This commit renames rcu_lockdep_assert() to RCU_LOCKDEP_WARN() for
consistency with the WARN() series of macros.  This also requires
inverting the sense of the conditional, which this commit also does.
Reported-by: NIngo Molnar <mingo@kernel.org>
Signed-off-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Reviewed-by: NIngo Molnar <mingo@kernel.org>

This patch reduces the kernel size by removing error messages that duplicate
the normal OOM message.

A simplified version of the semantic patch that finds this problem is as
follows: (http://coccinelle.lip6.fr)

@@
identifier f,print,l;
expression e;
constant char[] c;
@@

e = \(kzalloc\|kmalloc\|devm_kzalloc\|devm_kmalloc\)(...);
if (e == NULL) {
  <+...
-  print(...,c,...);
  ... when any
(
  goto l;
|
  return ...;
)
  ...+> }
Signed-off-by: NQuentin Lambert <lambert.quentin@gmail.com>
Acked-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Add lockdep asserts for holding the RCU lock when calling
dev_pm_opp_get_freq() and dev_pm_opp_get_voltage() to aid in detecting
RCU misuses.

These are called often after dev_pm_opp_find_freq_ceil/exact() which
already asserts for RCU lock. However one could make an error by
releasing lock too early - just after dev_pm_opp_find_freq_ceil().
Signed-off-by: NKrzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

kernel doc has gotten bit-rotted over time. Re-sync with Locking and
Return information. document all functions properly and ensure that
./scripts/kernel-doc -v  ./drivers/base/power/opp.c >/dev/null returns
no errors
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

All exported functions use dev_pm_* prefix and all static functions
are now standardized with _ prefix. This is better than having to deal
with multiple function naming styles within the same file.
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Allows user drivers such as devfreq to be modules.
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

A lot of callers are missing the fact that dev_pm_opp_get_opp_count
needs to be called under RCU lock. Given that RCU locks can safely be
nested, instead of providing *_locked() API, let's take RCU lock inside
dev_pm_opp_get_opp_count() and leave callers as is.
Signed-off-by: NDmitry Torokhov <dtor@chromium.org>
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Not having OPP defined for a device is not a crime, we should not splat
warning in this case. Also, it seems that we are ready to accept invalid
dev (find_device_opp will return ERR_PTR(-EINVAL) then) so let's not
crash in dev_name() in such case.
Signed-off-by: NDmitry Torokhov <dtor@chromium.org>
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Certain OPP APIs need to be called under RCU lock; let's add a few
rcu_lockdep_assert() calls to warn about potential misuse.
Signed-off-by: NDmitry Torokhov <dtor@chromium.org>
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

This makes it less error prone and moves common resource deallocation at a
single place.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Get the 'device_opp' allocation code into a separate routine to keep only the
necessary part in dev_pm_opp_add_dynamic().

Also do s/sizeof(struct device_opp)/sizeof(*dev_opp) and remove the print
message on kzalloc() failure as checkpatch warns for that.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Reuse find_device_opp() in opp_set_availability() instead of duplicating code.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Its a local routine and need not be accessible outside of opp.c.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Somehow one of the instance of freeing resources failed to use kfree_rcu() and
used kfree() instead. This might cause problems as the node might be referenced
by readers under rcu locks and we must wait for the rcu grace period as well.

While we are at it, also update comment over 'struct device_opp' to mention why
we are waiting for both rcu and srcu grace periods.

Fixes: 129eec55 (PM / OPP Introduce APIs to remove OPPs)
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

By mistake we called find_device_opp() twice in of_free_opp_table(), fix it.

Generated diff doesn't show the problem well and so here is the code snippet:

void of_free_opp_table(struct device *dev)
{
	struct device_opp *dev_opp = find_device_opp(dev);
	struct dev_pm_opp *opp, *tmp;

	/* Check for existing list for 'dev' */
	dev_opp = find_device_opp(dev);

	...
}
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

We find/allocate dev_opp after using its value to fill new_opp->dev_opp right
now. Move this to a later point where dev_opp is valid.

Fixes: a7470db6 (PM / OPP don't match for existing OPPs when list is empty)
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

This existed before we introduced call_srcu() in opp layer to synchronize with
srcu_notifier_call_chain() while removing OPPs. And is a potential bug which
wasn't noticed earlier.

Let fix it as well by using the right API to free OPP.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

OPPs are created statically (from DT) or dynamically. Currently we don't free
OPPs that are created statically, when the module unloads. And so if the module
is inserted back again, we get warning for duplicate OPPs as the same were
already present.

Also, there might be a need to remove dynamic OPPs in future and so API for that
is also added.

This patch adds helper APIs to remove/free existing static and dynamic OPPs.

Because the OPPs are used both under RCU and SRCU, we have to wait for grace
period of both. And so are using kfree_rcu() from within call_srcu().
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Static OPPs are the ones created from Device Tree entries and dynamic are the
ones created at runtime by calling dev_pm_opp_add().

There is a need to distinguish them as we need to free static OPPs from cpufreq
drivers when they are removed.

So, add another field 'dynamic' in 'struct dev_pm_opp' to keep this information.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

OPP list is guaranteed to be empty when 'dev_opp' is created. And so we don't
need to run the comparison loop with existing OPPs.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Both 'struct dev_pm_opp' and 'struct device_opp' have member with name 'head'
but with different types. This leads to confusion while reading the code.

Name them 'rcu_head' and 'srcu_head'.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

We don't have any protection against addition of duplicate OPPs currently and in
case some code tries to add them, it will end up corrupting OPP tables.

We need to handle some duplication cases separately as returning error might not
be the right thing always. The new list of return values for dev_pm_opp_add()
are:

 0:         On success OR
            Duplicate OPPs (both freq and volt are same) and opp->available
 -EEXIST:   Freq are same and volt are different OR
            Duplicate OPPs (both freq and volt are same) and !opp->available
 -ENOMEM:   Memory allocation failure
Acked-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NChander Kashyap <k.chander@samsung.com>
Signed-off-by: NInderpal Singh <inderpal.s@samsung.com>
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

In of_init_opp_table function, if a failure to add an OPP is
detected, the count of OPPs, yet to be added is not updated.
Fix this by decrementing this count on failure as well.
Signed-off-by: NChander Kashyap <k.chander@samsung.com>
Signed-off-by: NInderpal Singh <inderpal.s@samsung.com>
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Cc: 3.7+ <stable@vger.kernel.org> # 3.7+
Acked-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

CPUFreq specific helper functions for OPP (Operating Performance Points)
now use generic OPP functions that allow CPUFreq to be be moved back
into CPUFreq framework. This allows for independent modifications
or future enhancements as needed isolated to just CPUFreq framework
alone.

Here, we just move relevant code and documentation to make this part of
CPUFreq infrastructure.

Cc: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

CPUFREQ custom functions for OPP (Operating Performance Points)
currently exist inside the OPP library. These custom functions currently
depend on internal data structures to pick up OPP information to create
the cpufreq table.  For example, the cpufreq table is created precisely
in the same order of how OPP entries are stored inside the list implementation.

This kind of tight interdependency is purely artificial since the same
functionality can be achieved using the generic OPP functions
meant to do the same. This interdependency also limits the independent
modification of cpufreq and OPP library.

So use the generic dev_pm_opp_find_freq_ceil function that achieves the
table organization as we currently use.

As a result of this, we dont need to use the internal device_opp
structure anymore, and we hence we can switch over to rcu lock instead
of the mutex holding the internal list lock.

This breaking of dependency on internal data structure imposes no change
to usage of these.

NOTE: This change is a precursor to moving this cpufreq specific logic
out of the generic library into cpufreq.

Cc: Kevin Hilman <khilman@deeprootsystems.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

None of these files are actually using any __init type directives
and hence don't need to include <linux/init.h>.  Most are just a
left over from __devinit and __cpuinit removal, or simply due to
code getting copied from one driver to the next.

Cc: Len Brown <len.brown@intel.com>
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>
Acked-by: NPavel Machek <pavel@ucw.cz>
Acked-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: NMark Brown <broonie@linaro.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

Since Operating Performance Points (OPP) functions are specific
to device specific power management, be specific and rename opp.h
to pm_opp.h
Reported-by: NRandy Dunlap <rdunlap@infradead.org>
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Since Operating Performance Points (OPP) data structures are specific
to device specific power management, be specific and rename opp_* data
structures in OPP library with dev_pm_opp_* equivalent.

Affected structures are:
 struct opp
 enum opp_event

Minor checkpatch warning resulting of this change was fixed as well.
Reported-by: NRandy Dunlap <rdunlap@infradead.org>
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Since Operating Performance Points (OPP) functions are specific to
device specific power management, be specific and rename opp_*
accessors in OPP library with dev_pm_opp_* equivalent.

Affected functions are:
 opp_get_voltage
 opp_get_freq
 opp_get_opp_count
 opp_find_freq_exact
 opp_find_freq_floor
 opp_find_freq_ceil
 opp_add
 opp_enable
 opp_disable
 opp_get_notifier
 opp_init_cpufreq_table
 opp_free_cpufreq_table
Reported-by: NRandy Dunlap <rdunlap@infradead.org>
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NViresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Export opp_add() so that modules can use it.
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Acked-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

The "index" field of struct cpufreq_frequency_table was never an
index and isn't used at all by the cpufreq core.  It only is useful
for cpufreq drivers for their internal purposes.

Many people nowadays blindly set it in ascending order with the
assumption that the core will use it, which is a mistake.

Rename it to "driver_data" as that's what its purpose is. All of its
users are updated accordingly.

[rjw: Changelog]
Signed-off-by: NViresh Kumar <viresh.kumar@linaro.org>
Acked-by: NSimon Horman <horms+renesas@verge.net.au>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

commit dde8437d (PM / OPP: RCU reclaim) introduced rcu_head for
struct opp. This aids freeing using kfree_rcu. However, we missed
adding documentation for the same. This generates kernel doc warning:
Warning(drivers/base/power/opp.c:70): No description found for
parameter 'head'

Add documentation as appropriate.

[rjw: Changelog]
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Export cpufreq helpers in OPP to make the cpufreq-core0 and highbank-cpufreq
drivers loadable as modules.
Signed-off-by: NMark Langsdorf <mark.langsdorf@calxeda.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

We are GPLV2 library, so be clear in the symbols exported as well.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NShawn Guo <shawn.guo@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

The callback function of call_rcu() just calls a kfree(), so we
can use kfree_rcu() instead of call_rcu() + callback function.

dpatch engine is used to auto generate this patch.
(https://github.com/weiyj/dpatch)
Signed-off-by: NWei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Currently the opp_find* functions return -ENODEV when:
a) it cant find a device (e.g. request for an OPP search on device
   which was not registered)
b) When it cant find a match for the search strategy used

This makes life a little in-efficient for users such as devfreq
to make reasonable judgement before switching search strategies.

So, standardize the return results as following:
 -EINVAL for bad pointer parameters
 -ENODEV when device cannot be found
 -ERANGE when search fails

This has the following benefit for devfreq implementation:
The search fails when an unregistered device pointer is provided.
This is a trigger to change the search direction and search for
a better fit, however, if we cannot differentiate between a valid
search range failure Vs an unregistered device, second search goes
through the same fail return condition. This can be avoided by
appropriate handling of error return code.

With this change, we also fix devfreq for the improved search
strategy with updated error code.
Signed-off-by: NNishanth Menon <nm@ti.com>
Reviewed-by: NKevin Hilman <khilman@ti.com>
Acked-by: NMyungJoo Ham <myungjoo.ham@samsung.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Export the OPP functions for use by driver modules.

Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Kevin Hilman <khilman@ti.com>
Cc: linux-pm@vger.kernel.org
Cc: linux-kernel@vger.kernel.org

[nm@ti.com: expansion of functions exported]
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NLiam Girdwood <lrg@ti.com>
Acked-by: NKevin Hilman <khilman@ti.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

synchronize_rcu() blocks the caller of opp_enable/disbale
for a complete grace period. This blocking duration prevents
any intensive use of the functions. Replace synchronize_rcu()
by call_rcu() which will call our function for freeing the old
opp element.

The duration of opp_enable() and opp_disable() will be no more
dependant of the grace period.
Signed-off-by: NVincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

With a lot of devices booting from device tree nowadays, it requires
that OPP table can be initialized from device tree.  The patch adds
a helper function of_init_opp_table together with a binding doc for
that purpose.
Signed-off-by: NShawn Guo <shawn.guo@linaro.org>
Acked-by: NRob Herring <rob.herring@calxeda.com>
Signed-off-by: NRafael J. Wysocki <rjw@sisk.pl>

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>