GitHub/moto-9609/android_kernel_motorola_exynos9610.git
7 years agoMerge tag 'amlogic-clk' of git://git.kernel.org/pub/scm/linux/kernel/git/khilman...
Michael Turquette [Wed, 12 Apr 2017 16:51:43 +0000 (18:51 +0200)]
Merge tag 'amlogic-clk' of git://git./linux/kernel/git/khilman/linux-amlogic into clk-next

Same great taste as the previous pull request, but now with 50% less DT
bikeshedding!

Amlogic clock driver updates for v4.12
- meson8: add some new PLLs
- new clocks for Mali
- misc fixes.

7 years agoclk: aggregate return codes of notify chains
Peter De Schrijver [Tue, 21 Mar 2017 10:16:26 +0000 (12:16 +0200)]
clk: aggregate return codes of notify chains

In case there are multiple notify chains for the same clocks (because they
were registered by different users), we need to propagate potential failure
of any single one of them to the caller. Otherwise we eg risk violating the
V/f curve when a notifier is used for DVFS.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: add clk_possible_parents debugfs file
Peter De Schrijver [Tue, 21 Mar 2017 13:20:31 +0000 (15:20 +0200)]
clk: add clk_possible_parents debugfs file

For validation purposes, it's often useful to be able to retrieve the list
of possible parents in userspace. Add a debugfs file for every clock which
has more than 1 possible parent.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Jon Mayo <jmayo@nvidia.com>
[sboyd@codeaurora.org: Remove useless cast from void and extra
newline]
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: imx: correct uart4_serial clock name in driver for i.MX6UL
Robin van der Gracht [Mon, 6 Mar 2017 08:13:43 +0000 (09:13 +0100)]
clk: imx: correct uart4_serial clock name in driver for i.MX6UL

Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Robin van der Gracht <robin@protonic.nl>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: zte: Mark pll config tables as const
Stephen Boyd [Fri, 7 Apr 2017 19:21:33 +0000 (12:21 -0700)]
clk: zte: Mark pll config tables as const

These should be const.

Cc: Shawn Guo <shawn.guo@linaro.org>
Cc: Jun Nie <jun.nie@linaro.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: zte: add pll_vga clock for zx296718
Shawn Guo [Tue, 21 Mar 2017 08:38:23 +0000 (16:38 +0800)]
clk: zte: add pll_vga clock for zx296718

It adds zx296718 pll_vga clock for VGA support, so that VGA device can
get required pixel rate from clock driver for different display mode.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Reviewed-by: Jun Nie <jun.nie@linaro.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: zte: pd_bit is not 0 on zx296718
Shawn Guo [Tue, 21 Mar 2017 08:38:22 +0000 (16:38 +0800)]
clk: zte: pd_bit is not 0 on zx296718

The bit 0 of PLL_CFG0 register is not powerdown on zx296718, but part of
of postdiv2 field.  The consequence is that functions like hw_to_idx()
and zx_pll_enable() will end up tampering the postdiv2 of the PLL.

Let's fix it by defining pd_bit 0xff which is obviously invalid for a
bit position and having PLL driver check the validity before operating
on the bit.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Reviewed-by: Jun Nie <jun.nie@linaro.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: zte: set CLK_SET_RATE_PARENT for a few zx296718 clocks
Shawn Guo [Tue, 21 Mar 2017 08:38:21 +0000 (16:38 +0800)]
clk: zte: set CLK_SET_RATE_PARENT for a few zx296718 clocks

To support VOU VGA display driver with different modes, we need to set
flag for a few clocks, so that clk_set_rate() call in VOU driver can get
VGA device desired pixel rate.

While at it, the divider between pll_vga and clk_vga gets corrected, as
it's 1:1 instead of 1:2.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Reviewed-by: Jun Nie <jun.nie@linaro.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: imx: clk-imx6ul: The i.mx6ul has no aips_tz3 clock
Robin van der Gracht [Fri, 3 Mar 2017 14:14:05 +0000 (15:14 +0100)]
clk: imx: clk-imx6ul: The i.mx6ul has no aips_tz3 clock

The clock was mapped on CG15 (gpio2_clocks) in the CCRG0 register.

Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
Signed-off-by: Robin van der Gracht <robin@protonic.nl>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoMerge tag 'tegra-for-4.12-clk' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra...
Michael Turquette [Wed, 12 Apr 2017 16:51:01 +0000 (18:51 +0200)]
Merge tag 'tegra-for-4.12-clk' of git://git./linux/kernel/git/tegra/linux into clk-next

Pull Tegra clk driver updates from Thierry Reding:

This contains a bunch of fixes and cleanups, mostly to the Tegra210
clock driver.

* tag 'tegra-for-4.12-clk' of git://git.kernel.org/pub/scm/linux/kernel/git/tegra/linux: (24 commits)
  clk: tegra: Don't reset PLL-CX if it is already enabled
  clk: tegra: Add missing Tegra210 clocks
  clk: tegra: Propagate clk_out_x rate to parent
  clk: tegra: Fix build warnings on Tegra20/Tegra30
  clk: tegra: Mark TEGRA210_CLK_DBGAPB as always on
  clk: tegra: Add SATA seq input control
  clk: tegra: Add Tegra210 special resets
  clk: tegra: Rework pll_u
  clk: tegra: Implement reset control reset
  clk: tegra: Fix disable unused for clocks sharing enable bit
  clk: tegra: Handle UTMIPLL IDDQ
  clk: tegra: Add aclk
  clk: tegra: Add super clock mux/divider
  clk: tegra: Define Tegra210 DMIC clocks
  clk: tegra: Fix constness for peripheral clocks
  clk: tegra: Define Tegra210 DMIC sync clocks
  clk: tegra: Add CEC clock
  clk: tegra: Fix type for m field
  clk: tegra: Correct tegra210_pll_fixed_mdiv_cfg rate calculation
  clk: tegra: Don't warn for PLL defaults unnecessarily
  ...

7 years agocs-2000-cp: keep Reserved bit on each register
Kuninori Morimoto [Wed, 5 Apr 2017 05:27:21 +0000 (14:27 +0900)]
cs-2000-cp: keep Reserved bit on each register

Thus CS2000 datasheet is indicating below, this patch
follows it.

WARNING: All "Reserved" registers must maintain their default
         state to ensure proper functional operation.

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: qcom: msm8996: Fix the vfe1 powerdomain name
Rajendra Nayak [Thu, 23 Mar 2017 07:43:40 +0000 (13:13 +0530)]
clk: qcom: msm8996: Fix the vfe1 powerdomain name

Fix a typo which caused both vfe0 and vfe1 powerdomains to be
named as vfe0.

Signed-off-by: Rajendra Nayak <rnayak@codeaurora.org>
Fixes: 7e824d507909 ("clk: qcom: gdsc: Add mmcc gdscs for msm8996 family")
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: stm32f4: fix timeout management for pll and ready gate
Gabriel Fernandez [Thu, 16 Mar 2017 08:16:41 +0000 (09:16 +0100)]
clk: stm32f4: fix timeout management for pll and ready gate

Use a classic polling to test bit ready.
And the shift of the bit ready of LSE & LSI were wrongs.

Fixes: 861adc44d290 ("clk: stm32f4: Add LSI & LSE clocks")
Signed-off-by: Gabriel Fernandez <gabriel.fernandez@st.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoclk: iproc: Remove redundant check
Ray Jui [Wed, 5 Apr 2017 19:53:37 +0000 (12:53 -0700)]
clk: iproc: Remove redundant check

Remove the redundant check of 'rate' in the if statement of the
'pll_set_rate' function

Reported-by: David Binderman <dcb314@hotmail.com>
Signed-off-by: Ray Jui <ray.jui@broadcom.com>
Fixes: 5fe225c105fd ("clk: iproc: add initial common clock support")
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoMerge tag 'v4.12-rockchip-clk1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Michael Turquette [Wed, 12 Apr 2017 16:50:34 +0000 (18:50 +0200)]
Merge tag 'v4.12-rockchip-clk1' of git://git./linux/kernel/git/mmind/linux-rockchip into clk-next

Pull rockchip clk driver updates from Heiko Stuebner:

  General rockchip clock changes for 4.12. Contains some new clock-ids
  as well as fixups of the clock-ids on rk3368 timers, which were unused
  and completely wrong (more and differently named timers).
  Also there is one new clock on rk3328 using the muxgrf type, a fix for
  pll enablement which should wait for the pll to lock before continuing,
  some more critical clocks and the rename of the rk1108 to rv1108, as the
  soc seems to have been using a preliminary name before its actual release.
  The plan is to have the driver changes (pinctrl, clk) go through the
  respective maintainer trees and once everything landed in mainline do
  the rename of the devicetree files. With the dts-include change in the
  clock rename, we also keep everything compiling and thus bisectability.

* tag 'v4.12-rockchip-clk1' of git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip:
  clk: rockchip: add pll_wait_lock for pll_enable
  clk: rockchip: rename RK1108 to RV1108
  dt-bindings: rk1108-cru: rename RK1108 to RV1108
  clk: rockchip: mark some rk3368 core-clks as critical
  clk: rockchip: export SCLK_TIMERXX id for timers on rk3368
  clk: rockchip: describe clk_gmac using the new muxgrf type on rk3328
  clk: rockchip: add clock ids for timer10-15 of RK3368 SoCs
  clk: rockchip: fix up rk3368 timer-ids
  clk: rockchip: add rk3328 clk_mac2io_ext ID
  clk: rockchip: Set "ignore unused" for PMU M0 clocks on rk3399

7 years agoMerge tag 'clk-renesas-for-v4.12-tag2' of git://git.kernel.org/pub/scm/linux/kernel...
Michael Turquette [Wed, 12 Apr 2017 16:49:36 +0000 (18:49 +0200)]
Merge tag 'clk-renesas-for-v4.12-tag2' of git://git./linux/kernel/git/geert/renesas-drivers into clk-next

Pull Renesas clk driver updates from Geert Uytterhoeven:

  - Add support for the Clock Pulse Generator / Module Standby and
    Software Reset module on revision ES2.0 of the R-Car H3 SoC, which
    differs from ES1.x in some areas.
  - Add IMR clocks for R-Car H3 and M3-W,
  - Add workaround for PLL0/2/4 errata on R-Car H3 ES1.0,
  - Small fixes and cleanups.

* tag 'clk-renesas-for-v4.12-tag2' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-drivers:
  clk: renesas: rcar-gen3-cpg: Add support for RCLK on R-Car H3 ES2.0
  clk: renesas: r8a7795: Add support for R-Car H3 ES2.0
  clk: renesas: Add r8a7795 ES2.0 CPG Core Clock Definitions
  clk: renesas: cpg-mssr: Add support for fixing up clock tables
  clk: renesas: rcar-gen3: Add workaround for PLL0/2/4 errata on H3 ES1.0
  clk: renesas: rcar-gen3-cpg: Pass mode pins to rcar_gen3_cpg_init()
  clk: renesas: r8a7796: Reformat core clock table
  clk: renesas: r8a7795: Reformat core clock table
  clk: renesas: r8a7796: Correct name of watchdog clock
  clk: renesas: r8a7795: Correct name of watchdog clock
  clk: renesas: r8a7795: Correct parent clock and sort order for Audio DMACs
  clk: renesas: r8a7796: Add IMR clocks
  clk: renesas: r8a7795: Add IMR clocks

7 years agoMerge branch 'clk-fixes' into clk-next
Stephen Boyd [Wed, 5 Apr 2017 01:35:51 +0000 (18:35 -0700)]
Merge branch 'clk-fixes' into clk-next

* clk-fixes:
  clk: stm32f4: fix: exclude values 0 and 1 for PLLQ

7 years agoclk: stm32f4: fix: exclude values 0 and 1 for PLLQ
Gabriel Fernandez [Thu, 16 Mar 2017 08:16:40 +0000 (09:16 +0100)]
clk: stm32f4: fix: exclude values 0 and 1 for PLLQ

0000: PLLQ = 0, wrong configuration
0001: PLLQ = 1, wrong configuration
...
0010: PLLQ = 2
0011: PLLQ = 3
0100: PLLQ = 4
...
1111: PLLQ = 1

Use divider table to exclude 0 and 1 values.

Fixes: 83135ad3c517 ("clk: stm32f4: Add PLL_I2S & PLL_SAI for STM32F429/469 boards")

Signed-off-by: Gabriel Fernandez <gabriel.fernandez@st.com>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoMerge branch 'for-4.12-ti-clk-cleanups' of https://github.com/t-kristo/linux-pm into...
Stephen Boyd [Tue, 4 Apr 2017 22:33:18 +0000 (15:33 -0700)]
Merge branch 'for-4.12-ti-clk-cleanups' of https://github.com/t-kristo/linux-pm into clk-next

* 'for-4.12-ti-clk-cleanups' of https://github.com/t-kristo/linux-pm:
  clk: ti: convert to use proper register definition for all accesses
  clk: ti: dpll44xx: fix clksel register initialization
  clk: ti: gate: export gate_clk_ops locally
  clk: ti: divider: add driver internal API for parsing divider data
  clk: ti: divider: convert TI divider clock to use its own data representation
  clk: ti: mux: convert TI mux clock to use its internal data representation
  clk: ti: drop unnecessary MEMMAP_ADDRESSING flag
  clk: ti: omap4: cleanup unnecessary clock aliases
  clk: ti: enforce const types on string arrays
  clk: ti: move omap2_init_clk_clkdm under TI clock driver
  clk: ti: add clkdm_lookup to the exported functions
  clk: ti: use automatic clock alias generation framework
  clk: ti: add API for creating aliases automatically for simple clock types
  clk: ti: add support for automatic clock alias generation
  clk: ti: remove un-used definitions from public clk_hw_omap struct

7 years agoclk: hi6220: add debug APB clock
Leo Yan [Sat, 25 Mar 2017 18:23:15 +0000 (02:23 +0800)]
clk: hi6220: add debug APB clock

The debug APB clock is absent in hi6220 driver, so this patch is to add
support for it.

Signed-off-by: Leo Yan <leo.yan@linaro.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
7 years agoMerge branch 'v4.12/clk-drivers' into v4.12/clk
Kevin Hilman [Tue, 4 Apr 2017 22:58:11 +0000 (15:58 -0700)]
Merge branch 'v4.12/clk-drivers' into v4.12/clk

* v4.12/clk-drivers:
  clk: meson-gxbb: Add GXL/GXM GP0 Variant
  clk: meson-gxbb: Add GP0 PLL init parameters
  clk: meson: Add support for parameters for specific PLLs
  clk: meson-gxbb: Add MALI clocks
  clk: meson: mpll: correct N2 maximum value
  clk: meson8b: add the mplls clocks 0, 1 and 2
  clk: meson: gxbb: mpll: use rw operation
  clk: meson: mpll: add rw operation
  clk: gxbb: put dividers and muxes in tables
  clk: meson8b: put dividers and muxes in tables
  clk: meson: add missing const qualifiers on gate arrays
  clk: meson: fix SET_PARM macro

7 years agoclk: meson-gxbb: Add GXL/GXM GP0 Variant
Neil Armstrong [Wed, 22 Mar 2017 10:32:25 +0000 (11:32 +0100)]
clk: meson-gxbb: Add GXL/GXM GP0 Variant

The clock tree in the Amlogic GXBB and GXL/GXM SoCs is shared, but the GXL/GXM
SoCs embeds a different GP0 PLL, and needs different parameters with a vendor
provided reduced rate table.

This patch adds the GXL GP0 variant, and adds a GXL DT compatible in order
to use the GXL GP0 PLL instead of the GXBB specific one.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1490178747-14837-4-git-send-email-narmstrong@baylibre.com

7 years agoclk: meson-gxbb: Add GP0 PLL init parameters
Neil Armstrong [Wed, 22 Mar 2017 10:32:24 +0000 (11:32 +0100)]
clk: meson-gxbb: Add GP0 PLL init parameters

Tha Amlogic GXBB SoC GP0 PLL needs some vendor provided parameters to be
initializated in the the GP0 control registers before configuring the rate
with the rate table provided parameters.

GXBB GP0 PLL tweaks are also selected to respect the vendor init procedure.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1490178747-14837-3-git-send-email-narmstrong@baylibre.com

7 years agoclk: meson: Add support for parameters for specific PLLs
Neil Armstrong [Wed, 22 Mar 2017 10:32:23 +0000 (11:32 +0100)]
clk: meson: Add support for parameters for specific PLLs

In recent Amlogic GXBB, GXL and GXM SoCs, the GP0 PLL needs some specific
parameters in order to initialize and lock correctly.

This patch adds an optional PARAM table used to initialize the PLL to a
default value with it's parameters in order to achieve to desired frequency.

The GP0 PLL in GXBB, GXL/GXM also needs some tweaks in the initialization
steps, and these are exposed along the PARAM table.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1490178747-14837-2-git-send-email-narmstrong@baylibre.com

7 years agoclk: meson-gxbb: Add MALI clocks
Neil Armstrong [Wed, 22 Mar 2017 10:18:54 +0000 (11:18 +0100)]
clk: meson-gxbb: Add MALI clocks

The Mali is clocked by two identical clock paths behind a glitch free mux
to safely change frequency while running.

The two "mali_0" and "mali_1" clocks are composed of a mux, divider and gate.
Expose these two clocks trees using generic clocks.
Finally the glitch free mux is added as "mali" clock.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1490177935-9646-3-git-send-email-narmstrong@baylibre.com

7 years agodt-bindings: clock: gxbb-clkc: Add GXL compatible variant
Neil Armstrong [Wed, 22 Mar 2017 10:32:27 +0000 (11:32 +0100)]
dt-bindings: clock: gxbb-clkc: Add GXL compatible variant

Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1490178747-14837-6-git-send-email-narmstrong@baylibre.com

7 years agoclk: meson-gxbb: Expose GP0 dt-bindings clock id
Neil Armstrong [Wed, 22 Mar 2017 10:32:26 +0000 (11:32 +0100)]
clk: meson-gxbb: Expose GP0 dt-bindings clock id

This patch exposes the GP0 PLL clock id in the dt bindings.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1490178747-14837-5-git-send-email-narmstrong@baylibre.com

7 years agoclk: meson-gxbb: Add MALI clock IDS
Neil Armstrong [Wed, 22 Mar 2017 10:18:53 +0000 (11:18 +0100)]
clk: meson-gxbb: Add MALI clock IDS

Add missing MALI clock IDs and expose the muxes and gates in the dt-bindings.

Signed-off-by: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/1490177935-9646-2-git-send-email-narmstrong@baylibre.com

7 years agodt-bindings: clk: gxbb: expose i2s output clock gates
Jerome Brunet [Thu, 9 Mar 2017 10:41:54 +0000 (11:41 +0100)]
dt-bindings: clk: gxbb: expose i2s output clock gates

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-10-jbrunet@baylibre.com

7 years agoclk: tegra: Don't reset PLL-CX if it is already enabled
Jon Hunter [Mon, 27 Mar 2017 11:01:05 +0000 (12:01 +0100)]
clk: tegra: Don't reset PLL-CX if it is already enabled

Commit 8dce89a1c2cf ("clk: tegra: Don't warn for PLL defaults
unnecessarily") changed the tegra210_pllcx_set_defaults() function
causing the PLL to always be reset regardless of whether it is in-use.
This function was changed so that resetting of the PLL will only be
skipped if the PLL is enabled AND 'pllcx->params->defaults_set' is not
true. However, the 'pllcx->params->defaults_set' is always true and
hence, the PLL is now always reset. This causes the boot to fail on the
Tegra210 Smaug where the PLL is already enabled and in-use. Fix this by
only resetting the PLL if not in-use and only printing the warning that
the defaults are not set after we have checked the default settings.

Fixes: 8dce89a1c2cf ("clk: tegra: Don't warn for PLL defaults unnecessarily")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Acked-By: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Add missing Tegra210 clocks
Peter De Schrijver [Wed, 22 Mar 2017 14:23:16 +0000 (16:23 +0200)]
clk: tegra: Add missing Tegra210 clocks

iqc1, iqc2, tegra_clk_pll_a_out_adsp, tegra_clk_pll_a_out0_out_adsp, adsp
and adsp neon were not modelled. dp2 wasn't modelled for Tegra210.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Propagate clk_out_x rate to parent
Alex Frid [Wed, 22 Mar 2017 13:38:19 +0000 (15:38 +0200)]
clk: tegra: Propagate clk_out_x rate to parent

Given that externx can only be used as a parent for clk_out_x, it makes
sense to propagate requests to make clk_out_x easier to handle.

Signed-off-by: Alex Frid <afrid@nvidia.com>
Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: renesas: rcar-gen3-cpg: Add support for RCLK on R-Car H3 ES2.0
Geert Uytterhoeven [Fri, 10 Mar 2017 11:13:37 +0000 (12:13 +0100)]
clk: renesas: rcar-gen3-cpg: Add support for RCLK on R-Car H3 ES2.0

Starting with R-Car H3 ES2.0, the parent of RCLK is selected using MD28.

Add support for that, but retain the old behavior for R-Car H3 ES1.x and
M3-W ES1.0 using a quirk.

Inspired by a patch by Takeshi Kihara in the BSP.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: Takeshi Kihara <takeshi.kihara.df@renesas.com>
7 years agoclk: renesas: r8a7795: Add support for R-Car H3 ES2.0
Geert Uytterhoeven [Thu, 29 Sep 2016 12:36:11 +0000 (14:36 +0200)]
clk: renesas: r8a7795: Add support for R-Car H3 ES2.0

The Clock Pulse Generator / Module Standby and Software Reset module in
R-Car H3 ES2.0 differs from ES1.x in the following areas:
  - More core clocks (S0D2, S0D3, S0D6, S0D8, S0D12),
  - Different parent clocks for AUDMAC, EtherAVB, FCP, FDP, IMR,
    SYS-DMAC, VIN, VSPB, VSPI,
  - Removal of modules CSI21, FCPCI, FCPF2, FCPVD3, FCPVI2, FDP1-2,
    USB3-IF1, VSPD3, VSPI2,
  - Addition of modules EHCI3, HS-USB-IF3, USB-DMAC3-0, USB-DMAC3-1.

The goal is twofold:
  1. Support both the ES1.x and ES2.0 SoC revisions in a single binary
     for now,
  2. Make it clear which code supports ES1.x, so it can easily be
     identified and removed later, when production SoCs are deemed
     ubiquitous.

This is achieved by:
  - Updating the clock tables for the latest revision (ES2.0), but not
    removing clocks that only exist on earlier revisions (ES1.x),
  - Detecting the SoC revision at runtime using the new soc_device_match()
    API, and fixing up the clocks tables to match the actual SoC
    revision, by:
      - NULLifying core and module clocks of modules that do not exist,
      - Reparenting module clocks that have a different parent on ES1.x.

Based on R-Car Gen3 Hardware User's Manual rev. 0.53E.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
7 years agoclk: renesas: Add r8a7795 ES2.0 CPG Core Clock Definitions
Geert Uytterhoeven [Thu, 29 Sep 2016 11:06:15 +0000 (13:06 +0200)]
clk: renesas: Add r8a7795 ES2.0 CPG Core Clock Definitions

Add all R-Car H3 ES2.0 Clock Pulse Generator Core Clock Outputs, as
listed in Table 8.2a ("List of Clocks [R-Car H3]") of the R-Car Gen3
Hardware User's Manual rev. 0.53E.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
7 years agoclk: renesas: cpg-mssr: Add support for fixing up clock tables
Geert Uytterhoeven [Thu, 29 Sep 2016 12:47:58 +0000 (14:47 +0200)]
clk: renesas: cpg-mssr: Add support for fixing up clock tables

The same SoC may have different clocks and/or module clock parents,
depending on SoC revision.  One option is to use different sets of clock
tables for each SoC revision.  However, if the differences are small, it
is much more space-efficient to have a single set of clock tables, and
fix those up at runtime instead.

Hence provide three helpers:
  - Two helpers to NULLify core and module clocks that do not exist on
    some revisions (NULLified clocks are skipped during the registration
    phase),
  - One helper to reparent module clocks that have different clock
    parents.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
7 years agoclk: meson: mpll: correct N2 maximum value
Jerome Brunet [Thu, 9 Mar 2017 10:41:53 +0000 (11:41 +0100)]
clk: meson: mpll: correct N2 maximum value

Gxbb datasheet says N2 maximum value is 127 but the register field is
9 bits wide, the maximum value should 511.

Test shows value greater than 127, all the way to 511, works well

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-9-jbrunet@baylibre.com

7 years agoclk: meson8b: add the mplls clocks 0, 1 and 2
Jerome Brunet [Thu, 9 Mar 2017 10:41:52 +0000 (11:41 +0100)]
clk: meson8b: add the mplls clocks 0, 1 and 2

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-8-jbrunet@baylibre.com

7 years agoclk: meson: gxbb: mpll: use rw operation
Jerome Brunet [Thu, 9 Mar 2017 10:41:51 +0000 (11:41 +0100)]
clk: meson: gxbb: mpll: use rw operation

Use read/write operations for the mpll clocks instead of the
read-only ones.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-7-jbrunet@baylibre.com

7 years agoclk: meson: mpll: add rw operation
Jerome Brunet [Thu, 9 Mar 2017 10:41:50 +0000 (11:41 +0100)]
clk: meson: mpll: add rw operation

This patch adds new callbacks to the meson-mpll driver to control
and set the pll rate. For this, we also need to add the enable bit and
sdm enable bit. The corresponding parameters are added to mpll data
structure.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-6-jbrunet@baylibre.com

7 years agoclk: gxbb: put dividers and muxes in tables
Jerome Brunet [Thu, 9 Mar 2017 10:41:49 +0000 (11:41 +0100)]
clk: gxbb: put dividers and muxes in tables

Until now, there was only 2 dividers and 2 muxes declared for the gxbb
platform. With the ongoing work on various subsystem, including audio,
this is about to change. Use the same approach as gates for dividers and
muxes, putting them in tables to fix the register address at runtime.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-5-jbrunet@baylibre.com

7 years agoclk: meson8b: put dividers and muxes in tables
Jerome Brunet [Thu, 9 Mar 2017 10:41:48 +0000 (11:41 +0100)]
clk: meson8b: put dividers and muxes in tables

Until now, there was only 1 divider and 1 mux declared for the meson8b
platform. With the ongoing work on various system, including audio, this
is about to change. Use the same approach as gates for dividers and muxes,
putting them in tables to fix the register address at runtime.

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-4-jbrunet@baylibre.com

7 years agoclk: meson: add missing const qualifiers on gate arrays
Jerome Brunet [Thu, 9 Mar 2017 10:41:47 +0000 (11:41 +0100)]
clk: meson: add missing const qualifiers on gate arrays

Reported-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-3-jbrunet@baylibre.com

7 years agoclk: meson: fix SET_PARM macro
Jerome Brunet [Thu, 9 Mar 2017 10:41:46 +0000 (11:41 +0100)]
clk: meson: fix SET_PARM macro

parameter val is not enclosed in parenthesis which is buggy when given an
expression instead of a simple value

Signed-off-by: Jerome Brunet <jbrunet@baylibre.com>
Reviewed-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Michael Turquette <mturquette@baylibre.com>
Link: lkml.kernel.org/r/20170309104154.28295-2-jbrunet@baylibre.com

7 years agoMerge tag 'sunxi-clk-fixes-for-4.11' of https://git.kernel.org/pub/scm/linux/kernel...
Stephen Boyd [Thu, 23 Mar 2017 23:08:46 +0000 (16:08 -0700)]
Merge tag 'sunxi-clk-fixes-for-4.11' of https://git./linux/kernel/git/mripard/linux into clk-fixes

Pull Allwinner clock fixes from Maxime Ripard:

A few fixes for a bunch of clocks on a few SoCs. The most important one is
probably one that fixes the NKMP clock frequency calculation and could end
up with clocking the CPU frequency to out of bounds rates.

* tag 'sunxi-clk-fixes-for-4.11' of https://git.kernel.org/pub/scm/linux/kernel/git/mripard/linux:
  clk: sunxi-ng: fix recalc_rate formula of NKMP clocks
  clk: sunxi-ng: Fix div/mult settings for osc12M on A64
  clk: sunxi-ng: sun6i: Fix enable bit offset for hdmi-ddc module clock
  clk: sunxi: ccu-sun5i needs nkmp
  clk: sunxi-ng: mp: Adjust parent rate for pre-dividers

7 years agoclk: rockchip: add pll_wait_lock for pll_enable
Elaine Zhang [Wed, 22 Feb 2017 02:59:55 +0000 (10:59 +0800)]
clk: rockchip: add pll_wait_lock for pll_enable

If pll is power down,when power up pll need wait pll lock.
The reference documents section:
PLL frequency change and lock check

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agoclk: rockchip: rename RK1108 to RV1108
Andy Yan [Fri, 17 Mar 2017 17:18:38 +0000 (18:18 +0100)]
clk: rockchip: rename RK1108 to RV1108

Rockchip finally named the SOC as RV1108, so change it.

Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
[include rename in rk1108.dtsi to prevent compile errors]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agodt-bindings: rk1108-cru: rename RK1108 to RV1108
Andy Yan [Fri, 17 Mar 2017 17:18:37 +0000 (18:18 +0100)]
dt-bindings: rk1108-cru: rename RK1108 to RV1108

Rockchip finally named the SOC as RV1108, so change it.

Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agoclk: renesas: rcar-gen3: Add workaround for PLL0/2/4 errata on H3 ES1.0
Geert Uytterhoeven [Fri, 10 Mar 2017 10:46:10 +0000 (11:46 +0100)]
clk: renesas: rcar-gen3: Add workaround for PLL0/2/4 errata on H3 ES1.0

Add a workaround for errata on R-Car H3 ES1.0, where the PLL0, PLL2, and
PLL4 clock frequencies are off by a factor of two.

Inspired by a patch by Dien Pham in the BSP.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: Dien Pham <dien.pham.ry@renesas.com>
7 years agoclk: renesas: rcar-gen3-cpg: Pass mode pins to rcar_gen3_cpg_init()
Geert Uytterhoeven [Fri, 10 Mar 2017 10:36:33 +0000 (11:36 +0100)]
clk: renesas: rcar-gen3-cpg: Pass mode pins to rcar_gen3_cpg_init()

Pass the mode pin states from the SoC-specific CPG/MSSR driver to the
R-Car Gen3 CPG driver core, as their state will be needed to make some
core clock configuration decisions.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
7 years agoclk: renesas: r8a7796: Reformat core clock table
Geert Uytterhoeven [Thu, 10 Nov 2016 12:18:25 +0000 (13:18 +0100)]
clk: renesas: r8a7796: Reformat core clock table

For easier comparison with other clock drivers.
No functional changes.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
7 years agoclk: renesas: r8a7795: Reformat core clock table
Geert Uytterhoeven [Thu, 10 Nov 2016 12:16:57 +0000 (13:16 +0100)]
clk: renesas: r8a7795: Reformat core clock table

For easier comparison with other clock drivers.
No functional changes.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
7 years agoclk: renesas: r8a7796: Correct name of watchdog clock
Geert Uytterhoeven [Tue, 28 Feb 2017 16:18:08 +0000 (17:18 +0100)]
clk: renesas: r8a7796: Correct name of watchdog clock

There's only a single watchdog clock, and it's named "rwdt".

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
7 years agoclk: renesas: r8a7795: Correct name of watchdog clock
Geert Uytterhoeven [Tue, 28 Feb 2017 16:17:31 +0000 (17:17 +0100)]
clk: renesas: r8a7795: Correct name of watchdog clock

There's only a single watchdog clock, and it's named "rwdt".

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
7 years agoclk: renesas: r8a7795: Correct parent clock and sort order for Audio DMACs
Geert Uytterhoeven [Tue, 28 Feb 2017 16:31:59 +0000 (17:31 +0100)]
clk: renesas: r8a7795: Correct parent clock and sort order for Audio DMACs

The parent clock of the Audio DMACs is the "ZS" AXI bus clock, which
maps to S3D1 on R-Car H3 ES1.x.
All module clocks must be sorted by clock ID.

Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
7 years agoclk: tegra: Fix build warnings on Tegra20/Tegra30
Thierry Reding [Mon, 20 Mar 2017 16:14:14 +0000 (17:14 +0100)]
clk: tegra: Fix build warnings on Tegra20/Tegra30

The recent conversion of proper const usage was only partial and didn't
include Tegra20 and Tegra30 support. Fix that up.

Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Mark TEGRA210_CLK_DBGAPB as always on
Peter De Schrijver [Tue, 28 Feb 2017 15:19:50 +0000 (17:19 +0200)]
clk: tegra: Mark TEGRA210_CLK_DBGAPB as always on

This is needed to make the JTAG debugging interface work.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
[treding@nvidia.com: add TODO comment]
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Add SATA seq input control
Peter De Schrijver [Wed, 15 Mar 2017 15:42:05 +0000 (17:42 +0200)]
clk: tegra: Add SATA seq input control

This will be used by the powergating driver to ensure proper sequencer
state when the SATA domain is powergated.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Add Tegra210 special resets
Peter De Schrijver [Wed, 15 Mar 2017 12:59:32 +0000 (14:59 +0200)]
clk: tegra: Add Tegra210 special resets

Tegra210 has 2 special resets which don't follow the normal pattern:
DVCO and ADSP. Add them in this patch.

Changelog:

v2: add DT bindings file

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Rework pll_u
Peter De Schrijver [Tue, 14 Mar 2017 14:12:49 +0000 (16:12 +0200)]
clk: tegra: Rework pll_u

In normal operation pll_u is under hardware control and has a fixed rate
of 480MHz. Hardware will turn on pll_u on whenever any of the XUSB
powerdomains is on. From a software point of view we model this is if
pll_u is always on using a fixed rate clock. However the bootloader
might or might not have configured pll_u this way. So we will check the
current state of pll_u at boot and reconfigure it if required.

There are 3 possiblities at kernel boot:
1) pll_u is under hardware control: do nothing
2) pll_u is under hardware control and enabled: enable hardware control
3) pll_u is disabled: enable pll_u and enable hardware control

In all cases we also check if UTMIPLL is under hardware control at boot
and configure it for hardware control if that is not the case.
The same is done during SC7 resume.

Thanks to Joseph Lo <josephl@nvidia.com> for bug fixes.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Implement reset control reset
Mikko Perttunen [Thu, 2 Mar 2017 14:16:16 +0000 (16:16 +0200)]
clk: tegra: Implement reset control reset

For completeness, also implement this reset framework API for Tegra.

Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Reviewed-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Arto Merilainen <amerilainen@nvidia.com>
Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Fix disable unused for clocks sharing enable bit
Peter De Schrijver [Thu, 2 Mar 2017 13:22:05 +0000 (15:22 +0200)]
clk: tegra: Fix disable unused for clocks sharing enable bit

In case 2 clocks share an enable bit and one of them is enabled by a
driver and the other one is not, CCF will think it's enabled because it
will only look at the HW state. Therefore it will disable the clock and
thus also disable the other clock which was enabled. Solve this by
reading the initial state of the enable bit and incrementing the
refcount if it's set.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Handle UTMIPLL IDDQ
Peter De Schrijver [Tue, 28 Feb 2017 15:19:24 +0000 (17:19 +0200)]
clk: tegra: Handle UTMIPLL IDDQ

Export UTMIPLL IDDQ functions. These will be needed when powergating the
XUSB partition.

Signed-off-by: BH Hsieh <bhsieh@nvidia.com>
Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Add aclk
Peter De Schrijver [Tue, 28 Feb 2017 14:37:22 +0000 (16:37 +0200)]
clk: tegra: Add aclk

This clock clocks the ADSP Cortex-A9.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Add super clock mux/divider
Peter De Schrijver [Tue, 28 Feb 2017 14:37:21 +0000 (16:37 +0200)]
clk: tegra: Add super clock mux/divider

Add a super clock type which implements both mux and divider. This is
used for aclk.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Define Tegra210 DMIC clocks
Peter De Schrijver [Tue, 28 Feb 2017 14:37:20 +0000 (16:37 +0200)]
clk: tegra: Define Tegra210 DMIC clocks

Tegra210 has 3 inputs for Digital Microphones (DMICs). Provide the
required clocks for them.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Fix constness for peripheral clocks
Peter De Schrijver [Tue, 28 Feb 2017 14:37:19 +0000 (16:37 +0200)]
clk: tegra: Fix constness for peripheral clocks

checkpatch now warns for const ** and expects const * const * to be used
instead. This means we have to update the prototypes and function
declarations to handle this change.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Define Tegra210 DMIC sync clocks
Peter De Schrijver [Tue, 28 Feb 2017 14:37:18 +0000 (16:37 +0200)]
clk: tegra: Define Tegra210 DMIC sync clocks

Tegra210 has 3 DMIC inputs which can be clocked from the recovered clock
of several other audio inputs (eg. i2s0, i2s1, ...). To model this, we
add a 3 new clocks similar to the audio* clocks which handle the same
function for the I2S and SPDIF clocks.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Add CEC clock
Peter De Schrijver [Tue, 28 Feb 2017 14:37:17 +0000 (16:37 +0200)]
clk: tegra: Add CEC clock

This clock is used to clock the HDMI CEC interface.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Fix type for m field
Peter De Schrijver [Thu, 23 Feb 2017 10:44:44 +0000 (12:44 +0200)]
clk: tegra: Fix type for m field

When used as part of fractional ndiv calculations, the current range is
not enough because the denominator of the fraction is multiplied with m.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Correct tegra210_pll_fixed_mdiv_cfg rate calculation
Peter De Schrijver [Thu, 23 Feb 2017 10:44:43 +0000 (12:44 +0200)]
clk: tegra: Correct tegra210_pll_fixed_mdiv_cfg rate calculation

Return the actually achieved rate in cfg->output_rate rather than just
the requested rate. This is important to make clk_round_rate() return
the correct result.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Don't warn for PLL defaults unnecessarily
Peter De Schrijver [Thu, 23 Feb 2017 10:44:42 +0000 (12:44 +0200)]
clk: tegra: Don't warn for PLL defaults unnecessarily

If the PLL is on, only warn if the defaults are not yet set. Otherwise
be silent.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Remove non-existing pll_m_out1 clock
Peter De Schrijver [Thu, 23 Feb 2017 10:44:41 +0000 (12:44 +0200)]
clk: tegra: Remove non-existing pll_m_out1 clock

This clock doesn't actually exist, so remove it.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Correct afi clock parent
Peter De Schrijver [Thu, 23 Feb 2017 10:44:40 +0000 (12:44 +0200)]
clk: tegra: Correct afi clock parent

The parent for afi is actually mselect, not clk_m.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Fix ISP clock modelling
Peter De Schrijver [Thu, 23 Feb 2017 10:44:39 +0000 (12:44 +0200)]
clk: tegra: Fix ISP clock modelling

The 2 ISP clocks (ispa and ispb) share a mux/divider control. So model
this as 1 mux/divider clock and child gate clocks.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: tegra: Fix pll_a1 iddq register, add pll_a1
Peter De Schrijver [Thu, 23 Feb 2017 10:44:38 +0000 (12:44 +0200)]
clk: tegra: Fix pll_a1 iddq register, add pll_a1

pll_a1 was using CLK_RST_CONTROLLER_PLLA1_MISC_0 for IDDQ control rather
than the correct register CLK_RST_CONTROLLER_PLLA1_MISC_1. Also add
pll_a1 to the set of clocks defined for Tegra210.

Signed-off-by: Peter De Schrijver <pdeschrijver@nvidia.com>
Reviewed-by: Mikko Perttunen <mperttunen@nvidia.com>
Tested-by: Mikko Perttunen <mperttunen@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
7 years agoclk: sunxi-ng: fix recalc_rate formula of NKMP clocks
Icenowy Zheng [Fri, 17 Mar 2017 20:19:43 +0000 (04:19 +0800)]
clk: sunxi-ng: fix recalc_rate formula of NKMP clocks

In commit e66f81bbd746 ("clk: sunxi-ng: Implement factors offsets"), the
final formula of NKMP clocks' recalc_rate is refactored; however, the
refactored formula broke the calculation due to some C language operand
priority problem -- the priority of operand >> is lower than * and /,
makes the formula being parsed as "(parent_rate * n * k) >> (p / m)", but
it should be "(parent_rate * n * k >> p) / m".

Add the pair of parentheses to fix up this issue. This pair of
parentheses used to exist in the old formula.

Fixes: e66f81bbd746 ("clk: sunxi-ng: Implement factors offsets")
Signed-off-by: Icenowy Zheng <icenowy@aosc.xyz>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
7 years agoclk: sunxi-ng: Fix div/mult settings for osc12M on A64
Philipp Tomsich [Wed, 15 Mar 2017 11:23:58 +0000 (12:23 +0100)]
clk: sunxi-ng: Fix div/mult settings for osc12M on A64

The mult/div for osc12M was previously backwards (giving a 48M rate
for osc12M). Fix it.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Christoph Muellner <christoph.muellner@theobroma-systems.com>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
7 years agoclk: rockchip: mark some rk3368 core-clks as critical
Elaine Zhang [Tue, 7 Mar 2017 11:16:48 +0000 (19:16 +0800)]
clk: rockchip: mark some rk3368 core-clks as critical

Mark pclk_pd_alive, pclk_peri, hclk_peri as critical

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agoclk: rockchip: export SCLK_TIMERXX id for timers on rk3368
Elaine Zhang [Tue, 7 Mar 2017 11:05:18 +0000 (19:05 +0800)]
clk: rockchip: export SCLK_TIMERXX id for timers on rk3368

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agoclk: rockchip: describe clk_gmac using the new muxgrf type on rk3328
Elaine Zhang [Mon, 6 Feb 2017 02:50:35 +0000 (10:50 +0800)]
clk: rockchip: describe clk_gmac using the new muxgrf type on rk3328

With the newly introduced clk type for muxes in the grf we now can
describe some missing clocks, like the clk_gmac2io and clk_gmac2phy
that selects between clk_mac2io_src and gmac_clkin based on a bit
set in the general register files.

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agoMerge branch 'v4.12-shared/clkids' into v4.12-clk/next
Heiko Stuebner [Fri, 10 Mar 2017 10:18:53 +0000 (11:18 +0100)]
Merge branch 'v4.12-shared/clkids' into v4.12-clk/next

7 years agoclk: rockchip: add clock ids for timer10-15 of RK3368 SoCs
Elaine Zhang [Tue, 7 Mar 2017 11:05:17 +0000 (19:05 +0800)]
clk: rockchip: add clock ids for timer10-15 of RK3368 SoCs

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agoclk: rockchip: fix up rk3368 timer-ids
Elaine Zhang [Tue, 7 Mar 2017 11:05:15 +0000 (19:05 +0800)]
clk: rockchip: fix up rk3368 timer-ids

The timer-ids are wrong compared to the manual, probably due a simple
copy-paste mistake from the otherwise very similar rk3288. And there
are even more timers in the system than the ones wrongly listed here.

Timer-Ids were unused both in clock-driver as well as devicetree
till now, so fixing them won't break anything.

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agoclk: rockchip: add rk3328 clk_mac2io_ext ID
Elaine Zhang [Mon, 6 Feb 2017 02:50:34 +0000 (10:50 +0800)]
clk: rockchip: add rk3328 clk_mac2io_ext ID

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
7 years agoclk: ti: convert to use proper register definition for all accesses
Tero Kristo [Thu, 9 Feb 2017 09:24:37 +0000 (11:24 +0200)]
clk: ti: convert to use proper register definition for all accesses

Currently, TI clock driver uses an encapsulated struct that is cast into
a void pointer to store all register addresses. This can be considered
as rather nasty hackery, and prevents from expanding the register
address field also. Instead, replace all the code to use proper struct
in place for this, which contains all the previously used data.

This patch is rather large as it is touching multiple files, but this
can't be split up as we need to avoid any boot breakage.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: dpll44xx: fix clksel register initialization
Tero Kristo [Thu, 9 Feb 2017 09:25:28 +0000 (11:25 +0200)]
clk: ti: dpll44xx: fix clksel register initialization

clksel register pointer should be used for the DPLL-MX autoidle handling.
Currently this is not setup at all. Fix by adding proper handling for the
register.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: gate: export gate_clk_ops locally
Tero Kristo [Thu, 9 Feb 2017 09:10:19 +0000 (11:10 +0200)]
clk: ti: gate: export gate_clk_ops locally

These are going to be used by the clkctrl support that will be introduced
later.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: divider: add driver internal API for parsing divider data
Tero Kristo [Thu, 9 Feb 2017 12:46:53 +0000 (14:46 +0200)]
clk: ti: divider: add driver internal API for parsing divider data

This can be used from the divider itself, and also from the clkctrl
clocks once this is introduced.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: divider: convert TI divider clock to use its own data representation
Tero Kristo [Thu, 9 Feb 2017 12:45:45 +0000 (14:45 +0200)]
clk: ti: divider: convert TI divider clock to use its own data representation

Instead of using the generic clock driver data struct, use one internal
for the TI clock driver itself. This allows modifying the register access
parts in subsequent patch.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: mux: convert TI mux clock to use its internal data representation
Tero Kristo [Thu, 9 Feb 2017 12:40:40 +0000 (14:40 +0200)]
clk: ti: mux: convert TI mux clock to use its internal data representation

Instead of using the generic clock driver data struct, use one internal
for the TI clock driver itself. This allows modifying the register access
parts in subsequent patch.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: drop unnecessary MEMMAP_ADDRESSING flag
Tero Kristo [Mon, 30 Jan 2017 14:01:36 +0000 (16:01 +0200)]
clk: ti: drop unnecessary MEMMAP_ADDRESSING flag

This has been superceded by the usage of ti_clk_ll_ops for now.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: omap4: cleanup unnecessary clock aliases
Tero Kristo [Tue, 22 Mar 2016 14:10:55 +0000 (16:10 +0200)]
clk: ti: omap4: cleanup unnecessary clock aliases

Cleanup any unnecessary DT_CLK() alias entries from the OMAP4 clock file.
Most of these are now handled dynamically by the driver code.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: enforce const types on string arrays
Tero Kristo [Wed, 5 Oct 2016 12:37:02 +0000 (15:37 +0300)]
clk: ti: enforce const types on string arrays

Constant string arrays should use const char * const instead of just
const char *. Change the implementations using these to proper type.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: move omap2_init_clk_clkdm under TI clock driver
Tero Kristo [Fri, 30 Sep 2016 11:13:38 +0000 (14:13 +0300)]
clk: ti: move omap2_init_clk_clkdm under TI clock driver

This is not needed outside the driver, so move it inside it and remove
the prototype from the public header also.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: add clkdm_lookup to the exported functions
Tero Kristo [Fri, 30 Sep 2016 11:10:11 +0000 (14:10 +0300)]
clk: ti: add clkdm_lookup to the exported functions

This will be needed to move some additional clockdomain functionality
under clock driver.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: use automatic clock alias generation framework
Tero Kristo [Thu, 29 Sep 2016 09:06:40 +0000 (12:06 +0300)]
clk: ti: use automatic clock alias generation framework

Generate clock aliases automatically for all TI clock drivers.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: add API for creating aliases automatically for simple clock types
Tero Kristo [Thu, 29 Sep 2016 09:05:00 +0000 (12:05 +0300)]
clk: ti: add API for creating aliases automatically for simple clock types

This API generates clock aliases automatically for simple clock types
(fixed-clock, fixed-factor-clock), so that we don't need to add the data
for these statically into tables. Shall be called from the SoC specific
clock init.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: add support for automatic clock alias generation
Tero Kristo [Thu, 29 Sep 2016 09:00:57 +0000 (12:00 +0300)]
clk: ti: add support for automatic clock alias generation

Large portions of the OMAP framework still depend on the support of
having clock aliases in place, so add support functions for generating
these automatically.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
7 years agoclk: ti: remove un-used definitions from public clk_hw_omap struct
Tero Kristo [Fri, 18 Dec 2015 13:36:57 +0000 (15:36 +0200)]
clk: ti: remove un-used definitions from public clk_hw_omap struct

Clksel support has been deprecated a while back, so remove these from
the struct also.

Signed-off-by: Tero Kristo <t-kristo@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>