Snooze is a poll idle state in powernv and pseries platforms. Snooze
has a timeout so that if a CPU stays in snooze for more than target
residency of the next available idle state, then it would exit
thereby giving chance to the cpuidle governor to re-evaluate and
promote the CPU to a deeper idle state. Therefore whenever snooze
exits due to this timeout, its last_residency will be target_residency
of the next deeper state.
Commit e93e59ce5b "cpuidle: Replace ktime_get() with local_clock()"
changed the math around last_residency calculation. Specifically,
while converting last_residency value from nano- to microseconds, it
carries out right shift by 10. Because of that, in snooze timeout
exit scenarios last_residency calculated is roughly 2.3% less than
target_residency of the next available state. This pattern is picked
up by get_typical_interval() in the menu governor and therefore
expected_interval in menu_select() is frequently less than the
target_residency of any state other than snooze.
Due to this we are entering snooze at a higher rate, thereby
affecting the single thread performance.
Fix this by using more precise division via ktime_us_delta().
Fixes: e93e59ce5b "cpuidle: Replace ktime_get() with local_clock()"
Reported-by: Anton Blanchard <anton@samba.org>
Bisected-by: Shilpasri G Bhat <shilpa.bhat@linux.vnet.ibm.com>
Signed-off-by: Shreyas B. Prabhu <shreyas@linux.vnet.ibm.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Acked-by: Balbir Singh <bsingharora@gmail.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Commit 0b89e9aa28 (cpuidle: delay enabling interrupts until all
coupled CPUs leave idle) rightfully fixed a regression by letting
the coupled idle state framework to handle local interrupt enabling
when the CPU is exiting an idle state.
The current code checks if the idle state is coupled and, if so, it
will let the coupled code to enable interrupts. This way, it can
decrement the ready-count before handling the interrupt. This
mechanism prevents the other CPUs from waiting for a CPU which is
handling interrupts.
But the check is done against the state index returned by the back
end driver's ->enter functions which could be different from the
initial index passed as parameter to the cpuidle_enter_state()
function.
entered_state = target_state->enter(dev, drv, index);
[ ... ]
if (!cpuidle_state_is_coupled(drv, entered_state))
local_irq_enable();
[ ... ]
If the 'index' is referring to a coupled idle state but the
'entered_state' is *not* coupled, then the interrupts are enabled
again. All CPUs blocked on the sync barrier may busy loop longer
if the CPU has interrupts to handle before decrementing the
ready-count. That's consuming more energy than saving.
Fixes: 0b89e9aa28 (cpuidle: delay enabling interrupts until all coupled CPUs leave idle)
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: 3.15+ <stable@vger.kernel.org> # 3.15+
[ rjw: Subject & changelog ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The ktime_get() can have a non negligeable overhead, use local_clock()
instead.
In order to test the difference between ktime_get() and local_clock(),
a quick hack has been added to trigger, via debugfs, 10000 times a
call to ktime_get() and local_clock() and measure the elapsed time.
Then the average value, the min and max is computed for each call.
From userspace, the test above was called 100 times every 2 seconds.
So, ktime_get() and local_clock() have been called 1000000 times in
total.
The results are:
ktime_get():
============
* average: 101 ns (stddev: 27.4)
* maximum: 38313 ns
* minimum: 65 ns
local_clock():
==============
* average: 60 ns (stddev: 9.8)
* maximum: 13487 ns
* minimum: 46 ns
The local_clock() is faster and more stable.
Even if it is a drop in the ocean, changing the ktime_get() by the
local_clock() allows to save 80ns at idle time (entry + exit). And
in some circumstances, especially when there are several CPUs racing
for the clock access, we save tens of microseconds.
The idle duration resulting from a diff is converted from nanosec to
microsec. This could be done with integer division (div 1000) - which is
an expensive operation or by 10 bits shifting (div 1024) - which is fast
but unprecise.
The following table gives some results at the limits.
------------------------------------------
| nsec | div(1000) | div(1024) |
------------------------------------------
| 1e3 | 1 usec | 976 nsec |
------------------------------------------
| 1e6 | 1000 usec | 976 usec |
------------------------------------------
| 1e9 | 1000000 usec | 976562 usec |
------------------------------------------
There is a linear deviation of 2.34%. This loss of precision is acceptable
in the context of the resulting diff which is used for statistics. These
ones are processed to guess estimate an approximation of the duration of the
next idle period which ends up into an idle state selection. The selection
criteria takes into account the next duration based on large intervals,
represented by the idle state's target residency.
The 2^10 division is enough because the approximation regarding the 1e3
division is lost in all the approximations done for the next idle duration
computation.
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
[ rjw: Subject ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Currently the 'registered' member of the cpuidle_device struct is set
to 1 during cpuidle_register_device. In this same function there are
checks to see if the device is already registered to prevent duplicate
calls to register the device, but this value is never set to 0 even on
unregister of the device. Because of this, any attempt to call
cpuidle_register_device after a call to cpuidle_unregister_device will
fail which shouldn't be the case.
To prevent this, set registered to 0 when the device is unregistered.
Fixes: c878a52d3c (cpuidle: Check if device is already registered)
Signed-off-by: Dave Gerlach <d-gerlach@ti.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: All applicable <stable@vger.kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Commit 51164251f5 "sched / idle: Drop default_idle_call() fallback
from call_cpuidle()" made find_deepest_state() return non-negative
value and check all the states with index > 0. Also as a result,
find_deepest_state() returns 0 even when enter_freeze callbacks are not
implemented and enter_freeze_proper() is called which ends up crashing
the kernel.
This patch updates the check for index > 0 in cpuidle_enter_freeze and
cpuidle_idle_call(when idle_should_freeze is true) to restore the
suspend-to-idle functionality in absence of enter_freeze callback.
Fixes: 51164251f5 "sched / idle: Drop default_idle_call() fallback from call_cpuidle()"
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
After commit 9c4b2867ed (cpuidle: menu: Fix menu_select() for
CPUIDLE_DRIVER_STATE_START == 0) it is clear that menu_select()
cannot return negative values. Moreover, ladder_select_state()
will never return a negative value too, so make find_deepest_state()
return non-negative values too and drop the default_idle_call()
fallback from call_cpuidle().
This eliminates one branch from the idle loop and makes the governors
and find_deepest_state() handle the case when all states have been
disabled from sysfs consistently.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Pull power management and ACPI updates from Rafael Wysocki:
"From the number of commits perspective, the biggest items are ACPICA
and cpufreq changes with the latter taking the lead (over 50 commits).
On the cpufreq front, there are many cleanups and minor fixes in the
core and governors, driver updates etc. We also have a new cpufreq
driver for Mediatek MT8173 chips.
ACPICA mostly updates its debug infrastructure and adds a number of
fixes and cleanups for a good measure.
The Operating Performance Points (OPP) framework is updated with new
DT bindings and support for them among other things.
We have a few updates of the generic power domains framework and a
reorganization of the ACPI device enumeration code and bus type
operations.
And a lot of fixes and cleanups all over.
Included is one branch from the MFD tree as it contains some
PM-related driver core and ACPI PM changes a few other commits are
based on.
Specifics:
- ACPICA update to upstream revision 20150818 including method
tracing extensions to allow more in-depth AML debugging in the
kernel and a number of assorted fixes and cleanups (Bob Moore, Lv
Zheng, Markus Elfring).
- ACPI sysfs code updates and a documentation update related to AML
method tracing (Lv Zheng).
- ACPI EC driver fix related to serialized evaluations of _Qxx
methods and ACPI tools updates allowing the EC userspace tool to be
built from the kernel source (Lv Zheng).
- ACPI processor driver updates preparing it for future introduction
of CPPC support and ACPI PCC mailbox driver updates (Ashwin
Chaugule).
- ACPI interrupts enumeration fix for a regression related to the
handling of IRQ attribute conflicts between MADT and the ACPI
namespace (Jiang Liu).
- Fixes related to ACPI device PM (Mika Westerberg, Srinidhi
Kasagar).
- ACPI device registration code reorganization to separate the
sysfs-related code and bus type operations from the rest (Rafael J
Wysocki).
- Assorted cleanups in the ACPI core (Jarkko Nikula, Mathias Krause,
Andy Shevchenko, Rafael J Wysocki, Nicolas Iooss).
- ACPI cpufreq driver and ia64 cpufreq driver fixes and cleanups (Pan
Xinhui, Rafael J Wysocki).
- cpufreq core cleanups on top of the previous changes allowing it to
preseve its sysfs directories over system suspend/resume (Viresh
Kumar, Rafael J Wysocki, Sebastian Andrzej Siewior).
- cpufreq fixes and cleanups related to governors (Viresh Kumar).
- cpufreq updates (core and the cpufreq-dt driver) related to the
turbo/boost mode support (Viresh Kumar, Bartlomiej Zolnierkiewicz).
- New DT bindings for Operating Performance Points (OPP), support for
them in the OPP framework and in the cpufreq-dt driver plus related
OPP framework fixes and cleanups (Viresh Kumar).
- cpufreq powernv driver updates (Shilpasri G Bhat).
- New cpufreq driver for Mediatek MT8173 (Pi-Cheng Chen).
- Assorted cpufreq driver (speedstep-lib, sfi, integrator) cleanups
and fixes (Abhilash Jindal, Andrzej Hajda, Cristian Ardelean).
- intel_pstate driver updates including Skylake-S support, support
for enabling HW P-states per CPU and an additional vendor bypass
list entry (Kristen Carlson Accardi, Chen Yu, Ethan Zhao).
- cpuidle core fixes related to the handling of coupled idle states
(Xunlei Pang).
- intel_idle driver updates including Skylake Client support and
support for freeze-mode-specific idle states (Len Brown).
- Driver core updates related to power management (Andy Shevchenko,
Rafael J Wysocki).
- Generic power domains framework fixes and cleanups (Jon Hunter,
Geert Uytterhoeven, Rajendra Nayak, Ulf Hansson).
- Device PM QoS framework update to allow the latency tolerance
setting to be exposed to user space via sysfs (Mika Westerberg).
- devfreq support for PPMUv2 in Exynos5433 and a fix for an incorrect
exynos-ppmu DT binding (Chanwoo Choi, Javier Martinez Canillas).
- System sleep support updates (Alan Stern, Len Brown, SungEun Kim).
- rockchip-io AVS support updates (Heiko Stuebner).
- PM core clocks support fixup (Colin Ian King).
- Power capping RAPL driver update including support for Skylake H/S
and Broadwell-H (Radivoje Jovanovic, Seiichi Ikarashi).
- Generic device properties framework fixes related to the handling
of static (driver-provided) property sets (Andy Shevchenko).
- turbostat and cpupower updates (Len Brown, Shilpasri G Bhat,
Shreyas B Prabhu)"
* tag 'pm+acpi-4.3-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (180 commits)
cpufreq: speedstep-lib: Use monotonic clock
cpufreq: powernv: Increase the verbosity of OCC console messages
cpufreq: sfi: use kmemdup rather than duplicating its implementation
cpufreq: drop !cpufreq_driver check from cpufreq_parse_governor()
cpufreq: rename cpufreq_real_policy as cpufreq_user_policy
cpufreq: remove redundant 'policy' field from user_policy
cpufreq: remove redundant 'governor' field from user_policy
cpufreq: update user_policy.* on success
cpufreq: use memcpy() to copy policy
cpufreq: remove redundant CPUFREQ_INCOMPATIBLE notifier event
cpufreq: mediatek: Add MT8173 cpufreq driver
dt-bindings: mediatek: Add MT8173 CPU DVFS clock bindings
PM / Domains: Fix typo in description of genpd_dev_pm_detach()
PM / Domains: Remove unusable governor dummies
PM / Domains: Make pm_genpd_init() available to modules
PM / domains: Align column headers and data in pm_genpd_summary output
powercap / RAPL: disable the 2nd power limit properly
tools: cpupower: Fix error when running cpupower monitor
PM / OPP: Drop unlikely before IS_ERR(_OR_NULL)
PM / OPP: Fix static checker warning (broken 64bit big endian systems)
...
Pull scheduler updates from Ingo Molnar:
"The biggest change in this cycle is the rewrite of the main SMP load
balancing metric: the CPU load/utilization. The main goal was to make
the metric more precise and more representative - see the changelog of
this commit for the gory details:
9d89c257df ("sched/fair: Rewrite runnable load and utilization average tracking")
It is done in a way that significantly reduces complexity of the code:
5 files changed, 249 insertions(+), 494 deletions(-)
and the performance testing results are encouraging. Nevertheless we
need to keep an eye on potential regressions, since this potentially
affects every SMP workload in existence.
This work comes from Yuyang Du.
Other changes:
- SCHED_DL updates. (Andrea Parri)
- Simplify architecture callbacks by removing finish_arch_switch().
(Peter Zijlstra et al)
- cputime accounting: guarantee stime + utime == rtime. (Peter
Zijlstra)
- optimize idle CPU wakeups some more - inspired by Facebook server
loads. (Mike Galbraith)
- stop_machine fixes and updates. (Oleg Nesterov)
- Introduce the 'trace_sched_waking' tracepoint. (Peter Zijlstra)
- sched/numa tweaks. (Srikar Dronamraju)
- misc fixes and small cleanups"
* 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (44 commits)
sched/deadline: Fix comment in enqueue_task_dl()
sched/deadline: Fix comment in push_dl_tasks()
sched: Change the sched_class::set_cpus_allowed() calling context
sched: Make sched_class::set_cpus_allowed() unconditional
sched: Fix a race between __kthread_bind() and sched_setaffinity()
sched: Ensure a task has a non-normalized vruntime when returning back to CFS
sched/numa: Fix NUMA_DIRECT topology identification
tile: Reorganize _switch_to()
sched, sparc32: Update scheduler comments in copy_thread()
sched: Remove finish_arch_switch()
sched, tile: Remove finish_arch_switch
sched, sh: Fold finish_arch_switch() into switch_to()
sched, score: Remove finish_arch_switch()
sched, avr32: Remove finish_arch_switch()
sched, MIPS: Get rid of finish_arch_switch()
sched, arm: Remove finish_arch_switch()
sched/fair: Clean up load average references
sched/fair: Provide runnable_load_avg back to cfs_rq
sched/fair: Remove task and group entity load when they are dead
sched/fair: Init cfs_rq's sched_entity load average
...
Put tick_freeze() under RCU_NONIDLE() to prevent RCU from complaining
about suspicious RCU usage in idle by trace_suspend_resume() called
from there.
While at it, fix a comment related to another usage of RCU_NONIDLE()
in enter_freeze_proper().
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
* pm-sleep:
PM / sleep: trace_device_pm_callback coverage in dpm_prepare/complete
PM / wakeup: add a dummy wakeup_source to record statistics
PM / sleep: Make suspend-to-idle-specific code depend on CONFIG_SUSPEND
PM / sleep: Return -EBUSY from suspend_enter() on wakeup detection
PM / tick: Add tracepoints for suspend-to-idle diagnostics
PM / sleep: Fix symbol name in a comment in kernel/power/main.c
leds / PM: fix hibernation on arm when gpio-led used with CPU led trigger
ARM: omap-device: use SET_NOIRQ_SYSTEM_SLEEP_PM_OPS
bus: omap_l3_noc: add missed callbacks for suspend-to-disk
PM / sleep: Add macro to define common noirq system PM callbacks
PM / sleep: Refine diagnostic messages in enter_state()
PM / wakeup: validate wakeup source before activating it.
* pm-runtime:
PM / Runtime: Update last_busy in rpm_resume
PM / runtime: add note about re-calling in during device probe()
The CPUIDLE_DRIVER_STATE_START symbol is defined as 1 only if
CONFIG_ARCH_HAS_CPU_RELAX is set, otherwise it is defined as 0.
However, if CONFIG_ARCH_HAS_CPU_RELAX is set, the first (index 0)
entry in the cpuidle driver's table of states is overwritten with
the default "poll" entry by the core. The "state" defined by the
"poll" entry doesn't provide ->enter_dead and ->enter_freeze
callbacks and its exit_latency is 0.
For this reason, it is not necessary to use CPUIDLE_DRIVER_STATE_START
in cpuidle_play_dead() (->enter_dead is NULL, so the "poll state"
will be skipped by the loop).
It also is arguably unuseful to return states with exit_latency
equal to 0 from find_deepest_state(), so the function can be modified
to start the loop from index 0 and the "poll state" will be skipped by
it as a result of the check against latency_req.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Since idle_should_freeze() is defined to always return 'false'
for CONFIG_SUSPEND unset, all of the code depending on it in
cpuidle_idle_call() is not necessary in that case.
Make that code depend on CONFIG_SUSPEND too to avoid building it
when it is not going to be used.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
If tick_broadcast_enter() fails in cpuidle_enter_state(),
try to find another idle state to enter instead of invoking
default_idle_call() immediately and returning -EBUSY which
should increase the chances of saving some energy in those
cases.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Tested-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Kevin Hilman <khilman@linaro.org>
The check of the cpuidle_enter() return value against -EBUSY
made in call_cpuidle() will not be necessary any more if
cpuidle_enter_state() calls default_idle_call() directly when it
is about to return -EBUSY, so make that happen and eliminate the
check.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Tested-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Kevin Hilman <khilman@linaro.org>
Introduce a wrapper function around idle_set_state() called
sched_idle_set_state() that will pass this_rq() to it as the
first argument and make cpuidle_enter_state() call the new
function before and after entering the target state.
At the same time, remove direct invocations of idle_set_state()
from call_cpuidle().
This will allow the invocation of default_idle_call() to be
moved from call_cpuidle() to cpuidle_enter_state() safely
and call_cpuidle() to be simplified a bit as a result.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Tested-by: Preeti U Murthy <preeti@linux.vnet.ibm.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Kevin Hilman <khilman@linaro.org>
The kerneldoc comment for cpuidle_enter_state() doesn't match the
function's header any more, so fix it.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Avoid calling the governor's ->reflect method if the state index
passed to cpuidle_reflect() is negative.
This allows the analogous check to be dropped from menu_reflect(),
so do that too, and ensures that arbitrary error codes can be
passed to cpuidle_reflect() as the index with no adverse
consequences.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Commit 335f49196f (sched/idle: Use explicit broadcast oneshot
control function) replaced clockevents_notify() invocations in
cpuidle_idle_call() with direct calls to tick_broadcast_enter()
and tick_broadcast_exit(), but it overlooked the fact that
interrupts were already enabled before calling the latter which
led to functional breakage on systems using idle states with the
CPUIDLE_FLAG_TIMER_STOP flag set.
Fix that by moving the invocations of tick_broadcast_enter()
and tick_broadcast_exit() down into cpuidle_enter_state() where
interrupts are still disabled when tick_broadcast_exit() is
called. Also ensure that interrupts will be disabled before
running tick_broadcast_exit() even if they have been enabled by
the idle state's ->enter callback. Trigger a WARN_ON_ONCE() in
that case, as we generally don't want that to happen for states
with CPUIDLE_FLAG_TIMER_STOP set.
Fixes: 335f49196f (sched/idle: Use explicit broadcast oneshot control function)
Reported-and-tested-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Reported-and-tested-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Thomas Schlichter reports the following issue on his Samsung NC20:
"The C-states C1 and C2 to the OS when connected to AC, and additionally
provides the C3 C-state when disconnected from AC. However, the number
of C-states shown in sysfs is fixed to the number of C-states present
at boot.
If I boot with AC connected, I always only see the C-states up to C2
even if I disconnect AC.
The reason is commit 130a5f6924 (ACPI / cpuidle: remove dev->state_count
setting). It removes the update of dev->state_count, but sysfs uses
exactly this variable to show the C-states.
The fix is to use drv->state_count in sysfs. As this is currently the
last user of dev->state_count, this variable can be completely removed."
Remove dev->state_count as per the above.
Reported-by: Thomas Schlichter <thomas.schlichter@web.de>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: 3.14+ <stable@vger.kernel.org> # 3.14+
[ rjw: Changelog ]
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Commit 3810631332 (PM / sleep: Re-implement suspend-to-idle handling)
overlooked the fact that entering some sufficiently deep idle states
by CPUs may cause their local timers to stop and in those cases it
is necessary to switch over to a broadcast timer prior to entering
the idle state. If the cpuidle driver in use does not provide
the new ->enter_freeze callback for any of the idle states, that
problem affects suspend-to-idle too, but it is not taken into account
after the changes made by commit 3810631332.
Fix that by changing the definition of cpuidle_enter_freeze() and
re-arranging of the code in cpuidle_idle_call(), so the former does
not call cpuidle_enter() any more and the fallback case is handled
by cpuidle_idle_call() directly.
Fixes: 3810631332 (PM / sleep: Re-implement suspend-to-idle handling)
Reported-and-tested-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Modify cpuidle_enter_freeze() to do the sanity checks done by
cpuidle_select() to avoid crashing the suspend-to-idle code
path in case something is missing.
Fixes: 3810631332 (PM / sleep: Re-implement suspend-to-idle handling)
Original-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Disabling interrupts at the end of cpuidle_enter_freeze() is not
useful, because its caller, cpuidle_idle_call(), re-enables them
right away after invoking it.
To avoid that unnecessary back and forth dance with interrupts,
make cpuidle_enter_freeze() enable interrupts after calling
enter_freeze_proper() and drop the local_irq_disable() at its
end, so that all of the code paths in it end up with interrupts
enabled. Then, cpuidle_idle_call() will not need to re-enable
interrupts after calling cpuidle_enter_freeze() any more, because
the latter will return with interrupts enabled, in analogy with
cpuidle_enter().
Reported-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
The efficiency of suspend-to-idle depends on being able to keep CPUs
in the deepest available idle states for as much time as possible.
Ideally, they should only be brought out of idle by system wakeup
interrupts.
However, timer interrupts occurring periodically prevent that from
happening and it is not practical to chase all of the "misbehaving"
timers in a whack-a-mole fashion. A much more effective approach is
to suspend the local ticks for all CPUs and the entire timekeeping
along the lines of what is done during full suspend, which also
helps to keep suspend-to-idle and full suspend reasonably similar.
The idea is to suspend the local tick on each CPU executing
cpuidle_enter_freeze() and to make the last of them suspend the
entire timekeeping. That should prevent timer interrupts from
triggering until an IO interrupt wakes up one of the CPUs. It
needs to be done with interrupts disabled on all of the CPUs,
though, because otherwise the suspended clocksource might be
accessed by an interrupt handler which might lead to fatal
consequences.
Unfortunately, the existing ->enter callbacks provided by cpuidle
drivers generally cannot be used for implementing that, because some
of them re-enable interrupts temporarily and some idle entry methods
cause interrupts to be re-enabled automatically on exit. Also some
of these callbacks manipulate local clock event devices of the CPUs
which really shouldn't be done after suspending their ticks.
To overcome that difficulty, introduce a new cpuidle state callback,
->enter_freeze, that will be guaranteed (1) to keep interrupts
disabled all the time (and return with interrupts disabled) and (2)
not to touch the CPU timer devices. Modify cpuidle_enter_freeze() to
look for the deepest available idle state with ->enter_freeze present
and to make the CPU execute that callback with suspended tick (and the
last of the online CPUs to execute it with suspended timekeeping).
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Shreyas B. Prabhu
Daniel Lezcano
Dave Gerlach
Sudeep Holla
Rafael J. Wysocki
Linus Torvalds
Xunlei Pang
Lucas Stach
Bartlomiej Zolnierkiewicz