cpufreq: arm-big-little: use clk_get instead of clk_get_sys

Currently clk_get_sys is used with cpu-cluster.<n> as the device id
which is incorrect. It should be connection/consumer ID instead.

It is possible to specify input clock in the cpu device node along
with the optional clock-name. clk_get_sys can't handle that.

This patch replaces clk_get_sys with clk_get to extend support for
clocks specified in the device tree cpu node.

Signed-off-by: Sudeep KarkadaNagesha <[email protected]>
Acked-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: exynos: Show a list of available frequencies

This patch adds freq_attr to show a list of exynos5440 scaling
available frequencies through sysfs. Common exynos driver already
supports this attribute.

Signed-off-by: Jungseok Lee <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sa11x0: Fix build breakage after "Expose frequency table"

Fix build breakage introduced by commit 22c8b4f (cpufreq: sa11x0:
Expose frequency table).

[rjw: Changelog]
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: tegra: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: spear: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sa11x0: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: s5pv210: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: s3c: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pmac64: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pmac32: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Cc: Benjamin Herrenschmidt <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pasemi: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: omap: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

This also rearranges the code a bit to make it more sensible. Also removes some
unnecessary checks.

Cc: Santosh Shilimkar <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: maple: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: loongson2: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here. This driver wasn't setting transition_latency and so is getting set to 0
by default. Lets mark it explicitly by calling the generic routine with
transition_latency as 0.

Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Aaro Koskinen <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: kirkwood: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: imx6q: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: exynos: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: dbx500: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Linus Walleij <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: davinci: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Cc: Sekhar Nori <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: cris: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Cc: Mikael Starvik <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Jesper Nilsson <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: cpufreq-cpu0: use cpufreq_generic_init()

Use generic cpufreq_generic_init() routine instead of replicating the same code
here.

Cc: Shawn Guo <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: remove CONFIG_CPU_FREQ_TABLE

CONFIG_CPU_FREQ_TABLE will be always enabled when cpufreq framework is used, as
cpufreq core depends on it. So, we don't need this CONFIG option anymore as it
is not configurable. Remove CONFIG_CPU_FREQ_TABLE and update its users.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: create cpufreq_generic_init() routine

Many CPUFreq drivers for SMP system (where all cores share same clock lines), do
similar stuff in their ->init() part.

This patch creates a generic routine in cpufreq core which can be used by these
so that we can remove some redundant code.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: unicore2: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: tegra: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Stephen Warren <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: speedstep: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: David S. Miller <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: spear: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sh: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Paul Mundt <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sc520_freq: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sa11x0: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Russell King <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: s5pv210: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Kukjin Kim <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: s3c: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Kukjin Kim <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pxa: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Eric Miao <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: ppc: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: powernow: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pmac: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pcc: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: p4: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: David S. Miller <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: omap: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Santosh Shilimkar <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: maple: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Dmitry Eremin-Solenikov <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: loongson2: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: John Crispin <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: longhaul: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: kirkwood: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Tested-by: Andrew Lunn <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: integrator: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init(). And so we don't need to set policy->cur from driver anymore.

Over that it sets policy->min and max correctly. They were earlier set to
current frequency of CPU but they should be set to max and min freq of cpu.

Cc: Russell King <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: imx6q: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Shawn Guo <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: ia64-acpi: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Tony Luck <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: gx: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: exynos: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Acked-By: Amit Daniel Kachhap <[email protected]>
Acked-by: Kukjin Kim <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: elanfreq: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: e_powersaver: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: dbx500: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Linus Walleij <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: davinci: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Sekhar Nori <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: cris: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Mikael Starvik <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Jesper Nilsson <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: nforce2: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: cpu0: don't initialize part of policy set by core too

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Shawn Guo <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: blackfin: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Cc: Steven Miao <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: at32ap: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Acked-by: Hans-Christian Egtvedt <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: arm_big_little: don't initialize part of policy is set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: acpi-cpufreq: don't initialize part of policy set by core

Many common initializations of struct policy are moved to core now and hence
this driver doesn't need to do it. This patch removes such code.

Most recent of those changes is to call ->get() in the core after calling
->init().

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: call cpufreq_driver->get() after calling ->init()

Almost all drivers set policy->cur with current CPU frequency in their ->init()
part. This can be done for all of them at core level and so they wouldn't need
to do it.

This patch adds supporting code in cpufreq core for calling get() after we have
called init() for a policy.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: tegra: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the Tegra driver.

Cc: Stephen Warren <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: speedstep: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the speedstep driver.

Cc: David S. Miller <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: spear: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the spear driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sparc: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the sparc driver.

Cc: David S. Miller <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sh: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the sh driver.

Cc: Paul Mundt <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sc520: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the sc520 driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: sa11x0: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the sa11x0 driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: s5pv210: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the s5pv210 driver.

Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Kukjin Kim <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: s3cx4xx: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the s3cx4xx driver.

Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Kukjin Kim <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pxa: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the PXA driver.

Cc: Eric Miao <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: ppc_cbe: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the ppc_cbe driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: ppc-corenet: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the ppc-corenet driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: powernow: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the powernow driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pmac: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the pmac driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: pasemi: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the pasemi driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: p4-clockmod: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the p4-clockmod driver.

Cc: David S. Miller <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: omap: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the OMAP driver.

Cc: Santosh Shilimkar <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: maple: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the maple driver.

Cc: Dmitry Eremin-Solenikov <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: loongson2: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the loongson2 driver.

Cc: John Crispin <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: longhaul: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the longhaul driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: kirkwood: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the kirkwood driver.

Tested-by: Andrew Lunn <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: imx6q: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines for in the imx6q driver.

Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Shawn Guo <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: ia64-acpi: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the ia64-acpi driver.

Cc: Tony Luck <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: exynos: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the exynos driver.

Signed-off-by: Viresh Kumar <[email protected]>
Acked-By: Amit Daniel Kachhap <[email protected]>
Acked-by: Kukjin Kim <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: elanfreq: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the elanfreq driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: e_powersaver: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the e_powersaver driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: dbx500: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines for in the dbx500 driver.

Cc: Linus Walleij <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: davinci: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the davinci driver.

Cc: Sekhar Nori <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: cris: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the cris driver.

Cc: Jesper Nilsson <[email protected]>
Cc: Mikael Starvik <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Jesper Nilsson <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: cpufreq-cpu0: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the cpufreq-cpu0 driver.

Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Shawn Guo <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: blackfin: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the blackfin driver.

Cc: Steven Miao <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: at32ap: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the at32ap driver.

Acked-by: Hans-Christian Egtvedt <[email protected]>
Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: arm_big_little: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses these generic routines in the arm_big_little driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: acpi: Use generic cpufreq routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch uses the generic verify routine in the ACPI driver.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: define generic .attr, .exit() and .verify() routines

Most of the CPUFreq drivers do similar things in .exit() and .verify() routines
and .attr. So its better if we have generic routines for them which can be used
by cpufreq drivers then.

This patch introduces generic .attr, .exit() and .verify() cpufreq drivers.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: add new routine cpufreq_verify_within_cpu_limits()

Most of the users of cpufreq_verify_within_limits() calls it for
limiting with min/max from policy->cpuinfo. We can make that code
simple by introducing another routine which will do this for them
automatically.

This patch adds another routine cpufreq_verify_within_cpu_limits()
and updates others to use it.

Signed-off-by: Viresh Kumar <[email protected]>
Acked-by: Dirk Brandewie <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: use cpufreq_driver->flags to mark CPUFREQ_HAVE_GOVERNOR_PER_POLICY

Use cpufreq_driver->flags to mark CPUFREQ_HAVE_GOVERNOR_PER_POLICY instead
of a separate field within cpufreq_driver. This will save some bytes of
memory.

Signed-off-by: Viresh Kumar <[email protected]>
Reviewed-by: Srivatsa S. Bhat <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: rewrite cpufreq_driver->flags using shift operator

Currently cpufreq_driver's flags are defined directly using 0x1, 0x2, 0x4, 0x8,
etc.. As the list grows it becomes less readable..

Use bitwise shift operator << to generate these numbers for respective
positions.

Signed-off-by: Viresh Kumar <[email protected]>
Reviewed-by: Srivatsa S. Bhat <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: rename __cpufreq_set_policy() as cpufreq_set_policy()

Earlier there used to be two functions named __cpufreq_set_policy() and
cpufreq_set_policy(), but now we only have a single routine lets name it
cpufreq_set_policy() instead of __cpufreq_set_policy().

This also removes some invalid comments or fixes some incorrect comments.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>

cpufreq: Optimize cpufreq_frequency_table_verify()

cpufreq_frequency_table_verify() is rewritten here to make it more logical
and efficient.
 - merge multiple lines for variable declarations together.
 - quit early if any frequency between min/max is found.
 - don't call cpufreq_verify_within_limits() in case any valid freq is
   found as it is of no use.
 - rename the count variable as found and change its type to boolean.

Signed-off-by: Viresh Kumar <[email protected]>
Signed-off-by: Rafael J. Wysocki <[email protected]>