linux-toradex.git/drivers/base/power, branch v4.9.16

PM / wakeirq: Fix dedicated wakeirq for drivers not using autosuspend

2017-01-12T10:39:31+00:00

commit bed570307ed78f21b77cb04a1df781dee4a8f05a upstream.

I noticed some wakeirq flakeyness with consumer drivers not using
autosuspend. For drivers not using autosuspend, the wakeirq may never
get unmasked in rpm_suspend() because of irq desc->depth.

We are configuring dedicated wakeirqs to start with IRQ_NOAUTOEN as we
naturally don't want them running until rpm_suspend() is called.

However, when a consumer driver initially calls pm_runtime_get(), we
now wrongly start with disable_irq_nosync() call on the dedicated
wakeirq that is disabled to start with.

This causes desc->depth to toggle between 1 and 2 instead of the usual
0 and 1. This can prevent enable_irq() from unmasking the wakeirq as
that only happens at desc->depth 1.

This does not necessarily show up with drivers using autosuspend as
there is time for disable_irq_nosync() before rpm_suspend() gets called
after the autosuspend timeout.

Let's fix the issue by adding wirq->status that lazily gets set on
the first rpm_suspend(). We also need PM runtime core private functions
for dev_pm_enable_wake_irq_check() and dev_pm_disable_wake_irq_check()
so we can enable the dedicated wakeirq on the first rpm_suspend().

While at it, let's also fix the comments for dev_pm_enable_wake_irq()
and dev_pm_disable_wake_irq(). Those can still be used by the consumer
drivers as needed because the IRQ core manages the interrupt usecount
for us.

Fixes: 4990d4fe327b (PM / Wakeirq: Add automated device wake IRQ handling)
Signed-off-by: Tony Lindgren 
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

PM / OPP: Don't use OPP structure outside of rcu protected section

2017-01-06T09:40:15+00:00

commit dc39d06fcd7a4a82d72eae7b71e94e888b96d29e upstream.

The OPP structure must not be used out of the rcu protected section.
Cache the values to be used in separate variables instead.

Signed-off-by: Viresh Kumar 
Reviewed-by: Stephen Boyd 
Tested-by: Dave Gerlach 
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

PM / OPP: Pass opp_table to dev_pm_opp_put_regulator()

2017-01-06T09:40:15+00:00

commit 91291d9ad92faa65a56a9a19d658d8049b78d3d4 upstream.

Joonyoung Shim reported an interesting problem on his ARM octa-core
Odoroid-XU3 platform. During system suspend, dev_pm_opp_put_regulator()
was failing for a struct device for which dev_pm_opp_set_regulator() is
called earlier.

This happened because an earlier call to
dev_pm_opp_of_cpumask_remove_table() function (from cpufreq-dt.c file)
removed all the entries from opp_table->dev_list apart from the last CPU
device in the cpumask of CPUs sharing the OPP.

But both dev_pm_opp_set_regulator() and dev_pm_opp_put_regulator()
routines get CPU device for the first CPU in the cpumask. And so the OPP
core failed to find the OPP table for the struct device.

This patch attempts to fix this problem by returning a pointer to the
opp_table from dev_pm_opp_set_regulator() and using that as the
parameter to dev_pm_opp_put_regulator(). This ensures that the
dev_pm_opp_put_regulator() doesn't fail to find the opp table.

Note that similar design problem also exists with other
dev_pm_opp_put_*() APIs, but those aren't used currently by anyone and
so we don't need to update them for now.

Reported-by: Joonyoung Shim 
Signed-off-by: Stephen Boyd 
Signed-off-by: Viresh Kumar 
[ Viresh: Wrote commit log and tested on exynos 5250 ]
Signed-off-by: Rafael J. Wysocki 
Signed-off-by: Greg Kroah-Hartman

PM / sleep: don't suspend parent when async child suspend_{noirq, late} fails

2016-11-11T00:29:09+00:00

Consider two devices, A and B, where B is a child of A, and B utilizes
asynchronous suspend (it does not matter whether A is sync or async). If
B fails to suspend_noirq() or suspend_late(), or is interrupted by a
wakeup (pm_wakeup_pending()), then it aborts and sets the async_error
variable. However, device A does not (immediately) check the async_error
variable; it may continue to run its own suspend_noirq()/suspend_late()
callback. This is bad.

We can resolve this problem by doing our error and wakeup checking
(particularly, for the async_error flag) after waiting for children to
suspend, instead of before. This also helps align the logic for the noirq and
late suspend cases with the logic in __device_suspend().

It's easy to observe this erroneous behavior by, for example, forcing a
device to sleep a bit in its suspend_noirq() (to ensure the parent is
waiting for the child to complete), then return an error, and watch the
parent suspend_noirq() still get called. (Or similarly, fake a wakeup
event at the right (or is it wrong?) time.)

Fixes: de377b397272 (PM / sleep: Asynchronous threads for suspend_late)
Fixes: 28b6fd6e3779 (PM / sleep: Asynchronous threads for suspend_noirq)
Reported-by: Jeffy Chen 
Signed-off-by: Brian Norris 
Signed-off-by: Rafael J. Wysocki

Merge branches 'pm-cpuidle', 'pm-opp' and 'pm-avs'

2016-10-01T23:43:16+00:00

* pm-cpuidle:
  ARM: cpuidle: Fix error return code

* pm-opp:
  PM / OPP: Don't support OPP if it provides supported-hw but platform does not
  PM / OPP: avoid maybe-uninitialized warning

* pm-avs:
  PM / AVS: SmartReflex: Neaten logging

Merge branch 'pm-domains'

2016-10-01T23:41:29+00:00

* pm-domains:
  PM / Domains: Rename pm_genpd_sync_poweron|poweroff()
  PM / Domains: Don't measure latency of ->power_on|off() during system PM
  PM / Domains: Remove redundant system PM callbacks
  PM / Domains: Simplify detaching a device from its genpd
  PM / Domains: Allow holes in genpd_data.domains array
  PM / Domains: Add support for removing nested PM domains by provider
  PM / Domains: Add support for removing PM domains
  PM / Domains: Store the provider in the PM domain structure
  PM / Domains: Prepare for adding support to remove PM domains
  PM / Domains: Verify the PM domain is present when adding a provider
  PM / Domains: Don't expose xlate and provider helper functions
  PM / Domains: Don't expose generic_pm_domain structure to clients
  staging: board: Remove calls to of_genpd_get_from_provider()
  ARM: EXYNOS: Remove calls to of_genpd_get_from_provider()
  PM / Domains: Add new helper functions for device-tree
  PM / Domains: Always enable debugfs support if available

PM / OPP: Don't support OPP if it provides supported-hw but platform does not

2016-09-26T13:13:31+00:00

The OPP framework allows each OPP to set a opp-supported-hw property
which provides values that are matched against supported_hw values
provided by the platform to limit support for certain OPPs on specific
hardware. Currently, if the platform does not set supported_hw values,
all OPPs are interpreted as supported, even if they have provided their
own opp-supported-hw values.

If an OPP has provided opp-supported-hw, it is indicating that there is
some specific hardware configuration it is supported by. These constraints
should be honored, and if no supported_hw has been provided by the
platform, there is no way to determine if that OPP is actually supported,
so it should be marked as not supported.

Signed-off-by: Dave Gerlach 
Acked-by: Viresh Kumar 
Signed-off-by: Rafael J. Wysocki

PM / Domains: Rename pm_genpd_sync_poweron|poweroff()

2016-09-23T23:54:29+00:00

These are internal static functions to genpd. Let's conform to the naming
rules, by dropping the "pm_" prefix from these.

Signed-off-by: Ulf Hansson 
Signed-off-by: Rafael J. Wysocki

PM / Domains: Don't measure latency of ->power_on|off() during system PM

2016-09-23T23:54:29+00:00

Measure latency does by itself contribute to an increased latency, thus we
should avoid it when it isn't needed.

Currently genpd measures latencies in the system PM phase for the
->power_on|off() callbacks, except in the syscore case when it's not
allowed to use ktime_get() as timekeeping may be suspended.

Since there should be plenty of occasions during runtime PM to perform
these measurements, let's rely on that and drop them from system PM. This
will also make it consistent for how measurements are done of the runtime
PM callbacks (as those may be invoked during system PM).

Signed-off-by: Ulf Hansson 
Signed-off-by: Rafael J. Wysocki

PM / Domains: Remove redundant system PM callbacks

2016-09-23T23:54:28+00:00

In cases when the PM domain haven't assigned a system PM callback, the PM
core fall-backs to check for the callback at the driver level instead.
This makes it redundant to assign a pm_generic_* helper function to a
corresponding system PM callback at a PM domain level.

Therefore, let's remove these assignments in pm_genpd_init().

Signed-off-by: Ulf Hansson 
Signed-off-by: Rafael J. Wysocki