The critical clock handling in __clk_core_init isn't taking the enable lock
before calling clk_core_enable, which in turns triggers the warning in the
lockdep_assert_held call in that function when lockep is enabled.

Add the calls to clk_enable_lock/unlock to make sure it doesn't happen.

Fixes: 32b9b109 ("clk: Allow clocks to be marked as CRITICAL")
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Tested-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

The branch clocks MMSS_MMAGIC_AXI_CLK and MMAGIC_BIMC_AXI_CLK are
controlled by RPM when the APPs processor enable or disable the
RPM_MMAXI_CLK.

During the boot sequence, someone can enable the RPM_MMAXI_CLK, resulting
in register status bits showing that these clocks are enabled, our
clock driver may look at the enabled status of these clocks and try to
disable them since it thinks they are unused.

Don't make the clock driver touch these clocks.
Signed-off-by: Archit Taneja <architt@codeaurora.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

Merge tag 'imx-clk-fixes-4.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux into clk-next

Pull some non-critical i.MX clk fixes from Shawn Guo:

 * Fix the commit 3713e3f5 ("clk: imx35: define two clocks for rtc")
   which messed up the clock enumeration when adding new clock.

* tag 'imx-clk-fixes-4.6-2' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
  ARM: dts: imx35: restore existing used clock enumeration
  clk: imx6q: fix typo in CAN clock definition

The JZ4780's MSC clock divider registers multiply the clock divider by 2.
This means that MMC devices run at half their expected speed. Add the
ability to divide the clock divider in order to solve this.
Signed-off-by: Harvey Hunt <harvey.hunt@imgtec.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-clk@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

Merge tag 'v4.7-rockchip-clk4' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into clk-next

Pull rockchip clk updates from Heiko Stuebner:

Another small rk3399 fixup as well as simplifications around
our handling of the General-Register-Files syscon.

* tag 'v4.7-rockchip-clk4' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:
  clk: rockchip: drop old_rate calculation on pll rate changes
  clk: rockchip: simplify GRF handling in pll clocks
  clk: rockchip: lookup General Register Files in rockchip_clk_init
  clk: rockchip: fix the rk3399 sdmmc sample / drv name

The A10 SoCs and its relatives has a special clock controller to drive the
display engines (both frontend and backend), that have a lot in common with
the clock to drive the first TCON channel.

Add a driver to support both.
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Rob Herring <robh@kernel.org>
[sboyd@codeaurora.org: Silence variable sized array warning]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

Previously when everything happened in the set_rate callbacks itself we
needed the old_rate value for the possible rate rollback, so that made
it easy to also use it in the debug output.

Now with the param-handling being done in separate functions, reading and
recalculating the current pll rate only to use it in a debug message that
won't get displayed in regular cases anyway is quite a waste.

Therefore drop that value from the debug output. In the worst case that
previous rate will have been displayed on the rate change before.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>

With the previous commit, the clock drivers now know at init time if the
GRF regmap is available. That means if it isn't available then, it also
won't become available later and we can therefore switch PLLs, that need
the GRF for the lock-status, to read-only mode - similar behaviour as the
aborting of rate changes we did before.

This saves some conditionals on every rate change and we can also drop
the rockchip_clk_get_grf function completely.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>

In the distant past syscons were initialized pretty late and weren't
available at the time the clock init ran. As the GRF is mainly needed
for PLL lock-status checking, we had this lazy init that tried to grab
the syscon on PLL rate changes and denied these changes if it was not
available.

These days syscons are available very early and recent addition to
rockchip clocks, like the PLL clk_init actually also rely on them
being available at that time, so there is no need to keep that lazy
init around, as it will also result in some more simplifications in
other parts of the clock-code.
Signed-off-by: Heiko Stuebner <heiko@sntech.de>

The rk3399 clock table had a simple typo in it, calling the SDMMC sample
and drive clocks by the wrong name.  Fix this minor typo.
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Tested-by: Brian Norris <briannorris@chromium.org>
Acked-by: Brian Norris <briannorris@chromium.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>

The Armada CP110 system controller provides, amongst other things, a
number of clocks for the platform: a small number of core clocks, and
then a number of gatable clocks, derived from some of the core
clocks. Those clocks are configured via registers of the CP110 System
Controller.

The CP110 is the other core HW block (next to the AP806) used in the
Marvel Armada 7K and 8K SoCs.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[sboyd@codeaurora.org: Silence some checkpatch noise]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

This commit adds the DT binding documentation for the Marvell CP110
system controller, which is part of the CP110 HW block, itself used in
the Marvell Armada 7K and 8K SoCs.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Rob Herring <rob@kernel.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

The Armada AP806 system controller, amongst other things, provides a
number of clocks for the platform: the CPU cluster clocks, whose
frequencies are found by reading the Sample At Reset register, one
fixed clock, and another clock derived from the fixed clock, which is
the one used by most peripherals in AP806.

The AP806 is one of the two core HW blocks used in the Marvell 7K/8K
SoCs.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
[sboyd@codeaurora.org: Silence some checkpatch noise]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

* clk-hi3519:
  clk: hisilicon: add CRG driver for hi3519 soc
  clk: hisilicon: export some hisilicon APIs to modules
  reset: hisilicon: add reset controller driver for hisilicon SOCs

The CRG(Clock and Reset Generator) block provides clock
and reset signals for other modules in hi3519 soc.
Signed-off-by: Jiancheng Xue <xuejiancheng@hisilicon.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

Change some arguments to constant type.
Export some hisilicon APIs to modules.
Signed-off-by: Jiancheng Xue <xuejiancheng@huawei.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

In most of hisilicon SOCs, reset controller and clock provider are
combined together as a block named CRG (Clock and Reset Generator).
This patch mainly implements the reset function.
Signed-off-by: Jiancheng Xue <xuejiancheng@hisilicon.com>
Acked-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

When sizeof is applied to a pointer typed expression, it gives
the size of the pointer. So, here do not use sizeof on pointer
type. Also, silent checkpatch.pl by using kmalloc_array over
kmalloc.

Note that this has no effect on runtime because 'parent_names'
is a pointer to a pointer.

Problem found using Coccinelle.
Signed-off-by: Vaishali Thakkar <vaishali.thakkar@oracle.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

This patch adds the CLK_SET_RATE_PARENT flag for the crypto core and
ahb blocks.  Without this flag, clk_set_rate can fail for certain
frequency requests.
Signed-off-by: Andy Gross <andy.gross@linaro.org>
Fixes: 3966fab8 ("clk: qcom: Add MSM8916 Global Clock Controller support")
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

Merge tag 'imx-clk-4.7-2' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux into clk-next

Pull i.MX clk updates from Shawn Guo:

 - Update clk-pllv3 driver to get it return correct frequency for
   Ethernet PLL on i.MX7D.
 - Correct ahb clock mux settings for i.MX7D per latest hardware
   document.

* tag 'imx-clk-4.7-2' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
  clk: imx7d: fix ahb clock mux 1
  clk: imx: return correct frequency for Ethernet PLL

Failure to initialize this flag to 0 by default can result in
stack junk filling the clk_init_data structure and weird things
happen. Joachim noticed that the critical clk feature started
triggering for these clks causing boot failures, when it really
shouldn't have happened:

BUG: scheduling while atomic: swapper/0/0x00000002
CPU: 0 PID: 0 Comm: swapper Not tainted
4.6.0-rc6-next-20160505-00001-g5c8320450d1c #826
Hardware name: NXP LPC18xx/43xx (Device Tree)
[<2800be81>] (unwind_backtrace) from [<2800b22f>] (show_stack+0xb/0xc)
[<2800b22f>] (show_stack) from [<2801ea21>] (__schedule_bug+0x2d/0x44)
[<2801ea21>] (__schedule_bug) from [<281dc937>] (__schedule+0x3b/0x268)
[<281dc937>] (__schedule) from [<281dcbbb>] (schedule+0x57/0x64)
[<281dcbbb>] (schedule) from [<281de8ef>] (schedule_timeout+0xfb/0x120)
[<281de8ef>] (schedule_timeout) from [<28030fcd>] (msleep+0xf/0x12)
[<28030fcd>] (msleep) from [<28165a6d>] (clk_creg_32k_prepare+0x1f/0x24)
[<28165a6d>] (clk_creg_32k_prepare) from [<281620d5>]
(clk_core_prepare+0x1d/0x36)
[<281620d5>] (clk_core_prepare) from [<2816340b>] (clk_register+0x22f/0x318)
[<2816340b>] (clk_register) from [<282b06c9>] (lpc18xx_creg_clk_init+0x55/0x84)
[<282b06c9>] (lpc18xx_creg_clk_init) from [<282b0149>] (of_clk_init+0xc1/0x12c)
[<282b0149>] (of_clk_init) from [<282a665d>] (time_init+0x15/0x20)
[<282a665d>] (time_init) from [<282a457d>] (start_kernel+0x169/0x274)
[<282a457d>] (start_kernel) from [<28008025>] (0x28008025)
bad: scheduling from the idle thread!
CPU: 0 PID: 0 Comm: swapper Tainted: G W
4.6.0-rc6-next-20160505-00001-g5c8320450d1c #826

Fix this by initializing the flags member to 0.
Acked-by: Joachim Eastwood <manabian@gmail.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

The ARC SDP I2S clock can be programmed using a
specific PLL.

This patch has the goal of adding a clock driver
that programs this PLL.

At this moment the rate values are hardcoded in
a table but in the future it would be ideal to
use a function which determines the PLL values
given the desired rate.
Signed-off-by: Jose Abreu <joabreu@synopsys.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

The clock parent of the AHB root clock when using mux option 1
is the SYS PLL 270MHz clock. This is specified in  Table 5-11
Clock Root Table of the i.MX 7Dual Applications Processor
Reference Manual.

While it could be a documentation error, the 270MHz parent is
also mentioned in the boot ROM configuration in Table 6-28: The
clock is by default at 135MHz due to a POST_PODF value of 1
(=> divider of 2).
Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>

Unlike devm_clk_register(), devm_clk_hw_register() returns integer.
So, the statement "Clocks returned from this function ..." sounds
odd.  Adjust the comment for this new API.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>

Merge tag 'clk-renesas-for-v4.7-tag3' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-drivers into clk-next

Pull Renesas clk driver updates from Geert Uytterhoeven:

  - Support for CSI2 and VIN module clocks on R-Car H3,
  - Renesas CPG/MSTP and CPG/MSSR Clock Domain fixes.

* tag 'clk-renesas-for-v4.7-tag3' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-drivers:
  clk: renesas: cpg-mssr: Use always-on governor for Clock Domain
  clk: renesas: cpg-mssr: Postpone call to pm_genpd_init()
  clk: renesas: mstp: Use always-on governor for Clock Domain
  clk: renesas: mstp: Postpone call to pm_genpd_init()
  clk: renesas: r8a7795: Add VIN clocks
  clk: renesas: r8a7795: Add CSI2 clocks

Merge tag 'sunxi-clocks-for-4.7' of https://git.kernel.org/pub/scm/linux/kernel/git/mripard/linux into clk-next

Pull Allwinner clock driver updates from Maxime Ripard:

As usual, a bunch of clocks patches for 4.7, mostly fixes and cleanups, and
display-related clocks.

* tag 'sunxi-clocks-for-4.7' of https://git.kernel.org/pub/scm/linux/kernel/git/mripard/linux:
  clk: sunxi: Let divs clocks read the base factor clock name from devicetree
  clk: sunxi: Add TCON channel1 clock
  clk: sunxi: Add PLL3 clock
  dt-bindings: clk: sun5i: add DRAM gates compatible
  clk: sunxi: Use resource_size
  clk: sunxi: Add sun6i/8i display support
  clk: sunxi: mod1 clock should modify it's parent

Merge tag 'tegra-for-4.7-clk' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux into clk-next

Pull tegra clk driver changes from Thierry Reding:

This set of changes contains a bunch of cleanups and minor fixes along
with some new clocks, mainly on Tegra210, in preparation for supporting
DisplayPort and HDMI 2.0.

* tag 'tegra-for-4.7-clk' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux:
  clk: tegra: dfll: Reformat CVB frequency table
  clk: tegra: dfll: Properly clean up on failure and removal
  clk: tegra: dfll: Make code more comprehensible
  clk: tegra: dfll: Reference CVB table instead of copying data
  clk: tegra: dfll: Update kerneldoc
  clk: tegra: Fix PLL_U post divider and initial rate on Tegra30
  clk: tegra: Initialize PLL_C to sane rate on Tegra30
  clk: tegra: Fix pllre Tegra210 and add pll_re_out1
  clk: tegra: Add sor_safe clock
  clk: tegra: dpaux and dpaux1 are fixed factor clocks
  clk: tegra: Add dpaux1 clock
  clk: tegra: Use correct parent for dpaux clock
  clk: tegra: Add fixed factor peripheral clock type
  clk: tegra: Special-case mipi-cal parent on Tegra114
  clk: tegra: Remove trailing blank line
  clk: tegra: Constify peripheral clock registers
  clk: tegra: Add interface to enable hardware control of SATA/XUSB PLLs

Merge tag 'v4.7-rockchip-clk3' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip into clk-next

Pull rockchip clk updates from Heiko Stuebner:

A spelling fix and a bunch of rk3399 clock fixes.

* tag 'v4.7-rockchip-clk3' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:
  clk: rockchip: fix the rk3399 cifout clock
  clk: rockchip: drop unnecessary CLK_IGNORE_UNUSED flags from rk3399
  clk: rockchip: add some frequencies on the rk3399 PLL table
  clk: rockchip: assign more necessary rk3399 clock ids
  clk: rockchip: export some necessary rk3399 clock ids
  clk: rockchip: rename rga clock-id on rk3399
  clk: rockchip: add general gpu soft-reset on rk3399
  clk: rockchip: fix the gate bit for i2c4 and i2c8 on rk3399
  clk: rockchip: fix of spelling mistake on unsuccessful in pll clock type

Increase the readability of the CVB frequency table by reformatting it a
little.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Upon failure to probe the DFLL, the OPP table will not be cleaned up
properly. Fix this and while at it make sure the OPP table will also be
cleared upon driver removal.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Rename some variables and structure fields to make the code more
comprehensible. Also change the prototype of internal functions to be
more in line with the OPP core functions.
Signed-off-by: Thierry Reding <treding@nvidia.com>

Instead of copying parts of the CVB table into a separate structure,
keep track of the selected CVB table and directly reference data from
it.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The kerneldoc for struct tegra_dfll_soc_data is stale. Update it to
match the current structure definition.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The post divider value in the frequency table is wrong as it would lead
to the PLL producing an output rate of 960 MHz instead of the desired
480 MHz. This wasn't a problem as nothing used the table to actually
initialize the PLL rate, but the bootloader configuration was used
unaltered.

If the bootloader does not set up the PLL it will fail to come when used
under Linux. To fix this don't rely on the bootloader, but set the
correct rate in the clock driver.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Thierry Reding <treding@nvidia.com>

If the bootloader does not touch PLL_C it will stay in its reset state,
failing to lock when enabled. This leads to consumers of this clock to
fail probing. Fix this by always programming the PLL with a sane rate,
which allows it to lock, at startup.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Thierry Reding <treding@nvidia.com>

Use a new Tegra210 version of the pll_register_pllre function to
allow setting the proper settings for the m and n div fields.

Additionally define PLL_RE_OUT1 on Tegra210.
Signed-off-by: Rhyland Klein <rklein@nvidia.com>
[treding@nvidia.com: define PLLRE_OUT1 register offset]
Signed-off-by: Thierry Reding <treding@nvidia.com>

The sor_safe clock is a fixed factor (1:17) clock derived from pll_p. It
has a gate bit in the peripheral clock registers. While the SOR is being
powered up, sor_safe can be used as the source until the SOR brick can
generate its own clock.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The dpaux (on Tegra124 and Tegra210) and dpaux1 (on Tegra210) are fixed
factor clocks (1:17) and derived from pll_p_out0 (pll_p). They also have
a gate bit in the peripheral clock registers.
Signed-off-by: Thierry Reding <treding@nvidia.com>

This clock is of the same type as dpaux and is added to feed into the
second DPAUX block used in conjunction with SOR1.
Signed-off-by: Thierry Reding <treding@nvidia.com>

The dpaux clock is derived from pll_p_out0 (pll_p), not clk_m.
Signed-off-by: Thierry Reding <treding@nvidia.com>