While resuming a device belonging to a PM domain,
pm_genpd_runtime_resume() calls __pm_genpd_restore_device() to
restore its state, if necessary.  The latter starts the device,
using genpd_start_dev(), restores its state, using
genpd_restore_dev(), and then stops it, using genpd_stop_dev().
However, this last operation is not necessary, because the
device is supposed to be operational after pm_genpd_runtime_resume()
has returned and because of it pm_genpd_runtime_resume() has to
call genpd_start_dev() once again for the "restored" device, which
is inefficient.

To make things more efficient, remove the call to genpd_stop_dev()
from __pm_genpd_restore_device() and the direct call to
genpd_start_dev() from pm_genpd_runtime_resume().  [Of course,
genpd_start_dev() still has to be called by it for devices with the
power.irq_safe flag set, because __pm_genpd_restore_device() is not
executed for them.]

This change has been tested on the SH7372 Mackerel board.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Currently, the default "save state" and "restore state" routines
for generic PM domains, pm_genpd_default_save_state() and
pm_genpd_default_restore_state(), respectively, only use runtime PM
callbacks provided by device drivers, but in general those callbacks
need not provide the entire necessary functionality.  Namely, in
general it may be necessary to execute subsystem (i.e. device type,
device class or bus type) callbacks that will carry out all of the
necessary operations.

For this reason, modify pm_genpd_default_save_state() and
pm_genpd_default_restore_state() to execute subsystem callbacks,
if they are provided, and fall back to driver callbacks otherwise.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

__device_suspend() must always send a completion. Otherwise, parent
devices will wait forever.

Commit 1e2ef05b, "PM: Limit race conditions between runtime PM and
system sleep (v2)", introduced a regression by short-circuiting the
complete_all() for certain error cases.

This patch fixes the bug by always signalling a completion.

Addresses http://crosbug.com/31972

Tested by injecting an abort.
Signed-off-by: NMandeep Singh Baines <msb@chromium.org>
Cc: stable@vger.kernel.org
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The generic PM domains core code currently requires domains to be in
the "power on" state for adding devices to them, but this limitation
turns out to be inconvenient in some situations, so remove it.

For this purpose, make __pm_genpd_add_device() set the device's
need_restore flag if the domain is in the "power off" state, so that
the device's "restore state" (usually .runtime_resume()) callback
is executed when it is resumed after the domain has been turned on.
If the domain is in the "power on" state, the device's need_restore
flag will be cleared by __pm_genpd_add_device(), so that its "save
state" (usually .runtime_suspend()) callback is executed when the
domain is about to be turned off.  However, since that default
behavior need not be always desirable, add a helper function
pm_genpd_dev_need_restore() allowing a device's need_restore flag
to be set/unset at any time.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The default domain power off governor function for generic PM
domains, default_power_down_ok(), may violate subdomain maximum
off time limit by allowing the master domain to be off for too
long.  Namely, it only finds the minium of all device maximum
off times over the domain's devices and uses that to compute the
domain's maximum off time, but it should do the same for the
subdomains.

Fix this problem by modifying default_power_down_ok() to compute
the given domain's maximum off time as the difference between the
minimum off time over all devices and subdomains in the domain and
its power on latency.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Current pm_genpd_add_subdomain() will allow duplicated link between
master and slave domain.  This patch fixed it.

Because when current pm_genpd_add_subdomain() checks whether the link
between the master and slave generic PM domain already exists,
slave_links instead of master_links of master domain is used.
Signed-off-by: NHuang Ying <ying.huang@intel.com>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The results of the default device stop and domain power off governor
functions for generic PM domains, default_stop_ok() and
default_power_down_ok(), depend only on the timing data of devices,
which are static, and on their PM QoS constraints.  Thus, in theory,
these functions only need to carry out their computations, which may
be time consuming in general, when it is known that the PM QoS
constraint of at least one of the devices in question has changed.

Use the PM QoS notifiers of devices to implement that.  First,
introduce new fields, constraint_changed and max_off_time_changed,
into struct gpd_timing_data and struct generic_pm_domain,
respectively, and register a PM QoS notifier function when adding
a device into a domain that will set those fields to 'true' whenever
the device's PM QoS constraint is modified.  Second, make
default_stop_ok() and default_power_down_ok() use those fields to
decide whether or not to carry out their computations from scratch.

The device and PM domain hierarchies are taken into account in that
and the expense is that the changes of PM QoS constraints of
suspended devices will not be taken into account immediately, which
isn't guaranteed anyway in general.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The removal of a device from a PM domain doesn't have to browse
the domain's device list, because it can check directly if the
device belongs to the given domain.  Moreover, it should clear
the domain_data pointer in dev->power.subsys_data, because
dev_pm_put_subsys_data(dev) may not remove dev->power.subsys_data
and the stale domain data pointer may cause problems to happen.

Rework pm_genpd_remove_device() taking the above observations into
account.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The current behavior of dev_pm_qos_add_notifier() makes device PM QoS
notifiers less than useful.  Namely, it silently returns success when
called before any PM QoS constraints are added for the device, so the
caller will assume that the notifier has been registered, but when
someone actually adds some nontrivial constraints for the device
eventually, the previous callers of dev_pm_qos_add_notifier()
will not know about that and their notifier routines will not be
executed (contrary to their expectations).

To address this problem make dev_pm_qos_add_notifier() create the
constraints object for the device if it is not present when the
routine is called.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by : markgross <markgross@thegnar.org>

After the previous changes in default_stop_ok() and
default_power_down_ok() for PM domains, there are two fields in
struct dev_pm_info that aren't necessary any more,  suspend_time
and max_time_suspended_ns.

Remove those fields along with all of the code that accesses them,
which simplifies the runtime PM framework quite a bit.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The existing default domain power down governor function for PM
domains, default_power_down_ok(), is supposed to check whether or not
the PM QoS latency constraints of the devices in the domain will be
violated if the domain is turned off by pm_genpd_poweroff().
However, the computations carried out by it don't reflect the
definition of the PM QoS latency constrait in
Documentation/ABI/testing/sysfs-devices-power.

Make default_power_down_ok() follow the definition of the PM QoS
latency constrait.  In particular, make it only take latencies into
account, because it doesn't matter how much time has elapsed since
the domain's devices were suspended for the computation.

Remove the break_even_ns and power_off_time fields from
struct generic_pm_domain, because they are not necessary any more.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The existing default device stop governor function for PM domains,
default_stop_ok(), is supposed to check whether or not the device's
PM QoS latency constraint will be violated if the device is stopped
by pm_genpd_runtime_suspend().  However, the computations carried out
by it don't reflect the definition of the PM QoS latency constrait in
Documentation/ABI/testing/sysfs-devices-power.

Make default_stop_ok() follow the definition of the PM QoS latency
constrait.  In particular, make it take the device's start and stop
latencies correctly.

Add a new field, effective_constraint_ns, to struct gpd_timing_data
and use it to store the difference between the device's PM QoS
constraint and its resume latency for use by the device's parent
(the effective_constraint_ns values for the children are used for
computing the parent's one along with its PM QoS constraint).

Remove the break_even_ns field from struct gpd_timing_data, because
it's not used any more.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Android allows user space to manipulate wakelocks using two
sysfs file located in /sys/power/, wake_lock and wake_unlock.
Writing a wakelock name and optionally a timeout to the wake_lock
file causes the wakelock whose name was written to be acquired (it
is created before is necessary), optionally with the given timeout.
Writing the name of a wakelock to wake_unlock causes that wakelock
to be released.

Implement an analogous interface for user space using wakeup sources.
Add the /sys/power/wake_lock and /sys/power/wake_unlock files
allowing user space to create, activate and deactivate wakeup
sources, such that writing a name and optionally a timeout to
wake_lock causes the wakeup source of that name to be activated,
optionally with the given timeout.  If that wakeup source doesn't
exist, it will be created and then activated.  Writing a name to
wake_unlock causes the wakeup source of that name, if there is one,
to be deactivated.  Wakeup sources created with the help of
wake_lock that haven't been used for more than 5 minutes are garbage
collected and destroyed.  Moreover, there can be only WL_NUMBER_LIMIT
wakeup sources created with the help of wake_lock present at a time.

The data type used to track wakeup sources created by user space is
called "struct wakelock" to indicate the origins of this feature.

This version of the patch includes an rbtree manipulation fix from John Stultz.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: NNeilBrown <neilb@suse.de>

Android uses one wakelock statistics that is only necessary for
opportunistic sleep.  Namely, the prevent_suspend_time field
accumulates the total time the given wakelock has been locked
while "automatic suspend" was enabled.  Add an analogous field,
prevent_sleep_time, to wakeup sources and make it behave in a similar
way.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Introduce a mechanism by which the kernel can trigger global
transitions to a sleep state chosen by user space if there are no
active wakeup sources.

It consists of a new sysfs attribute, /sys/power/autosleep, that
can be written one of the strings returned by reads from
/sys/power/state, an ordered workqueue and a work item carrying out
the "suspend" operations.  If a string representing the system's
sleep state is written to /sys/power/autosleep, the work item
triggering transitions to that state is queued up and it requeues
itself after every execution until user space writes "off" to
/sys/power/autosleep.

That work item enables the detection of wakeup events using the
functions already defined in drivers/base/power/wakeup.c (with one
small modification) and calls either pm_suspend(), or hibernate() to
put the system into a sleep state.  If a wakeup event is reported
while the transition is in progress, it will abort the transition and
the "system suspend" work item will be queued up again.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: NNeilBrown <neilb@suse.de>

Add tracepoints to wakeup_source_activate and wakeup_source_deactivate.
Useful for checking that specific wakeup sources overlap as expected.
Signed-off-by: NArve Hjønnevåg <arve@android.com>
Acked-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Wakeup statistics used by Android are slightly different from what we
have in wakeup sources at the moment and there aren't any known
users of those statistics other than Android, so modify them to make
it easier for Android to switch to wakeup sources.

This removes the struct wakeup_source's hit_cout field, which is very
rough and therefore not very useful, and adds two new fields,
wakeup_count and expire_count.  The first one tracks how many times
the wakeup source is activated with events_check_enabled set (which
roughly corresponds to the situations when a system power transition
to a sleep state is in progress and would be aborted by this wakeup
source if it were the only active one at that time) and the second
one is the number of times the wakeup source has been activated with
a timeout that expired.

Additionally, the last_time field is now updated when the wakeup
source is deactivated too (previously it was only updated during
the wakeup source's activation), which seems to be what Android does
with the analogous counter for wakelocks.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The current wakeup source deactivation code doesn't do anything when
the counter of wakeup events in progress goes down to zero, which
requires pm_get_wakeup_count() to poll that counter periodically.
Although this reduces the average time it takes to deactivate a
wakeup source, it also may lead to a substantial amount of unnecessary
polling if there are extended periods of wakeup activity.  Thus it
seems reasonable to use a wait queue for signaling the "no wakeup
events in progress" condition and remove the polling.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: Nmark gross <markgross@thegnar.org>
Acked-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Currently, the device suspend code in drivers/base/power/main.c
only checks if there have been any wakeup events, and therefore the
ongoing system transition to a sleep state should be aborted, during
the first (i.e. "suspend") device suspend phase.  However, wakeup
events may be reported later as well, so it's reasonable to look for
them in the in the subsequent (i.e. "late suspend" and "suspend
noirq") phases.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Acked-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

This patch (as1535) fixes a bug in the runtime PM core.  When a
runtime suspend attempt completes, whether successfully or not, the
device's power.wait_queue is supposed to be signalled.  But this
doesn't happen in the failure pathway of rpm_suspend() when another
autosuspend attempt is rescheduled.  As a result, a task can get stuck
indefinitely on the wait queue (I have seen this happen in testing).

The patch fixes the problem by moving the wake_up_all() call up near
the start of the failure code.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
CC: <stable@vger.kernel.org>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Power domains that were off before hibernation shouldn't be turned on
during device restore, so prevent that from happening.

This change fixes up commit 65533bbf

    PM / Domains: Fix hibernation restore of devices, v2

that didn't include it by mistake.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The TMU device on the Mackerel board belongs to the A4R power domain
and loses power when the domain is turned off.  Unfortunately, the
TMU driver is not prepared to cope with such situations and crashes
the system when that happens.  To work around this problem introduce
a new helper function, pm_genpd_dev_always_on(), allowing a device
driver to mark its device as "always on" in case it belongs to a PM
domain, which will make the generic PM domains core code avoid
powering off the domain containing the device, both at run time and
during system suspend.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Tested-by: NSimon Horman <horms@verge.net.au>
Acked-by: NPaul Mundt <lethal@linux-sh.org>
Cc: stable@vger.kernel.org

During resume from hibernation pm_genpd_restore_noirq() should only
power off domains whose suspend_power_off flags are set once and
not every time it is called for a device in the given domain.
Moreover, it shouldn't decrement genpd->suspended_count, because
that field is not touched during device freezing and therefore it is
always equal to 0 when pm_genpd_restore_noirq() runs for the first
device in the given domain.

This means pm_genpd_restore_noirq() may use genpd->suspended_count
to determine whether or not it it has been called for the domain in
question already in this cycle (it only needs to increment that
field every time it runs for this purpose) and whether or not it
should check if the domain needs to be powered off.  For that to
work, though, pm_genpd_prepare() has to clear genpd->suspended_count
when it runs for the first device in the given domain (in which case
that flag need not be cleared during domain initialization).
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Cc: stable@vger.kernel.org

During system suspend pm_genpd_suspend_noirq() checks if the given
device is in a wakeup path (i.e. it appears to be needed for one or
more wakeup devices to work or is a wakeup device itself) and if it
needs to be "active" for wakeup to work.  If that is the case, the
function returns 0 without incrementing the device domain's counter
of suspended devices and without executing genpd_stop_dev() for the
device.  In consequence, the device is not stopped (e.g. its clock
isn't disabled) and power is always supplied to its domain in the
resulting system sleep state.

However, pm_genpd_resume_noirq() doesn't repeat that check and it
runs genpd_start_dev() and decrements the domain's counter of
suspended devices even for the wakeup device that weren't stopped by
pm_genpd_suspend_noirq().  As a result, the start callback may be run
unnecessarily for them and their domains' counters of suspended
devices may become negative.  Both outcomes aren't desirable, so fix
pm_genpd_resume_noirq() to look for wakeup devices that might not be
stopped by during system suspend.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Tested-by: NSimon Horman <horms@verge.net.au>
Cc: stable@vger.kernel.org

A runtime suspend of a device (e.g. an MMC controller) belonging to
a power domain or, in a more complicated scenario, a runtime suspend
of another device in the same power domain, may cause power to be
removed from the entire domain.  In that case, the amount of time
necessary to runtime-resume the given device (e.g. the MMC
controller) is often substantially greater than the time needed to
run its driver's runtime resume callback.  That may hurt performance
in some situations, because user data may need to wait for the
device to become operational, so we should make it possible to
prevent that from happening.

For this reason, introduce a new sysfs attribute for devices,
power/pm_qos_resume_latency_us, allowing user space to specify the
upper bound of the time necessary to bring the (runtime-suspended)
device up after the resume of it has been requested.  However, make
that attribute appear only for the devices whose drivers declare
support for it by calling the (new) dev_pm_qos_expose_latency_limit()
helper function with the appropriate initial value of the attribute.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Reviewed-by: NKevin Hilman <khilman@ti.com>
Reviewed-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Acked-by: NLinus Walleij <linus.walleij@linaro.org>

For files that are actively using linux/device.h, make sure
that they call it out.  This will allow us to clean up some
of the implicit uses of linux/device.h within include/*
without introducing build regressions.

Yes, this was created by "cheating" -- i.e. the headers were
cleaned up, and then the fallout was found and fixed, and then
the two commits were reordered.  This ensures we don't introduce
build regressions into the git history.
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>

The existing wakeup source initialization routines are not
particularly useful for wakeup sources that aren't created by
wakeup_source_create(), because their users have to open code
filling the objects with zeros and setting their names.  For this
reason, introduce routines that can be used for initializing, for
example, static wakeup source objects.
Requested-by: NArve Hjønnevåg <arve@android.com>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

If __pm_stay_awake() is called after __pm_wakeup_event() for the same
wakep source object before its timer expires, it won't cancel the
timer, so the wakeup source will be deactivated from the timer
function as scheduled by __pm_wakeup_event().  In that case
__pm_stay_awake() doesn't have any effect beyond incrementing
the wakeup source's event_count field, although it should cancel
the timer and make the wakeup source stay active until __pm_relax()
is called for it.

To fix this problem make __pm_stay_awake() delete the wakeup source's
timer and ensure that it won't be deactivated from the timer funtion
afterwards by clearing its timer_expires field.
Reported-by: NArve Hjønnevåg <arve@android.com>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

If __pm_wakeup_event() has been used (with a nonzero timeout) to
report a wakeup event and then __pm_relax() immediately followed by
__pm_stay_awake() is called or __pm_wakeup_event() is called once
again for the same wakeup source object before its timer expires, the
timer function pm_wakeup_timer_fn() may still be run as a result of
the previous __pm_wakeup_event() call.  In either of those cases it
may mistakenly deactivate the wakeup source that has just been
activated.

To prevent that from happening, make wakeup_source_deactivate()
clear the wakeup source's timer_expires field and make
pm_wakeup_timer_fn() check if timer_expires is different from zero
and if it's not in future before calling wakeup_source_deactivate()
(if timer_expires is 0, it means that the timer has just been
deleted and if timer_expires is in future, it means that the timer
has just been rescheduled to a different time).
Reported-by: NArve Hjønnevåg <arve@android.com>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

If wakeup_source_destroy() is called for an active wakeup source that
is never deactivated, it will spin forever.  To prevent that from
happening, make wakeup_source_destroy() call __pm_relax() for the
wakeup source object it is about to free instead of waiting until
it will be deactivated by someone else.  However, for this to work
it also needs to make sure that the timer function will not be
executed after the final __pm_relax(), so make it run
del_timer_sync() on the wakeup source's timer beforehand.

Additionally, update the kerneldoc comment to document the
requirement that __pm_stay_awake() and __pm_wakeup_event() must not
be run in parallel with wakeup_source_destroy().
Reported-by: NArve Hjønnevåg <arve@android.com>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Initialize wakeup source locks in wakeup_source_add() instead of
wakeup_source_create(), because otherwise the locks of the wakeup
sources that haven't been allocated with wakeup_source_create()
aren't initialized and handled properly.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

After the introduction of the late/early phases of device
suspend/resume during system-wide power transitions it is possible
to make the generic PM domains code execute its default late/early
device suspend/resume callbacks during those phases instead of the
corresponding _noirq phases.  The _noirq device suspend/resume
phases were only used for executing those callbacks, because this
was the only way it could be done, but now we can do better.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Introduce generic subsystem callbacks for the new phases of device
suspend/resume during system power transitions: "late suspend",
"early resume", "late freeze", "early thaw", "late poweroff",
"early restore".
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

The current device suspend/resume phases during system-wide power
transitions appear to be insufficient for some platforms that want
to use the same callback routines for saving device states and
related operations during runtime suspend/resume as well as during
system suspend/resume.  In principle, they could point their
.suspend_noirq() and .resume_noirq() to the same callback routines
as their .runtime_suspend() and .runtime_resume(), respectively,
but at least some of them require device interrupts to be enabled
while the code in those routines is running.

It also makes sense to have device suspend-resume callbacks that will
be executed with runtime PM disabled and with device interrupts
enabled in case someone needs to run some special code in that
context during system-wide power transitions.

Apart from this, .suspend_noirq() and .resume_noirq() were introduced
as a workaround for drivers using shared interrupts and failing to
prevent their interrupt handlers from accessing suspended hardware.
It appears to be better not to use them for other porposes, or we may
have to deal with some serious confusion (which seems to be happening
already).

For the above reasons, introduce new device suspend/resume phases,
"late suspend" and "early resume" (and analogously for hibernation)
whose callback will be executed with runtime PM disabled and with
device interrupts enabled and whose callback pointers generally may
point to runtime suspend/resume routines.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>
Reviewed-by: NMark Brown <broonie@opensource.wolfsonmicro.com>
Reviewed-by: NKevin Hilman <khilman@ti.com>

A device node pointer is added to generic pm domain structure to associate
the domain with a node in the device tree. The platform code parses the
device tree to find available nodes representing the generic power domain,
instantiates the available domains and initializes them by calling
pm_genpd_init().

Nodes representing the devices include a phandle of the power domain to
which it belongs. As these devices get instantiated, the driver code
checkes for availability of a power domain phandle, converts the phandle
to a device node and uses the new pm_genpd_of_add_device() api to
associate the device with a power domain.

pm_genpd_of_add_device() runs through its list of registered power domains
and matches the OF node of the domain with the one specified as the
parameter. If a match is found, the device is associated with the matched
domain.

Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Grant Likely <grant.likely@secretlab.ca>
Signed-off-by: NThomas Abraham <thomas.abraham@linaro.org>
Acked-by: NRafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

The governor functions in drivers/base/power/domain_governor.c
are only used if CONFIG_PM_RUNTIME is set and they refer to data
structures that are only present in that case.  For this reason,
they shouldn't be compiled at all when CONFIG_PM_RUNTIME is not set.
Reported-by: NKukjin Kim <kgene.kim@samsung.com>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Some callback functions defined in drivers/base/power/domain.c are
only necessary if CONFIG_PM_SLEEP is set and they call some other
functions that are only available in that case.  For this reason,
they should not be compiled at all when CONFIG_PM_SLEEP is not set.
Reported-by: NMagnus Damm <damm@opensource.se>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Some devices, like the I2C controller on SH7372, are not
necessary for providing power to their children or forwarding
wakeup signals (and generally interrupts) from them.  They are
only needed by their children when there's some data to transfer,
so they may be suspended for the majority of time and resumed
on demand, when the children have data to send or receive.  For this
purpose, however, their power.ignore_children flags have to be set,
or the PM core wouldn't allow them to be suspended while their
children were active.

Unfortunately, in some situations it may take too much time to
resume such devices so that they can assist their children in
transferring data.  For example, if such a device belongs to a PM
domain which goes to the "power off" state when that device is
suspended, it may take too much time to restore power to the
domain in response to the request from one of the device's
children.  In that case, if the parent's resume time is critical,
the domain should stay in the "power on" state, although it still may
be desirable to power manage the parent itself (e.g. by manipulating
its clock).

In general, device PM QoS may be used to address this problem.
Namely, if the device's children added PM QoS latency constraints
for it, they would be able to prevent it from being put into an
overly deep low-power state.  However, in some cases the devices
needing to be serviced are not the immediate children of a
"children-ignoring" device, but its grandchildren or even less
direct descendants.  In those cases, the entity wanting to add a
PM QoS request for a given device's ancestor that ignores its
children will have to find it in the first place, so introduce a new
helper function that may be used to achieve that.  This function,
dev_pm_qos_add_ancestor_request(), will search for the first
ancestor of the given device whose power.ignore_children flag is
set and will add a device PM QoS latency request for that ancestor
on behalf of the caller.  The request added this way may be removed
with the help of dev_pm_qos_remove_request() in the future, like
any other device PM QoS latency request.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Since the PM core is now going to execute driver callbacks directly
if the corresponding subsystem callbacks are not present,
forward-only subsystem callbacks (i.e. such that only execute the
corresponding driver callbacks) are not necessary any more.  Thus
it is possible to remove generic_subsys_pm_ops, because the only
callback in there that is not forward-only, .runtime_idle, is not
really used by the only user of generic_subsys_pm_ops, which is
vio_bus_type.

However, the generic callback routines themselves cannot be removed
from generic_ops.c, because they are used individually by a number
of subsystems.
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

Make the PM core execute driver PM callbacks directly if the
corresponding subsystem callbacks are not present.

There are three reasons for doing that.  First, it reflects the
behavior of drivers/base/dd.c:really_probe() that runs the driver's
.probe() callback directly if the bus type's one is not defined, so
this change will remove one arbitrary difference between the PM core
and the remaining parts of the driver core.  Second, it will allow
some subsystems, whose PM callbacks don't do anything except for
executing driver callbacks, to be simplified quite a bit by removing
those "forward-only" callbacks.  Finally, it will allow us to remove
one level of indirection in the system suspend and resume code paths
where it is not necessary, which is going to lead to less debug noise
with initcall_debug passed in the kernel command line (messages won't
be printed for driverless devices whose subsystems don't provide
PM callbacks among other things).
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>