| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Merge changes in Intel thermal control drivers for 6.7-rc1:
- Add power floor notifications support to the int340x thermal control
driver (Srinivas Pandruvada).
- Rework updating trip points in the int340x thermal driver so that it
does not access thermal zone internals directly (Rafael Wysocki).
- Use param_get_byte() instead of param_get_int() as the max_idle module
parameter .get() callback in the Intel powerclamp thermal driver to
avoid possible out-of-bounds access (David Arcari).
- Add workload hints support to the the int340x thermal driver (Srinivas
Pandruvada).
* thermal-intel:
selftests/thermel/intel: Add test to read power floor status
thermal: int340x: processor_thermal: Enable power floor support
thermal: int340x: processor_thermal: Handle power floor interrupts
thermal: int340x: processor_thermal: Support power floor notifications
thermal: int340x: processor_thermal: Set feature mask before proc_thermal_add
thermal: int340x: processor_thermal: Common function to clear SOC interrupt
thermal: int340x: processor_thermal: Move interrupt status MMIO offset to common header
thermal: intel: powerclamp: fix mismatch in get function for max_idle
thermal: int340x: Use thermal_zone_for_each_trip()
thermal: int340x: processor_thermal: Ack all PCI interrupts
thermal: int340x: Add ArrowLake-S PCI ID
selftests/thermel/intel: Add test to read workload hint
thermal: int340x: Handle workload hint interrupts
thermal: int340x: processor_thermal: Add workload type hint interface
thermal: int340x: Remove PROC_THERMAL_FEATURE_WLT_REQ for Meteor Lake
thermal: int340x: processor_thermal: Use non MSI interrupts by default
thermal: int340x: processor_thermal: Add interrupt configuration function
thermal: int340x: processor_thermal: Move mailbox code to common module
|
|
Enable power floor feature support for Meteor Lake processors.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
On thermal device interrupt, if the interrupt is generated for passing
power floor status, call the callback to pass notification to the user
space.
First call proc_thermal_check_power_floor_intr() to check interrupt, if
this callback returns true, wake the IRQ thread to call
proc_thermal_power_floor_intr_callback() to notify user space.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
When the hardware reduces the power to the minimum possible, the power
floor is notified via an interrupt.
This can happen when user space requests a power limit via powercap RAPL
interface, which forces the system to enter to the lowest power. This
power floor indication can be used as a hint to resort to other methods
of reducing power than via RAPL power limit.
Before power floor status can be read or the firmware can trigger
notifications regarding it, it needs to be configured via a mailbox
command. The actual power floor status is read via bit 39 of MMIO
offset 0x5B18 of the processor thermal PCI device.
To show the current power floor status and get notification
on a sysfs attribute, add 2 new attributes to
/sys/bus/pci/devices/0000\:00\:04.0/power_limits/
power_floor_enable : This attribute is present when power floor
notifications are supported. This attribute allows to enable/disable
power floor notifications.
power_floor_status : This attribute is present when power floor
notifications are supported. When enabled via power_floor_enable, this
attribute shows the current power floor status.
The power floor implementation provides interfaces which are called
from the sysfs callbacks to enable/disable and read power floor
status. It also provides two additional interfaces to check if the
current processor thermal device interrupt is for power floor status
and to send notifications to user space.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Changelog and documentation changes edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The function proc_thermal_add() adds sysfs entries for power limits.
The feature mask of available features is not present at that time, so
it cannot be used by proc_thermal_add() to selectively create sysfs
attributes.
The feature mask is set by proc_thermal_mmio_add(), so modify the code
to call it before proc_thermal_add() so as to allow the latter to use
the feature mask.
There is no functional impact with this change.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The SOC interrupt status register contains multiple interrupt sources
(workload hint interrupt and power floor interrupt). It is not possible
to clear individual interrupt source with read-modify-write, as it may
clear the new interrupt from the firmware after a read operation. It is
also not possible to set the interrupt status bit to 1 for the other
interrupt source, which is not part of clearing.
Hence, create a common function, to clear all status bits at once.
Call this function after processing all interrupt sources.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
common header
Move define SOC_WT_RES_INT_STATUS_OFFSET to processor_thermal_device.h.
This way it can be reused in other modules.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
KASAN reported this
[ 444.853098] BUG: KASAN: global-out-of-bounds in param_get_int+0x77/0x90
[ 444.853111] Read of size 4 at addr ffffffffc16c9220 by task cat/2105
...
[ 444.853442] The buggy address belongs to the variable:
[ 444.853443] max_idle+0x0/0xffffffffffffcde0 [intel_powerclamp]
There is a mismatch between the param_get_int and the definition of
max_idle. Replacing param_get_int with param_get_byte resolves this
issue.
Fixes: ebf519710218 ("thermal: intel: powerclamp: Add two module parameters")
Cc: 6.3+ <stable@vger.kernel.org> # 6.3+
Signed-off-by: David Arcari <darcari@redhat.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Modify int340x_thermal_update_trips() to use thermal_zone_for_each_trip()
for walking trips instead of using the trips[] table passed to the
thermal zone registration function.
For this purpose, store active trip point indices in the priv fieids of
the corresponding thermal_trip structures.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
|
|
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
All interrupts from the processor thermal PCI device require ACK. This
is done by writing 0x01 at offset 0xDC in the config space.
This is already done for the thereshold interrupt. Extend this for the
workload hint interrupt.
Fixes: e682b86211a1 ("thermal: int340x: Handle workload hint interrupts")
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Add ArrowLake-S PCI ID for processor thermal device.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
On thermal device interrupt, if the interrupt is generated for passing
workload hint, call the callback to pass notification to the user
space.
First call proc_thermal_check_wt_intr() to check interrupt, if this
callback returns true, wake the IRQ thread to call
proc_thermal_wt_intr_callback() to notify user space.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Subject adjustment, changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Prior to Meteor Lake processor generation, user space can pass workload
type request to the firmware. Then firmware can optimize power based on
the indicated workload type. User space also uses workload type requests
to implement its own heuristics.
The firmware in Meteor Lake processor generation is capable of predicting
workload type without software help.
To avoid duplicate processing, add a sysfs interface allowing user space
to obtain the workload hint from the firmware instead of trying to
predict the workload type by itself.
This workload hint is passed from the firmware via MMIO offset 0x5B18 of
the processor thermal PCI device. Before workload hints can be produced by
the firmware, it needs to be configured via a mailbox command. This
mailbox command turns ON the workload hint and it allows to program a
notification delay to control the rate of notifications.
The notification delay can be changed from user space vis sysfs.
Attribute group 'workload_hint' in sysfs is used for implementing the
workload hints interface between user space and the kernel.
It contains the following attributes:
workload_type_enable:
Enables/disables workload type hints from the firmware.
notification_delay_ms:
Notification delay in milliseconds.
workload_type_index:
The current workload type index predicted by the firmware (see
the documentation changes below for supported index values and
their meaning).
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Changelog edits, documentation edits, whitespace adjustments ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Meteor Lake processor supports firmware hints for predicting workload
type. So, remove support for passing workload hints to the firmware.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
[ rjw: Subject adjustment ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
There are issues in using MSI interrupts for processor thermal device.
The support is not consistent across generations. But the legacy PCI
interrupts work on all current generations.
Hence always use legacy PCI interrupts by default, instead of MSI.
Add a module param to use of MSI, so that MSI can be still used.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Some features like workload type prediction and power floor events
require interrupt support to avoid polling. Here interrupts are enabled
and disabled via sending mailbox commands. The mailbox command ID is
0x1E for read and 0x1F for write.
The interrupt configuration will require mutex protection as it involves
read-modify-write operation. Since mutex are already used in the mailbox
read/write functions: send_mbox_write_cmd() and send_mbox_read_cmd(),
there will be double locking. But, this can be avoided by moving mutexes
from mailbox read/write processing functions to the callers:
processor_thermal_send_mbox_[read|write]_cmd().
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Adjust subject, fix up computation ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The processor thermal mailbox is used for workload type request and
also in the processor thermal RFIM module. So, move the workload type
request code to its own module from the current processor thermal
mailbox module.
processor_thermal_mailbox.c contains only mailbox read/write related
source code. The source related to workload_types requests is moved to
a module processor_thermal_wt_req.c.
In addition
- Rename PROC_THERMAL_FEATURE_MBOX to PROC_THERMAL_FEATURE_WT_REQ.
- proc_thermal_mbox_add(), which adds workload type sysfs attribute group
is renamed to proc_thermal_wt_req_add().
- proc_thermal_mbox_remove() is renamed to proc_thermal_wt_req_remove().
While here, resolve check patch warnings for 100 columns for only modified
lines.
No functional changes are expected.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Reviewed-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
All of the remaining callers of thermal_zone_device_register()
can use thermal_tripless_zone_device_register(), so make them
do so in order to allow the former to be dropped.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Miquel Raynal <miquel.raynal@bootlin.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull thermal control updates from Rafael Wysocki:
"These rework the Intel DTS IOSF and the ACPI thermal drivers to pass
tables of generic trip point structures to the core during
initialization and make some requisite modifications in the thermal
core, fix a few issues elsewhere and clean up code.
This includes changes that are present in the ACPI updates too,
because they involve both ACPI and the thermal core. The list of
specific changes below is limited to thermal control, however.
Specifics:
- Make the ACPI thermal driver use its own Notify() handler (Michal
Wilczynski)
- Rework the ACPI thermal driver to use a table of generic trip point
structures on top of the internal representation of trip points and
remove thermal zone callbacks that are not necessary any more from
that driver (Rafael Wysocki)
- Fix a few issues in the Intel DTS IOSF thermal driver, clean up
code in it and make it pass tables of generic trip point structures
to the core during thermal zone registration (Rafael Wysocki)
- Drop a redundant check from the Intel DTS IOSF thermal driver's
"remove" routine (Zhang Rui)
- Use module_platform_driver() to replace an open-coded counterpart
of it in the int340x thermal driver (Yang Yingliang)
- Fix possible uninitialized value access in __thermal_of_bind() and
__thermal_of_unbind() (Peng Fan)
- Make the int3400 driver use thermal zone device wrappers (Daniel
Lezcano)
- Remove redundant thermal zone state check from the int340x thermal
driver (Daniel Lezcano)
- Drop non-functional nocrt parameter from ACPI thermal (Mario
Limonciello)
- Explicitly include correct DT includes in the thermal core and
drivers (Rob Herring)"
* tag 'thermal-6.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
thermal: intel: intel_soc_dts_iosf: Remove redundant check
thermal: intel: int340x: simplify the code with module_platform_driver()
thermal/of: Fix potential uninitialized value access
thermal: intel: intel_soc_dts_iosf: Use struct thermal_trip
thermal: intel: intel_soc_dts_iosf: Rework critical trip setup
thermal: intel: intel_soc_dts_iosf: Add helper for resetting trip points
thermal: intel: intel_soc_dts_iosf: Change initialization ordering
thermal: intel: intel_soc_dts_iosf: Pass sensors to update_trip_temp()
thermal: intel: intel_soc_dts_iosf: Untangle update_trip_temp()
thermal: intel: intel_soc_dts_iosf: Always assume notification support
thermal: intel: intel_soc_dts_iosf: Drop redundant symbol definition
thermal: intel: intel_soc_dts_iosf: Always use 2 trips
thermal: Explicitly include correct DT includes
thermal/drivers/int340x: Do not check the thermal zone state
thermal/drivers/int3400: Use thermal zone device wrappers
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf event updates from Ingo Molnar:
- AMD IBS improvements
- Intel PMU driver updates
- Extend core perf facilities & the ARM PMU driver to better handle ARM big.LITTLE events
- Micro-optimize software events and the ring-buffer code
- Misc cleanups & fixes
* tag 'perf-core-2023-08-28' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/uncore: Remove unnecessary ?: operator around pcibios_err_to_errno() call
perf/x86/intel: Add Crestmont PMU
x86/cpu: Update Hybrids
x86/cpu: Fix Crestmont uarch
x86/cpu: Fix Gracemont uarch
perf: Remove unused extern declaration arch_perf_get_page_size()
perf: Remove unused PERF_PMU_CAP_HETEROGENEOUS_CPUS capability
arm_pmu: Remove unused PERF_PMU_CAP_HETEROGENEOUS_CPUS capability
perf/x86: Remove unused PERF_PMU_CAP_HETEROGENEOUS_CPUS capability
arm_pmu: Add PERF_PMU_CAP_EXTENDED_HW_TYPE capability
perf/x86/ibs: Set mem_lvl_num, mem_remote and mem_hops for data_src
perf/mem: Add PERF_MEM_LVLNUM_NA to PERF_MEM_NA
perf/mem: Introduce PERF_MEM_LVLNUM_UNC
perf/ring_buffer: Use local_try_cmpxchg in __perf_output_begin
locking/arch: Avoid variable shadowing in local_try_cmpxchg()
perf/core: Use local64_try_cmpxchg in perf_swevent_set_period
perf/x86: Use local64_try_cmpxchg
perf/amd: Prevent grouping of IBS events
|
|
Remove the redundant check in remove_dts_thermal_zone() because all of
its existing callers pass a valid pointer as the argument.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The init/exit() of the driver only calls platform_driver_{un}register(),
so it can be simpilfied by using module_platform_driver().
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
[ rjw: Subject and changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Because the number of trip points in each thermal zone and their
types are known to intel_soc_dts_iosf_init() prior to the registration
of the thermal zones, make it create an array of struct thermal_trip
entries in each struct intel_soc_dts_sensor_entry object and make
add_dts_thermal_zone() use thermal_zone_device_register_with_trips()
for thermal zone registration and pass that array as its second
argument.
Drop the sys_get_trip_temp() and sys_get_trip_type() callback
functions along with the respective callback pointers in
tzone_ops, because they are not necessary any more.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
Critical trip points appear in the DTS thermal zones only after those
thermal zones have been registered via intel_soc_dts_iosf_init().
Moreover, they are "created" by changing the type of an existing trip
point from THERMAL_TRIP_PASSIVE to THERMAL_TRIP_CRITICAL via
intel_soc_dts_iosf_add_read_only_critical_trip(), the caller of which
has to be careful enough to pass at least 1 as the number of read-only
trip points to intel_soc_dts_iosf_init() beforehand.
This is questionable, because user space may have started to use the
trips at the time when intel_soc_dts_iosf_add_read_only_critical_trip()
runs and there is no synchronization between it and sys_set_trip_temp().
To address it, use the observation that nonzero number of read-only
trip points is only passed to intel_soc_dts_iosf_init() when critical
trip points are going to be used, so in fact that function may get all
of the information regarding the critical trip points upfront and it
can configure them before registering the corresponding thermal zones.
Accordingly, replace the read_only_trip_count argument of
intel_soc_dts_iosf_init() with a pair of new arguments related to
critical trip points: a bool one indicating whether or not critical
trip points are to be used at all and an int one representing the
critical trip point temperature offset relative to Tj_max. Use these
arguments to configure the critical trip points before the registration
of the thermal zones and to compute the number of writeable trip points
in add_dts_thermal_zone().
Modify both callers of intel_soc_dts_iosf_init() to take these changes
into account and drop the intel_soc_dts_iosf_add_read_only_critical_trip()
call, that is not necessary any more, from intel_soc_thermal_init(),
which also allows it to return success right after requesting the IRQ.
Finally, drop intel_soc_dts_iosf_add_read_only_critical_trip()
altogether, because it does not have any more users.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
Because trip points are reset for each sensor in two places in the
same way, add a helper function for that to reduce code duplication
a bit.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
The initial configuration of trip points in intel_soc_dts_iosf_init()
takes place after registering the sensor thermal zones which is
potentially problematic, because it may race with the setting of trip
point temperatures via sysfs, as there is no synchronization between it
and sys_set_trip_temp().
To address this, change the initialization ordering so that the trip
points are configured prior to the registration of thermal zones.
Accordingly, change the cleanup ordering in intel_soc_dts_iosf_exit()
to remove the thermal zones before resetting the trip points.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
After previous changes, update_trip_temp() only uses its dts argument to
get to the sensors field in the struct intel_soc_dts_sensor_entry object
pointed to by that argument, so pass the value of that field directly to
it instead.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
Function update_trip_temp() is currently used for the initialization
of trip points as well as for changing trip point temperatures in
sys_set_trip_temp().
This is quite confusing and passing the value of dts->trip_types[trip]
to it so that it can store that value in the same memory location is
not particularly useful, because it only is necessary to set the
trip point type once, at the initialization time.
For this reason, drop the last argument from update_trip_temp() and
introduce configure_trip() calling the former internally for the
initial configuration of trip points.
Modify the majority of update_trip_temp() callers to use
configure_trip() instead of it.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
None of the existing callers of intel_soc_dts_iosf_init() passes
INTEL_SOC_DTS_INTERRUPT_NONE as the first argument to it, so the
notification local variable in it is always true and the
notification_support argument of add_dts_thermal_zone() is always
true either.
For this reason, drop the notification local variable from
intel_soc_dts_iosf_init() and the notification_support argument from
add_dts_thermal_zone() and rearrange the latter to always set
writable_trip_cnt and trip_mask.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
SOC_MAX_DTS_SENSORS is already defined in intel_soc_dts_iosf.h which is
included in intel_soc_dts_iosf.c, so it does not need to be defined in
the latter again.
Drop the redundant definition of that symbol from intel_soc_dts_iosf.c.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
Both the existing callers of intel_soc_dts_iosf_init() pass 2 as the trip
count argument, so it can be replaced with SOC_MAX_DTS_TRIPS everywhere in
the code and the trip_count argument of that function can be dropped.
This also allows the trip_count field to be dropped from struct
intel_soc_dts_sensor_entry, as it is always equal to 2, and some
related code can be simplified.
Make changes accordingly.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
|
|
Alderlake N is an E-core only product using Gracemont
micro-architecture. It fits the pre-existing naming scheme perfectly
fine, adhere to it.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20230807150405.686834933@infradead.org
|
|
Depends on the interface used, the RAPL registers can be either MSR
indexes or memory mapped IO addresses. Current RAPL common code uses u64
to save both MSR and memory mapped IO registers. With this, when
handling register address with an __iomem annotation, it triggers a
sparse warning like below:
sparse warnings: (new ones prefixed by >>)
>> drivers/powercap/intel_rapl_tpmi.c:141:41: sparse: sparse: incorrect type in initializer (different address spaces) @@ expected unsigned long long [usertype] *tpmi_rapl_regs @@ got void [noderef] __iomem * @@
drivers/powercap/intel_rapl_tpmi.c:141:41: sparse: expected unsigned long long [usertype] *tpmi_rapl_regs
drivers/powercap/intel_rapl_tpmi.c:141:41: sparse: got void [noderef] __iomem *
Fix the problem by using a union to save the registers instead.
Suggested-by: David Laight <David.Laight@ACULAB.COM>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202307031405.dy3druuy-lkp@intel.com/
Tested-by: Wang Wendy <wendy.wang@intel.com>
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
[ rjw: Subject and changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The driver is accessing the thermal zone state to ensure the state is
different from the one we want to set.
We don't want the driver to access the thermal zone device internals.
Actually, the thermal core code already checks if the thermal zone's
state is different before calling this function, thus this check is
duplicate.
Remove it.
Acked-by: srinivas pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Use the thermal core API to access the thermal zone "type" field
instead of directly using the structure field. While here, remove
access to the temperature field, as this driver is reporting fake
temperature, which can be replaced with INT3400_FAKE_TEMP. Also
replace hardcoded 20C with INT3400_FAKE_TEMP
Acked-by: srinivas pandruvada <srinivas.pandruvada@linux.intel.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull thermal control updates from Rafael Wysocki:
"These extend the int340x thermal driver, add thermal DT bindings for
some Qcom platforms, add DT bindings and support for Armada AP807 and
MSM8909, allow selecting the bang-bang thermal governor as the default
one, address issues in several thermal drivers for ARM platforms and
clean up code.
Specifics:
- Add new IOCTLs to the int340x thermal driver to allow user space to
retrieve the Passive v2 thermal table (Srinivas Pandruvada)
- Add DT bindings for SM6375, MSM8226 and QCM2290 Qcom platforms
(Konrad Dybcio)
- Add DT bindings and support for QCom MSM8226 (Matti Lehtimäki)
- Add DT bindings for QCom ipq9574 (Praveenkumar I)
- Convert bcm2835 DT bindings to the yaml schema (Stefan Wahren)
- Allow selecting the bang-bang governor as default (Thierry Reding)
- Refactor and prepare the code to set the scene for RCar Gen4
(Wolfram Sang)
- Clean up and fix the QCom tsens drivers. Add DT bindings and
calibration for the MSM8909 platform (Stephan Gerhold)
- Revert a patch introducing a wrong usage of devm_of_iomap() on the
Mediatek platform (Ricardo Cañuelo)
- Fix the clock vs reset ordering in order to conform to the
documentation on the sun8i (Christophe JAILLET)
- Prevent setting up undocumented registers, enable the only
described sensors and add the version 2.1 on the Qoriq sensor (Peng
Fan)
- Add DT bindings and support for the Armada AP807 (Alex Leibovich)
- Update the mlx5 driver with the recent thermal changes (Daniel
Lezcano)
- Convert to platform remove callback returning void on STM32 (Uwe
Kleine-König)
- Add an error information printing for devm_thermal_add_hwmon_sysfs()
and remove the error from the Sun8i, Amlogic, i.MX, TI, K3, Tegra,
Qoriq, Mediateka and QCom (Yangtao Li)
- Register as hwmon sensor for the Generic ADC (Chen-Yu Tsai)
- Use the dev_err_probe() function in the QCom tsens alarm driver
(Luca Weiss)"
* tag 'thermal-6.5-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (39 commits)
thermal/drivers/qcom/temp-alarm: Use dev_err_probe
thermal/drivers/generic-adc: Register thermal zones as hwmon sensors
thermal/drivers/mediatek/lvts_thermal: Remove redundant msg in lvts_ctrl_start()
thermal/drivers/qcom: Remove redundant msg at probe time
thermal/drivers/ti-soc: Remove redundant msg in ti_thermal_expose_sensor()
thermal/drivers/qoriq: Remove redundant msg in qoriq_tmu_register_tmu_zone()
thermal/drivers/tegra: Remove redundant msg in tegra_tsensor_register_channel()
drivers/thermal/k3: Remove redundant msg in k3_bandgap_probe()
thermal/drivers/imx: Remove redundant msg in imx8mm_tmu_probe() and imx_sc_thermal_probe()
thermal/drivers/amlogic: Remove redundant msg in amlogic_thermal_probe()
thermal/drivers/sun8i: Remove redundant msg in sun8i_ths_register()
thermal/hwmon: Add error information printing for devm_thermal_add_hwmon_sysfs()
thermal/drivers/stm32: Convert to platform remove callback returning void
net/mlx5: Update the driver with the recent thermal changes
thermal/drivers/armada: Add support for AP807 thermal data
dt-bindings: armada-thermal: Add armada-ap807-thermal compatible
thermal/drivers/qoriq: Support version 2.1
thermal/drivers/qoriq: Only enable supported sensors
thermal/drivers/qoriq: No need to program site adjustment register
thermal/drivers/mediatek/lvts_thermal: Register thermal zones as hwmon sensors
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull power management updates from Rafael Wysocki:
"These add Intel TPMI (Topology Aware Register and PM Capsule
Interface) support to the power capping subsystem, extend the
intel_idle driver to work in VM guests where MWAIT is not available,
extend the system-wide power management diagnostics, fix bugs and
clean up code.
Specifics:
- Introduce power capping core support for Intel TPMI (Topology Aware
Register and PM Capsule Interface) and a TPMI interface driver for
Intel RAPL (Zhang Rui, Dan Carpenter)
- Fix CONFIG_IOSF_MBI dependency in the Intel RAPL power capping
driver (Zhang Rui)
- Fix invalid initialization for pl4_supported field in the Intel
RAPL power capping driver (Sumeet Pawnikar)
- Clean up the intel_idle driver, make it work with VM guests that
cannot use the MWAIT instruction and address the case in which the
host may enter a deep idle state when the guest is idle (Arjan van
de Ven)
- Prevent cpufreq drivers that provide the ->adjust_perf() callback
without a ->fast_switch() one which is used as a fallback from the
former in some cases (Wyes Karny)
- Fix some issues related to the AMD P-state cpufreq driver (Mario
Limonciello, Wyes Karny)
- Fix the energy_performance_preference attribute handling in the
intel_pstate driver in passive mode (Tero Kristo)
- Fix the handling of pm_suspend_target_state when CONFIG_PM is unset
(Kai-Heng Feng)
- Correct spelling mistake in a comment in the hibernation code (Wang
Honghui)
- Add arch_resume_nosmt() prototype to avoid a "missing prototypes"
build warning (Arnd Bergmann)
- Restrict pm_pr_dbg() to system-wide power transitions and use it in
a few additional places (Mario Limonciello)
- Drop verification of in-params from genpd_add_device() and ensure
that all of its callers will do it (Ulf Hansson)
- Prevent possible integer overflows from occurring in
genpd_parse_state() (Nikita Zhandarovich)
- Reorder fieldls in 'struct devfreq_dev_status' to reduce its size
somewhat (Christophe JAILLET)
- Ensure that the Exynos PPMU driver is already loaded before the
Exynos Bus driver starts probing so as to avoid a possible freeze
loading of the kernel modules (Marek Szyprowski)
- Fix variable deferencing before NULL check in the mtk-cci devfreq
driver (Sukrut Bellary)"
* tag 'pm-6.5-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (42 commits)
intel_idle: Add a "Long HLT" C1 state for the VM guest mode
cpufreq: intel_pstate: Fix energy_performance_preference for passive
cpufreq: amd-pstate: Add a kernel config option to set default mode
cpufreq: amd-pstate: Set a fallback policy based on preferred_profile
ACPI: CPPC: Add definition for undefined FADT preferred PM profile value
cpufreq: amd-pstate: Set default governor to schedutil
PM: domains: Move the verification of in-params from genpd_add_device()
cpufreq: amd-pstate: Make amd-pstate EPP driver name hyphenated
cpufreq: amd-pstate: Write CPPC enable bit per-socket
intel_idle: Add support for using intel_idle in a VM guest using just hlt
cpufreq: Fail driver register if it has adjust_perf without fast_switch
intel_idle: clean up the (new) state_update_enter_method function
intel_idle: refactor state->enter manipulation into its own function
platform/x86/amd: pmc: Use pm_pr_dbg() for suspend related messages
pinctrl: amd: Use pm_pr_dbg to show debugging messages
ACPI: x86: Add pm_debug_messages for LPS0 _DSM state tracking
include/linux/suspend.h: Only show pm_pr_dbg messages at suspend/resume
powercap: RAPL: Fix a NULL vs IS_ERR() bug
powercap: RAPL: Fix CONFIG_IOSF_MBI dependency
powercap: RAPL: fix invalid initialization for pl4_supported field
...
|
|
|
|
Merge power capping updates for 6.5-rc1:
- Introduce power capping core support for Intel TPMI (Topology Aware
Register and PM Capsule Interface) and a TPMI interface driver for
Intel RAPL (Zhang Rui, Dan Carpenter).
- Fix CONFIG_IOSF_MBI dependency in the Intel RAPL power capping
driver (Zhang Rui).
- Fix invalid initialization for pl4_supported field in the Intel RAPL
power capping driver (Sumeet Pawnikar).
* powercap:
powercap: RAPL: Fix a NULL vs IS_ERR() bug
powercap: RAPL: Fix CONFIG_IOSF_MBI dependency
powercap: RAPL: fix invalid initialization for pl4_supported field
powercap: intel_rapl: Introduce RAPL TPMI interface driver
powercap: intel_rapl: Introduce core support for TPMI interface
powercap: intel_rapl: Introduce RAPL I/F type
powercap: intel_rapl: Make cpu optional for rapl_package
powercap: intel_rapl: Remove redundant cpu parameter
powercap: intel_rapl: Add support for lock bit per Power Limit
powercap: intel_rapl: Cleanup Power Limits support
powercap: intel_rapl: Use bitmap for Power Limits
powercap: intel_rapl: Change primitive order
powercap: intel_rapl: Use index to initialize primitive information
powercap: intel_rapl: Support per domain energy/power/time unit
powercap: intel_rapl: Support per Interface primitive information
powercap: intel_rapl: Support per Interface rapl_defaults
powercap: intel_rapl: Allow probing without CPUID match
powercap: intel_rapl: Remove unused field in struct rapl_if_priv
|
|
Since commit 955fb8719efb ("thermal/intel/intel_soc_dts_iosf: Use Intel
TCC library") intel_soc_dts_iosf is reporting the wrong temperature.
The driver expects tj_max to be in milli-degrees-celcius but after
the switch to the TCC library this is now in degrees celcius so
instead of e.g. 90000 it is set to 90 causing a temperature 45
degrees below tj_max to be reported as -44910 milli-degrees
instead of as 45000 milli-degrees.
Fix this by adding back the lost factor of 1000.
Fixes: 955fb8719efb ("thermal/intel/intel_soc_dts_iosf: Use Intel TCC library")
Reported-by: Bernhard Krug <b.krug@elektronenpumpe.de>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Zhang Rui <rui.zhang@intel.com>
Cc: 6.3+ <stable@vger.kernel.org> # 6.3+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Export Passive version 2 table similar to the way _TRT and _ART tables
via IOCTLs.
This removes need for binary utility to read ACPI Passive 2 table by
providing open source support. This table already has open source
implementation in the user space thermald, when the table is part of
data vault exported by the int3400 sysfs.
This table is supported in some older platforms before Ice Lake
generation.
Passive 2 tables contain multiple entries. Each entry has following
fields:
* Source: Named Reference (String). This is the source device for
temperature.
* Target: Named Reference (String). This is the target device to
control.
* Priority: Priority of this device compared to others.
* SamplingPeriod: Time Period in 1/10 of seconds unit.
* PassiveTemp: Passive Temperature in 1/10 of Kelvin.
* SourceDomain: Domain for the source (00:Processor, others reserved).
* ControlKnob: Type of control knob (00:Power Limit 1, others: reserved)
* Limit: The target state to set on reaching passive temperature.
This can be a string "max", "min" or a power limit value.
* LimitStepSize: Step size during activation.
* UnLimitStepSize: Step size during deactivation.
* Reserved1: Reserved
Three IOCTLs are added similar to IOCTLs for reading TRT:
ACPI_THERMAL_GET_PSVT_COUNT: Number of passive 2 entries.
ACPI_THERMAL_GET_PSVT_LEN: Total return data size (count x each
passive 2 entry size).
ACPI_THERMAL_GET_PSVT: Get the data as an array of objects with
passive 2 entries.
This change is based on original development done by:
Todd Brandt <todd.e.brandt@linux.intel.com>
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
[ rjw: Changelog and subject edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Prior to the commit "763bd29fd3d1 ("thermal: int340x_thermal: Use
sysfs_emit_at() instead of scnprintf()", there was a new line after each
UUID string.
With the newline removed, existing user space like "thermald" fails to
compare each supported UUID as it is using getline() to read UUID and
apply correct thermal table.
To avoid breaking existing user space, add newline after each UUID string.
Signed-off-by: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Fixes: 763bd29fd3d1 ("thermal: int340x_thermal: Use sysfs_emit_at() instead of scnprintf()")
Cc: 6.3+ <stable@vger.kernel.org> # 6.3+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Different RAPL Interfaces may have different primitive information and
rapl_defaults calls.
To better distinguish this difference in the RAPL framework code,
introduce a new enum to represent different types of RAPL Interfaces.
No functional change.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Wang Wendy <wendy.wang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
MSR RAPL Interface always removes a rapl_package when all the CPUs in
that rapl_package are offlined. This is because it relies on an online
CPU to access the MSR.
But for RAPL Interface using MMIO registers, when all the cpus within
the rapl_package are offlined,
1. the register can still be accessed
2. monitoring and setting the Power Pimits for the rapl_package is still
meaningful because of uncore power.
This means that, a valid rapl_package doesn't rely on one or more cpus
being onlined.
For this sense, make cpu optional for rapl_package. A rapl_package can
be registered either using a CPU id to represent the physical
package/die, or using the physical package id directly.
Note that, the thermal throttling interrupt is not disabled via
MSR_IA32_PACKAGE_THERM_INTERRUPT for such rapl_package at the moment.
If it is still needed in the future, this can be achieved by selecting
an onlined CPU using the physical package id.
Note that, processor_thermal_rapl, the current MMIO RAPL Interface
driver, can also be converted to register using a package id instead.
But this is not done right now because processor_thermal_rapl driver
works on single-package systems only, and offlining the only package
will not happen. So keep the previous logic.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Wang Wendy <wendy.wang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Currently, a RAPL package is registered with the number of Power Limits
supported in each RAPL domain. But this doesn't tell which Power Limits
are available. Using the number of Power Limits supported to guess the
availability of each Power Limit is fragile.
Use bitmap to represent the availability of each Power Limit.
Note that PL1 is mandatory thus it does not need to be set explicitly by
the RAPL Interface drivers.
No functional change intended.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Wang Wendy <wendy.wang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
