| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc updates from Michael Ellerman:
- Add HOTPLUG_SMT support (/sys/devices/system/cpu/smt) and honour the
configured SMT state when hotplugging CPUs into the system
- Combine final TLB flush and lazy TLB mm shootdown IPIs when using the
Radix MMU to avoid a broadcast TLBIE flush on exit
- Drop the exclusion between ptrace/perf watchpoints, and drop the now
unused associated arch hooks
- Add support for the "nohlt" command line option to disable CPU idle
- Add support for -fpatchable-function-entry for ftrace, with GCC >=
13.1
- Rework memory block size determination, and support 256MB size on
systems with GPUs that have hotpluggable memory
- Various other small features and fixes
Thanks to Andrew Donnellan, Aneesh Kumar K.V, Arnd Bergmann, Athira
Rajeev, Benjamin Gray, Christophe Leroy, Frederic Barrat, Gautam
Menghani, Geoff Levand, Hari Bathini, Immad Mir, Jialin Zhang, Joel
Stanley, Jordan Niethe, Justin Stitt, Kajol Jain, Kees Cook, Krzysztof
Kozlowski, Laurent Dufour, Liang He, Linus Walleij, Mahesh Salgaonkar,
Masahiro Yamada, Michal Suchanek, Nageswara R Sastry, Nathan Chancellor,
Nathan Lynch, Naveen N Rao, Nicholas Piggin, Nick Desaulniers, Omar
Sandoval, Randy Dunlap, Reza Arbab, Rob Herring, Russell Currey, Sourabh
Jain, Thomas Gleixner, Trevor Woerner, Uwe Kleine-König, Vaibhav Jain,
Xiongfeng Wang, Yuan Tan, Zhang Rui, and Zheng Zengkai.
* tag 'powerpc-6.6-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (135 commits)
macintosh/ams: linux/platform_device.h is needed
powerpc/xmon: Reapply "Relax frame size for clang"
powerpc/mm/book3s64: Use 256M as the upper limit with coherent device memory attached
powerpc/mm/book3s64: Fix build error with SPARSEMEM disabled
powerpc/iommu: Fix notifiers being shared by PCI and VIO buses
powerpc/mpc5xxx: Add missing fwnode_handle_put()
powerpc/config: Disable SLAB_DEBUG_ON in skiroot
powerpc/pseries: Remove unused hcall tracing instruction
powerpc/pseries: Fix hcall tracepoints with JUMP_LABEL=n
powerpc: dts: add missing space before {
powerpc/eeh: Use pci_dev_id() to simplify the code
powerpc/64s: Move CPU -mtune options into Kconfig
powerpc/powermac: Fix unused function warning
powerpc/pseries: Rework lppaca_shared_proc() to avoid DEBUG_PREEMPT
powerpc: Don't include lppaca.h in paca.h
powerpc/pseries: Move hcall_vphn() prototype into vphn.h
powerpc/pseries: Move VPHN constants into vphn.h
cxl: Drop unused detach_spa()
powerpc: Drop zalloc_maybe_bootmem()
powerpc/powernv: Use struct opal_prd_msg in more places
...
|
|
Merge CPU power management updates for 6.6-rc1:
- Rework the menu and teo cpuidle governors to avoid calling
tick_nohz_get_sleep_length(), which is likely to become quite
expensive going forward, too often and improve making decisions
regarding whether or not to stop the scheduler tick in the teo
governor (Rafael Wysocki).
- Improve the performance of cpufreq_stats_create_table() in some
cases (Liao Chang).
- Fix two issues in the amd-pstate-ut cpufreq driver (Swapnil Sapkal).
- Use clamp() helper macro to improve the code readability in
cpufreq_verify_within_limits() (Liao Chang).
- Set stale CPU frequency to minimum in intel_pstate (Doug Smythies).
* pm-cpuidle:
cpuidle: teo: Avoid unnecessary variable assignments
cpuidle: menu: Skip tick_nohz_get_sleep_length() call in some cases
cpuidle: teo: Gather statistics regarding whether or not to stop the tick
cpuidle: teo: Skip tick_nohz_get_sleep_length() call in some cases
cpuidle: teo: Do not call tick_nohz_get_sleep_length() upfront
cpuidle: teo: Drop utilized from struct teo_cpu
cpuidle: teo: Avoid stopping the tick unnecessarily when bailing out
cpuidle: teo: Update idle duration estimate when choosing shallower state
* pm-cpufreq:
cpufreq: amd-pstate-ut: Fix kernel panic when loading the driver
cpufreq: amd-pstate-ut: Remove module parameter access
cpufreq: Use clamp() helper macro to improve the code readability
cpufreq: intel_pstate: set stale CPU frequency to minimum
cpufreq: stats: Improve the performance of cpufreq_stats_create_table()
|
|
lppaca_shared_proc() takes a pointer to the lppaca which is typically
accessed through get_lppaca(). With DEBUG_PREEMPT enabled, this leads
to checking if preemption is enabled, for example:
BUG: using smp_processor_id() in preemptible [00000000] code: grep/10693
caller is lparcfg_data+0x408/0x19a0
CPU: 4 PID: 10693 Comm: grep Not tainted 6.5.0-rc3 #2
Call Trace:
dump_stack_lvl+0x154/0x200 (unreliable)
check_preemption_disabled+0x214/0x220
lparcfg_data+0x408/0x19a0
...
This isn't actually a problem however, as it does not matter which
lppaca is accessed, the shared proc state will be the same.
vcpudispatch_stats_procfs_init() already works around this by disabling
preemption, but the lparcfg code does not, erroring any time
/proc/powerpc/lparcfg is accessed with DEBUG_PREEMPT enabled.
Instead of disabling preemption on the caller side, rework
lppaca_shared_proc() to not take a pointer and instead directly access
the lppaca, bypassing any potential preemption checks.
Fixes: f13c13a00512 ("powerpc: Stop using non-architected shared_proc field in lppaca")
Signed-off-by: Russell Currey <ruscur@russell.cc>
[mpe: Rework to avoid needing a definition in paca.h and lppaca.h]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230823055317.751786-4-mpe@ellerman.id.au
|
|
Notice that it is not necessary to assign tick_intercept_sum in every
iteration of the first loop over idle states in teo_select(), because
the intercept_sum value does not change after the assignment in a
given iteration of the loop, so its value after the last iteration of
the loop can be used for computing the tick_intercept_sum value
directly.
Modify the code accordingly.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Because the cost of calling tick_nohz_get_sleep_length() may increase
in the future, reorder the code in menu_select() so it first uses the
statistics to determine the expected idle duration. If that value is
higher than RESIDENCY_THRESHOLD_NS, tick_nohz_get_sleep_length() will
be called to obtain the time till the closest timer and refine the
idle duration prediction if necessary.
This causes the governor to always take the full overhead of
get_typical_interval() with the assumption that the cost will be
amortized by skipping the tick_nohz_get_sleep_length() call in the
cases when the predicted idle duration is relatively very small.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Doug Smythies <dsmythies@telus.net>
|
|
Currently, if the target residency of the deepest idle state is less than
the tick period length, which is quite likely for HZ=100, and the deepest
idle state is about to be selected by the TEO idle governor, the decision
on whether or not to stop the scheduler tick is based entirely on the
time till the closest timer. This is often insufficient, because timers
may not be in heavy use and there may be a plenty of other CPU wakeup
events between the deepest idle state's target residency and the closest
tick.
Allow the governor to count those events by making the deepest idle
state's bin effectively end at TICK_NSEC and introducing an additional
"bin" for collecting "hit" events (ie. the ones in which the measured
idle duration falls into the same bin as the time till the closest
timer) with idle duration values past TICK_NSEC.
This way the "intercepts" metric for the deepest idle state's bin
becomes nonzero in general, and so it can influence the decision on
whether or not to stop the tick possibly increasing the governor's
accuracy in that respect.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
Tested-by: Anna-Maria Behnsen <anna-maria@linutronix.de>
|
|
Make teo_select() avoid calling tick_nohz_get_sleep_length() if the
candidate idle state to return is state 0 or if state 0 is a polling
one and the target residency of the current candidate one is below
a certain threshold, in which cases it may be assumed that the CPU will
be woken up immediately by a non-timer wakeup source and the timers
are not likely to matter.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
Tested-by: Anna-Maria Behnsen <anna-maria@linutronix.de>
|
|
Because the cost of calling tick_nohz_get_sleep_length() may increase
in the future, reorder the code in teo_select() so it first uses the
statistics to pick up a candidate idle state and applies the utilization
heuristic to it and only then calls tick_nohz_get_sleep_length() to
obtain the sleep length value and refine the selection if necessary.
This change by itself does not cause tick_nohz_get_sleep_length() to
be called less often, but it prepares the code for subsequent changes
that will do so.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
Tested-by: Anna-Maria Behnsen <anna-maria@linutronix.de>
|
|
A switch from OSI to PC mode is only possible if all CPUs other than the
calling one are OFF, either through a call to CPU_OFF or not yet booted.
Currently OSI mode is enabled before power domains are created. In cases
where CPUidle states are not using hierarchical CPU topology the bail out
path tries to switch back to PC mode which gets denied by firmware since
other CPUs are online at this point and creates inconsistent state as
firmware is in OSI mode and Linux in PC mode.
This change moves enabling OSI mode after power domains are created,
this would makes sure that hierarchical CPU topology is used before
switching firmware to OSI mode.
Cc: stable@vger.kernel.org
Fixes: 70c179b49870 ("cpuidle: psci: Allow PM domain to be initialized even if no OSI mode")
Signed-off-by: Maulik Shah <quic_mkshah@quicinc.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
Genpd parent and child domain topology created using dt_idle_pd_init_topology()
needs to be removed during error cases.
Add new helper function dt_idle_pd_remove_topology() for same.
Cc: stable@vger.kernel.org
Reviewed-by: Ulf Hanssson <ulf.hansson@linaro.org>
Signed-off-by: Maulik Shah <quic_mkshah@quicinc.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
|
|
Because the utilized field in struct teo_cpu is only used locally in
teo_select(), replace it with a local variable in that function.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-and-tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
|
|
When teo_select() is going to return early in some special cases, make
it avoid stopping the tick if the idle state to be returned is shallow.
In particular, never stop the tick if state 0 is to be returned.
Link: https://lore.kernel.org/linux-pm/CAJZ5v0jJxHj65r2HXBTd3wfbZtsg=_StzwO1kA5STDnaPe_dWA@mail.gmail.com
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-and-tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
|
|
The TEO governor takes CPU utilization into account by refining idle state
selection when the utilization is above a certain threshold. This is done by
choosing an idle state shallower than the previously selected one.
However, when doing this, the idle duration estimate needs to be
adjusted so as to prevent the scheduler tick from being stopped when the
candidate idle state is shallow, which may lead to excessive energy
usage if the CPU is not woken up quickly enough going forward.
Moreover, if the scheduler tick has been stopped already and the new
idle duration estimate is too small, the replacement candidate state
cannot be used.
Modify the relevant code to take the above observations into account.
Fixes: 9ce0f7c4bc64 ("cpuidle: teo: Introduce util-awareness")
Link: https://lore.kernel.org/linux-pm/CAJZ5v0jJxHj65r2HXBTd3wfbZtsg=_StzwO1kA5STDnaPe_dWA@mail.gmail.com
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-and-tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
|
|
With the introduction of local_clock_noinstr(), local_clock() itself
is no longer marked noinstr, use the correct function.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Tested-by: Michael Kelley <mikelley@microsoft.com> # Hyper-V
Link: https://lore.kernel.org/r/20230519102716.045980863@infradead.org
|
|
sbi_probe_extension() is specified with "Returns 0 if the given SBI
extension ID (EID) is not available, or 1 if it is available unless
defined as any other non-zero value by the implementation."
Additionally, sbiret.value is a long. Fix the implementation to
ensure any nonzero long value is considered a success, rather
than only positive int values.
Fixes: b9dcd9e41587 ("RISC-V: Add basic support for SBI v0.2")
Signed-off-by: Andrew Jones <ajones@ventanamicro.com>
Reviewed-by: Conor Dooley <conor.dooley@microchip.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20230427163626.101042-1-ajones@ventanamicro.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux
Pull powerpc updates from Michael Ellerman:
- Add support for building the kernel using PC-relative addressing on
Power10.
- Allow HV KVM guests on Power10 to use prefixed instructions.
- Unify support for the P2020 CPU (85xx) into a single machine
description.
- Always build the 64-bit kernel with 128-bit long double.
- Drop support for several obsolete 2000's era development boards as
identified by Paul Gortmaker.
- A series fixing VFIO on Power since some generic changes.
- Various other small features and fixes.
Thanks to Alexey Kardashevskiy, Andrew Donnellan, Benjamin Gray, Bo Liu,
Christophe Leroy, Dan Carpenter, David Binderman, Ira Weiny, Joel
Stanley, Kajol Jain, Kautuk Consul, Liang He, Luis Chamberlain, Masahiro
Yamada, Michael Neuling, Nathan Chancellor, Nathan Lynch, Nicholas
Miehlbradt, Nicholas Piggin, Nick Desaulniers, Nysal Jan K.A, Pali
Rohár, Paul Gortmaker, Paul Mackerras, Petr Vaněk, Randy Dunlap, Rob
Herring, Sachin Sant, Sean Christopherson, Segher Boessenkool, and
Timothy Pearson.
* tag 'powerpc-6.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (156 commits)
powerpc/64s: Disable pcrel code model on Clang
powerpc: Fix merge conflict between pcrel and copy_thread changes
powerpc/configs/powernv: Add IGB=y
powerpc/configs/64s: Drop JFS Filesystem
powerpc/configs/64s: Use EXT4 to mount EXT2 filesystems
powerpc/configs: Make pseries_defconfig an alias for ppc64le_guest
powerpc/configs: Make pseries_le an alias for ppc64le_guest
powerpc/configs: Incorporate generic kvm_guest.config into guest configs
powerpc/configs: Add IBMVETH=y and IBMVNIC=y to guest configs
powerpc/configs/64s: Enable Device Mapper options
powerpc/configs/64s: Enable PSTORE
powerpc/configs/64s: Enable VLAN support
powerpc/configs/64s: Enable BLK_DEV_NVME
powerpc/configs/64s: Drop REISERFS
powerpc/configs/64s: Use SHA512 for module signatures
powerpc/configs/64s: Enable IO_STRICT_DEVMEM
powerpc/configs/64s: Enable SCHEDSTATS
powerpc/configs/64s: Enable DEBUG_VM & other options
powerpc/configs/64s: Enable EMULATED_STATS
powerpc/configs/64s: Enable KUNIT and most tests
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core
Pull driver core updates from Greg KH:
"Here is the large set of driver core changes for 6.4-rc1.
Once again, a busy development cycle, with lots of changes happening
in the driver core in the quest to be able to move "struct bus" and
"struct class" into read-only memory, a task now complete with these
changes.
This will make the future rust interactions with the driver core more
"provably correct" as well as providing more obvious lifetime rules
for all busses and classes in the kernel.
The changes required for this did touch many individual classes and
busses as many callbacks were changed to take const * parameters
instead. All of these changes have been submitted to the various
subsystem maintainers, giving them plenty of time to review, and most
of them actually did so.
Other than those changes, included in here are a small set of other
things:
- kobject logging improvements
- cacheinfo improvements and updates
- obligatory fw_devlink updates and fixes
- documentation updates
- device property cleanups and const * changes
- firwmare loader dependency fixes.
All of these have been in linux-next for a while with no reported
problems"
* tag 'driver-core-6.4-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (120 commits)
device property: make device_property functions take const device *
driver core: update comments in device_rename()
driver core: Don't require dynamic_debug for initcall_debug probe timing
firmware_loader: rework crypto dependencies
firmware_loader: Strip off \n from customized path
zram: fix up permission for the hot_add sysfs file
cacheinfo: Add use_arch[|_cache]_info field/function
arch_topology: Remove early cacheinfo error message if -ENOENT
cacheinfo: Check cache properties are present in DT
cacheinfo: Check sib_leaf in cache_leaves_are_shared()
cacheinfo: Allow early level detection when DT/ACPI info is missing/broken
cacheinfo: Add arm64 early level initializer implementation
cacheinfo: Add arch specific early level initializer
tty: make tty_class a static const structure
driver core: class: remove struct class_interface * from callbacks
driver core: class: mark the struct class in struct class_interface constant
driver core: class: make class_register() take a const *
driver core: class: mark class_release() as taking a const *
driver core: remove incorrect comment for device_create*
MIPS: vpe-cmp: remove module owner pointer from struct class usage.
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux
Pull more devicetree updates from Rob Herring:
- First part of DT header detangling dropping cpu.h from of_device.h
and replacing some includes with forward declarations. A handful of
drivers needed some adjustment to their includes as a result.
- Refactor of_device.h to be used by bus drivers rather than various
device drivers. This moves non-bus related functions out of
of_device.h. The end goal is for of_platform.h and of_device.h to
stop including each other.
- Refactor open coded parsing of "ranges" in some bus drivers to use DT
address parsing functions
- Add some new address parsing functions of_property_read_reg(),
of_range_count(), and of_range_to_resource() in preparation to
convert more open coded parsing of DT addresses to use them.
- Treewide clean-ups to use of_property_read_bool() and
of_property_present() as appropriate. The ones here are the ones that
didn't get picked up elsewhere.
* tag 'devicetree-for-6.4-2' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux: (34 commits)
bus: tegra-gmi: Replace of_platform.h with explicit includes
hte: Use of_property_present() for testing DT property presence
w1: w1-gpio: Use of_property_read_bool() for boolean properties
virt: fsl: Use of_property_present() for testing DT property presence
soc: fsl: Use of_property_present() for testing DT property presence
sbus: display7seg: Use of_property_read_bool() for boolean properties
sparc: Use of_property_read_bool() for boolean properties
sparc: Use of_property_present() for testing DT property presence
bus: mvebu-mbus: Remove open coded "ranges" parsing
of/address: Add of_property_read_reg() helper
of/address: Add of_range_count() helper
of/address: Add support for 3 address cell bus
of/address: Add of_range_to_resource() helper
of: unittest: Add bus address range parsing tests
of: Drop cpu.h include from of_device.h
OPP: Adjust includes to remove of_device.h
irqchip: loongson-eiointc: Add explicit include for cpuhotplug.h
cpuidle: Adjust includes to remove of_device.h
cpufreq: sun50i: Add explicit include for cpu.h
cpufreq: Adjust includes to remove of_device.h
...
|
|
Code in the idle path is not allowed to be instrumented because RCU is
disabled, see commit 0e985e9d2286 ("cpuidle: Add comments about
noinstr/__cpuidle usage").
Mark the cpuidle ->enter() callbacks as __cpuidle and use the
raw_local_irq_*() routines to ensure that is the case.
Reported-by: Sachin Sant <sachinp@linux.ibm.com>
Link: https://lore.kernel.org/all/4C073F6A-C812-4C4A-BB7A-ECD10B75FB88@linux.ibm.com/
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Tested-by: Sachin Sant <sachinp@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20230406144535.3786008-3-mpe@ellerman.id.au
|
|
Now that of_cpu_device_node_get() is defined in of.h, of_device.h is just
implicitly including other includes, and is no longer needed. Adjust the
include files with what was implicitly included by of_device.h (cpu.h,
cpuhotplug.h, of.h, and of_platform.h) and drop including of_device.h.
Acked-by: Anup Patel <anup@brainfault.org>
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Acked-by: Sudeep Holla <sudeep.holla@arm.com>
Link: https://lore.kernel.org/r/20230329-dt-cpu-header-cleanups-v1-16-581e2605fe47@kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
|
|
We need the fixes in here for testing, as well as the driver core
changes for documentation updates to build on.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
It is preferred to use typed property access functions (i.e.
of_property_read_<type> functions) rather than low-level
of_get_property/of_find_property functions for reading properties. As
part of this, convert of_get_property/of_find_property calls to the
recently added of_property_present() helper when we just want to test
for presence of a property and nothing more.
Signed-off-by: Rob Herring <robh@kernel.org>
Acked-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Direct access to the struct bus_type dev_root pointer is going away soon
so replace that with a call to bus_get_dev_root() instead, which is what
it is there for.
This allows us to clean up the cpuidle_add_interface() call a bit as it
was only called in one place, with the same argument so just put that
into the function itself. Note that cpuidle_remove_interface() should
also probably be removed in the future as there are no callers of it for
some reason.
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: linux-pm@vger.kernel.org
Acked-by: Rafael J. Wysocki <rafael@kernel.org>
Link: https://lore.kernel.org/r/20230322090557.2943479-1-gregkh@linuxfoundation.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
In case that psci_pd_init_topology() fails for some reason,
psci_pd_remove() will be responsible for deleting provider and removing
genpd from psci_pd_providers list. There will be a failure when removing
the cluster PD, because the cpu (child) PDs haven't been removed.
[ 0.050232] CPUidle PSCI: init PM domain cpu0
[ 0.050278] CPUidle PSCI: init PM domain cpu1
[ 0.050329] CPUidle PSCI: init PM domain cpu2
[ 0.050370] CPUidle PSCI: init PM domain cpu3
[ 0.050422] CPUidle PSCI: init PM domain cpu-cluster0
[ 0.050475] PM: genpd_remove: unable to remove cpu-cluster0
[ 0.051412] PM: genpd_remove: removed cpu3
[ 0.051449] PM: genpd_remove: removed cpu2
[ 0.051499] PM: genpd_remove: removed cpu1
[ 0.051546] PM: genpd_remove: removed cpu0
Fix the problem by iterating the provider list reversely, so that parent
PD gets removed after child's PDs like below.
[ 0.029052] CPUidle PSCI: init PM domain cpu0
[ 0.029076] CPUidle PSCI: init PM domain cpu1
[ 0.029103] CPUidle PSCI: init PM domain cpu2
[ 0.029124] CPUidle PSCI: init PM domain cpu3
[ 0.029151] CPUidle PSCI: init PM domain cpu-cluster0
[ 0.029647] PM: genpd_remove: removed cpu0
[ 0.029666] PM: genpd_remove: removed cpu1
[ 0.029690] PM: genpd_remove: removed cpu2
[ 0.029714] PM: genpd_remove: removed cpu3
[ 0.029738] PM: genpd_remove: removed cpu-cluster0
Fixes: a65a397f2451 ("cpuidle: psci: Add support for PM domains by using genpd")
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Cc: 5.10+ <stable@vger.kernel.org> # 5.10+
Signed-off-by: Rafael J. Wysocki <rjw@rjwysocki.net>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull ARM SoC driver updates from Arnd Bergmann:
"As usual, there are lots of minor driver changes across SoC platforms
from NXP, Amlogic, AMD Zynq, Mediatek, Qualcomm, Apple and Samsung.
These usually add support for additional chip variations in existing
drivers, but also add features or bugfixes.
The SCMI firmware subsystem gains a unified raw userspace interface
through debugfs, which can be used for validation purposes.
Newly added drivers include:
- New power management drivers for StarFive JH7110, Allwinner D1 and
Renesas RZ/V2M
- A driver for Qualcomm battery and power supply status
- A SoC device driver for identifying Nuvoton WPCM450 chips
- A regulator coupler driver for Mediatek MT81xxv"
* tag 'soc-drivers-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (165 commits)
power: supply: Introduce Qualcomm PMIC GLINK power supply
soc: apple: rtkit: Do not copy the reg state structure to the stack
soc: sunxi: SUN20I_PPU should depend on PM
memory: renesas-rpc-if: Remove redundant division of dummy
soc: qcom: socinfo: Add IDs for IPQ5332 and its variant
dt-bindings: arm: qcom,ids: Add IDs for IPQ5332 and its variant
dt-bindings: power: qcom,rpmpd: add RPMH_REGULATOR_LEVEL_LOW_SVS_L1
firmware: qcom_scm: Move qcom_scm.h to include/linux/firmware/qcom/
MAINTAINERS: Update qcom CPR maintainer entry
dt-bindings: firmware: document Qualcomm SM8550 SCM
dt-bindings: firmware: qcom,scm: add qcom,scm-sa8775p compatible
soc: qcom: socinfo: Add Soc IDs for IPQ8064 and variants
dt-bindings: arm: qcom,ids: Add Soc IDs for IPQ8064 and variants
soc: qcom: socinfo: Add support for new field in revision 17
soc: qcom: smd-rpm: Add IPQ9574 compatible
soc: qcom: pmic_glink: remove redundant calculation of svid
soc: qcom: stats: Populate all subsystem debugfs files
dt-bindings: soc: qcom,rpmh-rsc: Update to allow for generic nodes
soc: qcom: pmic_glink: add CONFIG_NET/CONFIG_OF dependencies
soc: qcom: pmic_glink: Introduce altmode support
...
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull power management updates from Rafael Wysocki:
"These add EPP support to the AMD P-state cpufreq driver, add support
for new platforms to the Intel RAPL power capping driver, intel_idle
and the Qualcomm cpufreq driver, enable thermal cooling for Tegra194,
drop the custom cpufreq driver for loongson1 that is not necessary any
more (and the corresponding cpufreq platform device), fix assorted
issues and clean up code.
Specifics:
- Add EPP support to the AMD P-state cpufreq driver (Perry Yuan, Wyes
Karny, Arnd Bergmann, Bagas Sanjaya)
- Drop the custom cpufreq driver for loongson1 that is not necessary
any more and the corresponding cpufreq platform device (Keguang
Zhang)
- Remove "select SRCU" from system sleep, cpufreq and OPP Kconfig
entries (Paul E. McKenney)
- Enable thermal cooling for Tegra194 (Yi-Wei Wang)
- Register module device table and add missing compatibles for
cpufreq-qcom-hw (Nícolas F. R. A. Prado, Abel Vesa and Luca Weiss)
- Various dt binding updates for qcom-cpufreq-nvmem and
opp-v2-kryo-cpu (Christian Marangi)
- Make kobj_type structure in the cpufreq core constant (Thomas
Weißschuh)
- Make cpufreq_unregister_driver() return void (Uwe Kleine-König)
- Make the TEO cpuidle governor check CPU utilization in order to
refine idle state selection (Kajetan Puchalski)
- Make Kconfig select the haltpoll cpuidle governor when the haltpoll
cpuidle driver is selected and replace a default_idle() call in
that driver with arch_cpu_idle() to allow MWAIT to be used (Li
RongQing)
- Add Emerald Rapids Xeon support to the intel_idle driver (Artem
Bityutskiy)
- Add ARCH_SUSPEND_POSSIBLE dependencies for ARMv4 cpuidle drivers to
avoid randconfig build failures (Arnd Bergmann)
- Make kobj_type structures used in the cpuidle sysfs interface
constant (Thomas Weißschuh)
- Make the cpuidle driver registration code update microsecond values
of idle state parameters in accordance with their nanosecond values
if they are provided (Rafael Wysocki)
- Make the PSCI cpuidle driver prevent topology CPUs from being
suspended on PREEMPT_RT (Krzysztof Kozlowski)
- Document that pm_runtime_force_suspend() cannot be used with
DPM_FLAG_SMART_SUSPEND (Richard Fitzgerald)
- Add EXPORT macros for exporting PM functions from drivers (Richard
Fitzgerald)
- Remove /** from non-kernel-doc comments in hibernation code (Randy
Dunlap)
- Fix possible name leak in powercap_register_zone() (Yang Yingliang)
- Add Meteor Lake and Emerald Rapids support to the intel_rapl power
capping driver (Zhang Rui)
- Modify the idle_inject power capping facility to support 100% idle
injection (Srinivas Pandruvada)
- Fix large time windows handling in the intel_rapl power capping
driver (Zhang Rui)
- Fix memory leaks with using debugfs_lookup() in the generic PM
domains and Energy Model code (Greg Kroah-Hartman)
- Add missing 'cache-unified' property in the example for kryo OPP
bindings (Rob Herring)
- Fix error checking in opp_migrate_dentry() (Qi Zheng)
- Let qcom,opp-fuse-level be a 2-long array for qcom SoCs (Konrad
Dybcio)
- Modify some power management utilities to use the canonical ftrace
path (Ross Zwisler)
- Correct spelling problems for Documentation/power/ as reported by
codespell (Randy Dunlap)"
* tag 'pm-6.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (53 commits)
Documentation: amd-pstate: disambiguate user space sections
cpufreq: amd-pstate: Fix invalid write to MSR_AMD_CPPC_REQ
dt-bindings: opp: opp-v2-kryo-cpu: enlarge opp-supported-hw maximum
dt-bindings: cpufreq: qcom-cpufreq-nvmem: make cpr bindings optional
dt-bindings: cpufreq: qcom-cpufreq-nvmem: specify supported opp tables
PM: Add EXPORT macros for exporting PM functions
cpuidle: psci: Do not suspend topology CPUs on PREEMPT_RT
MIPS: loongson32: Drop obsolete cpufreq platform device
powercap: intel_rapl: Fix handling for large time window
cpuidle: driver: Update microsecond values of state parameters as needed
cpuidle: sysfs: make kobj_type structures constant
cpuidle: add ARCH_SUSPEND_POSSIBLE dependencies
PM: EM: fix memory leak with using debugfs_lookup()
PM: domains: fix memory leak with using debugfs_lookup()
cpufreq: Make kobj_type structure constant
cpufreq: davinci: Fix clk use after free
cpufreq: amd-pstate: avoid uninitialized variable use
cpufreq: Make cpufreq_unregister_driver() return void
OPP: fix error checking in opp_migrate_dentry()
dt-bindings: cpufreq: cpufreq-qcom-hw: Add SM8550 compatible
...
|
|
The runtime Power Management of CPU topology is not compatible with
PREEMPT_RT:
1. Core cpuidle path disables IRQs.
2. Core cpuidle calls cpuidle-psci.
3. cpuidle-psci in __psci_enter_domain_idle_state() calls
pm_runtime_put_sync_suspend() and pm_runtime_get_sync() which use
spinlocks (which are sleeping on PREEMPT_RT).
Deep sleep modes are not a priority of Realtime kernels because the
latencies might become unpredictable. On the other hand the PSCI CPU
idle power domain is a parent of other devices and power domain
controllers, thus it cannot be simply skipped (e.g. on Qualcomm SM8250).
Disable the idle callbacks in cpuidle-psci and mark the domain as
always on. This is a trade-off between making PREEMPT_RT working and
still having a proper power domain hierarchy in the system.
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Tested-by: Adrien Thierry <athierry@redhat.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
If the cpuidle driver provides the target residency and exit latency in
nanoseconds, the corresponding values in microseconds need to be set to
reflect the provided numbers in order for the sysfs interface to show
them correctly, so make __cpuidle_driver_init() do that.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
|
|
Since commit ee6d3dd4ed48 ("driver core: make kobj_type constant.")
the driver core allows the usage of const struct kobj_type.
Take advantage of this to constify the structure definitions to prevent
modification at runtime.
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Some ARMv4 processors don't support suspend, which leads
to a build failure with the tegra and qualcomm cpuidle driver:
WARNING: unmet direct dependencies detected for ARM_CPU_SUSPEND
Depends on [n]: ARCH_SUSPEND_POSSIBLE [=n]
Selected by [y]:
- ARM_TEGRA_CPUIDLE [=y] && CPU_IDLE [=y] && (ARM [=y] || ARM64) && (ARCH_TEGRA [=n] || COMPILE_TEST [=y]) && !ARM64 && MMU [=y]
arch/arm/kernel/sleep.o: in function `__cpu_suspend':
(.text+0x68): undefined reference to `cpu_sa110_suspend_size'
(.text+0x68): undefined reference to `cpu_fa526_suspend_size'
Add an explicit dependency to make randconfig builds avoid
this combination.
Fixes: faae6c9f2e68 ("cpuidle: tegra: Enable compile testing")
Fixes: a871be6b8eee ("cpuidle: Convert Qualcomm SPM driver to a generic CPUidle driver")
Link: https://lore.kernel.org/all/20211013160125.772873-1-arnd@kernel.org/
Cc: All applicable <stable@vger.kernel.org>
Reviewed-by: Dmitry Osipenko <digetx@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Move include/linux/qcom_scm.h to include/linux/firmware/qcom/qcom_scm.h.
This removes 1 of a few remaining Qualcomm-specific headers into a more
approciate subdirectory under include/.
Suggested-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Elliot Berman <quic_eberman@quicinc.com>
Reviewed-by: Guru Das Srinagesh <quic_gurus@quicinc.com>
Acked-by: Mukesh Ojha <quic_mojha@quicinc.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20230203210956.3580811-1-quic_eberman@quicinc.com
|
|
The instrumentation_begin()/end() annotations in poll_idle() were
complete nonsense. Specifically they caused tracing to happen in the
middle of noinstr code, resulting in RCU splats.
Now that local_clock() is noinstr, mark up the rest and let it rip.
Fixes: 00717eb8c955 ("cpuidle: Annotate poll_idle()")
Reported-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/oe-lkp/202301192148.58ece903-oliver.sang@intel.com
Link: https://lore.kernel.org/r/20230126151323.819534689@infradead.org
|
|
When a KVM guest has MWAIT, mwait_idle() is used as the default idle
function.
However, the cpuidle-haltpoll driver calls default_idle() from
default_enter_idle() directly and that one uses HLT instead of MWAIT,
which may affect performance adversely, because MWAIT is preferred to
HLT as explained by the changelog of commit aebef63cf7ff ("x86: Remove
vendor checks from prefer_mwait_c1_over_halt").
Make default_enter_idle() call arch_cpu_idle(), which can use MWAIT,
instead of default_idle() to address this issue.
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Suggested-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Li RongQing <lirongqing@baidu.com>
[ rjw: Changelog rewrite ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
The recent cpuidle changes started triggering RCU splats on
Juno development boards:
| =============================
| WARNING: suspicious RCU usage
| -----------------------------
| include/trace/events/ipi.h:19 suspicious rcu_dereference_check() usage!
Fix cpuidle on ARM64:
- ... by introducing a new 'is_rcu' flag to the cpuidle helpers & make
ARM64 use it, as ARM64 wants to keep RCU active longer and wants to
do the ct_cpuidle_enter()/exit() dance itself.
- Also update the PSCI driver accordingly.
- This also removes the last known RCU_NONIDLE() user as a bonus.
Reported-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Link: https://lore.kernel.org/r/Y8Z31UbzG3LJgAXE@hirez.programming.kicks-ass.net
--
|
|
The added '.flags' value is sometimes ignored here because
it gets overwritten by another initialization:
drivers/cpuidle/cpuidle-mvebu-v7.c:24:33: error: initialized field overwritten [-Werror=override-init]
24 | #define MVEBU_V7_FLAG_DEEP_IDLE 0x10000
| ^~~~~~~
drivers/cpuidle/cpuidle-mvebu-v7.c:69:43: note: in expansion of macro 'MVEBU_V7_FLAG_DEEP_IDLE'
...
Merge the two fields into one.
Fixes: 4ce40e9dbe83 ("cpuidle, armada: Push RCU-idle into driver")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20230117164642.1672784-1-arnd@kernel.org
|
|
The haltpoll cpuidle driver should select the haltpoll governor, so
as to ensure that they work together.
Signed-off-by: Li RongQing <lirongqing@baidu.com>
[ rjw: Changelog edits ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Add a few words on noinstr / __cpuidle usage.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230112195542.397238052@infradead.org
|
|
For all cpuidle drivers that use CPUIDLE_FLAG_RCU_IDLE, ensure that
all functions that call ct_cpuidle_enter() are marked __cpuidle.
( due to lack of noinstr validation on these platforms it is entirely
possible this isn't complete )
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230112195542.274096325@infradead.org
|
|
Tracing (kprobes included) and other compiler instrumentation relies
on a normal kernel runtime. Therefore all functions that disable RCU
should be noinstr, as should all functions that are called while RCU
is disabled.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20230112195542.212914195@infradead.org
|
|
The __cpuidle functions will become a noinstr class, as such they need
explicit annotations.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195540.312601331@infradead.org
|
|
The whole disable-RCU, enable-IRQS dance is very intricate since
changing IRQ state is traced, which depends on RCU.
Add two helpers for the cpuidle case that mirror the entry code:
ct_cpuidle_enter()
ct_cpuidle_exit()
And fix all the cases where the enter/exit dance was buggy.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195540.130014793@infradead.org
|
|
Doing RCU-idle outside the driver, only to then temporarily enable it
again before going idle is suboptimal.
Notably: this converts all dt_init_idle_driver() and
__CPU_PM_CPU_IDLE_ENTER() users for they are inextrably intertwined.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195540.068981667@infradead.org
|
|
Doing RCU-idle outside the driver, only to then temporarily enable it
again before going idle is suboptimal.
Notably the cpu_pm_*() calls implicitly re-enable RCU for a bit.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.946630819@infradead.org
|
|
Doing RCU-idle outside the driver, only to then temporarily enable it
again, at least twice, before going idle is suboptimal.
Notably once implicitly through the cpu_pm_*() calls and once
explicitly doing ct_irq_*_irqon().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Reviewed-by: Guo Ren <guoren@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.760296658@infradead.org
|
|
Doing RCU-idle outside the driver, only to then temporarily enable it
again, at least twice, before going idle is suboptimal.
Notably once implicitly through the cpu_pm_*() calls and once
explicitly doing RCU_NONIDLE().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.699546331@infradead.org
|
|
Doing RCU-idle outside the driver, only to then temporarily enable it
again, at least twice, before going idle is suboptimal.
That is, once implicitly through the cpu_pm_*() calls and once
explicitly doing ct_irq_*_irqon().
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20230112195539.637185846@infradead.org
|
|
Make cpuidle_enter_state() consistent with the s2idle variant and
verify ->enter() always returns with interrupts disabled.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195539.576412812@infradead.org
|
|
Make cpuidle_state::enter() methods IRQ state invariant on exit.
Additionally make sure to use raw_local_irq_*() methods since this
cpuidle callback will be called with RCU already disabled.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Link: https://lore.kernel.org/r/20230112195539.515253662@infradead.org
|
|
Modern interactive systems, such as recent Android phones, tend to have
power efficient shallow idle states. Selecting deeper idle states on a
device while a latency-sensitive workload is running can adversely
impact performance due to increased latency. Additionally, if the CPU
wakes up from a deeper sleep before its target residency as is often the
case, it results in a waste of energy on top of that.
At the moment, none of the available idle governors take any scheduling
information into account. They also tend to overestimate the idle
duration quite often, which causes them to select excessively deep idle
states, thus leading to increased wakeup latency and lower performance
with no power saving. For 'menu' while web browsing on Android for
instance, those types of wakeups ('too deep') account for over 24% of
all wakeups.
At the same time, on some platforms idle state 0 can be power efficient
enough to warrant wanting to prefer it over idle state 1. This is
because the power usage of the two states can be so close that
sufficient amounts of too deep state 1 sleeps can completely offset the
state 1 power saving to the point where it would've been more power
efficient to just use state 0 instead. This is, of course, for systems
where state 0 is not a polling state, such as arm-based devices.
Sleeps that happened in state 0 while they could have used state 1 ('too
shallow') only save less power than they otherwise could have. Too deep
sleeps, on the other hand, harm performance and nullify the potential
power saving from using state 1 in the first place. While taking this
into account, it is clear that on balance it is preferable for an idle
governor to have more too shallow sleeps instead of more too deep sleeps
on those kinds of platforms.
This patch specifically tunes TEO to prefer shallower idle states in
order to reduce wakeup latency and achieve better performance.
To this end, before selecting the next idle state it uses the avg_util
signal of a CPU's runqueue in order to determine to what extent the CPU
is being utilized. This util value is then compared to a threshold
defined as a percentage of the CPU's capacity (capacity >> 6 ie. ~1.5%
in the current implementation). If the util is above the threshold, the
index of the idle state selected by TEO metrics will be reduced by 1,
thus selecting a shallower state. If the util is below the threshold,
the governor defaults to the TEO metrics mechanism to try to select the
deepest available idle state based on the closest timer event and its
own correctness.
The main goal of this is to reduce latency and increase performance for
some workloads. Under some workloads it will result in an increase in
power usage (Geekbench 5) while for other workloads it will also result
in a decrease in power usage compared to TEO (PCMark Web, Jankbench,
Speedometer).
It can provide drastically decreased latency and performance benefits in
certain types of workloads that are sensitive to latency.
Example test results:
1. GB5 (better score, latency & more power usage)
| metric | menu | teo | teo-util-aware |
| ------------------------------------- | -------------- | ----------------- | ----------------- |
| gmean score | 2826.5 (0.0%) | 2764.8 (-2.18%) | 2865 (1.36%) |
| gmean power usage [mW] | 2551.4 (0.0%) | 2606.8 (2.17%) | 2722.3 (6.7%) |
| gmean too deep % | 14.99% | 9.65% | 4.02% |
| gmean too shallow % | 2.5% | 5.96% | 14.59% |
| gmean task wakeup latency (asynctask) | 78.16μs (0.0%) | 61.60μs (-21.19%) | 54.45μs (-30.34%) |
2. Jankbench (better score, latency & less power usage)
| metric | menu | teo | teo-util-aware |
| ------------------------------------- | -------------- | ----------------- | ----------------- |
| gmean frame duration | 13.9 (0.0%) | 14.7 (6.0%) | 12.6 (-9.0%) |
| gmean jank percentage | 1.5 (0.0%) | 2.1 (36.99%) | 1.3 (-17.37%) |
| gmean power usage [mW] | 144.6 (0.0%) | 136.9 (-5.27%) | 121.3 (-16.08%) |
| gmean too deep % | 26.00% | 11.00% | 2.54% |
| gmean too shallow % | 4.74% | 11.89% | 21.93% |
| gmean wakeup latency (RenderThread) | 139.5μs (0.0%) | 116.5μs (-16.49%) | 91.11μs (-34.7%) |
| gmean wakeup latency (surfaceflinger) | 124.0μs (0.0%) | 151.9μs (22.47%) | 87.65μs (-29.33%) |
Signed-off-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
[ rjw: Comment edits and white space adjustments ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
|
Add a no_poll flag to teo_find_shallower_state() that will let the
function optionally not consider polling states.
This allows the caller to guard against the function inadvertently
resulting in TEO putting the CPU in a polling state when that
behaviour is undesirable.
Signed-off-by: Kajetan Puchalski <kajetan.puchalski@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
|
