video: rockchip: rk322x: check hwc_size before H/W register config

Check hwc_size after hwc registor config, if check fail, would cause
unexpect problem, iommu crash.

Change-Id: I2e18ea86e9e27e13ccce0737d9d48befcbe345fb
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

video: rockchip: vop: 3399: add win2 and win3 when get dsp_info

Change-Id: Ib1ae76de66656b0531683eede117346945587008
Signed-off-by: Huang Jiachai <hjc@rock-chips.com>

video: rockchip: vop: 3399: add to config YUYV and UYVY data format

Change-Id: I5762f691d724449035098333a732095774c96513
Signed-off-by: Huang Jiachai <hjc@rock-chips.com>

video: rockchip: fb: add support data format YUYV and UYVY

Change-Id: Iaef1a0e6f80e4246cfe6b2692bdb6085ea5355b0
Signed-off-by: Huang Jiachai <hjc@rock-chips.com>

video: rockchip: vop: 3399: add support afbdc

Change-Id: I2e796809baeef99c3463c4789a65eb1057cb577f
Signed-off-by: Huang Jiachai <hjc@rock-chips.com>

thermal: rockchip: fixes the code_to_temp for tsadc driver

We should judge the table.id[mid].code insearch algorithm on matter the
adc value increment or decrement.

Or otherwise, the temperature return the incorrect value in some cases.
[    1.438589] adc_val=402,temp=-40000
[    1.438903] adc_val=403,temp=-39375
[    1.439217] adc_val=404,temp=-38750
...
[    1.441102] adc_val=410,temp=-40000
[    1.441416] adc_val=411,temp=-34445
[    1.441737] adc_val=412,temp=-33889
...

Let's fix it right now.

Fixes commit 020ba95
"thermal: rockchip: Add the sort mode for adc value increment or decrement"

Change-Id: Icac84d06ebf463439ca11db5a19d629b4b2b865c
Reported-by: Rocky Hao <rocky.hao@rock-chips.com>
Signed-off-by: Caesar Wang <wxt@rock-chips.com>

ARM64: dts: rk3399: the USB 2.0 vbus GPIO is board specific

A GPIO was put in rk3399.dtsi that doesn't belong there.  Specifically
this GPIO isn't the same for all rk3399 boards.  I presume it belongs in
rk3399-tb.dts, so move it there.

Change-Id: I0b3272655da565eb6b348a33401f7517224db5fa
Fixes: 3ed499f07cf7 ("ARM64: dts: rockchip: rk3399: add usb2.0 phy node")
Signed-off-by: Douglas Anderson <dianders@chromium.org>

ARM64: rockchip_cros_defconfig: Add support DRM for cros

And removed:
----
-CONFIG_FB=y
-CONFIG_LCDC_RK3368=y
-CONFIG_LCDC_LITE_RK3X=y
-CONFIG_RK_IOMMU=y
-CONFIG_RK_IOVMM=y
----
which are unused on the chromeos.

Change-Id: Icd521b56b6285099d72d3bf25575466792b6d353
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

vcodec_service/rockchip: revise build failure

there will be a build failure when CONFIG_RK_IOMMU
disabled.

Change-Id: Ifd56e39b9cb3021f308f195087304a1d1ec2c599
Signed-off-by: alpha lin <alpha.lin@rock-chips.com>

ARM64: dts: rk3399-tb: fix up the pwm regulator node

add pwm init voltage and id for uboot.
fix up the pwms node and add pwm polarity.

Change-Id: I4159c97ae498411ab958c2b1e1223139ac670452
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>

Makefile: hack out the double version of SUBLEVEL

Portage's linux-info.eclass (in getfilevar_noexec) looks for the
definition of SUBLEVEL, and it doesn't expect 2 definitions. We could
fix the eclass, but let's hack this out for now.

See strongswan's emerge output:

...
 * Found sources for kernel version:
 *     4.4.6
 * 0
/mnt/host/source/src/third_party/portage-stable/eclass/linux-info.eclass: line 388: 6
0: syntax error in expression (error token is "0")
...

Change-Id: I6964e6731ed461ca3a8c4afde0ddfe48e0105627
Signed-off-by: Brian Norris <briannorris@chromium.org>
Reviewed-on: https://chrome-internal-review.googlesource.com/252620
Reviewed-by: Doug Anderson <dianders@google.com>
Tested-by: Doug Anderson <dianders@google.com>

ARM64: rockchip_cros_defcofnig: turn on the chrome platform

This config should be opened since the config used for chromeos.

Change-Id: I52dd22b1c1a707e6d27311337a5be6f0041cb7f9
Signed-off-by: Caesar Wang <wxt@rock-chips.com>

clk: rockchip: rk3399: remove unnecessary critical clocks

Change-Id: If1f3cf9eb91f89ad38f034b5a9d90571c486efc9
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

clk: rockchip: rk3399: add all of NOCs into critical clocks

We need to declare that we enable all NOCs which are critical clocks
always and clearly and explicitly show that we have enabled them at
clk_summary.

Change-Id: I859664692b4d1bb0dda0ee38295dfcbc3cc70019
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

clk: rockchip: rk3399: Keep DMAC1 enable always for SPI5

Change-Id: I4b2b8bdf7649b0c5209852160597ad2737ed5a7b
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

clk: rockchip: rk3399: remove unnecessary CLK_IGNORE_UNUSED flags

Change-Id: I87dddf2ceb14e5d094b320568530ae8976fdbf14
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

ARM64: dts: rk3399: remove clk_ignore_unused

Change-Id: I48874e2b82487d5e9ae6e83c954ea2bd06960c8f
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

Input: touchscreen-gt9xx: enable the gt9xx SLOT REPORT

On the moment, the gt9xx touchscreen driver can't work on chromeos.
Since the driver report event has *not* judge correct by the chromeos.

We need report the singel point touch information for event firstly,
otherwise the chromeos will force a signel point to work.

That's seem a chromeos issue/leak.
Anyway, we can report the point including the signal information
to workaround.

Verify on rk3399evb board with chromeos.

root@localhost / # evtest
No device specified, trying to scan all of /dev/input/event*
Available devices:
/dev/input/event0:      goodix-ts
/dev/input/event1:      rk29-keypad
Select the device event number [0-1]: 0
Input driver version is 1.0.1
Input device ID: bus 0x18 vendor 0xdead product 0xbeef version 0x28bb
Input device name: "goodix-ts"
Supported events:
  Event type 0 (EV_SYN)
  Event type 1 (EV_KEY)
  Event code 330 (BTN_TOUCH)
  Event type 3 (EV_ABS)
  Event code 0 (ABS_X)
  ...

  Event: time 1450321044.293221, type 3 (EV_ABS), code 57 (ABS_MT_TRACKING_ID), value 0
  Event: time 1450321044.293221, type 3 (EV_ABS), code 53 (ABS_MT_POSITION_X), value 4095
  Event: time 1450321044.293221, type 3 (EV_ABS), code 54 (ABS_MT_POSITION_Y), value 4083
  Event: time 1450321044.293221, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 8
  Event: time 1450321044.293221, type 3 (EV_ABS), code 50 (ABS_MT_WIDTH_MAJOR), value 8
  Event: time 1450321044.293221, -------------- SYN_REPORT ------------
  Event: time 1450321044.384655, type 3 (EV_ABS), code 57 (ABS_MT_TRACKING_ID), value -1
  Event: time 1450321044.384655, -------------- SYN_REPORT ------------

Change-Id: Ic41327a673632e471429ded35b68eecbbd7f3069
Signed-off-by: Caesar Wang <wxt@rock-chips.com>

ARM64: dts: rk3399: chrome: enable mipi node

Change-Id: Icc169b97ec985b5e7332ed1ed5ed78d20c717062
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

ARM64/cros_defconfig: remove IEP and RK_VCODEC

Remove CONFIG_IEP and CONFIG_RK_VCODEC definition for
they aren't required in rockchip chromeos.

Change-Id: I3a0bce0943931a7546378fb7c7e663e1317b93da
Signed-off-by: alpha lin <alpha.lin@rock-chips.com>

iep/rockchip: revise build fail when CONFIG_RK_IOMMU disabled

When CONFIG_RK_IOMMU disabled, iep build will throw out error
information for some mismatch definitions.

Change-Id: I0fb22550eaaebd62523d794e45de7b94fae8db63
Signed-off-by: alpha lin <alpha.lin@rock-chips.com>

ARM64: config: enable DRM relevant config

Change-Id: Id86ee59190b4f45d5e0e3e8e114920ec28b3fa8e
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

ARM64: dts: rk3399: add mipi node

Change-Id: I06562ff3b62efa38f84ac892513725dcf4559471
Signed-off-by: Chris Zhong <zyw@rock-chips.com>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

Document: update dw_mipi_dsi document for RK3399

There is a phy config clock in RK3399, it must be control by mipi
driver.

Change-Id: I5c029b79ae5867b652ab761dc7416f78f8e070d2
Signed-off-by: Chris Zhong <zyw@rock-chips.com>

DRM/rockchip: mipi: add a phy config clock control

Thers is a phy config clock in RK3399, it must be enable before phy
init, and be disable after phy init.

Change-Id: Idb2d4c85f5284065c3f1d540d9e2fddf5565040d
Signed-off-by: Chris Zhong <zyw@rock-chips.com>

drm/rockchip: fix compile warning

fix warning:
    warning: format '%x' expects argument of type 'unsigned int',
    but argument 2 has type 'size_t'

Change-Id: Ifab0d16f0229aa3d5fc244678298fa2138bd4aa1
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

ARM64: dts: rk3399-tb: add wifi/sdio/sdcard support

Change-Id: Id5b97f2eb3b1bd2eeb42882743a3e64f59d45128
Signed-off-by: xiaoyao <xiaoyao@rock-chips.com>

DRM: mipi: support rk3399 mipi dsi

The vopb/vopl switch register of rk3399 mipi is different from rk3288,
the default setting for mipi dsi mode is different too, so add a
of_device_id structure to distinguish them, and make sure set the
correct mode before mipi phy init.

Change-Id: I54542752dddd1b28fc0500c0a763f14c29fe98f0
Signed-off-by: Chris Zhong <zyw@rock-chips.com>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

ARM64: rockchip: rk3399: clk: fix i2c4 and i2c8 gate-register

Fix a typo making the sclk_i2c4 and sclk_i2c8 access a
wrong register bit offset to handle its gate.

Change-Id: I836244b8e14aa34ef44241c8ff24ebd6b63ed23a
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>

ARM64: dts: rk3399: chrome: enable vop and vop iommu node

Change-Id: I0a92ac277c76d086703fe844bc0fa34a68a428c3
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

UPSTREAM: thermal: of-thermal: allow setting trip_temp on hardware

In current of-thermal, the .set_trip_temp only support to
set trip_temp for SW. But some sensors support to set
trip_temp on hardware, so that can trigger interrupt,
shutdown or any other events.
This patch adds .set_trip_temp() callback in
thermal_zone_of_device_ops{}, so that the sensor device can
use it to set trip_temp on hardware.

Change-Id: I879ea144a9ac21a5032dd885118887b4942c96cf
Signed-off-by: Wei Ni <wni@nvidia.com>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
(cherry picked from git.kernel.org next/linux-next.git master
 commit 60f9ce3ada53498409d51da06502392884933f89)

UPSTREAM: thermal: convert rockchip_thermal to use devm_thermal_zone_of_sensor_register

This changes the driver to use the devm_ version
of thermal_zone_of_sensor_register and cleans
up the  local points and unregister calls.

Change-Id: I9c13e9c91a626a0ef8df69edf3f383c494536bab
Cc: Zhang Rui <rui.zhang@intel.com>
Cc: Heiko Stuebner <heiko@sntech.de>
Cc: linux-pm@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-rockchip@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Tested-by: Caesar Wang <wxt@rock-chips.com>
Reviewed-by: Caesar Wang <wxt@rock-chips.com>
Reviewed-by: Heiko Stuebner <heiko@sntech.de>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
(cherry picked from git.kernel.org next/linux-next.git master
 commit 93c2f85df90edfc515f99cacd7720e286861f744)

UPSTREAM: thermal: of-thermal: Add devm version of thermal_zone_of_sensor_register

Add resource managed version of thermal_zone_of_sensor_register() and
thermal_zone_of_sensor_unregister().

This helps in reducing the code size in error path, remove of
driver remove callbacks and making proper sequence for deallocations.

Change-Id: Ie2e47805f398ff7b4f1dcc71e0581221d79ba9d2
Signed-off-by: Laxman Dewangan <ldewangan@nvidia.com>
Signed-off-by: Eduardo Valentin <edubezval@gmail.com>
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
(cherry picked from git.kernel.org next/linux-next.git master
 commit e498b4984db82b4ba3ceea7dba813222a31e9c2e)

ARM64: dts: rockchip: enable the gt9xx touchscreen on rk3399-tb dts

The gt9xx touchscreen should be used by rk3399 evb board.
The rk3399-monkey.dtsi and rk3399-chrome.dtsi just run the different OS.
So... we should move the touchscreen node into the rk3399-tb.dtsi.

Change-Id: Ida8203e045e0fc0eb49e8a37e4ad609c230e040f
Signed-off-by: Caesar Wang <wxt@rock-chips.com>

ARM64: rockchip_cros_defconfig: add the gpio regulator

The gru/kevin board works the iodomain with gpio regulator.
That's useful for rk3399 kevin/gru board.

Change-Id: I70b6185e3a21a038347b9f8ccd679908817184dd
Signed-off-by: Caesar Wang <wxt@rock-chips.com>

ARM64: rockchip_cros_defconfig: Turn on kgdb

Technically emerging the kernel with USE=kgdb is supposed to get most of
this.  ...and the kernel command line is supposed to come from
elsewhere.  Until we get that happier, maybe this CL is useful.

Change-Id: Idd0100b623eb88b4a4a26922754b71f5cbcd602a
Signed-off-by: Douglas Anderson <dianders@chromium.org>

ARM64: rockchip_cros_defconfig: turn on dwc3 / XHCI

Because USB is important.

Change-Id: I8131a57277922bf13b215fef56e23f3e06cb20d2
Signed-off-by: Douglas Anderson <dianders@chromium.org>

ARM64: rockchip_cros_defconfig: turn on REGULATOR_DEBUG

It's useful. Might be something we can leave on even in a real kernel.
We'll have to see. We have common clock debugging...

Change-Id: I3c1cb55b067a4c54bb425d23664c5f1474016c92
Signed-off-by: Douglas Anderson <dianders@chromium.org>

ARM64: configs: rockchip_cros_defconfig: enable es8316 codec

Change-Id: Iea0fa47a1072a4193ade1e87da68d7cc58c59ece
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

ARM64: dts: rk3399: add VOP and VOP iommu node

Change-Id: I65a9a8797408cf0a77238971852f3b8ebba217b0
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

drm/rockchip: rewrite IOMMU support code

This patch is learn from Marek Szyprowski's patch:
  (drm/exynos: rewrite IOMMU support code)

The patch replaces usage of ARM-specific IOMMU/DMA-mapping related calls
with new generic code for managing DMA-IOMMU integration layer. It also
removes all the hacks, which were needed to configure common DMA/IO address
space on the virtual rockchip-drm device.

Change-Id: I5d2b90002bf135a72ce30cc8503a7d06769835f3
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

dt-bindings: add document for rk3399-vop

Change-Id: I2605d9364cc78c3e5c64ea29bb4c1dfe1fe2f97c
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

drm/rockchip: vop: add rk3399 vop support

There are two VOP in rk3399 chip, respectively VOP_BIG and VOP_LIT.
most registers layout of this two vop is same, their framework arm both
VOP_FULL, the Major differences of this two is that:

VOP_BIG max output resolution is 4096x2160.
VOP_LIT max output resolution is 2560x1600

VOP_BIG support four windows.
VOP_LIT only support two windows.

RK3399 vop register layout is similar with rk3288, so some feature can reuse
with rk3288.

Change-Id: I76f94c93b0e63e4fbba51755e92c604211613e8b
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

ARM64: dts: rockchip: support the gru/kevin boards for rk3399 SoCs

This initial patch adds to support the gru/kevin board for rk3399 SoCs.

It builds at least:
  make -j32 ARCH=arm64 CROSS_COMPILE=aarch64-cros-linux-gnu- dtbs

Change-Id: I7f3841513da130c107aca0d6b393b2bf269a5396
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Caesar Wang <wxt@rock-chips.com>

UPSTREAM: ASoC: rockchip: add bindings for rk3399 i2s

Add devicetree bindings for i2s controller found on rk3399
processors from rockchip.

It's helpful to add full set of compatible strings for serials
of Rockchip SoCs (rk3066, rk3188, rk3288, rk3399).

Change-Id: Ida3f9ffecda52d81016bdb6edb640568ed8c872a
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Jianqun Xu <jay.xu@rock-chips.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
(cherry picked from commit 255edcdfab3d1a45b78d1dacd845d6da5fc64b03)

UPSTREAM: ASoC: rockchip: i2s: Add SNDRV_PCM_FMTBIT_S32_LE support

Change-Id: I9b769b160410cdf0a06c2d6b3b77d563d3672beb
Signed-off-by: Michael Trimarchi <michael@amarulasolutions.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sugar Zhang <sugar.zhang@rock-chips.com>
(cherry picked from commit 4ab936d1aca69978dc738592a00e34f836bda1c3)

UPSTREAM: ASoC: rockchip: i2s: compatible with different chips

there maybe more than one i2s module inside chip, and these i2s modules
have different channels features.

for example: there are 3 i2s in rk3066, one support 8 channels playback
and 2 channels capture, but the others only support 2 channels playback
and 2 channels capture.

in order to compatible with these various chips, we add playback and
capture property to specify these values.

there are default channels configuration in driver: 8 channels playback
and 2 channels capture. if not add property, we use the default values.

Change-Id: I45f3214160877223ba9722bd38a36584e416b14d
Signed-off-by: Sugar Zhang <sugar.zhang@rock-chips.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
(cherry picked from commit c4f9374ddc461ed76be30f4d354a6d1ecb94dfa5)

UPSTREAM: ASoC: rockchip: i2s: change bclk and lrck according to sample rates

This patch sets the dividers autonomously.

when i2s works on master mode, and sample rates changed. We need to change
bclk and lrck at the same time for cpu internal side.

As the input source clock to the module is MCLK_I2S,
and by the divider of the module, the clock generator generates
SCLK and LRCK to transmitter and receiver.

Change-Id: I377f0f08656659787b980785fab0b69197b7b80b
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
(cherry picked from commit 2458c37779ddb91b4109949d86f5a5e193ba415b)

ARM64: dts: rk336x: fix enable incorrect HCLK_I2Sx when startup

This patch like below:
----
commit 3860aa1ccfe01adb6c3fd09e880d812ceb408e5c
Author: Heiko Stuebner <heiko@sntech.de>
Date:   Sat Jan 9 03:18:51 2016 +0100

    ARM: dts: rockchip: swap i2s clock ordering on rk3036

    For sound setups using the simple-card mechanism, the main clock
    (sysclk) is expected to be the first element. For the i2s-driver
    itself it doesn't matter, as it uses named clocks, so we can just
    swap them.
----
If we set HCLK_I2Sx at first, rockchip_i2s_set_sysclk will set the
HCLK_I2S freq (from example is 100MHz) to set the i2s_div, it is
incorrect.

Change-Id: Iab69d541c47d1293a784ebffc23f6c1ceaf9c0b1
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>

drm/rockchip: get rid of rockchip_drm_crtc_mode_config

We need to take care of the vop status when use rockchip_drm_crtc_mode_config,
if vop is disabled, the function would failed, that is terrible.

Save connector type and output mode on drm_display_mode->private_flags on
connector mode_fixup, then we can configure the type and mode safely
on crtc mode_set.

Change-Id: I129cf8a2f100fc19fe96f1d8985e905bea477e28
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

ARM64: dts: rk3399: fix enable incorrect HCLK_I2Sx when startup

This patch like below:
----
commit 3860aa1ccfe01adb6c3fd09e880d812ceb408e5c
Author: Heiko Stuebner <heiko@sntech.de>
Date:   Sat Jan 9 03:18:51 2016 +0100

    ARM: dts: rockchip: swap i2s clock ordering on rk3036

    For sound setups using the simple-card mechanism, the main clock
    (sysclk) is expected to be the first element. For the i2s-driver
    itself it doesn't matter, as it uses named clocks, so we can just
    swap them.
----
If we set HCLK_I2Sx at first, rockchip_i2s_set_sysclk will set the
HCLK_I2S freq (from example is 100MHz) to set the i2s_div, it is
incorrect.

Change-Id: I2b424ded3845b8ccd3ef233e43c5f9f915544547
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

arm64: dts: set right backlight pwm polarity for rk3399-tb

Change-Id: Icaf5a12d3979ec6e31b8dbeaa5df46c6128f62c6
Signed-off-by: David Wu <david.wu@rock-chips.com>

iommu/rockchip: enable upstream iommu on ARM64 platform

Change-Id: I8470750d63b890322d48bc47c07c516f2e3ba270
Signed-off-by: Simon <xxm@rock-chips.com>

iommu/rockchip: enable upstream iommu build on ARM64 platform

Change-Id: Ied95b0a8d531210729a960b90dab81f1e25e9035
Signed-off-by: Simon <xxm@rock-chips.com>

iommu/rockchip: add ARM64 cache flush operation for iommu

Change-Id: I5848ca01a38e902a436b5b4abccfe235e2746cc1
Signed-off-by: Simon <xxm@rock-chips.com>

Documentation: add rockchip,rk3399-lcdc to rockchip_lcdc.txt

Change-Id: I4a82256650e2bc1e4861202a6002b5c2dca07cb1
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

ARM64: dts: rk3399: Add vpu and rkvdec resource node

Add vpu and rkvdec resource node to enable video codec
supporting for rk3399 android.

Change-Id: I1689955858355b6061957dc43eea17f9b8d71096
Signed-off-by: alpha.lin <alpha.lin@rock-chips.com>

FROMLIST: drm/rockchip: dw_hdmi: Don't call platform_set_drvdata()

The Rockchip dw_hdmi driver just called platform_set_drvdata() to get
your hopes up that maybe, somehow, you'd be able to retrieve the 'struct
rockchip_hdmi' from a pointer to the 'struct device'.  You can't.  When
we call dw_hdmi_bind() the main driver calls dev_set_drvdata(), which
clobbers our setting.

Let's just remove the platform_set_drvdata() to avoid dashing people's
hopes.

(am from https://patchwork.kernel.org/patch/8523401/)

Change-Id: I28c4dcff37f6800b841e0492eb2613dcff7d1c81
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

FROMLIST: drm/rockchip: vop: Fix vop crtc cleanup

This fixes a few problems in the vop crtc cleanup (handling error
paths and cleanup upon exit):

* The vop_create_crtc() error path had an unsafe version of the
  iterator used for iterating over all planes (though it was
  destroying planes in the iterator so should have used the safe
  version)

* vop_destroy_crtc() - wasn't calling vop_plane_destroy(), which made
  slub_debug unhappy, at least if we ended up running this due to a
  deferred probe.

* In vop_create_crtc() if we were missing the "port" device tree node
  we would fail but not return an error (found by code inspection).

Fix these problems.

(am from https://patchwork.kernel.org/patch/8523361/)

Change-Id: I3c00faca6e2fc10edc5b4576012ac28b6809a2f3
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

FROMLIST: drm/rockchip: dw_hdmi: Call drm_encoder_cleanup() in error path

The drm_encoder_cleanup() was missing both from the error path of
dw_hdmi_rockchip_bind().  This caused a crash when slub_debug was
enabled and we ended up deferring probe of HDMI at boot.

This call isn't needed from unbind() because if dw_hdmi_bind() returns
no error then it takes over the job of freeing the encoder (in
dw_hdmi_unbind).

(am from https://patchwork.kernel.org/patch/8523331/)

Change-Id: Ibf5c39a5db304177a9f16d8dc691221512002348
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

FROMLIST: drm/rockchip: vop: Disable planes when disabling CRTC

When a VOP is re-enabled, it will start scanning right away the
framebuffers that were configured from the last time, even if those have
been destroyed already.

To prevent the VOP from trying to access freed memory, disable all its
windows when the CRTC is being disabled, then each window will get a
valid framebuffer address before it's enabled again.

Link: http://lkml.kernel.org/g/CAAObsKAv+05ih5U+=4kic_NsjGMhfxYheHR8xXXmacZs+p5SHw@mail.gmail.com
(am from https://patchwork.kernel.org/patch/8643631/)

Change-Id: Iaacb1624b4351a94d663ec73d9174b0fd4bc4b54
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

FROMLIST: drm/rockchip: vop: Don't reject empty modesets

So that when DRM_IOCTL_MODE_SETCRTC is called without a FB nor mode, the
CRTC gets disabled.

Link: http://lkml.kernel.org/g/CAAObsKAv+05ih5U+=4kic_NsjGMhfxYheHR8xXXmacZs+p5SHw@mail.gmail.com
(am from https://patchwork.kernel.org/patch/8618471/)

Change-Id: I3a36e20f727f2087d718acec886766b743de2d9b
Signed-off-by: Tomeu Vizoso <tomeu.vizoso@collabora.com>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

FROMLIST: drm/rockchip: cancel pending vblanks on close

When closing the DRM device while a vblank is pending, we access
file_priv after it has been free'd, which gives:

  Unable to handle kernel NULL pointer dereference at virtual address 00000000
  ...
  PC is at __list_add+0x5c/0xe8
  LR is at send_vblank_event+0x54/0x1f0
  ...
  [<c02952e8>] (__list_add) from [<c031a7b4>] (send_vblank_event+0x54/0x1f0)
  [<c031a760>] (send_vblank_event) from [<c031a9c0>] (drm_send_vblank_event+0x70/0x78)
  [<c031a950>] (drm_send_vblank_event) from [<c031a9f8>] (drm_crtc_send_vblank_event+0x30/0x34)
  [<c031a9c8>] (drm_crtc_send_vblank_event) from [<c0339ad8>] (vop_isr+0x224/0x28c)
  [<c03398b4>] (vop_isr) from [<c0081780>] (handle_irq_event_percpu+0x12c/0x3e4)

This can be triggered somewhat reliably with:

modetest -M rockchip -v -s ...

Add a preclose hook to the driver so that we can discard any pending
vblank events when the device is closed.

(am from https://patchwork.kernel.org/patch/8568111)

Change-Id: Icce075cf22f3a9d7b2157c29a47b370160b0c8d8
Signed-off-by: John Keeping <john@metanate.com>
Signed-off-by: Mark Yao <mark.yao@rock-chips.com>

arm64: dts: add i2c0, i2c1, i2c4 rise and fall time for rk3399-tb

Change-Id: Ibbd7d7ee7bfce25fe104b03ea742c0bab2b4b586
Signed-off-by: David Wu <david.wu@rock-chips.com>

arm64: dtsi: rk3399-tb: add node of GPU

Change-Id: I52f15bf127b360a2bfe1356de1ed7f4d97b4534b
Signed-off-by: chenzhen <chenzhen@rock-chips.com>

drm/rockchip: analogix_dp: add rk3399 eDP support

RK3399 and RK3288 shared the same eDP IP controller, so this time we
just need to append the RK3399 compatible name to analogix_dp documentation
and driver code.

Change-Id: I3fee6893c56698ee2948b9df2f3ffb7729fe75ef
Signed-off-by: Yakir Yang <ykk@rock-chips.com>

drm: bridge: analogix_dp: correct the register bit define error in ANALOGIX_DP_PLL_REG_1

There're an register define error in ANALOGIX_DP_PLL_REG_1 which introduced
by commit 45970584ead0a5dfe27a6edef198ede536ad37ba (FROMLIST: drm: bridge:
analogix/dp: add some rk3288 special registers setting).

The PHY PLL input clock source is selected by ANALOGIX_DP_PLL_REG_1
BIT 0, not BIT 1.

Change-Id: I8cb806d23144697225f626aaa2af19e6379dfe51
Signed-off-by: Yakir Yang <ykk@rock-chips.com>

Merge branch 'android-4.4' of https://android.googlesource.com/kernel/common

* android-4.4: (34 commits)
  sdcardfs: remove unneeded __init and __exit
  sdcardfs: Remove unused code
  fs: Export d_absolute_path
  sdcardfs: remove effectless config option
  inotify: Fix erroneous update of bit count
  fs: sdcardfs: Declare LOOKUP_CASE_INSENSITIVE unconditionally
  trace: cpufreq: fix typo in min/max cpufreq
  sdcardfs: Add support for d_canonical_path
  vfs: add d_canonical_path for stacked filesystem support
  sdcardfs: Bring up to date with Android M permissions:
  Changed type-casting in packagelist management
  Port of sdcardfs to 4.4
  Included sdcardfs source code for kernel 3.0
  ANDROID: usb: gadget: Add support for MTP OS desc
  CHROMIUM: usb: gadget: f_accessory: add .raw_request callback
  CHROMIUM: usb: gadget: audio_source: add .free_func callback
  CHROMIUM: usb: gadget: f_mtp: fix usb_ss_ep_comp_descriptor
  CHROMIUM: usb: gadget: f_mtp: Add SuperSpeed support
  FROMLIST: mmc: block: fix ABI regression of mmc_blk_ioctl
  FROMLIST: mm: ASLR: use get_random_long()
  ...

Change-Id: I88fa8a7f6bfc80bee98503b9ede0fee08d62814c

Revert "mmc: block: ioctl return EINVAL if cmd unknown"

This reverts commit 7c1abbfe2a27bae41202d4b46687f82eba0c1d5b.
Will fixes by aosp commit "mmc: block: fix ABI regression of mmc_blk_ioctl"

Signed-off-by: Huang, Tao <huangtao@rock-chips.com>

clk: rockchip: rk3399: Add more pll rates for HDMI

Modify the 594MHz parameter for higher VCO freq to reduce the jitter.

Change-Id: I78784210b69d6895758192c84724b982fcc9e72d
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

sdcardfs: remove unneeded __init and __exit

Change-Id: I2a2d45d52f891332174c3000e8681c5167c1564f

sdcardfs: Remove unused code

Change-Id: Ie97cba27ce44818ac56cfe40954f164ad44eccf6

pinctrl: rockchip: fix pull setting error for rk3399

Change-Id: Ie52226e751510373abc75754ab756a36ae4f3aaa
Signed-off-by: David Wu <david.wu@rock-chips.com>

ARM64: dts: rk3399-tb: add sound node

Change-Id: Ibcb2e6baae871c6456e1cbe6b145cb64bb007c7d
Signed-off-by: Sugar Zhang <sugar.zhang@rock-chips.com>

ARM64: configs: rockchip_defconfig: enable es8316 codec

Change-Id: I799d23f12b26b99da8d77c3c67c30393e461c589
Signed-off-by: Sugar Zhang <sugar.zhang@rock-chips.com>

ASoC: es8316: add codec driver

Change-Id: I6896036a7af922747180535939074f399f8c38d1
Signed-off-by: Sugar Zhang <sugar.zhang@rock-chips.com>

ARM64: rockchip_defconfig: enable mp8865 regulator

Change-Id: I612fbe41e97f4e06dc154f76a256c7996d7ea09a
Signed-off-by: Zain Wang <wzz@rock-chips.com>

regulator: add devicetree bindings for mp8865 regulators

Change-Id: Ie837faba6a9f599dcfc5cdbdaffe12ca81060957
Signed-off-by: Zain Wang <wzz@rock-chips.com>

ARM64: dts: rk3399-tb: add support for mp8865

Change-Id: I71072722fde924f151986d6d2dc916cb868e7b98
Signed-off-by: Zain Wang <wzz@rock-chips.com>

regulator: mp8865: add regulator driver for MP8865

Change-Id: I5fa8423e5d1e301a008dcdfd60f93c442f6211a8
Signed-off-by: Zain Wang <wzz@rock-chips.com>

UPSTREAM: soc: rockchip: power-domain: Modify power domain driver for rk3399

This driver is modified to support RK3399 SoC.

Change-Id: Id4d040a78e7e18d78e9644b14e0611fc6ab44abe
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
[small indentation fixups]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
(cherry picked from git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git
 v4.7-armsoc/drivers commit fd8b62cc38b356bcdf20ac8f1a647db7b11240cf)

UPSTREAM: dt-bindings: add binding for rk3399 power domains

Add binding documentation for the power domains
found on Rockchip RK3399 SoCs

Change-Id: I51d70a08c86b5361ac5d51151711e07ffa3046ef
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Acked-by: Kevin Hilman <khilman@baylibre.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
(cherry picked from git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git
 v4.7-armsoc/drivers commit e6e270aecbb0ed605b7267fdf6f828945014de24)

UPSTREAM: soc: rockchip: power-domain: add support for sub-power domains

This patch adds support for making one power domain a sub-domain of
other domain. This is useful for modeling power dependences,
which needs to have more than one power domain enabled to be operational.

Change-Id: Ie2a79aab6e1e073157c06d44252ef327caee5261
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
[restructured error handling in subdomain-addition]
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
(cherry picked from git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git
 v4.7-armsoc/drivers commit 6be05b5ec16132f3df3f1d857ab01e30f726b542)

UPSTREAM: soc: rockchip: power-domain: allow domains only handling idle requests

On some Rockchip SoC there exist child-domains only handling their
idle state with the actual power-state handled by a (shared) parent-
domain.

So allow such types of domains. For them, we can determine their
state (on/off) by checking the inverse idle-state instead.

There exist one special case if both idle as well power handling
were set as not present, but as the domain-data is defined in the
code itself, we can expect the reasonable developer to define them
in a correct way, without adding more checks.

Change-Id: Ic82cb387565a39043d5d52e62a94910de10d4bbd
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
(cherry picked from git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git
 v4.7-armsoc/drivers commit 1fe767a56c32730a947fc2bfcac5d14490bde6ff)

UPSTREAM: soc: rockchip: power-domain: make idle handling optional

Not all new socs need to handle idle states on domain state changes,
so add the possibility to make them optional.

Change-Id: I46d869e1de9e03ec0664518effbcf2642053391e
Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
(cherry picked from git://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git
 v4.7-armsoc/drivers commit 6aa841c8097fee1b17b343719c45694e3166ca34)

ARM64: dts: rk3399-tb: enable gpu node

Change-Id: I1199eae97ec4400a96bab02ca8ed1db522fa742d
Signed-off-by: Jianqun Xu <jay.xu@rock-chips.com>

video: rockchip: mipi: 3399: add support rk3399

Change-Id: I52fbecb5c6739b8ed1e35c3c3be7778a189874a6
Signed-off-by: xubilv <xbl@rock-chips.com>

arm64: dtsi: rk3399: add node of GPU

Change-Id: If48ed7f58aa00c28122ac77a8d79ab675f3a6208
Signed-off-by: chenzhen <chenzhen@rock-chips.com>

phy: rockchip-usb: Fixed the port cannot be resumed after suspended.

selecting utmi interface signals from utmi interface of usb20
host0 controller to usb2phy, when phy is resumed.

Change-Id: I487e836b89177cd8bc2dc56400f4dc277c8d2bf0
Signed-off-by: Frank Wang <frank.wang@rock-chips.com>

ARM64: rockchip_defconfig: open rk322x vop

Change-Id: I9df6706a34fecadd5f149cd566b31323dafb51c2
Signed-off-by: Huang Jiachai <hjc@rock-chips.com>

ARM64: dts: rk3399-monkey: add fb node

Change-Id: I1afb5fde325e9227e30fcdd5efbea4f44818de59
Signed-off-by: Huang Jiachai <hjc@rock-chips.com>
Signed-off-by: Jianqun Xu <jay.xu@rock-chips.com>

dtsi: screen-timing: lcd-tv080wum-nl0-mipi: for rk3399 monkey

Change-Id: I01ca2bd999973820317bcf46f8a7d3e85d59a606
Signed-off-by: Huang Jiachai <hjc@rock-chips.com>

video: rockchip: vop: 3399: add support rk3399

Change-Id: Icbccfdd4fb841df67c0ade1cfd141fb574d837f0
Signed-off-by: Huang Jiachai <hjc@rock-chips.com>

clk: rockchip: rk3399: use the FRACMUX type for spdif and i2s0~2

Change-Id: I1f0f60185e70bb894010137db8a35dda218201f2
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

video: rockchip: mipi: enable clk before write register

Change-Id: I717ed1143c53e7c2cd04697e8cc3984f89e29504
Signed-off-by: xubilv <xbl@rock-chips.com>

clk: rockchip: rk3399: pclkin_isp1_wrapper should source from pclkin_cif

Change-Id: Id2ccc4003fcd286594fa2c2fc366f7c922417883
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

ARM64: dts: rk3399: remove the reference pmugrf and grf

We don't need to reference the pmugrf/grf in the clock driver any more.

Change-Id: Ibda203163c84ab4004e1225e5868267024069199
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

ARM64: dts: rk3399-monkey: support the gt9xx touchscreen

Change-Id: Iabe2264bf9cffe09259b17912d6391dfb87ee4dd
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
Signed-off-by: Jianqun Xu <jay.xu@rock-chips.com>

clk: rockchip: rk3399: remove re-enable pmucru clk_gate/pclk_alive

These are not gating default when the SoC startup, so we don't need
to re-enable them.

Change-Id: I956a31345fe7f24b973db6c9e49d87a2988ac7d6
Signed-off-by: Xing Zheng <zhengxing@rock-chips.com>

iommu/rockchip: add more judgement for virtual device who using iommu

A virtual device like "drm" call iommu_attach_device may fail.
Current only judge if a device has "group", this is not enough,
"group->iommu_data" is needed

Change-Id: I1a66d6016dfef867d83aa4cccaf223ced4e07161
Signed-off-by: Simon <xxm@rock-chips.com>