Instead use ppmu_events[i].id
which must be equal to it.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Devicetree naming convention requires device node names to use hyphens
instead of underscore, so Exynos5422 devfreq event name
"ppmu-event3-dmc0_0" should be "ppmu-event3-dmc0-0". Newly introduced
dtschema enforces this, however the driver still expects old name with
an underscore.
Add new events for Exynos5422 while still accepting old name for
backwards compatibility.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@canonical.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
EMC clock is always-on and can't be zero. Check whether clk_round_rate()
returns zero rate and error out if it does. It can return zero if clock
tree isn't initialized properly.
Acked-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
HZ unit conversion macros are available in units.h, use them and remove
the duplicate definition.
The new macro has an unsigned long type.
All the code is dealing with unsigned long and the code using the macro is
doing a coercitive cast to unsigned long.
Link: https://lkml.kernel.org/r/20210816114732.1834145-5-daniel.lezcano@linaro.org
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Christian Eggers <ceggers@arri.de>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Chanwoo Choi <cw00.choi@samsung.com>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: Jonathan Cameron <jic23@kernel.org>
Cc: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Cc: Kyungmin Park <kyungmin.park@samsung.com>
Cc: Lars-Peter Clausen <lars@metafoo.de>
Cc: Lukasz Luba <lukasz.luba@arm.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Miquel Raynal <miquel.raynal@bootlin.com>
Cc: MyungJoo Ham <myungjoo.ham@samsung.com>
Cc: Peter Meerwald <pmeerw@pmeerw.net>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The 86ad9a24f2 ("PM / devfreq: Add required OPPs support to passive governor")
supported the required-opp property for using devfreq passive governor.
But, 86ad9a24f2 has caused the problem on use-case when required-opp
is not used such as exynos-bus.c devfreq driver. So that fix the
get_target_freq of passive governor for supporting the case of when
required-opp is not used.
Fixes: 86ad9a24f2 ("PM / devfreq: Add required OPPs support to passive governor")
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Use DEVICE_ATTR_RW helper instead of plain DEVICE_ATTR,
which makes the code a bit shorter and easier to read.
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
The driver can't support simple ondemand governor due to missing
.get_dev_status() capability.
Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Expose ACTMON devfreq device as a cooling device in order to throttle
memory freq on overheat. Throttling of memory freq has a significant
cooling effect on NVIDIA Tegra SoCs since higher memory freqs require
higher SoC core voltage which is one of the main causes of the heating.
Signed-off-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Current driver actually does not support simple ondemand governor
as it's unable to provide device load information. So removing
the unnecessary callback to avoid confusing.
Right now the driver is using userspace governor by default.
polling_ms was also dropped as it's not needed for non-ondemand
governor.
Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Set err code in the error path before jumping to the end of the function.
Fixes: 4dc3bab868 ("PM / devfreq: Add support delayed timer for polling mode")
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
i.MX is a DT-only platform, so of_match_ptr() can be safely
removed.
Remove the unneeded of_match_ptr().
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
i.MX is a DT-only platform, so of_match_ptr() can be safely
removed.
Remove the unneeded of_match_ptr().
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Current driver actually does not support simple ondemand governor
as it's unable to provide device load information. So removing
the unnecessary callback to avoid confusing.
Right now the driver is using userspace governor by default.
polling_ms was also dropped as it's not needed for non-ondemand
governor.
Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
First of all, no_central_polling was removed since
commit 7e6fdd4bad ("PM / devfreq: Core updates to support devices
which can idle")
Secondly, get_target_freq() is not only called only with update_devfreq()
notified by OPP now, but also min/max freq qos notifier.
So remove this invalid description now to avoid confusing.
Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Check .get_dev_status() in devfreq_update_stats in case it's abused
when a device does not provide it.
Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Common pattern of handling deferred probe can be simplified with
dev_err_probe(). Less code and the error value gets printed.
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Use the more accurate returned new_freq as resume_freq.
It's the same as how devfreq->previous_freq was updated.
Fixes: 83f8ca45af ("PM / devfreq: add support for suspend/resume of a devfreq device")
Signed-off-by: Dong Aisheng <aisheng.dong@nxp.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
The devfreq->lock is held for time of setup. Release the lock in the
error path, before jumping to the end of the function.
Change the goto destination which frees the allocated memory.
Cc: v5.9+ <stable@vger.kernel.org> # v5.9+
Fixes: 4dc3bab868 ("PM / devfreq: Add support delayed timer for polling mode")
Signed-off-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Currently the default behavior is to manually having the devfreq
backend to register themselves as a devfreq cooling device.
Instead of adding the code in the drivers for the thermal cooling
device registering, let's provide a flag in the devfreq's profile to
tell the common devfreq code to register the newly created devfreq as
a cooling device.
Suggested-by: Chanwoo Choi <cwchoi00@gmail.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Lukasz Luba <lukasz.luba@arm.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
* pm-devfreq:
PM / devfreq: rk3399_dmc: Remove unneeded semicolon
PM / devfreq: Replace devfreq->dev.parent as dev in devfreq_add_device
PM / devfreq: Correct spelling in a comment
* pm-tools:
cpupower: Add cpuid cap flag for MSR_AMD_HWCR support
cpupower: Remove family arg to decode_pstates()
cpupower: Condense pstate enabled bit checks in decode_pstates()
cpupower: Update family checks when decoding HW pstates
cpupower: Remove unused pscur variable.
cpupower: Add CPUPOWER_CAP_AMD_HW_PSTATE cpuid caps flag
cpupower: Correct macro name for CPB caps flag
cpupower: Update msr_pstate union struct naming
cpupower: add Makefile dependencies for install targets
