The OPP list needs to be protected against concurrent accesses. Using
simple RCU read locks does the trick and gets rid of the following
lockdep warning:

	===============================
	[ INFO: suspicious RCU usage. ]
	4.2.0-next-20150908 #1 Not tainted
	-------------------------------
	drivers/base/power/opp.c:460 Missing rcu_read_lock() or dev_opp_list_lock protection!

	other info that might help us debug this:

	rcu_scheduler_active = 1, debug_locks = 0
	4 locks held by kworker/u8:0/6:
	 #0:  ("%s""deferwq"){++++.+}, at: [<c0040d8c>] process_one_work+0x118/0x4bc
	 #1:  (deferred_probe_work){+.+.+.}, at: [<c0040d8c>] process_one_work+0x118/0x4bc
	 #2:  (&dev->mutex){......}, at: [<c03b8194>] __device_attach+0x20/0x118
	 #3:  (prepare_lock){+.+...}, at: [<c054bc08>] clk_prepare_lock+0x10/0xf8

	stack backtrace:
	CPU: 2 PID: 6 Comm: kworker/u8:0 Not tainted 4.2.0-next-20150908 #1
	Hardware name: NVIDIA Tegra SoC (Flattened Device Tree)
	Workqueue: deferwq deferred_probe_work_func
	[<c001802c>] (unwind_backtrace) from [<c00135a4>] (show_stack+0x10/0x14)
	[<c00135a4>] (show_stack) from [<c02a8418>] (dump_stack+0x94/0xd4)
	[<c02a8418>] (dump_stack) from [<c03c6f6c>] (dev_pm_opp_find_freq_ceil+0x108/0x114)
	[<c03c6f6c>] (dev_pm_opp_find_freq_ceil) from [<c0551a3c>] (dfll_calculate_rate_request+0xb8/0x170)
	[<c0551a3c>] (dfll_calculate_rate_request) from [<c0551b10>] (dfll_clk_round_rate+0x1c/0x2c)
	[<c0551b10>] (dfll_clk_round_rate) from [<c054de2c>] (clk_calc_new_rates+0x1b8/0x228)
	[<c054de2c>] (clk_calc_new_rates) from [<c054e44c>] (clk_core_set_rate_nolock+0x44/0xac)
	[<c054e44c>] (clk_core_set_rate_nolock) from [<c054e4d8>] (clk_set_rate+0x24/0x34)
	[<c054e4d8>] (clk_set_rate) from [<c0512460>] (tegra124_cpufreq_probe+0x120/0x230)
	[<c0512460>] (tegra124_cpufreq_probe) from [<c03b9cbc>] (platform_drv_probe+0x44/0xac)
	[<c03b9cbc>] (platform_drv_probe) from [<c03b84c8>] (driver_probe_device+0x218/0x304)
	[<c03b84c8>] (driver_probe_device) from [<c03b69b0>] (bus_for_each_drv+0x60/0x94)
	[<c03b69b0>] (bus_for_each_drv) from [<c03b8228>] (__device_attach+0xb4/0x118)
	ata1: SATA link down (SStatus 0 SControl 300)
	[<c03b8228>] (__device_attach) from [<c03b77c8>] (bus_probe_device+0x88/0x90)
	[<c03b77c8>] (bus_probe_device) from [<c03b7be8>] (deferred_probe_work_func+0x58/0x8c)
	[<c03b7be8>] (deferred_probe_work_func) from [<c0040dfc>] (process_one_work+0x188/0x4bc)
	[<c0040dfc>] (process_one_work) from [<c004117c>] (worker_thread+0x4c/0x4f4)
	[<c004117c>] (worker_thread) from [<c0047230>] (kthread+0xe4/0xf8)
	[<c0047230>] (kthread) from [<c000f7d0>] (ret_from_fork+0x14/0x24)
Signed-off-by: NThierry Reding <treding@nvidia.com>
Fixes: c4fe70ad ("clk: tegra: Add closed loop support for the DFLL")
[vince.h@nvidia.com: Unlock rcu on error path]
Signed-off-by: NVince Hsu <vince.h@nvidia.com>
[sboyd@codeaurora.org: Dropped second hunk that nested the rcu
read lock unnecessarily]
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Again a result of the gpio-clock-liberation the rk3368 needs the
pclk_pd_pmu marked as critical, to boot successfully.
Reported-by: NMark Rutland <mark.rutland@arm.com>
Signed-off-by: NHeiko Stuebner <heiko@sntech.de>
Tested-by: NMark Rutland <mark.rutland@arm.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Now that the rockchip clock subsystem does clock gating with GPIO banks,
these are no longer enabled once during probe and no longer stay enabled
for eternity. When all these clocks are disabled, the parent clock pclk_peri
might be disabled too, as no other child claims it. So, we need to add pclk_peri
to the critical clocks.
Signed-off-by: NRomain Perier <romain.perier@gmail.com>
Tested-by: NMichael Niewoehner <linux@mniewoehner.de>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

pclk_cpu needs to keep running because it is needed for devices like
the act8865 regulator but with the recent gpio clock handling this is
not always the case anymore. So add it to the list of critical clocks.
Signed-off-by: NMichael Niewoehner <linux@mniewoehner.de>
Reviewed-by: NHeiko Stuebner <heiko@sntech.de>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Currently the registration of critical clocks is done in the function
shared between rk3066 and rk3188 clock trees. That results in them
getting handled maybe before all of them are registered.

Therefore move the critical clock handling down to the end of the soc-
specific clock registration function, so that all clocks are registered
before they're maybe handled as critical clock.
Signed-off-by: NHeiko Stuebner <heiko@sntech.de>
Tested-by: NMichael Niewoehner <linux@mniewoehner.de>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

The previous code, kernel builds Hi6220's common clock driver and stub
clock driver together. Stub clock driver has introduced the dependency
with CONFIG_MAILBOX, so kernel will not build Hi6220's common clock
driver due ARM64's defconfig have not enabled CONFIG_MAILBOX by default.

So separately build stub clock driver and common clock driver for
Hi6220; and only let stub clock driver has the dependency with
CONFIG_MAILBOX.
Signed-off-by: NLeo Yan <leo.yan@linaro.org>
Tested-by: NKevin Hilman <khilman@linaro.org>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Recent cleanup removed some include files without checking if the cleaned
up code still compiles. This results in the following compile error.

drivers/clk/h8300/clk-h8s2678.c: In function ‘h8s2678_pll_clk_setup’:
drivers/clk/h8300/clk-h8s2678.c:99:14: error:
	implicit declaration of function ‘kzalloc’
drivers/clk/h8300/clk-h8s2678.c:138:2: error:
      implicit declaration of function ‘kfree’

Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Signed-off-by: NGuenter Roeck <linux@roeck-us.net>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Commit d5e136a2 ("clk: samsung: Register
clk provider only after registering its all clocks", merged to v3.17-rc1)
modified a way that driver registers registers to core framework. This
change has not been applied to s5pv210 clocks driver, which has been
merged in parallel to that commit. This patch adds a missing call to
samsung_clk_of_add_provider(), so the driver is operational again.
Signed-off-by: NMarek Szyprowski <m.szyprowski@samsung.com>
CC: stable@vger.kernel.org  # v3.17+
Acked-by: NTomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: NMichael Turquette <mturquette@baylibre.com>

Current critical clock list for pistachio enables
only mips and sys clocks by default but there are
also other clocks that are not claimed by anyone and
needs to be enabled by default.

This patch updates the critical clocks that need
to be enabled by default.

Add a separate struct to distinguish the critical clocks
as listed:
1.) core clocks:
	a.) mips clock
2.) peripheral system clocks:
	a.) sys clock
	b.) sys_bus clock
	c.) DDR clock
	d.) ROM clock

Fixes: b35d7c33("CLK: Pistachio: Register core clocks")
Cc: <stable@vger.kernel.org> # 4.1
Reviewed-by: NAndrew Bresticker <abrestic@chromium.org>
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@imgtec.com>
Signed-off-by: NDamien.Horsley <Damien.Horsley@imgtec.com>
Signed-off-by: NGovindraj Raja <govindraj.raja@imgtec.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

.recalc_rate callback for the fractional PLL doesn't take operating
mode into account when calculating PLL rate. This results in
the incorrect PLL rates when PLL is operating in integer mode.

Operating mode of fractional PLL is based on the value of the
fractional divider. Currently it assumes that the PLL will always
be configured in fractional mode which may not be
the case. This may result in the wrong output frequency.

Also vco was calculated based on the current operating mode which
makes no sense because .set_rate is setting operating mode. Instead,
vco should be calculated using PLL settings that are about to be set.

Fixes: 43049b0c("CLK: Pistachio: Add PLL driver")
Cc: <stable@vger.kernel.org> # 4.1
Reviewed-by: NAndrew Bresticker <abrestic@chromium.org>
Signed-off-by: NZdenko Pulitika <zdenko.pulitika@imgtec.com>
Signed-off-by: NGovindraj Raja <govindraj.raja@imgtec.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

PLL enable callbacks are overriding PLL mode (int/frac) and
Noise reduction (on/off) settings set by the boot loader which
results in the incorrect clock rate.

PLL mode and noise reduction are defined by the DSMPD and DACPD bits
of the PLL control register. PLL .enable() callbacks enable PLL
by deasserting all power-down bits of the PLL control register,
including DSMPD and DACPD bits, which is not necessary since
these bits don't actually enable/disable PLL.

This commit fixes the problem by removing DSMPD and DACPD bits
from the "PLL enable" mask.

Fixes: 43049b0c("CLK: Pistachio: Add PLL driver")
Cc: <stable@vger.kernel.org> # 4.1
Reviewed-by: NAndrew Bresitcker <abrestic@chromium.org>
Signed-off-by: NZdenko Pulitika <zdenko.pulitika@imgtec.com>
Signed-off-by: NGovindraj Raja <govindraj.raja@imgtec.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

This commit fixes 32bit integer overflows throughout the pll driver
(i.e. wherever the result of integer multiplication may exceed the
range of u32).

One of the functions affected by this problem is .recalc_rate. It
returns incorrect rate for some pll settings (not for all though)
which in turn results in the incorrect rate setup of pll's child
clocks.

Fixes: 43049b0c("CLK: Pistachio: Add PLL driver")
Cc: <stable@vger.kernel.org> # 4.1
Reviewed-by: NAndrew Bresticker <abrestic@chromium.org>
Signed-off-by: NZdenko Pulitika <zdenko.pulitika@imgtec.com>
Signed-off-by: NGovindraj Raja <govindraj.raja@imgtec.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

The latest Tegra clk pull had some problems. Fix them.

drivers/clk/tegra/clk-tegra124.c:1450:6: warning: symbol 'tegra124_clock_assert_dfll_dvco_reset' was not declared. Should it be static?
drivers/clk/tegra/clk-tegra124.c:1466:6: warning: symbol 'tegra124_clock_deassert_dfll_dvco_reset' was not declared. Should it be static?
drivers/clk/tegra/clk-tegra124.c:1476:5: warning: symbol 'tegra124_reset_assert' was not declared. Should it be static?
drivers/clk/tegra/clk-tegra124.c:1486:5: warning: symbol 'tegra124_reset_deassert' was not declared. Should it be static?
drivers/clk/tegra/clk-dfll.c:590 dfll_load_i2c_lut() warn: inconsistent indenting
drivers/clk/tegra/clk-dfll.c:1448 dfll_build_i2c_lut() warn: unsigned 'td->i2c_lut[0]' is never less than zero.
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Merge tag 'tegra-for-4.3-clk' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux into clk-next

clk: tegra: Changes for v4.3-rc1

This contains the DFLL driver needed to implement CPU frequency scaling
on Tegra.

Fix the enable bit of the pseudorandom number generator clock.
Reported-by: NStanimir Varbanov <stanimir.varbanov@linaro.org>
Fixes: 3966fab8 "clk: qcom: Add MSM8916 Global Clock Controller support"
Signed-off-by: NGeorgi Djakov <georgi.djakov@linaro.org>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

of_clk_set_defaults uses the type 'bool', but clk-conf.h does not
include its definition.

This results in a compile error when only clk-conf.h is used.
Signed-off-by: NChen-Yu Tsai <wens@csie.org>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Init data marked const should be annotated with __initconst for
correctness and not __initdata.  This also fixes LTO builds that
otherwise fail with section mismatch errors.
Signed-off-by: NNicolas Pitre <nico@linaro.org>
Signed-off-by: NMichael Turquette <mturquette@baylibre.com>
[sboyd@codeaurora.org: Dropped hunks that moved const char *
arrays to const char * const]
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

The structure is xin24m -> pll -> pll-mux (xin24m,pll,xin32k). The pll
does have an init callback to make sure the boot-selected frequency is
using the expected pll settings and resets the same frequency using
the values provided in the driver if necessary.

The setting itself also involves remuxing the pll-mux temporarily to
the xin24m source to let the new pll rate settle. Until now this worked
flawlessly, even when it had the flaw of accessing the mux settings
before the mux actually got registered.

With the recent clock-core conversions this flaw became apparent in
null pointer dereference in
[<c03fc400>] (clk_hw_get_num_parents) from [<c0400df0>] (clk_mux_get_parent+0x14/0xc8)
[<c0400ddc>] (clk_mux_get_parent) from [<c040246c>] (rockchip_rk3066_pll_set_rate+0xd8/0x320)

So to fix that, simply register the pll-mux before the pll, so that
it will be fully initialized when the pll clock executes its init-
callback and possibly touches the pll-mux clock.
Signed-off-by: NHeiko Stuebner <heiko@sntech.de>
Signed-off-by: NMichael Turquette <mturquette@baylibre.com>

These Ux500 clocks have been around for years and were never
properly documented. Add the proper binding documentation.

Cc: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NMichael Turquette <mturquette@baylibre.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

The base addresses for the Ux500 PRCC controllers are hardcoded,
let's move them to the clock node in the device tree and delete
the constants.

Cc: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Acked-by: NOlof Johansson <olof@lixom.net>
Acked-by: NMichael Turquette <mturquette@baylibre.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

__clk_set_parent_after() actually used the second argument then we
could put this duplicate logic in there and call it with a different
order of arguments in the success vs. error paths in this function.

Cc: Mike Turquette <mturquette@linaro.org>
Suggested-by: NStephen Boyd <sboyd@codeaurora.org>
Signed-off-by: NDong Aisheng <aisheng.dong@freescale.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Use the provider based method to get a clock's name so that we
can get rid of the clk member in struct clk_hw one day. Mostly
converted with the following coccinelle script.

@@
struct clk_hw *E;
@@

-__clk_get_name(E->clk)
+clk_hw_get_name(E)
Acked-by: NHeiko Stuebner <heiko@sntech.de>
Cc: Sylwester Nawrocki <s.nawrocki@samsung.com>
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Cc: Peter De Schrijver <pdeschrijver@nvidia.com>
Cc: Prashant Gaikwad <pgaikwad@nvidia.com>
Cc: Stephen Warren <swarren@wwwdotorg.org>
Acked-by: NThierry Reding <treding@nvidia.com>
Cc: Thierry Reding <thierry.reding@gmail.com>
Cc: Alexandre Courbot <gnurou@gmail.com>
Cc: Tero Kristo <t-kristo@ti.com>
Cc: Ulf Hansson <ulf.hansson@linaro.org>
Acked-by: NSebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Acked-by: NAndrew Bresticker <abrestic@chromium.org>
Cc: Ezequiel Garcia <ezequiel.garcia@imgtec.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Kevin Cernekee <cernekee@chromium.org>
Acked-by: NGeert Uytterhoeven <geert+renesas@glider.be>
Cc: Ulrich Hecht <ulrich.hecht+renesas@gmail.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-rockchip@lists.infradead.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: linux-tegra@vger.kernel.org
Cc: linux-omap@vger.kernel.org
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

We don't modify the clk_hw argument in these functions, so it's
safe to mark it as const.
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

On Hi6220, there have some clocks which can use mailbox channel to send
messages to power controller to change frequency; this includes CPU, GPU
and DDR clocks.

For dynamic frequency scaling, firstly need write the frequency value to
SRAM region, and then send message to mailbox to trigger power controller
to handle this requirement. This driver will use syscon APIs to pass SRAM
memory region and use common mailbox APIs for channels accessing.

This init driver will support cpu frequency change firstly.
Signed-off-by: NLeo Yan <leo.yan@linaro.org>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Document the new compatible for stub clock driver which is used for CPU
and DDR's dynamic frequency scaling.
Signed-off-by: NLeo Yan <leo.yan@linaro.org>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Document "hisilicon,hi6220-sramctrl" for SRAM controller.
Signed-off-by: NLeo Yan <leo.yan@linaro.org>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

PLL VCO frequency is given by Fvco = Fref * 2 * NF / NR
in integer-N mode, or by Fvco = Fref * SSN / NR in Spread
Spectrum (fractional-N) mode. Thus fix the missing part
of NR
Signed-off-by: NGuo Zeng <guo.zeng@csr.com>
Signed-off-by: NBarry Song <Baohua.Song@csr.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

bit4 thcpum_cpudiv4_clken 	r/w
thcpum_cpudiv4_clk clock enable (default: 1)
Root clock CPU_CLK must be enabled for this clock to be enabled

bit3 coresight_tpiu_clken 	r/w
coresight_tpiu_clk clock enable (default: 0)
Root clock TPIU_CLK must be enabled for this clock to be enabled
Signed-off-by: NGuo Zeng <guo.zeng@csr.com>
Signed-off-by: NBarry Song <Baohua.Song@csr.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Signed-off-by: NYibo Cai <yibo.cai@csr.com>
Signed-off-by: NBarry Song <Baohua.Song@csr.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

I cannot believe that I spend quite a lot time in finding this bug.
It seems a pitfall people tend to fall in.

In "int64 = int32 * int32", conversion from 32-bits to 64-bits comes
after the multiplication. So this statement may not work as expected.
Signed-off-by: NYibo Cai <yibo.cai@csr.com>
Signed-off-by: NBarry Song <Baohua.Song@csr.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

this patch makes mux clks can be referenced by device drivers.
Signed-off-by: NGuo Zeng <guo.zeng@csr.com>
Signed-off-by: NBarry Song <Barry.Song@csr.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Signed-off-by: NGuo Zeng <Guo.Zeng@csr.com>
Signed-off-by: NBarry Song <Baohua.Song@csr.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

The debug base clock can be bypassed from the main PLL to the OSC1 clock.
The bypass register is the staysoc1(0x10) register that is in the clock
manager.

This patch adds the option to get the correct parent for the debug base
clock.
Signed-off-by: NDinh Nguyen <dinguyen@opensource.altera.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

This code is unused and not coming back. Let's kill it off.

Cc: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Remove these APIs now that we've converted all users to the
replacement struct clk_hw based versions.
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

This driver uses __clk_get_name() when it's dealing with struct
clk_hw pointers. Use the simpler form so that we can get rid of
the clk member in struct clk_hw one day.
Acked-by: NSebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

We're removing struct clk from the clk provider API, so switch
this code to using the clk_hw based provider APIs.

Cc: Wentao Xu <wentaox@codeaurora.org>
Cc: Rob Clark <robdclark@gmail.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

We're removing struct clk from the clk provider API. This code is
calling the consumer APIs to change the parent to a 1 MHz fixed
rate clock for each of the clocks that the driver provides. Move
to using the assigned-clock-parents DT property for this instead.
Because this is an ABI break, detect if the property is missing
and fall back to setting the parent explicitly before the clocks
are registered.
Acked-by: NPawel Moll <pawel.moll@arm.com>
Cc: Linus Walleij <linus.walleij@linaro.org>
Tested-by: NSudeep Holla <sudeep.holla@arm.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

Signed-off-by: NPawel Moll <pawel.moll@arm.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>

We're removing struct clk from the clk provider API, so switch
this code to using the clk_hw based provider APIs.
Acked-by: NTero Kristo <t-kristo@ti.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>