Currently intel_idle and acpi_idle driver show double cpu_idle "exit idle"
events -> this patch fixes it and makes cpu_idle events throwing less complex.
It also introduces cpu_idle events for all architectures which use
the cpuidle subsystem, namely:
- arch/arm/mach-at91/cpuidle.c
- arch/arm/mach-davinci/cpuidle.c
- arch/arm/mach-kirkwood/cpuidle.c
- arch/arm/mach-omap2/cpuidle34xx.c
- arch/drivers/acpi/processor_idle.c (for all cases, not only mwait)
- arch/x86/kernel/process.c (did throw events before, but was a mess)
- drivers/idle/intel_idle.c (did throw events before)
Convention should be:
Fire cpu_idle events inside the current pm_idle function (not somewhere
down the the callee tree) to keep things easy.
Current possible pm_idle functions in X86:
c1e_idle, poll_idle, cpuidle_idle_call, mwait_idle, default_idle
-> this is really easy is now.
This affects userspace:
The type field of the cpu_idle power event can now direclty get
mapped to:
/sys/devices/system/cpu/cpuX/cpuidle/stateX/{name,desc,usage,time,...}
instead of throwing very CPU/mwait specific values.
This change is not visible for the intel_idle driver.
For the acpi_idle driver it should only be visible if the vendor
misses out C-states in his BIOS.
Another (perf timechart) patch reads out cpuidle info of cpu_idle
events from:
/sys/.../cpuidle/stateX/*, then the cpuidle events are mapped
to the correct C-/cpuidle state again, even if e.g. vendors miss
out C-states in their BIOS and for example only export C1 and C3.
-> everything is fine.
Signed-off-by: Thomas Renninger <trenn@suse.de>
CC: Robert Schoene <robert.schoene@tu-dresden.de>
CC: Jean Pihet <j-pihet@ti.com>
CC: Arjan van de Ven <arjan@linux.intel.com>
CC: Ingo Molnar <mingo@elte.hu>
CC: Frederic Weisbecker <fweisbec@gmail.com>
CC: linux-pm@lists.linux-foundation.org
CC: linux-acpi@vger.kernel.org
CC: linux-kernel@vger.kernel.org
CC: linux-perf-users@vger.kernel.org
CC: linux-omap@vger.kernel.org
Signed-off-by: Len Brown <len.brown@intel.com>
C0 means and is well know as "not idle".
All documentation out there uses this term as "running"/"not idle"
state. Also Linux userspace tools (e.g. cpufreq-aperf and turbostat)
show C0 residency which there is correct, but means something totally
else than cpuidle "POLL" state.
Signed-off-by: Thomas Renninger <trenn@suse.de>
Signed-off-by: Len Brown <len.brown@intel.com>
The following scenario is possible with the current cpuidle code and
the ACPI cpuidle driver:
(1) acpi_processor_cst_has_changed() is called,
(2) cpuidle_disable_device() is called,
(3) cpuidle_remove_state_sysfs() is called to remove the (presumably
outdated) states info from sysfs,
(3) acpi_processor_get_power_info() is called, the first entry in the
pr->power.states[] table is filled with zeros,
(4) acpi_processor_setup_cpuidle() is called and it doesn't fill the
first entry in pr->power.states[],
(5) cpuidle_enable_device() is called,
(6) __cpuidle_register_device() is _not_ called, since the device has
already been registered,
(7) Consequently, poll_idle_init() is _not_ called either,
(8) cpuidle_add_state_sysfs() is called to create the sysfs attributes
for the new states and it uses the bogus first table entry from
acpi_processor_get_power_info() for creating state0.
This problem is avoided if cpuidle_enable_device()
unconditionally calls poll_idle_init().
Reported-by: Len Brown <len.brown@intel.com>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Len Brown <len.brown@intel.com>
cc: stable@kernel.org
* 'for-2.6.38' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu: (30 commits)
gameport: use this_cpu_read instead of lookup
x86: udelay: Use this_cpu_read to avoid address calculation
x86: Use this_cpu_inc_return for nmi counter
x86: Replace uses of current_cpu_data with this_cpu ops
x86: Use this_cpu_ops to optimize code
vmstat: User per cpu atomics to avoid interrupt disable / enable
irq_work: Use per cpu atomics instead of regular atomics
cpuops: Use cmpxchg for xchg to avoid lock semantics
x86: this_cpu_cmpxchg and this_cpu_xchg operations
percpu: Generic this_cpu_cmpxchg() and this_cpu_xchg support
percpu,x86: relocate this_cpu_add_return() and friends
connector: Use this_cpu operations
xen: Use this_cpu_inc_return
taskstats: Use this_cpu_ops
random: Use this_cpu_inc_return
fs: Use this_cpu_inc_return in buffer.c
highmem: Use this_cpu_xx_return() operations
vmstat: Use this_cpu_inc_return for vm statistics
x86: Support for this_cpu_add, sub, dec, inc_return
percpu: Generic support for this_cpu_add, sub, dec, inc_return
...
Fixed up conflicts: in arch/x86/kernel/{apic/nmi.c, apic/x2apic_uv_x.c, process.c}
as per Tejun.
Add these new power trace events:
power:cpu_idle
power:cpu_frequency
power:machine_suspend
The old C-state/idle accounting events:
power:power_start
power:power_end
Have now a replacement (but we are still keeping the old
tracepoints for compatibility):
power:cpu_idle
and
power:power_frequency
is replaced with:
power:cpu_frequency
power:machine_suspend is newly introduced.
Jean Pihet has a patch integrated into the generic layer
(kernel/power/suspend.c) which will make use of it.
the type= field got removed from both, it was never
used and the type is differed by the event type itself.
perf timechart userspace tool gets adjusted in a separate patch.
Signed-off-by: Thomas Renninger <trenn@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Arjan van de Ven <arjan@linux.intel.com>
Acked-by: Jean Pihet <jean.pihet@newoldbits.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: rjw@sisk.pl
LKML-Reference: <1294073445-14812-3-git-send-email-trenn@suse.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
LKML-Reference: <1290072314-31155-2-git-send-email-trenn@suse.de>
__get_cpu_var() can be replaced with this_cpu_read and will then use a single
read instruction with implied address calculation to access the correct per cpu
instance.
However, the address of a per cpu variable passed to __this_cpu_read() cannot be
determed (since its an implied address conversion through segment prefixes).
Therefore apply this only to uses of __get_cpu_var where the addres of the
variable is not used.
V3->V4:
- Move one instance of this_cpu_inc_return to a later patch
so that this one can go in without percpu infrastructrure
changes.
Sedat: fixed compile failure caused by an extra ')'.
Cc: Neil Horman <nhorman@tuxdriver.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Sedat Dilek <sedat.dilek@gmail.com>
Acked-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
On some SoC chips, HW resources may be in use during any particular idle
period. As a consequence, the cpuidle states that the SoC is safe to
enter can change from idle period to idle period. In addition, the
latency and threshold of each cpuidle state can vary, depending on the
operating condition when the CPU becomes idle, e.g. the current cpu
frequency, the current state of the HW blocks, etc.
cpuidle core and the menu governor, in the current form, are geared
towards cpuidle states that are static, i.e. the availabiltiy of the
states, their latencies, their thresholds are non-changing during run
time. cpuidle does not provide any hook that cpuidle drivers can use to
adjust those values on the fly for the current idle period before the menu
governor selects the target cpuidle state.
This patch extends cpuidle core and the menu governor to handle states
that are dynamic. There are three additions in the patch and the patch
maintains backwards-compatibility with existing cpuidle drivers.
1) add prepare() to struct cpuidle_device. A cpuidle driver can hook
into the callback and cpuidle will call prepare() before calling the
governor's select function. The callback gives the cpuidle driver a
chance to update the dynamic information of the cpuidle states for the
current idle period, e.g. state availability, latencies, thresholds,
power values, etc.
2) add CPUIDLE_FLAG_IGNORE as one of the state flags. In the prepare()
function, a cpuidle driver can set/clear the flag to indicate to the
menu governor whether a cpuidle state should be ignored, i.e. not
available, during the current idle period.
3) add power_specified bit to struct cpuidle_device. The menu governor
currently assumes that the cpuidle states are arranged in the order of
increasing latency, threshold, and power savings. This is true or can
be made true for static states. Once the state parameters are dynamic,
the latencies, thresholds, and power savings for the cpuidle states can
increase or decrease by different amounts from idle period to idle
period. So the assumption of increasing latency, threshold, and power
savings from Cn to C(n+1) can no longer be guaranteed.
It can be straightforward to calculate the power consumption of each
available state and to specify it in power_usage for the idle period.
Using the power_usage fields, the menu governor then selects the state
that has the lowest power consumption and that still satisfies all other
critieria. The power_specified bit defaults to 0. For existing cpuidle
drivers, cpuidle detects that power_specified is 0 and fills in a dummy
set of power_usage values.
Signed-off-by: Ai Li <aili@codeaurora.org>
Cc: Len Brown <len.brown@intel.com>
Acked-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Venkatesh Pallipadi <venki@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
and fix the broken case if a core's frequency depends on others.
trace_power_frequency was only implemented in a rather ungeneric way
in acpi-cpufreq driver's target() function only.
-> Move the call to trace_power_frequency to
cpufreq.c:cpufreq_notify_transition() where CPUFREQ_POSTCHANGE
notifier is triggered.
This will support power frequency tracing by all cpufreq drivers
trace_power_frequency did not trace frequency changes correctly when
the userspace governor was used or when CPU cores' frequency depend
on each other.
-> Moving this into the CPUFREQ_POSTCHANGE notifier and pass the cpu
which gets switched automatically fixes this.
Robert Schoene provided some important fixes on top of my initial
quick shot version which are integrated in this patch:
- Forgot some changes in power_end trace (TP_printk/variable names)
- Variable dummy in power_end must now be cpu_id
- Use static 64 bit variable instead of unsigned int for cpu_id
Signed-off-by: Thomas Renninger <trenn@suse.de>
CC: davej@redhat.com
CC: arjan@infradead.org
CC: linux-kernel@vger.kernel.org
CC: robert.schoene@tu-dresden.de
Tested-by: robert.schoene@tu-dresden.de
Signed-off-by: Dave Jones <davej@redhat.com>
cpuidle_register_driver() sets cpuidle_curr_driver
cpuidle_unregister_driver() clears cpuidle_curr_driver
We should't expose cpuidle_curr_driver to
potential modification except via these interfaces.
So make it static and create cpuidle_get_driver() to observe it.
Signed-off-by: Len Brown <len.brown@intel.com>
In the case where cpuidle_idle_call() returns before changing state due to
a need_resched(), it was returning with IRQs disabled.
The idle path assumes that the platform specific idle code returns with
interrupts enabled (although this too is undocumented AFAICT) and on ARM
we have a WARN_ON(!(irqs_disabled()) when returning from the idle loop, so
the user-visible effects were only a warning since interrupts were
eventually re-enabled later.
On x86, this same problem exists, but there is no WARN_ON() to detect it.
As on ARM, the interrupts are eventually re-enabled, so I'm not sure of
any actual bugs triggered by this. It's primarily a
correctness/consistency fix.
This patch ensures IRQs are (re)enabled before returning.
Reported-by: Hemanth V <hemanthv@ti.com>
Signed-off-by: Kevin Hilman <khilman@deeprootsystems.com>
Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Len Brown <len.brown@intel.com>
Cc: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Tested-by: Martin Michlmayr <tbm@cyrius.com>
Cc: <stable@kernel.org> [2.6.31.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The "end of a C state" trace point currently happens before
the code runs that corrects the TSC for having stopped during idle.
The result of this is that the timestamp of the end-of-C-state event
is garbage on cpus where the TSC stops during idle.
This patch moves the end point of the C state to after the timekeeping
engine of the kernel has been corrected.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Cc: Len Brown <len.brown@intel.com>
Cc: fweisbec@gmail.com
Cc: peterz@infradead.org
Cc: Paul Mackerras <paulus@samba.org>
LKML-Reference: <20090919133533.139c2a46@infradead.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
It's showing up as regressions; disabling it very likely just papers
over an underlying issue, but time is running out for 2.6.28, lets get
back to this for 2.6.29
Fixes: #11826 and #11893
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* 'v28-range-hrtimers-for-linus-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (37 commits)
hrtimers: add missing docbook comments to struct hrtimer
hrtimers: simplify hrtimer_peek_ahead_timers()
hrtimers: fix docbook comments
DECLARE_PER_CPU needs linux/percpu.h
hrtimers: fix typo
rangetimers: fix the bug reported by Ingo for real
rangetimer: fix BUG_ON reported by Ingo
rangetimer: fix x86 build failure for the !HRTIMERS case
select: fix alpha OSF wrapper
select: fix alpha OSF wrapper
hrtimer: peek at the timer queue just before going idle
hrtimer: make the futex() system call use the per process slack value
hrtimer: make the nanosleep() syscall use the per process slack
hrtimer: fix signed/unsigned bug in slack estimator
hrtimer: show the timer ranges in /proc/timer_list
hrtimer: incorporate feedback from Peter Zijlstra
hrtimer: add a hrtimer_start_range() function
hrtimer: another build fix
hrtimer: fix build bug found by Ingo
hrtimer: make select() and poll() use the hrtimer range feature
...
cpuidle accounts the idle time for the C-state it was trying to enter and
not to the actual state that the driver eventually entered. The driver may
select a different state than the one chosen by cpuidle due to
constraints like bus-mastering, etc.
Change the time acounting code to look at the dev->last_state after
returning from target_state->enter(). Driver can modify dev->last_state
internally, inside the enter routine to reflect the actual C-state
entered.
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Tested-by: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: Len Brown <len.brown@intel.com>
As part of going idle, we already look at the time of the next timer event to determine
which C-state to select etc.
This patch adds functionality that causes the timers that are past their
soft expire time, to fire at this time, before we calculate the next wakeup
time. This functionality will thus avoid wakeups by running timers before
going idle rather than specially waking up for it.
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
pm_idle_save resp. pm_idle_old can be NULL when the restore code in
acpi_processor_cst_has_changed() resp. cpuidle_uninstall_idle_handler()
is called. This can set pm_idle unconditinally to NULL, which causes the
kernel to panic when calling pm_idle in the x86 idle code. This was
covered by an extra check for !pm_idle in the x86 idle code, which was
removed during the x86 idle code refactoring.
Instead of restoring the pm_idle check in the x86 code prevent the
acpi/cpuidle code to set pm_idle to NULL.
Reported by: Dhaval Giani http://lkml.org/lkml/2008/7/2/309
Based on a debug patch from Ingo Molnar
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
It's never used and the comments refer to nonatomic and retry
interchangably. So get rid of it.
Acked-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
cpuidle and acpi driver interaction bug with the way cpuidle_register_driver()
is called. Due to this bug, there will be oops on
AC<->DC on some systems, where they support C-states in one DC and not in AC.
The current code does
ON BOOT:
Look at CST and other C-state info to see whether more than C1 is
supported. If it is, then acpi processor_idle does a
cpuidle_register_driver() call, which internally enables the device.
ON CST change notification (AC<->DC) and on suspend-resume:
acpi driver temporarily disables device, updates the device with
any new C-states, and reenables the device.
The problem is is on boot, there are no C2, C3 states supported and we skip
the register. Later on AC<->DC, we may get a CST notification and we try
to reevaluate CST and enabled the device, without actually registering it.
This causes breakage as we try to create /sys fs sub directory, without the
parent directory which is created at register time.
Thanks to Sanjeev for reporting the problem here.
http://bugzilla.kernel.org/show_bug.cgi?id=10394
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
commit 9b12e18cdc
'ACPI: cpuidle: Support C1 idle time accounting'
was implicated in a 100% C0 idle regression.
http://bugzilla.kernel.org/show_bug.cgi?id=10076
It pointed out a potential problem where the menu governor
may get confused by the C-state residency time from poll
idle or C1 idle, where this timing info is not accurate.
This inaccuracy is due to interrupts being handled
before we account for C-state exit.
Do not mark TIME_VALID for CO poll state.
Mark C1 time as valid only with the MWAIT (CSTATE_FFH) entry method.
This makes governors use the timing information only when it is correct and
eliminates any wrong policy decisions that may result from invalid timing
information.
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
cpuidle C-state sysfs node time and usage are very easy to overflow because
they are all of unsigned int type, time will overflow within about two hours,
usage will take longer time to overflow, but they are increasing for ever.
This patch will convert them to unsigned long long.
Signed-off-by: Yi Yang <yi.y.yang@intel.com>
Acked-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Add a new sysfs entry under cpuidle states. desc - can be used by driver to
communicate to userspace any specific information about the state.
This helps in identifying the exact hardware C-states behind the ACPI C-state
definition.
Idea is to export this through powertop, which will help to map the C-state
reported by powertop to actual hardware C-state.
Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
Len Brown
Thomas Renninger
Rafael J. Wysocki
Linus Torvalds
Christoph Lameter
Ai Li
Kevin Hilman
Arjan van de Ven
Arjan van de Ven
Venkatesh Pallipadi
Thomas Gleixner
Jens Axboe
Yi Yang