ftrace: Synchronize the initialization of ftrace_opsgrafted

There's some internal state that ftrace_ops needs to have set, but since
it can be declared outside of the ftrace.c code, it calls
ftrace_ops_init() on the ops in every global function. The issue is that
if two tasks call it on the same ops at the same time it is possible to
have the initialization of one corrupt the initialization of the other
call.

Create a ops_mutex to use to synchronize every initialization of the
ftrace_ops. The mutex is taken within checking the ftrace_ops flag that
states it was initializied but the flag is checked again after the mutex
has been taken. Checking first outside the mutex allows it to shortcut
having to take the mutex. But then the check needs to be done again after
the mute is taken in case of races.

Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260902095501.6b59af20@gandalf.local.home
Fixes: f04f24fb7e48d ("ftrace, kprobes: Fix a deadlock on ftrace_regex_lock")
Reported-by: sashiko-bot@kernel.org
Close: https://lore.kernel.org/all/20260829025528.49A831F000E9@smtp.kernel.org/
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>

49 files changed, 25928 insertions, 0 deletions

diff --git a/drivers/gpu/drm/tegra/Kconfig b/drivers/gpu/drm/tegra/Kconfig
new file mode 100644
index 000000000..8a3b16aac
--- /dev/null
+++ b/drivers/gpu/drm/tegra/Kconfig
@@ -0,0 +1,46 @@
+# SPDX-License-Identifier: GPL-2.0-only
+config DRM_TEGRA
+	tristate "NVIDIA Tegra DRM"
+	depends on ARCH_TEGRA || COMPILE_TEST
+	depends on COMMON_CLK
+	depends on DRM
+	depends on OF
+	select DRM_CLIENT_SELECTION
+	select DRM_DISPLAY_DP_HELPER
+	select DRM_DISPLAY_HDMI_HELPER
+	select DRM_DISPLAY_HELPER
+	select DRM_BRIDGE_CONNECTOR
+	select DRM_DISPLAY_DP_AUX_BUS
+	select DRM_KMS_HELPER
+	select DRM_MIPI_DSI
+	select DRM_PANEL
+	select FB_DMAMEM_HELPERS if DRM_FBDEV_EMULATION
+	select TEGRA_HOST1X
+	select INTERCONNECT
+	select IOMMU_IOVA
+	select CEC_CORE if CEC_NOTIFIER
+	select SND_SIMPLE_CARD if SND_SOC_TEGRA20_SPDIF
+	select SND_SOC_HDMI_CODEC if SND_SOC_TEGRA20_SPDIF
+	select SND_AUDIO_GRAPH_CARD if SND_SOC_TEGRA20_SPDIF
+	help
+	  Choose this option if you have an NVIDIA Tegra SoC.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called tegra-drm.
+
+if DRM_TEGRA
+
+config DRM_TEGRA_DEBUG
+	bool "NVIDIA Tegra DRM debug support"
+	help
+	  Say yes here to enable debugging support.
+
+config DRM_TEGRA_STAGING
+	bool "Enable HOST1X interface"
+	depends on STAGING
+	help
+	  Say yes if HOST1X should be available for userspace DRM users.
+
+	  If unsure, choose N.
+
+endif
diff --git a/drivers/gpu/drm/tegra/Makefile b/drivers/gpu/drm/tegra/Makefile
new file mode 100644
index 000000000..e399b40d6
--- /dev/null
+++ b/drivers/gpu/drm/tegra/Makefile
@@ -0,0 +1,35 @@
+# SPDX-License-Identifier: GPL-2.0
+ccflags-$(CONFIG_DRM_TEGRA_DEBUG) += -DDEBUG
+
+tegra-drm-y := \
+	drm.o \
+	uapi.o \
+	submit.o \
+	firewall.o \
+	gem.o \
+	fb.o \
+	dp.o \
+	hub.o \
+	plane.o \
+	dc.o \
+	output.o \
+	rgb.o \
+	hda.o \
+	hdmi.o \
+	mipi-phy.o \
+	dsi.o \
+	sor.o \
+	dpaux.o \
+	gr2d.o \
+	gr3d.o \
+	falcon.o \
+	vic.o \
+	nvdec.o \
+	nvjpg.o \
+	riscv.o
+
+tegra-drm-y += trace.o
+
+tegra-drm-$(CONFIG_DRM_FBDEV_EMULATION) += fbdev.o
+
+obj-$(CONFIG_DRM_TEGRA) += tegra-drm.o
diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c
new file mode 100644
index 000000000..0b3fcc701
--- /dev/null
+++ b/drivers/gpu/drm/tegra/dc.c
@@ -0,0 +1,3424 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#include <linux/clk.h>
+#include <linux/debugfs.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/iommu.h>
+#include <linux/interconnect.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_domain.h>
+#include <linux/pm_opp.h>
+#include <linux/pm_runtime.h>
+#include <linux/reset.h>
+
+#include <soc/tegra/common.h>
+#include <soc/tegra/pmc.h>
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_blend.h>
+#include <drm/drm_debugfs.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_print.h>
+#include <drm/drm_vblank.h>
+
+#include "dc.h"
+#include "drm.h"
+#include "gem.h"
+#include "hub.h"
+#include "plane.h"
+
+static const u16 default_srgb_lut[] = {
+	0x6000, 0x60CE, 0x619D, 0x626C, 0x632D, 0x63D4, 0x6469, 0x64F0, 0x656B, 0x65DF, 0x664A,
+	0x66B0, 0x6711, 0x676D, 0x67C4, 0x6819, 0x686A, 0x68B8, 0x6904, 0x694D, 0x6994, 0x69D8,
+	0x6A1B, 0x6A5D, 0x6A9C, 0x6ADA, 0x6B17, 0x6B52, 0x6B8C, 0x6BC5, 0x6BFD, 0x6C33, 0x6C69,
+	0x6C9E, 0x6CD1, 0x6D04, 0x6D36, 0x6D67, 0x6D98, 0x6DC7, 0x6DF6, 0x6E25, 0x6E52, 0x6E7F,
+	0x6EAC, 0x6ED7, 0x6F03, 0x6F2D, 0x6F58, 0x6F81, 0x6FAA, 0x6FD3, 0x6FFB, 0x7023, 0x704B,
+	0x7071, 0x7098, 0x70BE, 0x70E4, 0x7109, 0x712E, 0x7153, 0x7177, 0x719B, 0x71BF, 0x71E2,
+	0x7205, 0x7227, 0x724A, 0x726C, 0x728E, 0x72AF, 0x72D0, 0x72F1, 0x7312, 0x7333, 0x7353,
+	0x7373, 0x7392, 0x73B2, 0x73D1, 0x73F0, 0x740F, 0x742D, 0x744C, 0x746A, 0x7488, 0x74A6,
+	0x74C3, 0x74E0, 0x74FE, 0x751B, 0x7537, 0x7554, 0x7570, 0x758D, 0x75A9, 0x75C4, 0x75E0,
+	0x75FC, 0x7617, 0x7632, 0x764D, 0x7668, 0x7683, 0x769E, 0x76B8, 0x76D3, 0x76ED, 0x7707,
+	0x7721, 0x773B, 0x7754, 0x776E, 0x7787, 0x77A0, 0x77B9, 0x77D2, 0x77EB, 0x7804, 0x781D,
+	0x7835, 0x784E, 0x7866, 0x787E, 0x7896, 0x78AE, 0x78C6, 0x78DD, 0x78F5, 0x790D, 0x7924,
+	0x793B, 0x7952, 0x796A, 0x7981, 0x7997, 0x79AE, 0x79C5, 0x79DB, 0x79F2, 0x7A08, 0x7A1F,
+	0x7A35, 0x7A4B, 0x7A61, 0x7A77, 0x7A8D, 0x7AA3, 0x7AB8, 0x7ACE, 0x7AE3, 0x7AF9, 0x7B0E,
+	0x7B24, 0x7B39, 0x7B4E, 0x7B63, 0x7B78, 0x7B8D, 0x7BA2, 0x7BB6, 0x7BCB, 0x7BE0, 0x7BF4,
+	0x7C08, 0x7C1D, 0x7C31, 0x7C45, 0x7C59, 0x7C6E, 0x7C82, 0x7C96, 0x7CA9, 0x7CBD, 0x7CD1,
+	0x7CE5, 0x7CF8, 0x7D0C, 0x7D1F, 0x7D33, 0x7D46, 0x7D59, 0x7D6D, 0x7D80, 0x7D93, 0x7DA6,
+	0x7DB9, 0x7DCC, 0x7DDF, 0x7DF2, 0x7E04, 0x7E17, 0x7E2A, 0x7E3C, 0x7E4F, 0x7E61, 0x7E74,
+	0x7E86, 0x7E98, 0x7EAB, 0x7EBD, 0x7ECF, 0x7EE1, 0x7EF3, 0x7F05, 0x7F17, 0x7F29, 0x7F3B,
+	0x7F4D, 0x7F5E, 0x7F70, 0x7F82, 0x7F93, 0x7FA5, 0x7FB6, 0x7FC8, 0x7FD9, 0x7FEB, 0x7FFC,
+	0x800D, 0x801E, 0x8030, 0x8041, 0x8052, 0x8063, 0x8074, 0x8085, 0x8096, 0x80A7, 0x80B7,
+	0x80C8, 0x80D9, 0x80EA, 0x80FA, 0x810B, 0x811C, 0x812C, 0x813D, 0x814D, 0x815D, 0x816E,
+	0x817E, 0x818E, 0x819F, 0x81AF, 0x81BF, 0x81CF, 0x81DF, 0x81EF, 0x81FF, 0x820F, 0x821F,
+	0x822F, 0x823F, 0x824F, 0x825F, 0x826F, 0x827E, 0x828E, 0x829E, 0x82AD, 0x82BD, 0x82CC,
+	0x82DC, 0x82EB, 0x82FB, 0x830A, 0x831A, 0x8329, 0x8338, 0x8348, 0x8357, 0x8366, 0x8375,
+	0x8385, 0x8394, 0x83A3, 0x83B2, 0x83C1, 0x83D0, 0x83DF, 0x83EE, 0x83FD, 0x840C, 0x841A,
+	0x8429, 0x8438, 0x8447, 0x8455, 0x8464, 0x8473, 0x8481, 0x8490, 0x849F, 0x84AD, 0x84BC,
+	0x84CA, 0x84D9, 0x84E7, 0x84F5, 0x8504, 0x8512, 0x8521, 0x852F, 0x853D, 0x854B, 0x855A,
+	0x8568, 0x8576, 0x8584, 0x8592, 0x85A0, 0x85AE, 0x85BC, 0x85CA, 0x85D8, 0x85E6, 0x85F4,
+	0x8602, 0x8610, 0x861E, 0x862C, 0x8639, 0x8647, 0x8655, 0x8663, 0x8670, 0x867E, 0x868C,
+	0x8699, 0x86A7, 0x86B5, 0x86C2, 0x86D0, 0x86DD, 0x86EB, 0x86F8, 0x8705, 0x8713, 0x8720,
+	0x872E, 0x873B, 0x8748, 0x8756, 0x8763, 0x8770, 0x877D, 0x878B, 0x8798, 0x87A5, 0x87B2,
+	0x87BF, 0x87CC, 0x87D9, 0x87E6, 0x87F3, 0x8801, 0x880E, 0x881A, 0x8827, 0x8834, 0x8841,
+	0x884E, 0x885B, 0x8868, 0x8875, 0x8882, 0x888E, 0x889B, 0x88A8, 0x88B5, 0x88C1, 0x88CE,
+	0x88DB, 0x88E7, 0x88F4, 0x8900, 0x890D, 0x891A, 0x8926, 0x8933, 0x893F, 0x894C, 0x8958,
+	0x8965, 0x8971, 0x897D, 0x898A, 0x8996, 0x89A3, 0x89AF, 0x89BB, 0x89C8, 0x89D4, 0x89E0,
+	0x89EC, 0x89F9, 0x8A05, 0x8A11, 0x8A1D, 0x8A29, 0x8A36, 0x8A42, 0x8A4E, 0x8A5A, 0x8A66,
+	0x8A72, 0x8A7E, 0x8A8A, 0x8A96, 0x8AA2, 0x8AAE, 0x8ABA, 0x8AC6, 0x8AD2, 0x8ADE, 0x8AEA,
+	0x8AF5, 0x8B01, 0x8B0D, 0x8B19, 0x8B25, 0x8B31, 0x8B3C, 0x8B48, 0x8B54, 0x8B60, 0x8B6B,
+	0x8B77, 0x8B83, 0x8B8E, 0x8B9A, 0x8BA6, 0x8BB1, 0x8BBD, 0x8BC8, 0x8BD4, 0x8BDF, 0x8BEB,
+	0x8BF6, 0x8C02, 0x8C0D, 0x8C19, 0x8C24, 0x8C30, 0x8C3B, 0x8C47, 0x8C52, 0x8C5D, 0x8C69,
+	0x8C74, 0x8C80, 0x8C8B, 0x8C96, 0x8CA1, 0x8CAD, 0x8CB8, 0x8CC3, 0x8CCF, 0x8CDA, 0x8CE5,
+	0x8CF0, 0x8CFB, 0x8D06, 0x8D12, 0x8D1D, 0x8D28, 0x8D33, 0x8D3E, 0x8D49, 0x8D54, 0x8D5F,
+	0x8D6A, 0x8D75, 0x8D80, 0x8D8B, 0x8D96, 0x8DA1, 0x8DAC, 0x8DB7, 0x8DC2, 0x8DCD, 0x8DD8,
+	0x8DE3, 0x8DEE, 0x8DF9, 0x8E04, 0x8E0E, 0x8E19, 0x8E24, 0x8E2F, 0x8E3A, 0x8E44, 0x8E4F,
+	0x8E5A, 0x8E65, 0x8E6F, 0x8E7A, 0x8E85, 0x8E90, 0x8E9A, 0x8EA5, 0x8EB0, 0x8EBA, 0x8EC5,
+	0x8ECF, 0x8EDA, 0x8EE5, 0x8EEF, 0x8EFA, 0x8F04, 0x8F0F, 0x8F19, 0x8F24, 0x8F2E, 0x8F39,
+	0x8F43, 0x8F4E, 0x8F58, 0x8F63, 0x8F6D, 0x8F78, 0x8F82, 0x8F8C, 0x8F97, 0x8FA1, 0x8FAC,
+	0x8FB6, 0x8FC0, 0x8FCB, 0x8FD5, 0x8FDF, 0x8FEA, 0x8FF4, 0x8FFE, 0x9008, 0x9013, 0x901D,
+	0x9027, 0x9031, 0x903C, 0x9046, 0x9050, 0x905A, 0x9064, 0x906E, 0x9079, 0x9083, 0x908D,
+	0x9097, 0x90A1, 0x90AB, 0x90B5, 0x90BF, 0x90C9, 0x90D3, 0x90DD, 0x90E7, 0x90F1, 0x90FB,
+	0x9105, 0x910F, 0x9119, 0x9123, 0x912D, 0x9137, 0x9141, 0x914B, 0x9155, 0x915F, 0x9169,
+	0x9173, 0x917D, 0x9186, 0x9190, 0x919A, 0x91A4, 0x91AE, 0x91B8, 0x91C1, 0x91CB, 0x91D5,
+	0x91DF, 0x91E9, 0x91F2, 0x91FC, 0x9206, 0x9210, 0x9219, 0x9223, 0x922D, 0x9236, 0x9240,
+	0x924A, 0x9253, 0x925D, 0x9267, 0x9270, 0x927A, 0x9283, 0x928D, 0x9297, 0x92A0, 0x92AA,
+	0x92B3, 0x92BD, 0x92C6, 0x92D0, 0x92DA, 0x92E3, 0x92ED, 0x92F6, 0x9300, 0x9309, 0x9313,
+	0x931C, 0x9325, 0x932F, 0x9338, 0x9342, 0x934B, 0x9355, 0x935E, 0x9367, 0x9371, 0x937A,
+	0x9384, 0x938D, 0x9396, 0x93A0, 0x93A9, 0x93B2, 0x93BC, 0x93C5, 0x93CE, 0x93D7, 0x93E1,
+	0x93EA, 0x93F3, 0x93FC, 0x9406, 0x940F, 0x9418, 0x9421, 0x942B, 0x9434, 0x943D, 0x9446,
+	0x944F, 0x9459, 0x9462, 0x946B, 0x9474, 0x947D, 0x9486, 0x948F, 0x9499, 0x94A2, 0x94AB,
+	0x94B4, 0x94BD, 0x94C6, 0x94CF, 0x94D8, 0x94E1, 0x94EA, 0x94F3, 0x94FC, 0x9505, 0x950E,
+	0x9517, 0x9520, 0x9529, 0x9532, 0x953B, 0x9544, 0x954D, 0x9556, 0x955F, 0x9568, 0x9571,
+	0x957A, 0x9583, 0x958C, 0x9595, 0x959D, 0x95A6, 0x95AF, 0x95B8, 0x95C1, 0x95CA, 0x95D3,
+	0x95DB, 0x95E4, 0x95ED, 0x95F6, 0x95FF, 0x9608, 0x9610, 0x9619, 0x9622, 0x962B, 0x9633,
+	0x963C, 0x9645, 0x964E, 0x9656, 0x965F, 0x9668, 0x9671, 0x9679, 0x9682, 0x968B, 0x9693,
+	0x969C, 0x96A5, 0x96AD, 0x96B6, 0x96BF, 0x96C7, 0x96D0, 0x96D9, 0x96E1, 0x96EA, 0x96F2,
+	0x96FB, 0x9704, 0x970C, 0x9715, 0x971D, 0x9726, 0x972E, 0x9737, 0x9740, 0x9748, 0x9751,
+	0x9759, 0x9762, 0x976A, 0x9773, 0x977B, 0x9784, 0x978C, 0x9795, 0x979D, 0x97A6, 0x97AE,
+	0x97B6, 0x97BF, 0x97C7, 0x97D0, 0x97D8, 0x97E1, 0x97E9, 0x97F1, 0x97FA, 0x9802, 0x980B,
+	0x9813, 0x981B, 0x9824, 0x982C, 0x9834, 0x983D, 0x9845, 0x984D, 0x9856, 0x985E, 0x9866,
+	0x986F, 0x9877, 0x987F, 0x9888, 0x9890, 0x9898, 0x98A0, 0x98A9, 0x98B1, 0x98B9, 0x98C1,
+	0x98CA, 0x98D2, 0x98DA, 0x98E2, 0x98EB, 0x98F3, 0x98FB, 0x9903, 0x990B, 0x9914, 0x991C,
+	0x9924, 0x992C, 0x9934, 0x993C, 0x9945, 0x994D, 0x9955, 0x995D, 0x9965, 0x996D, 0x9975,
+	0x997D, 0x9986, 0x998E, 0x9996, 0x999E, 0x99A6, 0x99AE, 0x99B6, 0x99BE, 0x99C6, 0x99CE,
+	0x99D6, 0x99DE, 0x99E6, 0x99EE, 0x99F6, 0x99FE, 0x9A06, 0x9A0E, 0x9A16, 0x9A1E, 0x9A26,
+	0x9A2E, 0x9A36, 0x9A3E, 0x9A46, 0x9A4E, 0x9A56, 0x9A5E, 0x9A66, 0x9A6E, 0x9A76, 0x9A7E,
+	0x9A86, 0x9A8E, 0x9A96, 0x9A9D, 0x9AA5, 0x9AAD, 0x9AB5, 0x9ABD, 0x9AC5, 0x9ACD, 0x9AD5,
+	0x9ADC, 0x9AE4, 0x9AEC, 0x9AF4, 0x9AFC, 0x9B04, 0x9B0C, 0x9B13, 0x9B1B, 0x9B23, 0x9B2B,
+	0x9B33, 0x9B3A, 0x9B42, 0x9B4A, 0x9B52, 0x9B59, 0x9B61, 0x9B69, 0x9B71, 0x9B79, 0x9B80,
+	0x9B88, 0x9B90, 0x9B97, 0x9B9F, 0x9BA7, 0x9BAF, 0x9BB6, 0x9BBE, 0x9BC6, 0x9BCD, 0x9BD5,
+	0x9BDD, 0x9BE5, 0x9BEC, 0x9BF4, 0x9BFC, 0x9C03, 0x9C0B, 0x9C12, 0x9C1A, 0x9C22, 0x9C29,
+	0x9C31, 0x9C39, 0x9C40, 0x9C48, 0x9C50, 0x9C57, 0x9C5F, 0x9C66, 0x9C6E, 0x9C75, 0x9C7D,
+	0x9C85, 0x9C8C, 0x9C94, 0x9C9B, 0x9CA3, 0x9CAA, 0x9CB2, 0x9CBA, 0x9CC1, 0x9CC9, 0x9CD0,
+	0x9CD8, 0x9CDF, 0x9CE7, 0x9CEE, 0x9CF6, 0x9CFD, 0x9D05, 0x9D0C, 0x9D14, 0x9D1B, 0x9D23,
+	0x9D2A, 0x9D32, 0x9D39, 0x9D40, 0x9D48, 0x9D4F, 0x9D57, 0x9D5E, 0x9D66, 0x9D6D, 0x9D75,
+	0x9D7C, 0x9D83, 0x9D8B, 0x9D92, 0x9D9A, 0x9DA1, 0x9DA8, 0x9DB0, 0x9DB7, 0x9DBE, 0x9DC6,
+	0x9DCD, 0x9DD5, 0x9DDC, 0x9DE3, 0x9DEB, 0x9DF2, 0x9DF9, 0x9E01, 0x9E08, 0x9E0F, 0x9E17,
+	0x9E1E, 0x9E25, 0x9E2D, 0x9E34, 0x9E3B, 0x9E43, 0x9E4A, 0x9E51, 0x9E58, 0x9E60, 0x9E67,
+	0x9E6E, 0x9E75, 0x9E7D, 0x9E84, 0x9E8B, 0x9E92, 0x9E9A, 0x9EA1, 0x9EA8, 0x9EAF, 0x9EB7,
+	0x9EBE, 0x9EC5, 0x9ECC, 0x9ED4, 0x9EDB, 0x9EE2, 0x9EE9, 0x9EF0, 0x9EF7, 0x9EFF, 0x9F06,
+	0x9F0D, 0x9F14, 0x9F1B, 0x9F23, 0x9F2A, 0x9F31, 0x9F38, 0x9F3F, 0x9F46, 0x9F4D, 0x9F55,
+	0x9F5C, 0x9F63, 0x9F6A, 0x9F71, 0x9F78, 0x9F7F, 0x9F86, 0x9F8D, 0x9F95, 0x9F9C, 0x9FA3,
+	0x9FAA, 0x9FB1, 0x9FB8, 0x9FBF, 0x9FC6, 0x9FCD, 0x9FD4, 0x9FDB, 0x9FE2, 0x9FE9, 0x9FF0,
+	0x9FF7, 0x9FFF,
+};
+
+static void tegra_crtc_atomic_destroy_state(struct drm_crtc *crtc,
+					    struct drm_crtc_state *state);
+
+static void tegra_dc_stats_reset(struct tegra_dc_stats *stats)
+{
+	stats->frames = 0;
+	stats->vblank = 0;
+	stats->underflow = 0;
+	stats->overflow = 0;
+}
+
+/* Reads the active copy of a register. */
+static u32 tegra_dc_readl_active(struct tegra_dc *dc, unsigned long offset)
+{
+	u32 value;
+
+	tegra_dc_writel(dc, READ_MUX, DC_CMD_STATE_ACCESS);
+	value = tegra_dc_readl(dc, offset);
+	tegra_dc_writel(dc, 0, DC_CMD_STATE_ACCESS);
+
+	return value;
+}
+
+static inline unsigned int tegra_plane_offset(struct tegra_plane *plane,
+					      unsigned int offset)
+{
+	if (offset >= 0x500 && offset <= 0x638) {
+		offset = 0x000 + (offset - 0x500);
+		return plane->offset + offset;
+	}
+
+	if (offset >= 0x700 && offset <= 0x719) {
+		offset = 0x180 + (offset - 0x700);
+		return plane->offset + offset;
+	}
+
+	if (offset >= 0x800 && offset <= 0x839) {
+		offset = 0x1c0 + (offset - 0x800);
+		return plane->offset + offset;
+	}
+
+	dev_WARN(plane->dc->dev, "invalid offset: %x\n", offset);
+
+	return plane->offset + offset;
+}
+
+static inline u32 tegra_plane_readl(struct tegra_plane *plane,
+				    unsigned int offset)
+{
+	return tegra_dc_readl(plane->dc, tegra_plane_offset(plane, offset));
+}
+
+static inline void tegra_plane_writel(struct tegra_plane *plane, u32 value,
+				      unsigned int offset)
+{
+	tegra_dc_writel(plane->dc, value, tegra_plane_offset(plane, offset));
+}
+
+bool tegra_dc_has_output(struct tegra_dc *dc, struct device *dev)
+{
+	struct device_node *np = dc->dev->of_node;
+	struct of_phandle_iterator it;
+	int err;
+
+	of_for_each_phandle(&it, err, np, "nvidia,outputs", NULL, 0)
+		if (it.node == dev->of_node) {
+			of_node_put(it.node);
+			return true;
+		}
+
+	return false;
+}
+
+/*
+ * Double-buffered registers have two copies: ASSEMBLY and ACTIVE. When the
+ * *_ACT_REQ bits are set the ASSEMBLY copy is latched into the ACTIVE copy.
+ * Latching happens mmediately if the display controller is in STOP mode or
+ * on the next frame boundary otherwise.
+ *
+ * Triple-buffered registers have three copies: ASSEMBLY, ARM and ACTIVE. The
+ * ASSEMBLY copy is latched into the ARM copy immediately after *_UPDATE bits
+ * are written. When the *_ACT_REQ bits are written, the ARM copy is latched
+ * into the ACTIVE copy, either immediately if the display controller is in
+ * STOP mode, or at the next frame boundary otherwise.
+ */
+void tegra_dc_commit(struct tegra_dc *dc)
+{
+	tegra_dc_writel(dc, GENERAL_ACT_REQ << 8, DC_CMD_STATE_CONTROL);
+	tegra_dc_writel(dc, GENERAL_ACT_REQ, DC_CMD_STATE_CONTROL);
+}
+
+static inline u32 compute_dda_inc(unsigned int in, unsigned int out, bool v,
+				  unsigned int bpp)
+{
+	fixed20_12 outf = dfixed_init(out);
+	fixed20_12 inf = dfixed_init(in);
+	u32 dda_inc;
+	int max;
+
+	if (v)
+		max = 15;
+	else {
+		switch (bpp) {
+		case 2:
+			max = 8;
+			break;
+
+		default:
+			WARN_ON_ONCE(1);
+			fallthrough;
+		case 4:
+			max = 4;
+			break;
+		}
+	}
+
+	outf.full = max_t(u32, outf.full - dfixed_const(1), dfixed_const(1));
+	inf.full -= dfixed_const(1);
+
+	dda_inc = dfixed_div(inf, outf);
+	dda_inc = min_t(u32, dda_inc, dfixed_const(max));
+
+	return dda_inc;
+}
+
+static inline u32 compute_initial_dda(unsigned int in)
+{
+	fixed20_12 inf = dfixed_init(in);
+	return dfixed_frac(inf);
+}
+
+static void tegra_plane_setup_blending_legacy(struct tegra_plane *plane)
+{
+	u32 background[3] = {
+		BLEND_WEIGHT1(0) | BLEND_WEIGHT0(0) | BLEND_COLOR_KEY_NONE,
+		BLEND_WEIGHT1(0) | BLEND_WEIGHT0(0) | BLEND_COLOR_KEY_NONE,
+		BLEND_WEIGHT1(0) | BLEND_WEIGHT0(0) | BLEND_COLOR_KEY_NONE,
+	};
+	u32 foreground = BLEND_WEIGHT1(255) | BLEND_WEIGHT0(255) |
+			 BLEND_COLOR_KEY_NONE;
+	u32 blendnokey = BLEND_WEIGHT1(255) | BLEND_WEIGHT0(255);
+	struct tegra_plane_state *state;
+	u32 blending[2];
+	unsigned int i;
+
+	/* disable blending for non-overlapping case */
+	tegra_plane_writel(plane, blendnokey, DC_WIN_BLEND_NOKEY);
+	tegra_plane_writel(plane, foreground, DC_WIN_BLEND_1WIN);
+
+	state = to_tegra_plane_state(plane->base.state);
+
+	if (state->opaque) {
+		/*
+		 * Since custom fix-weight blending isn't utilized and weight
+		 * of top window is set to max, we can enforce dependent
+		 * blending which in this case results in transparent bottom
+		 * window if top window is opaque and if top window enables
+		 * alpha blending, then bottom window is getting alpha value
+		 * of 1 minus the sum of alpha components of the overlapping
+		 * plane.
+		 */
+		background[0] |= BLEND_CONTROL_DEPENDENT;
+		background[1] |= BLEND_CONTROL_DEPENDENT;
+
+		/*
+		 * The region where three windows overlap is the intersection
+		 * of the two regions where two windows overlap. It contributes
+		 * to the area if all of the windows on top of it have an alpha
+		 * component.
+		 */
+		switch (state->base.normalized_zpos) {
+		case 0:
+			if (state->blending[0].alpha &&
+			    state->blending[1].alpha)
+				background[2] |= BLEND_CONTROL_DEPENDENT;
+			break;
+
+		case 1:
+			background[2] |= BLEND_CONTROL_DEPENDENT;
+			break;
+		}
+	} else {
+		/*
+		 * Enable alpha blending if pixel format has an alpha
+		 * component.
+		 */
+		foreground |= BLEND_CONTROL_ALPHA;
+
+		/*
+		 * If any of the windows on top of this window is opaque, it
+		 * will completely conceal this window within that area. If
+		 * top window has an alpha component, it is blended over the
+		 * bottom window.
+		 */
+		for (i = 0; i < 2; i++) {
+			if (state->blending[i].alpha &&
+			    state->blending[i].top)
+				background[i] |= BLEND_CONTROL_DEPENDENT;
+		}
+
+		switch (state->base.normalized_zpos) {
+		case 0:
+			if (state->blending[0].alpha &&
+			    state->blending[1].alpha)
+				background[2] |= BLEND_CONTROL_DEPENDENT;
+			break;
+
+		case 1:
+			/*
+			 * When both middle and topmost windows have an alpha,
+			 * these windows a mixed together and then the result
+			 * is blended over the bottom window.
+			 */
+			if (state->blending[0].alpha &&
+			    state->blending[0].top)
+				background[2] |= BLEND_CONTROL_ALPHA;
+
+			if (state->blending[1].alpha &&
+			    state->blending[1].top)
+				background[2] |= BLEND_CONTROL_ALPHA;
+			break;
+		}
+	}
+
+	switch (state->base.normalized_zpos) {
+	case 0:
+		tegra_plane_writel(plane, background[0], DC_WIN_BLEND_2WIN_X);
+		tegra_plane_writel(plane, background[1], DC_WIN_BLEND_2WIN_Y);
+		tegra_plane_writel(plane, background[2], DC_WIN_BLEND_3WIN_XY);
+		break;
+
+	case 1:
+		/*
+		 * If window B / C is topmost, then X / Y registers are
+		 * matching the order of blending[...] state indices,
+		 * otherwise a swap is required.
+		 */
+		if (!state->blending[0].top && state->blending[1].top) {
+			blending[0] = foreground;
+			blending[1] = background[1];
+		} else {
+			blending[0] = background[0];
+			blending[1] = foreground;
+		}
+
+		tegra_plane_writel(plane, blending[0], DC_WIN_BLEND_2WIN_X);
+		tegra_plane_writel(plane, blending[1], DC_WIN_BLEND_2WIN_Y);
+		tegra_plane_writel(plane, background[2], DC_WIN_BLEND_3WIN_XY);
+		break;
+
+	case 2:
+		tegra_plane_writel(plane, foreground, DC_WIN_BLEND_2WIN_X);
+		tegra_plane_writel(plane, foreground, DC_WIN_BLEND_2WIN_Y);
+		tegra_plane_writel(plane, foreground, DC_WIN_BLEND_3WIN_XY);
+		break;
+	}
+}
+
+static void tegra_plane_setup_blending(struct tegra_plane *plane,
+				       const struct tegra_dc_window *window)
+{
+	u32 value;
+
+	value = BLEND_FACTOR_DST_ALPHA_ZERO | BLEND_FACTOR_SRC_ALPHA_K2 |
+		BLEND_FACTOR_DST_COLOR_NEG_K1_TIMES_SRC |
+		BLEND_FACTOR_SRC_COLOR_K1_TIMES_SRC;
+	tegra_plane_writel(plane, value, DC_WIN_BLEND_MATCH_SELECT);
+
+	value = BLEND_FACTOR_DST_ALPHA_ZERO | BLEND_FACTOR_SRC_ALPHA_K2 |
+		BLEND_FACTOR_DST_COLOR_NEG_K1_TIMES_SRC |
+		BLEND_FACTOR_SRC_COLOR_K1_TIMES_SRC;
+	tegra_plane_writel(plane, value, DC_WIN_BLEND_NOMATCH_SELECT);
+
+	value = K2(255) | K1(255) | WINDOW_LAYER_DEPTH(255 - window->zpos);
+	tegra_plane_writel(plane, value, DC_WIN_BLEND_LAYER_CONTROL);
+}
+
+static bool
+tegra_plane_use_horizontal_filtering(struct tegra_plane *plane,
+				     const struct tegra_dc_window *window)
+{
+	struct tegra_dc *dc = plane->dc;
+
+	if (window->src.w == window->dst.w)
+		return false;
+
+	if (plane->index == 0 && dc->soc->has_win_a_without_filters)
+		return false;
+
+	return true;
+}
+
+static bool
+tegra_plane_use_vertical_filtering(struct tegra_plane *plane,
+				   const struct tegra_dc_window *window)
+{
+	struct tegra_dc *dc = plane->dc;
+
+	if (window->src.h == window->dst.h)
+		return false;
+
+	if (plane->index == 0 && dc->soc->has_win_a_without_filters)
+		return false;
+
+	if (plane->index == 2 && dc->soc->has_win_c_without_vert_filter)
+		return false;
+
+	return true;
+}
+
+static void tegra_dc_setup_window(struct tegra_plane *plane,
+				  const struct tegra_dc_window *window)
+{
+	unsigned h_offset, v_offset, h_size, v_size, h_dda, v_dda, bpp;
+	struct tegra_dc *dc = plane->dc;
+	unsigned int planes;
+	u32 value;
+	bool yuv;
+
+	/*
+	 * For YUV planar modes, the number of bytes per pixel takes into
+	 * account only the luma component and therefore is 1.
+	 */
+	yuv = tegra_plane_format_is_yuv(window->format, &planes, NULL);
+	if (!yuv)
+		bpp = window->bits_per_pixel / 8;
+	else
+		bpp = (planes > 1) ? 1 : 2;
+
+	tegra_plane_writel(plane, window->format, DC_WIN_COLOR_DEPTH);
+	tegra_plane_writel(plane, window->swap, DC_WIN_BYTE_SWAP);
+
+	value = V_POSITION(window->dst.y) | H_POSITION(window->dst.x);
+	tegra_plane_writel(plane, value, DC_WIN_POSITION);
+
+	value = V_SIZE(window->dst.h) | H_SIZE(window->dst.w);
+	tegra_plane_writel(plane, value, DC_WIN_SIZE);
+
+	h_offset = window->src.x * bpp;
+	v_offset = window->src.y;
+	h_size = window->src.w * bpp;
+	v_size = window->src.h;
+
+	if (window->reflect_x)
+		h_offset += (window->src.w - 1) * bpp;
+
+	if (window->reflect_y)
+		v_offset += window->src.h - 1;
+
+	value = V_PRESCALED_SIZE(v_size) | H_PRESCALED_SIZE(h_size);
+	tegra_plane_writel(plane, value, DC_WIN_PRESCALED_SIZE);
+
+	/*
+	 * For DDA computations the number of bytes per pixel for YUV planar
+	 * modes needs to take into account all Y, U and V components.
+	 */
+	if (yuv && planes > 1)
+		bpp = 2;
+
+	h_dda = compute_dda_inc(window->src.w, window->dst.w, false, bpp);
+	v_dda = compute_dda_inc(window->src.h, window->dst.h, true, bpp);
+
+	value = V_DDA_INC(v_dda) | H_DDA_INC(h_dda);
+	tegra_plane_writel(plane, value, DC_WIN_DDA_INC);
+
+	h_dda = compute_initial_dda(window->src.x);
+	v_dda = compute_initial_dda(window->src.y);
+
+	tegra_plane_writel(plane, h_dda, DC_WIN_H_INITIAL_DDA);
+	tegra_plane_writel(plane, v_dda, DC_WIN_V_INITIAL_DDA);
+
+	tegra_plane_writel(plane, 0, DC_WIN_UV_BUF_STRIDE);
+	tegra_plane_writel(plane, 0, DC_WIN_BUF_STRIDE);
+
+	tegra_plane_writel(plane, window->base[0], DC_WINBUF_START_ADDR);
+
+	if (yuv && planes > 1) {
+		tegra_plane_writel(plane, window->base[1], DC_WINBUF_START_ADDR_U);
+
+		if (planes > 2)
+			tegra_plane_writel(plane, window->base[2], DC_WINBUF_START_ADDR_V);
+
+		value = window->stride[1] << 16 | window->stride[0];
+		tegra_plane_writel(plane, value, DC_WIN_LINE_STRIDE);
+	} else {
+		tegra_plane_writel(plane, window->stride[0], DC_WIN_LINE_STRIDE);
+	}
+
+	tegra_plane_writel(plane, h_offset, DC_WINBUF_ADDR_H_OFFSET);
+	tegra_plane_writel(plane, v_offset, DC_WINBUF_ADDR_V_OFFSET);
+
+	if (dc->soc->supports_block_linear) {
+		unsigned long height = window->tiling.value;
+
+		switch (window->tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WINBUF_SURFACE_KIND_PITCH;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WINBUF_SURFACE_KIND_TILED;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			value = DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(height) |
+				DC_WINBUF_SURFACE_KIND_BLOCK;
+			break;
+		}
+
+		tegra_plane_writel(plane, value, DC_WINBUF_SURFACE_KIND);
+	} else {
+		switch (window->tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV |
+				DC_WIN_BUFFER_ADDR_MODE_LINEAR;
+			break;
+
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WIN_BUFFER_ADDR_MODE_TILE_UV |
+				DC_WIN_BUFFER_ADDR_MODE_TILE;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			/*
+			 * No need to handle this here because ->atomic_check
+			 * will already have filtered it out.
+			 */
+			break;
+		}
+
+		tegra_plane_writel(plane, value, DC_WIN_BUFFER_ADDR_MODE);
+	}
+
+	value = WIN_ENABLE;
+
+	if (yuv) {
+		/* setup default colorspace conversion coefficients */
+		tegra_plane_writel(plane, 0x00f0, DC_WIN_CSC_YOF);
+		tegra_plane_writel(plane, 0x012a, DC_WIN_CSC_KYRGB);
+		tegra_plane_writel(plane, 0x0000, DC_WIN_CSC_KUR);
+		tegra_plane_writel(plane, 0x0198, DC_WIN_CSC_KVR);
+		tegra_plane_writel(plane, 0x039b, DC_WIN_CSC_KUG);
+		tegra_plane_writel(plane, 0x032f, DC_WIN_CSC_KVG);
+		tegra_plane_writel(plane, 0x0204, DC_WIN_CSC_KUB);
+		tegra_plane_writel(plane, 0x0000, DC_WIN_CSC_KVB);
+
+		value |= CSC_ENABLE;
+	} else if (window->bits_per_pixel < 24) {
+		value |= COLOR_EXPAND;
+	}
+
+	if (window->reflect_x)
+		value |= H_DIRECTION;
+
+	if (window->reflect_y)
+		value |= V_DIRECTION;
+
+	if (tegra_plane_use_horizontal_filtering(plane, window)) {
+		/*
+		 * Enable horizontal 6-tap filter and set filtering
+		 * coefficients to the default values defined in TRM.
+		 */
+		tegra_plane_writel(plane, 0x00008000, DC_WIN_H_FILTER_P(0));
+		tegra_plane_writel(plane, 0x3e087ce1, DC_WIN_H_FILTER_P(1));
+		tegra_plane_writel(plane, 0x3b117ac1, DC_WIN_H_FILTER_P(2));
+		tegra_plane_writel(plane, 0x591b73aa, DC_WIN_H_FILTER_P(3));
+		tegra_plane_writel(plane, 0x57256d9a, DC_WIN_H_FILTER_P(4));
+		tegra_plane_writel(plane, 0x552f668b, DC_WIN_H_FILTER_P(5));
+		tegra_plane_writel(plane, 0x73385e8b, DC_WIN_H_FILTER_P(6));
+		tegra_plane_writel(plane, 0x72435583, DC_WIN_H_FILTER_P(7));
+		tegra_plane_writel(plane, 0x714c4c8b, DC_WIN_H_FILTER_P(8));
+		tegra_plane_writel(plane, 0x70554393, DC_WIN_H_FILTER_P(9));
+		tegra_plane_writel(plane, 0x715e389b, DC_WIN_H_FILTER_P(10));
+		tegra_plane_writel(plane, 0x71662faa, DC_WIN_H_FILTER_P(11));
+		tegra_plane_writel(plane, 0x536d25ba, DC_WIN_H_FILTER_P(12));
+		tegra_plane_writel(plane, 0x55731bca, DC_WIN_H_FILTER_P(13));
+		tegra_plane_writel(plane, 0x387a11d9, DC_WIN_H_FILTER_P(14));
+		tegra_plane_writel(plane, 0x3c7c08f1, DC_WIN_H_FILTER_P(15));
+
+		value |= H_FILTER;
+	}
+
+	if (tegra_plane_use_vertical_filtering(plane, window)) {
+		unsigned int i, k;
+
+		/*
+		 * Enable vertical 2-tap filter and set filtering
+		 * coefficients to the default values defined in TRM.
+		 */
+		for (i = 0, k = 128; i < 16; i++, k -= 8)
+			tegra_plane_writel(plane, k, DC_WIN_V_FILTER_P(i));
+
+		value |= V_FILTER;
+	}
+
+	tegra_plane_writel(plane, value, DC_WIN_WIN_OPTIONS);
+
+	if (dc->soc->has_legacy_blending)
+		tegra_plane_setup_blending_legacy(plane);
+	else
+		tegra_plane_setup_blending(plane, window);
+}
+
+static const u32 tegra20_primary_formats[] = {
+	DRM_FORMAT_ARGB4444,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_RGBA5551,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_ARGB8888,
+	/* non-native formats */
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGBX5551,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_XRGB8888,
+};
+
+static const u64 tegra20_modifiers[] = {
+	DRM_FORMAT_MOD_LINEAR,
+	DRM_FORMAT_MOD_NVIDIA_TEGRA_TILED,
+	DRM_FORMAT_MOD_INVALID
+};
+
+static const u32 tegra114_primary_formats[] = {
+	DRM_FORMAT_ARGB4444,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_RGBA5551,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_ARGB8888,
+	/* new on Tegra114 */
+	DRM_FORMAT_ABGR4444,
+	DRM_FORMAT_ABGR1555,
+	DRM_FORMAT_BGRA5551,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGBX5551,
+	DRM_FORMAT_XBGR1555,
+	DRM_FORMAT_BGRX5551,
+	DRM_FORMAT_BGR565,
+	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_RGBA8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+};
+
+static const u32 tegra124_primary_formats[] = {
+	DRM_FORMAT_ARGB4444,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_RGBA5551,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_ARGB8888,
+	/* new on Tegra114 */
+	DRM_FORMAT_ABGR4444,
+	DRM_FORMAT_ABGR1555,
+	DRM_FORMAT_BGRA5551,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGBX5551,
+	DRM_FORMAT_XBGR1555,
+	DRM_FORMAT_BGRX5551,
+	DRM_FORMAT_BGR565,
+	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_RGBA8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	/* new on Tegra124 */
+	DRM_FORMAT_RGBX8888,
+	DRM_FORMAT_BGRX8888,
+};
+
+static const u64 tegra124_modifiers[] = {
+	DRM_FORMAT_MOD_LINEAR,
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(0),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(1),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(2),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(3),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(4),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(5),
+	DRM_FORMAT_MOD_INVALID
+};
+
+static int tegra_plane_atomic_check(struct drm_plane *plane,
+				    struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
+										 plane);
+	struct tegra_plane_state *plane_state = to_tegra_plane_state(new_plane_state);
+	unsigned int supported_rotation = DRM_MODE_ROTATE_0 |
+					  DRM_MODE_REFLECT_X |
+					  DRM_MODE_REFLECT_Y;
+	unsigned int rotation = new_plane_state->rotation;
+	struct tegra_bo_tiling *tiling = &plane_state->tiling;
+	struct tegra_plane *tegra = to_tegra_plane(plane);
+	struct tegra_dc *dc = to_tegra_dc(new_plane_state->crtc);
+	int err;
+
+	plane_state->peak_memory_bandwidth = 0;
+	plane_state->avg_memory_bandwidth = 0;
+
+	/* no need for further checks if the plane is being disabled */
+	if (!new_plane_state->crtc) {
+		plane_state->total_peak_memory_bandwidth = 0;
+		return 0;
+	}
+
+	err = tegra_plane_format(new_plane_state->fb->format->format,
+				 &plane_state->format,
+				 &plane_state->swap);
+	if (err < 0)
+		return err;
+
+	/*
+	 * Tegra20 and Tegra30 are special cases here because they support
+	 * only variants of specific formats with an alpha component, but not
+	 * the corresponding opaque formats. However, the opaque formats can
+	 * be emulated by disabling alpha blending for the plane.
+	 */
+	if (dc->soc->has_legacy_blending) {
+		err = tegra_plane_setup_legacy_state(tegra, plane_state);
+		if (err < 0)
+			return err;
+	}
+
+	err = tegra_fb_get_tiling(new_plane_state->fb, tiling);
+	if (err < 0)
+		return err;
+
+	if (tiling->mode == TEGRA_BO_TILING_MODE_BLOCK &&
+	    !dc->soc->supports_block_linear) {
+		DRM_ERROR("hardware doesn't support block linear mode\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * Older userspace used custom BO flag in order to specify the Y
+	 * reflection, while modern userspace uses the generic DRM rotation
+	 * property in order to achieve the same result.  The legacy BO flag
+	 * duplicates the DRM rotation property when both are set.
+	 */
+	if (tegra_fb_is_bottom_up(new_plane_state->fb))
+		rotation |= DRM_MODE_REFLECT_Y;
+
+	rotation = drm_rotation_simplify(rotation, supported_rotation);
+
+	if (rotation & DRM_MODE_REFLECT_X)
+		plane_state->reflect_x = true;
+	else
+		plane_state->reflect_x = false;
+
+	if (rotation & DRM_MODE_REFLECT_Y)
+		plane_state->reflect_y = true;
+	else
+		plane_state->reflect_y = false;
+
+	/*
+	 * Tegra doesn't support different strides for U and V planes so we
+	 * error out if the user tries to display a framebuffer with such a
+	 * configuration.
+	 */
+	if (new_plane_state->fb->format->num_planes > 2) {
+		if (new_plane_state->fb->pitches[2] != new_plane_state->fb->pitches[1]) {
+			DRM_ERROR("unsupported UV-plane configuration\n");
+			return -EINVAL;
+		}
+	}
+
+	err = tegra_plane_state_add(tegra, new_plane_state);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static void tegra_plane_atomic_disable(struct drm_plane *plane,
+				       struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *old_state = drm_atomic_get_old_plane_state(state,
+									   plane);
+	struct tegra_plane *p = to_tegra_plane(plane);
+	u32 value;
+
+	/* rien ne va plus */
+	if (!old_state || !old_state->crtc)
+		return;
+
+	value = tegra_plane_readl(p, DC_WIN_WIN_OPTIONS);
+	value &= ~WIN_ENABLE;
+	tegra_plane_writel(p, value, DC_WIN_WIN_OPTIONS);
+}
+
+static void tegra_plane_atomic_update(struct drm_plane *plane,
+				      struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state,
+									   plane);
+	struct tegra_plane_state *tegra_plane_state = to_tegra_plane_state(new_state);
+	struct drm_framebuffer *fb = new_state->fb;
+	struct tegra_plane *p = to_tegra_plane(plane);
+	struct tegra_dc_window window;
+	unsigned int i;
+
+	/* rien ne va plus */
+	if (!new_state->crtc || !new_state->fb)
+		return;
+
+	if (!new_state->visible)
+		return tegra_plane_atomic_disable(plane, state);
+
+	memset(&window, 0, sizeof(window));
+	window.src.x = new_state->src.x1 >> 16;
+	window.src.y = new_state->src.y1 >> 16;
+	window.src.w = drm_rect_width(&new_state->src) >> 16;
+	window.src.h = drm_rect_height(&new_state->src) >> 16;
+	window.dst.x = new_state->dst.x1;
+	window.dst.y = new_state->dst.y1;
+	window.dst.w = drm_rect_width(&new_state->dst);
+	window.dst.h = drm_rect_height(&new_state->dst);
+	window.bits_per_pixel = fb->format->cpp[0] * 8;
+	window.reflect_x = tegra_plane_state->reflect_x;
+	window.reflect_y = tegra_plane_state->reflect_y;
+
+	/* copy from state */
+	window.zpos = new_state->normalized_zpos;
+	window.tiling = tegra_plane_state->tiling;
+	window.format = tegra_plane_state->format;
+	window.swap = tegra_plane_state->swap;
+
+	for (i = 0; i < fb->format->num_planes; i++) {
+		window.base[i] = tegra_plane_state->iova[i] + fb->offsets[i];
+
+		/*
+		 * Tegra uses a shared stride for UV planes. Framebuffers are
+		 * already checked for this in the tegra_plane_atomic_check()
+		 * function, so it's safe to ignore the V-plane pitch here.
+		 */
+		if (i < 2)
+			window.stride[i] = fb->pitches[i];
+	}
+
+	tegra_dc_setup_window(p, &window);
+}
+
+static const struct drm_plane_helper_funcs tegra_plane_helper_funcs = {
+	.prepare_fb = tegra_plane_prepare_fb,
+	.cleanup_fb = tegra_plane_cleanup_fb,
+	.atomic_check = tegra_plane_atomic_check,
+	.atomic_disable = tegra_plane_atomic_disable,
+	.atomic_update = tegra_plane_atomic_update,
+};
+
+static unsigned long tegra_plane_get_possible_crtcs(struct drm_device *drm)
+{
+	/*
+	 * Ideally this would use drm_crtc_mask(), but that would require the
+	 * CRTC to already be in the mode_config's list of CRTCs. However, it
+	 * will only be added to that list in the drm_crtc_init_with_planes()
+	 * (in tegra_dc_init()), which in turn requires registration of these
+	 * planes. So we have ourselves a nice little chicken and egg problem
+	 * here.
+	 *
+	 * We work around this by manually creating the mask from the number
+	 * of CRTCs that have been registered, and should therefore always be
+	 * the same as drm_crtc_index() after registration.
+	 */
+	return 1 << drm->mode_config.num_crtc;
+}
+
+static struct drm_plane *tegra_primary_plane_create(struct drm_device *drm,
+						    struct tegra_dc *dc)
+{
+	unsigned long possible_crtcs = tegra_plane_get_possible_crtcs(drm);
+	enum drm_plane_type type = DRM_PLANE_TYPE_PRIMARY;
+	struct tegra_plane *plane;
+	unsigned int num_formats;
+	const u64 *modifiers;
+	const u32 *formats;
+	int err;
+
+	plane = kzalloc_obj(*plane);
+	if (!plane)
+		return ERR_PTR(-ENOMEM);
+
+	/* Always use window A as primary window */
+	plane->offset = 0xa00;
+	plane->index = 0;
+	plane->dc = dc;
+
+	num_formats = dc->soc->num_primary_formats;
+	formats = dc->soc->primary_formats;
+	modifiers = dc->soc->modifiers;
+
+	err = tegra_plane_interconnect_init(plane);
+	if (err) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	err = drm_universal_plane_init(drm, &plane->base, possible_crtcs,
+				       &tegra_plane_funcs, formats,
+				       num_formats, modifiers, type, NULL);
+	if (err < 0) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	drm_plane_helper_add(&plane->base, &tegra_plane_helper_funcs);
+	drm_plane_create_zpos_property(&plane->base, plane->index, 0, 255);
+
+	err = drm_plane_create_rotation_property(&plane->base,
+						 DRM_MODE_ROTATE_0,
+						 DRM_MODE_ROTATE_0 |
+						 DRM_MODE_ROTATE_180 |
+						 DRM_MODE_REFLECT_X |
+						 DRM_MODE_REFLECT_Y);
+	if (err < 0)
+		dev_err(dc->dev, "failed to create rotation property: %d\n",
+			err);
+
+	return &plane->base;
+}
+
+static const u32 tegra_legacy_cursor_plane_formats[] = {
+	DRM_FORMAT_RGBA8888,
+};
+
+static const u32 tegra_cursor_plane_formats[] = {
+	DRM_FORMAT_ARGB8888,
+};
+
+static int tegra_cursor_atomic_check(struct drm_plane *plane,
+				     struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
+										 plane);
+	struct tegra_plane_state *plane_state = to_tegra_plane_state(new_plane_state);
+	struct tegra_plane *tegra = to_tegra_plane(plane);
+	int err;
+
+	plane_state->peak_memory_bandwidth = 0;
+	plane_state->avg_memory_bandwidth = 0;
+
+	/* no need for further checks if the plane is being disabled */
+	if (!new_plane_state->crtc) {
+		plane_state->total_peak_memory_bandwidth = 0;
+		return 0;
+	}
+
+	/* scaling not supported for cursor */
+	if ((new_plane_state->src_w >> 16 != new_plane_state->crtc_w) ||
+	    (new_plane_state->src_h >> 16 != new_plane_state->crtc_h))
+		return -EINVAL;
+
+	/* only square cursors supported */
+	if (new_plane_state->src_w != new_plane_state->src_h)
+		return -EINVAL;
+
+	if (new_plane_state->crtc_w != 32 && new_plane_state->crtc_w != 64 &&
+	    new_plane_state->crtc_w != 128 && new_plane_state->crtc_w != 256)
+		return -EINVAL;
+
+	err = tegra_plane_state_add(tegra, new_plane_state);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static void __tegra_cursor_atomic_update(struct drm_plane *plane,
+					 struct drm_plane_state *new_state)
+{
+	struct tegra_plane_state *tegra_plane_state = to_tegra_plane_state(new_state);
+	struct tegra_dc *dc = to_tegra_dc(new_state->crtc);
+	struct tegra_drm *tegra = plane->dev->dev_private;
+#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
+	u64 dma_mask = *dc->dev->dma_mask;
+#endif
+	unsigned int x, y;
+	u32 value = 0;
+
+	/* rien ne va plus */
+	if (!new_state->crtc || !new_state->fb)
+		return;
+
+	/*
+	 * Legacy display supports hardware clipping of the cursor, but
+	 * nvdisplay relies on software to clip the cursor to the screen.
+	 */
+	if (!dc->soc->has_nvdisplay)
+		value |= CURSOR_CLIP_DISPLAY;
+
+	switch (new_state->crtc_w) {
+	case 32:
+		value |= CURSOR_SIZE_32x32;
+		break;
+
+	case 64:
+		value |= CURSOR_SIZE_64x64;
+		break;
+
+	case 128:
+		value |= CURSOR_SIZE_128x128;
+		break;
+
+	case 256:
+		value |= CURSOR_SIZE_256x256;
+		break;
+
+	default:
+		WARN(1, "cursor size %ux%u not supported\n",
+		     new_state->crtc_w, new_state->crtc_h);
+		return;
+	}
+
+	value |= (tegra_plane_state->iova[0] >> 10) & 0x3fffff;
+	tegra_dc_writel(dc, value, DC_DISP_CURSOR_START_ADDR);
+
+#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
+	value = (tegra_plane_state->iova[0] >> 32) & (dma_mask >> 32);
+	tegra_dc_writel(dc, value, DC_DISP_CURSOR_START_ADDR_HI);
+#endif
+
+	/* enable cursor and set blend mode */
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+	value |= CURSOR_ENABLE;
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+	value = tegra_dc_readl(dc, DC_DISP_BLEND_CURSOR_CONTROL);
+	value &= ~CURSOR_DST_BLEND_MASK;
+	value &= ~CURSOR_SRC_BLEND_MASK;
+
+	if (dc->soc->has_nvdisplay)
+		value &= ~CURSOR_COMPOSITION_MODE_XOR;
+	else
+		value |= CURSOR_MODE_NORMAL;
+
+	value |= CURSOR_DST_BLEND_NEG_K1_TIMES_SRC;
+	value |= CURSOR_SRC_BLEND_K1_TIMES_SRC;
+	value |= CURSOR_ALPHA;
+	tegra_dc_writel(dc, value, DC_DISP_BLEND_CURSOR_CONTROL);
+
+	/* nvdisplay relies on software for clipping */
+	if (dc->soc->has_nvdisplay) {
+		struct drm_rect src;
+
+		x = new_state->dst.x1;
+		y = new_state->dst.y1;
+
+		drm_rect_fp_to_int(&src, &new_state->src);
+
+		value = (src.y1 & tegra->vmask) << 16 | (src.x1 & tegra->hmask);
+		tegra_dc_writel(dc, value, DC_DISP_PCALC_HEAD_SET_CROPPED_POINT_IN_CURSOR);
+
+		value = (drm_rect_height(&src) & tegra->vmask) << 16 |
+			(drm_rect_width(&src) & tegra->hmask);
+		tegra_dc_writel(dc, value, DC_DISP_PCALC_HEAD_SET_CROPPED_SIZE_IN_CURSOR);
+	} else {
+		x = new_state->crtc_x;
+		y = new_state->crtc_y;
+	}
+
+	/* position the cursor */
+	value = ((y & tegra->vmask) << 16) | (x & tegra->hmask);
+	tegra_dc_writel(dc, value, DC_DISP_CURSOR_POSITION);
+}
+
+static void tegra_cursor_atomic_update(struct drm_plane *plane,
+				       struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state, plane);
+
+	__tegra_cursor_atomic_update(plane, new_state);
+}
+
+static void tegra_cursor_atomic_disable(struct drm_plane *plane,
+					struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *old_state = drm_atomic_get_old_plane_state(state,
+									   plane);
+	struct tegra_dc *dc;
+	u32 value;
+
+	/* rien ne va plus */
+	if (!old_state || !old_state->crtc)
+		return;
+
+	dc = to_tegra_dc(old_state->crtc);
+
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+	value &= ~CURSOR_ENABLE;
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+}
+
+static int tegra_cursor_atomic_async_check(struct drm_plane *plane, struct drm_atomic_commit *state,
+					   bool flip)
+{
+	struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state, plane);
+	struct drm_crtc_state *crtc_state;
+	int min_scale, max_scale;
+	int err;
+
+	crtc_state = drm_atomic_get_new_crtc_state(state, new_state->crtc);
+	if (WARN_ON(!crtc_state))
+		return -EINVAL;
+
+	if (!crtc_state->active)
+		return -EINVAL;
+
+	if (plane->state->crtc != new_state->crtc ||
+	    plane->state->src_w != new_state->src_w ||
+	    plane->state->src_h != new_state->src_h ||
+	    plane->state->crtc_w != new_state->crtc_w ||
+	    plane->state->crtc_h != new_state->crtc_h ||
+	    plane->state->fb != new_state->fb ||
+	    plane->state->fb == NULL)
+		return -EINVAL;
+
+	min_scale = (1 << 16) / 8;
+	max_scale = (8 << 16) / 1;
+
+	err = drm_atomic_helper_check_plane_state(new_state, crtc_state, min_scale, max_scale,
+						  true, true);
+	if (err < 0)
+		return err;
+
+	if (new_state->visible != plane->state->visible)
+		return -EINVAL;
+
+	return 0;
+}
+
+static void tegra_cursor_atomic_async_update(struct drm_plane *plane,
+					     struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state, plane);
+	struct tegra_dc *dc = to_tegra_dc(new_state->crtc);
+
+	plane->state->src_x = new_state->src_x;
+	plane->state->src_y = new_state->src_y;
+	plane->state->crtc_x = new_state->crtc_x;
+	plane->state->crtc_y = new_state->crtc_y;
+
+	if (new_state->visible) {
+		struct tegra_plane *p = to_tegra_plane(plane);
+		u32 value;
+
+		__tegra_cursor_atomic_update(plane, new_state);
+
+		value = (WIN_A_ACT_REQ << p->index) << 8 | GENERAL_UPDATE;
+		tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
+		(void)tegra_dc_readl(dc, DC_CMD_STATE_CONTROL);
+
+		value = (WIN_A_ACT_REQ << p->index) | GENERAL_ACT_REQ;
+		tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
+		(void)tegra_dc_readl(dc, DC_CMD_STATE_CONTROL);
+	}
+}
+
+static const struct drm_plane_helper_funcs tegra_cursor_plane_helper_funcs = {
+	.prepare_fb = tegra_plane_prepare_fb,
+	.cleanup_fb = tegra_plane_cleanup_fb,
+	.atomic_check = tegra_cursor_atomic_check,
+	.atomic_update = tegra_cursor_atomic_update,
+	.atomic_disable = tegra_cursor_atomic_disable,
+	.atomic_async_check = tegra_cursor_atomic_async_check,
+	.atomic_async_update = tegra_cursor_atomic_async_update,
+};
+
+static const uint64_t linear_modifiers[] = {
+	DRM_FORMAT_MOD_LINEAR,
+	DRM_FORMAT_MOD_INVALID
+};
+
+static struct drm_plane *tegra_dc_cursor_plane_create(struct drm_device *drm,
+						      struct tegra_dc *dc)
+{
+	unsigned long possible_crtcs = tegra_plane_get_possible_crtcs(drm);
+	struct tegra_plane *plane;
+	unsigned int num_formats;
+	const u32 *formats;
+	int err;
+
+	plane = kzalloc_obj(*plane);
+	if (!plane)
+		return ERR_PTR(-ENOMEM);
+
+	/*
+	 * This index is kind of fake. The cursor isn't a regular plane, but
+	 * its update and activation request bits in DC_CMD_STATE_CONTROL do
+	 * use the same programming. Setting this fake index here allows the
+	 * code in tegra_add_plane_state() to do the right thing without the
+	 * need to special-casing the cursor plane.
+	 */
+	plane->index = 6;
+	plane->dc = dc;
+
+	if (!dc->soc->has_nvdisplay) {
+		num_formats = ARRAY_SIZE(tegra_legacy_cursor_plane_formats);
+		formats = tegra_legacy_cursor_plane_formats;
+
+		err = tegra_plane_interconnect_init(plane);
+		if (err) {
+			kfree(plane);
+			return ERR_PTR(err);
+		}
+	} else {
+		num_formats = ARRAY_SIZE(tegra_cursor_plane_formats);
+		formats = tegra_cursor_plane_formats;
+	}
+
+	err = drm_universal_plane_init(drm, &plane->base, possible_crtcs,
+				       &tegra_plane_funcs, formats,
+				       num_formats, linear_modifiers,
+				       DRM_PLANE_TYPE_CURSOR, NULL);
+	if (err < 0) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	drm_plane_helper_add(&plane->base, &tegra_cursor_plane_helper_funcs);
+	drm_plane_create_zpos_immutable_property(&plane->base, 255);
+
+	return &plane->base;
+}
+
+static const u32 tegra20_overlay_formats[] = {
+	DRM_FORMAT_ARGB4444,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_RGBA5551,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_ARGB8888,
+	/* non-native formats */
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGBX5551,
+	DRM_FORMAT_XBGR8888,
+	DRM_FORMAT_XRGB8888,
+	/* planar formats */
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YUV420,
+	DRM_FORMAT_YUV422,
+};
+
+static const u32 tegra114_overlay_formats[] = {
+	DRM_FORMAT_ARGB4444,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_RGBA5551,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_ARGB8888,
+	/* new on Tegra114 */
+	DRM_FORMAT_ABGR4444,
+	DRM_FORMAT_ABGR1555,
+	DRM_FORMAT_BGRA5551,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGBX5551,
+	DRM_FORMAT_XBGR1555,
+	DRM_FORMAT_BGRX5551,
+	DRM_FORMAT_BGR565,
+	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_RGBA8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	/* planar formats */
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YUV420,
+	DRM_FORMAT_YUV422,
+	/* semi-planar formats */
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_NV21,
+	DRM_FORMAT_NV16,
+	DRM_FORMAT_NV61,
+	DRM_FORMAT_NV24,
+	DRM_FORMAT_NV42,
+};
+
+static const u32 tegra124_overlay_formats[] = {
+	DRM_FORMAT_ARGB4444,
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_RGBA5551,
+	DRM_FORMAT_ABGR8888,
+	DRM_FORMAT_ARGB8888,
+	/* new on Tegra114 */
+	DRM_FORMAT_ABGR4444,
+	DRM_FORMAT_ABGR1555,
+	DRM_FORMAT_BGRA5551,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGBX5551,
+	DRM_FORMAT_XBGR1555,
+	DRM_FORMAT_BGRX5551,
+	DRM_FORMAT_BGR565,
+	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_RGBA8888,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	/* new on Tegra124 */
+	DRM_FORMAT_RGBX8888,
+	DRM_FORMAT_BGRX8888,
+	/* planar formats */
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YVYU,
+	DRM_FORMAT_VYUY,
+	DRM_FORMAT_YUV420, /* YU12 */
+	DRM_FORMAT_YUV422, /* YU16 */
+	DRM_FORMAT_YUV444, /* YU24 */
+	/* semi-planar formats */
+	DRM_FORMAT_NV12,
+	DRM_FORMAT_NV21,
+	DRM_FORMAT_NV16,
+	DRM_FORMAT_NV61,
+	DRM_FORMAT_NV24,
+	DRM_FORMAT_NV42,
+};
+
+static struct drm_plane *tegra_dc_overlay_plane_create(struct drm_device *drm,
+						       struct tegra_dc *dc,
+						       unsigned int index,
+						       bool cursor)
+{
+	unsigned long possible_crtcs = tegra_plane_get_possible_crtcs(drm);
+	struct tegra_plane *plane;
+	unsigned int num_formats;
+	enum drm_plane_type type;
+	const u32 *formats;
+	int err;
+
+	plane = kzalloc_obj(*plane);
+	if (!plane)
+		return ERR_PTR(-ENOMEM);
+
+	plane->offset = 0xa00 + 0x200 * index;
+	plane->index = index;
+	plane->dc = dc;
+
+	num_formats = dc->soc->num_overlay_formats;
+	formats = dc->soc->overlay_formats;
+
+	err = tegra_plane_interconnect_init(plane);
+	if (err) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	if (!cursor)
+		type = DRM_PLANE_TYPE_OVERLAY;
+	else
+		type = DRM_PLANE_TYPE_CURSOR;
+
+	err = drm_universal_plane_init(drm, &plane->base, possible_crtcs,
+				       &tegra_plane_funcs, formats,
+				       num_formats, linear_modifiers,
+				       type, NULL);
+	if (err < 0) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	drm_plane_helper_add(&plane->base, &tegra_plane_helper_funcs);
+	drm_plane_create_zpos_property(&plane->base, plane->index, 0, 255);
+
+	err = drm_plane_create_rotation_property(&plane->base,
+						 DRM_MODE_ROTATE_0,
+						 DRM_MODE_ROTATE_0 |
+						 DRM_MODE_ROTATE_180 |
+						 DRM_MODE_REFLECT_X |
+						 DRM_MODE_REFLECT_Y);
+	if (err < 0)
+		dev_err(dc->dev, "failed to create rotation property: %d\n",
+			err);
+
+	return &plane->base;
+}
+
+static struct drm_plane *tegra_dc_add_shared_planes(struct drm_device *drm,
+						    struct tegra_dc *dc)
+{
+	struct drm_plane *plane, *primary = NULL;
+	unsigned int i, j;
+
+	for (i = 0; i < dc->soc->num_wgrps; i++) {
+		const struct tegra_windowgroup_soc *wgrp = &dc->soc->wgrps[i];
+
+		if (wgrp->dc == dc->pipe) {
+			for (j = 0; j < wgrp->num_windows; j++) {
+				unsigned int index = wgrp->windows[j];
+				enum drm_plane_type type;
+
+				if (primary)
+					type = DRM_PLANE_TYPE_OVERLAY;
+				else
+					type = DRM_PLANE_TYPE_PRIMARY;
+
+				plane = tegra_shared_plane_create(drm, dc,
+								  wgrp->index,
+								  index, type);
+				if (IS_ERR(plane))
+					return plane;
+
+				/*
+				 * Choose the first shared plane owned by this
+				 * head as the primary plane.
+				 */
+				if (!primary)
+					primary = plane;
+			}
+		}
+	}
+
+	return primary;
+}
+
+static struct drm_plane *tegra_dc_add_planes(struct drm_device *drm,
+					     struct tegra_dc *dc)
+{
+	struct drm_plane *planes[2], *primary;
+	unsigned int planes_num;
+	unsigned int i;
+	int err;
+
+	primary = tegra_primary_plane_create(drm, dc);
+	if (IS_ERR(primary))
+		return primary;
+
+	if (dc->soc->supports_cursor)
+		planes_num = 2;
+	else
+		planes_num = 1;
+
+	for (i = 0; i < planes_num; i++) {
+		planes[i] = tegra_dc_overlay_plane_create(drm, dc, 1 + i,
+							  false);
+		if (IS_ERR(planes[i])) {
+			err = PTR_ERR(planes[i]);
+
+			while (i--)
+				planes[i]->funcs->destroy(planes[i]);
+
+			primary->funcs->destroy(primary);
+			return ERR_PTR(err);
+		}
+	}
+
+	return primary;
+}
+
+static void tegra_dc_destroy(struct drm_crtc *crtc)
+{
+	drm_crtc_cleanup(crtc);
+}
+
+static void tegra_crtc_reset(struct drm_crtc *crtc)
+{
+	struct tegra_dc_state *state = kzalloc_obj(*state);
+
+	if (crtc->state)
+		tegra_crtc_atomic_destroy_state(crtc, crtc->state);
+
+	if (state)
+		__drm_atomic_helper_crtc_reset(crtc, &state->base);
+	else
+		__drm_atomic_helper_crtc_reset(crtc, NULL);
+}
+
+static struct drm_crtc_state *
+tegra_crtc_atomic_duplicate_state(struct drm_crtc *crtc)
+{
+	struct tegra_dc_state *state = to_dc_state(crtc->state);
+	struct tegra_dc_state *copy;
+
+	copy = kmalloc_obj(*copy);
+	if (!copy)
+		return NULL;
+
+	__drm_atomic_helper_crtc_duplicate_state(crtc, &copy->base);
+	copy->clk = state->clk;
+	copy->pclk = state->pclk;
+	copy->div = state->div;
+	copy->planes = state->planes;
+
+	return &copy->base;
+}
+
+static void tegra_crtc_atomic_destroy_state(struct drm_crtc *crtc,
+					    struct drm_crtc_state *state)
+{
+	__drm_atomic_helper_crtc_destroy_state(state);
+	kfree(state);
+}
+
+#define DEBUGFS_REG32(_name) { .name = #_name, .offset = _name }
+
+static const struct debugfs_reg32 tegra_dc_regs[] = {
+	DEBUGFS_REG32(DC_CMD_GENERAL_INCR_SYNCPT),
+	DEBUGFS_REG32(DC_CMD_GENERAL_INCR_SYNCPT_CNTRL),
+	DEBUGFS_REG32(DC_CMD_GENERAL_INCR_SYNCPT_ERROR),
+	DEBUGFS_REG32(DC_CMD_WIN_A_INCR_SYNCPT),
+	DEBUGFS_REG32(DC_CMD_WIN_A_INCR_SYNCPT_CNTRL),
+	DEBUGFS_REG32(DC_CMD_WIN_A_INCR_SYNCPT_ERROR),
+	DEBUGFS_REG32(DC_CMD_WIN_B_INCR_SYNCPT),
+	DEBUGFS_REG32(DC_CMD_WIN_B_INCR_SYNCPT_CNTRL),
+	DEBUGFS_REG32(DC_CMD_WIN_B_INCR_SYNCPT_ERROR),
+	DEBUGFS_REG32(DC_CMD_WIN_C_INCR_SYNCPT),
+	DEBUGFS_REG32(DC_CMD_WIN_C_INCR_SYNCPT_CNTRL),
+	DEBUGFS_REG32(DC_CMD_WIN_C_INCR_SYNCPT_ERROR),
+	DEBUGFS_REG32(DC_CMD_CONT_SYNCPT_VSYNC),
+	DEBUGFS_REG32(DC_CMD_DISPLAY_COMMAND_OPTION0),
+	DEBUGFS_REG32(DC_CMD_DISPLAY_COMMAND),
+	DEBUGFS_REG32(DC_CMD_SIGNAL_RAISE),
+	DEBUGFS_REG32(DC_CMD_DISPLAY_POWER_CONTROL),
+	DEBUGFS_REG32(DC_CMD_INT_STATUS),
+	DEBUGFS_REG32(DC_CMD_INT_MASK),
+	DEBUGFS_REG32(DC_CMD_INT_ENABLE),
+	DEBUGFS_REG32(DC_CMD_INT_TYPE),
+	DEBUGFS_REG32(DC_CMD_INT_POLARITY),
+	DEBUGFS_REG32(DC_CMD_SIGNAL_RAISE1),
+	DEBUGFS_REG32(DC_CMD_SIGNAL_RAISE2),
+	DEBUGFS_REG32(DC_CMD_SIGNAL_RAISE3),
+	DEBUGFS_REG32(DC_CMD_STATE_ACCESS),
+	DEBUGFS_REG32(DC_CMD_STATE_CONTROL),
+	DEBUGFS_REG32(DC_CMD_DISPLAY_WINDOW_HEADER),
+	DEBUGFS_REG32(DC_CMD_REG_ACT_CONTROL),
+	DEBUGFS_REG32(DC_COM_CRC_CONTROL),
+	DEBUGFS_REG32(DC_COM_CRC_CHECKSUM),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_ENABLE(0)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_ENABLE(1)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_ENABLE(2)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_ENABLE(3)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_POLARITY(0)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_POLARITY(1)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_POLARITY(2)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_POLARITY(3)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_DATA(0)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_DATA(1)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_DATA(2)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_DATA(3)),
+	DEBUGFS_REG32(DC_COM_PIN_INPUT_ENABLE(0)),
+	DEBUGFS_REG32(DC_COM_PIN_INPUT_ENABLE(1)),
+	DEBUGFS_REG32(DC_COM_PIN_INPUT_ENABLE(2)),
+	DEBUGFS_REG32(DC_COM_PIN_INPUT_ENABLE(3)),
+	DEBUGFS_REG32(DC_COM_PIN_INPUT_DATA(0)),
+	DEBUGFS_REG32(DC_COM_PIN_INPUT_DATA(1)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_SELECT(0)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_SELECT(1)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_SELECT(2)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_SELECT(3)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_SELECT(4)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_SELECT(5)),
+	DEBUGFS_REG32(DC_COM_PIN_OUTPUT_SELECT(6)),
+	DEBUGFS_REG32(DC_COM_PIN_MISC_CONTROL),
+	DEBUGFS_REG32(DC_COM_PIN_PM0_CONTROL),
+	DEBUGFS_REG32(DC_COM_PIN_PM0_DUTY_CYCLE),
+	DEBUGFS_REG32(DC_COM_PIN_PM1_CONTROL),
+	DEBUGFS_REG32(DC_COM_PIN_PM1_DUTY_CYCLE),
+	DEBUGFS_REG32(DC_COM_SPI_CONTROL),
+	DEBUGFS_REG32(DC_COM_SPI_START_BYTE),
+	DEBUGFS_REG32(DC_COM_HSPI_WRITE_DATA_AB),
+	DEBUGFS_REG32(DC_COM_HSPI_WRITE_DATA_CD),
+	DEBUGFS_REG32(DC_COM_HSPI_CS_DC),
+	DEBUGFS_REG32(DC_COM_SCRATCH_REGISTER_A),
+	DEBUGFS_REG32(DC_COM_SCRATCH_REGISTER_B),
+	DEBUGFS_REG32(DC_COM_GPIO_CTRL),
+	DEBUGFS_REG32(DC_COM_GPIO_DEBOUNCE_COUNTER),
+	DEBUGFS_REG32(DC_COM_CRC_CHECKSUM_LATCHED),
+	DEBUGFS_REG32(DC_DISP_DISP_SIGNAL_OPTIONS0),
+	DEBUGFS_REG32(DC_DISP_DISP_SIGNAL_OPTIONS1),
+	DEBUGFS_REG32(DC_DISP_DISP_WIN_OPTIONS),
+	DEBUGFS_REG32(DC_DISP_DISP_MEM_HIGH_PRIORITY),
+	DEBUGFS_REG32(DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER),
+	DEBUGFS_REG32(DC_DISP_DISP_TIMING_OPTIONS),
+	DEBUGFS_REG32(DC_DISP_REF_TO_SYNC),
+	DEBUGFS_REG32(DC_DISP_SYNC_WIDTH),
+	DEBUGFS_REG32(DC_DISP_BACK_PORCH),
+	DEBUGFS_REG32(DC_DISP_ACTIVE),
+	DEBUGFS_REG32(DC_DISP_FRONT_PORCH),
+	DEBUGFS_REG32(DC_DISP_H_PULSE0_CONTROL),
+	DEBUGFS_REG32(DC_DISP_H_PULSE0_POSITION_A),
+	DEBUGFS_REG32(DC_DISP_H_PULSE0_POSITION_B),
+	DEBUGFS_REG32(DC_DISP_H_PULSE0_POSITION_C),
+	DEBUGFS_REG32(DC_DISP_H_PULSE0_POSITION_D),
+	DEBUGFS_REG32(DC_DISP_H_PULSE1_CONTROL),
+	DEBUGFS_REG32(DC_DISP_H_PULSE1_POSITION_A),
+	DEBUGFS_REG32(DC_DISP_H_PULSE1_POSITION_B),
+	DEBUGFS_REG32(DC_DISP_H_PULSE1_POSITION_C),
+	DEBUGFS_REG32(DC_DISP_H_PULSE1_POSITION_D),
+	DEBUGFS_REG32(DC_DISP_H_PULSE2_CONTROL),
+	DEBUGFS_REG32(DC_DISP_H_PULSE2_POSITION_A),
+	DEBUGFS_REG32(DC_DISP_H_PULSE2_POSITION_B),
+	DEBUGFS_REG32(DC_DISP_H_PULSE2_POSITION_C),
+	DEBUGFS_REG32(DC_DISP_H_PULSE2_POSITION_D),
+	DEBUGFS_REG32(DC_DISP_V_PULSE0_CONTROL),
+	DEBUGFS_REG32(DC_DISP_V_PULSE0_POSITION_A),
+	DEBUGFS_REG32(DC_DISP_V_PULSE0_POSITION_B),
+	DEBUGFS_REG32(DC_DISP_V_PULSE0_POSITION_C),
+	DEBUGFS_REG32(DC_DISP_V_PULSE1_CONTROL),
+	DEBUGFS_REG32(DC_DISP_V_PULSE1_POSITION_A),
+	DEBUGFS_REG32(DC_DISP_V_PULSE1_POSITION_B),
+	DEBUGFS_REG32(DC_DISP_V_PULSE1_POSITION_C),
+	DEBUGFS_REG32(DC_DISP_V_PULSE2_CONTROL),
+	DEBUGFS_REG32(DC_DISP_V_PULSE2_POSITION_A),
+	DEBUGFS_REG32(DC_DISP_V_PULSE3_CONTROL),
+	DEBUGFS_REG32(DC_DISP_V_PULSE3_POSITION_A),
+	DEBUGFS_REG32(DC_DISP_M0_CONTROL),
+	DEBUGFS_REG32(DC_DISP_M1_CONTROL),
+	DEBUGFS_REG32(DC_DISP_DI_CONTROL),
+	DEBUGFS_REG32(DC_DISP_PP_CONTROL),
+	DEBUGFS_REG32(DC_DISP_PP_SELECT_A),
+	DEBUGFS_REG32(DC_DISP_PP_SELECT_B),
+	DEBUGFS_REG32(DC_DISP_PP_SELECT_C),
+	DEBUGFS_REG32(DC_DISP_PP_SELECT_D),
+	DEBUGFS_REG32(DC_DISP_DISP_CLOCK_CONTROL),
+	DEBUGFS_REG32(DC_DISP_DISP_INTERFACE_CONTROL),
+	DEBUGFS_REG32(DC_DISP_DISP_COLOR_CONTROL),
+	DEBUGFS_REG32(DC_DISP_SHIFT_CLOCK_OPTIONS),
+	DEBUGFS_REG32(DC_DISP_DATA_ENABLE_OPTIONS),
+	DEBUGFS_REG32(DC_DISP_SERIAL_INTERFACE_OPTIONS),
+	DEBUGFS_REG32(DC_DISP_LCD_SPI_OPTIONS),
+	DEBUGFS_REG32(DC_DISP_BORDER_COLOR),
+	DEBUGFS_REG32(DC_DISP_COLOR_KEY0_LOWER),
+	DEBUGFS_REG32(DC_DISP_COLOR_KEY0_UPPER),
+	DEBUGFS_REG32(DC_DISP_COLOR_KEY1_LOWER),
+	DEBUGFS_REG32(DC_DISP_COLOR_KEY1_UPPER),
+	DEBUGFS_REG32(DC_DISP_CURSOR_FOREGROUND),
+	DEBUGFS_REG32(DC_DISP_CURSOR_BACKGROUND),
+	DEBUGFS_REG32(DC_DISP_CURSOR_START_ADDR),
+	DEBUGFS_REG32(DC_DISP_CURSOR_START_ADDR_NS),
+	DEBUGFS_REG32(DC_DISP_CURSOR_POSITION),
+	DEBUGFS_REG32(DC_DISP_CURSOR_POSITION_NS),
+	DEBUGFS_REG32(DC_DISP_INIT_SEQ_CONTROL),
+	DEBUGFS_REG32(DC_DISP_SPI_INIT_SEQ_DATA_A),
+	DEBUGFS_REG32(DC_DISP_SPI_INIT_SEQ_DATA_B),
+	DEBUGFS_REG32(DC_DISP_SPI_INIT_SEQ_DATA_C),
+	DEBUGFS_REG32(DC_DISP_SPI_INIT_SEQ_DATA_D),
+	DEBUGFS_REG32(DC_DISP_DC_MCCIF_FIFOCTRL),
+	DEBUGFS_REG32(DC_DISP_MCCIF_DISPLAY0A_HYST),
+	DEBUGFS_REG32(DC_DISP_MCCIF_DISPLAY0B_HYST),
+	DEBUGFS_REG32(DC_DISP_MCCIF_DISPLAY1A_HYST),
+	DEBUGFS_REG32(DC_DISP_MCCIF_DISPLAY1B_HYST),
+	DEBUGFS_REG32(DC_DISP_DAC_CRT_CTRL),
+	DEBUGFS_REG32(DC_DISP_DISP_MISC_CONTROL),
+	DEBUGFS_REG32(DC_DISP_SD_CONTROL),
+	DEBUGFS_REG32(DC_DISP_SD_CSC_COEFF),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(0)),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(1)),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(2)),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(3)),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(4)),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(5)),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(6)),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(7)),
+	DEBUGFS_REG32(DC_DISP_SD_LUT(8)),
+	DEBUGFS_REG32(DC_DISP_SD_FLICKER_CONTROL),
+	DEBUGFS_REG32(DC_DISP_DC_PIXEL_COUNT),
+	DEBUGFS_REG32(DC_DISP_SD_HISTOGRAM(0)),
+	DEBUGFS_REG32(DC_DISP_SD_HISTOGRAM(1)),
+	DEBUGFS_REG32(DC_DISP_SD_HISTOGRAM(2)),
+	DEBUGFS_REG32(DC_DISP_SD_HISTOGRAM(3)),
+	DEBUGFS_REG32(DC_DISP_SD_HISTOGRAM(4)),
+	DEBUGFS_REG32(DC_DISP_SD_HISTOGRAM(5)),
+	DEBUGFS_REG32(DC_DISP_SD_HISTOGRAM(6)),
+	DEBUGFS_REG32(DC_DISP_SD_HISTOGRAM(7)),
+	DEBUGFS_REG32(DC_DISP_SD_BL_TF(0)),
+	DEBUGFS_REG32(DC_DISP_SD_BL_TF(1)),
+	DEBUGFS_REG32(DC_DISP_SD_BL_TF(2)),
+	DEBUGFS_REG32(DC_DISP_SD_BL_TF(3)),
+	DEBUGFS_REG32(DC_DISP_SD_BL_CONTROL),
+	DEBUGFS_REG32(DC_DISP_SD_HW_K_VALUES),
+	DEBUGFS_REG32(DC_DISP_SD_MAN_K_VALUES),
+	DEBUGFS_REG32(DC_DISP_CURSOR_START_ADDR_HI),
+	DEBUGFS_REG32(DC_DISP_BLEND_CURSOR_CONTROL),
+	DEBUGFS_REG32(DC_WIN_WIN_OPTIONS),
+	DEBUGFS_REG32(DC_WIN_BYTE_SWAP),
+	DEBUGFS_REG32(DC_WIN_BUFFER_CONTROL),
+	DEBUGFS_REG32(DC_WIN_COLOR_DEPTH),
+	DEBUGFS_REG32(DC_WIN_POSITION),
+	DEBUGFS_REG32(DC_WIN_SIZE),
+	DEBUGFS_REG32(DC_WIN_PRESCALED_SIZE),
+	DEBUGFS_REG32(DC_WIN_H_INITIAL_DDA),
+	DEBUGFS_REG32(DC_WIN_V_INITIAL_DDA),
+	DEBUGFS_REG32(DC_WIN_DDA_INC),
+	DEBUGFS_REG32(DC_WIN_LINE_STRIDE),
+	DEBUGFS_REG32(DC_WIN_BUF_STRIDE),
+	DEBUGFS_REG32(DC_WIN_UV_BUF_STRIDE),
+	DEBUGFS_REG32(DC_WIN_BUFFER_ADDR_MODE),
+	DEBUGFS_REG32(DC_WIN_DV_CONTROL),
+	DEBUGFS_REG32(DC_WIN_BLEND_NOKEY),
+	DEBUGFS_REG32(DC_WIN_BLEND_1WIN),
+	DEBUGFS_REG32(DC_WIN_BLEND_2WIN_X),
+	DEBUGFS_REG32(DC_WIN_BLEND_2WIN_Y),
+	DEBUGFS_REG32(DC_WIN_BLEND_3WIN_XY),
+	DEBUGFS_REG32(DC_WIN_HP_FETCH_CONTROL),
+	DEBUGFS_REG32(DC_WINBUF_START_ADDR),
+	DEBUGFS_REG32(DC_WINBUF_START_ADDR_NS),
+	DEBUGFS_REG32(DC_WINBUF_START_ADDR_U),
+	DEBUGFS_REG32(DC_WINBUF_START_ADDR_U_NS),
+	DEBUGFS_REG32(DC_WINBUF_START_ADDR_V),
+	DEBUGFS_REG32(DC_WINBUF_START_ADDR_V_NS),
+	DEBUGFS_REG32(DC_WINBUF_ADDR_H_OFFSET),
+	DEBUGFS_REG32(DC_WINBUF_ADDR_H_OFFSET_NS),
+	DEBUGFS_REG32(DC_WINBUF_ADDR_V_OFFSET),
+	DEBUGFS_REG32(DC_WINBUF_ADDR_V_OFFSET_NS),
+	DEBUGFS_REG32(DC_WINBUF_UFLOW_STATUS),
+	DEBUGFS_REG32(DC_WINBUF_AD_UFLOW_STATUS),
+	DEBUGFS_REG32(DC_WINBUF_BD_UFLOW_STATUS),
+	DEBUGFS_REG32(DC_WINBUF_CD_UFLOW_STATUS),
+};
+
+static int tegra_dc_show_regs(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = s->private;
+	struct tegra_dc *dc = node->info_ent->data;
+	unsigned int i;
+	int err = 0;
+
+	drm_modeset_lock(&dc->base.mutex, NULL);
+
+	if (!dc->base.state->active) {
+		err = -EBUSY;
+		goto unlock;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(tegra_dc_regs); i++) {
+		unsigned int offset = tegra_dc_regs[i].offset;
+
+		seq_printf(s, "%-40s %#05x %08x\n", tegra_dc_regs[i].name,
+			   offset, tegra_dc_readl(dc, offset));
+	}
+
+unlock:
+	drm_modeset_unlock(&dc->base.mutex);
+	return err;
+}
+
+static int tegra_dc_show_crc(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = s->private;
+	struct tegra_dc *dc = node->info_ent->data;
+	int err = 0;
+	u32 value;
+
+	drm_modeset_lock(&dc->base.mutex, NULL);
+
+	if (!dc->base.state->active) {
+		err = -EBUSY;
+		goto unlock;
+	}
+
+	value = DC_COM_CRC_CONTROL_ACTIVE_DATA | DC_COM_CRC_CONTROL_ENABLE;
+	tegra_dc_writel(dc, value, DC_COM_CRC_CONTROL);
+	tegra_dc_commit(dc);
+
+	drm_crtc_wait_one_vblank(&dc->base);
+	drm_crtc_wait_one_vblank(&dc->base);
+
+	value = tegra_dc_readl(dc, DC_COM_CRC_CHECKSUM);
+	seq_printf(s, "%08x\n", value);
+
+	tegra_dc_writel(dc, 0, DC_COM_CRC_CONTROL);
+
+unlock:
+	drm_modeset_unlock(&dc->base.mutex);
+	return err;
+}
+
+static int tegra_dc_show_stats(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = s->private;
+	struct tegra_dc *dc = node->info_ent->data;
+
+	seq_printf(s, "frames: %lu\n", dc->stats.frames);
+	seq_printf(s, "vblank: %lu\n", dc->stats.vblank);
+	seq_printf(s, "underflow: %lu\n", dc->stats.underflow);
+	seq_printf(s, "overflow: %lu\n", dc->stats.overflow);
+
+	seq_printf(s, "frames total: %lu\n", dc->stats.frames_total);
+	seq_printf(s, "vblank total: %lu\n", dc->stats.vblank_total);
+	seq_printf(s, "underflow total: %lu\n", dc->stats.underflow_total);
+	seq_printf(s, "overflow total: %lu\n", dc->stats.overflow_total);
+
+	return 0;
+}
+
+static struct drm_info_list debugfs_files[] = {
+	{ "regs", tegra_dc_show_regs, 0, NULL },
+	{ "crc", tegra_dc_show_crc, 0, NULL },
+	{ "stats", tegra_dc_show_stats, 0, NULL },
+};
+
+static int tegra_dc_late_register(struct drm_crtc *crtc)
+{
+	unsigned int i, count = ARRAY_SIZE(debugfs_files);
+	struct drm_minor *minor = crtc->dev->primary;
+	struct dentry *root;
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+
+#ifdef CONFIG_DEBUG_FS
+	root = crtc->debugfs_entry;
+#else
+	root = NULL;
+#endif
+
+	dc->debugfs_files = kmemdup(debugfs_files, sizeof(debugfs_files),
+				    GFP_KERNEL);
+	if (!dc->debugfs_files)
+		return -ENOMEM;
+
+	for (i = 0; i < count; i++)
+		dc->debugfs_files[i].data = dc;
+
+	drm_debugfs_create_files(dc->debugfs_files, count, root, minor);
+
+	return 0;
+}
+
+static void tegra_dc_early_unregister(struct drm_crtc *crtc)
+{
+	unsigned int count = ARRAY_SIZE(debugfs_files);
+	struct drm_minor *minor = crtc->dev->primary;
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	struct dentry *root;
+
+#ifdef CONFIG_DEBUG_FS
+	root = crtc->debugfs_entry;
+#else
+	root = NULL;
+#endif
+
+	drm_debugfs_remove_files(dc->debugfs_files, count, root, minor);
+	kfree(dc->debugfs_files);
+	dc->debugfs_files = NULL;
+}
+
+static u32 tegra_dc_get_vblank_counter(struct drm_crtc *crtc)
+{
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+
+	/* XXX vblank syncpoints don't work with nvdisplay yet */
+	if (dc->syncpt && !dc->soc->has_nvdisplay)
+		return host1x_syncpt_read(dc->syncpt);
+
+	/* fallback to software emulated VBLANK counter */
+	return (u32)drm_crtc_vblank_count(&dc->base);
+}
+
+static int tegra_dc_enable_vblank(struct drm_crtc *crtc)
+{
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	u32 value;
+
+	value = tegra_dc_readl(dc, DC_CMD_INT_MASK);
+	value |= VBLANK_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
+
+	return 0;
+}
+
+static void tegra_dc_disable_vblank(struct drm_crtc *crtc)
+{
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	u32 value;
+
+	value = tegra_dc_readl(dc, DC_CMD_INT_MASK);
+	value &= ~VBLANK_INT;
+	tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
+}
+
+static const struct drm_crtc_funcs tegra_crtc_funcs = {
+	.page_flip = drm_atomic_helper_page_flip,
+	.set_config = drm_atomic_helper_set_config,
+	.destroy = tegra_dc_destroy,
+	.reset = tegra_crtc_reset,
+	.atomic_duplicate_state = tegra_crtc_atomic_duplicate_state,
+	.atomic_destroy_state = tegra_crtc_atomic_destroy_state,
+	.late_register = tegra_dc_late_register,
+	.early_unregister = tegra_dc_early_unregister,
+	.get_vblank_counter = tegra_dc_get_vblank_counter,
+	.enable_vblank = tegra_dc_enable_vblank,
+	.disable_vblank = tegra_dc_disable_vblank,
+};
+
+static int tegra_dc_set_timings(struct tegra_dc *dc,
+				struct drm_display_mode *mode)
+{
+	unsigned int h_ref_to_sync = 1;
+	unsigned int v_ref_to_sync = 1;
+	unsigned long value;
+
+	if (!dc->soc->has_nvdisplay) {
+		tegra_dc_writel(dc, 0x0, DC_DISP_DISP_TIMING_OPTIONS);
+
+		value = (v_ref_to_sync << 16) | h_ref_to_sync;
+		tegra_dc_writel(dc, value, DC_DISP_REF_TO_SYNC);
+	}
+
+	value = ((mode->vsync_end - mode->vsync_start) << 16) |
+		((mode->hsync_end - mode->hsync_start) <<  0);
+	tegra_dc_writel(dc, value, DC_DISP_SYNC_WIDTH);
+
+	value = ((mode->vtotal - mode->vsync_end) << 16) |
+		((mode->htotal - mode->hsync_end) <<  0);
+	tegra_dc_writel(dc, value, DC_DISP_BACK_PORCH);
+
+	value = ((mode->vsync_start - mode->vdisplay) << 16) |
+		((mode->hsync_start - mode->hdisplay) <<  0);
+	tegra_dc_writel(dc, value, DC_DISP_FRONT_PORCH);
+
+	value = (mode->vdisplay << 16) | mode->hdisplay;
+	tegra_dc_writel(dc, value, DC_DISP_ACTIVE);
+
+	return 0;
+}
+
+/**
+ * tegra_dc_state_setup_clock - check clock settings and store them in atomic
+ *     state
+ * @dc: display controller
+ * @crtc_state: CRTC atomic state
+ * @clk: parent clock for display controller
+ * @pclk: pixel clock
+ * @div: shift clock divider
+ *
+ * Returns:
+ * 0 on success or a negative error-code on failure.
+ */
+int tegra_dc_state_setup_clock(struct tegra_dc *dc,
+			       struct drm_crtc_state *crtc_state,
+			       struct clk *clk, unsigned long pclk,
+			       unsigned int div)
+{
+	struct tegra_dc_state *state = to_dc_state(crtc_state);
+
+	if (!clk_has_parent(dc->clk, clk))
+		return -EINVAL;
+
+	state->clk = clk;
+	state->pclk = pclk;
+	state->div = div;
+
+	return 0;
+}
+
+static void tegra_dc_update_voltage_state(struct tegra_dc *dc,
+					  struct tegra_dc_state *state)
+{
+	unsigned long rate, pstate;
+	struct dev_pm_opp *opp;
+	int err;
+
+	if (!dc->has_opp_table)
+		return;
+
+	/* calculate actual pixel clock rate which depends on internal divider */
+	rate = DIV_ROUND_UP(clk_get_rate(dc->clk) * 2, state->div + 2);
+
+	/* find suitable OPP for the rate */
+	opp = dev_pm_opp_find_freq_ceil(dc->dev, &rate);
+
+	/*
+	 * Very high resolution modes may results in a clock rate that is
+	 * above the characterized maximum. In this case it's okay to fall
+	 * back to the characterized maximum.
+	 */
+	if (opp == ERR_PTR(-ERANGE))
+		opp = dev_pm_opp_find_freq_floor(dc->dev, &rate);
+
+	if (IS_ERR(opp)) {
+		dev_err(dc->dev, "failed to find OPP for %luHz: %pe\n",
+			rate, opp);
+		return;
+	}
+
+	pstate = dev_pm_opp_get_required_pstate(opp, 0);
+	dev_pm_opp_put(opp);
+
+	/*
+	 * The minimum core voltage depends on the pixel clock rate (which
+	 * depends on internal clock divider of the CRTC) and not on the
+	 * rate of the display controller clock. This is why we're not using
+	 * dev_pm_opp_set_rate() API and instead controlling the power domain
+	 * directly.
+	 */
+	err = dev_pm_genpd_set_performance_state(dc->dev, pstate);
+	if (err)
+		dev_err(dc->dev, "failed to set power domain state to %lu: %d\n",
+			pstate, err);
+}
+
+static void tegra_dc_set_clock_rate(struct tegra_dc *dc,
+				    struct tegra_dc_state *state)
+{
+	int err;
+
+	err = clk_set_parent(dc->clk, state->clk);
+	if (err < 0)
+		dev_err(dc->dev, "failed to set parent clock: %d\n", err);
+
+	/*
+	 * Outputs may not want to change the parent clock rate. This is only
+	 * relevant to Tegra20 where only a single display PLL is available.
+	 * Since that PLL would typically be used for HDMI, an internal LVDS
+	 * panel would need to be driven by some other clock such as PLL_P
+	 * which is shared with other peripherals. Changing the clock rate
+	 * should therefore be avoided.
+	 */
+	if (state->pclk > 0) {
+		err = clk_set_rate(state->clk, state->pclk);
+		if (err < 0)
+			dev_err(dc->dev,
+				"failed to set clock rate to %lu Hz\n",
+				state->pclk);
+
+		err = clk_set_rate(dc->clk, state->pclk);
+		if (err < 0)
+			dev_err(dc->dev, "failed to set clock %pC to %lu Hz: %d\n",
+				dc->clk, state->pclk, err);
+	}
+
+	DRM_DEBUG_KMS("rate: %lu, div: %u\n", clk_get_rate(dc->clk),
+		      state->div);
+	DRM_DEBUG_KMS("pclk: %lu\n", state->pclk);
+
+	tegra_dc_update_voltage_state(dc, state);
+}
+
+static void tegra_dc_stop(struct tegra_dc *dc)
+{
+	u32 value;
+
+	/* stop the display controller */
+	value = tegra_dc_readl(dc, DC_CMD_DISPLAY_COMMAND);
+	value &= ~DISP_CTRL_MODE_MASK;
+	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_COMMAND);
+
+	tegra_dc_commit(dc);
+}
+
+static bool tegra_dc_idle(struct tegra_dc *dc)
+{
+	u32 value;
+
+	value = tegra_dc_readl_active(dc, DC_CMD_DISPLAY_COMMAND);
+
+	return (value & DISP_CTRL_MODE_MASK) == 0;
+}
+
+static int tegra_dc_wait_idle(struct tegra_dc *dc, unsigned long timeout)
+{
+	timeout = jiffies + msecs_to_jiffies(timeout);
+
+	while (time_before(jiffies, timeout)) {
+		if (tegra_dc_idle(dc))
+			return 0;
+
+		usleep_range(1000, 2000);
+	}
+
+	dev_dbg(dc->dev, "timeout waiting for DC to become idle\n");
+	return -ETIMEDOUT;
+}
+
+static void
+tegra_crtc_update_memory_bandwidth(struct drm_crtc *crtc,
+				   struct drm_atomic_commit *state,
+				   bool prepare_bandwidth_transition)
+{
+	const struct tegra_plane_state *old_tegra_state, *new_tegra_state;
+	u32 i, new_avg_bw, old_avg_bw, new_peak_bw, old_peak_bw;
+	const struct drm_plane_state *old_plane_state;
+	const struct drm_crtc_state *old_crtc_state;
+	struct tegra_dc_window window, old_window;
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	struct tegra_plane *tegra;
+	struct drm_plane *plane;
+
+	if (dc->soc->has_nvdisplay)
+		return;
+
+	old_crtc_state = drm_atomic_get_old_crtc_state(state, crtc);
+
+	if (!crtc->state->active) {
+		if (!old_crtc_state->active)
+			return;
+
+		/*
+		 * When CRTC is disabled on DPMS, the state of attached planes
+		 * is kept unchanged. Hence we need to enforce removal of the
+		 * bandwidths from the ICC paths.
+		 */
+		drm_atomic_crtc_for_each_plane(plane, crtc) {
+			tegra = to_tegra_plane(plane);
+
+			icc_set_bw(tegra->icc_mem, 0, 0);
+			icc_set_bw(tegra->icc_mem_vfilter, 0, 0);
+		}
+
+		return;
+	}
+
+	for_each_old_plane_in_state(old_crtc_state->state, plane,
+				    old_plane_state, i) {
+		old_tegra_state = to_const_tegra_plane_state(old_plane_state);
+		new_tegra_state = to_const_tegra_plane_state(plane->state);
+		tegra = to_tegra_plane(plane);
+
+		/*
+		 * We're iterating over the global atomic state and it contains
+		 * planes from another CRTC, hence we need to filter out the
+		 * planes unrelated to this CRTC.
+		 */
+		if (tegra->dc != dc)
+			continue;
+
+		new_avg_bw = new_tegra_state->avg_memory_bandwidth;
+		old_avg_bw = old_tegra_state->avg_memory_bandwidth;
+
+		new_peak_bw = new_tegra_state->total_peak_memory_bandwidth;
+		old_peak_bw = old_tegra_state->total_peak_memory_bandwidth;
+
+		/*
+		 * See the comment related to !crtc->state->active above,
+		 * which explains why bandwidths need to be updated when
+		 * CRTC is turning ON.
+		 */
+		if (new_avg_bw == old_avg_bw && new_peak_bw == old_peak_bw &&
+		    old_crtc_state->active)
+			continue;
+
+		window.src.h = drm_rect_height(&plane->state->src) >> 16;
+		window.dst.h = drm_rect_height(&plane->state->dst);
+
+		old_window.src.h = drm_rect_height(&old_plane_state->src) >> 16;
+		old_window.dst.h = drm_rect_height(&old_plane_state->dst);
+
+		/*
+		 * During the preparation phase (atomic_begin), the memory
+		 * freq should go high before the DC changes are committed
+		 * if bandwidth requirement goes up, otherwise memory freq
+		 * should to stay high if BW requirement goes down.  The
+		 * opposite applies to the completion phase (post_commit).
+		 */
+		if (prepare_bandwidth_transition) {
+			new_avg_bw = max(old_avg_bw, new_avg_bw);
+			new_peak_bw = max(old_peak_bw, new_peak_bw);
+
+			if (tegra_plane_use_vertical_filtering(tegra, &old_window))
+				window = old_window;
+		}
+
+		icc_set_bw(tegra->icc_mem, new_avg_bw, new_peak_bw);
+
+		if (tegra_plane_use_vertical_filtering(tegra, &window))
+			icc_set_bw(tegra->icc_mem_vfilter, new_avg_bw, new_peak_bw);
+		else
+			icc_set_bw(tegra->icc_mem_vfilter, 0, 0);
+	}
+}
+
+static void tegra_crtc_atomic_disable(struct drm_crtc *crtc,
+				      struct drm_atomic_commit *state)
+{
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	u32 value;
+	int err;
+
+	if (!tegra_dc_idle(dc)) {
+		tegra_dc_stop(dc);
+
+		/*
+		 * Ignore the return value, there isn't anything useful to do
+		 * in case this fails.
+		 */
+		tegra_dc_wait_idle(dc, 100);
+	}
+
+	/*
+	 * This should really be part of the RGB encoder driver, but clearing
+	 * these bits has the side-effect of stopping the display controller.
+	 * When that happens no VBLANK interrupts will be raised. At the same
+	 * time the encoder is disabled before the display controller, so the
+	 * above code is always going to timeout waiting for the controller
+	 * to go idle.
+	 *
+	 * Given the close coupling between the RGB encoder and the display
+	 * controller doing it here is still kind of okay. None of the other
+	 * encoder drivers require these bits to be cleared.
+	 *
+	 * XXX: Perhaps given that the display controller is switched off at
+	 * this point anyway maybe clearing these bits isn't even useful for
+	 * the RGB encoder?
+	 */
+	if (dc->rgb) {
+		value = tegra_dc_readl(dc, DC_CMD_DISPLAY_POWER_CONTROL);
+		value &= ~(PW0_ENABLE | PW1_ENABLE | PW2_ENABLE | PW3_ENABLE |
+			   PW4_ENABLE | PM0_ENABLE | PM1_ENABLE);
+		tegra_dc_writel(dc, value, DC_CMD_DISPLAY_POWER_CONTROL);
+	}
+
+	tegra_dc_stats_reset(&dc->stats);
+	drm_crtc_vblank_off(crtc);
+
+	spin_lock_irq(&crtc->dev->event_lock);
+
+	if (crtc->state->event) {
+		drm_crtc_send_vblank_event(crtc, crtc->state->event);
+		crtc->state->event = NULL;
+	}
+
+	spin_unlock_irq(&crtc->dev->event_lock);
+
+	err = host1x_client_suspend(&dc->client);
+	if (err < 0)
+		dev_err(dc->dev, "failed to suspend: %d\n", err);
+
+	if (dc->has_opp_table) {
+		err = dev_pm_genpd_set_performance_state(dc->dev, 0);
+		if (err)
+			dev_err(dc->dev,
+				"failed to clear power domain state: %d\n", err);
+	}
+}
+
+static void tegra_crtc_atomic_enable(struct drm_crtc *crtc,
+				     struct drm_atomic_commit *state)
+{
+	struct drm_display_mode *mode = &crtc->state->adjusted_mode;
+	struct tegra_dc_state *crtc_state = to_dc_state(crtc->state);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	u32 value;
+	int err;
+
+	/* apply PLL changes */
+	tegra_dc_set_clock_rate(dc, crtc_state);
+
+	err = host1x_client_resume(&dc->client);
+	if (err < 0) {
+		dev_err(dc->dev, "failed to resume: %d\n", err);
+		return;
+	}
+
+	/* initialize display controller */
+	if (dc->syncpt) {
+		u32 syncpt = host1x_syncpt_id(dc->syncpt), enable;
+
+		if (dc->soc->has_nvdisplay)
+			enable = 1 << 31;
+		else
+			enable = 1 << 8;
+
+		value = SYNCPT_CNTRL_NO_STALL;
+		tegra_dc_writel(dc, value, DC_CMD_GENERAL_INCR_SYNCPT_CNTRL);
+
+		value = enable | syncpt;
+		tegra_dc_writel(dc, value, DC_CMD_CONT_SYNCPT_VSYNC);
+	}
+
+	if (dc->soc->has_nvdisplay) {
+		value = DSC_TO_UF_INT | DSC_BBUF_UF_INT | DSC_RBUF_UF_INT |
+			DSC_OBUF_UF_INT;
+		tegra_dc_writel(dc, value, DC_CMD_INT_TYPE);
+
+		value = DSC_TO_UF_INT | DSC_BBUF_UF_INT | DSC_RBUF_UF_INT |
+			DSC_OBUF_UF_INT | SD3_BUCKET_WALK_DONE_INT |
+			HEAD_UF_INT | MSF_INT | REG_TMOUT_INT |
+			REGION_CRC_INT | V_PULSE2_INT | V_PULSE3_INT |
+			VBLANK_INT | FRAME_END_INT;
+		tegra_dc_writel(dc, value, DC_CMD_INT_POLARITY);
+
+		value = SD3_BUCKET_WALK_DONE_INT | HEAD_UF_INT | VBLANK_INT |
+			FRAME_END_INT;
+		tegra_dc_writel(dc, value, DC_CMD_INT_ENABLE);
+
+		value = HEAD_UF_INT | REG_TMOUT_INT | FRAME_END_INT;
+		tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
+
+		tegra_dc_writel(dc, READ_MUX, DC_CMD_STATE_ACCESS);
+	} else {
+		value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
+			WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
+		tegra_dc_writel(dc, value, DC_CMD_INT_TYPE);
+
+		value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
+			WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
+		tegra_dc_writel(dc, value, DC_CMD_INT_POLARITY);
+
+		/* initialize timer */
+		value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(0x20) |
+			WINDOW_B_THRESHOLD(0x20) | WINDOW_C_THRESHOLD(0x20);
+		tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY);
+
+		value = CURSOR_THRESHOLD(0) | WINDOW_A_THRESHOLD(1) |
+			WINDOW_B_THRESHOLD(1) | WINDOW_C_THRESHOLD(1);
+		tegra_dc_writel(dc, value, DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER);
+
+		value = VBLANK_INT | WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
+			WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
+		tegra_dc_writel(dc, value, DC_CMD_INT_ENABLE);
+
+		value = WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT |
+			WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT;
+		tegra_dc_writel(dc, value, DC_CMD_INT_MASK);
+	}
+
+	if (dc->soc->supports_background_color)
+		tegra_dc_writel(dc, 0, DC_DISP_BLEND_BACKGROUND_COLOR);
+	else
+		tegra_dc_writel(dc, 0, DC_DISP_BORDER_COLOR);
+
+	/* apply pixel clock changes */
+	if (!dc->soc->has_nvdisplay) {
+		value = SHIFT_CLK_DIVIDER(crtc_state->div) | PIXEL_CLK_DIVIDER_PCD1;
+		tegra_dc_writel(dc, value, DC_DISP_DISP_CLOCK_CONTROL);
+	}
+
+	/* program display mode */
+	tegra_dc_set_timings(dc, mode);
+
+	/* interlacing isn't supported yet, so disable it */
+	if (dc->soc->supports_interlacing) {
+		value = tegra_dc_readl(dc, DC_DISP_INTERLACE_CONTROL);
+		value &= ~INTERLACE_ENABLE;
+		tegra_dc_writel(dc, value, DC_DISP_INTERLACE_CONTROL);
+	}
+
+	value = tegra_dc_readl(dc, DC_CMD_DISPLAY_COMMAND);
+	value &= ~DISP_CTRL_MODE_MASK;
+	value |= DISP_CTRL_MODE_C_DISPLAY;
+	tegra_dc_writel(dc, value, DC_CMD_DISPLAY_COMMAND);
+
+	if (!dc->soc->has_nvdisplay) {
+		value = tegra_dc_readl(dc, DC_CMD_DISPLAY_POWER_CONTROL);
+		value |= PW0_ENABLE | PW1_ENABLE | PW2_ENABLE | PW3_ENABLE |
+			 PW4_ENABLE | PM0_ENABLE | PM1_ENABLE;
+		tegra_dc_writel(dc, value, DC_CMD_DISPLAY_POWER_CONTROL);
+	}
+
+	/* enable underflow reporting and display red for missing pixels */
+	if (dc->soc->has_nvdisplay) {
+		value = UNDERFLOW_MODE_RED | UNDERFLOW_REPORT_ENABLE;
+		tegra_dc_writel(dc, value, DC_COM_RG_UNDERFLOW);
+	}
+
+	if (dc->rgb) {
+		/* XXX: parameterize? */
+		value = SC0_H_QUALIFIER_NONE | SC1_H_QUALIFIER_NONE;
+		tegra_dc_writel(dc, value, DC_DISP_SHIFT_CLOCK_OPTIONS);
+	}
+
+	tegra_dc_commit(dc);
+
+	drm_crtc_vblank_on(crtc);
+}
+
+static void tegra_crtc_atomic_begin(struct drm_crtc *crtc,
+				    struct drm_atomic_commit *state)
+{
+	unsigned long flags;
+
+	tegra_crtc_update_memory_bandwidth(crtc, state, true);
+
+	if (crtc->state->event) {
+		spin_lock_irqsave(&crtc->dev->event_lock, flags);
+
+		if (drm_crtc_vblank_get(crtc) != 0)
+			drm_crtc_send_vblank_event(crtc, crtc->state->event);
+		else
+			drm_crtc_arm_vblank_event(crtc, crtc->state->event);
+
+		spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
+
+		crtc->state->event = NULL;
+	}
+}
+
+static void tegra_crtc_atomic_flush(struct drm_crtc *crtc,
+				    struct drm_atomic_commit *state)
+{
+	struct drm_crtc_state *crtc_state = drm_atomic_get_new_crtc_state(state,
+									  crtc);
+	struct tegra_dc_state *dc_state = to_dc_state(crtc_state);
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	u32 value;
+
+	value = dc_state->planes << 8 | GENERAL_UPDATE;
+	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
+	value = tegra_dc_readl(dc, DC_CMD_STATE_CONTROL);
+
+	value = dc_state->planes | GENERAL_ACT_REQ;
+	tegra_dc_writel(dc, value, DC_CMD_STATE_CONTROL);
+	value = tegra_dc_readl(dc, DC_CMD_STATE_CONTROL);
+}
+
+static bool tegra_plane_is_cursor(const struct drm_plane_state *state)
+{
+	const struct tegra_dc_soc_info *soc = to_tegra_dc(state->crtc)->soc;
+	const struct drm_format_info *fmt = state->fb->format;
+	unsigned int src_w = drm_rect_width(&state->src) >> 16;
+	unsigned int dst_w = drm_rect_width(&state->dst);
+
+	if (state->plane->type != DRM_PLANE_TYPE_CURSOR)
+		return false;
+
+	if (soc->supports_cursor)
+		return true;
+
+	if (src_w != dst_w || fmt->num_planes != 1 || src_w * fmt->cpp[0] > 256)
+		return false;
+
+	return true;
+}
+
+static unsigned long
+tegra_plane_overlap_mask(struct drm_crtc_state *state,
+			 const struct drm_plane_state *plane_state)
+{
+	const struct drm_plane_state *other_state;
+	const struct tegra_plane *tegra;
+	unsigned long overlap_mask = 0;
+	struct drm_plane *plane;
+	struct drm_rect rect;
+
+	if (!plane_state->visible || !plane_state->fb)
+		return 0;
+
+	/*
+	 * Data-prefetch FIFO will easily help to overcome temporal memory
+	 * pressure if other plane overlaps with the cursor plane.
+	 */
+	if (tegra_plane_is_cursor(plane_state))
+		return 0;
+
+	drm_atomic_crtc_state_for_each_plane_state(plane, other_state, state) {
+		rect = plane_state->dst;
+
+		tegra = to_tegra_plane(other_state->plane);
+
+		if (!other_state->visible || !other_state->fb)
+			continue;
+
+		/*
+		 * Ignore cursor plane overlaps because it's not practical to
+		 * assume that it contributes to the bandwidth in overlapping
+		 * area if window width is small.
+		 */
+		if (tegra_plane_is_cursor(other_state))
+			continue;
+
+		if (drm_rect_intersect(&rect, &other_state->dst))
+			overlap_mask |= BIT(tegra->index);
+	}
+
+	return overlap_mask;
+}
+
+static int tegra_crtc_calculate_memory_bandwidth(struct drm_crtc *crtc,
+						 struct drm_atomic_commit *state)
+{
+	ulong overlap_mask[TEGRA_DC_LEGACY_PLANES_NUM] = {}, mask;
+	u32 plane_peak_bw[TEGRA_DC_LEGACY_PLANES_NUM] = {};
+	bool all_planes_overlap_simultaneously = true;
+	const struct tegra_plane_state *tegra_state;
+	const struct drm_plane_state *plane_state;
+	struct tegra_dc *dc = to_tegra_dc(crtc);
+	struct drm_crtc_state *new_state;
+	struct tegra_plane *tegra;
+	struct drm_plane *plane;
+
+	/*
+	 * The nv-display uses shared planes.  The algorithm below assumes
+	 * maximum 3 planes per-CRTC, this assumption isn't applicable to
+	 * the nv-display.  Note that T124 support has additional windows,
+	 * but currently they aren't supported by the driver.
+	 */
+	if (dc->soc->has_nvdisplay)
+		return 0;
+
+	new_state = drm_atomic_get_new_crtc_state(state, crtc);
+
+	/*
+	 * For overlapping planes pixel's data is fetched for each plane at
+	 * the same time, hence bandwidths are accumulated in this case.
+	 * This needs to be taken into account for calculating total bandwidth
+	 * consumed by all planes.
+	 *
+	 * Here we get the overlapping state of each plane, which is a
+	 * bitmask of plane indices telling with what planes there is an
+	 * overlap. Note that bitmask[plane] includes BIT(plane) in order
+	 * to make further code nicer and simpler.
+	 */
+	drm_atomic_crtc_state_for_each_plane_state(plane, plane_state, new_state) {
+		tegra_state = to_const_tegra_plane_state(plane_state);
+		tegra = to_tegra_plane(plane);
+
+		if (WARN_ON_ONCE(tegra->index >= TEGRA_DC_LEGACY_PLANES_NUM))
+			return -EINVAL;
+
+		plane_peak_bw[tegra->index] = tegra_state->peak_memory_bandwidth;
+		mask = tegra_plane_overlap_mask(new_state, plane_state);
+		overlap_mask[tegra->index] = mask;
+
+		if (hweight_long(mask) != 3)
+			all_planes_overlap_simultaneously = false;
+	}
+
+	/*
+	 * Then we calculate maximum bandwidth of each plane state.
+	 * The bandwidth includes the plane BW + BW of the "simultaneously"
+	 * overlapping planes, where "simultaneously" means areas where DC
+	 * fetches from the planes simultaneously during of scan-out process.
+	 *
+	 * For example, if plane A overlaps with planes B and C, but B and C
+	 * don't overlap, then the peak bandwidth will be either in area where
+	 * A-and-B or A-and-C planes overlap.
+	 *
+	 * The plane_peak_bw[] contains peak memory bandwidth values of
+	 * each plane, this information is needed by interconnect provider
+	 * in order to set up latency allowance based on the peak BW, see
+	 * tegra_crtc_update_memory_bandwidth().
+	 */
+	drm_atomic_crtc_state_for_each_plane_state(plane, plane_state, new_state) {
+		u32 i, old_peak_bw, new_peak_bw, overlap_bw = 0;
+
+		/*
+		 * Note that plane's atomic check doesn't touch the
+		 * total_peak_memory_bandwidth of enabled plane, hence the
+		 * current state contains the old bandwidth state from the
+		 * previous CRTC commit.
+		 */
+		tegra_state = to_const_tegra_plane_state(plane_state);
+		tegra = to_tegra_plane(plane);
+
+		for_each_set_bit(i, &overlap_mask[tegra->index], 3) {
+			if (i == tegra->index)
+				continue;
+
+			if (all_planes_overlap_simultaneously)
+				overlap_bw += plane_peak_bw[i];
+			else
+				overlap_bw = max(overlap_bw, plane_peak_bw[i]);
+		}
+
+		new_peak_bw = plane_peak_bw[tegra->index] + overlap_bw;
+		old_peak_bw = tegra_state->total_peak_memory_bandwidth;
+
+		/*
+		 * If plane's peak bandwidth changed (for example plane isn't
+		 * overlapped anymore) and plane isn't in the atomic state,
+		 * then add plane to the state in order to have the bandwidth
+		 * updated.
+		 */
+		if (old_peak_bw != new_peak_bw) {
+			struct tegra_plane_state *new_tegra_state;
+			struct drm_plane_state *new_plane_state;
+
+			new_plane_state = drm_atomic_get_plane_state(state, plane);
+			if (IS_ERR(new_plane_state))
+				return PTR_ERR(new_plane_state);
+
+			new_tegra_state = to_tegra_plane_state(new_plane_state);
+			new_tegra_state->total_peak_memory_bandwidth = new_peak_bw;
+		}
+	}
+
+	return 0;
+}
+
+static int tegra_crtc_atomic_check(struct drm_crtc *crtc,
+				   struct drm_atomic_commit *state)
+{
+	int err;
+
+	err = tegra_crtc_calculate_memory_bandwidth(crtc, state);
+	if (err)
+		return err;
+
+	return 0;
+}
+
+void tegra_crtc_atomic_post_commit(struct drm_crtc *crtc,
+				   struct drm_atomic_commit *state)
+{
+	/*
+	 * Display bandwidth is allowed to go down only once hardware state
+	 * is known to be armed, i.e. state was committed and VBLANK event
+	 * received.
+	 */
+	tegra_crtc_update_memory_bandwidth(crtc, state, false);
+}
+
+static const struct drm_crtc_helper_funcs tegra_crtc_helper_funcs = {
+	.atomic_check = tegra_crtc_atomic_check,
+	.atomic_begin = tegra_crtc_atomic_begin,
+	.atomic_flush = tegra_crtc_atomic_flush,
+	.atomic_enable = tegra_crtc_atomic_enable,
+	.atomic_disable = tegra_crtc_atomic_disable,
+};
+
+static irqreturn_t tegra_dc_irq(int irq, void *data)
+{
+	struct tegra_dc *dc = data;
+	unsigned long status;
+
+	status = tegra_dc_readl(dc, DC_CMD_INT_STATUS);
+	tegra_dc_writel(dc, status, DC_CMD_INT_STATUS);
+
+	if (status & FRAME_END_INT) {
+		/*
+		dev_dbg(dc->dev, "%s(): frame end\n", __func__);
+		*/
+		dc->stats.frames_total++;
+		dc->stats.frames++;
+	}
+
+	if (status & VBLANK_INT) {
+		/*
+		dev_dbg(dc->dev, "%s(): vertical blank\n", __func__);
+		*/
+		drm_crtc_handle_vblank(&dc->base);
+		dc->stats.vblank_total++;
+		dc->stats.vblank++;
+	}
+
+	if (status & (WIN_A_UF_INT | WIN_B_UF_INT | WIN_C_UF_INT)) {
+		/*
+		dev_dbg(dc->dev, "%s(): underflow\n", __func__);
+		*/
+		dc->stats.underflow_total++;
+		dc->stats.underflow++;
+	}
+
+	if (status & (WIN_A_OF_INT | WIN_B_OF_INT | WIN_C_OF_INT)) {
+		/*
+		dev_dbg(dc->dev, "%s(): overflow\n", __func__);
+		*/
+		dc->stats.overflow_total++;
+		dc->stats.overflow++;
+	}
+
+	if (status & HEAD_UF_INT) {
+		dev_dbg_ratelimited(dc->dev, "%s(): head underflow\n", __func__);
+		dc->stats.underflow_total++;
+		dc->stats.underflow++;
+	}
+
+	return IRQ_HANDLED;
+}
+
+static bool tegra_dc_has_window_groups(struct tegra_dc *dc)
+{
+	unsigned int i;
+
+	if (!dc->soc->wgrps)
+		return true;
+
+	for (i = 0; i < dc->soc->num_wgrps; i++) {
+		const struct tegra_windowgroup_soc *wgrp = &dc->soc->wgrps[i];
+
+		if (wgrp->dc == dc->pipe && wgrp->num_windows > 0)
+			return true;
+	}
+
+	return false;
+}
+
+static int tegra_dc_early_init(struct host1x_client *client)
+{
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = drm->dev_private;
+
+	tegra->num_crtcs++;
+
+	return 0;
+}
+
+static int tegra_dc_init(struct host1x_client *client)
+{
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	unsigned long flags = HOST1X_SYNCPT_CLIENT_MANAGED;
+	struct tegra_dc *dc = host1x_client_to_dc(client);
+	struct tegra_drm *tegra = drm->dev_private;
+	struct drm_plane *primary = NULL;
+	struct drm_plane *cursor = NULL;
+	int err;
+
+	/*
+	 * DC has been reset by now, so VBLANK syncpoint can be released
+	 * for general use.
+	 */
+	host1x_syncpt_release_vblank_reservation(client, 26 + dc->pipe);
+
+	/*
+	 * XXX do not register DCs with no window groups because we cannot
+	 * assign a primary plane to them, which in turn will cause KMS to
+	 * crash.
+	 */
+	if (!tegra_dc_has_window_groups(dc))
+		return 0;
+
+	/*
+	 * Set the display hub as the host1x client parent for the display
+	 * controller. This is needed for the runtime reference counting that
+	 * ensures the display hub is always powered when any of the display
+	 * controllers are.
+	 */
+	if (dc->soc->has_nvdisplay)
+		client->parent = &tegra->hub->client;
+
+	dc->syncpt = host1x_syncpt_request(client, flags);
+	if (!dc->syncpt)
+		dev_warn(dc->dev, "failed to allocate syncpoint\n");
+
+	err = host1x_client_iommu_attach(client);
+	if (err < 0 && err != -ENODEV) {
+		dev_err(client->dev, "failed to attach to domain: %d\n", err);
+		return err;
+	}
+
+	if (dc->soc->wgrps)
+		primary = tegra_dc_add_shared_planes(drm, dc);
+	else
+		primary = tegra_dc_add_planes(drm, dc);
+
+	if (IS_ERR(primary)) {
+		err = PTR_ERR(primary);
+		goto cleanup;
+	}
+
+	if (dc->soc->supports_cursor) {
+		cursor = tegra_dc_cursor_plane_create(drm, dc);
+		if (IS_ERR(cursor)) {
+			err = PTR_ERR(cursor);
+			goto cleanup;
+		}
+	} else {
+		/* dedicate one overlay to mouse cursor */
+		cursor = tegra_dc_overlay_plane_create(drm, dc, 2, true);
+		if (IS_ERR(cursor)) {
+			err = PTR_ERR(cursor);
+			goto cleanup;
+		}
+	}
+
+	err = drm_crtc_init_with_planes(drm, &dc->base, primary, cursor,
+					&tegra_crtc_funcs, NULL);
+	if (err < 0)
+		goto cleanup;
+
+	drm_crtc_helper_add(&dc->base, &tegra_crtc_helper_funcs);
+
+	/*
+	 * Keep track of the minimum pitch alignment across all display
+	 * controllers.
+	 */
+	if (dc->soc->pitch_align > tegra->pitch_align)
+		tegra->pitch_align = dc->soc->pitch_align;
+
+	/* track maximum resolution */
+	if (dc->soc->has_nvdisplay)
+		drm->mode_config.max_width = drm->mode_config.max_height = 16384;
+	else
+		drm->mode_config.max_width = drm->mode_config.max_height = 4096;
+
+	err = tegra_dc_rgb_init(drm, dc);
+	if (err < 0 && err != -ENODEV) {
+		dev_err(dc->dev, "failed to initialize RGB output: %d\n", err);
+		goto cleanup;
+	}
+
+	err = devm_request_irq(dc->dev, dc->irq, tegra_dc_irq, 0,
+			       dev_name(dc->dev), dc);
+	if (err < 0) {
+		dev_err(dc->dev, "failed to request IRQ#%u: %d\n", dc->irq,
+			err);
+		goto cleanup;
+	}
+
+	/*
+	 * Inherit the DMA parameters (such as maximum segment size) from the
+	 * parent host1x device.
+	 */
+	client->dev->dma_parms = client->host->dma_parms;
+
+	return 0;
+
+cleanup:
+	if (!IS_ERR_OR_NULL(cursor))
+		drm_plane_cleanup(cursor);
+
+	if (!IS_ERR(primary))
+		drm_plane_cleanup(primary);
+
+	host1x_client_iommu_detach(client);
+	host1x_syncpt_put(dc->syncpt);
+
+	return err;
+}
+
+static int tegra_dc_exit(struct host1x_client *client)
+{
+	struct tegra_dc *dc = host1x_client_to_dc(client);
+	int err;
+
+	if (!tegra_dc_has_window_groups(dc))
+		return 0;
+
+	/* avoid a dangling pointer just in case this disappears */
+	client->dev->dma_parms = NULL;
+
+	devm_free_irq(dc->dev, dc->irq, dc);
+
+	err = tegra_dc_rgb_exit(dc);
+	if (err) {
+		dev_err(dc->dev, "failed to shutdown RGB output: %d\n", err);
+		return err;
+	}
+
+	host1x_client_iommu_detach(client);
+	host1x_syncpt_put(dc->syncpt);
+
+	return 0;
+}
+
+static int tegra_dc_late_exit(struct host1x_client *client)
+{
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = drm->dev_private;
+
+	tegra->num_crtcs--;
+
+	return 0;
+}
+
+static int tegra_dc_runtime_suspend(struct host1x_client *client)
+{
+	struct tegra_dc *dc = host1x_client_to_dc(client);
+	struct device *dev = client->dev;
+	int err;
+
+	err = reset_control_assert(dc->rst);
+	if (err < 0) {
+		dev_err(dev, "failed to assert reset: %d\n", err);
+		return err;
+	}
+
+	if (dc->soc->has_powergate)
+		tegra_pmc_powergate_power_off(dc->pmc, dc->powergate);
+
+	clk_disable_unprepare(dc->clk);
+	pm_runtime_put_sync(dev);
+
+	return 0;
+}
+
+static int tegra_dc_runtime_resume(struct host1x_client *client)
+{
+	struct tegra_dc *dc = host1x_client_to_dc(client);
+	struct device *dev = client->dev;
+	int err;
+
+	err = pm_runtime_resume_and_get(dev);
+	if (err < 0) {
+		dev_err(dev, "failed to get runtime PM: %d\n", err);
+		return err;
+	}
+
+	if (dc->soc->has_powergate) {
+		err = tegra_pmc_powergate_sequence_power_up(dc->pmc,
+							    dc->powergate,
+							    dc->clk, dc->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to power partition: %d\n", err);
+			goto put_rpm;
+		}
+	} else {
+		err = clk_prepare_enable(dc->clk);
+		if (err < 0) {
+			dev_err(dev, "failed to enable clock: %d\n", err);
+			goto put_rpm;
+		}
+
+		err = reset_control_deassert(dc->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to deassert reset: %d\n", err);
+			goto disable_clk;
+		}
+	}
+
+	return 0;
+
+disable_clk:
+	clk_disable_unprepare(dc->clk);
+put_rpm:
+	pm_runtime_put_sync(dev);
+	return err;
+}
+
+static const struct host1x_client_ops dc_client_ops = {
+	.early_init = tegra_dc_early_init,
+	.init = tegra_dc_init,
+	.exit = tegra_dc_exit,
+	.late_exit = tegra_dc_late_exit,
+	.suspend = tegra_dc_runtime_suspend,
+	.resume = tegra_dc_runtime_resume,
+};
+
+static const struct tegra_dc_soc_info tegra20_dc_soc_info = {
+	.supports_background_color = false,
+	.supports_interlacing = false,
+	.supports_cursor = false,
+	.supports_block_linear = false,
+	.supports_sector_layout = false,
+	.has_legacy_blending = true,
+	.pitch_align = 8,
+	.has_powergate = false,
+	.coupled_pm = true,
+	.has_nvdisplay = false,
+	.num_primary_formats = ARRAY_SIZE(tegra20_primary_formats),
+	.primary_formats = tegra20_primary_formats,
+	.num_overlay_formats = ARRAY_SIZE(tegra20_overlay_formats),
+	.overlay_formats = tegra20_overlay_formats,
+	.modifiers = tegra20_modifiers,
+	.has_win_a_without_filters = true,
+	.has_win_b_vfilter_mem_client = true,
+	.has_win_c_without_vert_filter = true,
+	.plane_tiled_memory_bandwidth_x2 = false,
+	.has_pll_d2_out0 = false,
+};
+
+static const struct tegra_dc_soc_info tegra30_dc_soc_info = {
+	.supports_background_color = false,
+	.supports_interlacing = false,
+	.supports_cursor = false,
+	.supports_block_linear = false,
+	.supports_sector_layout = false,
+	.has_legacy_blending = true,
+	.pitch_align = 8,
+	.has_powergate = false,
+	.coupled_pm = false,
+	.has_nvdisplay = false,
+	.num_primary_formats = ARRAY_SIZE(tegra20_primary_formats),
+	.primary_formats = tegra20_primary_formats,
+	.num_overlay_formats = ARRAY_SIZE(tegra20_overlay_formats),
+	.overlay_formats = tegra20_overlay_formats,
+	.modifiers = tegra20_modifiers,
+	.has_win_a_without_filters = false,
+	.has_win_b_vfilter_mem_client = true,
+	.has_win_c_without_vert_filter = false,
+	.plane_tiled_memory_bandwidth_x2 = true,
+	.has_pll_d2_out0 = true,
+};
+
+static const struct tegra_dc_soc_info tegra114_dc_soc_info = {
+	.supports_background_color = false,
+	.supports_interlacing = false,
+	.supports_cursor = false,
+	.supports_block_linear = false,
+	.supports_sector_layout = false,
+	.has_legacy_blending = true,
+	.pitch_align = 64,
+	.has_powergate = true,
+	.coupled_pm = false,
+	.has_nvdisplay = false,
+	.num_primary_formats = ARRAY_SIZE(tegra114_primary_formats),
+	.primary_formats = tegra114_primary_formats,
+	.num_overlay_formats = ARRAY_SIZE(tegra114_overlay_formats),
+	.overlay_formats = tegra114_overlay_formats,
+	.modifiers = tegra20_modifiers,
+	.has_win_a_without_filters = false,
+	.has_win_b_vfilter_mem_client = false,
+	.has_win_c_without_vert_filter = false,
+	.plane_tiled_memory_bandwidth_x2 = true,
+	.has_pll_d2_out0 = true,
+};
+
+static const struct tegra_dc_soc_info tegra124_dc_soc_info = {
+	.supports_background_color = true,
+	.supports_interlacing = true,
+	.supports_cursor = true,
+	.supports_block_linear = true,
+	.supports_sector_layout = false,
+	.has_legacy_blending = false,
+	.pitch_align = 64,
+	.has_powergate = true,
+	.coupled_pm = false,
+	.has_nvdisplay = false,
+	.num_primary_formats = ARRAY_SIZE(tegra124_primary_formats),
+	.primary_formats = tegra124_primary_formats,
+	.num_overlay_formats = ARRAY_SIZE(tegra124_overlay_formats),
+	.overlay_formats = tegra124_overlay_formats,
+	.modifiers = tegra124_modifiers,
+	.has_win_a_without_filters = false,
+	.has_win_b_vfilter_mem_client = false,
+	.has_win_c_without_vert_filter = false,
+	.plane_tiled_memory_bandwidth_x2 = false,
+	.has_pll_d2_out0 = true,
+};
+
+static const struct tegra_dc_soc_info tegra210_dc_soc_info = {
+	.supports_background_color = true,
+	.supports_interlacing = true,
+	.supports_cursor = true,
+	.supports_block_linear = true,
+	.supports_sector_layout = false,
+	.has_legacy_blending = false,
+	.pitch_align = 64,
+	.has_powergate = true,
+	.coupled_pm = false,
+	.has_nvdisplay = false,
+	.num_primary_formats = ARRAY_SIZE(tegra114_primary_formats),
+	.primary_formats = tegra114_primary_formats,
+	.num_overlay_formats = ARRAY_SIZE(tegra114_overlay_formats),
+	.overlay_formats = tegra114_overlay_formats,
+	.modifiers = tegra124_modifiers,
+	.has_win_a_without_filters = false,
+	.has_win_b_vfilter_mem_client = false,
+	.has_win_c_without_vert_filter = false,
+	.plane_tiled_memory_bandwidth_x2 = false,
+	.has_pll_d2_out0 = true,
+};
+
+static const struct tegra_windowgroup_soc tegra186_dc_wgrps[] = {
+	{
+		.index = 0,
+		.dc = 0,
+		.windows = (const unsigned int[]) { 0 },
+		.num_windows = 1,
+	}, {
+		.index = 1,
+		.dc = 1,
+		.windows = (const unsigned int[]) { 1 },
+		.num_windows = 1,
+	}, {
+		.index = 2,
+		.dc = 1,
+		.windows = (const unsigned int[]) { 2 },
+		.num_windows = 1,
+	}, {
+		.index = 3,
+		.dc = 2,
+		.windows = (const unsigned int[]) { 3 },
+		.num_windows = 1,
+	}, {
+		.index = 4,
+		.dc = 2,
+		.windows = (const unsigned int[]) { 4 },
+		.num_windows = 1,
+	}, {
+		.index = 5,
+		.dc = 2,
+		.windows = (const unsigned int[]) { 5 },
+		.num_windows = 1,
+	},
+};
+
+static const struct tegra_dc_soc_info tegra186_dc_soc_info = {
+	.supports_background_color = true,
+	.supports_interlacing = true,
+	.supports_cursor = true,
+	.supports_block_linear = true,
+	.supports_sector_layout = false,
+	.has_legacy_blending = false,
+	.pitch_align = 64,
+	.has_powergate = false,
+	.coupled_pm = false,
+	.has_nvdisplay = true,
+	.wgrps = tegra186_dc_wgrps,
+	.num_wgrps = ARRAY_SIZE(tegra186_dc_wgrps),
+	.plane_tiled_memory_bandwidth_x2 = false,
+	.has_pll_d2_out0 = false,
+};
+
+static const struct tegra_windowgroup_soc tegra194_dc_wgrps[] = {
+	{
+		.index = 0,
+		.dc = 0,
+		.windows = (const unsigned int[]) { 0 },
+		.num_windows = 1,
+	}, {
+		.index = 1,
+		.dc = 1,
+		.windows = (const unsigned int[]) { 1 },
+		.num_windows = 1,
+	}, {
+		.index = 2,
+		.dc = 1,
+		.windows = (const unsigned int[]) { 2 },
+		.num_windows = 1,
+	}, {
+		.index = 3,
+		.dc = 2,
+		.windows = (const unsigned int[]) { 3 },
+		.num_windows = 1,
+	}, {
+		.index = 4,
+		.dc = 2,
+		.windows = (const unsigned int[]) { 4 },
+		.num_windows = 1,
+	}, {
+		.index = 5,
+		.dc = 2,
+		.windows = (const unsigned int[]) { 5 },
+		.num_windows = 1,
+	},
+};
+
+static const struct tegra_dc_soc_info tegra194_dc_soc_info = {
+	.supports_background_color = true,
+	.supports_interlacing = true,
+	.supports_cursor = true,
+	.supports_block_linear = true,
+	.supports_sector_layout = true,
+	.has_legacy_blending = false,
+	.pitch_align = 64,
+	.has_powergate = false,
+	.coupled_pm = false,
+	.has_nvdisplay = true,
+	.wgrps = tegra194_dc_wgrps,
+	.num_wgrps = ARRAY_SIZE(tegra194_dc_wgrps),
+	.plane_tiled_memory_bandwidth_x2 = false,
+	.has_pll_d2_out0 = false,
+};
+
+static const struct of_device_id tegra_dc_of_match[] = {
+	{
+		.compatible = "nvidia,tegra194-dc",
+		.data = &tegra194_dc_soc_info,
+	}, {
+		.compatible = "nvidia,tegra186-dc",
+		.data = &tegra186_dc_soc_info,
+	}, {
+		.compatible = "nvidia,tegra210-dc",
+		.data = &tegra210_dc_soc_info,
+	}, {
+		.compatible = "nvidia,tegra124-dc",
+		.data = &tegra124_dc_soc_info,
+	}, {
+		.compatible = "nvidia,tegra114-dc",
+		.data = &tegra114_dc_soc_info,
+	}, {
+		.compatible = "nvidia,tegra30-dc",
+		.data = &tegra30_dc_soc_info,
+	}, {
+		.compatible = "nvidia,tegra20-dc",
+		.data = &tegra20_dc_soc_info,
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(of, tegra_dc_of_match);
+
+static int tegra_dc_parse_dt(struct tegra_dc *dc)
+{
+	struct device_node *np;
+	u32 value = 0;
+	int err;
+
+	err = of_property_read_u32(dc->dev->of_node, "nvidia,head", &value);
+	if (err < 0) {
+		dev_err(dc->dev, "missing \"nvidia,head\" property\n");
+
+		/*
+		 * If the nvidia,head property isn't present, try to find the
+		 * correct head number by looking up the position of this
+		 * display controller's node within the device tree. Assuming
+		 * that the nodes are ordered properly in the DTS file and
+		 * that the translation into a flattened device tree blob
+		 * preserves that ordering this will actually yield the right
+		 * head number.
+		 *
+		 * If those assumptions don't hold, this will still work for
+		 * cases where only a single display controller is used.
+		 */
+		for_each_matching_node(np, tegra_dc_of_match) {
+			if (np == dc->dev->of_node) {
+				of_node_put(np);
+				break;
+			}
+
+			value++;
+		}
+	}
+
+	dc->pipe = value;
+
+	return 0;
+}
+
+static int tegra_dc_match_by_pipe(struct device *dev, const void *data)
+{
+	struct tegra_dc *dc = dev_get_drvdata(dev);
+	unsigned int pipe = (unsigned long)(void *)data;
+
+	return dc->pipe == pipe;
+}
+
+static int tegra_dc_couple(struct tegra_dc *dc)
+{
+	/*
+	 * On Tegra20, DC1 requires DC0 to be taken out of reset in order to
+	 * be enabled, otherwise CPU hangs on writing to CMD_DISPLAY_COMMAND /
+	 * POWER_CONTROL registers during CRTC enabling.
+	 */
+	if (dc->soc->coupled_pm && dc->pipe == 1) {
+		struct device *companion;
+		struct tegra_dc *parent;
+
+		companion = driver_find_device(dc->dev->driver, NULL, (const void *)0,
+					       tegra_dc_match_by_pipe);
+		if (!companion)
+			return -EPROBE_DEFER;
+
+		parent = dev_get_drvdata(companion);
+		dc->client.parent = &parent->client;
+
+		dev_dbg(dc->dev, "coupled to %s\n", dev_name(companion));
+		put_device(companion);
+	}
+
+	return 0;
+}
+
+static int tegra_dc_init_opp_table(struct tegra_dc *dc)
+{
+	struct tegra_core_opp_params opp_params = {};
+	int err;
+
+	err = devm_tegra_core_dev_init_opp_table(dc->dev, &opp_params);
+	if (err && err != -ENODEV)
+		return err;
+
+	if (err)
+		dc->has_opp_table = false;
+	else
+		dc->has_opp_table = true;
+
+	return 0;
+}
+
+static int tegra_dc_probe(struct platform_device *pdev)
+{
+	u64 dma_mask = dma_get_mask(pdev->dev.parent);
+	struct tegra_dc *dc;
+	int err;
+
+	err = dma_coerce_mask_and_coherent(&pdev->dev, dma_mask);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set DMA mask: %d\n", err);
+		return err;
+	}
+
+	dc = devm_kzalloc(&pdev->dev, sizeof(*dc), GFP_KERNEL);
+	if (!dc)
+		return -ENOMEM;
+
+	dc->soc = of_device_get_match_data(&pdev->dev);
+
+	INIT_LIST_HEAD(&dc->list);
+	dc->dev = &pdev->dev;
+
+	err = tegra_dc_parse_dt(dc);
+	if (err < 0)
+		return err;
+
+	err = tegra_dc_couple(dc);
+	if (err < 0)
+		return err;
+
+	dc->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(dc->clk)) {
+		dev_err(&pdev->dev, "failed to get clock\n");
+		return PTR_ERR(dc->clk);
+	}
+
+	dc->rst = devm_reset_control_get(&pdev->dev, "dc");
+	if (IS_ERR(dc->rst)) {
+		dev_err(&pdev->dev, "failed to get reset\n");
+		return PTR_ERR(dc->rst);
+	}
+
+	/* assert reset and disable clock */
+	err = clk_prepare_enable(dc->clk);
+	if (err < 0)
+		return err;
+
+	usleep_range(2000, 4000);
+
+	err = reset_control_assert(dc->rst);
+	if (err < 0) {
+		clk_disable_unprepare(dc->clk);
+		return err;
+	}
+
+	usleep_range(2000, 4000);
+
+	clk_disable_unprepare(dc->clk);
+
+	if (dc->soc->has_powergate) {
+		dc->pmc = devm_tegra_pmc_get(dc->dev);
+		if (IS_ERR(dc->pmc))
+			return dev_err_probe(dc->dev, PTR_ERR(dc->pmc),
+					     "failed to get PMC\n");
+
+		if (dc->pipe == 0)
+			dc->powergate = TEGRA_POWERGATE_DIS;
+		else
+			dc->powergate = TEGRA_POWERGATE_DISB;
+
+		tegra_pmc_powergate_power_off(dc->pmc, dc->powergate);
+	}
+
+	err = tegra_dc_init_opp_table(dc);
+	if (err < 0)
+		return err;
+
+	dc->regs = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(dc->regs))
+		return PTR_ERR(dc->regs);
+
+	dc->irq = platform_get_irq(pdev, 0);
+	if (dc->irq < 0)
+		return -ENXIO;
+
+	if (dc->soc->has_nvdisplay) {
+		unsigned int i;
+		u64 r;
+
+		dc->cmu_output_lut =
+			dmam_alloc_coherent(dc->dev, ARRAY_SIZE(default_srgb_lut) * sizeof(u64),
+					    &dc->cmu_output_lut_phys, GFP_KERNEL);
+
+		if (!dc->cmu_output_lut) {
+			dev_err(dc->dev, "failed to allocate lut for cmu\n");
+			return -ENOMEM;
+		}
+
+		for (i = 0; i < ARRAY_SIZE(default_srgb_lut); i++) {
+			r = default_srgb_lut[i];
+			dc->cmu_output_lut[i] = (r << 32) | (r << 16) | r;
+		}
+	}
+
+	err = tegra_dc_rgb_probe(dc);
+	if (err < 0 && err != -ENODEV)
+		return dev_err_probe(&pdev->dev, err,
+				     "failed to probe RGB output\n");
+
+	platform_set_drvdata(pdev, dc);
+	pm_runtime_enable(&pdev->dev);
+
+	INIT_LIST_HEAD(&dc->client.list);
+	dc->client.ops = &dc_client_ops;
+	dc->client.dev = &pdev->dev;
+
+	err = host1x_client_register(&dc->client);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to register host1x client: %d\n",
+			err);
+		goto disable_pm;
+	}
+
+	return 0;
+
+disable_pm:
+	pm_runtime_disable(&pdev->dev);
+	tegra_dc_rgb_remove(dc);
+
+	return err;
+}
+
+static void tegra_dc_remove(struct platform_device *pdev)
+{
+	struct tegra_dc *dc = platform_get_drvdata(pdev);
+
+	host1x_client_unregister(&dc->client);
+
+	tegra_dc_rgb_remove(dc);
+
+	pm_runtime_disable(&pdev->dev);
+}
+
+struct platform_driver tegra_dc_driver = {
+	.driver = {
+		.name = "tegra-dc",
+		.of_match_table = tegra_dc_of_match,
+	},
+	.probe = tegra_dc_probe,
+	.remove = tegra_dc_remove,
+};
diff --git a/drivers/gpu/drm/tegra/dc.h b/drivers/gpu/drm/tegra/dc.h
new file mode 100644
index 000000000..0cb051596
--- /dev/null
+++ b/drivers/gpu/drm/tegra/dc.h
@@ -0,0 +1,854 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#ifndef TEGRA_DC_H
+#define TEGRA_DC_H 1
+
+#include <linux/host1x.h>
+
+#include <drm/drm_crtc.h>
+
+#include "drm.h"
+
+struct tegra_output;
+
+#define TEGRA_DC_LEGACY_PLANES_NUM	7
+
+struct tegra_dc_state {
+	struct drm_crtc_state base;
+
+	struct clk *clk;
+	unsigned long pclk;
+	unsigned int div;
+
+	u32 planes;
+};
+
+static inline struct tegra_dc_state *to_dc_state(struct drm_crtc_state *state)
+{
+	if (state)
+		return container_of(state, struct tegra_dc_state, base);
+
+	return NULL;
+}
+
+struct tegra_dc_stats {
+	unsigned long frames;
+	unsigned long vblank;
+	unsigned long underflow;
+	unsigned long overflow;
+
+	unsigned long frames_total;
+	unsigned long vblank_total;
+	unsigned long underflow_total;
+	unsigned long overflow_total;
+};
+
+struct tegra_windowgroup_soc {
+	unsigned int index;
+	unsigned int dc;
+	const unsigned int *windows;
+	unsigned int num_windows;
+};
+
+struct tegra_dc_soc_info {
+	bool supports_background_color;
+	bool supports_interlacing;
+	bool supports_cursor;
+	bool supports_block_linear;
+	bool supports_sector_layout;
+	bool has_legacy_blending;
+	unsigned int pitch_align;
+	bool has_powergate;
+	bool coupled_pm;
+	bool has_nvdisplay;
+	const struct tegra_windowgroup_soc *wgrps;
+	unsigned int num_wgrps;
+	const u32 *primary_formats;
+	unsigned int num_primary_formats;
+	const u32 *overlay_formats;
+	unsigned int num_overlay_formats;
+	const u64 *modifiers;
+	bool has_win_a_without_filters;
+	bool has_win_b_vfilter_mem_client;
+	bool has_win_c_without_vert_filter;
+	bool plane_tiled_memory_bandwidth_x2;
+	bool has_pll_d2_out0;
+};
+
+struct tegra_dc {
+	struct host1x_client client;
+	struct host1x_syncpt *syncpt;
+	struct device *dev;
+
+	struct drm_crtc base;
+	unsigned int powergate;
+	int pipe;
+
+	struct clk *clk;
+	struct reset_control *rst;
+	void __iomem *regs;
+	int irq;
+
+	struct tegra_output *rgb;
+	struct tegra_pmc *pmc;
+
+	struct tegra_dc_stats stats;
+	struct list_head list;
+
+	struct drm_info_list *debugfs_files;
+
+	const struct tegra_dc_soc_info *soc;
+
+	bool has_opp_table;
+
+	u64 *cmu_output_lut;
+	dma_addr_t cmu_output_lut_phys;
+};
+
+static inline struct tegra_dc *
+host1x_client_to_dc(struct host1x_client *client)
+{
+	return container_of(client, struct tegra_dc, client);
+}
+
+static inline struct tegra_dc *to_tegra_dc(struct drm_crtc *crtc)
+{
+	return crtc ? container_of(crtc, struct tegra_dc, base) : NULL;
+}
+
+static inline void tegra_dc_writel(struct tegra_dc *dc, u32 value,
+				   unsigned int offset)
+{
+	trace_dc_writel(dc->dev, offset, value);
+	writel(value, dc->regs + (offset << 2));
+}
+
+static inline u32 tegra_dc_readl(struct tegra_dc *dc, unsigned int offset)
+{
+	u32 value = readl(dc->regs + (offset << 2));
+
+	trace_dc_readl(dc->dev, offset, value);
+
+	return value;
+}
+
+struct tegra_dc_window {
+	struct {
+		unsigned int x;
+		unsigned int y;
+		unsigned int w;
+		unsigned int h;
+	} src;
+	struct {
+		unsigned int x;
+		unsigned int y;
+		unsigned int w;
+		unsigned int h;
+	} dst;
+	unsigned int bits_per_pixel;
+	unsigned int stride[2];
+	unsigned long base[3];
+	unsigned int zpos;
+	bool reflect_x;
+	bool reflect_y;
+
+	struct tegra_bo_tiling tiling;
+	u32 format;
+	u32 swap;
+};
+
+/* from dc.c */
+bool tegra_dc_has_output(struct tegra_dc *dc, struct device *dev);
+void tegra_dc_commit(struct tegra_dc *dc);
+int tegra_dc_state_setup_clock(struct tegra_dc *dc,
+			       struct drm_crtc_state *crtc_state,
+			       struct clk *clk, unsigned long pclk,
+			       unsigned int div);
+void tegra_crtc_atomic_post_commit(struct drm_crtc *crtc,
+				   struct drm_atomic_commit *state);
+
+/* from rgb.c */
+int tegra_dc_rgb_probe(struct tegra_dc *dc);
+void tegra_dc_rgb_remove(struct tegra_dc *dc);
+int tegra_dc_rgb_init(struct drm_device *drm, struct tegra_dc *dc);
+int tegra_dc_rgb_exit(struct tegra_dc *dc);
+
+#define DC_CMD_GENERAL_INCR_SYNCPT		0x000
+#define DC_CMD_GENERAL_INCR_SYNCPT_CNTRL	0x001
+#define  SYNCPT_CNTRL_NO_STALL   (1 << 8)
+#define  SYNCPT_CNTRL_SOFT_RESET (1 << 0)
+#define DC_CMD_GENERAL_INCR_SYNCPT_ERROR	0x002
+#define DC_CMD_WIN_A_INCR_SYNCPT		0x008
+#define DC_CMD_WIN_A_INCR_SYNCPT_CNTRL		0x009
+#define DC_CMD_WIN_A_INCR_SYNCPT_ERROR		0x00a
+#define DC_CMD_WIN_B_INCR_SYNCPT		0x010
+#define DC_CMD_WIN_B_INCR_SYNCPT_CNTRL		0x011
+#define DC_CMD_WIN_B_INCR_SYNCPT_ERROR		0x012
+#define DC_CMD_WIN_C_INCR_SYNCPT		0x018
+#define DC_CMD_WIN_C_INCR_SYNCPT_CNTRL		0x019
+#define DC_CMD_WIN_C_INCR_SYNCPT_ERROR		0x01a
+#define DC_CMD_CONT_SYNCPT_VSYNC		0x028
+#define  SYNCPT_VSYNC_ENABLE (1 << 8)
+#define DC_CMD_DISPLAY_COMMAND_OPTION0		0x031
+#define DC_CMD_DISPLAY_COMMAND			0x032
+#define DISP_CTRL_MODE_STOP (0 << 5)
+#define DISP_CTRL_MODE_C_DISPLAY (1 << 5)
+#define DISP_CTRL_MODE_NC_DISPLAY (2 << 5)
+#define DISP_CTRL_MODE_MASK (3 << 5)
+#define DC_CMD_SIGNAL_RAISE			0x033
+#define DC_CMD_DISPLAY_POWER_CONTROL		0x036
+#define PW0_ENABLE (1 <<  0)
+#define PW1_ENABLE (1 <<  2)
+#define PW2_ENABLE (1 <<  4)
+#define PW3_ENABLE (1 <<  6)
+#define PW4_ENABLE (1 <<  8)
+#define PM0_ENABLE (1 << 16)
+#define PM1_ENABLE (1 << 18)
+
+#define DC_CMD_INT_STATUS			0x037
+#define DC_CMD_INT_MASK				0x038
+#define DC_CMD_INT_ENABLE			0x039
+#define DC_CMD_INT_TYPE				0x03a
+#define DC_CMD_INT_POLARITY			0x03b
+#define CTXSW_INT                (1 << 0)
+#define FRAME_END_INT            (1 << 1)
+#define VBLANK_INT               (1 << 2)
+#define V_PULSE3_INT             (1 << 4)
+#define V_PULSE2_INT             (1 << 5)
+#define REGION_CRC_INT           (1 << 6)
+#define REG_TMOUT_INT            (1 << 7)
+#define WIN_A_UF_INT             (1 << 8)
+#define WIN_B_UF_INT             (1 << 9)
+#define WIN_C_UF_INT             (1 << 10)
+#define MSF_INT                  (1 << 12)
+#define WIN_A_OF_INT             (1 << 14)
+#define WIN_B_OF_INT             (1 << 15)
+#define WIN_C_OF_INT             (1 << 16)
+#define HEAD_UF_INT              (1 << 23)
+#define SD3_BUCKET_WALK_DONE_INT (1 << 24)
+#define DSC_OBUF_UF_INT          (1 << 26)
+#define DSC_RBUF_UF_INT          (1 << 27)
+#define DSC_BBUF_UF_INT          (1 << 28)
+#define DSC_TO_UF_INT            (1 << 29)
+
+#define DC_CMD_SIGNAL_RAISE1			0x03c
+#define DC_CMD_SIGNAL_RAISE2			0x03d
+#define DC_CMD_SIGNAL_RAISE3			0x03e
+
+#define DC_CMD_STATE_ACCESS			0x040
+#define READ_MUX  (1 << 0)
+#define WRITE_MUX (1 << 2)
+
+#define DC_CMD_STATE_CONTROL			0x041
+#define GENERAL_ACT_REQ (1 <<  0)
+#define WIN_A_ACT_REQ   (1 <<  1)
+#define WIN_B_ACT_REQ   (1 <<  2)
+#define WIN_C_ACT_REQ   (1 <<  3)
+#define CURSOR_ACT_REQ  (1 <<  7)
+#define GENERAL_UPDATE  (1 <<  8)
+#define WIN_A_UPDATE    (1 <<  9)
+#define WIN_B_UPDATE    (1 << 10)
+#define WIN_C_UPDATE    (1 << 11)
+#define CURSOR_UPDATE   (1 << 15)
+#define COMMON_ACTREQ   (1 << 16)
+#define COMMON_UPDATE   (1 << 17)
+#define NC_HOST_TRIG    (1 << 24)
+
+#define DC_CMD_DISPLAY_WINDOW_HEADER		0x042
+#define WINDOW_A_SELECT (1 << 4)
+#define WINDOW_B_SELECT (1 << 5)
+#define WINDOW_C_SELECT (1 << 6)
+
+#define DC_CMD_REG_ACT_CONTROL			0x043
+
+#define DC_COM_CRC_CONTROL			0x300
+#define  DC_COM_CRC_CONTROL_ALWAYS (1 << 3)
+#define  DC_COM_CRC_CONTROL_FULL_FRAME  (0 << 2)
+#define  DC_COM_CRC_CONTROL_ACTIVE_DATA (1 << 2)
+#define  DC_COM_CRC_CONTROL_WAIT (1 << 1)
+#define  DC_COM_CRC_CONTROL_ENABLE (1 << 0)
+#define DC_COM_CRC_CHECKSUM			0x301
+#define DC_COM_PIN_OUTPUT_ENABLE(x) (0x302 + (x))
+#define DC_COM_PIN_OUTPUT_POLARITY(x) (0x306 + (x))
+#define LVS_OUTPUT_POLARITY_LOW (1 << 28)
+#define LHS_OUTPUT_POLARITY_LOW (1 << 30)
+#define DC_COM_PIN_OUTPUT_DATA(x) (0x30a + (x))
+#define DC_COM_PIN_INPUT_ENABLE(x) (0x30e + (x))
+#define DC_COM_PIN_INPUT_DATA(x) (0x312 + (x))
+#define DC_COM_PIN_OUTPUT_SELECT(x) (0x314 + (x))
+
+#define DC_COM_PIN_MISC_CONTROL			0x31b
+#define DC_COM_PIN_PM0_CONTROL			0x31c
+#define DC_COM_PIN_PM0_DUTY_CYCLE		0x31d
+#define DC_COM_PIN_PM1_CONTROL			0x31e
+#define DC_COM_PIN_PM1_DUTY_CYCLE		0x31f
+
+#define DC_COM_SPI_CONTROL			0x320
+#define DC_COM_SPI_START_BYTE			0x321
+#define DC_COM_HSPI_WRITE_DATA_AB		0x322
+#define DC_COM_HSPI_WRITE_DATA_CD		0x323
+#define DC_COM_HSPI_CS_DC			0x324
+#define DC_COM_SCRATCH_REGISTER_A		0x325
+#define DC_COM_SCRATCH_REGISTER_B		0x326
+#define DC_COM_GPIO_CTRL			0x327
+#define DC_COM_GPIO_DEBOUNCE_COUNTER		0x328
+#define DC_COM_CRC_CHECKSUM_LATCHED		0x329
+
+#define DC_COM_RG_UNDERFLOW			0x365
+#define  UNDERFLOW_MODE_RED      (1 << 8)
+#define  UNDERFLOW_REPORT_ENABLE (1 << 0)
+
+#define DC_DISP_DISP_SIGNAL_OPTIONS0		0x400
+#define H_PULSE0_ENABLE (1 <<  8)
+#define H_PULSE1_ENABLE (1 << 10)
+#define H_PULSE2_ENABLE (1 << 12)
+
+#define DC_DISP_DISP_SIGNAL_OPTIONS1		0x401
+
+#define DC_DISP_DISP_WIN_OPTIONS		0x402
+#define HDMI_ENABLE	(1 << 30)
+#define DSI_ENABLE	(1 << 29)
+#define SOR1_TIMING_CYA	(1 << 27)
+#define CURSOR_ENABLE	(1 << 16)
+
+#define SOR_ENABLE(x)	(1 << (25 + (((x) > 1) ? ((x) + 1) : (x))))
+
+#define DC_DISP_DISP_MEM_HIGH_PRIORITY		0x403
+#define CURSOR_THRESHOLD(x)   (((x) & 0x03) << 24)
+#define WINDOW_A_THRESHOLD(x) (((x) & 0x7f) << 16)
+#define WINDOW_B_THRESHOLD(x) (((x) & 0x7f) <<  8)
+#define WINDOW_C_THRESHOLD(x) (((x) & 0xff) <<  0)
+
+#define DC_DISP_DISP_MEM_HIGH_PRIORITY_TIMER	0x404
+#define CURSOR_DELAY(x)   (((x) & 0x3f) << 24)
+#define WINDOW_A_DELAY(x) (((x) & 0x3f) << 16)
+#define WINDOW_B_DELAY(x) (((x) & 0x3f) <<  8)
+#define WINDOW_C_DELAY(x) (((x) & 0x3f) <<  0)
+
+#define DC_DISP_DISP_TIMING_OPTIONS		0x405
+#define VSYNC_H_POSITION(x) ((x) & 0xfff)
+
+#define DC_DISP_REF_TO_SYNC			0x406
+#define DC_DISP_SYNC_WIDTH			0x407
+#define DC_DISP_BACK_PORCH			0x408
+#define DC_DISP_ACTIVE				0x409
+#define DC_DISP_FRONT_PORCH			0x40a
+#define DC_DISP_H_PULSE0_CONTROL		0x40b
+#define DC_DISP_H_PULSE0_POSITION_A		0x40c
+#define DC_DISP_H_PULSE0_POSITION_B		0x40d
+#define DC_DISP_H_PULSE0_POSITION_C		0x40e
+#define DC_DISP_H_PULSE0_POSITION_D		0x40f
+#define DC_DISP_H_PULSE1_CONTROL		0x410
+#define DC_DISP_H_PULSE1_POSITION_A		0x411
+#define DC_DISP_H_PULSE1_POSITION_B		0x412
+#define DC_DISP_H_PULSE1_POSITION_C		0x413
+#define DC_DISP_H_PULSE1_POSITION_D		0x414
+#define DC_DISP_H_PULSE2_CONTROL		0x415
+#define DC_DISP_H_PULSE2_POSITION_A		0x416
+#define DC_DISP_H_PULSE2_POSITION_B		0x417
+#define DC_DISP_H_PULSE2_POSITION_C		0x418
+#define DC_DISP_H_PULSE2_POSITION_D		0x419
+#define DC_DISP_V_PULSE0_CONTROL		0x41a
+#define DC_DISP_V_PULSE0_POSITION_A		0x41b
+#define DC_DISP_V_PULSE0_POSITION_B		0x41c
+#define DC_DISP_V_PULSE0_POSITION_C		0x41d
+#define DC_DISP_V_PULSE1_CONTROL		0x41e
+#define DC_DISP_V_PULSE1_POSITION_A		0x41f
+#define DC_DISP_V_PULSE1_POSITION_B		0x420
+#define DC_DISP_V_PULSE1_POSITION_C		0x421
+#define DC_DISP_V_PULSE2_CONTROL		0x422
+#define DC_DISP_V_PULSE2_POSITION_A		0x423
+#define DC_DISP_V_PULSE3_CONTROL		0x424
+#define DC_DISP_V_PULSE3_POSITION_A		0x425
+#define DC_DISP_M0_CONTROL			0x426
+#define DC_DISP_M1_CONTROL			0x427
+#define DC_DISP_DI_CONTROL			0x428
+#define DC_DISP_PP_CONTROL			0x429
+#define DC_DISP_PP_SELECT_A			0x42a
+#define DC_DISP_PP_SELECT_B			0x42b
+#define DC_DISP_PP_SELECT_C			0x42c
+#define DC_DISP_PP_SELECT_D			0x42d
+
+#define PULSE_MODE_NORMAL    (0 << 3)
+#define PULSE_MODE_ONE_CLOCK (1 << 3)
+#define PULSE_POLARITY_HIGH  (0 << 4)
+#define PULSE_POLARITY_LOW   (1 << 4)
+#define PULSE_QUAL_ALWAYS    (0 << 6)
+#define PULSE_QUAL_VACTIVE   (2 << 6)
+#define PULSE_QUAL_VACTIVE1  (3 << 6)
+#define PULSE_LAST_START_A   (0 << 8)
+#define PULSE_LAST_END_A     (1 << 8)
+#define PULSE_LAST_START_B   (2 << 8)
+#define PULSE_LAST_END_B     (3 << 8)
+#define PULSE_LAST_START_C   (4 << 8)
+#define PULSE_LAST_END_C     (5 << 8)
+#define PULSE_LAST_START_D   (6 << 8)
+#define PULSE_LAST_END_D     (7 << 8)
+
+#define PULSE_START(x) (((x) & 0xfff) <<  0)
+#define PULSE_END(x)   (((x) & 0xfff) << 16)
+
+#define DC_DISP_DISP_CLOCK_CONTROL		0x42e
+#define PIXEL_CLK_DIVIDER_PCD1  (0 << 8)
+#define PIXEL_CLK_DIVIDER_PCD1H (1 << 8)
+#define PIXEL_CLK_DIVIDER_PCD2  (2 << 8)
+#define PIXEL_CLK_DIVIDER_PCD3  (3 << 8)
+#define PIXEL_CLK_DIVIDER_PCD4  (4 << 8)
+#define PIXEL_CLK_DIVIDER_PCD6  (5 << 8)
+#define PIXEL_CLK_DIVIDER_PCD8  (6 << 8)
+#define PIXEL_CLK_DIVIDER_PCD9  (7 << 8)
+#define PIXEL_CLK_DIVIDER_PCD12 (8 << 8)
+#define PIXEL_CLK_DIVIDER_PCD16 (9 << 8)
+#define PIXEL_CLK_DIVIDER_PCD18 (10 << 8)
+#define PIXEL_CLK_DIVIDER_PCD24 (11 << 8)
+#define PIXEL_CLK_DIVIDER_PCD13 (12 << 8)
+#define SHIFT_CLK_DIVIDER(x)    ((x) & 0xff)
+
+#define DC_DISP_DISP_INTERFACE_CONTROL		0x42f
+#define DISP_DATA_FORMAT_DF1P1C    (0 << 0)
+#define DISP_DATA_FORMAT_DF1P2C24B (1 << 0)
+#define DISP_DATA_FORMAT_DF1P2C18B (2 << 0)
+#define DISP_DATA_FORMAT_DF1P2C16B (3 << 0)
+#define DISP_DATA_FORMAT_DF2S      (4 << 0)
+#define DISP_DATA_FORMAT_DF3S      (5 << 0)
+#define DISP_DATA_FORMAT_DFSPI     (6 << 0)
+#define DISP_DATA_FORMAT_DF1P3C24B (7 << 0)
+#define DISP_DATA_FORMAT_DF1P3C18B (8 << 0)
+#define DISP_ALIGNMENT_MSB         (0 << 8)
+#define DISP_ALIGNMENT_LSB         (1 << 8)
+#define DISP_ORDER_RED_BLUE        (0 << 9)
+#define DISP_ORDER_BLUE_RED        (1 << 9)
+
+#define DC_DISP_DISP_COLOR_CONTROL		0x430
+#define BASE_COLOR_SIZE666     ( 0 << 0)
+#define BASE_COLOR_SIZE111     ( 1 << 0)
+#define BASE_COLOR_SIZE222     ( 2 << 0)
+#define BASE_COLOR_SIZE333     ( 3 << 0)
+#define BASE_COLOR_SIZE444     ( 4 << 0)
+#define BASE_COLOR_SIZE555     ( 5 << 0)
+#define BASE_COLOR_SIZE565     ( 6 << 0)
+#define BASE_COLOR_SIZE332     ( 7 << 0)
+#define BASE_COLOR_SIZE888     ( 8 << 0)
+#define BASE_COLOR_SIZE101010  (10 << 0)
+#define BASE_COLOR_SIZE121212  (12 << 0)
+#define DITHER_CONTROL_MASK    (3 << 8)
+#define DITHER_CONTROL_DISABLE (0 << 8)
+#define DITHER_CONTROL_ORDERED (2 << 8)
+#define DITHER_CONTROL_ERRDIFF (3 << 8)
+#define BASE_COLOR_SIZE_MASK   (0xf << 0)
+#define BASE_COLOR_SIZE_666    (  0 << 0)
+#define BASE_COLOR_SIZE_111    (  1 << 0)
+#define BASE_COLOR_SIZE_222    (  2 << 0)
+#define BASE_COLOR_SIZE_333    (  3 << 0)
+#define BASE_COLOR_SIZE_444    (  4 << 0)
+#define BASE_COLOR_SIZE_555    (  5 << 0)
+#define BASE_COLOR_SIZE_565    (  6 << 0)
+#define BASE_COLOR_SIZE_332    (  7 << 0)
+#define BASE_COLOR_SIZE_888    (  8 << 0)
+#define BASE_COLOR_SIZE_101010 ( 10 << 0)
+#define BASE_COLOR_SIZE_121212 ( 12 << 0)
+#define CMU_ENABLE_ENABLE      (1 << 20)
+
+#define DC_DISP_SHIFT_CLOCK_OPTIONS		0x431
+#define  SC1_H_QUALIFIER_NONE	(1 << 16)
+#define  SC0_H_QUALIFIER_NONE	(1 <<  0)
+
+#define DC_DISP_DATA_ENABLE_OPTIONS		0x432
+#define DE_SELECT_ACTIVE_BLANK  (0 << 0)
+#define DE_SELECT_ACTIVE        (1 << 0)
+#define DE_SELECT_ACTIVE_IS     (2 << 0)
+#define DE_CONTROL_ONECLK       (0 << 2)
+#define DE_CONTROL_NORMAL       (1 << 2)
+#define DE_CONTROL_EARLY_EXT    (2 << 2)
+#define DE_CONTROL_EARLY        (3 << 2)
+#define DE_CONTROL_ACTIVE_BLANK (4 << 2)
+
+#define DC_DISP_SERIAL_INTERFACE_OPTIONS	0x433
+#define DC_DISP_LCD_SPI_OPTIONS			0x434
+#define DC_DISP_BORDER_COLOR			0x435
+#define DC_DISP_COLOR_KEY0_LOWER		0x436
+#define DC_DISP_COLOR_KEY0_UPPER		0x437
+#define DC_DISP_COLOR_KEY1_LOWER		0x438
+#define DC_DISP_COLOR_KEY1_UPPER		0x439
+
+#define DC_DISP_CURSOR_FOREGROUND		0x43c
+#define DC_DISP_CURSOR_BACKGROUND		0x43d
+
+#define DC_DISP_CURSOR_START_ADDR		0x43e
+#define CURSOR_CLIP_DISPLAY	(0 << 28)
+#define CURSOR_CLIP_WIN_A	(1 << 28)
+#define CURSOR_CLIP_WIN_B	(2 << 28)
+#define CURSOR_CLIP_WIN_C	(3 << 28)
+#define CURSOR_SIZE_32x32	(0 << 24)
+#define CURSOR_SIZE_64x64	(1 << 24)
+#define CURSOR_SIZE_128x128	(2 << 24)
+#define CURSOR_SIZE_256x256	(3 << 24)
+#define DC_DISP_CURSOR_START_ADDR_NS		0x43f
+
+#define DC_DISP_CURSOR_POSITION			0x440
+#define DC_DISP_CURSOR_POSITION_NS		0x441
+
+#define DC_DISP_INIT_SEQ_CONTROL		0x442
+#define DC_DISP_SPI_INIT_SEQ_DATA_A		0x443
+#define DC_DISP_SPI_INIT_SEQ_DATA_B		0x444
+#define DC_DISP_SPI_INIT_SEQ_DATA_C		0x445
+#define DC_DISP_SPI_INIT_SEQ_DATA_D		0x446
+
+#define DC_DISP_DC_MCCIF_FIFOCTRL		0x480
+#define DC_DISP_MCCIF_DISPLAY0A_HYST		0x481
+#define DC_DISP_MCCIF_DISPLAY0B_HYST		0x482
+#define DC_DISP_MCCIF_DISPLAY1A_HYST		0x483
+#define DC_DISP_MCCIF_DISPLAY1B_HYST		0x484
+
+#define DC_DISP_DAC_CRT_CTRL			0x4c0
+#define DC_DISP_DISP_MISC_CONTROL		0x4c1
+#define DC_DISP_SD_CONTROL			0x4c2
+#define DC_DISP_SD_CSC_COEFF			0x4c3
+#define DC_DISP_SD_LUT(x)			(0x4c4 + (x))
+#define DC_DISP_SD_FLICKER_CONTROL		0x4cd
+#define DC_DISP_DC_PIXEL_COUNT			0x4ce
+#define DC_DISP_SD_HISTOGRAM(x)			(0x4cf + (x))
+#define DC_DISP_SD_BL_PARAMETERS		0x4d7
+#define DC_DISP_SD_BL_TF(x)			(0x4d8 + (x))
+#define DC_DISP_SD_BL_CONTROL			0x4dc
+#define DC_DISP_SD_HW_K_VALUES			0x4dd
+#define DC_DISP_SD_MAN_K_VALUES			0x4de
+
+#define DC_DISP_BLEND_BACKGROUND_COLOR		0x4e4
+#define  BACKGROUND_COLOR_ALPHA(x) (((x) & 0xff) << 24)
+#define  BACKGROUND_COLOR_BLUE(x)  (((x) & 0xff) << 16)
+#define  BACKGROUND_COLOR_GREEN(x) (((x) & 0xff) << 8)
+#define  BACKGROUND_COLOR_RED(x)   (((x) & 0xff) << 0)
+
+#define DC_DISP_INTERLACE_CONTROL		0x4e5
+#define  INTERLACE_STATUS (1 << 2)
+#define  INTERLACE_START  (1 << 1)
+#define  INTERLACE_ENABLE (1 << 0)
+
+#define DC_DISP_CURSOR_START_ADDR_HI		0x4ec
+#define DC_DISP_BLEND_CURSOR_CONTROL		0x4f1
+#define CURSOR_COMPOSITION_MODE_BLEND		(0 << 25)
+#define CURSOR_COMPOSITION_MODE_XOR		(1 << 25)
+#define CURSOR_MODE_LEGACY			(0 << 24)
+#define CURSOR_MODE_NORMAL			(1 << 24)
+#define CURSOR_DST_BLEND_ZERO			(0 << 16)
+#define CURSOR_DST_BLEND_K1			(1 << 16)
+#define CURSOR_DST_BLEND_NEG_K1_TIMES_SRC	(2 << 16)
+#define CURSOR_DST_BLEND_MASK			(3 << 16)
+#define CURSOR_SRC_BLEND_K1			(0 << 8)
+#define CURSOR_SRC_BLEND_K1_TIMES_SRC		(1 << 8)
+#define CURSOR_SRC_BLEND_MASK			(3 << 8)
+#define CURSOR_ALPHA				0xff
+
+#define DC_WIN_CORE_ACT_CONTROL 0x50e
+#define  VCOUNTER (0 << 0)
+#define  HCOUNTER (1 << 0)
+
+#define DC_WIN_CORE_IHUB_WGRP_LATENCY_CTLA 0x543
+#define  LATENCY_CTL_MODE_ENABLE (1 << 2)
+
+#define DC_WIN_CORE_IHUB_WGRP_LATENCY_CTLB 0x544
+#define  WATERMARK_MASK 0x1fffffff
+
+#define DC_WIN_CORE_PRECOMP_WGRP_PIPE_METER 0x560
+#define  PIPE_METER_INT(x)  (((x) & 0xff) << 8)
+#define  PIPE_METER_FRAC(x) (((x) & 0xff) << 0)
+
+#define DC_WIN_CORE_IHUB_WGRP_POOL_CONFIG 0x561
+#define  MEMPOOL_ENTRIES(x) (((x) & 0xffff) << 0)
+
+#define DC_WIN_CORE_IHUB_WGRP_FETCH_METER 0x562
+#define  SLOTS(x) (((x) & 0xff) << 0)
+
+#define DC_WIN_CORE_IHUB_LINEBUF_CONFIG 0x563
+#define  MODE_TWO_LINES  (0 << 14)
+#define  MODE_FOUR_LINES (1 << 14)
+
+#define DC_WIN_CORE_IHUB_THREAD_GROUP 0x568
+#define  THREAD_NUM_MASK (0x1f << 1)
+#define  THREAD_NUM(x) (((x) & 0x1f) << 1)
+#define  THREAD_GROUP_ENABLE (1 << 0)
+
+#define DC_WIN_H_FILTER_P(p)			(0x601 + (p))
+#define DC_WIN_V_FILTER_P(p)			(0x619 + (p))
+
+#define DC_WIN_CSC_YOF				0x611
+#define DC_WIN_CSC_KYRGB			0x612
+#define DC_WIN_CSC_KUR				0x613
+#define DC_WIN_CSC_KVR				0x614
+#define DC_WIN_CSC_KUG				0x615
+#define DC_WIN_CSC_KVG				0x616
+#define DC_WIN_CSC_KUB				0x617
+#define DC_WIN_CSC_KVB				0x618
+
+#define DC_WIN_WIN_OPTIONS			0x700
+#define H_DIRECTION  (1 <<  0)
+#define V_DIRECTION  (1 <<  2)
+#define COLOR_EXPAND (1 <<  6)
+#define H_FILTER     (1 <<  8)
+#define V_FILTER     (1 << 10)
+#define CSC_ENABLE   (1 << 18)
+#define WIN_ENABLE   (1 << 30)
+
+#define DC_WIN_BYTE_SWAP			0x701
+#define BYTE_SWAP_NOSWAP  (0 << 0)
+#define BYTE_SWAP_SWAP2   (1 << 0)
+#define BYTE_SWAP_SWAP4   (2 << 0)
+#define BYTE_SWAP_SWAP4HW (3 << 0)
+
+#define DC_WIN_BUFFER_CONTROL			0x702
+#define BUFFER_CONTROL_HOST  (0 << 0)
+#define BUFFER_CONTROL_VI    (1 << 0)
+#define BUFFER_CONTROL_EPP   (2 << 0)
+#define BUFFER_CONTROL_MPEGE (3 << 0)
+#define BUFFER_CONTROL_SB2D  (4 << 0)
+
+#define DC_WIN_COLOR_DEPTH			0x703
+#define WIN_COLOR_DEPTH_P1              0
+#define WIN_COLOR_DEPTH_P2              1
+#define WIN_COLOR_DEPTH_P4              2
+#define WIN_COLOR_DEPTH_P8              3
+#define WIN_COLOR_DEPTH_B4G4R4A4        4
+#define WIN_COLOR_DEPTH_B5G5R5A1        5
+#define WIN_COLOR_DEPTH_B5G6R5          6
+#define WIN_COLOR_DEPTH_A1B5G5R5        7
+#define WIN_COLOR_DEPTH_B8G8R8A8       12
+#define WIN_COLOR_DEPTH_R8G8B8A8       13
+#define WIN_COLOR_DEPTH_B6x2G6x2R6x2A8 14
+#define WIN_COLOR_DEPTH_R6x2G6x2B6x2A8 15
+#define WIN_COLOR_DEPTH_YCbCr422       16
+#define WIN_COLOR_DEPTH_YUV422         17
+#define WIN_COLOR_DEPTH_YCbCr420P      18
+#define WIN_COLOR_DEPTH_YUV420P        19
+#define WIN_COLOR_DEPTH_YCbCr422P      20
+#define WIN_COLOR_DEPTH_YUV422P        21
+#define WIN_COLOR_DEPTH_YCbCr422R      22
+#define WIN_COLOR_DEPTH_YUV422R        23
+#define WIN_COLOR_DEPTH_YCbCr422RA     24
+#define WIN_COLOR_DEPTH_YUV422RA       25
+#define WIN_COLOR_DEPTH_R4G4B4A4       27
+#define WIN_COLOR_DEPTH_R5G5B5A        28
+#define WIN_COLOR_DEPTH_AR5G5B5        29
+#define WIN_COLOR_DEPTH_B5G5R5X1       30
+#define WIN_COLOR_DEPTH_X1B5G5R5       31
+#define WIN_COLOR_DEPTH_R5G5B5X1       32
+#define WIN_COLOR_DEPTH_X1R5G5B5       33
+#define WIN_COLOR_DEPTH_R5G6B5         34
+#define WIN_COLOR_DEPTH_A8R8G8B8       35
+#define WIN_COLOR_DEPTH_A8B8G8R8       36
+#define WIN_COLOR_DEPTH_B8G8R8X8       37
+#define WIN_COLOR_DEPTH_R8G8B8X8       38
+#define WIN_COLOR_DEPTH_YCbCr444P      41
+#define WIN_COLOR_DEPTH_YCrCb420SP     42
+#define WIN_COLOR_DEPTH_YCbCr420SP     43
+#define WIN_COLOR_DEPTH_YCrCb422SP     44
+#define WIN_COLOR_DEPTH_YCbCr422SP     45
+#define WIN_COLOR_DEPTH_YCrCb444SP     48
+#define WIN_COLOR_DEPTH_YCbCr444SP     49
+#define WIN_COLOR_DEPTH_X8B8G8R8       65
+#define WIN_COLOR_DEPTH_X8R8G8B8       66
+
+#define DC_WIN_POSITION				0x704
+#define H_POSITION(x) (((x) & 0x1fff) <<  0) /* XXX 0x7fff on Tegra186 */
+#define V_POSITION(x) (((x) & 0x1fff) << 16) /* XXX 0x7fff on Tegra186 */
+
+#define DC_WIN_SIZE				0x705
+#define H_SIZE(x) (((x) & 0x1fff) <<  0) /* XXX 0x7fff on Tegra186 */
+#define V_SIZE(x) (((x) & 0x1fff) << 16) /* XXX 0x7fff on Tegra186 */
+
+#define DC_WIN_PRESCALED_SIZE			0x706
+#define H_PRESCALED_SIZE(x) (((x) & 0x7fff) <<  0)
+#define V_PRESCALED_SIZE(x) (((x) & 0x1fff) << 16) /* XXX 0x7fff on Tegra186 */
+
+#define DC_WIN_H_INITIAL_DDA			0x707
+#define DC_WIN_V_INITIAL_DDA			0x708
+#define DC_WIN_DDA_INC				0x709
+#define H_DDA_INC(x) (((x) & 0xffff) <<  0)
+#define V_DDA_INC(x) (((x) & 0xffff) << 16)
+
+#define DC_WIN_LINE_STRIDE			0x70a
+#define DC_WIN_BUF_STRIDE			0x70b
+#define DC_WIN_UV_BUF_STRIDE			0x70c
+#define DC_WIN_BUFFER_ADDR_MODE			0x70d
+#define DC_WIN_BUFFER_ADDR_MODE_LINEAR		(0 <<  0)
+#define DC_WIN_BUFFER_ADDR_MODE_TILE		(1 <<  0)
+#define DC_WIN_BUFFER_ADDR_MODE_LINEAR_UV	(0 << 16)
+#define DC_WIN_BUFFER_ADDR_MODE_TILE_UV		(1 << 16)
+
+#define DC_WIN_DV_CONTROL			0x70e
+
+#define DC_WIN_BLEND_NOKEY			0x70f
+#define  BLEND_WEIGHT1(x) (((x) & 0xff) << 16)
+#define  BLEND_WEIGHT0(x) (((x) & 0xff) <<  8)
+
+#define DC_WIN_BLEND_1WIN			0x710
+#define  BLEND_CONTROL_FIX    (0 << 2)
+#define  BLEND_CONTROL_ALPHA  (1 << 2)
+#define  BLEND_COLOR_KEY_NONE (0 << 0)
+#define  BLEND_COLOR_KEY_0    (1 << 0)
+#define  BLEND_COLOR_KEY_1    (2 << 0)
+#define  BLEND_COLOR_KEY_BOTH (3 << 0)
+
+#define DC_WIN_BLEND_2WIN_X			0x711
+#define  BLEND_CONTROL_DEPENDENT (2 << 2)
+
+#define DC_WIN_BLEND_2WIN_Y			0x712
+#define DC_WIN_BLEND_3WIN_XY			0x713
+
+#define DC_WIN_HP_FETCH_CONTROL			0x714
+
+#define DC_WINBUF_START_ADDR			0x800
+#define DC_WINBUF_START_ADDR_NS			0x801
+#define DC_WINBUF_START_ADDR_U			0x802
+#define DC_WINBUF_START_ADDR_U_NS		0x803
+#define DC_WINBUF_START_ADDR_V			0x804
+#define DC_WINBUF_START_ADDR_V_NS		0x805
+
+#define DC_WINBUF_ADDR_H_OFFSET			0x806
+#define DC_WINBUF_ADDR_H_OFFSET_NS		0x807
+#define DC_WINBUF_ADDR_V_OFFSET			0x808
+#define DC_WINBUF_ADDR_V_OFFSET_NS		0x809
+
+#define DC_WINBUF_UFLOW_STATUS			0x80a
+#define DC_WINBUF_SURFACE_KIND			0x80b
+#define DC_WINBUF_SURFACE_KIND_PITCH	(0 << 0)
+#define DC_WINBUF_SURFACE_KIND_TILED	(1 << 0)
+#define DC_WINBUF_SURFACE_KIND_BLOCK	(2 << 0)
+#define DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(x) (((x) & 0x7) << 4)
+
+#define DC_WINBUF_START_ADDR_HI			0x80d
+
+#define DC_WINBUF_START_ADDR_HI_U		0x80f
+#define DC_WINBUF_START_ADDR_HI_V		0x811
+
+#define DC_WINBUF_CDE_CONTROL			0x82f
+#define  ENABLE_SURFACE (1 << 0)
+
+#define DC_WINBUF_AD_UFLOW_STATUS		0xbca
+#define DC_WINBUF_BD_UFLOW_STATUS		0xdca
+#define DC_WINBUF_CD_UFLOW_STATUS		0xfca
+
+/* Tegra186 and later */
+#define DC_DISP_CORE_SOR_SET_CONTROL(x)		(0x403 + (x))
+#define PROTOCOL_MASK (0xf << 8)
+#define PROTOCOL_SINGLE_TMDS_A (0x1 << 8)
+
+#define DC_DISP_CORE_HEAD_SET_CONTROL_OUTPUT_LUT	0x431
+#define  OUTPUT_LUT_MODE_MASK        (3 << 5)
+#define  OUTPUT_LUT_MODE_INTERPOLATE (1 << 5)
+#define  OUTPUT_LUT_SIZE_MASK        (3 << 1)
+#define  OUTPUT_LUT_SIZE_SIZE_1025   (2 << 1)
+
+#define DC_DISP_COREPVT_HEAD_SET_OUTPUT_LUT_BASE	0x432
+#define DC_DISP_COREPVT_HEAD_SET_OUTPUT_LUT_BASE_HI	0x433
+
+#define DC_DISP_PCALC_HEAD_SET_CROPPED_POINT_IN_CURSOR	0x442
+#define DC_DISP_PCALC_HEAD_SET_CROPPED_SIZE_IN_CURSOR	0x446
+
+#define DC_WINC_PRECOMP_WGRP_PIPE_CAPA 0x500
+#define DC_WINC_PRECOMP_WGRP_PIPE_CAPB 0x501
+#define DC_WINC_PRECOMP_WGRP_PIPE_CAPC 0x502
+#define  MAX_PIXELS_5TAP444(x) ((x) & 0xffff)
+#define DC_WINC_PRECOMP_WGRP_PIPE_CAPD 0x503
+#define DC_WINC_PRECOMP_WGRP_PIPE_CAPE 0x504
+#define  MAX_PIXELS_2TAP444(x) ((x) & 0xffff)
+#define DC_WINC_PRECOMP_WGRP_PIPE_CAPF 0x505
+
+#define DC_WIN_CORE_WINDOWGROUP_SET_CONTROL	0x702
+#define OWNER_MASK (0xf << 0)
+#define OWNER(x) (((x) & 0xf) << 0)
+
+#define DC_WIN_CROPPED_SIZE			0x706
+
+#define DC_WIN_SET_INPUT_SCALER_H_START_PHASE	0x707
+#define DC_WIN_SET_INPUT_SCALER_V_START_PHASE	0x708
+
+#define DC_WIN_PLANAR_STORAGE			0x709
+#define PITCH(x) (((x) >> 6) & 0x1fff)
+
+#define DC_WIN_PLANAR_STORAGE_UV		0x70a
+#define  PITCH_U(x) ((((x) >> 6) & 0x1fff) <<  0)
+#define  PITCH_V(x) ((((x) >> 6) & 0x1fff) << 16)
+
+#define DC_WIN_SET_INPUT_SCALER_HPHASE_INCR	0x70b
+#define DC_WIN_SET_INPUT_SCALER_VPHASE_INCR	0x70c
+
+#define DC_WIN_SET_PARAMS			0x70d
+#define  CLAMP_BEFORE_BLEND (1 << 15)
+#define  DEGAMMA_NONE (0 << 13)
+#define  DEGAMMA_SRGB (1 << 13)
+#define  DEGAMMA_YUV8_10 (2 << 13)
+#define  DEGAMMA_YUV12 (3 << 13)
+#define  INPUT_RANGE_BYPASS (0 << 10)
+#define  INPUT_RANGE_LIMITED (1 << 10)
+#define  INPUT_RANGE_FULL (2 << 10)
+#define  COLOR_SPACE_RGB (0 << 8)
+#define  COLOR_SPACE_YUV_601 (1 << 8)
+#define  COLOR_SPACE_YUV_709 (2 << 8)
+#define  COLOR_SPACE_YUV_2020 (3 << 8)
+
+#define DC_WIN_WINDOWGROUP_SET_CONTROL_INPUT_SCALER	0x70e
+#define  HORIZONTAL_TAPS_2 (1 << 3)
+#define  HORIZONTAL_TAPS_5 (4 << 3)
+#define  VERTICAL_TAPS_2 (1 << 0)
+#define  VERTICAL_TAPS_5 (4 << 0)
+
+#define DC_WIN_WINDOWGROUP_SET_INPUT_SCALER_COEFF 0x70f
+#define  COEFF_INDEX(x) (((x) & 0xff) << 15)
+#define  COEFF_DATA(x) (((x) & 0x3ff) << 0)
+
+#define DC_WIN_WINDOWGROUP_SET_INPUT_SCALER_USAGE	0x711
+#define  INPUT_SCALER_USE422  (1 << 2)
+#define  INPUT_SCALER_VBYPASS (1 << 1)
+#define  INPUT_SCALER_HBYPASS (1 << 0)
+
+#define DC_WIN_BLEND_LAYER_CONTROL		0x716
+#define  COLOR_KEY_NONE (0 << 25)
+#define  COLOR_KEY_SRC (1 << 25)
+#define  COLOR_KEY_DST (2 << 25)
+#define  BLEND_BYPASS (1 << 24)
+#define  K2(x) (((x) & 0xff) << 16)
+#define  K1(x) (((x) & 0xff) << 8)
+#define  WINDOW_LAYER_DEPTH(x) (((x) & 0xff) << 0)
+
+#define DC_WIN_BLEND_MATCH_SELECT		0x717
+#define  BLEND_FACTOR_DST_ALPHA_ZERO			(0 << 12)
+#define  BLEND_FACTOR_DST_ALPHA_ONE			(1 << 12)
+#define  BLEND_FACTOR_DST_ALPHA_NEG_K1_TIMES_SRC	(2 << 12)
+#define  BLEND_FACTOR_DST_ALPHA_K2			(3 << 12)
+#define  BLEND_FACTOR_SRC_ALPHA_ZERO			(0 << 8)
+#define  BLEND_FACTOR_SRC_ALPHA_K1			(1 << 8)
+#define  BLEND_FACTOR_SRC_ALPHA_K2			(2 << 8)
+#define  BLEND_FACTOR_SRC_ALPHA_NEG_K1_TIMES_DST	(3 << 8)
+#define  BLEND_FACTOR_DST_COLOR_ZERO			(0 << 4)
+#define  BLEND_FACTOR_DST_COLOR_ONE			(1 << 4)
+#define  BLEND_FACTOR_DST_COLOR_K1			(2 << 4)
+#define  BLEND_FACTOR_DST_COLOR_K2			(3 << 4)
+#define  BLEND_FACTOR_DST_COLOR_K1_TIMES_DST		(4 << 4)
+#define  BLEND_FACTOR_DST_COLOR_NEG_K1_TIMES_DST	(5 << 4)
+#define  BLEND_FACTOR_DST_COLOR_NEG_K1_TIMES_SRC	(6 << 4)
+#define  BLEND_FACTOR_DST_COLOR_NEG_K1			(7 << 4)
+#define  BLEND_FACTOR_SRC_COLOR_ZERO			(0 << 0)
+#define  BLEND_FACTOR_SRC_COLOR_ONE			(1 << 0)
+#define  BLEND_FACTOR_SRC_COLOR_K1			(2 << 0)
+#define  BLEND_FACTOR_SRC_COLOR_K1_TIMES_DST		(3 << 0)
+#define  BLEND_FACTOR_SRC_COLOR_NEG_K1_TIMES_DST	(4 << 0)
+#define  BLEND_FACTOR_SRC_COLOR_K1_TIMES_SRC		(5 << 0)
+
+#define DC_WIN_BLEND_NOMATCH_SELECT		0x718
+
+#define DC_WIN_PRECOMP_WGRP_PARAMS		0x724
+#define  SWAP_UV (1 << 0)
+
+#define DC_WIN_WINDOW_SET_CONTROL		0x730
+#define  CONTROL_CSC_ENABLE (1 << 5)
+
+#define DC_WINBUF_CROPPED_POINT			0x806
+#define OFFSET_Y(x) (((x) & 0xffff) << 16)
+#define OFFSET_X(x) (((x) & 0xffff) << 0)
+
+#endif /* TEGRA_DC_H */
diff --git a/drivers/gpu/drm/tegra/dp.c b/drivers/gpu/drm/tegra/dp.c
new file mode 100644
index 000000000..3cb194f5c
--- /dev/null
+++ b/drivers/gpu/drm/tegra/dp.c
@@ -0,0 +1,820 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright (C) 2013-2019 NVIDIA Corporation
+ * Copyright (C) 2015 Rob Clark
+ */
+
+#include <linux/string_choices.h>
+
+#include <drm/display/drm_dp_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_print.h>
+
+#include "dp.h"
+
+static const u8 drm_dp_edp_revisions[] = { 0x11, 0x12, 0x13, 0x14 };
+
+static void drm_dp_link_caps_reset(struct drm_dp_link_caps *caps)
+{
+	caps->enhanced_framing = false;
+	caps->tps3_supported = false;
+	caps->fast_training = false;
+	caps->channel_coding = false;
+	caps->alternate_scrambler_reset = false;
+}
+
+void drm_dp_link_caps_copy(struct drm_dp_link_caps *dest,
+			   const struct drm_dp_link_caps *src)
+{
+	dest->enhanced_framing = src->enhanced_framing;
+	dest->tps3_supported = src->tps3_supported;
+	dest->fast_training = src->fast_training;
+	dest->channel_coding = src->channel_coding;
+	dest->alternate_scrambler_reset = src->alternate_scrambler_reset;
+}
+
+static void drm_dp_link_reset(struct drm_dp_link *link)
+{
+	unsigned int i;
+
+	if (!link)
+		return;
+
+	link->revision = 0;
+	link->max_rate = 0;
+	link->max_lanes = 0;
+
+	drm_dp_link_caps_reset(&link->caps);
+	link->aux_rd_interval.cr = 0;
+	link->aux_rd_interval.ce = 0;
+	link->edp = 0;
+
+	link->rate = 0;
+	link->lanes = 0;
+
+	for (i = 0; i < DP_MAX_SUPPORTED_RATES; i++)
+		link->rates[i] = 0;
+
+	link->num_rates = 0;
+}
+
+/**
+ * drm_dp_link_add_rate() - add a rate to the list of supported rates
+ * @link: the link to add the rate to
+ * @rate: the rate to add
+ *
+ * Add a link rate to the list of supported link rates.
+ *
+ * Returns:
+ * 0 on success or one of the following negative error codes on failure:
+ * - ENOSPC if the maximum number of supported rates has been reached
+ * - EEXISTS if the link already supports this rate
+ *
+ * See also:
+ * drm_dp_link_remove_rate()
+ */
+int drm_dp_link_add_rate(struct drm_dp_link *link, unsigned long rate)
+{
+	unsigned int i, pivot;
+
+	if (link->num_rates == DP_MAX_SUPPORTED_RATES)
+		return -ENOSPC;
+
+	for (pivot = 0; pivot < link->num_rates; pivot++)
+		if (rate <= link->rates[pivot])
+			break;
+
+	if (pivot != link->num_rates && rate == link->rates[pivot])
+		return -EEXIST;
+
+	for (i = link->num_rates; i > pivot; i--)
+		link->rates[i] = link->rates[i - 1];
+
+	link->rates[pivot] = rate;
+	link->num_rates++;
+
+	return 0;
+}
+
+/**
+ * drm_dp_link_remove_rate() - remove a rate from the list of supported rates
+ * @link: the link from which to remove the rate
+ * @rate: the rate to remove
+ *
+ * Removes a link rate from the list of supported link rates.
+ *
+ * Returns:
+ * 0 on success or one of the following negative error codes on failure:
+ * - EINVAL if the specified rate is not among the supported rates
+ *
+ * See also:
+ * drm_dp_link_add_rate()
+ */
+int drm_dp_link_remove_rate(struct drm_dp_link *link, unsigned long rate)
+{
+	unsigned int i;
+
+	for (i = 0; i < link->num_rates; i++)
+		if (rate == link->rates[i])
+			break;
+
+	if (i == link->num_rates)
+		return -EINVAL;
+
+	link->num_rates--;
+
+	while (i < link->num_rates) {
+		link->rates[i] = link->rates[i + 1];
+		i++;
+	}
+
+	return 0;
+}
+
+/**
+ * drm_dp_link_update_rates() - normalize the supported link rates array
+ * @link: the link for which to normalize the supported link rates
+ *
+ * Users should call this function after they've manually modified the array
+ * of supported link rates. This function removes any stale entries, compacts
+ * the array and updates the supported link rate count. Note that calling the
+ * drm_dp_link_remove_rate() function already does this janitorial work.
+ *
+ * See also:
+ * drm_dp_link_add_rate(), drm_dp_link_remove_rate()
+ */
+void drm_dp_link_update_rates(struct drm_dp_link *link)
+{
+	unsigned int i, count = 0;
+
+	for (i = 0; i < link->num_rates; i++) {
+		if (link->rates[i] != 0)
+			link->rates[count++] = link->rates[i];
+	}
+
+	for (i = count; i < link->num_rates; i++)
+		link->rates[i] = 0;
+
+	link->num_rates = count;
+}
+
+/**
+ * drm_dp_link_probe() - probe a DisplayPort link for capabilities
+ * @aux: DisplayPort AUX channel
+ * @link: pointer to structure in which to return link capabilities
+ *
+ * The structure filled in by this function can usually be passed directly
+ * into drm_dp_link_power_up() and drm_dp_link_configure() to power up and
+ * configure the link based on the link's capabilities.
+ *
+ * Returns 0 on success or a negative error code on failure.
+ */
+int drm_dp_link_probe(struct drm_dp_aux *aux, struct drm_dp_link *link)
+{
+	u8 dpcd[DP_RECEIVER_CAP_SIZE], value;
+	unsigned int rd_interval;
+	int err;
+
+	drm_dp_link_reset(link);
+
+	err = drm_dp_dpcd_read(aux, DP_DPCD_REV, dpcd, sizeof(dpcd));
+	if (err < 0)
+		return err;
+
+	link->revision = dpcd[DP_DPCD_REV];
+	link->max_rate = drm_dp_max_link_rate(dpcd);
+	link->max_lanes = drm_dp_max_lane_count(dpcd);
+
+	link->caps.enhanced_framing = drm_dp_enhanced_frame_cap(dpcd);
+	link->caps.tps3_supported = drm_dp_tps3_supported(dpcd);
+	link->caps.fast_training = drm_dp_fast_training_cap(dpcd);
+	link->caps.channel_coding = drm_dp_channel_coding_supported(dpcd);
+
+	if (drm_dp_alternate_scrambler_reset_cap(dpcd)) {
+		link->caps.alternate_scrambler_reset = true;
+
+		err = drm_dp_dpcd_readb(aux, DP_EDP_DPCD_REV, &value);
+		if (err < 0)
+			return err;
+
+		if (value >= ARRAY_SIZE(drm_dp_edp_revisions))
+			DRM_ERROR("unsupported eDP version: %02x\n", value);
+		else
+			link->edp = drm_dp_edp_revisions[value];
+	}
+
+	/*
+	 * The DPCD stores the AUX read interval in units of 4 ms. There are
+	 * two special cases:
+	 *
+	 *   1) if the TRAINING_AUX_RD_INTERVAL field is 0, the clock recovery
+	 *      and channel equalization should use 100 us or 400 us AUX read
+	 *      intervals, respectively
+	 *
+	 *   2) for DP v1.4 and above, clock recovery should always use 100 us
+	 *      AUX read intervals
+	 */
+	rd_interval = dpcd[DP_TRAINING_AUX_RD_INTERVAL] &
+			   DP_TRAINING_AUX_RD_MASK;
+
+	if (rd_interval > 4) {
+		DRM_DEBUG_KMS("AUX interval %u out of range (max. 4)\n",
+			      rd_interval);
+		rd_interval = 4;
+	}
+
+	rd_interval *= 4 * USEC_PER_MSEC;
+
+	if (rd_interval == 0 || link->revision >= DP_DPCD_REV_14)
+		link->aux_rd_interval.cr = 100;
+
+	if (rd_interval == 0)
+		link->aux_rd_interval.ce = 400;
+
+	link->rate = link->max_rate;
+	link->lanes = link->max_lanes;
+
+	/* Parse SUPPORTED_LINK_RATES from eDP 1.4 */
+	if (link->edp >= 0x14) {
+		u8 supported_rates[DP_MAX_SUPPORTED_RATES * 2];
+		unsigned int i;
+		u16 rate;
+
+		err = drm_dp_dpcd_read(aux, DP_SUPPORTED_LINK_RATES,
+				       supported_rates,
+				       sizeof(supported_rates));
+		if (err < 0)
+			return err;
+
+		for (i = 0; i < DP_MAX_SUPPORTED_RATES; i++) {
+			rate = supported_rates[i * 2 + 1] << 8 |
+			       supported_rates[i * 2 + 0];
+
+			drm_dp_link_add_rate(link, rate * 200);
+		}
+	}
+
+	return 0;
+}
+
+/**
+ * drm_dp_link_configure() - configure a DisplayPort link
+ * @aux: DisplayPort AUX channel
+ * @link: pointer to a structure containing the link configuration
+ *
+ * Returns 0 on success or a negative error code on failure.
+ */
+int drm_dp_link_configure(struct drm_dp_aux *aux, struct drm_dp_link *link)
+{
+	u8 values[2], value;
+	int err;
+
+	if (link->ops && link->ops->configure) {
+		err = link->ops->configure(link);
+		if (err < 0) {
+			DRM_ERROR("failed to configure DP link: %d\n", err);
+			return err;
+		}
+	}
+
+	values[0] = drm_dp_link_rate_to_bw_code(link->rate);
+	values[1] = link->lanes;
+
+	if (link->caps.enhanced_framing)
+		values[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
+
+	err = drm_dp_dpcd_write(aux, DP_LINK_BW_SET, values, sizeof(values));
+	if (err < 0)
+		return err;
+
+	if (link->caps.channel_coding)
+		value = DP_SET_ANSI_8B10B;
+	else
+		value = 0;
+
+	err = drm_dp_dpcd_writeb(aux, DP_MAIN_LINK_CHANNEL_CODING_SET, value);
+	if (err < 0)
+		return err;
+
+	if (link->caps.alternate_scrambler_reset) {
+		err = drm_dp_dpcd_writeb(aux, DP_EDP_CONFIGURATION_SET,
+					 DP_ALTERNATE_SCRAMBLER_RESET_ENABLE);
+		if (err < 0)
+			return err;
+	}
+
+	return 0;
+}
+
+/**
+ * drm_dp_link_choose() - choose the lowest possible configuration for a mode
+ * @link: DRM DP link object
+ * @mode: DRM display mode
+ * @info: DRM display information
+ *
+ * According to the eDP specification, a source should select a configuration
+ * with the lowest number of lanes and the lowest possible link rate that can
+ * match the bitrate requirements of a video mode. However it must ensure not
+ * to exceed the capabilities of the sink.
+ *
+ * Returns: 0 on success or a negative error code on failure.
+ */
+int drm_dp_link_choose(struct drm_dp_link *link,
+		       const struct drm_display_mode *mode,
+		       const struct drm_display_info *info)
+{
+	/* available link symbol clock rates */
+	static const unsigned int rates[3] = { 162000, 270000, 540000 };
+	/* available number of lanes */
+	static const unsigned int lanes[3] = { 1, 2, 4 };
+	unsigned long requirement, capacity;
+	unsigned int rate = link->max_rate;
+	unsigned int i, j;
+
+	/* bandwidth requirement */
+	requirement = mode->clock * info->bpc * 3;
+
+	for (i = 0; i < ARRAY_SIZE(lanes) && lanes[i] <= link->max_lanes; i++) {
+		for (j = 0; j < ARRAY_SIZE(rates) && rates[j] <= rate; j++) {
+			/*
+			 * Capacity for this combination of lanes and rate,
+			 * factoring in the ANSI 8B/10B encoding.
+			 *
+			 * Link rates in the DRM DP helpers are really link
+			 * symbol frequencies, so a tenth of the actual rate
+			 * of the link.
+			 */
+			capacity = lanes[i] * (rates[j] * 10) * 8 / 10;
+
+			if (capacity >= requirement) {
+				DRM_DEBUG_KMS("using %u lanes at %u kHz (%lu/%lu kbps)\n",
+					      lanes[i], rates[j], requirement,
+					      capacity);
+				link->lanes = lanes[i];
+				link->rate = rates[j];
+				return 0;
+			}
+		}
+	}
+
+	return -ERANGE;
+}
+
+/**
+ * DOC: Link training
+ *
+ * These functions contain common logic and helpers to implement DisplayPort
+ * link training.
+ */
+
+/**
+ * drm_dp_link_train_init() - initialize DisplayPort link training state
+ * @train: DisplayPort link training state
+ */
+void drm_dp_link_train_init(struct drm_dp_link_train *train)
+{
+	struct drm_dp_link_train_set *request = &train->request;
+	struct drm_dp_link_train_set *adjust = &train->adjust;
+	unsigned int i;
+
+	for (i = 0; i < 4; i++) {
+		request->voltage_swing[i] = 0;
+		adjust->voltage_swing[i] = 0;
+
+		request->pre_emphasis[i] = 0;
+		adjust->pre_emphasis[i] = 0;
+
+		request->post_cursor[i] = 0;
+		adjust->post_cursor[i] = 0;
+	}
+
+	train->pattern = DP_TRAINING_PATTERN_DISABLE;
+	train->clock_recovered = false;
+	train->channel_equalized = false;
+}
+
+static bool drm_dp_link_train_valid(const struct drm_dp_link_train *train)
+{
+	return train->clock_recovered && train->channel_equalized;
+}
+
+static int drm_dp_link_apply_training(struct drm_dp_link *link)
+{
+	struct drm_dp_link_train_set *request = &link->train.request;
+	unsigned int lanes = link->lanes, *vs, *pe, *pc, i;
+	struct drm_dp_aux *aux = link->aux;
+	u8 values[4], pattern = 0;
+	int err;
+
+	err = link->ops->apply_training(link);
+	if (err < 0) {
+		DRM_ERROR("failed to apply link training: %d\n", err);
+		return err;
+	}
+
+	vs = request->voltage_swing;
+	pe = request->pre_emphasis;
+	pc = request->post_cursor;
+
+	/* write currently selected voltage-swing and pre-emphasis levels */
+	for (i = 0; i < lanes; i++)
+		values[i] = DP_TRAIN_VOLTAGE_SWING_LEVEL(vs[i]) |
+			    DP_TRAIN_PRE_EMPHASIS_LEVEL(pe[i]);
+
+	err = drm_dp_dpcd_write(aux, DP_TRAINING_LANE0_SET, values, lanes);
+	if (err < 0) {
+		DRM_ERROR("failed to set training parameters: %d\n", err);
+		return err;
+	}
+
+	/* write currently selected post-cursor level (if supported) */
+	if (link->revision >= 0x12 && link->rate == 540000) {
+		values[0] = values[1] = 0;
+
+		for (i = 0; i < lanes; i++)
+			values[i / 2] |= DP_LANE_POST_CURSOR(i, pc[i]);
+
+		err = drm_dp_dpcd_write(aux, DP_TRAINING_LANE0_1_SET2, values,
+					DIV_ROUND_UP(lanes, 2));
+		if (err < 0) {
+			DRM_ERROR("failed to set post-cursor: %d\n", err);
+			return err;
+		}
+	}
+
+	/* write link pattern */
+	if (link->train.pattern != DP_TRAINING_PATTERN_DISABLE)
+		pattern |= DP_LINK_SCRAMBLING_DISABLE;
+
+	pattern |= link->train.pattern;
+
+	err = drm_dp_dpcd_writeb(aux, DP_TRAINING_PATTERN_SET, pattern);
+	if (err < 0) {
+		DRM_ERROR("failed to set training pattern: %d\n", err);
+		return err;
+	}
+
+	return 0;
+}
+
+static void drm_dp_link_train_wait(struct drm_dp_link *link)
+{
+	unsigned long min = 0;
+
+	switch (link->train.pattern) {
+	case DP_TRAINING_PATTERN_1:
+		min = link->aux_rd_interval.cr;
+		break;
+
+	case DP_TRAINING_PATTERN_2:
+	case DP_TRAINING_PATTERN_3:
+		min = link->aux_rd_interval.ce;
+		break;
+
+	default:
+		break;
+	}
+
+	if (min > 0)
+		usleep_range(min, 2 * min);
+}
+
+static void drm_dp_link_get_adjustments(struct drm_dp_link *link,
+					u8 status[DP_LINK_STATUS_SIZE])
+{
+	struct drm_dp_link_train_set *adjust = &link->train.adjust;
+	unsigned int i;
+	u8 post_cursor;
+	int err;
+
+	err = drm_dp_dpcd_read(link->aux, DP_ADJUST_REQUEST_POST_CURSOR2,
+			       &post_cursor, sizeof(post_cursor));
+	if (err < 0) {
+		DRM_ERROR("failed to read post_cursor2: %d\n", err);
+		post_cursor = 0;
+	}
+
+	for (i = 0; i < link->lanes; i++) {
+		adjust->voltage_swing[i] =
+			drm_dp_get_adjust_request_voltage(status, i) >>
+				DP_TRAIN_VOLTAGE_SWING_SHIFT;
+
+		adjust->pre_emphasis[i] =
+			drm_dp_get_adjust_request_pre_emphasis(status, i) >>
+				DP_TRAIN_PRE_EMPHASIS_SHIFT;
+
+		adjust->post_cursor[i] =
+			(post_cursor >> (i << 1)) & 0x3;
+	}
+}
+
+static void drm_dp_link_train_adjust(struct drm_dp_link_train *train)
+{
+	struct drm_dp_link_train_set *request = &train->request;
+	struct drm_dp_link_train_set *adjust = &train->adjust;
+	unsigned int i;
+
+	for (i = 0; i < 4; i++)
+		if (request->voltage_swing[i] != adjust->voltage_swing[i])
+			request->voltage_swing[i] = adjust->voltage_swing[i];
+
+	for (i = 0; i < 4; i++)
+		if (request->pre_emphasis[i] != adjust->pre_emphasis[i])
+			request->pre_emphasis[i] = adjust->pre_emphasis[i];
+
+	for (i = 0; i < 4; i++)
+		if (request->post_cursor[i] != adjust->post_cursor[i])
+			request->post_cursor[i] = adjust->post_cursor[i];
+}
+
+static int drm_dp_link_recover_clock(struct drm_dp_link *link)
+{
+	u8 status[DP_LINK_STATUS_SIZE];
+	int err;
+
+	err = drm_dp_link_apply_training(link);
+	if (err < 0)
+		return err;
+
+	drm_dp_link_train_wait(link);
+
+	err = drm_dp_dpcd_read_link_status(link->aux, status);
+	if (err < 0) {
+		DRM_ERROR("failed to read link status: %d\n", err);
+		return err;
+	}
+
+	if (!drm_dp_clock_recovery_ok(status, link->lanes))
+		drm_dp_link_get_adjustments(link, status);
+	else
+		link->train.clock_recovered = true;
+
+	return 0;
+}
+
+static int drm_dp_link_clock_recovery(struct drm_dp_link *link)
+{
+	unsigned int repeat;
+	int err;
+
+	/* start clock recovery using training pattern 1 */
+	link->train.pattern = DP_TRAINING_PATTERN_1;
+
+	for (repeat = 1; repeat < 5; repeat++) {
+		err = drm_dp_link_recover_clock(link);
+		if (err < 0) {
+			DRM_ERROR("failed to recover clock: %d\n", err);
+			return err;
+		}
+
+		if (link->train.clock_recovered)
+			break;
+
+		drm_dp_link_train_adjust(&link->train);
+	}
+
+	return 0;
+}
+
+static int drm_dp_link_equalize_channel(struct drm_dp_link *link)
+{
+	struct drm_dp_aux *aux = link->aux;
+	u8 status[DP_LINK_STATUS_SIZE];
+	int err;
+
+	err = drm_dp_link_apply_training(link);
+	if (err < 0)
+		return err;
+
+	drm_dp_link_train_wait(link);
+
+	err = drm_dp_dpcd_read_link_status(aux, status);
+	if (err < 0) {
+		DRM_ERROR("failed to read link status: %d\n", err);
+		return err;
+	}
+
+	if (!drm_dp_clock_recovery_ok(status, link->lanes)) {
+		DRM_ERROR("clock recovery lost while equalizing channel\n");
+		link->train.clock_recovered = false;
+		return 0;
+	}
+
+	if (!drm_dp_channel_eq_ok(status, link->lanes))
+		drm_dp_link_get_adjustments(link, status);
+	else
+		link->train.channel_equalized = true;
+
+	return 0;
+}
+
+static int drm_dp_link_channel_equalization(struct drm_dp_link *link)
+{
+	unsigned int repeat;
+	int err;
+
+	/* start channel equalization using pattern 2 or 3 */
+	if (link->caps.tps3_supported)
+		link->train.pattern = DP_TRAINING_PATTERN_3;
+	else
+		link->train.pattern = DP_TRAINING_PATTERN_2;
+
+	for (repeat = 1; repeat < 5; repeat++) {
+		err = drm_dp_link_equalize_channel(link);
+		if (err < 0) {
+			DRM_ERROR("failed to equalize channel: %d\n", err);
+			return err;
+		}
+
+		if (link->train.channel_equalized)
+			break;
+
+		drm_dp_link_train_adjust(&link->train);
+	}
+
+	return 0;
+}
+
+static int drm_dp_link_downgrade(struct drm_dp_link *link)
+{
+	switch (link->rate) {
+	case 162000:
+		return -EINVAL;
+
+	case 270000:
+		link->rate = 162000;
+		break;
+
+	case 540000:
+		link->rate = 270000;
+		return 0;
+	}
+
+	return 0;
+}
+
+static void drm_dp_link_train_disable(struct drm_dp_link *link)
+{
+	int err;
+
+	link->train.pattern = DP_TRAINING_PATTERN_DISABLE;
+
+	err = drm_dp_link_apply_training(link);
+	if (err < 0)
+		DRM_ERROR("failed to disable link training: %d\n", err);
+}
+
+static int drm_dp_link_train_full(struct drm_dp_link *link)
+{
+	int err;
+
+retry:
+	DRM_DEBUG_KMS("full-training link: %u lane%s at %u MHz\n",
+		      link->lanes, str_plural(link->lanes),
+		      link->rate / 100);
+
+	err = drm_dp_link_configure(link->aux, link);
+	if (err < 0) {
+		DRM_ERROR("failed to configure DP link: %d\n", err);
+		return err;
+	}
+
+	err = drm_dp_link_clock_recovery(link);
+	if (err < 0) {
+		DRM_ERROR("clock recovery failed: %d\n", err);
+		goto out;
+	}
+
+	if (!link->train.clock_recovered) {
+		DRM_ERROR("clock recovery failed, downgrading link\n");
+
+		err = drm_dp_link_downgrade(link);
+		if (err < 0)
+			goto out;
+
+		goto retry;
+	}
+
+	DRM_DEBUG_KMS("clock recovery succeeded\n");
+
+	err = drm_dp_link_channel_equalization(link);
+	if (err < 0) {
+		DRM_ERROR("channel equalization failed: %d\n", err);
+		goto out;
+	}
+
+	if (!link->train.channel_equalized) {
+		DRM_ERROR("channel equalization failed, downgrading link\n");
+
+		err = drm_dp_link_downgrade(link);
+		if (err < 0)
+			goto out;
+
+		goto retry;
+	}
+
+	DRM_DEBUG_KMS("channel equalization succeeded\n");
+
+out:
+	drm_dp_link_train_disable(link);
+	return err;
+}
+
+static int drm_dp_link_train_fast(struct drm_dp_link *link)
+{
+	u8 status[DP_LINK_STATUS_SIZE];
+	int err;
+
+	DRM_DEBUG_KMS("fast-training link: %u lane%s at %u MHz\n",
+		      link->lanes, str_plural(link->lanes),
+		      link->rate / 100);
+
+	err = drm_dp_link_configure(link->aux, link);
+	if (err < 0) {
+		DRM_ERROR("failed to configure DP link: %d\n", err);
+		return err;
+	}
+
+	/* transmit training pattern 1 for 500 microseconds */
+	link->train.pattern = DP_TRAINING_PATTERN_1;
+
+	err = drm_dp_link_apply_training(link);
+	if (err < 0)
+		goto out;
+
+	usleep_range(500, 1000);
+
+	/* transmit training pattern 2 or 3 for 500 microseconds */
+	if (link->caps.tps3_supported)
+		link->train.pattern = DP_TRAINING_PATTERN_3;
+	else
+		link->train.pattern = DP_TRAINING_PATTERN_2;
+
+	err = drm_dp_link_apply_training(link);
+	if (err < 0)
+		goto out;
+
+	usleep_range(500, 1000);
+
+	err = drm_dp_dpcd_read_link_status(link->aux, status);
+	if (err < 0) {
+		DRM_ERROR("failed to read link status: %d\n", err);
+		goto out;
+	}
+
+	if (!drm_dp_clock_recovery_ok(status, link->lanes)) {
+		DRM_ERROR("clock recovery failed\n");
+		err = -EIO;
+	}
+
+	if (!drm_dp_channel_eq_ok(status, link->lanes)) {
+		DRM_ERROR("channel equalization failed\n");
+		err = -EIO;
+	}
+
+out:
+	drm_dp_link_train_disable(link);
+	return err;
+}
+
+/**
+ * drm_dp_link_train() - perform DisplayPort link training
+ * @link: a DP link object
+ *
+ * Uses the context stored in the DP link object to perform link training. It
+ * is expected that drivers will call drm_dp_link_probe() to obtain the link
+ * capabilities before performing link training.
+ *
+ * If the sink supports fast link training (no AUX CH handshake) and valid
+ * training settings are available, this function will try to perform fast
+ * link training and fall back to full link training on failure.
+ *
+ * Returns: 0 on success or a negative error code on failure.
+ */
+int drm_dp_link_train(struct drm_dp_link *link)
+{
+	int err;
+
+	drm_dp_link_train_init(&link->train);
+
+	if (link->caps.fast_training) {
+		if (drm_dp_link_train_valid(&link->train)) {
+			err = drm_dp_link_train_fast(link);
+			if (err < 0)
+				DRM_ERROR("fast link training failed: %d\n",
+					  err);
+			else
+				return 0;
+		} else {
+			DRM_DEBUG_KMS("training parameters not available\n");
+		}
+	} else {
+		DRM_DEBUG_KMS("fast link training not supported\n");
+	}
+
+	err = drm_dp_link_train_full(link);
+	if (err < 0)
+		DRM_ERROR("full link training failed: %d\n", err);
+
+	return err;
+}
diff --git a/drivers/gpu/drm/tegra/dp.h b/drivers/gpu/drm/tegra/dp.h
new file mode 100644
index 000000000..774fb13c0
--- /dev/null
+++ b/drivers/gpu/drm/tegra/dp.h
@@ -0,0 +1,173 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2013-2019 NVIDIA Corporation.
+ * Copyright (C) 2015 Rob Clark
+ */
+
+#ifndef DRM_TEGRA_DP_H
+#define DRM_TEGRA_DP_H 1
+
+#include <linux/types.h>
+
+struct drm_display_info;
+struct drm_display_mode;
+struct drm_dp_aux;
+struct drm_dp_link;
+
+/**
+ * struct drm_dp_link_caps - DP link capabilities
+ */
+struct drm_dp_link_caps {
+	/**
+	 * @enhanced_framing:
+	 *
+	 * enhanced framing capability (mandatory as of DP 1.2)
+	 */
+	bool enhanced_framing;
+
+	/**
+	 * @tps3_supported:
+	 *
+	 * training pattern sequence 3 supported for equalization
+	 */
+	bool tps3_supported;
+
+	/**
+	 * @fast_training:
+	 *
+	 * AUX CH handshake not required for link training
+	 */
+	bool fast_training;
+
+	/**
+	 * @channel_coding:
+	 *
+	 * ANSI 8B/10B channel coding capability
+	 */
+	bool channel_coding;
+
+	/**
+	 * @alternate_scrambler_reset:
+	 *
+	 * eDP alternate scrambler reset capability
+	 */
+	bool alternate_scrambler_reset;
+};
+
+void drm_dp_link_caps_copy(struct drm_dp_link_caps *dest,
+			   const struct drm_dp_link_caps *src);
+
+/**
+ * struct drm_dp_link_ops - DP link operations
+ */
+struct drm_dp_link_ops {
+	/**
+	 * @apply_training: apply the link training
+	 */
+	int (*apply_training)(struct drm_dp_link *link);
+
+	/**
+	 * @configure: configure the DP link
+	 */
+	int (*configure)(struct drm_dp_link *link);
+};
+
+#define DP_TRAIN_VOLTAGE_SWING_LEVEL(x) ((x) << 0)
+#define DP_TRAIN_PRE_EMPHASIS_LEVEL(x) ((x) << 3)
+#define DP_LANE_POST_CURSOR(i, x) (((x) & 0x3) << (((i) & 1) << 2))
+
+/**
+ * struct drm_dp_link_train_set - link training settings
+ * @voltage_swing: per-lane voltage swing
+ * @pre_emphasis: per-lane pre-emphasis
+ * @post_cursor: per-lane post-cursor
+ */
+struct drm_dp_link_train_set {
+	unsigned int voltage_swing[4];
+	unsigned int pre_emphasis[4];
+	unsigned int post_cursor[4];
+};
+
+/**
+ * struct drm_dp_link_train - link training state information
+ * @request: currently requested settings
+ * @adjust: adjustments requested by sink
+ * @pattern: currently requested training pattern
+ * @clock_recovered: flag to track if clock recovery has completed
+ * @channel_equalized: flag to track if channel equalization has completed
+ */
+struct drm_dp_link_train {
+	struct drm_dp_link_train_set request;
+	struct drm_dp_link_train_set adjust;
+
+	unsigned int pattern;
+
+	bool clock_recovered;
+	bool channel_equalized;
+};
+
+/**
+ * struct drm_dp_link - DP link capabilities and configuration
+ * @revision: DP specification revision supported on the link
+ * @max_rate: maximum clock rate supported on the link
+ * @max_lanes: maximum number of lanes supported on the link
+ * @caps: capabilities supported on the link (see &drm_dp_link_caps)
+ * @aux_rd_interval: AUX read interval to use for training (in microseconds)
+ * @aux_rd_interval.cr: clock recovery read interval
+ * @aux_rd_interval.ce: channel equalization read interval
+ * @edp: eDP revision (0x11: eDP 1.1, 0x12: eDP 1.2, ...)
+ * @rate: currently configured link rate
+ * @lanes: currently configured number of lanes
+ * @rates: additional supported link rates in kHz (eDP 1.4)
+ * @num_rates: number of additional supported link rates (eDP 1.4)
+ */
+struct drm_dp_link {
+	unsigned char revision;
+	unsigned int max_rate;
+	unsigned int max_lanes;
+
+	struct drm_dp_link_caps caps;
+
+	struct {
+		unsigned int cr;
+		unsigned int ce;
+	} aux_rd_interval;
+
+	unsigned char edp;
+
+	unsigned int rate;
+	unsigned int lanes;
+
+	unsigned long rates[DP_MAX_SUPPORTED_RATES];
+	unsigned int num_rates;
+
+	/**
+	 * @ops: DP link operations
+	 */
+	const struct drm_dp_link_ops *ops;
+
+	/**
+	 * @aux: DP AUX channel
+	 */
+	struct drm_dp_aux *aux;
+
+	/**
+	 * @train: DP link training state
+	 */
+	struct drm_dp_link_train train;
+};
+
+int drm_dp_link_add_rate(struct drm_dp_link *link, unsigned long rate);
+int drm_dp_link_remove_rate(struct drm_dp_link *link, unsigned long rate);
+void drm_dp_link_update_rates(struct drm_dp_link *link);
+
+int drm_dp_link_probe(struct drm_dp_aux *aux, struct drm_dp_link *link);
+int drm_dp_link_configure(struct drm_dp_aux *aux, struct drm_dp_link *link);
+int drm_dp_link_choose(struct drm_dp_link *link,
+		       const struct drm_display_mode *mode,
+		       const struct drm_display_info *info);
+
+void drm_dp_link_train_init(struct drm_dp_link_train *train);
+int drm_dp_link_train(struct drm_dp_link *link);
+
+#endif
diff --git a/drivers/gpu/drm/tegra/dpaux.c b/drivers/gpu/drm/tegra/dpaux.c
new file mode 100644
index 000000000..e5297ac5c
--- /dev/null
+++ b/drivers/gpu/drm/tegra/dpaux.c
@@ -0,0 +1,822 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/pinctrl/pinconf-generic.h>
+#include <linux/pinctrl/pinctrl.h>
+#include <linux/pinctrl/pinmux.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/reset.h>
+#include <linux/workqueue.h>
+
+#include <drm/display/drm_dp_helper.h>
+#include <drm/display/drm_dp_aux_bus.h>
+#include <drm/drm_panel.h>
+
+#include "dp.h"
+#include "dpaux.h"
+#include "drm.h"
+#include "trace.h"
+
+static DEFINE_MUTEX(dpaux_lock);
+static LIST_HEAD(dpaux_list);
+
+struct tegra_dpaux_soc {
+	unsigned int cmh;
+	unsigned int drvz;
+	unsigned int drvi;
+};
+
+struct tegra_dpaux {
+	struct drm_dp_aux aux;
+	struct device *dev;
+
+	const struct tegra_dpaux_soc *soc;
+
+	void __iomem *regs;
+	int irq;
+
+	struct tegra_output *output;
+
+	struct reset_control *rst;
+	struct clk *clk_parent;
+	struct clk *clk;
+
+	struct regulator *vdd;
+
+	struct completion complete;
+	struct work_struct work;
+	struct list_head list;
+
+#ifdef CONFIG_GENERIC_PINCONF
+	struct pinctrl_dev *pinctrl;
+	struct pinctrl_desc desc;
+#endif
+};
+
+static inline struct tegra_dpaux *to_dpaux(struct drm_dp_aux *aux)
+{
+	return container_of(aux, struct tegra_dpaux, aux);
+}
+
+static inline struct tegra_dpaux *work_to_dpaux(struct work_struct *work)
+{
+	return container_of(work, struct tegra_dpaux, work);
+}
+
+static inline u32 tegra_dpaux_readl(struct tegra_dpaux *dpaux,
+				    unsigned int offset)
+{
+	u32 value = readl(dpaux->regs + (offset << 2));
+
+	trace_dpaux_readl(dpaux->dev, offset, value);
+
+	return value;
+}
+
+static inline void tegra_dpaux_writel(struct tegra_dpaux *dpaux,
+				      u32 value, unsigned int offset)
+{
+	trace_dpaux_writel(dpaux->dev, offset, value);
+	writel(value, dpaux->regs + (offset << 2));
+}
+
+static void tegra_dpaux_write_fifo(struct tegra_dpaux *dpaux, const u8 *buffer,
+				   size_t size)
+{
+	size_t i, j;
+
+	for (i = 0; i < DIV_ROUND_UP(size, 4); i++) {
+		size_t num = min_t(size_t, size - i * 4, 4);
+		u32 value = 0;
+
+		for (j = 0; j < num; j++)
+			value |= buffer[i * 4 + j] << (j * 8);
+
+		tegra_dpaux_writel(dpaux, value, DPAUX_DP_AUXDATA_WRITE(i));
+	}
+}
+
+static void tegra_dpaux_read_fifo(struct tegra_dpaux *dpaux, u8 *buffer,
+				  size_t size)
+{
+	size_t i, j;
+
+	for (i = 0; i < DIV_ROUND_UP(size, 4); i++) {
+		size_t num = min_t(size_t, size - i * 4, 4);
+		u32 value;
+
+		value = tegra_dpaux_readl(dpaux, DPAUX_DP_AUXDATA_READ(i));
+
+		for (j = 0; j < num; j++)
+			buffer[i * 4 + j] = value >> (j * 8);
+	}
+}
+
+static ssize_t tegra_dpaux_transfer(struct drm_dp_aux *aux,
+				    struct drm_dp_aux_msg *msg)
+{
+	unsigned long timeout = msecs_to_jiffies(250);
+	struct tegra_dpaux *dpaux = to_dpaux(aux);
+	unsigned long status;
+	ssize_t ret = 0;
+	u8 reply = 0;
+	u32 value;
+
+	/* Tegra has 4x4 byte DP AUX transmit and receive FIFOs. */
+	if (msg->size > 16)
+		return -EINVAL;
+
+	/*
+	 * Allow zero-sized messages only for I2C, in which case they specify
+	 * address-only transactions.
+	 */
+	if (msg->size < 1) {
+		switch (msg->request & ~DP_AUX_I2C_MOT) {
+		case DP_AUX_I2C_WRITE_STATUS_UPDATE:
+		case DP_AUX_I2C_WRITE:
+		case DP_AUX_I2C_READ:
+			value = DPAUX_DP_AUXCTL_CMD_ADDRESS_ONLY;
+			break;
+
+		default:
+			return -EINVAL;
+		}
+	} else {
+		/* For non-zero-sized messages, set the CMDLEN field. */
+		value = DPAUX_DP_AUXCTL_CMDLEN(msg->size - 1);
+	}
+
+	switch (msg->request & ~DP_AUX_I2C_MOT) {
+	case DP_AUX_I2C_WRITE:
+		if (msg->request & DP_AUX_I2C_MOT)
+			value |= DPAUX_DP_AUXCTL_CMD_MOT_WR;
+		else
+			value |= DPAUX_DP_AUXCTL_CMD_I2C_WR;
+
+		break;
+
+	case DP_AUX_I2C_READ:
+		if (msg->request & DP_AUX_I2C_MOT)
+			value |= DPAUX_DP_AUXCTL_CMD_MOT_RD;
+		else
+			value |= DPAUX_DP_AUXCTL_CMD_I2C_RD;
+
+		break;
+
+	case DP_AUX_I2C_WRITE_STATUS_UPDATE:
+		if (msg->request & DP_AUX_I2C_MOT)
+			value |= DPAUX_DP_AUXCTL_CMD_MOT_RQ;
+		else
+			value |= DPAUX_DP_AUXCTL_CMD_I2C_RQ;
+
+		break;
+
+	case DP_AUX_NATIVE_WRITE:
+		value |= DPAUX_DP_AUXCTL_CMD_AUX_WR;
+		break;
+
+	case DP_AUX_NATIVE_READ:
+		value |= DPAUX_DP_AUXCTL_CMD_AUX_RD;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	tegra_dpaux_writel(dpaux, msg->address, DPAUX_DP_AUXADDR);
+	tegra_dpaux_writel(dpaux, value, DPAUX_DP_AUXCTL);
+
+	if ((msg->request & DP_AUX_I2C_READ) == 0) {
+		tegra_dpaux_write_fifo(dpaux, msg->buffer, msg->size);
+		ret = msg->size;
+	}
+
+	/* start transaction */
+	value = tegra_dpaux_readl(dpaux, DPAUX_DP_AUXCTL);
+	value |= DPAUX_DP_AUXCTL_TRANSACTREQ;
+	tegra_dpaux_writel(dpaux, value, DPAUX_DP_AUXCTL);
+
+	status = wait_for_completion_timeout(&dpaux->complete, timeout);
+	if (!status)
+		return -ETIMEDOUT;
+
+	/* read status and clear errors */
+	value = tegra_dpaux_readl(dpaux, DPAUX_DP_AUXSTAT);
+	tegra_dpaux_writel(dpaux, 0xf00, DPAUX_DP_AUXSTAT);
+
+	if (value & DPAUX_DP_AUXSTAT_TIMEOUT_ERROR)
+		return -ETIMEDOUT;
+
+	if ((value & DPAUX_DP_AUXSTAT_RX_ERROR) ||
+	    (value & DPAUX_DP_AUXSTAT_SINKSTAT_ERROR) ||
+	    (value & DPAUX_DP_AUXSTAT_NO_STOP_ERROR))
+		return -EIO;
+
+	switch ((value & DPAUX_DP_AUXSTAT_REPLY_TYPE_MASK) >> 16) {
+	case 0x00:
+		reply = DP_AUX_NATIVE_REPLY_ACK;
+		break;
+
+	case 0x01:
+		reply = DP_AUX_NATIVE_REPLY_NACK;
+		break;
+
+	case 0x02:
+		reply = DP_AUX_NATIVE_REPLY_DEFER;
+		break;
+
+	case 0x04:
+		reply = DP_AUX_I2C_REPLY_NACK;
+		break;
+
+	case 0x08:
+		reply = DP_AUX_I2C_REPLY_DEFER;
+		break;
+	}
+
+	if ((msg->size > 0) && (msg->reply == DP_AUX_NATIVE_REPLY_ACK)) {
+		if (msg->request & DP_AUX_I2C_READ) {
+			size_t count = value & DPAUX_DP_AUXSTAT_REPLY_MASK;
+
+			/*
+			 * There might be a smarter way to do this, but since
+			 * the DP helpers will already retry transactions for
+			 * an -EBUSY return value, simply reuse that instead.
+			 */
+			if (count != msg->size) {
+				ret = -EBUSY;
+				goto out;
+			}
+
+			tegra_dpaux_read_fifo(dpaux, msg->buffer, count);
+			ret = count;
+		}
+	}
+
+	msg->reply = reply;
+
+out:
+	return ret;
+}
+
+static void tegra_dpaux_hotplug(struct work_struct *work)
+{
+	struct tegra_dpaux *dpaux = work_to_dpaux(work);
+
+	if (dpaux->output)
+		drm_helper_hpd_irq_event(dpaux->output->connector.dev);
+}
+
+static irqreturn_t tegra_dpaux_irq(int irq, void *data)
+{
+	struct tegra_dpaux *dpaux = data;
+	u32 value;
+
+	/* clear interrupts */
+	value = tegra_dpaux_readl(dpaux, DPAUX_INTR_AUX);
+	tegra_dpaux_writel(dpaux, value, DPAUX_INTR_AUX);
+
+	if (value & (DPAUX_INTR_PLUG_EVENT | DPAUX_INTR_UNPLUG_EVENT))
+		schedule_work(&dpaux->work);
+
+	if (value & DPAUX_INTR_IRQ_EVENT) {
+		/* TODO: handle this */
+	}
+
+	if (value & DPAUX_INTR_AUX_DONE)
+		complete(&dpaux->complete);
+
+	return IRQ_HANDLED;
+}
+
+enum tegra_dpaux_functions {
+	DPAUX_PADCTL_FUNC_AUX,
+	DPAUX_PADCTL_FUNC_I2C,
+	DPAUX_PADCTL_FUNC_OFF,
+};
+
+static void tegra_dpaux_pad_power_down(struct tegra_dpaux *dpaux)
+{
+	u32 value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE);
+
+	value |= DPAUX_HYBRID_SPARE_PAD_POWER_DOWN;
+
+	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE);
+}
+
+static void tegra_dpaux_pad_power_up(struct tegra_dpaux *dpaux)
+{
+	u32 value = tegra_dpaux_readl(dpaux, DPAUX_HYBRID_SPARE);
+
+	value &= ~DPAUX_HYBRID_SPARE_PAD_POWER_DOWN;
+
+	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_SPARE);
+}
+
+static int tegra_dpaux_pad_config(struct tegra_dpaux *dpaux, unsigned function)
+{
+	u32 value;
+
+	switch (function) {
+	case DPAUX_PADCTL_FUNC_AUX:
+		value = DPAUX_HYBRID_PADCTL_AUX_CMH(dpaux->soc->cmh) |
+			DPAUX_HYBRID_PADCTL_AUX_DRVZ(dpaux->soc->drvz) |
+			DPAUX_HYBRID_PADCTL_AUX_DRVI(dpaux->soc->drvi) |
+			DPAUX_HYBRID_PADCTL_AUX_INPUT_RCV |
+			DPAUX_HYBRID_PADCTL_MODE_AUX;
+		break;
+
+	case DPAUX_PADCTL_FUNC_I2C:
+		value = DPAUX_HYBRID_PADCTL_I2C_SDA_INPUT_RCV |
+			DPAUX_HYBRID_PADCTL_I2C_SCL_INPUT_RCV |
+			DPAUX_HYBRID_PADCTL_AUX_CMH(dpaux->soc->cmh) |
+			DPAUX_HYBRID_PADCTL_AUX_DRVZ(dpaux->soc->drvz) |
+			DPAUX_HYBRID_PADCTL_AUX_DRVI(dpaux->soc->drvi) |
+			DPAUX_HYBRID_PADCTL_MODE_I2C;
+		break;
+
+	case DPAUX_PADCTL_FUNC_OFF:
+		tegra_dpaux_pad_power_down(dpaux);
+		return 0;
+
+	default:
+		return -ENOTSUPP;
+	}
+
+	tegra_dpaux_writel(dpaux, value, DPAUX_HYBRID_PADCTL);
+	tegra_dpaux_pad_power_up(dpaux);
+
+	return 0;
+}
+
+#ifdef CONFIG_GENERIC_PINCONF
+static const struct pinctrl_pin_desc tegra_dpaux_pins[] = {
+	PINCTRL_PIN(0, "DP_AUX_CHx_P"),
+	PINCTRL_PIN(1, "DP_AUX_CHx_N"),
+};
+
+static const unsigned tegra_dpaux_pin_numbers[] = { 0, 1 };
+
+static const char * const tegra_dpaux_groups[] = {
+	"dpaux-io",
+};
+
+static const char * const tegra_dpaux_functions[] = {
+	"aux",
+	"i2c",
+	"off",
+};
+
+static int tegra_dpaux_get_groups_count(struct pinctrl_dev *pinctrl)
+{
+	return ARRAY_SIZE(tegra_dpaux_groups);
+}
+
+static const char *tegra_dpaux_get_group_name(struct pinctrl_dev *pinctrl,
+					      unsigned int group)
+{
+	return tegra_dpaux_groups[group];
+}
+
+static int tegra_dpaux_get_group_pins(struct pinctrl_dev *pinctrl,
+				      unsigned group, const unsigned **pins,
+				      unsigned *num_pins)
+{
+	*pins = tegra_dpaux_pin_numbers;
+	*num_pins = ARRAY_SIZE(tegra_dpaux_pin_numbers);
+
+	return 0;
+}
+
+static const struct pinctrl_ops tegra_dpaux_pinctrl_ops = {
+	.get_groups_count = tegra_dpaux_get_groups_count,
+	.get_group_name = tegra_dpaux_get_group_name,
+	.get_group_pins = tegra_dpaux_get_group_pins,
+	.dt_node_to_map = pinconf_generic_dt_node_to_map_group,
+	.dt_free_map = pinconf_generic_dt_free_map,
+};
+
+static int tegra_dpaux_get_functions_count(struct pinctrl_dev *pinctrl)
+{
+	return ARRAY_SIZE(tegra_dpaux_functions);
+}
+
+static const char *tegra_dpaux_get_function_name(struct pinctrl_dev *pinctrl,
+						 unsigned int function)
+{
+	return tegra_dpaux_functions[function];
+}
+
+static int tegra_dpaux_get_function_groups(struct pinctrl_dev *pinctrl,
+					   unsigned int function,
+					   const char * const **groups,
+					   unsigned * const num_groups)
+{
+	*num_groups = ARRAY_SIZE(tegra_dpaux_groups);
+	*groups = tegra_dpaux_groups;
+
+	return 0;
+}
+
+static int tegra_dpaux_set_mux(struct pinctrl_dev *pinctrl,
+			       unsigned int function, unsigned int group)
+{
+	struct tegra_dpaux *dpaux = pinctrl_dev_get_drvdata(pinctrl);
+
+	return tegra_dpaux_pad_config(dpaux, function);
+}
+
+static const struct pinmux_ops tegra_dpaux_pinmux_ops = {
+	.get_functions_count = tegra_dpaux_get_functions_count,
+	.get_function_name = tegra_dpaux_get_function_name,
+	.get_function_groups = tegra_dpaux_get_function_groups,
+	.set_mux = tegra_dpaux_set_mux,
+};
+#endif
+
+static int tegra_dpaux_probe(struct platform_device *pdev)
+{
+	struct tegra_dpaux *dpaux;
+	u32 value;
+	int err;
+
+	dpaux = devm_kzalloc(&pdev->dev, sizeof(*dpaux), GFP_KERNEL);
+	if (!dpaux)
+		return -ENOMEM;
+
+	dpaux->soc = of_device_get_match_data(&pdev->dev);
+	INIT_WORK(&dpaux->work, tegra_dpaux_hotplug);
+	init_completion(&dpaux->complete);
+	INIT_LIST_HEAD(&dpaux->list);
+	dpaux->dev = &pdev->dev;
+
+	dpaux->regs = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(dpaux->regs))
+		return PTR_ERR(dpaux->regs);
+
+	dpaux->irq = platform_get_irq(pdev, 0);
+	if (dpaux->irq < 0)
+		return dpaux->irq;
+
+	if (!pdev->dev.pm_domain) {
+		dpaux->rst = devm_reset_control_get(&pdev->dev, "dpaux");
+		if (IS_ERR(dpaux->rst)) {
+			dev_err(&pdev->dev,
+				"failed to get reset control: %ld\n",
+				PTR_ERR(dpaux->rst));
+			return PTR_ERR(dpaux->rst);
+		}
+	}
+
+	dpaux->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(dpaux->clk)) {
+		dev_err(&pdev->dev, "failed to get module clock: %ld\n",
+			PTR_ERR(dpaux->clk));
+		return PTR_ERR(dpaux->clk);
+	}
+
+	dpaux->clk_parent = devm_clk_get(&pdev->dev, "parent");
+	if (IS_ERR(dpaux->clk_parent)) {
+		dev_err(&pdev->dev, "failed to get parent clock: %ld\n",
+			PTR_ERR(dpaux->clk_parent));
+		return PTR_ERR(dpaux->clk_parent);
+	}
+
+	err = clk_set_rate(dpaux->clk_parent, 270000000);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set clock to 270 MHz: %d\n",
+			err);
+		return err;
+	}
+
+	dpaux->vdd = devm_regulator_get_optional(&pdev->dev, "vdd");
+	if (IS_ERR(dpaux->vdd)) {
+		if (PTR_ERR(dpaux->vdd) != -ENODEV)
+			return dev_err_probe(&pdev->dev, PTR_ERR(dpaux->vdd),
+					     "failed to get VDD supply\n");
+
+		dpaux->vdd = NULL;
+	}
+
+	platform_set_drvdata(pdev, dpaux);
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_get_sync(&pdev->dev);
+
+	err = devm_request_irq(dpaux->dev, dpaux->irq, tegra_dpaux_irq, 0,
+			       dev_name(dpaux->dev), dpaux);
+	if (err < 0) {
+		dev_err(dpaux->dev, "failed to request IRQ#%u: %d\n",
+			dpaux->irq, err);
+		goto err_pm_disable;
+	}
+
+	disable_irq(dpaux->irq);
+
+	dpaux->aux.transfer = tegra_dpaux_transfer;
+	dpaux->aux.dev = &pdev->dev;
+
+	drm_dp_aux_init(&dpaux->aux);
+
+	/*
+	 * Assume that by default the DPAUX/I2C pads will be used for HDMI,
+	 * so power them up and configure them in I2C mode.
+	 *
+	 * The DPAUX code paths reconfigure the pads in AUX mode, but there
+	 * is no possibility to perform the I2C mode configuration in the
+	 * HDMI path.
+	 */
+	err = tegra_dpaux_pad_config(dpaux, DPAUX_PADCTL_FUNC_I2C);
+	if (err < 0)
+		goto err_pm_disable;
+
+#ifdef CONFIG_GENERIC_PINCONF
+	dpaux->desc.name = dev_name(&pdev->dev);
+	dpaux->desc.pins = tegra_dpaux_pins;
+	dpaux->desc.npins = ARRAY_SIZE(tegra_dpaux_pins);
+	dpaux->desc.pctlops = &tegra_dpaux_pinctrl_ops;
+	dpaux->desc.pmxops = &tegra_dpaux_pinmux_ops;
+	dpaux->desc.owner = THIS_MODULE;
+
+	dpaux->pinctrl = devm_pinctrl_register(&pdev->dev, &dpaux->desc, dpaux);
+	if (IS_ERR(dpaux->pinctrl)) {
+		dev_err(&pdev->dev, "failed to register pincontrol\n");
+		err = PTR_ERR(dpaux->pinctrl);
+		goto err_pm_disable;
+	}
+#endif
+	/* enable and clear all interrupts */
+	value = DPAUX_INTR_AUX_DONE | DPAUX_INTR_IRQ_EVENT |
+		DPAUX_INTR_UNPLUG_EVENT | DPAUX_INTR_PLUG_EVENT;
+	tegra_dpaux_writel(dpaux, value, DPAUX_INTR_EN_AUX);
+	tegra_dpaux_writel(dpaux, value, DPAUX_INTR_AUX);
+
+	mutex_lock(&dpaux_lock);
+	list_add_tail(&dpaux->list, &dpaux_list);
+	mutex_unlock(&dpaux_lock);
+
+	err = devm_of_dp_aux_populate_ep_devices(&dpaux->aux);
+	if (err < 0) {
+		dev_err(dpaux->dev, "failed to populate AUX bus: %d\n", err);
+		goto err_pm_disable;
+	}
+
+	return 0;
+
+err_pm_disable:
+	pm_runtime_put_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+	return err;
+}
+
+static void tegra_dpaux_remove(struct platform_device *pdev)
+{
+	struct tegra_dpaux *dpaux = platform_get_drvdata(pdev);
+
+	cancel_work_sync(&dpaux->work);
+
+	/* make sure pads are powered down when not in use */
+	tegra_dpaux_pad_power_down(dpaux);
+
+	pm_runtime_put_sync(&pdev->dev);
+	pm_runtime_disable(&pdev->dev);
+
+	mutex_lock(&dpaux_lock);
+	list_del(&dpaux->list);
+	mutex_unlock(&dpaux_lock);
+}
+
+static int tegra_dpaux_suspend(struct device *dev)
+{
+	struct tegra_dpaux *dpaux = dev_get_drvdata(dev);
+	int err = 0;
+
+	if (dpaux->rst) {
+		err = reset_control_assert(dpaux->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to assert reset: %d\n", err);
+			return err;
+		}
+	}
+
+	usleep_range(1000, 2000);
+
+	clk_disable_unprepare(dpaux->clk_parent);
+	clk_disable_unprepare(dpaux->clk);
+
+	return err;
+}
+
+static int tegra_dpaux_resume(struct device *dev)
+{
+	struct tegra_dpaux *dpaux = dev_get_drvdata(dev);
+	int err;
+
+	err = clk_prepare_enable(dpaux->clk);
+	if (err < 0) {
+		dev_err(dev, "failed to enable clock: %d\n", err);
+		return err;
+	}
+
+	err = clk_prepare_enable(dpaux->clk_parent);
+	if (err < 0) {
+		dev_err(dev, "failed to enable parent clock: %d\n", err);
+		goto disable_clk;
+	}
+
+	usleep_range(1000, 2000);
+
+	if (dpaux->rst) {
+		err = reset_control_deassert(dpaux->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to deassert reset: %d\n", err);
+			goto disable_parent;
+		}
+
+		usleep_range(1000, 2000);
+	}
+
+	return 0;
+
+disable_parent:
+	clk_disable_unprepare(dpaux->clk_parent);
+disable_clk:
+	clk_disable_unprepare(dpaux->clk);
+	return err;
+}
+
+static const struct dev_pm_ops tegra_dpaux_pm_ops = {
+	RUNTIME_PM_OPS(tegra_dpaux_suspend, tegra_dpaux_resume, NULL)
+};
+
+static const struct tegra_dpaux_soc tegra124_dpaux_soc = {
+	.cmh = 0x02,
+	.drvz = 0x04,
+	.drvi = 0x18,
+};
+
+static const struct tegra_dpaux_soc tegra210_dpaux_soc = {
+	.cmh = 0x02,
+	.drvz = 0x04,
+	.drvi = 0x30,
+};
+
+static const struct tegra_dpaux_soc tegra194_dpaux_soc = {
+	.cmh = 0x02,
+	.drvz = 0x04,
+	.drvi = 0x2c,
+};
+
+static const struct of_device_id tegra_dpaux_of_match[] = {
+	{ .compatible = "nvidia,tegra194-dpaux", .data = &tegra194_dpaux_soc },
+	{ .compatible = "nvidia,tegra186-dpaux", .data = &tegra210_dpaux_soc },
+	{ .compatible = "nvidia,tegra210-dpaux", .data = &tegra210_dpaux_soc },
+	{ .compatible = "nvidia,tegra124-dpaux", .data = &tegra124_dpaux_soc },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, tegra_dpaux_of_match);
+
+struct platform_driver tegra_dpaux_driver = {
+	.driver = {
+		.name = "tegra-dpaux",
+		.of_match_table = tegra_dpaux_of_match,
+		.pm = pm_ptr(&tegra_dpaux_pm_ops),
+	},
+	.probe = tegra_dpaux_probe,
+	.remove = tegra_dpaux_remove,
+};
+
+struct drm_dp_aux *drm_dp_aux_find_by_of_node(struct device_node *np)
+{
+	struct tegra_dpaux *dpaux;
+
+	mutex_lock(&dpaux_lock);
+
+	list_for_each_entry(dpaux, &dpaux_list, list)
+		if (np == dpaux->dev->of_node) {
+			mutex_unlock(&dpaux_lock);
+			return &dpaux->aux;
+		}
+
+	mutex_unlock(&dpaux_lock);
+
+	return NULL;
+}
+
+int drm_dp_aux_attach(struct drm_dp_aux *aux, struct tegra_output *output)
+{
+	struct tegra_dpaux *dpaux = to_dpaux(aux);
+	unsigned long timeout;
+	int err;
+
+	aux->drm_dev = output->connector.dev;
+	err = drm_dp_aux_register(aux);
+	if (err < 0)
+		return err;
+
+	output->connector.polled = DRM_CONNECTOR_POLL_HPD;
+	dpaux->output = output;
+
+	if (output->panel) {
+		enum drm_connector_status status;
+
+		if (dpaux->vdd) {
+			err = regulator_enable(dpaux->vdd);
+			if (err < 0)
+				return err;
+		}
+
+		timeout = jiffies + msecs_to_jiffies(250);
+
+		while (time_before(jiffies, timeout)) {
+			status = drm_dp_aux_detect(aux);
+
+			if (status == connector_status_connected)
+				break;
+
+			usleep_range(1000, 2000);
+		}
+
+		if (status != connector_status_connected)
+			return -ETIMEDOUT;
+	}
+
+	enable_irq(dpaux->irq);
+	return 0;
+}
+
+int drm_dp_aux_detach(struct drm_dp_aux *aux)
+{
+	struct tegra_dpaux *dpaux = to_dpaux(aux);
+	unsigned long timeout;
+	int err;
+
+	drm_dp_aux_unregister(aux);
+	disable_irq(dpaux->irq);
+
+	if (dpaux->output->panel) {
+		enum drm_connector_status status;
+
+		if (dpaux->vdd) {
+			err = regulator_disable(dpaux->vdd);
+			if (err < 0)
+				return err;
+		}
+
+		timeout = jiffies + msecs_to_jiffies(250);
+
+		while (time_before(jiffies, timeout)) {
+			status = drm_dp_aux_detect(aux);
+
+			if (status == connector_status_disconnected)
+				break;
+
+			usleep_range(1000, 2000);
+		}
+
+		if (status != connector_status_disconnected)
+			return -ETIMEDOUT;
+
+		dpaux->output = NULL;
+	}
+
+	return 0;
+}
+
+enum drm_connector_status drm_dp_aux_detect(struct drm_dp_aux *aux)
+{
+	struct tegra_dpaux *dpaux = to_dpaux(aux);
+	u32 value;
+
+	value = tegra_dpaux_readl(dpaux, DPAUX_DP_AUXSTAT);
+
+	if (value & DPAUX_DP_AUXSTAT_HPD_STATUS)
+		return connector_status_connected;
+
+	return connector_status_disconnected;
+}
+
+int drm_dp_aux_enable(struct drm_dp_aux *aux)
+{
+	struct tegra_dpaux *dpaux = to_dpaux(aux);
+
+	return tegra_dpaux_pad_config(dpaux, DPAUX_PADCTL_FUNC_AUX);
+}
+
+int drm_dp_aux_disable(struct drm_dp_aux *aux)
+{
+	struct tegra_dpaux *dpaux = to_dpaux(aux);
+
+	tegra_dpaux_pad_power_down(dpaux);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/tegra/dpaux.h b/drivers/gpu/drm/tegra/dpaux.h
new file mode 100644
index 000000000..5eced10fa
--- /dev/null
+++ b/drivers/gpu/drm/tegra/dpaux.h
@@ -0,0 +1,73 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#ifndef DRM_TEGRA_DPAUX_H
+#define DRM_TEGRA_DPAUX_H
+
+#define DPAUX_CTXSW 0x00
+
+#define DPAUX_INTR_EN_AUX 0x01
+#define DPAUX_INTR_AUX 0x05
+#define DPAUX_INTR_AUX_DONE (1 << 3)
+#define DPAUX_INTR_IRQ_EVENT (1 << 2)
+#define DPAUX_INTR_UNPLUG_EVENT (1 << 1)
+#define DPAUX_INTR_PLUG_EVENT (1 << 0)
+
+#define DPAUX_DP_AUXDATA_WRITE(x) (0x09 + ((x) << 2))
+#define DPAUX_DP_AUXDATA_READ(x) (0x19 + ((x) << 2))
+#define DPAUX_DP_AUXADDR 0x29
+
+#define DPAUX_DP_AUXCTL 0x2d
+#define DPAUX_DP_AUXCTL_TRANSACTREQ (1 << 16)
+#define DPAUX_DP_AUXCTL_CMD_AUX_RD (9 << 12)
+#define DPAUX_DP_AUXCTL_CMD_AUX_WR (8 << 12)
+#define DPAUX_DP_AUXCTL_CMD_MOT_RQ (6 << 12)
+#define DPAUX_DP_AUXCTL_CMD_MOT_RD (5 << 12)
+#define DPAUX_DP_AUXCTL_CMD_MOT_WR (4 << 12)
+#define DPAUX_DP_AUXCTL_CMD_I2C_RQ (2 << 12)
+#define DPAUX_DP_AUXCTL_CMD_I2C_RD (1 << 12)
+#define DPAUX_DP_AUXCTL_CMD_I2C_WR (0 << 12)
+#define DPAUX_DP_AUXCTL_CMD_ADDRESS_ONLY (1 << 8)
+#define DPAUX_DP_AUXCTL_CMDLEN(x) ((x) & 0xff)
+
+#define DPAUX_DP_AUXSTAT 0x31
+#define DPAUX_DP_AUXSTAT_HPD_STATUS (1 << 28)
+#define DPAUX_DP_AUXSTAT_REPLY_TYPE_MASK (0xf0000)
+#define DPAUX_DP_AUXSTAT_NO_STOP_ERROR (1 << 11)
+#define DPAUX_DP_AUXSTAT_SINKSTAT_ERROR (1 << 10)
+#define DPAUX_DP_AUXSTAT_RX_ERROR (1 << 9)
+#define DPAUX_DP_AUXSTAT_TIMEOUT_ERROR (1 << 8)
+#define DPAUX_DP_AUXSTAT_REPLY_MASK (0xff)
+
+#define DPAUX_DP_AUX_SINKSTAT_LO 0x35
+#define DPAUX_DP_AUX_SINKSTAT_HI 0x39
+
+#define DPAUX_HPD_CONFIG 0x3d
+#define DPAUX_HPD_CONFIG_UNPLUG_MIN_TIME(x) (((x) & 0xffff) << 16)
+#define DPAUX_HPD_CONFIG_PLUG_MIN_TIME(x) ((x) & 0xffff)
+
+#define DPAUX_HPD_IRQ_CONFIG 0x41
+#define DPAUX_HPD_IRQ_CONFIG_MIN_LOW_TIME(x) ((x) & 0xffff)
+
+#define DPAUX_DP_AUX_CONFIG 0x45
+
+#define DPAUX_HYBRID_PADCTL 0x49
+#define DPAUX_HYBRID_PADCTL_I2C_SDA_INPUT_RCV (1 << 15)
+#define DPAUX_HYBRID_PADCTL_I2C_SCL_INPUT_RCV (1 << 14)
+#define DPAUX_HYBRID_PADCTL_AUX_CMH(x) (((x) & 0x3) << 12)
+#define DPAUX_HYBRID_PADCTL_AUX_DRVZ(x) (((x) & 0x7) << 8)
+#define DPAUX_HYBRID_PADCTL_AUX_DRVI(x) (((x) & 0x3f) << 2)
+#define DPAUX_HYBRID_PADCTL_AUX_INPUT_RCV (1 << 1)
+#define DPAUX_HYBRID_PADCTL_MODE_I2C (1 << 0)
+#define DPAUX_HYBRID_PADCTL_MODE_AUX (0 << 0)
+
+#define DPAUX_HYBRID_SPARE 0x4d
+#define DPAUX_HYBRID_SPARE_PAD_POWER_DOWN (1 << 0)
+
+#define DPAUX_SCRATCH_REG0 0x51
+#define DPAUX_SCRATCH_REG1 0x55
+#define DPAUX_SCRATCH_REG2 0x59
+
+#endif
diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c
new file mode 100644
index 000000000..324480c87
--- /dev/null
+++ b/drivers/gpu/drm/tegra/drm.c
@@ -0,0 +1,1462 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012-2016 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#include <linux/aperture.h>
+#include <linux/bitops.h>
+#include <linux/host1x.h>
+#include <linux/idr.h>
+#include <linux/iommu.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+#include <drm/clients/drm_client_setup.h>
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_debugfs.h>
+#include <drm/drm_drv.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_ioctl.h>
+#include <drm/drm_prime.h>
+#include <drm/drm_print.h>
+#include <drm/drm_vblank.h>
+
+#if IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU)
+#include <asm/dma-iommu.h>
+#endif
+
+#include "dc.h"
+#include "drm.h"
+#include "gem.h"
+#include "uapi.h"
+
+#define DRIVER_NAME "tegra"
+#define DRIVER_DESC "NVIDIA Tegra graphics"
+#define DRIVER_MAJOR 1
+#define DRIVER_MINOR 0
+#define DRIVER_PATCHLEVEL 0
+
+#define CARVEOUT_SZ SZ_64M
+#define CDMA_GATHER_FETCHES_MAX_NB 16383
+
+static int tegra_atomic_check(struct drm_device *drm,
+			      struct drm_atomic_commit *state)
+{
+	int err;
+
+	err = drm_atomic_helper_check(drm, state);
+	if (err < 0)
+		return err;
+
+	return tegra_display_hub_atomic_check(drm, state);
+}
+
+static const struct drm_mode_config_funcs tegra_drm_mode_config_funcs = {
+	.fb_create = tegra_fb_create,
+	.atomic_check = tegra_atomic_check,
+	.atomic_commit = drm_atomic_helper_commit,
+};
+
+static void tegra_atomic_post_commit(struct drm_device *drm,
+				     struct drm_atomic_commit *old_state)
+{
+	struct drm_crtc_state *old_crtc_state __maybe_unused;
+	struct drm_crtc *crtc;
+	unsigned int i;
+
+	for_each_old_crtc_in_state(old_state, crtc, old_crtc_state, i)
+		tegra_crtc_atomic_post_commit(crtc, old_state);
+}
+
+static void tegra_atomic_commit_tail(struct drm_atomic_commit *old_state)
+{
+	struct drm_device *drm = old_state->dev;
+	struct tegra_drm *tegra = drm->dev_private;
+
+	if (tegra->hub) {
+		bool fence_cookie = dma_fence_begin_signalling();
+
+		drm_atomic_helper_commit_modeset_disables(drm, old_state);
+		tegra_display_hub_atomic_commit(drm, old_state);
+		drm_atomic_helper_commit_planes(drm, old_state, 0);
+		drm_atomic_helper_commit_modeset_enables(drm, old_state);
+		drm_atomic_helper_commit_hw_done(old_state);
+		dma_fence_end_signalling(fence_cookie);
+		drm_atomic_helper_wait_for_vblanks(drm, old_state);
+		drm_atomic_helper_cleanup_planes(drm, old_state);
+	} else {
+		drm_atomic_helper_commit_tail_rpm(old_state);
+	}
+
+	tegra_atomic_post_commit(drm, old_state);
+}
+
+static const struct drm_mode_config_helper_funcs
+tegra_drm_mode_config_helpers = {
+	.atomic_commit_tail = tegra_atomic_commit_tail,
+};
+
+static int tegra_drm_open(struct drm_device *drm, struct drm_file *filp)
+{
+	struct tegra_drm_file *fpriv;
+
+	fpriv = kzalloc_obj(*fpriv);
+	if (!fpriv)
+		return -ENOMEM;
+
+	idr_init_base(&fpriv->legacy_contexts, 1);
+	xa_init_flags(&fpriv->contexts, XA_FLAGS_ALLOC1);
+	xa_init(&fpriv->syncpoints);
+	mutex_init(&fpriv->lock);
+	filp->driver_priv = fpriv;
+
+	return 0;
+}
+
+static void tegra_drm_context_free(struct tegra_drm_context *context)
+{
+	context->client->ops->close_channel(context);
+	pm_runtime_put(context->client->base.dev);
+	kfree(context);
+}
+
+static int host1x_reloc_copy_from_user(struct host1x_reloc *dest,
+				       struct drm_tegra_reloc __user *src,
+				       struct drm_device *drm,
+				       struct drm_file *file)
+{
+	u32 cmdbuf, target;
+	int err;
+
+	err = get_user(cmdbuf, &src->cmdbuf.handle);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->cmdbuf.offset, &src->cmdbuf.offset);
+	if (err < 0)
+		return err;
+
+	err = get_user(target, &src->target.handle);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->target.offset, &src->target.offset);
+	if (err < 0)
+		return err;
+
+	err = get_user(dest->shift, &src->shift);
+	if (err < 0)
+		return err;
+
+	dest->flags = HOST1X_RELOC_READ | HOST1X_RELOC_WRITE;
+
+	dest->cmdbuf.bo = tegra_gem_lookup(file, cmdbuf);
+	if (!dest->cmdbuf.bo)
+		return -ENOENT;
+
+	dest->target.bo = tegra_gem_lookup(file, target);
+	if (!dest->target.bo)
+		return -ENOENT;
+
+	return 0;
+}
+
+int tegra_drm_submit(struct tegra_drm_context *context,
+		     struct drm_tegra_submit *args, struct drm_device *drm,
+		     struct drm_file *file)
+{
+	struct host1x_client *client = &context->client->base;
+	unsigned int num_cmdbufs = args->num_cmdbufs;
+	unsigned int num_relocs = args->num_relocs;
+	struct drm_tegra_cmdbuf __user *user_cmdbufs;
+	struct drm_tegra_reloc __user *user_relocs;
+	struct drm_tegra_syncpt __user *user_syncpt;
+	struct drm_tegra_syncpt syncpt;
+	struct host1x *host1x = dev_get_drvdata(drm->dev->parent);
+	struct drm_gem_object **refs;
+	struct host1x_syncpt *sp = NULL;
+	struct host1x_job *job;
+	unsigned int num_refs;
+	int err;
+
+	user_cmdbufs = u64_to_user_ptr(args->cmdbufs);
+	user_relocs = u64_to_user_ptr(args->relocs);
+	user_syncpt = u64_to_user_ptr(args->syncpts);
+
+	/* We don't yet support other than one syncpt_incr struct per submit */
+	if (args->num_syncpts != 1)
+		return -EINVAL;
+
+	/* We don't yet support waitchks */
+	if (args->num_waitchks != 0)
+		return -EINVAL;
+
+	job = host1x_job_alloc(context->channel, args->num_cmdbufs,
+			       args->num_relocs, false);
+	if (!job)
+		return -ENOMEM;
+
+	job->num_relocs = args->num_relocs;
+	job->client = client;
+	job->class = client->class;
+	job->serialize = true;
+	job->syncpt_recovery = true;
+
+	/*
+	 * Track referenced BOs so that they can be unreferenced after the
+	 * submission is complete.
+	 */
+	num_refs = num_cmdbufs + num_relocs * 2;
+
+	refs = kmalloc_objs(*refs, num_refs);
+	if (!refs) {
+		err = -ENOMEM;
+		goto put;
+	}
+
+	/* reuse as an iterator later */
+	num_refs = 0;
+
+	while (num_cmdbufs) {
+		struct drm_tegra_cmdbuf cmdbuf;
+		struct host1x_bo *bo;
+		struct tegra_bo *obj;
+		u64 offset;
+
+		if (copy_from_user(&cmdbuf, user_cmdbufs, sizeof(cmdbuf))) {
+			err = -EFAULT;
+			goto fail;
+		}
+
+		/*
+		 * The maximum number of CDMA gather fetches is 16383, a higher
+		 * value means the words count is malformed.
+		 */
+		if (cmdbuf.words > CDMA_GATHER_FETCHES_MAX_NB) {
+			err = -EINVAL;
+			goto fail;
+		}
+
+		bo = tegra_gem_lookup(file, cmdbuf.handle);
+		if (!bo) {
+			err = -ENOENT;
+			goto fail;
+		}
+
+		offset = (u64)cmdbuf.offset + (u64)cmdbuf.words * sizeof(u32);
+		obj = host1x_to_tegra_bo(bo);
+		refs[num_refs++] = &obj->gem;
+
+		/*
+		 * Gather buffer base address must be 4-bytes aligned,
+		 * unaligned offset is malformed and cause commands stream
+		 * corruption on the buffer address relocation.
+		 */
+		if (offset & 3 || offset > obj->gem.size) {
+			err = -EINVAL;
+			goto fail;
+		}
+
+		host1x_job_add_gather(job, bo, cmdbuf.words, cmdbuf.offset);
+		num_cmdbufs--;
+		user_cmdbufs++;
+	}
+
+	/* copy and resolve relocations from submit */
+	while (num_relocs--) {
+		struct host1x_reloc *reloc;
+		struct tegra_bo *obj;
+
+		err = host1x_reloc_copy_from_user(&job->relocs[num_relocs],
+						  &user_relocs[num_relocs], drm,
+						  file);
+		if (err < 0)
+			goto fail;
+
+		reloc = &job->relocs[num_relocs];
+		obj = host1x_to_tegra_bo(reloc->cmdbuf.bo);
+		refs[num_refs++] = &obj->gem;
+
+		/*
+		 * The unaligned cmdbuf offset will cause an unaligned write
+		 * during of the relocations patching, corrupting the commands
+		 * stream.
+		 */
+		if (reloc->cmdbuf.offset & 3 ||
+		    reloc->cmdbuf.offset >= obj->gem.size) {
+			err = -EINVAL;
+			goto fail;
+		}
+
+		obj = host1x_to_tegra_bo(reloc->target.bo);
+		refs[num_refs++] = &obj->gem;
+
+		if (reloc->target.offset >= obj->gem.size) {
+			err = -EINVAL;
+			goto fail;
+		}
+	}
+
+	if (copy_from_user(&syncpt, user_syncpt, sizeof(syncpt))) {
+		err = -EFAULT;
+		goto fail;
+	}
+
+	/* Syncpoint ref will be dropped on job release. */
+	sp = host1x_syncpt_get_by_id(host1x, syncpt.id);
+	if (!sp) {
+		err = -ENOENT;
+		goto fail;
+	}
+
+	job->is_addr_reg = context->client->ops->is_addr_reg;
+	job->is_valid_class = context->client->ops->is_valid_class;
+	job->syncpt_incrs = syncpt.incrs;
+	job->syncpt = sp;
+	job->timeout = 10000;
+
+	if (args->timeout && args->timeout < 10000)
+		job->timeout = args->timeout;
+
+	err = host1x_job_pin(job, context->client->base.dev);
+	if (err)
+		goto fail;
+
+	err = host1x_job_submit(job);
+	if (err) {
+		host1x_job_unpin(job);
+		goto fail;
+	}
+
+	args->fence = job->syncpt_end;
+
+fail:
+	while (num_refs--)
+		drm_gem_object_put(refs[num_refs]);
+
+	kfree(refs);
+
+put:
+	host1x_job_put(job);
+	return err;
+}
+
+
+#ifdef CONFIG_DRM_TEGRA_STAGING
+static int tegra_gem_create(struct drm_device *drm, void *data,
+			    struct drm_file *file)
+{
+	struct drm_tegra_gem_create *args = data;
+	struct tegra_bo *bo;
+
+	bo = tegra_bo_create_with_handle(file, drm, args->size, args->flags,
+					 &args->handle);
+	if (IS_ERR(bo))
+		return PTR_ERR(bo);
+
+	return 0;
+}
+
+static int tegra_gem_mmap(struct drm_device *drm, void *data,
+			  struct drm_file *file)
+{
+	struct drm_tegra_gem_mmap *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	gem = drm_gem_object_lookup(file, args->handle);
+	if (!gem)
+		return -EINVAL;
+
+	bo = to_tegra_bo(gem);
+
+	args->offset = drm_vma_node_offset_addr(&bo->gem.vma_node);
+
+	drm_gem_object_put(gem);
+
+	return 0;
+}
+
+static int tegra_syncpt_read(struct drm_device *drm, void *data,
+			     struct drm_file *file)
+{
+	struct host1x *host = dev_get_drvdata(drm->dev->parent);
+	struct drm_tegra_syncpt_read *args = data;
+	struct host1x_syncpt *sp;
+
+	sp = host1x_syncpt_get_by_id_noref(host, args->id);
+	if (!sp)
+		return -EINVAL;
+
+	args->value = host1x_syncpt_read_min(sp);
+	return 0;
+}
+
+static int tegra_syncpt_incr(struct drm_device *drm, void *data,
+			     struct drm_file *file)
+{
+	struct host1x *host1x = dev_get_drvdata(drm->dev->parent);
+	struct drm_tegra_syncpt_incr *args = data;
+	struct host1x_syncpt *sp;
+
+	sp = host1x_syncpt_get_by_id_noref(host1x, args->id);
+	if (!sp)
+		return -EINVAL;
+
+	return host1x_syncpt_incr(sp);
+}
+
+static int tegra_syncpt_wait(struct drm_device *drm, void *data,
+			     struct drm_file *file)
+{
+	struct host1x *host1x = dev_get_drvdata(drm->dev->parent);
+	struct drm_tegra_syncpt_wait *args = data;
+	struct host1x_syncpt *sp;
+
+	sp = host1x_syncpt_get_by_id_noref(host1x, args->id);
+	if (!sp)
+		return -EINVAL;
+
+	return host1x_syncpt_wait(sp, args->thresh,
+				  msecs_to_jiffies(args->timeout),
+				  &args->value);
+}
+
+static int tegra_client_open(struct tegra_drm_file *fpriv,
+			     struct tegra_drm_client *client,
+			     struct tegra_drm_context *context)
+{
+	int err;
+
+	err = pm_runtime_resume_and_get(client->base.dev);
+	if (err)
+		return err;
+
+	err = client->ops->open_channel(client, context);
+	if (err < 0) {
+		pm_runtime_put(client->base.dev);
+		return err;
+	}
+
+	err = idr_alloc(&fpriv->legacy_contexts, context, 1, 0, GFP_KERNEL);
+	if (err < 0) {
+		client->ops->close_channel(context);
+		pm_runtime_put(client->base.dev);
+		return err;
+	}
+
+	context->client = client;
+	context->id = err;
+
+	return 0;
+}
+
+static int tegra_open_channel(struct drm_device *drm, void *data,
+			      struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct tegra_drm *tegra = drm->dev_private;
+	struct drm_tegra_open_channel *args = data;
+	struct tegra_drm_context *context;
+	struct tegra_drm_client *client;
+	int err = -ENODEV;
+
+	context = kzalloc_obj(*context);
+	if (!context)
+		return -ENOMEM;
+
+	mutex_lock(&fpriv->lock);
+
+	list_for_each_entry(client, &tegra->clients, list)
+		if (client->base.class == args->client) {
+			err = tegra_client_open(fpriv, client, context);
+			if (err < 0)
+				break;
+
+			args->context = context->id;
+			break;
+		}
+
+	if (err < 0)
+		kfree(context);
+
+	mutex_unlock(&fpriv->lock);
+	return err;
+}
+
+static int tegra_close_channel(struct drm_device *drm, void *data,
+			       struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_close_channel *args = data;
+	struct tegra_drm_context *context;
+	int err = 0;
+
+	mutex_lock(&fpriv->lock);
+
+	context = idr_find(&fpriv->legacy_contexts, args->context);
+	if (!context) {
+		err = -EINVAL;
+		goto unlock;
+	}
+
+	idr_remove(&fpriv->legacy_contexts, context->id);
+	tegra_drm_context_free(context);
+
+unlock:
+	mutex_unlock(&fpriv->lock);
+	return err;
+}
+
+static int tegra_get_syncpt(struct drm_device *drm, void *data,
+			    struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_get_syncpt *args = data;
+	struct tegra_drm_context *context;
+	struct host1x_syncpt *syncpt;
+	int err = 0;
+
+	mutex_lock(&fpriv->lock);
+
+	context = idr_find(&fpriv->legacy_contexts, args->context);
+	if (!context) {
+		err = -ENODEV;
+		goto unlock;
+	}
+
+	if (args->index >= context->client->base.num_syncpts) {
+		err = -EINVAL;
+		goto unlock;
+	}
+
+	syncpt = context->client->base.syncpts[args->index];
+	args->id = host1x_syncpt_id(syncpt);
+
+unlock:
+	mutex_unlock(&fpriv->lock);
+	return err;
+}
+
+static int tegra_submit(struct drm_device *drm, void *data,
+			struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_submit *args = data;
+	struct tegra_drm_context *context;
+	int err;
+
+	mutex_lock(&fpriv->lock);
+
+	context = idr_find(&fpriv->legacy_contexts, args->context);
+	if (!context) {
+		err = -ENODEV;
+		goto unlock;
+	}
+
+	err = context->client->ops->submit(context, args, drm, file);
+
+unlock:
+	mutex_unlock(&fpriv->lock);
+	return err;
+}
+
+static int tegra_get_syncpt_base(struct drm_device *drm, void *data,
+				 struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_get_syncpt_base *args = data;
+	struct tegra_drm_context *context;
+	struct host1x_syncpt_base *base;
+	struct host1x_syncpt *syncpt;
+	int err = 0;
+
+	mutex_lock(&fpriv->lock);
+
+	context = idr_find(&fpriv->legacy_contexts, args->context);
+	if (!context) {
+		err = -ENODEV;
+		goto unlock;
+	}
+
+	if (args->syncpt >= context->client->base.num_syncpts) {
+		err = -EINVAL;
+		goto unlock;
+	}
+
+	syncpt = context->client->base.syncpts[args->syncpt];
+
+	base = host1x_syncpt_get_base(syncpt);
+	if (!base) {
+		err = -ENXIO;
+		goto unlock;
+	}
+
+	args->id = host1x_syncpt_base_id(base);
+
+unlock:
+	mutex_unlock(&fpriv->lock);
+	return err;
+}
+
+static int tegra_gem_set_tiling(struct drm_device *drm, void *data,
+				struct drm_file *file)
+{
+	struct drm_tegra_gem_set_tiling *args = data;
+	enum tegra_bo_tiling_mode mode;
+	struct drm_gem_object *gem;
+	unsigned long value = 0;
+	struct tegra_bo *bo;
+
+	switch (args->mode) {
+	case DRM_TEGRA_GEM_TILING_MODE_PITCH:
+		mode = TEGRA_BO_TILING_MODE_PITCH;
+
+		if (args->value != 0)
+			return -EINVAL;
+
+		break;
+
+	case DRM_TEGRA_GEM_TILING_MODE_TILED:
+		mode = TEGRA_BO_TILING_MODE_TILED;
+
+		if (args->value != 0)
+			return -EINVAL;
+
+		break;
+
+	case DRM_TEGRA_GEM_TILING_MODE_BLOCK:
+		mode = TEGRA_BO_TILING_MODE_BLOCK;
+
+		if (args->value > 5)
+			return -EINVAL;
+
+		value = args->value;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	gem = drm_gem_object_lookup(file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+
+	bo->tiling.mode = mode;
+	bo->tiling.value = value;
+
+	drm_gem_object_put(gem);
+
+	return 0;
+}
+
+static int tegra_gem_get_tiling(struct drm_device *drm, void *data,
+				struct drm_file *file)
+{
+	struct drm_tegra_gem_get_tiling *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+	int err = 0;
+
+	gem = drm_gem_object_lookup(file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+
+	switch (bo->tiling.mode) {
+	case TEGRA_BO_TILING_MODE_PITCH:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_PITCH;
+		args->value = 0;
+		break;
+
+	case TEGRA_BO_TILING_MODE_TILED:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_TILED;
+		args->value = 0;
+		break;
+
+	case TEGRA_BO_TILING_MODE_BLOCK:
+		args->mode = DRM_TEGRA_GEM_TILING_MODE_BLOCK;
+		args->value = bo->tiling.value;
+		break;
+
+	default:
+		err = -EINVAL;
+		break;
+	}
+
+	drm_gem_object_put(gem);
+
+	return err;
+}
+
+static int tegra_gem_set_flags(struct drm_device *drm, void *data,
+			       struct drm_file *file)
+{
+	struct drm_tegra_gem_set_flags *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	if (args->flags & ~DRM_TEGRA_GEM_FLAGS)
+		return -EINVAL;
+
+	gem = drm_gem_object_lookup(file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+	bo->flags = 0;
+
+	if (args->flags & DRM_TEGRA_GEM_BOTTOM_UP)
+		bo->flags |= TEGRA_BO_BOTTOM_UP;
+
+	drm_gem_object_put(gem);
+
+	return 0;
+}
+
+static int tegra_gem_get_flags(struct drm_device *drm, void *data,
+			       struct drm_file *file)
+{
+	struct drm_tegra_gem_get_flags *args = data;
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	gem = drm_gem_object_lookup(file, args->handle);
+	if (!gem)
+		return -ENOENT;
+
+	bo = to_tegra_bo(gem);
+	args->flags = 0;
+
+	if (bo->flags & TEGRA_BO_BOTTOM_UP)
+		args->flags |= DRM_TEGRA_GEM_BOTTOM_UP;
+
+	drm_gem_object_put(gem);
+
+	return 0;
+}
+#endif
+
+static const struct drm_ioctl_desc tegra_drm_ioctls[] = {
+#ifdef CONFIG_DRM_TEGRA_STAGING
+	DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_OPEN, tegra_drm_ioctl_channel_open,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_CLOSE, tegra_drm_ioctl_channel_close,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_MAP, tegra_drm_ioctl_channel_map,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_UNMAP, tegra_drm_ioctl_channel_unmap,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_CHANNEL_SUBMIT, tegra_drm_ioctl_channel_submit,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_ALLOCATE, tegra_drm_ioctl_syncpoint_allocate,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_FREE, tegra_drm_ioctl_syncpoint_free,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPOINT_WAIT, tegra_drm_ioctl_syncpoint_wait,
+			  DRM_RENDER_ALLOW),
+
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_CREATE, tegra_gem_create, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_MMAP, tegra_gem_mmap, DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_READ, tegra_syncpt_read,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_INCR, tegra_syncpt_incr,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_SYNCPT_WAIT, tegra_syncpt_wait,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_OPEN_CHANNEL, tegra_open_channel,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_CLOSE_CHANNEL, tegra_close_channel,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_GET_SYNCPT, tegra_get_syncpt,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_SUBMIT, tegra_submit,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_GET_SYNCPT_BASE, tegra_get_syncpt_base,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_SET_TILING, tegra_gem_set_tiling,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_GET_TILING, tegra_gem_get_tiling,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_SET_FLAGS, tegra_gem_set_flags,
+			  DRM_RENDER_ALLOW),
+	DRM_IOCTL_DEF_DRV(TEGRA_GEM_GET_FLAGS, tegra_gem_get_flags,
+			  DRM_RENDER_ALLOW),
+#endif
+};
+
+static const struct file_operations tegra_drm_fops = {
+	.owner = THIS_MODULE,
+	.open = drm_open,
+	.release = drm_release,
+	.unlocked_ioctl = drm_ioctl,
+	.mmap = tegra_drm_mmap,
+	.poll = drm_poll,
+	.read = drm_read,
+	.compat_ioctl = drm_compat_ioctl,
+	.llseek = noop_llseek,
+	.fop_flags = FOP_UNSIGNED_OFFSET,
+};
+
+static int tegra_drm_context_cleanup(int id, void *p, void *data)
+{
+	struct tegra_drm_context *context = p;
+
+	tegra_drm_context_free(context);
+
+	return 0;
+}
+
+static void tegra_drm_postclose(struct drm_device *drm, struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+
+	mutex_lock(&fpriv->lock);
+	idr_for_each(&fpriv->legacy_contexts, tegra_drm_context_cleanup, NULL);
+	tegra_drm_uapi_close_file(fpriv);
+	mutex_unlock(&fpriv->lock);
+
+	idr_destroy(&fpriv->legacy_contexts);
+	mutex_destroy(&fpriv->lock);
+	kfree(fpriv);
+}
+
+#ifdef CONFIG_DEBUG_FS
+static int tegra_debugfs_framebuffers(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)s->private;
+	struct drm_device *drm = node->minor->dev;
+	struct drm_framebuffer *fb;
+
+	mutex_lock(&drm->mode_config.fb_lock);
+
+	list_for_each_entry(fb, &drm->mode_config.fb_list, head) {
+		seq_printf(s, "%3d: user size: %d x %d, depth %d, %d bpp, refcount %d\n",
+			   fb->base.id, fb->width, fb->height,
+			   fb->format->depth,
+			   fb->format->cpp[0] * 8,
+			   drm_framebuffer_read_refcount(fb));
+	}
+
+	mutex_unlock(&drm->mode_config.fb_lock);
+
+	return 0;
+}
+
+static int tegra_debugfs_iova(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *)s->private;
+	struct drm_device *drm = node->minor->dev;
+	struct tegra_drm *tegra = drm->dev_private;
+	struct drm_printer p = drm_seq_file_printer(s);
+
+	if (tegra->domain) {
+		mutex_lock(&tegra->mm_lock);
+		drm_mm_print(&tegra->mm, &p);
+		mutex_unlock(&tegra->mm_lock);
+	}
+
+	return 0;
+}
+
+static struct drm_info_list tegra_debugfs_list[] = {
+	{ "framebuffers", tegra_debugfs_framebuffers, 0 },
+	{ "iova", tegra_debugfs_iova, 0 },
+};
+
+static void tegra_debugfs_init(struct drm_minor *minor)
+{
+	drm_debugfs_create_files(tegra_debugfs_list,
+				 ARRAY_SIZE(tegra_debugfs_list),
+				 minor->debugfs_root, minor);
+}
+#endif
+
+static const struct drm_driver tegra_drm_driver = {
+	.driver_features = DRIVER_MODESET | DRIVER_GEM |
+			   DRIVER_ATOMIC | DRIVER_RENDER | DRIVER_SYNCOBJ,
+	.open = tegra_drm_open,
+	.postclose = tegra_drm_postclose,
+
+#if defined(CONFIG_DEBUG_FS)
+	.debugfs_init = tegra_debugfs_init,
+#endif
+
+	.gem_prime_import = tegra_gem_prime_import,
+
+	.dumb_create = tegra_bo_dumb_create,
+
+	TEGRA_FBDEV_DRIVER_OPS,
+
+	.ioctls = tegra_drm_ioctls,
+	.num_ioctls = ARRAY_SIZE(tegra_drm_ioctls),
+	.fops = &tegra_drm_fops,
+
+	.name = DRIVER_NAME,
+	.desc = DRIVER_DESC,
+	.major = DRIVER_MAJOR,
+	.minor = DRIVER_MINOR,
+	.patchlevel = DRIVER_PATCHLEVEL,
+};
+
+int tegra_drm_register_client(struct tegra_drm *tegra,
+			      struct tegra_drm_client *client)
+{
+	/*
+	 * When MLOCKs are implemented, change to allocate a shared channel
+	 * only when MLOCKs are disabled.
+	 */
+	client->shared_channel = host1x_channel_request(&client->base);
+	if (!client->shared_channel)
+		return -EBUSY;
+
+	mutex_lock(&tegra->clients_lock);
+	list_add_tail(&client->list, &tegra->clients);
+	client->drm = tegra;
+	mutex_unlock(&tegra->clients_lock);
+
+	return 0;
+}
+
+int tegra_drm_unregister_client(struct tegra_drm *tegra,
+				struct tegra_drm_client *client)
+{
+	mutex_lock(&tegra->clients_lock);
+	list_del_init(&client->list);
+	client->drm = NULL;
+	mutex_unlock(&tegra->clients_lock);
+
+	if (client->shared_channel)
+		host1x_channel_put(client->shared_channel);
+
+	return 0;
+}
+
+int host1x_client_iommu_attach(struct host1x_client *client)
+{
+	struct iommu_domain *domain = iommu_get_domain_for_dev(client->dev);
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = drm->dev_private;
+	struct iommu_group *group = NULL;
+	int err;
+
+#if IS_ENABLED(CONFIG_ARM_DMA_USE_IOMMU)
+	if (client->dev->archdata.mapping) {
+		struct dma_iommu_mapping *mapping =
+				to_dma_iommu_mapping(client->dev);
+		arm_iommu_detach_device(client->dev);
+		arm_iommu_release_mapping(mapping);
+
+		domain = iommu_get_domain_for_dev(client->dev);
+	}
+#endif
+
+	/*
+	 * If the host1x client is already attached to an IOMMU domain that is
+	 * not the shared IOMMU domain, don't try to attach it to a different
+	 * domain. This allows using the IOMMU-backed DMA API.
+	 */
+	if (domain && domain->type != IOMMU_DOMAIN_IDENTITY &&
+	    domain != tegra->domain)
+		return 0;
+
+	if (tegra->domain) {
+		group = iommu_group_get(client->dev);
+		if (!group)
+			return -ENODEV;
+
+		if (domain != tegra->domain) {
+			err = iommu_attach_group(tegra->domain, group);
+			if (err < 0) {
+				iommu_group_put(group);
+				return err;
+			}
+		}
+
+		tegra->use_explicit_iommu = true;
+	}
+
+	client->group = group;
+
+	return 0;
+}
+
+void host1x_client_iommu_detach(struct host1x_client *client)
+{
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = drm->dev_private;
+	struct iommu_domain *domain;
+
+	if (client->group) {
+		/*
+		 * Devices that are part of the same group may no longer be
+		 * attached to a domain at this point because their group may
+		 * have been detached by an earlier client.
+		 */
+		domain = iommu_get_domain_for_dev(client->dev);
+		if (domain)
+			iommu_detach_group(tegra->domain, client->group);
+
+		iommu_group_put(client->group);
+		client->group = NULL;
+	}
+}
+
+void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *dma)
+{
+	struct iova *alloc;
+	void *virt;
+	gfp_t gfp;
+	int err;
+
+	if (tegra->domain)
+		size = iova_align(&tegra->carveout.domain, size);
+	else
+		size = PAGE_ALIGN(size);
+
+	gfp = GFP_KERNEL | __GFP_ZERO;
+	if (!tegra->domain) {
+		/*
+		 * Many units only support 32-bit addresses, even on 64-bit
+		 * SoCs. If there is no IOMMU to translate into a 32-bit IO
+		 * virtual address space, force allocations to be in the
+		 * lower 32-bit range.
+		 */
+		gfp |= GFP_DMA;
+	}
+
+	virt = (void *)__get_free_pages(gfp, get_order(size));
+	if (!virt)
+		return ERR_PTR(-ENOMEM);
+
+	if (!tegra->domain) {
+		/*
+		 * If IOMMU is disabled, devices address physical memory
+		 * directly.
+		 */
+		*dma = virt_to_phys(virt);
+		return virt;
+	}
+
+	alloc = alloc_iova(&tegra->carveout.domain,
+			   size >> tegra->carveout.shift,
+			   tegra->carveout.limit, true);
+	if (!alloc) {
+		err = -EBUSY;
+		goto free_pages;
+	}
+
+	*dma = iova_dma_addr(&tegra->carveout.domain, alloc);
+	err = iommu_map(tegra->domain, *dma, virt_to_phys(virt),
+			size, IOMMU_READ | IOMMU_WRITE, GFP_KERNEL);
+	if (err < 0)
+		goto free_iova;
+
+	return virt;
+
+free_iova:
+	__free_iova(&tegra->carveout.domain, alloc);
+free_pages:
+	free_pages((unsigned long)virt, get_order(size));
+
+	return ERR_PTR(err);
+}
+
+void tegra_drm_free(struct tegra_drm *tegra, size_t size, void *virt,
+		    dma_addr_t dma)
+{
+	if (tegra->domain)
+		size = iova_align(&tegra->carveout.domain, size);
+	else
+		size = PAGE_ALIGN(size);
+
+	if (tegra->domain) {
+		iommu_unmap(tegra->domain, dma, size);
+		free_iova(&tegra->carveout.domain,
+			  iova_pfn(&tegra->carveout.domain, dma));
+	}
+
+	free_pages((unsigned long)virt, get_order(size));
+}
+
+static bool host1x_drm_wants_iommu(struct host1x_device *dev)
+{
+	struct host1x *host1x = dev_get_drvdata(dev->dev.parent);
+	struct iommu_domain *domain;
+
+	/* Our IOMMU usage policy doesn't currently play well with GART */
+	if (of_machine_is_compatible("nvidia,tegra20"))
+		return false;
+
+	/*
+	 * If the Tegra DRM clients are backed by an IOMMU, push buffers are
+	 * likely to be allocated beyond the 32-bit boundary if sufficient
+	 * system memory is available. This is problematic on earlier Tegra
+	 * generations where host1x supports a maximum of 32 address bits in
+	 * the GATHER opcode. In this case, unless host1x is behind an IOMMU
+	 * as well it won't be able to process buffers allocated beyond the
+	 * 32-bit boundary.
+	 *
+	 * The DMA API will use bounce buffers in this case, so that could
+	 * perhaps still be made to work, even if less efficient, but there
+	 * is another catch: in order to perform cache maintenance on pages
+	 * allocated for discontiguous buffers we need to map and unmap the
+	 * SG table representing these buffers. This is fine for something
+	 * small like a push buffer, but it exhausts the bounce buffer pool
+	 * (typically on the order of a few MiB) for framebuffers (many MiB
+	 * for any modern resolution).
+	 *
+	 * Work around this by making sure that Tegra DRM clients only use
+	 * an IOMMU if the parent host1x also uses an IOMMU.
+	 *
+	 * Note that there's still a small gap here that we don't cover: if
+	 * the DMA API is backed by an IOMMU there's no way to control which
+	 * device is attached to an IOMMU and which isn't, except via wiring
+	 * up the device tree appropriately. This is considered an problem
+	 * of integration, so care must be taken for the DT to be consistent.
+	 */
+	domain = iommu_get_domain_for_dev(dev->dev.parent);
+
+	/*
+	 * Tegra20 and Tegra30 don't support addressing memory beyond the
+	 * 32-bit boundary, so the regular GATHER opcodes will always be
+	 * sufficient and whether or not the host1x is attached to an IOMMU
+	 * doesn't matter.
+	 */
+	if (!domain && host1x_get_dma_mask(host1x) <= DMA_BIT_MASK(32))
+		return true;
+
+	return domain != NULL;
+}
+
+static int host1x_drm_probe(struct host1x_device *dev)
+{
+	struct device *dma_dev = dev->dev.parent;
+	struct tegra_drm *tegra;
+	struct drm_device *drm;
+	int err;
+
+	drm = drm_dev_alloc(&tegra_drm_driver, &dev->dev);
+	if (IS_ERR(drm))
+		return PTR_ERR(drm);
+
+	tegra = kzalloc_obj(*tegra);
+	if (!tegra) {
+		err = -ENOMEM;
+		goto put;
+	}
+
+	if (host1x_drm_wants_iommu(dev) && device_iommu_mapped(dma_dev)) {
+		tegra->domain = iommu_paging_domain_alloc(dma_dev);
+		if (IS_ERR(tegra->domain)) {
+			err = PTR_ERR(tegra->domain);
+			goto free;
+		}
+
+		err = iova_cache_get();
+		if (err < 0)
+			goto domain;
+	}
+
+	mutex_init(&tegra->clients_lock);
+	INIT_LIST_HEAD(&tegra->clients);
+
+	dev_set_drvdata(&dev->dev, drm);
+	drm->dev_private = tegra;
+	tegra->drm = drm;
+
+	drm_mode_config_init(drm);
+
+	drm->mode_config.min_width = 0;
+	drm->mode_config.min_height = 0;
+	drm->mode_config.max_width = 0;
+	drm->mode_config.max_height = 0;
+
+	drm->mode_config.normalize_zpos = true;
+
+	drm->mode_config.funcs = &tegra_drm_mode_config_funcs;
+	drm->mode_config.helper_private = &tegra_drm_mode_config_helpers;
+
+	drm_kms_helper_poll_init(drm);
+
+	err = host1x_device_init(dev);
+	if (err < 0)
+		goto poll;
+
+	/*
+	 * Now that all display controller have been initialized, the maximum
+	 * supported resolution is known and the bitmask for horizontal and
+	 * vertical bitfields can be computed.
+	 */
+	tegra->hmask = drm->mode_config.max_width - 1;
+	tegra->vmask = drm->mode_config.max_height - 1;
+
+	if (tegra->use_explicit_iommu) {
+		u64 carveout_start, carveout_end, gem_start, gem_end;
+		u64 dma_mask = dma_get_mask(&dev->dev);
+		dma_addr_t start, end;
+		unsigned long order;
+
+		start = tegra->domain->geometry.aperture_start & dma_mask;
+		end = tegra->domain->geometry.aperture_end & dma_mask;
+
+		gem_start = start;
+		gem_end = end - CARVEOUT_SZ;
+		carveout_start = gem_end + 1;
+		carveout_end = end;
+
+		order = __ffs(tegra->domain->pgsize_bitmap);
+		init_iova_domain(&tegra->carveout.domain, 1UL << order,
+				 carveout_start >> order);
+
+		tegra->carveout.shift = iova_shift(&tegra->carveout.domain);
+		tegra->carveout.limit = carveout_end >> tegra->carveout.shift;
+
+		drm_mm_init(&tegra->mm, gem_start, gem_end - gem_start + 1);
+		mutex_init(&tegra->mm_lock);
+
+		DRM_DEBUG_DRIVER("IOMMU apertures:\n");
+		DRM_DEBUG_DRIVER("  GEM: %#llx-%#llx\n", gem_start, gem_end);
+		DRM_DEBUG_DRIVER("  Carveout: %#llx-%#llx\n", carveout_start,
+				 carveout_end);
+	} else if (tegra->domain) {
+		iommu_domain_free(tegra->domain);
+		tegra->domain = NULL;
+		iova_cache_put();
+	}
+
+	if (tegra->hub) {
+		err = tegra_display_hub_prepare(tegra->hub);
+		if (err < 0)
+			goto device;
+	}
+
+	/* syncpoints are used for full 32-bit hardware VBLANK counters */
+	drm->max_vblank_count = 0xffffffff;
+
+	err = drm_vblank_init(drm, drm->mode_config.num_crtc);
+	if (err < 0)
+		goto hub;
+
+	drm_mode_config_reset(drm);
+
+	/*
+	 * Only take over from a potential firmware framebuffer if any CRTCs
+	 * have been registered. This must not be a fatal error because there
+	 * are other accelerators that are exposed via this driver.
+	 *
+	 * Another case where this happens is on Tegra234 where the display
+	 * hardware is no longer part of the host1x complex, so this driver
+	 * will not expose any modesetting features.
+	 */
+	if (drm->mode_config.num_crtc > 0) {
+		err = aperture_remove_all_conflicting_devices(tegra_drm_driver.name);
+		if (err < 0)
+			goto hub;
+	} else {
+		/*
+		 * Indicate to userspace that this doesn't expose any display
+		 * capabilities.
+		 */
+		drm->driver_features &= ~(DRIVER_MODESET | DRIVER_ATOMIC);
+	}
+
+	err = drm_dev_register(drm, 0);
+	if (err < 0)
+		goto hub;
+
+	drm_client_setup(drm, NULL);
+
+	return 0;
+
+hub:
+	if (tegra->hub)
+		tegra_display_hub_cleanup(tegra->hub);
+device:
+	if (tegra->domain) {
+		mutex_destroy(&tegra->mm_lock);
+		drm_mm_takedown(&tegra->mm);
+		put_iova_domain(&tegra->carveout.domain);
+		iova_cache_put();
+	}
+
+	host1x_device_exit(dev);
+poll:
+	drm_kms_helper_poll_fini(drm);
+	drm_mode_config_cleanup(drm);
+domain:
+	if (tegra->domain)
+		iommu_domain_free(tegra->domain);
+free:
+	kfree(tegra);
+put:
+	drm_dev_put(drm);
+	return err;
+}
+
+static void host1x_drm_remove(struct host1x_device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(&dev->dev);
+	struct tegra_drm *tegra = drm->dev_private;
+	int err;
+
+	drm_dev_unregister(drm);
+
+	drm_kms_helper_poll_fini(drm);
+	drm_atomic_helper_shutdown(drm);
+	drm_mode_config_cleanup(drm);
+
+	if (tegra->hub)
+		tegra_display_hub_cleanup(tegra->hub);
+
+	err = host1x_device_exit(dev);
+	if (err < 0)
+		dev_err(&dev->dev, "host1x device cleanup failed: %d\n", err);
+
+	if (tegra->domain) {
+		mutex_destroy(&tegra->mm_lock);
+		drm_mm_takedown(&tegra->mm);
+		put_iova_domain(&tegra->carveout.domain);
+		iova_cache_put();
+		iommu_domain_free(tegra->domain);
+	}
+
+	kfree(tegra);
+	drm_dev_put(drm);
+}
+
+static void host1x_drm_shutdown(struct host1x_device *dev)
+{
+	drm_atomic_helper_shutdown(dev_get_drvdata(&dev->dev));
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int host1x_drm_suspend(struct device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(dev);
+
+	return drm_mode_config_helper_suspend(drm);
+}
+
+static int host1x_drm_resume(struct device *dev)
+{
+	struct drm_device *drm = dev_get_drvdata(dev);
+
+	return drm_mode_config_helper_resume(drm);
+}
+#endif
+
+static SIMPLE_DEV_PM_OPS(host1x_drm_pm_ops, host1x_drm_suspend,
+			 host1x_drm_resume);
+
+static const struct of_device_id host1x_drm_subdevs[] = {
+	{ .compatible = "nvidia,tegra20-dc", },
+	{ .compatible = "nvidia,tegra20-dsi", },
+	{ .compatible = "nvidia,tegra20-hdmi", },
+	{ .compatible = "nvidia,tegra20-gr2d", },
+	{ .compatible = "nvidia,tegra20-gr3d", },
+	{ .compatible = "nvidia,tegra30-dc", },
+	{ .compatible = "nvidia,tegra30-dsi", },
+	{ .compatible = "nvidia,tegra30-hdmi", },
+	{ .compatible = "nvidia,tegra30-gr2d", },
+	{ .compatible = "nvidia,tegra30-gr3d", },
+	{ .compatible = "nvidia,tegra114-dc", },
+	{ .compatible = "nvidia,tegra114-dsi", },
+	{ .compatible = "nvidia,tegra114-hdmi", },
+	{ .compatible = "nvidia,tegra114-gr2d", },
+	{ .compatible = "nvidia,tegra114-gr3d", },
+	{ .compatible = "nvidia,tegra124-dc", },
+	{ .compatible = "nvidia,tegra124-sor", },
+	{ .compatible = "nvidia,tegra124-hdmi", },
+	{ .compatible = "nvidia,tegra124-dsi", },
+	{ .compatible = "nvidia,tegra124-vic", },
+	{ .compatible = "nvidia,tegra132-dsi", },
+	{ .compatible = "nvidia,tegra210-dc", },
+	{ .compatible = "nvidia,tegra210-dsi", },
+	{ .compatible = "nvidia,tegra210-sor", },
+	{ .compatible = "nvidia,tegra210-sor1", },
+	{ .compatible = "nvidia,tegra210-vic", },
+	{ .compatible = "nvidia,tegra210-nvdec", },
+	{ .compatible = "nvidia,tegra210-nvjpg", },
+	{ .compatible = "nvidia,tegra186-display", },
+	{ .compatible = "nvidia,tegra186-dc", },
+	{ .compatible = "nvidia,tegra186-sor", },
+	{ .compatible = "nvidia,tegra186-sor1", },
+	{ .compatible = "nvidia,tegra186-vic", },
+	{ .compatible = "nvidia,tegra186-nvdec", },
+	{ .compatible = "nvidia,tegra194-display", },
+	{ .compatible = "nvidia,tegra194-dc", },
+	{ .compatible = "nvidia,tegra194-sor", },
+	{ .compatible = "nvidia,tegra194-vic", },
+	{ .compatible = "nvidia,tegra194-nvdec", },
+	{ .compatible = "nvidia,tegra234-vic", },
+	{ .compatible = "nvidia,tegra234-nvdec", },
+	{ .compatible = "nvidia,tegra264-vic", },
+	{ /* sentinel */ }
+};
+
+static struct host1x_driver host1x_drm_driver = {
+	.driver = {
+		.name = "drm",
+		.pm = &host1x_drm_pm_ops,
+	},
+	.probe = host1x_drm_probe,
+	.remove = host1x_drm_remove,
+	.shutdown = host1x_drm_shutdown,
+	.subdevs = host1x_drm_subdevs,
+};
+
+static struct platform_driver * const drivers[] = {
+	&tegra_display_hub_driver,
+	&tegra_dc_driver,
+	&tegra_hdmi_driver,
+	&tegra_dsi_driver,
+	&tegra_dpaux_driver,
+	&tegra_sor_driver,
+	&tegra_gr2d_driver,
+	&tegra_gr3d_driver,
+	&tegra_vic_driver,
+	&tegra_nvdec_driver,
+	&tegra_nvjpg_driver,
+};
+
+static int __init host1x_drm_init(void)
+{
+	int err;
+
+	if (drm_firmware_drivers_only())
+		return -ENODEV;
+
+	err = host1x_driver_register(&host1x_drm_driver);
+	if (err < 0)
+		return err;
+
+	err = platform_register_drivers(drivers, ARRAY_SIZE(drivers));
+	if (err < 0)
+		goto unregister_host1x;
+
+	return 0;
+
+unregister_host1x:
+	host1x_driver_unregister(&host1x_drm_driver);
+	return err;
+}
+module_init(host1x_drm_init);
+
+static void __exit host1x_drm_exit(void)
+{
+	platform_unregister_drivers(drivers, ARRAY_SIZE(drivers));
+	host1x_driver_unregister(&host1x_drm_driver);
+}
+module_exit(host1x_drm_exit);
+
+MODULE_AUTHOR("Thierry Reding <thierry.reding@avionic-design.de>");
+MODULE_DESCRIPTION("NVIDIA Tegra DRM driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h
new file mode 100644
index 000000000..bc720ae8d
--- /dev/null
+++ b/drivers/gpu/drm/tegra/drm.h
@@ -0,0 +1,214 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012-2013 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#ifndef HOST1X_DRM_H
+#define HOST1X_DRM_H 1
+
+#include <linux/host1x.h>
+#include <linux/iova.h>
+#include <linux/gpio/consumer.h>
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_bridge.h>
+#include <drm/drm_encoder.h>
+#include <drm/drm_fixed.h>
+#include <drm/drm_probe_helper.h>
+#include <uapi/drm/tegra_drm.h>
+
+#include "gem.h"
+#include "hub.h"
+#include "trace.h"
+
+/* XXX move to include/uapi/drm/drm_fourcc.h? */
+#define DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT BIT_ULL(22)
+
+struct drm_fb_helper;
+struct drm_fb_helper_surface_size;
+
+struct edid;
+struct reset_control;
+
+struct tegra_drm {
+	struct drm_device *drm;
+
+	struct iommu_domain *domain;
+	bool use_explicit_iommu;
+	struct mutex mm_lock;
+	struct drm_mm mm;
+
+	struct {
+		struct iova_domain domain;
+		unsigned long shift;
+		unsigned long limit;
+	} carveout;
+
+	struct mutex clients_lock;
+	struct list_head clients;
+
+	unsigned int hmask, vmask;
+	unsigned int pitch_align;
+	unsigned int num_crtcs;
+
+	struct tegra_display_hub *hub;
+};
+
+static inline struct host1x *tegra_drm_to_host1x(struct tegra_drm *tegra)
+{
+	return dev_get_drvdata(tegra->drm->dev->parent);
+}
+
+struct tegra_drm_client;
+
+struct tegra_drm_context {
+	struct tegra_drm_client *client;
+	struct host1x_channel *channel;
+
+	/* Only used by legacy UAPI. */
+	unsigned int id;
+
+	/* Only used by new UAPI. */
+	struct xarray mappings;
+	struct host1x_memory_context *memory_context;
+};
+
+struct tegra_drm_client_ops {
+	int (*open_channel)(struct tegra_drm_client *client,
+			    struct tegra_drm_context *context);
+	void (*close_channel)(struct tegra_drm_context *context);
+	int (*is_addr_reg)(struct device *dev, u32 class, u32 offset);
+	int (*is_valid_class)(u32 class);
+	int (*submit)(struct tegra_drm_context *context,
+		      struct drm_tegra_submit *args, struct drm_device *drm,
+		      struct drm_file *file);
+	int (*get_streamid_offset)(struct tegra_drm_client *client, u32 *offset);
+	int (*can_use_memory_ctx)(struct tegra_drm_client *client, bool *supported);
+};
+
+int tegra_drm_submit(struct tegra_drm_context *context,
+		     struct drm_tegra_submit *args, struct drm_device *drm,
+		     struct drm_file *file);
+
+static inline int
+tegra_drm_get_streamid_offset_thi(struct tegra_drm_client *client, u32 *offset)
+{
+	*offset = 0x30;
+
+	return 0;
+}
+
+struct tegra_drm_client {
+	struct host1x_client base;
+	struct list_head list;
+	struct tegra_drm *drm;
+	struct host1x_channel *shared_channel;
+
+	/* Set by driver */
+	unsigned int version;
+	const struct tegra_drm_client_ops *ops;
+};
+
+static inline struct tegra_drm_client *
+host1x_to_drm_client(struct host1x_client *client)
+{
+	return container_of(client, struct tegra_drm_client, base);
+}
+
+int tegra_drm_register_client(struct tegra_drm *tegra,
+			      struct tegra_drm_client *client);
+int tegra_drm_unregister_client(struct tegra_drm *tegra,
+				struct tegra_drm_client *client);
+int host1x_client_iommu_attach(struct host1x_client *client);
+void host1x_client_iommu_detach(struct host1x_client *client);
+
+void *tegra_drm_alloc(struct tegra_drm *tegra, size_t size, dma_addr_t *iova);
+void tegra_drm_free(struct tegra_drm *tegra, size_t size, void *virt,
+		    dma_addr_t iova);
+
+struct cec_notifier;
+
+struct tegra_output {
+	struct device_node *of_node;
+	struct device *dev;
+
+	struct drm_bridge *bridge;
+	struct drm_panel *panel;
+	struct i2c_adapter *ddc;
+	const struct drm_edid *drm_edid;
+	struct cec_notifier *cec;
+	unsigned int hpd_irq;
+	struct gpio_desc *hpd_gpio;
+
+	struct drm_encoder encoder;
+	struct drm_connector connector;
+};
+
+static inline struct tegra_output *encoder_to_output(struct drm_encoder *e)
+{
+	return container_of(e, struct tegra_output, encoder);
+}
+
+static inline struct tegra_output *connector_to_output(struct drm_connector *c)
+{
+	return container_of(c, struct tegra_output, connector);
+}
+
+/* from output.c */
+int tegra_output_probe(struct tegra_output *output);
+void tegra_output_remove(struct tegra_output *output);
+int tegra_output_init(struct drm_device *drm, struct tegra_output *output);
+void tegra_output_exit(struct tegra_output *output);
+void tegra_output_find_possible_crtcs(struct tegra_output *output,
+				      struct drm_device *drm);
+int tegra_output_suspend(struct tegra_output *output);
+int tegra_output_resume(struct tegra_output *output);
+
+int tegra_output_connector_get_modes(struct drm_connector *connector);
+enum drm_connector_status
+tegra_output_connector_detect(struct drm_connector *connector, bool force);
+void tegra_output_connector_destroy(struct drm_connector *connector);
+
+/* from dpaux.c */
+struct drm_dp_aux *drm_dp_aux_find_by_of_node(struct device_node *np);
+enum drm_connector_status drm_dp_aux_detect(struct drm_dp_aux *aux);
+int drm_dp_aux_attach(struct drm_dp_aux *aux, struct tegra_output *output);
+int drm_dp_aux_detach(struct drm_dp_aux *aux);
+int drm_dp_aux_enable(struct drm_dp_aux *aux);
+int drm_dp_aux_disable(struct drm_dp_aux *aux);
+
+/* from fb.c */
+struct tegra_bo *tegra_fb_get_plane(struct drm_framebuffer *framebuffer,
+				    unsigned int index);
+bool tegra_fb_is_bottom_up(struct drm_framebuffer *framebuffer);
+int tegra_fb_get_tiling(struct drm_framebuffer *framebuffer,
+			struct tegra_bo_tiling *tiling);
+struct drm_framebuffer *tegra_fb_create(struct drm_device *drm,
+					struct drm_file *file,
+					const struct drm_format_info *info,
+					const struct drm_mode_fb_cmd2 *cmd);
+
+#ifdef CONFIG_DRM_FBDEV_EMULATION
+int tegra_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
+				   struct drm_fb_helper_surface_size *sizes);
+#define TEGRA_FBDEV_DRIVER_OPS \
+	.fbdev_probe = tegra_fbdev_driver_fbdev_probe
+#else
+#define TEGRA_FBDEV_DRIVER_OPS \
+	.fbdev_probe = NULL
+#endif
+
+extern struct platform_driver tegra_display_hub_driver;
+extern struct platform_driver tegra_dc_driver;
+extern struct platform_driver tegra_hdmi_driver;
+extern struct platform_driver tegra_dsi_driver;
+extern struct platform_driver tegra_dpaux_driver;
+extern struct platform_driver tegra_sor_driver;
+extern struct platform_driver tegra_gr2d_driver;
+extern struct platform_driver tegra_gr3d_driver;
+extern struct platform_driver tegra_vic_driver;
+extern struct platform_driver tegra_nvdec_driver;
+extern struct platform_driver tegra_nvjpg_driver;
+
+#endif /* HOST1X_DRM_H */
diff --git a/drivers/gpu/drm/tegra/dsi.c b/drivers/gpu/drm/tegra/dsi.c
new file mode 100644
index 000000000..cb88aafbd
--- /dev/null
+++ b/drivers/gpu/drm/tegra/dsi.c
@@ -0,0 +1,1776 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#include <linux/clk.h>
+#include <linux/debugfs.h>
+#include <linux/delay.h>
+#include <linux/host1x.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/reset.h>
+#include <linux/tegra-mipi-cal.h>
+
+#include <video/mipi_display.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_debugfs.h>
+#include <drm/drm_file.h>
+#include <drm/drm_mipi_dsi.h>
+#include <drm/drm_panel.h>
+#include <drm/drm_print.h>
+#include <drm/drm_simple_kms_helper.h>
+
+#include "dc.h"
+#include "drm.h"
+#include "dsi.h"
+#include "mipi-phy.h"
+#include "trace.h"
+
+struct tegra_dsi_state {
+	struct drm_connector_state base;
+
+	struct mipi_dphy_timing timing;
+	unsigned long period;
+
+	unsigned int vrefresh;
+	unsigned int lanes;
+	unsigned long pclk;
+	unsigned long bclk;
+
+	enum tegra_dsi_format format;
+	unsigned int mul;
+	unsigned int div;
+};
+
+static inline struct tegra_dsi_state *
+to_dsi_state(struct drm_connector_state *state)
+{
+	return container_of(state, struct tegra_dsi_state, base);
+}
+
+struct tegra_dsi_config {
+	bool has_multiple_pad_controls;
+	bool has_mux_parent_clk;
+};
+
+struct tegra_dsi {
+	struct host1x_client client;
+	struct tegra_output output;
+	struct device *dev;
+
+	void __iomem *regs;
+
+	struct reset_control *rst;
+	struct clk *clk_parent;
+	struct clk *clk_lp;
+	struct clk *clk;
+
+	struct drm_info_list *debugfs_files;
+
+	unsigned long flags;
+	enum mipi_dsi_pixel_format format;
+	unsigned int lanes;
+
+	struct tegra_mipi_device *mipi;
+	struct mipi_dsi_host host;
+
+	struct regulator *vdd;
+
+	unsigned int video_fifo_depth;
+	unsigned int host_fifo_depth;
+
+	/* for ganged-mode support */
+	struct tegra_dsi *master;
+	struct tegra_dsi *slave;
+
+	const struct tegra_dsi_config *config;
+};
+
+static inline struct tegra_dsi *
+host1x_client_to_dsi(struct host1x_client *client)
+{
+	return container_of(client, struct tegra_dsi, client);
+}
+
+static inline struct tegra_dsi *host_to_tegra(struct mipi_dsi_host *host)
+{
+	return container_of(host, struct tegra_dsi, host);
+}
+
+static inline struct tegra_dsi *to_dsi(struct tegra_output *output)
+{
+	return container_of(output, struct tegra_dsi, output);
+}
+
+static struct tegra_dsi_state *tegra_dsi_get_state(struct tegra_dsi *dsi)
+{
+	return to_dsi_state(dsi->output.connector.state);
+}
+
+static inline u32 tegra_dsi_readl(struct tegra_dsi *dsi, unsigned int offset)
+{
+	u32 value = readl(dsi->regs + (offset << 2));
+
+	trace_dsi_readl(dsi->dev, offset, value);
+
+	return value;
+}
+
+static inline void tegra_dsi_writel(struct tegra_dsi *dsi, u32 value,
+				    unsigned int offset)
+{
+	trace_dsi_writel(dsi->dev, offset, value);
+	writel(value, dsi->regs + (offset << 2));
+}
+
+#define DEBUGFS_REG32(_name) { .name = #_name, .offset = _name }
+
+static const struct debugfs_reg32 tegra_dsi_regs[] = {
+	DEBUGFS_REG32(DSI_INCR_SYNCPT),
+	DEBUGFS_REG32(DSI_INCR_SYNCPT_CONTROL),
+	DEBUGFS_REG32(DSI_INCR_SYNCPT_ERROR),
+	DEBUGFS_REG32(DSI_CTXSW),
+	DEBUGFS_REG32(DSI_RD_DATA),
+	DEBUGFS_REG32(DSI_WR_DATA),
+	DEBUGFS_REG32(DSI_POWER_CONTROL),
+	DEBUGFS_REG32(DSI_INT_ENABLE),
+	DEBUGFS_REG32(DSI_INT_STATUS),
+	DEBUGFS_REG32(DSI_INT_MASK),
+	DEBUGFS_REG32(DSI_HOST_CONTROL),
+	DEBUGFS_REG32(DSI_CONTROL),
+	DEBUGFS_REG32(DSI_SOL_DELAY),
+	DEBUGFS_REG32(DSI_MAX_THRESHOLD),
+	DEBUGFS_REG32(DSI_TRIGGER),
+	DEBUGFS_REG32(DSI_TX_CRC),
+	DEBUGFS_REG32(DSI_STATUS),
+	DEBUGFS_REG32(DSI_INIT_SEQ_CONTROL),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_0),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_1),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_2),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_3),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_4),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_5),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_6),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_7),
+	DEBUGFS_REG32(DSI_PKT_SEQ_0_LO),
+	DEBUGFS_REG32(DSI_PKT_SEQ_0_HI),
+	DEBUGFS_REG32(DSI_PKT_SEQ_1_LO),
+	DEBUGFS_REG32(DSI_PKT_SEQ_1_HI),
+	DEBUGFS_REG32(DSI_PKT_SEQ_2_LO),
+	DEBUGFS_REG32(DSI_PKT_SEQ_2_HI),
+	DEBUGFS_REG32(DSI_PKT_SEQ_3_LO),
+	DEBUGFS_REG32(DSI_PKT_SEQ_3_HI),
+	DEBUGFS_REG32(DSI_PKT_SEQ_4_LO),
+	DEBUGFS_REG32(DSI_PKT_SEQ_4_HI),
+	DEBUGFS_REG32(DSI_PKT_SEQ_5_LO),
+	DEBUGFS_REG32(DSI_PKT_SEQ_5_HI),
+	DEBUGFS_REG32(DSI_DCS_CMDS),
+	DEBUGFS_REG32(DSI_PKT_LEN_0_1),
+	DEBUGFS_REG32(DSI_PKT_LEN_2_3),
+	DEBUGFS_REG32(DSI_PKT_LEN_4_5),
+	DEBUGFS_REG32(DSI_PKT_LEN_6_7),
+	DEBUGFS_REG32(DSI_PHY_TIMING_0),
+	DEBUGFS_REG32(DSI_PHY_TIMING_1),
+	DEBUGFS_REG32(DSI_PHY_TIMING_2),
+	DEBUGFS_REG32(DSI_BTA_TIMING),
+	DEBUGFS_REG32(DSI_TIMEOUT_0),
+	DEBUGFS_REG32(DSI_TIMEOUT_1),
+	DEBUGFS_REG32(DSI_TO_TALLY),
+	DEBUGFS_REG32(DSI_PAD_CONTROL_0),
+	DEBUGFS_REG32(DSI_PAD_CONTROL_CD),
+	DEBUGFS_REG32(DSI_PAD_CD_STATUS),
+	DEBUGFS_REG32(DSI_VIDEO_MODE_CONTROL),
+	DEBUGFS_REG32(DSI_PAD_CONTROL_1),
+	DEBUGFS_REG32(DSI_PAD_CONTROL_2),
+	DEBUGFS_REG32(DSI_PAD_CONTROL_3),
+	DEBUGFS_REG32(DSI_PAD_CONTROL_4),
+	DEBUGFS_REG32(DSI_GANGED_MODE_CONTROL),
+	DEBUGFS_REG32(DSI_GANGED_MODE_START),
+	DEBUGFS_REG32(DSI_GANGED_MODE_SIZE),
+	DEBUGFS_REG32(DSI_RAW_DATA_BYTE_COUNT),
+	DEBUGFS_REG32(DSI_ULTRA_LOW_POWER_CONTROL),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_8),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_9),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_10),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_11),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_12),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_13),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_14),
+	DEBUGFS_REG32(DSI_INIT_SEQ_DATA_15),
+};
+
+static int tegra_dsi_show_regs(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = s->private;
+	struct tegra_dsi *dsi = node->info_ent->data;
+	struct drm_crtc *crtc = dsi->output.encoder.crtc;
+	struct drm_device *drm = node->minor->dev;
+	unsigned int i;
+	int err = 0;
+
+	drm_modeset_lock_all(drm);
+
+	if (!crtc || !crtc->state->active) {
+		err = -EBUSY;
+		goto unlock;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(tegra_dsi_regs); i++) {
+		unsigned int offset = tegra_dsi_regs[i].offset;
+
+		seq_printf(s, "%-32s %#05x %08x\n", tegra_dsi_regs[i].name,
+			   offset, tegra_dsi_readl(dsi, offset));
+	}
+
+unlock:
+	drm_modeset_unlock_all(drm);
+	return err;
+}
+
+static struct drm_info_list debugfs_files[] = {
+	{ "regs", tegra_dsi_show_regs, 0, NULL },
+};
+
+static int tegra_dsi_late_register(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	unsigned int i, count = ARRAY_SIZE(debugfs_files);
+	struct drm_minor *minor = connector->dev->primary;
+	struct dentry *root = connector->debugfs_entry;
+	struct tegra_dsi *dsi = to_dsi(output);
+
+	dsi->debugfs_files = kmemdup(debugfs_files, sizeof(debugfs_files),
+				     GFP_KERNEL);
+	if (!dsi->debugfs_files)
+		return -ENOMEM;
+
+	for (i = 0; i < count; i++)
+		dsi->debugfs_files[i].data = dsi;
+
+	drm_debugfs_create_files(dsi->debugfs_files, count, root, minor);
+
+	return 0;
+}
+
+static void tegra_dsi_early_unregister(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	unsigned int count = ARRAY_SIZE(debugfs_files);
+	struct tegra_dsi *dsi = to_dsi(output);
+
+	drm_debugfs_remove_files(dsi->debugfs_files, count,
+				 connector->debugfs_entry,
+				 connector->dev->primary);
+	kfree(dsi->debugfs_files);
+	dsi->debugfs_files = NULL;
+}
+
+#define PKT_ID0(id)	((((id) & 0x3f) <<  3) | (1 <<  9))
+#define PKT_LEN0(len)	(((len) & 0x07) <<  0)
+#define PKT_ID1(id)	((((id) & 0x3f) << 13) | (1 << 19))
+#define PKT_LEN1(len)	(((len) & 0x07) << 10)
+#define PKT_ID2(id)	((((id) & 0x3f) << 23) | (1 << 29))
+#define PKT_LEN2(len)	(((len) & 0x07) << 20)
+
+#define PKT_LP		(1 << 30)
+#define NUM_PKT_SEQ	12
+
+/*
+ * non-burst mode with sync pulses
+ */
+static const u32 pkt_seq_video_non_burst_sync_pulses[NUM_PKT_SEQ] = {
+	[ 0] = PKT_ID0(MIPI_DSI_V_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_BLANKING_PACKET) | PKT_LEN1(1) |
+	       PKT_ID2(MIPI_DSI_H_SYNC_END) | PKT_LEN2(0) |
+	       PKT_LP,
+	[ 1] = 0,
+	[ 2] = PKT_ID0(MIPI_DSI_V_SYNC_END) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_BLANKING_PACKET) | PKT_LEN1(1) |
+	       PKT_ID2(MIPI_DSI_H_SYNC_END) | PKT_LEN2(0) |
+	       PKT_LP,
+	[ 3] = 0,
+	[ 4] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_BLANKING_PACKET) | PKT_LEN1(1) |
+	       PKT_ID2(MIPI_DSI_H_SYNC_END) | PKT_LEN2(0) |
+	       PKT_LP,
+	[ 5] = 0,
+	[ 6] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_BLANKING_PACKET) | PKT_LEN1(1) |
+	       PKT_ID2(MIPI_DSI_H_SYNC_END) | PKT_LEN2(0),
+	[ 7] = PKT_ID0(MIPI_DSI_BLANKING_PACKET) | PKT_LEN0(2) |
+	       PKT_ID1(MIPI_DSI_PACKED_PIXEL_STREAM_24) | PKT_LEN1(3) |
+	       PKT_ID2(MIPI_DSI_BLANKING_PACKET) | PKT_LEN2(4),
+	[ 8] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_BLANKING_PACKET) | PKT_LEN1(1) |
+	       PKT_ID2(MIPI_DSI_H_SYNC_END) | PKT_LEN2(0) |
+	       PKT_LP,
+	[ 9] = 0,
+	[10] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_BLANKING_PACKET) | PKT_LEN1(1) |
+	       PKT_ID2(MIPI_DSI_H_SYNC_END) | PKT_LEN2(0),
+	[11] = PKT_ID0(MIPI_DSI_BLANKING_PACKET) | PKT_LEN0(2) |
+	       PKT_ID1(MIPI_DSI_PACKED_PIXEL_STREAM_24) | PKT_LEN1(3) |
+	       PKT_ID2(MIPI_DSI_BLANKING_PACKET) | PKT_LEN2(4),
+};
+
+/*
+ * non-burst mode with sync events
+ */
+static const u32 pkt_seq_video_non_burst_sync_events[NUM_PKT_SEQ] = {
+	[ 0] = PKT_ID0(MIPI_DSI_V_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_END_OF_TRANSMISSION) | PKT_LEN1(7) |
+	       PKT_LP,
+	[ 1] = 0,
+	[ 2] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_END_OF_TRANSMISSION) | PKT_LEN1(7) |
+	       PKT_LP,
+	[ 3] = 0,
+	[ 4] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_END_OF_TRANSMISSION) | PKT_LEN1(7) |
+	       PKT_LP,
+	[ 5] = 0,
+	[ 6] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_BLANKING_PACKET) | PKT_LEN1(2) |
+	       PKT_ID2(MIPI_DSI_PACKED_PIXEL_STREAM_24) | PKT_LEN2(3),
+	[ 7] = PKT_ID0(MIPI_DSI_BLANKING_PACKET) | PKT_LEN0(4),
+	[ 8] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_END_OF_TRANSMISSION) | PKT_LEN1(7) |
+	       PKT_LP,
+	[ 9] = 0,
+	[10] = PKT_ID0(MIPI_DSI_H_SYNC_START) | PKT_LEN0(0) |
+	       PKT_ID1(MIPI_DSI_BLANKING_PACKET) | PKT_LEN1(2) |
+	       PKT_ID2(MIPI_DSI_PACKED_PIXEL_STREAM_24) | PKT_LEN2(3),
+	[11] = PKT_ID0(MIPI_DSI_BLANKING_PACKET) | PKT_LEN0(4),
+};
+
+static const u32 pkt_seq_command_mode[NUM_PKT_SEQ] = {
+	[ 0] = 0,
+	[ 1] = 0,
+	[ 2] = 0,
+	[ 3] = 0,
+	[ 4] = 0,
+	[ 5] = 0,
+	[ 6] = PKT_ID0(MIPI_DSI_DCS_LONG_WRITE) | PKT_LEN0(3) | PKT_LP,
+	[ 7] = 0,
+	[ 8] = 0,
+	[ 9] = 0,
+	[10] = PKT_ID0(MIPI_DSI_DCS_LONG_WRITE) | PKT_LEN0(5) | PKT_LP,
+	[11] = 0,
+};
+
+static void tegra_dsi_set_phy_timing(struct tegra_dsi *dsi,
+				     unsigned long period,
+				     const struct mipi_dphy_timing *timing)
+{
+	u32 value;
+
+	value = DSI_TIMING_FIELD(timing->hsexit, period, 1) << 24 |
+		DSI_TIMING_FIELD(timing->hstrail, period, 0) << 16 |
+		DSI_TIMING_FIELD(timing->hszero, period, 3) << 8 |
+		DSI_TIMING_FIELD(timing->hsprepare, period, 1);
+	tegra_dsi_writel(dsi, value, DSI_PHY_TIMING_0);
+
+	value = DSI_TIMING_FIELD(timing->clktrail, period, 1) << 24 |
+		DSI_TIMING_FIELD(timing->clkpost, period, 1) << 16 |
+		DSI_TIMING_FIELD(timing->clkzero, period, 1) << 8 |
+		DSI_TIMING_FIELD(timing->lpx, period, 1);
+	tegra_dsi_writel(dsi, value, DSI_PHY_TIMING_1);
+
+	value = DSI_TIMING_FIELD(timing->clkprepare, period, 1) << 16 |
+		DSI_TIMING_FIELD(timing->clkpre, period, 1) << 8 |
+		DSI_TIMING_FIELD(0xff * period, period, 0) << 0;
+	tegra_dsi_writel(dsi, value, DSI_PHY_TIMING_2);
+
+	value = DSI_TIMING_FIELD(timing->taget, period, 1) << 16 |
+		DSI_TIMING_FIELD(timing->tasure, period, 1) << 8 |
+		DSI_TIMING_FIELD(timing->tago, period, 1);
+	tegra_dsi_writel(dsi, value, DSI_BTA_TIMING);
+
+	if (dsi->slave)
+		tegra_dsi_set_phy_timing(dsi->slave, period, timing);
+}
+
+static int tegra_dsi_get_muldiv(enum mipi_dsi_pixel_format format,
+				unsigned int *mulp, unsigned int *divp)
+{
+	switch (format) {
+	case MIPI_DSI_FMT_RGB666_PACKED:
+	case MIPI_DSI_FMT_RGB888:
+		*mulp = 3;
+		*divp = 1;
+		break;
+
+	case MIPI_DSI_FMT_RGB565:
+		*mulp = 2;
+		*divp = 1;
+		break;
+
+	case MIPI_DSI_FMT_RGB666:
+		*mulp = 9;
+		*divp = 4;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int tegra_dsi_get_format(enum mipi_dsi_pixel_format format,
+				enum tegra_dsi_format *fmt)
+{
+	switch (format) {
+	case MIPI_DSI_FMT_RGB888:
+		*fmt = TEGRA_DSI_FORMAT_24P;
+		break;
+
+	case MIPI_DSI_FMT_RGB666:
+		*fmt = TEGRA_DSI_FORMAT_18NP;
+		break;
+
+	case MIPI_DSI_FMT_RGB666_PACKED:
+		*fmt = TEGRA_DSI_FORMAT_18P;
+		break;
+
+	case MIPI_DSI_FMT_RGB565:
+		*fmt = TEGRA_DSI_FORMAT_16P;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static void tegra_dsi_ganged_enable(struct tegra_dsi *dsi, unsigned int start,
+				    unsigned int size)
+{
+	u32 value;
+
+	tegra_dsi_writel(dsi, start, DSI_GANGED_MODE_START);
+	tegra_dsi_writel(dsi, size << 16 | size, DSI_GANGED_MODE_SIZE);
+
+	value = DSI_GANGED_MODE_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_GANGED_MODE_CONTROL);
+}
+
+static void tegra_dsi_enable(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+	value |= DSI_POWER_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+
+	if (dsi->slave)
+		tegra_dsi_enable(dsi->slave);
+}
+
+static unsigned int tegra_dsi_get_lanes(struct tegra_dsi *dsi)
+{
+	if (dsi->master)
+		return dsi->master->lanes + dsi->lanes;
+
+	if (dsi->slave)
+		return dsi->lanes + dsi->slave->lanes;
+
+	return dsi->lanes;
+}
+
+static void tegra_dsi_configure(struct tegra_dsi *dsi, unsigned int pipe,
+				const struct drm_display_mode *mode)
+{
+	unsigned int hact, hsw, hbp, hfp, i, mul, div;
+	struct tegra_dsi_state *state;
+	const u32 *pkt_seq;
+	u32 value;
+
+	/* XXX: pass in state into this function? */
+	if (dsi->master)
+		state = tegra_dsi_get_state(dsi->master);
+	else
+		state = tegra_dsi_get_state(dsi);
+
+	mul = state->mul;
+	div = state->div;
+
+	if (dsi->flags & MIPI_DSI_MODE_VIDEO_SYNC_PULSE) {
+		DRM_DEBUG_KMS("Non-burst video mode with sync pulses\n");
+		pkt_seq = pkt_seq_video_non_burst_sync_pulses;
+	} else if (dsi->flags & MIPI_DSI_MODE_VIDEO) {
+		DRM_DEBUG_KMS("Non-burst video mode with sync events\n");
+		pkt_seq = pkt_seq_video_non_burst_sync_events;
+	} else {
+		DRM_DEBUG_KMS("Command mode\n");
+		pkt_seq = pkt_seq_command_mode;
+	}
+
+	value = DSI_CONTROL_CHANNEL(0) |
+		DSI_CONTROL_FORMAT(state->format) |
+		DSI_CONTROL_LANES(dsi->lanes - 1) |
+		DSI_CONTROL_SOURCE(pipe);
+	tegra_dsi_writel(dsi, value, DSI_CONTROL);
+
+	tegra_dsi_writel(dsi, dsi->video_fifo_depth, DSI_MAX_THRESHOLD);
+
+	value = DSI_HOST_CONTROL_HS;
+	tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
+
+	value = tegra_dsi_readl(dsi, DSI_CONTROL);
+
+	if (dsi->flags & MIPI_DSI_CLOCK_NON_CONTINUOUS)
+		value |= DSI_CONTROL_HS_CLK_CTRL;
+
+	value &= ~DSI_CONTROL_TX_TRIG(3);
+
+	/* enable DCS commands for command mode */
+	if (dsi->flags & MIPI_DSI_MODE_VIDEO)
+		value &= ~DSI_CONTROL_DCS_ENABLE;
+	else
+		value |= DSI_CONTROL_DCS_ENABLE;
+
+	value |= DSI_CONTROL_VIDEO_ENABLE;
+	value &= ~DSI_CONTROL_HOST_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_CONTROL);
+
+	for (i = 0; i < NUM_PKT_SEQ; i++)
+		tegra_dsi_writel(dsi, pkt_seq[i], DSI_PKT_SEQ_0_LO + i);
+
+	if (dsi->flags & MIPI_DSI_MODE_VIDEO) {
+		/* horizontal active pixels */
+		hact = mode->hdisplay * mul / div;
+
+		/* horizontal sync width */
+		hsw = (mode->hsync_end - mode->hsync_start) * mul / div;
+
+		/* horizontal back porch */
+		hbp = (mode->htotal - mode->hsync_end) * mul / div;
+
+		/* horizontal front porch */
+		hfp = (mode->hsync_start - mode->hdisplay) * mul / div;
+
+		if (dsi->master || dsi->slave) {
+			hact /= 2;
+			hsw /= 2;
+			hbp /= 2;
+			hfp /= 2;
+		}
+
+		if ((dsi->flags & MIPI_DSI_MODE_VIDEO_SYNC_PULSE) == 0)
+			hbp += hsw;
+
+		/* subtract packet overhead */
+		hsw -= 10;
+		hbp -= 14;
+		hfp -= 8;
+
+		tegra_dsi_writel(dsi, hsw << 16 | 0, DSI_PKT_LEN_0_1);
+		tegra_dsi_writel(dsi, hact << 16 | hbp, DSI_PKT_LEN_2_3);
+		tegra_dsi_writel(dsi, hfp, DSI_PKT_LEN_4_5);
+		tegra_dsi_writel(dsi, 0x0f0f << 16, DSI_PKT_LEN_6_7);
+	} else {
+		u16 bytes;
+
+		if (dsi->master || dsi->slave) {
+			/*
+			 * For ganged mode, assume symmetric left-right mode.
+			 */
+			bytes = 1 + (mode->hdisplay / 2) * mul / div;
+		} else {
+			/* 1 byte (DCS command) + pixel data */
+			bytes = 1 + mode->hdisplay * mul / div;
+		}
+
+		tegra_dsi_writel(dsi, 0, DSI_PKT_LEN_0_1);
+		tegra_dsi_writel(dsi, bytes << 16, DSI_PKT_LEN_2_3);
+		tegra_dsi_writel(dsi, bytes << 16, DSI_PKT_LEN_4_5);
+		tegra_dsi_writel(dsi, 0, DSI_PKT_LEN_6_7);
+
+		value = MIPI_DCS_WRITE_MEMORY_START << 8 |
+			MIPI_DCS_WRITE_MEMORY_CONTINUE;
+		tegra_dsi_writel(dsi, value, DSI_DCS_CMDS);
+	}
+
+	/* set SOL delay */
+	if (dsi->master || dsi->slave) {
+		unsigned long delay, bclk, bclk_ganged;
+		unsigned int lanes = state->lanes;
+
+		/* SOL to valid, valid to FIFO and FIFO write delay */
+		delay = 4 + 4 + 2;
+		delay = DIV_ROUND_UP(delay * mul, div * lanes);
+		/* FIFO read delay */
+		delay = delay + 6;
+
+		bclk = DIV_ROUND_UP(mode->htotal * mul, div * lanes);
+		bclk_ganged = DIV_ROUND_UP(bclk * lanes / 2, lanes);
+		value = bclk - bclk_ganged + delay + 20;
+	} else {
+		value = 8 * mul / div;
+	}
+
+	tegra_dsi_writel(dsi, value, DSI_SOL_DELAY);
+
+	if (dsi->slave) {
+		tegra_dsi_configure(dsi->slave, pipe, mode);
+
+		/*
+		 * TODO: Support modes other than symmetrical left-right
+		 * split.
+		 */
+		tegra_dsi_ganged_enable(dsi, 0, mode->hdisplay / 2);
+		tegra_dsi_ganged_enable(dsi->slave, mode->hdisplay / 2,
+					mode->hdisplay / 2);
+	}
+}
+
+static int tegra_dsi_wait_idle(struct tegra_dsi *dsi, unsigned long timeout)
+{
+	u32 value;
+
+	timeout = jiffies + msecs_to_jiffies(timeout);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_dsi_readl(dsi, DSI_STATUS);
+		if (value & DSI_STATUS_IDLE)
+			return 0;
+
+		usleep_range(1000, 2000);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static void tegra_dsi_video_disable(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	value = tegra_dsi_readl(dsi, DSI_CONTROL);
+	value &= ~DSI_CONTROL_VIDEO_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_CONTROL);
+
+	if (dsi->slave)
+		tegra_dsi_video_disable(dsi->slave);
+}
+
+static void tegra_dsi_ganged_disable(struct tegra_dsi *dsi)
+{
+	tegra_dsi_writel(dsi, 0, DSI_GANGED_MODE_START);
+	tegra_dsi_writel(dsi, 0, DSI_GANGED_MODE_SIZE);
+	tegra_dsi_writel(dsi, 0, DSI_GANGED_MODE_CONTROL);
+}
+
+static int tegra_dsi_pad_enable(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	if (dsi->config->has_multiple_pad_controls) {
+		/*
+		 * XXX Is this still needed? The module reset is deasserted right
+		 * before this function is called.
+		 */
+		tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_0);
+		tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_1);
+		tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_2);
+		tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_3);
+		tegra_dsi_writel(dsi, 0, DSI_PAD_CONTROL_4);
+
+		value = DSI_PAD_CONTROL_VS1_PULLDN(0) | DSI_PAD_CONTROL_VS1_PDIO(0);
+		tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_0);
+
+		value = DSI_PAD_SLEW_UP(0x7) | DSI_PAD_SLEW_DN(0x7) |
+			DSI_PAD_LP_UP(0x1) | DSI_PAD_LP_DN(0x1) |
+			DSI_PAD_OUT_CLK(0x0);
+		tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_2);
+
+		value = DSI_PAD_PREEMP_PD_CLK(0x3) | DSI_PAD_PREEMP_PU_CLK(0x3) |
+			DSI_PAD_PREEMP_PD(0x03) | DSI_PAD_PREEMP_PU(0x3);
+		tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_3);
+	} else {
+		value = DSI_PAD_CONTROL_LPUPADJ(0x1) | DSI_PAD_CONTROL_LPDNADJ(0x1) |
+			DSI_PAD_CONTROL_PREEMP_EN(0x1) | DSI_PAD_CONTROL_SLEWDNADJ(0x6) |
+			DSI_PAD_CONTROL_SLEWUPADJ(0x6) | DSI_PAD_CONTROL_PDIO(0) |
+			DSI_PAD_CONTROL_PDIO_CLK(0) | DSI_PAD_CONTROL_PULLDN_ENAB(0);
+		tegra_dsi_writel(dsi, value, DSI_PAD_CONTROL_0);
+	}
+
+	return 0;
+}
+
+static int tegra_dsi_pad_calibrate(struct tegra_dsi *dsi)
+{
+	int err;
+
+	/* start calibration */
+	tegra_dsi_pad_enable(dsi);
+
+	err = tegra_mipi_start_calibration(dsi->mipi);
+	if (err < 0)
+		return err;
+
+	return tegra_mipi_finish_calibration(dsi->mipi);
+}
+
+static void tegra_dsi_set_timeout(struct tegra_dsi *dsi, unsigned long bclk,
+				  unsigned int vrefresh)
+{
+	unsigned int timeout;
+	u32 value;
+
+	/* one frame high-speed transmission timeout */
+	timeout = (bclk / vrefresh) / 512;
+	value = DSI_TIMEOUT_LRX(0x2000) | DSI_TIMEOUT_HTX(timeout);
+	tegra_dsi_writel(dsi, value, DSI_TIMEOUT_0);
+
+	/* 2 ms peripheral timeout for panel */
+	timeout = 2 * bclk / 512 * 1000;
+	value = DSI_TIMEOUT_PR(timeout) | DSI_TIMEOUT_TA(0x2000);
+	tegra_dsi_writel(dsi, value, DSI_TIMEOUT_1);
+
+	value = DSI_TALLY_TA(0) | DSI_TALLY_LRX(0) | DSI_TALLY_HTX(0);
+	tegra_dsi_writel(dsi, value, DSI_TO_TALLY);
+
+	if (dsi->slave)
+		tegra_dsi_set_timeout(dsi->slave, bclk, vrefresh);
+}
+
+static void tegra_dsi_disable(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	if (dsi->slave) {
+		tegra_dsi_ganged_disable(dsi->slave);
+		tegra_dsi_ganged_disable(dsi);
+	}
+
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+	value &= ~DSI_POWER_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+
+	if (dsi->slave)
+		tegra_dsi_disable(dsi->slave);
+
+	usleep_range(5000, 10000);
+}
+
+static void tegra_dsi_soft_reset(struct tegra_dsi *dsi)
+{
+	u32 value;
+
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+	value &= ~DSI_POWER_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+
+	usleep_range(300, 1000);
+
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+	value |= DSI_POWER_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+
+	usleep_range(300, 1000);
+
+	value = tegra_dsi_readl(dsi, DSI_TRIGGER);
+	if (value)
+		tegra_dsi_writel(dsi, 0, DSI_TRIGGER);
+
+	if (dsi->slave)
+		tegra_dsi_soft_reset(dsi->slave);
+}
+
+static void tegra_dsi_connector_reset(struct drm_connector *connector)
+{
+	struct tegra_dsi_state *state = kzalloc_obj(*state);
+
+	if (!state)
+		return;
+
+	if (connector->state) {
+		__drm_atomic_helper_connector_destroy_state(connector->state);
+		kfree(connector->state);
+	}
+
+	__drm_atomic_helper_connector_reset(connector, &state->base);
+}
+
+static struct drm_connector_state *
+tegra_dsi_connector_duplicate_state(struct drm_connector *connector)
+{
+	struct tegra_dsi_state *state = to_dsi_state(connector->state);
+	struct tegra_dsi_state *copy;
+
+	copy = kmemdup(state, sizeof(*state), GFP_KERNEL);
+	if (!copy)
+		return NULL;
+
+	__drm_atomic_helper_connector_duplicate_state(connector,
+						      &copy->base);
+
+	return &copy->base;
+}
+
+static const struct drm_connector_funcs tegra_dsi_connector_funcs = {
+	.reset = tegra_dsi_connector_reset,
+	.detect = tegra_output_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = tegra_output_connector_destroy,
+	.atomic_duplicate_state = tegra_dsi_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.late_register = tegra_dsi_late_register,
+	.early_unregister = tegra_dsi_early_unregister,
+};
+
+static enum drm_mode_status
+tegra_dsi_connector_mode_valid(struct drm_connector *connector,
+			       const struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static const struct drm_connector_helper_funcs tegra_dsi_connector_helper_funcs = {
+	.get_modes = tegra_output_connector_get_modes,
+	.mode_valid = tegra_dsi_connector_mode_valid,
+};
+
+static void tegra_dsi_unprepare(struct tegra_dsi *dsi)
+{
+	int err;
+
+	if (dsi->slave)
+		tegra_dsi_unprepare(dsi->slave);
+
+	err = tegra_mipi_disable(dsi->mipi);
+	if (err < 0)
+		dev_err(dsi->dev, "failed to disable MIPI calibration: %d\n",
+			err);
+
+	err = host1x_client_suspend(&dsi->client);
+	if (err < 0)
+		dev_err(dsi->dev, "failed to suspend: %d\n", err);
+}
+
+static void tegra_dsi_encoder_disable(struct drm_encoder *encoder)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_dsi *dsi = to_dsi(output);
+	u32 value;
+	int err;
+
+	if (output->panel)
+		drm_panel_disable(output->panel);
+
+	tegra_dsi_video_disable(dsi);
+
+	/*
+	 * The following accesses registers of the display controller, so make
+	 * sure it's only executed when the output is attached to one.
+	 */
+	if (dc) {
+		value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+		value &= ~DSI_ENABLE;
+		tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+		tegra_dc_commit(dc);
+	}
+
+	err = tegra_dsi_wait_idle(dsi, 100);
+	if (err < 0)
+		dev_dbg(dsi->dev, "failed to idle DSI: %d\n", err);
+
+	tegra_dsi_soft_reset(dsi);
+
+	if (output->panel)
+		drm_panel_unprepare(output->panel);
+
+	tegra_dsi_disable(dsi);
+
+	tegra_dsi_unprepare(dsi);
+}
+
+static int tegra_dsi_prepare(struct tegra_dsi *dsi)
+{
+	int err;
+
+	err = host1x_client_resume(&dsi->client);
+	if (err < 0) {
+		dev_err(dsi->dev, "failed to resume: %d\n", err);
+		return err;
+	}
+
+	err = tegra_mipi_enable(dsi->mipi);
+	if (err < 0)
+		dev_err(dsi->dev, "failed to enable MIPI calibration: %d\n",
+			err);
+
+	err = tegra_dsi_pad_calibrate(dsi);
+	if (err < 0)
+		dev_err(dsi->dev, "MIPI calibration failed: %d\n", err);
+
+	if (dsi->slave)
+		tegra_dsi_prepare(dsi->slave);
+
+	return 0;
+}
+
+static void tegra_dsi_encoder_enable(struct drm_encoder *encoder)
+{
+	struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_dsi *dsi = to_dsi(output);
+	struct tegra_dsi_state *state;
+	u32 value;
+	int err;
+
+	err = tegra_dsi_prepare(dsi);
+	if (err < 0) {
+		dev_err(dsi->dev, "failed to prepare: %d\n", err);
+		return;
+	}
+
+	/* If the bootloader enabled DSI it needs to be disabled
+	 * in order for the panel initialization commands to be
+	 * properly sent.
+	 */
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+
+	if (value & DSI_POWER_CONTROL_ENABLE)
+		tegra_dsi_disable(dsi);
+
+	state = tegra_dsi_get_state(dsi);
+
+	tegra_dsi_set_timeout(dsi, state->bclk, state->vrefresh);
+
+	/*
+	 * The D-PHY timing fields are expressed in byte-clock cycles, so
+	 * multiply the period by 8.
+	 */
+	tegra_dsi_set_phy_timing(dsi, state->period * 8, &state->timing);
+
+	if (output->panel)
+		drm_panel_prepare(output->panel);
+
+	tegra_dsi_configure(dsi, dc->pipe, mode);
+
+	/* enable display controller */
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+	value |= DSI_ENABLE;
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+	tegra_dc_commit(dc);
+
+	/* enable DSI controller */
+	tegra_dsi_enable(dsi);
+
+	if (output->panel)
+		drm_panel_enable(output->panel);
+}
+
+static int
+tegra_dsi_encoder_atomic_check(struct drm_encoder *encoder,
+			       struct drm_crtc_state *crtc_state,
+			       struct drm_connector_state *conn_state)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dsi_state *state = to_dsi_state(conn_state);
+	struct tegra_dc *dc = to_tegra_dc(conn_state->crtc);
+	struct tegra_dsi *dsi = to_dsi(output);
+	unsigned int scdiv;
+	unsigned long plld;
+	int err;
+
+	state->pclk = crtc_state->mode.clock * 1000;
+
+	err = tegra_dsi_get_muldiv(dsi->format, &state->mul, &state->div);
+	if (err < 0)
+		return err;
+
+	state->lanes = tegra_dsi_get_lanes(dsi);
+
+	err = tegra_dsi_get_format(dsi->format, &state->format);
+	if (err < 0)
+		return err;
+
+	state->vrefresh = drm_mode_vrefresh(&crtc_state->mode);
+
+	/* compute byte clock */
+	state->bclk = (state->pclk * state->mul) / (state->div * state->lanes);
+
+	DRM_DEBUG_KMS("mul: %u, div: %u, lanes: %u\n", state->mul, state->div,
+		      state->lanes);
+	DRM_DEBUG_KMS("format: %u, vrefresh: %u\n", state->format,
+		      state->vrefresh);
+	DRM_DEBUG_KMS("bclk: %lu\n", state->bclk);
+
+	/*
+	 * Compute bit clock and round up to the next MHz.
+	 */
+	plld = DIV_ROUND_UP(state->bclk * 8, USEC_PER_SEC) * USEC_PER_SEC;
+	state->period = DIV_ROUND_CLOSEST(NSEC_PER_SEC, plld);
+
+	err = mipi_dphy_timing_get_default(&state->timing, state->period);
+	if (err < 0)
+		return err;
+
+	err = mipi_dphy_timing_validate(&state->timing, state->period);
+	if (err < 0) {
+		dev_err(dsi->dev, "failed to validate D-PHY timing: %d\n", err);
+		return err;
+	}
+
+	/*
+	 * We divide the frequency by two here, but we make up for that by
+	 * setting the shift clock divider (further below) to half of the
+	 * correct value.
+	 */
+	plld /= 2;
+
+	/*
+	 * Derive pixel clock from bit clock using the shift clock divider.
+	 * Note that this is only half of what we would expect, but we need
+	 * that to make up for the fact that we divided the bit clock by a
+	 * factor of two above.
+	 *
+	 * It's not clear exactly why this is necessary, but the display is
+	 * not working properly otherwise. Perhaps the PLLs cannot generate
+	 * frequencies sufficiently high.
+	 */
+	scdiv = ((8 * state->mul) / (state->div * state->lanes)) - 2;
+
+	err = tegra_dc_state_setup_clock(dc, crtc_state, dsi->clk_parent,
+					 plld, scdiv);
+	if (err < 0) {
+		dev_err(output->dev, "failed to setup CRTC state: %d\n", err);
+		return err;
+	}
+
+	return err;
+}
+
+static const struct drm_encoder_helper_funcs tegra_dsi_encoder_helper_funcs = {
+	.disable = tegra_dsi_encoder_disable,
+	.enable = tegra_dsi_encoder_enable,
+	.atomic_check = tegra_dsi_encoder_atomic_check,
+};
+
+static int tegra_dsi_init(struct host1x_client *client)
+{
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	struct tegra_dsi *dsi = host1x_client_to_dsi(client);
+	int err;
+
+	/* Gangsters must not register their own outputs. */
+	if (!dsi->master) {
+		dsi->output.dev = client->dev;
+
+		drm_connector_init(drm, &dsi->output.connector,
+				   &tegra_dsi_connector_funcs,
+				   DRM_MODE_CONNECTOR_DSI);
+		drm_connector_helper_add(&dsi->output.connector,
+					 &tegra_dsi_connector_helper_funcs);
+		dsi->output.connector.dpms = DRM_MODE_DPMS_OFF;
+
+		drm_simple_encoder_init(drm, &dsi->output.encoder,
+					DRM_MODE_ENCODER_DSI);
+		drm_encoder_helper_add(&dsi->output.encoder,
+				       &tegra_dsi_encoder_helper_funcs);
+
+		drm_connector_attach_encoder(&dsi->output.connector,
+						  &dsi->output.encoder);
+		drm_connector_register(&dsi->output.connector);
+
+		err = tegra_output_init(drm, &dsi->output);
+		if (err < 0)
+			dev_err(dsi->dev, "failed to initialize output: %d\n",
+				err);
+
+		dsi->output.encoder.possible_crtcs = 0x3;
+	}
+
+	return 0;
+}
+
+static int tegra_dsi_exit(struct host1x_client *client)
+{
+	struct tegra_dsi *dsi = host1x_client_to_dsi(client);
+
+	tegra_output_exit(&dsi->output);
+
+	return 0;
+}
+
+static int tegra_dsi_runtime_suspend(struct host1x_client *client)
+{
+	struct tegra_dsi *dsi = host1x_client_to_dsi(client);
+	struct device *dev = client->dev;
+	int err;
+
+	if (dsi->rst) {
+		err = reset_control_assert(dsi->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to assert reset: %d\n", err);
+			return err;
+		}
+	}
+
+	usleep_range(1000, 2000);
+
+	clk_disable_unprepare(dsi->clk_lp);
+	clk_disable_unprepare(dsi->clk);
+
+	regulator_disable(dsi->vdd);
+	pm_runtime_put_sync(dev);
+
+	return 0;
+}
+
+static int tegra_dsi_runtime_resume(struct host1x_client *client)
+{
+	struct tegra_dsi *dsi = host1x_client_to_dsi(client);
+	struct device *dev = client->dev;
+	int err;
+
+	err = pm_runtime_resume_and_get(dev);
+	if (err < 0) {
+		dev_err(dev, "failed to get runtime PM: %d\n", err);
+		return err;
+	}
+
+	err = regulator_enable(dsi->vdd);
+	if (err < 0) {
+		dev_err(dev, "failed to enable VDD supply: %d\n", err);
+		goto put_rpm;
+	}
+
+	err = clk_prepare_enable(dsi->clk);
+	if (err < 0) {
+		dev_err(dev, "cannot enable DSI clock: %d\n", err);
+		goto disable_vdd;
+	}
+
+	err = clk_prepare_enable(dsi->clk_lp);
+	if (err < 0) {
+		dev_err(dev, "cannot enable low-power clock: %d\n", err);
+		goto disable_clk;
+	}
+
+	usleep_range(1000, 2000);
+
+	if (dsi->rst) {
+		err = reset_control_deassert(dsi->rst);
+		if (err < 0) {
+			dev_err(dev, "cannot assert reset: %d\n", err);
+			goto disable_clk_lp;
+		}
+	}
+
+	return 0;
+
+disable_clk_lp:
+	clk_disable_unprepare(dsi->clk_lp);
+disable_clk:
+	clk_disable_unprepare(dsi->clk);
+disable_vdd:
+	regulator_disable(dsi->vdd);
+put_rpm:
+	pm_runtime_put_sync(dev);
+	return err;
+}
+
+static const struct host1x_client_ops dsi_client_ops = {
+	.init = tegra_dsi_init,
+	.exit = tegra_dsi_exit,
+	.suspend = tegra_dsi_runtime_suspend,
+	.resume = tegra_dsi_runtime_resume,
+};
+
+static int tegra_dsi_setup_clocks(struct tegra_dsi *dsi)
+{
+	struct clk *parent;
+	int err;
+
+	/*
+	 * Tegra124+ uses a clock gate, not a mux, so this step
+	 * should be redundant for configuration; yet, DSI refuses
+	 * to work without it.
+	 */
+
+	parent = clk_get_parent(dsi->clk);
+	if (!parent)
+		return -EINVAL;
+
+	err = clk_set_parent(parent, dsi->clk_parent);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static const char * const error_report[16] = {
+	"SoT Error",
+	"SoT Sync Error",
+	"EoT Sync Error",
+	"Escape Mode Entry Command Error",
+	"Low-Power Transmit Sync Error",
+	"Peripheral Timeout Error",
+	"False Control Error",
+	"Contention Detected",
+	"ECC Error, single-bit",
+	"ECC Error, multi-bit",
+	"Checksum Error",
+	"DSI Data Type Not Recognized",
+	"DSI VC ID Invalid",
+	"Invalid Transmission Length",
+	"Reserved",
+	"DSI Protocol Violation",
+};
+
+static ssize_t tegra_dsi_read_response(struct tegra_dsi *dsi,
+				       const struct mipi_dsi_msg *msg,
+				       size_t count)
+{
+	u8 *rx = msg->rx_buf;
+	unsigned int i, j, k;
+	size_t size = 0;
+	u16 errors;
+	u32 value;
+
+	/* read and parse packet header */
+	value = tegra_dsi_readl(dsi, DSI_RD_DATA);
+
+	switch (value & 0x3f) {
+	case MIPI_DSI_RX_ACKNOWLEDGE_AND_ERROR_REPORT:
+		errors = (value >> 8) & 0xffff;
+		dev_dbg(dsi->dev, "Acknowledge and error report: %04x\n",
+			errors);
+		for (i = 0; i < ARRAY_SIZE(error_report); i++)
+			if (errors & BIT(i))
+				dev_dbg(dsi->dev, "  %2u: %s\n", i,
+					error_report[i]);
+		break;
+
+	case MIPI_DSI_RX_DCS_SHORT_READ_RESPONSE_1BYTE:
+		rx[0] = (value >> 8) & 0xff;
+		size = 1;
+		break;
+
+	case MIPI_DSI_RX_DCS_SHORT_READ_RESPONSE_2BYTE:
+		rx[0] = (value >>  8) & 0xff;
+		rx[1] = (value >> 16) & 0xff;
+		size = 2;
+		break;
+
+	case MIPI_DSI_RX_DCS_LONG_READ_RESPONSE:
+		size = ((value >> 8) & 0xff00) | ((value >> 8) & 0xff);
+		break;
+
+	case MIPI_DSI_RX_GENERIC_LONG_READ_RESPONSE:
+		size = ((value >> 8) & 0xff00) | ((value >> 8) & 0xff);
+		break;
+
+	default:
+		dev_err(dsi->dev, "unhandled response type: %02x\n",
+			value & 0x3f);
+		return -EPROTO;
+	}
+
+	size = min(size, msg->rx_len);
+
+	if (msg->rx_buf && size > 0) {
+		for (i = 0, j = 0; i < count - 1; i++, j += 4) {
+			u8 *rx = msg->rx_buf + j;
+
+			value = tegra_dsi_readl(dsi, DSI_RD_DATA);
+
+			for (k = 0; k < 4 && (j + k) < msg->rx_len; k++)
+				rx[j + k] = (value >> (k << 3)) & 0xff;
+		}
+	}
+
+	return size;
+}
+
+static int tegra_dsi_transmit(struct tegra_dsi *dsi, unsigned long timeout)
+{
+	tegra_dsi_writel(dsi, DSI_TRIGGER_HOST, DSI_TRIGGER);
+
+	timeout = jiffies + msecs_to_jiffies(timeout);
+
+	while (time_before(jiffies, timeout)) {
+		u32 value = tegra_dsi_readl(dsi, DSI_TRIGGER);
+		if ((value & DSI_TRIGGER_HOST) == 0)
+			return 0;
+
+		usleep_range(1000, 2000);
+	}
+
+	DRM_DEBUG_KMS("timeout waiting for transmission to complete\n");
+	return -ETIMEDOUT;
+}
+
+static int tegra_dsi_wait_for_response(struct tegra_dsi *dsi,
+				       unsigned long timeout)
+{
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	while (time_before(jiffies, timeout)) {
+		u32 value = tegra_dsi_readl(dsi, DSI_STATUS);
+		u8 count = value & 0x1f;
+
+		if (count > 0)
+			return count;
+
+		usleep_range(1000, 2000);
+	}
+
+	DRM_DEBUG_KMS("peripheral returned no data\n");
+	return -ETIMEDOUT;
+}
+
+static void tegra_dsi_writesl(struct tegra_dsi *dsi, unsigned long offset,
+			      const void *buffer, size_t size)
+{
+	const u8 *buf = buffer;
+	size_t i, j;
+	u32 value;
+
+	for (j = 0; j < size; j += 4) {
+		value = 0;
+
+		for (i = 0; i < 4 && j + i < size; i++)
+			value |= buf[j + i] << (i << 3);
+
+		tegra_dsi_writel(dsi, value, DSI_WR_DATA);
+	}
+}
+
+static ssize_t tegra_dsi_host_transfer(struct mipi_dsi_host *host,
+				       const struct mipi_dsi_msg *msg)
+{
+	struct tegra_dsi *dsi = host_to_tegra(host);
+	struct mipi_dsi_packet packet;
+	const u8 *header;
+	size_t count;
+	ssize_t err;
+	u32 value;
+
+	err = mipi_dsi_create_packet(&packet, msg);
+	if (err < 0)
+		return err;
+
+	header = packet.header;
+
+	/* maximum FIFO depth is 1920 words */
+	if (packet.size > dsi->video_fifo_depth * 4)
+		return -ENOSPC;
+
+	/* reset underflow/overflow flags */
+	value = tegra_dsi_readl(dsi, DSI_STATUS);
+	if (value & (DSI_STATUS_UNDERFLOW | DSI_STATUS_OVERFLOW)) {
+		value = DSI_HOST_CONTROL_FIFO_RESET;
+		tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
+		usleep_range(10, 20);
+	}
+
+	value = tegra_dsi_readl(dsi, DSI_POWER_CONTROL);
+	value |= DSI_POWER_CONTROL_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_POWER_CONTROL);
+
+	usleep_range(5000, 10000);
+
+	value = DSI_HOST_CONTROL_CRC_RESET | DSI_HOST_CONTROL_TX_TRIG_HOST |
+		DSI_HOST_CONTROL_CS | DSI_HOST_CONTROL_ECC;
+
+	if ((msg->flags & MIPI_DSI_MSG_USE_LPM) == 0)
+		value |= DSI_HOST_CONTROL_HS;
+
+	/*
+	 * The host FIFO has a maximum of 64 words, so larger transmissions
+	 * need to use the video FIFO.
+	 */
+	if (packet.size > dsi->host_fifo_depth * 4)
+		value |= DSI_HOST_CONTROL_FIFO_SEL;
+
+	tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
+
+	/*
+	 * For reads and messages with explicitly requested ACK, generate a
+	 * BTA sequence after the transmission of the packet.
+	 */
+	if ((msg->flags & MIPI_DSI_MSG_REQ_ACK) ||
+	    (msg->rx_buf && msg->rx_len > 0)) {
+		value = tegra_dsi_readl(dsi, DSI_HOST_CONTROL);
+		value |= DSI_HOST_CONTROL_PKT_BTA;
+		tegra_dsi_writel(dsi, value, DSI_HOST_CONTROL);
+	}
+
+	value = DSI_CONTROL_LANES(0) | DSI_CONTROL_HOST_ENABLE;
+	tegra_dsi_writel(dsi, value, DSI_CONTROL);
+
+	/* write packet header, ECC is generated by hardware */
+	value = header[2] << 16 | header[1] << 8 | header[0];
+	tegra_dsi_writel(dsi, value, DSI_WR_DATA);
+
+	/* write payload (if any) */
+	if (packet.payload_length > 0)
+		tegra_dsi_writesl(dsi, DSI_WR_DATA, packet.payload,
+				  packet.payload_length);
+
+	err = tegra_dsi_transmit(dsi, 250);
+	if (err < 0)
+		return err;
+
+	if ((msg->flags & MIPI_DSI_MSG_REQ_ACK) ||
+	    (msg->rx_buf && msg->rx_len > 0)) {
+		err = tegra_dsi_wait_for_response(dsi, 250);
+		if (err < 0)
+			return err;
+
+		count = err;
+
+		value = tegra_dsi_readl(dsi, DSI_RD_DATA);
+		switch (value) {
+		case 0x84:
+			/*
+			dev_dbg(dsi->dev, "ACK\n");
+			*/
+			break;
+
+		case 0x87:
+			/*
+			dev_dbg(dsi->dev, "ESCAPE\n");
+			*/
+			break;
+
+		default:
+			dev_err(dsi->dev, "unknown status: %08x\n", value);
+			break;
+		}
+
+		if (count > 1) {
+			err = tegra_dsi_read_response(dsi, msg, count);
+			if (err < 0)
+				dev_err(dsi->dev,
+					"failed to parse response: %zd\n",
+					err);
+			else {
+				/*
+				 * For read commands, return the number of
+				 * bytes returned by the peripheral.
+				 */
+				count = err;
+			}
+		}
+	} else {
+		/*
+		 * For write commands, we have transmitted the 4-byte header
+		 * plus the variable-length payload.
+		 */
+		count = 4 + packet.payload_length;
+	}
+
+	return count;
+}
+
+static int tegra_dsi_ganged_setup(struct tegra_dsi *dsi)
+{
+	struct clk *parent;
+	int err;
+
+	/* make sure both DSI controllers share the same PLL */
+	parent = clk_get_parent(dsi->slave->clk);
+	if (!parent)
+		return -EINVAL;
+
+	err = clk_set_parent(parent, dsi->clk_parent);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static int tegra_dsi_host_attach(struct mipi_dsi_host *host,
+				 struct mipi_dsi_device *device)
+{
+	struct tegra_dsi *dsi = host_to_tegra(host);
+
+	dsi->flags = device->mode_flags;
+	dsi->format = device->format;
+	dsi->lanes = device->lanes;
+
+	if (dsi->slave) {
+		int err;
+
+		dev_dbg(dsi->dev, "attaching dual-channel device %s\n",
+			dev_name(&device->dev));
+
+		err = tegra_dsi_ganged_setup(dsi);
+		if (err < 0) {
+			dev_err(dsi->dev, "failed to set up ganged mode: %d\n",
+				err);
+			return err;
+		}
+	}
+
+	/*
+	 * Slaves don't have a panel associated with them, so they provide
+	 * merely the second channel.
+	 */
+	if (!dsi->master) {
+		struct tegra_output *output = &dsi->output;
+
+		/*
+		 * tegra_output_probe() never populates a panel for DSI
+		 * outputs, so output->panel is always NULL here.
+		 */
+		output->panel = of_drm_find_panel(device->dev.of_node);
+		if (IS_ERR(output->panel))
+			output->panel = NULL;
+
+		if (output->panel && output->connector.dev)
+			drm_helper_hpd_irq_event(output->connector.dev);
+	}
+
+	return 0;
+}
+
+static int tegra_dsi_host_detach(struct mipi_dsi_host *host,
+				 struct mipi_dsi_device *device)
+{
+	struct tegra_dsi *dsi = host_to_tegra(host);
+	struct tegra_output *output = &dsi->output;
+
+	if (output->panel && &device->dev == output->panel->dev) {
+		drm_panel_put(output->panel);
+		output->panel = NULL;
+
+		if (output->connector.dev)
+			drm_helper_hpd_irq_event(output->connector.dev);
+	}
+
+	return 0;
+}
+
+static const struct mipi_dsi_host_ops tegra_dsi_host_ops = {
+	.attach = tegra_dsi_host_attach,
+	.detach = tegra_dsi_host_detach,
+	.transfer = tegra_dsi_host_transfer,
+};
+
+static int tegra_dsi_ganged_probe(struct tegra_dsi *dsi)
+{
+	struct device_node *np;
+
+	np = of_parse_phandle(dsi->dev->of_node, "nvidia,ganged-mode", 0);
+	if (np) {
+		struct platform_device *gangster = of_find_device_by_node(np);
+		of_node_put(np);
+		if (!gangster)
+			return -EPROBE_DEFER;
+
+		dsi->slave = platform_get_drvdata(gangster);
+		put_device(&gangster->dev);
+		if (!dsi->slave)
+			return -EPROBE_DEFER;
+
+		dsi->slave->master = dsi;
+	}
+
+	return 0;
+}
+
+static int tegra_dsi_probe(struct platform_device *pdev)
+{
+	struct tegra_dsi *dsi;
+	int err;
+
+	dsi = devm_kzalloc(&pdev->dev, sizeof(*dsi), GFP_KERNEL);
+	if (!dsi)
+		return -ENOMEM;
+
+	dsi->config = device_get_match_data(&pdev->dev);
+	if (!dsi->config)
+		return -ENODEV;
+
+	dsi->output.dev = dsi->dev = &pdev->dev;
+	dsi->video_fifo_depth = 1920;
+	dsi->host_fifo_depth = 64;
+
+	err = tegra_dsi_ganged_probe(dsi);
+	if (err < 0)
+		return err;
+
+	err = tegra_output_probe(&dsi->output);
+	if (err < 0)
+		return err;
+
+	dsi->output.connector.polled = DRM_CONNECTOR_POLL_HPD;
+
+	/*
+	 * Assume these values by default. When a DSI peripheral driver
+	 * attaches to the DSI host, the parameters will be taken from
+	 * the attached device.
+	 */
+	dsi->flags = MIPI_DSI_MODE_VIDEO;
+	dsi->format = MIPI_DSI_FMT_RGB888;
+	dsi->lanes = 4;
+
+	if (!pdev->dev.pm_domain) {
+		dsi->rst = devm_reset_control_get(&pdev->dev, "dsi");
+		if (IS_ERR(dsi->rst)) {
+			err = PTR_ERR(dsi->rst);
+			goto remove;
+		}
+	}
+
+	dsi->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(dsi->clk)) {
+		err = dev_err_probe(&pdev->dev, PTR_ERR(dsi->clk),
+				    "cannot get DSI clock\n");
+		goto remove;
+	}
+
+	dsi->clk_lp = devm_clk_get_optional(&pdev->dev, "lp");
+	if (IS_ERR(dsi->clk_lp)) {
+		err = dev_err_probe(&pdev->dev, PTR_ERR(dsi->clk_lp),
+				    "cannot get low-power clock\n");
+		goto remove;
+	}
+
+	dsi->clk_parent = devm_clk_get(&pdev->dev, "parent");
+	if (IS_ERR(dsi->clk_parent)) {
+		err = dev_err_probe(&pdev->dev, PTR_ERR(dsi->clk_parent),
+				    "cannot get parent clock\n");
+		goto remove;
+	}
+
+	dsi->vdd = devm_regulator_get(&pdev->dev, "avdd-dsi-csi");
+	if (IS_ERR(dsi->vdd)) {
+		err = dev_err_probe(&pdev->dev, PTR_ERR(dsi->vdd),
+				    "cannot get VDD supply\n");
+		goto remove;
+	}
+
+	if (dsi->config->has_mux_parent_clk) {
+		err = tegra_dsi_setup_clocks(dsi);
+		if (err < 0) {
+			dev_err(&pdev->dev, "cannot setup clocks\n");
+			goto remove;
+		}
+	}
+
+	dsi->regs = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(dsi->regs)) {
+		err = PTR_ERR(dsi->regs);
+		goto remove;
+	}
+
+	dsi->mipi = tegra_mipi_request(&pdev->dev, pdev->dev.of_node);
+	if (IS_ERR(dsi->mipi)) {
+		err = PTR_ERR(dsi->mipi);
+		goto remove;
+	}
+
+	dsi->host.ops = &tegra_dsi_host_ops;
+	dsi->host.dev = &pdev->dev;
+
+	err = mipi_dsi_host_register(&dsi->host);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to register DSI host: %d\n", err);
+		goto mipi_free;
+	}
+
+	platform_set_drvdata(pdev, dsi);
+	pm_runtime_enable(&pdev->dev);
+
+	INIT_LIST_HEAD(&dsi->client.list);
+	dsi->client.ops = &dsi_client_ops;
+	dsi->client.dev = &pdev->dev;
+
+	err = host1x_client_register(&dsi->client);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to register host1x client: %d\n",
+			err);
+		goto unregister;
+	}
+
+	return 0;
+
+unregister:
+	pm_runtime_disable(&pdev->dev);
+	mipi_dsi_host_unregister(&dsi->host);
+mipi_free:
+	tegra_mipi_free(dsi->mipi);
+remove:
+	tegra_output_remove(&dsi->output);
+	return err;
+}
+
+static void tegra_dsi_remove(struct platform_device *pdev)
+{
+	struct tegra_dsi *dsi = platform_get_drvdata(pdev);
+
+	pm_runtime_disable(&pdev->dev);
+
+	host1x_client_unregister(&dsi->client);
+
+	tegra_output_remove(&dsi->output);
+
+	mipi_dsi_host_unregister(&dsi->host);
+	tegra_mipi_free(dsi->mipi);
+}
+
+static const struct tegra_dsi_config tegra20_dsi_config = {
+	.has_multiple_pad_controls = false,
+	.has_mux_parent_clk = false,
+};
+
+/*
+ * Tegra30 allows DSIA/DSIB to be muxed to either PLL_D or PLL_D2; this is
+ * simply not modeled in the clock driver yet. If this functionality is
+ * required, the has_mux_parent_clk flag can be set to true once the clock
+ * driver is patched.
+ */
+static const struct tegra_dsi_config tegra30_dsi_config = {
+	.has_multiple_pad_controls = false,
+	.has_mux_parent_clk = false,
+};
+
+static const struct tegra_dsi_config tegra114_dsi_config = {
+	.has_multiple_pad_controls = true,
+	.has_mux_parent_clk = true,
+};
+
+/* TODO: figure out why has_mux_parent_clk = true is necessary on Tegra124+ */
+static const struct tegra_dsi_config tegra124_dsi_config = {
+	.has_multiple_pad_controls = true,
+	.has_mux_parent_clk = true,
+};
+
+static const struct of_device_id tegra_dsi_of_match[] = {
+	{ .compatible = "nvidia,tegra210-dsi", .data = &tegra124_dsi_config },
+	{ .compatible = "nvidia,tegra132-dsi", .data = &tegra124_dsi_config },
+	{ .compatible = "nvidia,tegra124-dsi", .data = &tegra124_dsi_config },
+	{ .compatible = "nvidia,tegra114-dsi", .data = &tegra114_dsi_config },
+	{ .compatible = "nvidia,tegra30-dsi", .data = &tegra30_dsi_config },
+	{ .compatible = "nvidia,tegra20-dsi", .data = &tegra20_dsi_config },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, tegra_dsi_of_match);
+
+struct platform_driver tegra_dsi_driver = {
+	.driver = {
+		.name = "tegra-dsi",
+		.of_match_table = tegra_dsi_of_match,
+	},
+	.probe = tegra_dsi_probe,
+	.remove = tegra_dsi_remove,
+};
diff --git a/drivers/gpu/drm/tegra/dsi.h b/drivers/gpu/drm/tegra/dsi.h
new file mode 100644
index 000000000..d834ac0c4
--- /dev/null
+++ b/drivers/gpu/drm/tegra/dsi.h
@@ -0,0 +1,153 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#ifndef DRM_TEGRA_DSI_H
+#define DRM_TEGRA_DSI_H
+
+#define DSI_INCR_SYNCPT			0x00
+#define DSI_INCR_SYNCPT_CONTROL		0x01
+#define DSI_INCR_SYNCPT_ERROR		0x02
+#define DSI_CTXSW			0x08
+#define DSI_RD_DATA			0x09
+#define DSI_WR_DATA			0x0a
+#define DSI_POWER_CONTROL		0x0b
+#define DSI_POWER_CONTROL_ENABLE	(1 << 0)
+#define DSI_INT_ENABLE			0x0c
+#define DSI_INT_STATUS			0x0d
+#define DSI_INT_MASK			0x0e
+#define DSI_HOST_CONTROL		0x0f
+#define DSI_HOST_CONTROL_FIFO_RESET	(1 << 21)
+#define DSI_HOST_CONTROL_CRC_RESET	(1 << 20)
+#define DSI_HOST_CONTROL_TX_TRIG_SOL	(0 << 12)
+#define DSI_HOST_CONTROL_TX_TRIG_FIFO	(1 << 12)
+#define DSI_HOST_CONTROL_TX_TRIG_HOST	(2 << 12)
+#define DSI_HOST_CONTROL_RAW		(1 << 6)
+#define DSI_HOST_CONTROL_HS		(1 << 5)
+#define DSI_HOST_CONTROL_FIFO_SEL	(1 << 4)
+#define DSI_HOST_CONTROL_IMM_BTA	(1 << 3)
+#define DSI_HOST_CONTROL_PKT_BTA	(1 << 2)
+#define DSI_HOST_CONTROL_CS		(1 << 1)
+#define DSI_HOST_CONTROL_ECC		(1 << 0)
+#define DSI_CONTROL			0x10
+#define DSI_CONTROL_HS_CLK_CTRL		(1 << 20)
+#define DSI_CONTROL_CHANNEL(c)		(((c) & 0x3) << 16)
+#define DSI_CONTROL_FORMAT(f)		(((f) & 0x3) << 12)
+#define DSI_CONTROL_TX_TRIG(x)		(((x) & 0x3) <<  8)
+#define DSI_CONTROL_LANES(n)		(((n) & 0x3) <<  4)
+#define DSI_CONTROL_DCS_ENABLE		(1 << 3)
+#define DSI_CONTROL_SOURCE(s)		(((s) & 0x1) <<  2)
+#define DSI_CONTROL_VIDEO_ENABLE	(1 << 1)
+#define DSI_CONTROL_HOST_ENABLE		(1 << 0)
+#define DSI_SOL_DELAY			0x11
+#define DSI_MAX_THRESHOLD		0x12
+#define DSI_TRIGGER			0x13
+#define DSI_TRIGGER_HOST		(1 << 1)
+#define DSI_TRIGGER_VIDEO		(1 << 0)
+#define DSI_TX_CRC			0x14
+#define DSI_STATUS			0x15
+#define DSI_STATUS_IDLE			(1 << 10)
+#define DSI_STATUS_UNDERFLOW		(1 <<  9)
+#define DSI_STATUS_OVERFLOW		(1 <<  8)
+#define DSI_INIT_SEQ_CONTROL		0x1a
+#define DSI_INIT_SEQ_DATA_0		0x1b
+#define DSI_INIT_SEQ_DATA_1		0x1c
+#define DSI_INIT_SEQ_DATA_2		0x1d
+#define DSI_INIT_SEQ_DATA_3		0x1e
+#define DSI_INIT_SEQ_DATA_4		0x1f
+#define DSI_INIT_SEQ_DATA_5		0x20
+#define DSI_INIT_SEQ_DATA_6		0x21
+#define DSI_INIT_SEQ_DATA_7		0x22
+#define DSI_PKT_SEQ_0_LO		0x23
+#define DSI_PKT_SEQ_0_HI		0x24
+#define DSI_PKT_SEQ_1_LO		0x25
+#define DSI_PKT_SEQ_1_HI		0x26
+#define DSI_PKT_SEQ_2_LO		0x27
+#define DSI_PKT_SEQ_2_HI		0x28
+#define DSI_PKT_SEQ_3_LO		0x29
+#define DSI_PKT_SEQ_3_HI		0x2a
+#define DSI_PKT_SEQ_4_LO		0x2b
+#define DSI_PKT_SEQ_4_HI		0x2c
+#define DSI_PKT_SEQ_5_LO		0x2d
+#define DSI_PKT_SEQ_5_HI		0x2e
+#define DSI_DCS_CMDS			0x33
+#define DSI_PKT_LEN_0_1			0x34
+#define DSI_PKT_LEN_2_3			0x35
+#define DSI_PKT_LEN_4_5			0x36
+#define DSI_PKT_LEN_6_7			0x37
+#define DSI_PHY_TIMING_0		0x3c
+#define DSI_PHY_TIMING_1		0x3d
+#define DSI_PHY_TIMING_2		0x3e
+#define DSI_BTA_TIMING			0x3f
+
+#define DSI_TIMING_FIELD(value, period, hwinc) \
+	((DIV_ROUND_CLOSEST(value, period) - (hwinc)) & 0xff)
+
+#define DSI_TIMEOUT_0			0x44
+#define DSI_TIMEOUT_LRX(x)		(((x) & 0xffff) << 16)
+#define DSI_TIMEOUT_HTX(x)		(((x) & 0xffff) <<  0)
+#define DSI_TIMEOUT_1			0x45
+#define DSI_TIMEOUT_PR(x)		(((x) & 0xffff) << 16)
+#define DSI_TIMEOUT_TA(x)		(((x) & 0xffff) <<  0)
+#define DSI_TO_TALLY			0x46
+#define DSI_TALLY_TA(x)			(((x) & 0xff) << 16)
+#define DSI_TALLY_LRX(x)		(((x) & 0xff) <<  8)
+#define DSI_TALLY_HTX(x)		(((x) & 0xff) <<  0)
+#define DSI_PAD_CONTROL_0		0x4b
+/* Tegra20/Tegra30 */
+#define DSI_PAD_CONTROL_PULLDN_ENAB(x)	(((x) & 0x1) << 28)
+#define DSI_PAD_CONTROL_SLEWUPADJ(x)	(((x) & 0x7) << 24)
+#define DSI_PAD_CONTROL_SLEWDNADJ(x)	(((x) & 0x7) << 20)
+#define DSI_PAD_CONTROL_PREEMP_EN(x)	(((x) & 0x1) << 19)
+#define DSI_PAD_CONTROL_PDIO_CLK(x)	(((x) & 0x1) << 18)
+#define DSI_PAD_CONTROL_PDIO(x)		(((x) & 0x3) << 16)
+#define DSI_PAD_CONTROL_LPUPADJ(x)	(((x) & 0x3) << 14)
+#define DSI_PAD_CONTROL_LPDNADJ(x)	(((x) & 0x3) << 12)
+/* Tegra114+ */
+#define DSI_PAD_CONTROL_VS1_PDIO(x)	(((x) & 0xf) <<  0)
+#define DSI_PAD_CONTROL_VS1_PDIO_CLK	(1 <<  8)
+#define DSI_PAD_CONTROL_VS1_PULLDN(x)	(((x) & 0xf) << 16)
+#define DSI_PAD_CONTROL_VS1_PULLDN_CLK	(1 << 24)
+#define DSI_PAD_CONTROL_CD		0x4c
+#define DSI_PAD_CD_STATUS		0x4d
+#define DSI_VIDEO_MODE_CONTROL		0x4e
+#define DSI_PAD_CONTROL_1		0x4f
+#define DSI_PAD_CONTROL_2		0x50
+#define DSI_PAD_OUT_CLK(x)		(((x) & 0x7) <<  0)
+#define DSI_PAD_LP_DN(x)		(((x) & 0x7) <<  4)
+#define DSI_PAD_LP_UP(x)		(((x) & 0x7) <<  8)
+#define DSI_PAD_SLEW_DN(x)		(((x) & 0x7) << 12)
+#define DSI_PAD_SLEW_UP(x)		(((x) & 0x7) << 16)
+#define DSI_PAD_CONTROL_3		0x51
+#define  DSI_PAD_PREEMP_PD_CLK(x)	(((x) & 0x3) << 12)
+#define  DSI_PAD_PREEMP_PU_CLK(x)	(((x) & 0x3) << 8)
+#define  DSI_PAD_PREEMP_PD(x)		(((x) & 0x3) << 4)
+#define  DSI_PAD_PREEMP_PU(x)		(((x) & 0x3) << 0)
+#define DSI_PAD_CONTROL_4		0x52
+#define DSI_GANGED_MODE_CONTROL		0x53
+#define DSI_GANGED_MODE_CONTROL_ENABLE	(1 << 0)
+#define DSI_GANGED_MODE_START		0x54
+#define DSI_GANGED_MODE_SIZE		0x55
+#define DSI_RAW_DATA_BYTE_COUNT		0x56
+#define DSI_ULTRA_LOW_POWER_CONTROL	0x57
+#define DSI_INIT_SEQ_DATA_8		0x58
+#define DSI_INIT_SEQ_DATA_9		0x59
+#define DSI_INIT_SEQ_DATA_10		0x5a
+#define DSI_INIT_SEQ_DATA_11		0x5b
+#define DSI_INIT_SEQ_DATA_12		0x5c
+#define DSI_INIT_SEQ_DATA_13		0x5d
+#define DSI_INIT_SEQ_DATA_14		0x5e
+#define DSI_INIT_SEQ_DATA_15		0x5f
+
+/*
+ * pixel format as used in the DSI_CONTROL_FORMAT field
+ */
+enum tegra_dsi_format {
+	TEGRA_DSI_FORMAT_16P,
+	TEGRA_DSI_FORMAT_18NP,
+	TEGRA_DSI_FORMAT_18P,
+	TEGRA_DSI_FORMAT_24P,
+};
+
+#endif
diff --git a/drivers/gpu/drm/tegra/falcon.c b/drivers/gpu/drm/tegra/falcon.c
new file mode 100644
index 000000000..1172356b6
--- /dev/null
+++ b/drivers/gpu/drm/tegra/falcon.c
@@ -0,0 +1,295 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2015, NVIDIA Corporation.
+ */
+
+#include <linux/platform_device.h>
+#include <linux/dma-mapping.h>
+#include <linux/firmware.h>
+#include <linux/pci_ids.h>
+#include <linux/iopoll.h>
+
+#include "falcon.h"
+#include "drm.h"
+
+enum falcon_memory {
+	FALCON_MEMORY_IMEM,
+	FALCON_MEMORY_DATA,
+};
+
+static void falcon_writel(struct falcon *falcon, u32 value, u32 offset)
+{
+	writel(value, falcon->regs + offset);
+}
+
+int falcon_wait_idle(struct falcon *falcon)
+{
+	u32 value;
+
+	if (falcon->riscv)
+		return readl_poll_timeout(falcon->regs + RISCV_CPUCTL, value,
+					  (value & RISCV_CPUCTL_ACTIVE_STAT_ACTIVE), 10, 100000);
+	else
+		return readl_poll_timeout(falcon->regs + FALCON_IDLESTATE, value,
+					  (value == 0), 10, 100000);
+}
+
+static int falcon_dma_wait_not_full(struct falcon *falcon)
+{
+	u32 value;
+
+	return readl_poll_timeout(falcon->regs + FALCON_DMATRFCMD, value,
+				  !(value & FALCON_DMATRFCMD_FULL), 10, 100000);
+}
+
+static int falcon_dma_wait_idle(struct falcon *falcon)
+{
+	u32 value;
+
+	return readl_poll_timeout(falcon->regs + FALCON_DMATRFCMD, value,
+				  (value & FALCON_DMATRFCMD_IDLE), 10, 100000);
+}
+
+static int falcon_copy_chunk(struct falcon *falcon,
+			     phys_addr_t base,
+			     unsigned long offset,
+			     enum falcon_memory target)
+{
+	u32 cmd = FALCON_DMATRFCMD_SIZE_256B;
+	int err;
+
+	if (target == FALCON_MEMORY_IMEM)
+		cmd |= FALCON_DMATRFCMD_IMEM;
+
+	/*
+	 * Use second DMA context (i.e. the one for firmware). Strictly
+	 * speaking, at this point both DMA contexts point to the firmware
+	 * stream ID, but this register's value will be reused by the firmware
+	 * for later DMA transactions, so we need to use the correct value.
+	 */
+	cmd |= FALCON_DMATRFCMD_DMACTX(1);
+
+	err = falcon_dma_wait_not_full(falcon);
+	if (err < 0)
+		return err;
+
+	falcon_writel(falcon, offset, FALCON_DMATRFMOFFS);
+	falcon_writel(falcon, base, FALCON_DMATRFFBOFFS);
+	falcon_writel(falcon, cmd, FALCON_DMATRFCMD);
+
+	return 0;
+}
+
+static void falcon_copy_firmware_image(struct falcon *falcon,
+				       const struct firmware *firmware)
+{
+	u32 *virt = falcon->firmware.virt;
+	size_t i;
+
+	/* copy the whole thing taking into account endianness */
+	for (i = 0; i < firmware->size / sizeof(u32); i++)
+		virt[i] = le32_to_cpu(((__le32 *)firmware->data)[i]);
+}
+
+static int falcon_parse_firmware_image(struct falcon *falcon)
+{
+	struct falcon_fw_bin_header_v1 *bin = (void *)falcon->firmware.virt;
+	struct falcon_fw_os_header_v1 *os;
+
+	/* endian problems would show up right here */
+	if (bin->magic != PCI_VENDOR_ID_NVIDIA && bin->magic != 0x10fe) {
+		dev_err(falcon->dev, "incorrect firmware magic\n");
+		return -EINVAL;
+	}
+
+	/* currently only version 1 is supported */
+	if (bin->version != 1) {
+		dev_err(falcon->dev, "unsupported firmware version\n");
+		return -EINVAL;
+	}
+
+	/* check that the firmware size is consistent */
+	if (bin->size > falcon->firmware.size) {
+		dev_err(falcon->dev, "firmware image size inconsistency\n");
+		return -EINVAL;
+	}
+
+	os = falcon->firmware.virt + bin->os_header_offset;
+
+	falcon->firmware.bin_data.size = bin->os_size;
+	falcon->firmware.bin_data.offset = bin->os_data_offset;
+	falcon->firmware.code.offset = os->code_offset;
+	falcon->firmware.code.size = os->code_size;
+	falcon->firmware.data.offset = os->data_offset;
+	falcon->firmware.data.size = os->data_size;
+
+	return 0;
+}
+
+static void falcon_parse_firmware_desc(struct falcon *falcon)
+{
+	struct falcon_fw_riscv_desc *desc =
+		(struct falcon_fw_riscv_desc *)falcon->firmware.desc_firmware->data;
+
+	falcon->firmware.code.offset = desc->code_offset;
+	falcon->firmware.code.size = desc->code_size;
+	falcon->firmware.data.offset = desc->data_offset;
+	falcon->firmware.data.size = desc->data_size;
+}
+
+int falcon_read_firmware(struct falcon *falcon, const char *name)
+{
+	int err;
+
+	/* request_firmware prints error if it fails */
+	err = request_firmware(&falcon->firmware.firmware, name, falcon->dev);
+	if (err < 0)
+		return err;
+
+	falcon->firmware.size = falcon->firmware.firmware->size;
+
+	if (falcon->riscv) {
+		/* Load separate descriptor */
+		char desc_name[128];
+
+		scnprintf(desc_name, sizeof(desc_name), "%s.desc", name);
+		err = request_firmware(&falcon->firmware.desc_firmware, desc_name, falcon->dev);
+		if (err < 0)
+			goto release_firmware;
+	}
+
+	return 0;
+
+release_firmware:
+	release_firmware(falcon->firmware.firmware);
+	falcon->firmware.firmware = NULL;
+
+	return err;
+}
+
+int falcon_load_firmware(struct falcon *falcon)
+{
+	const struct firmware *firmware = falcon->firmware.firmware;
+	int err;
+
+	/* copy firmware image into local area. this also ensures endianness */
+	falcon_copy_firmware_image(falcon, firmware);
+
+	if (falcon->riscv) {
+		falcon_parse_firmware_desc(falcon);
+	} else {
+		err = falcon_parse_firmware_image(falcon);
+		if (err < 0) {
+			dev_err(falcon->dev, "failed to parse firmware image\n");
+			return err;
+		}
+	}
+
+	release_firmware(firmware);
+	falcon->firmware.firmware = NULL;
+
+	release_firmware(falcon->firmware.desc_firmware);
+	falcon->firmware.desc_firmware = NULL;
+
+	return 0;
+}
+
+int falcon_init(struct falcon *falcon)
+{
+	falcon->firmware.virt = NULL;
+
+	return 0;
+}
+
+void falcon_exit(struct falcon *falcon)
+{
+	if (falcon->firmware.firmware)
+		release_firmware(falcon->firmware.firmware);
+
+	if (falcon->firmware.desc_firmware)
+		release_firmware(falcon->firmware.desc_firmware);
+}
+
+int falcon_boot(struct falcon *falcon)
+{
+	unsigned long offset;
+	u32 value;
+	int err;
+
+	if (!falcon->firmware.virt)
+		return -EINVAL;
+
+	err = readl_poll_timeout(falcon->regs + FALCON_DMACTL, value,
+				 (value & (FALCON_DMACTL_IMEM_SCRUBBING |
+					   FALCON_DMACTL_DMEM_SCRUBBING)) == 0,
+				 10, 10000);
+	if (err < 0)
+		return err;
+
+	falcon_writel(falcon, 0, FALCON_DMACTL);
+
+	/* setup the address of the binary data so Falcon can access it later */
+	falcon_writel(falcon, (falcon->firmware.iova +
+			       falcon->firmware.bin_data.offset) >> 8,
+		      FALCON_DMATRFBASE);
+
+	/* copy the data segment into Falcon internal memory */
+	for (offset = 0; offset < falcon->firmware.data.size; offset += 256)
+		falcon_copy_chunk(falcon,
+				  falcon->firmware.data.offset + offset,
+				  offset, FALCON_MEMORY_DATA);
+
+	/* copy the code segment into Falcon internal memory */
+	for (offset = 0; offset < falcon->firmware.code.size; offset += 256)
+		falcon_copy_chunk(falcon, falcon->firmware.code.offset + offset,
+				  offset, FALCON_MEMORY_IMEM);
+
+	/* wait for DMA to complete */
+	err = falcon_dma_wait_idle(falcon);
+	if (err < 0)
+		return err;
+
+	/* setup falcon interrupts */
+	falcon_writel(falcon, FALCON_IRQMSET_EXT(0xff) |
+			      FALCON_IRQMSET_SWGEN1 |
+			      FALCON_IRQMSET_SWGEN0 |
+			      FALCON_IRQMSET_EXTERR |
+			      FALCON_IRQMSET_HALT |
+			      FALCON_IRQMSET_WDTMR,
+		      FALCON_IRQMSET);
+	falcon_writel(falcon, FALCON_IRQDEST_EXT(0xff) |
+			      FALCON_IRQDEST_SWGEN1 |
+			      FALCON_IRQDEST_SWGEN0 |
+			      FALCON_IRQDEST_EXTERR |
+			      FALCON_IRQDEST_HALT,
+		      FALCON_IRQDEST);
+
+	/* enable interface */
+	falcon_writel(falcon, FALCON_ITFEN_MTHDEN |
+			      FALCON_ITFEN_CTXEN,
+		      FALCON_ITFEN);
+
+	if (falcon->riscv) {
+		falcon_writel(falcon, RISCV_BCR_CTRL_CORE_SELECT_RISCV, RISCV_BCR_CTRL);
+		falcon_writel(falcon, 0x0, RISCV_BOOT_VECTOR_HI);
+		falcon_writel(falcon, 0x100000, RISCV_BOOT_VECTOR_LO);
+		falcon_writel(falcon, RISCV_CPUCTL_STARTCPU, RISCV_CPUCTL);
+	} else {
+		falcon_writel(falcon, 0x00000000, FALCON_BOOTVEC);
+		falcon_writel(falcon, FALCON_CPUCTL_STARTCPU, FALCON_CPUCTL);
+	}
+
+	err = falcon_wait_idle(falcon);
+	if (err < 0) {
+		dev_err(falcon->dev, "Falcon boot failed due to timeout\n");
+		return err;
+	}
+
+	return 0;
+}
+
+void falcon_execute_method(struct falcon *falcon, u32 method, u32 data)
+{
+	falcon_writel(falcon, method >> 2, FALCON_UCLASS_METHOD_OFFSET);
+	falcon_writel(falcon, data, FALCON_UCLASS_METHOD_DATA);
+}
diff --git a/drivers/gpu/drm/tegra/falcon.h b/drivers/gpu/drm/tegra/falcon.h
new file mode 100644
index 000000000..37a17c613
--- /dev/null
+++ b/drivers/gpu/drm/tegra/falcon.h
@@ -0,0 +1,139 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2015, NVIDIA Corporation.
+ */
+
+#ifndef _FALCON_H_
+#define _FALCON_H_
+
+#include <linux/types.h>
+
+#define FALCON_UCLASS_METHOD_OFFSET		0x00000040
+
+#define FALCON_UCLASS_METHOD_DATA		0x00000044
+
+#define FALCON_IRQMSET				0x00001010
+#define FALCON_IRQMSET_WDTMR			(1 << 1)
+#define FALCON_IRQMSET_HALT			(1 << 4)
+#define FALCON_IRQMSET_EXTERR			(1 << 5)
+#define FALCON_IRQMSET_SWGEN0			(1 << 6)
+#define FALCON_IRQMSET_SWGEN1			(1 << 7)
+#define FALCON_IRQMSET_EXT(v)			(((v) & 0xff) << 8)
+
+#define FALCON_IRQDEST				0x0000101c
+#define FALCON_IRQDEST_HALT			(1 << 4)
+#define FALCON_IRQDEST_EXTERR			(1 << 5)
+#define FALCON_IRQDEST_SWGEN0			(1 << 6)
+#define FALCON_IRQDEST_SWGEN1			(1 << 7)
+#define FALCON_IRQDEST_EXT(v)			(((v) & 0xff) << 8)
+
+#define FALCON_ITFEN				0x00001048
+#define FALCON_ITFEN_CTXEN			(1 << 0)
+#define FALCON_ITFEN_MTHDEN			(1 << 1)
+
+#define FALCON_IDLESTATE			0x0000104c
+
+#define FALCON_CPUCTL				0x00001100
+#define FALCON_CPUCTL_STARTCPU			(1 << 1)
+
+#define FALCON_BOOTVEC				0x00001104
+
+#define FALCON_DMACTL				0x0000110c
+#define FALCON_DMACTL_DMEM_SCRUBBING		(1 << 1)
+#define FALCON_DMACTL_IMEM_SCRUBBING		(1 << 2)
+
+#define FALCON_DMATRFBASE			0x00001110
+
+#define FALCON_DMATRFMOFFS			0x00001114
+
+#define FALCON_DMATRFCMD			0x00001118
+#define FALCON_DMATRFCMD_FULL			(1 << 0)
+#define FALCON_DMATRFCMD_IDLE			(1 << 1)
+#define FALCON_DMATRFCMD_IMEM			(1 << 4)
+#define FALCON_DMATRFCMD_SIZE_256B		(6 << 8)
+#define FALCON_DMATRFCMD_DMACTX(v)		(((v) & 0x7) << 12)
+
+#define FALCON_DMATRFFBOFFS			0x0000111c
+
+#define RISCV_BOOT_VECTOR_LO			0x00001780
+#define RISCV_BOOT_VECTOR_HI			0x00001784
+
+#define RISCV_CPUCTL				0x00001788
+#define RISCV_CPUCTL_STARTCPU			(1 << 0)
+#define RISCV_CPUCTL_ACTIVE_STAT_ACTIVE		(1 << 7)
+
+#define RISCV_BCR_CTRL				0x00001a68
+#define RISCV_BCR_CTRL_CORE_SELECT_RISCV	(1 << 4)
+
+struct falcon_fw_bin_header_v1 {
+	u32 magic;		/* 0x10de */
+	u32 version;		/* version of bin format (1) */
+	u32 size;		/* entire image size including this header */
+	u32 os_header_offset;
+	u32 os_data_offset;
+	u32 os_size;
+};
+
+struct falcon_fw_os_app_v1 {
+	u32 offset;
+	u32 size;
+};
+
+struct falcon_fw_os_header_v1 {
+	u32 code_offset;
+	u32 code_size;
+	u32 data_offset;
+	u32 data_size;
+};
+
+struct falcon_fw_riscv_desc {
+	u32 reserved[74];
+	u32 data_offset;
+	u32 data_size;
+	u32 code_offset;
+	u32 code_size;
+};
+
+struct falcon_firmware_section {
+	unsigned long offset;
+	size_t size;
+};
+
+struct falcon_firmware {
+	/* Firmware after it is read but not loaded */
+	const struct firmware *firmware;
+	/* RISC-V firmware descriptor */
+	const struct firmware *desc_firmware;
+
+	/* Raw firmware data */
+	dma_addr_t iova;
+	dma_addr_t phys;
+	void *virt;
+	size_t size;
+
+	/* Parsed firmware information */
+	struct falcon_firmware_section bin_data;
+	struct falcon_firmware_section data;
+	struct falcon_firmware_section code;
+};
+
+struct falcon {
+	/* Set by falcon client */
+	struct device *dev;
+	void __iomem *regs;
+
+	/* Peregrine falcon, external boot */
+	bool riscv;
+
+	struct falcon_firmware firmware;
+};
+
+int falcon_init(struct falcon *falcon);
+void falcon_exit(struct falcon *falcon);
+int falcon_read_firmware(struct falcon *falcon, const char *firmware_name);
+int falcon_load_firmware(struct falcon *falcon);
+int falcon_boot(struct falcon *falcon);
+void falcon_execute_method(struct falcon *falcon, u32 method, u32 data);
+int falcon_wait_idle(struct falcon *falcon);
+
+#endif /* _FALCON_H_ */
diff --git a/drivers/gpu/drm/tegra/fb.c b/drivers/gpu/drm/tegra/fb.c
new file mode 100644
index 000000000..0a7cb9e46
--- /dev/null
+++ b/drivers/gpu/drm/tegra/fb.c
@@ -0,0 +1,185 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012-2013 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ *
+ * Based on the KMS/FB DMA helpers
+ *   Copyright (C) 2012 Analog Devices Inc.
+ */
+
+#include <linux/console.h>
+
+#include <drm/drm_fourcc.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_gem_framebuffer_helper.h>
+#include <drm/drm_modeset_helper.h>
+#include <drm/drm_print.h>
+
+#include "drm.h"
+#include "gem.h"
+
+struct tegra_bo *tegra_fb_get_plane(struct drm_framebuffer *framebuffer,
+				    unsigned int index)
+{
+	return to_tegra_bo(drm_gem_fb_get_obj(framebuffer, index));
+}
+
+bool tegra_fb_is_bottom_up(struct drm_framebuffer *framebuffer)
+{
+	struct tegra_bo *bo = tegra_fb_get_plane(framebuffer, 0);
+
+	if (bo->flags & TEGRA_BO_BOTTOM_UP)
+		return true;
+
+	return false;
+}
+
+int tegra_fb_get_tiling(struct drm_framebuffer *framebuffer,
+			struct tegra_bo_tiling *tiling)
+{
+	uint64_t modifier = framebuffer->modifier;
+
+	if (fourcc_mod_is_vendor(modifier, NVIDIA)) {
+		if ((modifier & DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT) == 0)
+			tiling->sector_layout = TEGRA_BO_SECTOR_LAYOUT_TEGRA;
+		else
+			tiling->sector_layout = TEGRA_BO_SECTOR_LAYOUT_GPU;
+
+		modifier &= ~DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT;
+	}
+
+	switch (modifier) {
+	case DRM_FORMAT_MOD_LINEAR:
+		tiling->mode = TEGRA_BO_TILING_MODE_PITCH;
+		tiling->value = 0;
+		break;
+
+	case DRM_FORMAT_MOD_NVIDIA_TEGRA_TILED:
+		tiling->mode = TEGRA_BO_TILING_MODE_TILED;
+		tiling->value = 0;
+		break;
+
+	case DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(0):
+		tiling->mode = TEGRA_BO_TILING_MODE_BLOCK;
+		tiling->value = 0;
+		break;
+
+	case DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(1):
+		tiling->mode = TEGRA_BO_TILING_MODE_BLOCK;
+		tiling->value = 1;
+		break;
+
+	case DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(2):
+		tiling->mode = TEGRA_BO_TILING_MODE_BLOCK;
+		tiling->value = 2;
+		break;
+
+	case DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(3):
+		tiling->mode = TEGRA_BO_TILING_MODE_BLOCK;
+		tiling->value = 3;
+		break;
+
+	case DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(4):
+		tiling->mode = TEGRA_BO_TILING_MODE_BLOCK;
+		tiling->value = 4;
+		break;
+
+	case DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(5):
+		tiling->mode = TEGRA_BO_TILING_MODE_BLOCK;
+		tiling->value = 5;
+		break;
+
+	default:
+		DRM_DEBUG_KMS("unknown format modifier: %llx\n", modifier);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static const struct drm_framebuffer_funcs tegra_fb_funcs = {
+	.destroy = drm_gem_fb_destroy,
+	.create_handle = drm_gem_fb_create_handle,
+};
+
+static
+struct drm_framebuffer *tegra_fb_alloc(struct drm_device *drm,
+				       const struct drm_format_info *info,
+				       const struct drm_mode_fb_cmd2 *mode_cmd,
+				       struct tegra_bo **planes,
+				       unsigned int num_planes)
+{
+	struct drm_framebuffer *fb;
+	unsigned int i;
+	int err;
+
+	fb = kzalloc_obj(*fb);
+	if (!fb)
+		return ERR_PTR(-ENOMEM);
+
+	drm_helper_mode_fill_fb_struct(drm, fb, info, mode_cmd);
+
+	for (i = 0; i < fb->format->num_planes; i++)
+		fb->obj[i] = &planes[i]->gem;
+
+	err = drm_framebuffer_init(drm, fb, &tegra_fb_funcs);
+	if (err < 0) {
+		dev_err(drm->dev, "failed to initialize framebuffer: %d\n",
+			err);
+		kfree(fb);
+		return ERR_PTR(err);
+	}
+
+	return fb;
+}
+
+struct drm_framebuffer *tegra_fb_create(struct drm_device *drm,
+					struct drm_file *file,
+					const struct drm_format_info *info,
+					const struct drm_mode_fb_cmd2 *cmd)
+{
+	struct tegra_bo *planes[4];
+	struct drm_gem_object *gem;
+	struct drm_framebuffer *fb;
+	unsigned int i;
+	int err;
+
+	for (i = 0; i < info->num_planes; i++) {
+		unsigned int width = cmd->width / (i ? info->hsub : 1);
+		unsigned int height = cmd->height / (i ? info->vsub : 1);
+		unsigned int size, bpp;
+
+		gem = drm_gem_object_lookup(file, cmd->handles[i]);
+		if (!gem) {
+			err = -ENXIO;
+			goto unreference;
+		}
+
+		bpp = info->cpp[i];
+
+		size = (height - 1) * cmd->pitches[i] +
+		       width * bpp + cmd->offsets[i];
+
+		if (gem->size < size) {
+			err = -EINVAL;
+			drm_gem_object_put(gem);
+			goto unreference;
+		}
+
+		planes[i] = to_tegra_bo(gem);
+	}
+
+	fb = tegra_fb_alloc(drm, info, cmd, planes, i);
+	if (IS_ERR(fb)) {
+		err = PTR_ERR(fb);
+		goto unreference;
+	}
+
+	return fb;
+
+unreference:
+	while (i--)
+		drm_gem_object_put(&planes[i]->gem);
+
+	return ERR_PTR(err);
+}
diff --git a/drivers/gpu/drm/tegra/fbdev.c b/drivers/gpu/drm/tegra/fbdev.c
new file mode 100644
index 000000000..003e80cf7
--- /dev/null
+++ b/drivers/gpu/drm/tegra/fbdev.c
@@ -0,0 +1,147 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012-2013 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ *
+ * Based on the KMS/FB DMA helpers
+ *   Copyright (C) 2012 Analog Devices Inc.
+ */
+
+#include <linux/console.h>
+#include <linux/fb.h>
+#include <linux/vmalloc.h>
+
+#include <drm/drm_drv.h>
+#include <drm/drm_crtc_helper.h>
+#include <drm/drm_fb_helper.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_gem_framebuffer_helper.h>
+#include <drm/drm_modeset_helper.h>
+
+#include "drm.h"
+#include "gem.h"
+
+static int tegra_fb_mmap(struct fb_info *info, struct vm_area_struct *vma)
+{
+	struct drm_fb_helper *helper = info->par;
+	struct tegra_bo *bo;
+	int err;
+
+	bo = tegra_fb_get_plane(helper->fb, 0);
+
+	err = drm_gem_mmap_obj(&bo->gem, bo->gem.size, vma);
+	if (err < 0)
+		return err;
+
+	return __tegra_gem_mmap(&bo->gem, vma);
+}
+
+static void tegra_fbdev_fb_destroy(struct fb_info *info)
+{
+	struct drm_fb_helper *helper = info->par;
+	struct tegra_bo *bo = tegra_fb_get_plane(helper->fb, 0);
+
+	drm_fb_helper_fini(helper);
+
+	/* Undo the special mapping we made in fbdev probe. */
+	if (bo->pages) {
+		vunmap(bo->vaddr);
+		bo->vaddr = NULL;
+	}
+
+	drm_client_buffer_delete(helper->buffer);
+	drm_client_release(&helper->client);
+}
+
+static const struct fb_ops tegra_fb_ops = {
+	.owner = THIS_MODULE,
+	__FB_DEFAULT_DMAMEM_OPS_RDWR,
+	DRM_FB_HELPER_DEFAULT_OPS,
+	__FB_DEFAULT_DMAMEM_OPS_DRAW,
+	.fb_mmap = tegra_fb_mmap,
+	.fb_destroy = tegra_fbdev_fb_destroy,
+};
+
+static const struct drm_fb_helper_funcs tegra_fbdev_helper_funcs = {
+};
+
+int tegra_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
+				   struct drm_fb_helper_surface_size *sizes)
+{
+	struct drm_client_dev *client = &helper->client;
+	struct drm_device *drm = client->dev;
+	struct drm_file *file = client->file;
+	struct tegra_drm *tegra = drm->dev_private;
+	struct fb_info *info = helper->info;
+	u32 fourcc, pitch;
+	u64 size;
+	const struct drm_format_info *format;
+	struct tegra_bo *bo;
+	struct drm_gem_object *gem;
+	u32 handle;
+	struct drm_client_buffer *buffer;
+	int err;
+
+	fourcc = drm_mode_legacy_fb_format(sizes->surface_bpp, sizes->surface_depth);
+	format = drm_get_format_info(drm, fourcc, DRM_FORMAT_MOD_LINEAR);
+	pitch = round_up(drm_format_info_min_pitch(format, 0, sizes->surface_width),
+			 tegra->pitch_align);
+	size = ALIGN(pitch * sizes->surface_height, PAGE_SIZE);
+
+	bo = tegra_bo_create(drm, size, 0);
+	if (IS_ERR(bo))
+		return PTR_ERR(bo);
+	gem = &bo->gem;
+
+	err = drm_gem_handle_create(file, gem, &handle);
+	if (err)
+		goto err_drm_gem_object_put;
+
+	buffer = drm_client_buffer_create(client, sizes->surface_width, sizes->surface_height,
+					  fourcc, handle, pitch);
+	if (IS_ERR(buffer)) {
+		err = PTR_ERR(buffer);
+		goto err_drm_gem_handle_delete;
+	}
+
+	helper->funcs = &tegra_fbdev_helper_funcs;
+	helper->buffer = buffer;
+	helper->fb = buffer->fb;
+
+	info->fbops = &tegra_fb_ops;
+
+	drm_fb_helper_fill_info(info, helper, sizes);
+
+	if (bo->pages) {
+		bo->vaddr = vmap(bo->pages, bo->num_pages, VM_MAP,
+				 pgprot_writecombine(PAGE_KERNEL));
+		if (!bo->vaddr) {
+			dev_err(drm->dev, "failed to vmap() framebuffer\n");
+			err = -ENOMEM;
+			goto err_drm_client_buffer_delete;
+		}
+	}
+
+	info->flags |= FBINFO_VIRTFB;
+	info->screen_buffer = bo->vaddr;
+	info->screen_size = gem->size;
+	info->fix.smem_start = (unsigned long)(bo->iova);
+	info->fix.smem_len = gem->size;
+
+	/* The handle is only needed for creating the framebuffer. */
+	drm_gem_handle_delete(file, handle);
+
+	/* The framebuffer still holds a reference on the GEM object. */
+	drm_gem_object_put(gem);
+
+	return 0;
+
+err_drm_client_buffer_delete:
+	drm_client_buffer_delete(buffer);
+err_drm_gem_handle_delete:
+	drm_gem_handle_delete(file, handle);
+err_drm_gem_object_put:
+	drm_gem_object_put(gem);
+	return err;
+}
diff --git a/drivers/gpu/drm/tegra/firewall.c b/drivers/gpu/drm/tegra/firewall.c
new file mode 100644
index 000000000..d53f890fa
--- /dev/null
+++ b/drivers/gpu/drm/tegra/firewall.c
@@ -0,0 +1,257 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (c) 2010-2020 NVIDIA Corporation */
+
+#include "drm.h"
+#include "submit.h"
+#include "uapi.h"
+
+struct tegra_drm_firewall {
+	struct tegra_drm_submit_data *submit;
+	struct tegra_drm_client *client;
+	u32 *data;
+	u32 pos;
+	u32 end;
+	u32 class;
+};
+
+static int fw_next(struct tegra_drm_firewall *fw, u32 *word)
+{
+	if (fw->pos == fw->end)
+		return -EINVAL;
+
+	*word = fw->data[fw->pos++];
+
+	return 0;
+}
+
+static bool fw_check_addr_valid(struct tegra_drm_firewall *fw, u32 offset)
+{
+	u32 i;
+
+	for (i = 0; i < fw->submit->num_used_mappings; i++) {
+		struct tegra_drm_mapping *m = fw->submit->used_mappings[i].mapping;
+
+		if (offset >= m->iova && offset <= m->iova_end)
+			return true;
+	}
+
+	return false;
+}
+
+static int fw_check_reg(struct tegra_drm_firewall *fw, u32 offset)
+{
+	bool is_addr;
+	u32 word;
+	int err;
+
+	err = fw_next(fw, &word);
+	if (err)
+		return err;
+
+	if (!fw->client->ops->is_addr_reg)
+		return 0;
+
+	is_addr = fw->client->ops->is_addr_reg(fw->client->base.dev, fw->class,
+					       offset);
+
+	if (!is_addr)
+		return 0;
+
+	if (!fw_check_addr_valid(fw, word))
+		return -EINVAL;
+
+	return 0;
+}
+
+static int fw_check_regs_seq(struct tegra_drm_firewall *fw, u32 offset,
+			     u32 count, bool incr)
+{
+	u32 i;
+
+	for (i = 0; i < count; i++) {
+		if (fw_check_reg(fw, offset))
+			return -EINVAL;
+
+		if (incr)
+			offset++;
+	}
+
+	return 0;
+}
+
+static int fw_check_regs_mask(struct tegra_drm_firewall *fw, u32 offset,
+			      u16 mask)
+{
+	unsigned long bmask = mask;
+	unsigned int bit;
+
+	for_each_set_bit(bit, &bmask, 16) {
+		if (fw_check_reg(fw, offset+bit))
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int fw_check_regs_imm(struct tegra_drm_firewall *fw, u32 offset)
+{
+	bool is_addr;
+
+	if (!fw->client->ops->is_addr_reg)
+		return 0;
+
+	is_addr = fw->client->ops->is_addr_reg(fw->client->base.dev, fw->class,
+					       offset);
+	if (is_addr)
+		return -EINVAL;
+
+	return 0;
+}
+
+static int fw_check_class(struct tegra_drm_firewall *fw, u32 class)
+{
+	if (!fw->client->ops->is_valid_class) {
+		if (class == fw->client->base.class)
+			return 0;
+		else
+			return -EINVAL;
+	}
+
+	if (!fw->client->ops->is_valid_class(class))
+		return -EINVAL;
+
+	return 0;
+}
+
+enum {
+	HOST1X_OPCODE_SETCLASS  = 0x00,
+	HOST1X_OPCODE_INCR      = 0x01,
+	HOST1X_OPCODE_NONINCR   = 0x02,
+	HOST1X_OPCODE_MASK      = 0x03,
+	HOST1X_OPCODE_IMM       = 0x04,
+	HOST1X_OPCODE_RESTART   = 0x05,
+	HOST1X_OPCODE_GATHER    = 0x06,
+	HOST1X_OPCODE_SETSTRMID = 0x07,
+	HOST1X_OPCODE_SETAPPID  = 0x08,
+	HOST1X_OPCODE_SETPYLD   = 0x09,
+	HOST1X_OPCODE_INCR_W    = 0x0a,
+	HOST1X_OPCODE_NONINCR_W = 0x0b,
+	HOST1X_OPCODE_GATHER_W  = 0x0c,
+	HOST1X_OPCODE_RESTART_W = 0x0d,
+	HOST1X_OPCODE_EXTEND    = 0x0e,
+};
+
+int tegra_drm_fw_validate(struct tegra_drm_client *client, u32 *data, u32 start,
+			  u32 words, struct tegra_drm_submit_data *submit,
+			  u32 *job_class)
+{
+	struct tegra_drm_firewall fw = {
+		.submit = submit,
+		.client = client,
+		.data = data,
+		.pos = start,
+		.end = start+words,
+		.class = *job_class,
+	};
+	bool payload_valid = false;
+	u32 payload;
+	int err;
+
+	while (fw.pos != fw.end) {
+		u32 word, opcode, offset, count, mask, class;
+
+		err = fw_next(&fw, &word);
+		if (err)
+			return err;
+
+		opcode = (word & 0xf0000000) >> 28;
+
+		switch (opcode) {
+		case HOST1X_OPCODE_SETCLASS:
+			offset = word >> 16 & 0xfff;
+			mask = word & 0x3f;
+			class = (word >> 6) & 0x3ff;
+			err = fw_check_class(&fw, class);
+			fw.class = class;
+			*job_class = class;
+			if (!err)
+				err = fw_check_regs_mask(&fw, offset, mask);
+			if (err)
+				dev_warn(client->base.dev,
+					 "illegal SETCLASS(offset=0x%x, mask=0x%x, class=0x%x) at word %u",
+					 offset, mask, class, fw.pos-1);
+			break;
+		case HOST1X_OPCODE_INCR:
+			offset = (word >> 16) & 0xfff;
+			count = word & 0xffff;
+			err = fw_check_regs_seq(&fw, offset, count, true);
+			if (err)
+				dev_warn(client->base.dev,
+					 "illegal INCR(offset=0x%x, count=%u) in class 0x%x at word %u",
+					 offset, count, fw.class, fw.pos-1);
+			break;
+		case HOST1X_OPCODE_NONINCR:
+			offset = (word >> 16) & 0xfff;
+			count = word & 0xffff;
+			err = fw_check_regs_seq(&fw, offset, count, false);
+			if (err)
+				dev_warn(client->base.dev,
+					 "illegal NONINCR(offset=0x%x, count=%u) in class 0x%x at word %u",
+					 offset, count, fw.class, fw.pos-1);
+			break;
+		case HOST1X_OPCODE_MASK:
+			offset = (word >> 16) & 0xfff;
+			mask = word & 0xffff;
+			err = fw_check_regs_mask(&fw, offset, mask);
+			if (err)
+				dev_warn(client->base.dev,
+					 "illegal MASK(offset=0x%x, mask=0x%x) in class 0x%x at word %u",
+					 offset, mask, fw.class, fw.pos-1);
+			break;
+		case HOST1X_OPCODE_IMM:
+			/* IMM cannot reasonably be used to write a pointer */
+			offset = (word >> 16) & 0xfff;
+			err = fw_check_regs_imm(&fw, offset);
+			if (err)
+				dev_warn(client->base.dev,
+					 "illegal IMM(offset=0x%x) in class 0x%x at word %u",
+					 offset, fw.class, fw.pos-1);
+			break;
+		case HOST1X_OPCODE_SETPYLD:
+			payload = word & 0xffff;
+			payload_valid = true;
+			break;
+		case HOST1X_OPCODE_INCR_W:
+			if (!payload_valid)
+				return -EINVAL;
+
+			offset = word & 0x3fffff;
+			err = fw_check_regs_seq(&fw, offset, payload, true);
+			if (err)
+				dev_warn(client->base.dev,
+					 "illegal INCR_W(offset=0x%x) in class 0x%x at word %u",
+					 offset, fw.class, fw.pos-1);
+			break;
+		case HOST1X_OPCODE_NONINCR_W:
+			if (!payload_valid)
+				return -EINVAL;
+
+			offset = word & 0x3fffff;
+			err = fw_check_regs_seq(&fw, offset, payload, false);
+			if (err)
+				dev_warn(client->base.dev,
+					 "illegal NONINCR(offset=0x%x) in class 0x%x at word %u",
+					 offset, fw.class, fw.pos-1);
+			break;
+		default:
+			dev_warn(client->base.dev, "illegal opcode at word %u",
+				 fw.pos-1);
+			return -EINVAL;
+		}
+
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/tegra/gem.c b/drivers/gpu/drm/tegra/gem.c
new file mode 100644
index 000000000..f76af733e
--- /dev/null
+++ b/drivers/gpu/drm/tegra/gem.c
@@ -0,0 +1,800 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * NVIDIA Tegra DRM GEM helper functions
+ *
+ * Copyright (C) 2012 Sascha Hauer, Pengutronix
+ * Copyright (C) 2013-2015 NVIDIA CORPORATION, All rights reserved.
+ *
+ * Based on the GEM/CMA helpers
+ *
+ * Copyright (c) 2011 Samsung Electronics Co., Ltd.
+ */
+
+#include <linux/dma-buf.h>
+#include <linux/iommu.h>
+#include <linux/module.h>
+#include <linux/pagemap.h>
+#include <linux/vmalloc.h>
+
+#include <drm/drm_drv.h>
+#include <drm/drm_dumb_buffers.h>
+#include <drm/drm_prime.h>
+
+#include "drm.h"
+#include "gem.h"
+
+MODULE_IMPORT_NS("DMA_BUF");
+
+static unsigned int sg_dma_count_chunks(struct scatterlist *sgl, unsigned int nents)
+{
+	dma_addr_t next = ~(dma_addr_t)0;
+	unsigned int count = 0, i;
+	struct scatterlist *s;
+
+	for_each_sg(sgl, s, nents, i) {
+		/* sg_dma_address(s) is only valid for entries that have sg_dma_len(s) != 0. */
+		if (!sg_dma_len(s))
+			continue;
+
+		if (sg_dma_address(s) != next) {
+			next = sg_dma_address(s) + sg_dma_len(s);
+			count++;
+		}
+	}
+
+	return count;
+}
+
+static inline unsigned int sgt_dma_count_chunks(struct sg_table *sgt)
+{
+	return sg_dma_count_chunks(sgt->sgl, sgt->nents);
+}
+
+static void tegra_bo_put(struct host1x_bo *bo)
+{
+	struct tegra_bo *obj = host1x_to_tegra_bo(bo);
+
+	drm_gem_object_put(&obj->gem);
+}
+
+static struct host1x_bo_mapping *tegra_bo_pin(struct device *dev, struct host1x_bo *bo,
+					      enum dma_data_direction direction)
+{
+	struct tegra_bo *obj = host1x_to_tegra_bo(bo);
+	struct drm_gem_object *gem = &obj->gem;
+	struct host1x_bo_mapping *map;
+	int err;
+
+	map = kzalloc_obj(*map);
+	if (!map)
+		return ERR_PTR(-ENOMEM);
+
+	kref_init(&map->ref);
+	map->bo = bo;
+	map->direction = direction;
+	map->dev = dev;
+
+	/*
+	 * Imported buffers need special treatment to satisfy the semantics of DMA-BUF.
+	 */
+	if (obj->dma_buf) {
+		struct dma_buf *buf = obj->dma_buf;
+
+		map->attach = dma_buf_attach(buf, dev);
+		if (IS_ERR(map->attach)) {
+			err = PTR_ERR(map->attach);
+			goto free;
+		}
+
+		map->sgt = dma_buf_map_attachment_unlocked(map->attach, direction);
+		if (IS_ERR(map->sgt)) {
+			dma_buf_detach(buf, map->attach);
+			err = PTR_ERR(map->sgt);
+			map->sgt = NULL;
+			goto free;
+		}
+
+		err = sgt_dma_count_chunks(map->sgt);
+		map->size = gem->size;
+
+		goto out;
+	}
+
+	/*
+	 * If we don't have a mapping for this buffer yet, return an SG table
+	 * so that host1x can do the mapping for us via the DMA API.
+	 */
+	map->sgt = kzalloc_obj(*map->sgt);
+	if (!map->sgt) {
+		err = -ENOMEM;
+		goto free;
+	}
+
+	if (obj->pages) {
+		/*
+		 * If the buffer object was allocated from the explicit IOMMU
+		 * API code paths, construct an SG table from the pages.
+		 */
+		err = sg_alloc_table_from_pages(map->sgt, obj->pages, obj->num_pages, 0, gem->size,
+						GFP_KERNEL);
+		if (err < 0)
+			goto free;
+	} else {
+		/*
+		 * If the buffer object had no pages allocated and if it was
+		 * not imported, it had to be allocated with the DMA API, so
+		 * the DMA API helper can be used.
+		 */
+		err = dma_get_sgtable(dev, map->sgt, obj->vaddr, obj->iova, gem->size);
+		if (err < 0)
+			goto free;
+	}
+
+	err = dma_map_sgtable(dev, map->sgt, direction, 0);
+	if (err)
+		goto free_sgt;
+
+out:
+	/*
+	 * If we've manually mapped the buffer object through the IOMMU, make sure to return the
+	 * existing IOVA address of our mapping.
+	 */
+	if (!obj->mm) {
+		map->phys = sg_dma_address(map->sgt->sgl);
+		map->chunks = err;
+	} else {
+		map->phys = obj->iova;
+		map->chunks = 1;
+	}
+
+	map->size = gem->size;
+
+	return map;
+
+free_sgt:
+	sg_free_table(map->sgt);
+free:
+	kfree(map->sgt);
+	kfree(map);
+	return ERR_PTR(err);
+}
+
+static void tegra_bo_unpin(struct host1x_bo_mapping *map)
+{
+	if (map->attach) {
+		dma_buf_unmap_attachment_unlocked(map->attach, map->sgt,
+						  map->direction);
+		dma_buf_detach(map->attach->dmabuf, map->attach);
+	} else {
+		dma_unmap_sgtable(map->dev, map->sgt, map->direction, 0);
+		sg_free_table(map->sgt);
+		kfree(map->sgt);
+	}
+
+	kfree(map);
+}
+
+static void *tegra_bo_mmap(struct host1x_bo *bo)
+{
+	struct tegra_bo *obj = host1x_to_tegra_bo(bo);
+	struct iosys_map map = { 0 };
+	void *vaddr;
+	int ret;
+
+	if (obj->vaddr)
+		return obj->vaddr;
+
+	if (obj->dma_buf) {
+		ret = dma_buf_vmap_unlocked(obj->dma_buf, &map);
+		if (ret < 0)
+			return ERR_PTR(ret);
+
+		return map.vaddr;
+	}
+
+	vaddr = vmap(obj->pages, obj->num_pages, VM_MAP,
+		     pgprot_writecombine(PAGE_KERNEL));
+	if (!vaddr)
+		return ERR_PTR(-ENOMEM);
+
+	return vaddr;
+}
+
+static void tegra_bo_munmap(struct host1x_bo *bo, void *addr)
+{
+	struct tegra_bo *obj = host1x_to_tegra_bo(bo);
+	struct iosys_map map = IOSYS_MAP_INIT_VADDR(addr);
+
+	if (obj->vaddr)
+		return;
+
+	if (obj->dma_buf)
+		return dma_buf_vunmap_unlocked(obj->dma_buf, &map);
+
+	vunmap(addr);
+}
+
+static struct host1x_bo *tegra_bo_get(struct host1x_bo *bo)
+{
+	struct tegra_bo *obj = host1x_to_tegra_bo(bo);
+
+	drm_gem_object_get(&obj->gem);
+
+	return bo;
+}
+
+static const struct host1x_bo_ops tegra_bo_ops = {
+	.get = tegra_bo_get,
+	.put = tegra_bo_put,
+	.pin = tegra_bo_pin,
+	.unpin = tegra_bo_unpin,
+	.mmap = tegra_bo_mmap,
+	.munmap = tegra_bo_munmap,
+};
+
+static int tegra_bo_iommu_map(struct tegra_drm *tegra, struct tegra_bo *bo)
+{
+	int prot = IOMMU_READ | IOMMU_WRITE;
+	ssize_t size;
+	int err;
+
+	if (bo->mm)
+		return -EBUSY;
+
+	bo->mm = kzalloc_obj(*bo->mm);
+	if (!bo->mm)
+		return -ENOMEM;
+
+	mutex_lock(&tegra->mm_lock);
+
+	err = drm_mm_insert_node_generic(&tegra->mm,
+					 bo->mm, bo->gem.size, PAGE_SIZE, 0, 0);
+	if (err < 0) {
+		dev_err(tegra->drm->dev, "out of I/O virtual memory: %d\n",
+			err);
+		goto unlock;
+	}
+
+	bo->iova = bo->mm->start;
+
+	size = iommu_map_sgtable(tegra->domain, bo->iova, bo->sgt, prot);
+	if (size < 0) {
+		dev_err(tegra->drm->dev, "failed to map buffer\n");
+		err = size;
+		goto remove;
+	}
+
+	bo->size = size;
+
+	mutex_unlock(&tegra->mm_lock);
+
+	return 0;
+
+remove:
+	drm_mm_remove_node(bo->mm);
+unlock:
+	mutex_unlock(&tegra->mm_lock);
+	kfree(bo->mm);
+	return err;
+}
+
+static int tegra_bo_iommu_unmap(struct tegra_drm *tegra, struct tegra_bo *bo)
+{
+	if (!bo->mm)
+		return 0;
+
+	mutex_lock(&tegra->mm_lock);
+	iommu_unmap(tegra->domain, bo->iova, bo->size);
+	drm_mm_remove_node(bo->mm);
+	mutex_unlock(&tegra->mm_lock);
+
+	kfree(bo->mm);
+
+	return 0;
+}
+
+static const struct drm_gem_object_funcs tegra_gem_object_funcs = {
+	.free = tegra_bo_free_object,
+	.export = tegra_gem_prime_export,
+	.vm_ops = &tegra_bo_vm_ops,
+};
+
+static struct tegra_bo *tegra_bo_alloc_object(struct drm_device *drm,
+					      size_t size)
+{
+	struct tegra_bo *bo;
+	int err;
+
+	bo = kzalloc_obj(*bo);
+	if (!bo)
+		return ERR_PTR(-ENOMEM);
+
+	bo->gem.funcs = &tegra_gem_object_funcs;
+
+	host1x_bo_init(&bo->base, &tegra_bo_ops);
+	size = round_up(size, PAGE_SIZE);
+
+	err = drm_gem_object_init(drm, &bo->gem, size);
+	if (err < 0)
+		goto free;
+
+	err = drm_gem_create_mmap_offset(&bo->gem);
+	if (err < 0)
+		goto release;
+
+	return bo;
+
+release:
+	drm_gem_object_release(&bo->gem);
+free:
+	kfree(bo);
+	return ERR_PTR(err);
+}
+
+static void tegra_bo_free(struct drm_device *drm, struct tegra_bo *bo)
+{
+	if (bo->pages) {
+		dma_unmap_sgtable(drm->dev, bo->sgt, DMA_FROM_DEVICE, 0);
+		drm_gem_put_pages(&bo->gem, bo->pages, true, true);
+		sg_free_table(bo->sgt);
+		kfree(bo->sgt);
+	} else if (bo->vaddr) {
+		dma_free_wc(drm->dev, bo->gem.size, bo->vaddr, bo->iova);
+	}
+}
+
+static int tegra_bo_get_pages(struct drm_device *drm, struct tegra_bo *bo)
+{
+	int err;
+
+	bo->pages = drm_gem_get_pages(&bo->gem);
+	if (IS_ERR(bo->pages))
+		return PTR_ERR(bo->pages);
+
+	bo->num_pages = bo->gem.size >> PAGE_SHIFT;
+
+	bo->sgt = drm_prime_pages_to_sg(bo->gem.dev, bo->pages, bo->num_pages);
+	if (IS_ERR(bo->sgt)) {
+		err = PTR_ERR(bo->sgt);
+		goto put_pages;
+	}
+
+	err = dma_map_sgtable(drm->dev, bo->sgt, DMA_FROM_DEVICE, 0);
+	if (err)
+		goto free_sgt;
+
+	return 0;
+
+free_sgt:
+	sg_free_table(bo->sgt);
+	kfree(bo->sgt);
+put_pages:
+	drm_gem_put_pages(&bo->gem, bo->pages, false, false);
+	return err;
+}
+
+static int tegra_bo_alloc(struct drm_device *drm, struct tegra_bo *bo)
+{
+	struct tegra_drm *tegra = drm->dev_private;
+	int err;
+
+	if (tegra->domain) {
+		err = tegra_bo_get_pages(drm, bo);
+		if (err < 0)
+			return err;
+
+		err = tegra_bo_iommu_map(tegra, bo);
+		if (err < 0) {
+			tegra_bo_free(drm, bo);
+			return err;
+		}
+	} else {
+		size_t size = bo->gem.size;
+
+		bo->vaddr = dma_alloc_wc(drm->dev, size, &bo->iova,
+					 GFP_KERNEL | __GFP_NOWARN);
+		if (!bo->vaddr) {
+			dev_err(drm->dev,
+				"failed to allocate buffer of size %zu\n",
+				size);
+			return -ENOMEM;
+		}
+	}
+
+	return 0;
+}
+
+struct tegra_bo *tegra_bo_create(struct drm_device *drm, size_t size,
+				 unsigned long flags)
+{
+	struct tegra_bo *bo;
+	int err;
+
+	bo = tegra_bo_alloc_object(drm, size);
+	if (IS_ERR(bo))
+		return bo;
+
+	err = tegra_bo_alloc(drm, bo);
+	if (err < 0)
+		goto release;
+
+	if (flags & DRM_TEGRA_GEM_CREATE_TILED)
+		bo->tiling.mode = TEGRA_BO_TILING_MODE_TILED;
+
+	if (flags & DRM_TEGRA_GEM_CREATE_BOTTOM_UP)
+		bo->flags |= TEGRA_BO_BOTTOM_UP;
+
+	return bo;
+
+release:
+	drm_gem_object_release(&bo->gem);
+	kfree(bo);
+	return ERR_PTR(err);
+}
+
+struct tegra_bo *tegra_bo_create_with_handle(struct drm_file *file,
+					     struct drm_device *drm,
+					     size_t size,
+					     unsigned long flags,
+					     u32 *handle)
+{
+	struct tegra_bo *bo;
+	int err;
+
+	bo = tegra_bo_create(drm, size, flags);
+	if (IS_ERR(bo))
+		return bo;
+
+	err = drm_gem_handle_create(file, &bo->gem, handle);
+	if (err) {
+		tegra_bo_free_object(&bo->gem);
+		return ERR_PTR(err);
+	}
+
+	drm_gem_object_put(&bo->gem);
+
+	return bo;
+}
+
+static struct tegra_bo *tegra_bo_import(struct drm_device *drm,
+					struct dma_buf *buf)
+{
+	struct tegra_drm *tegra = drm->dev_private;
+	struct dma_buf_attachment *attach;
+	struct tegra_bo *bo;
+	int err;
+
+	bo = tegra_bo_alloc_object(drm, buf->size);
+	if (IS_ERR(bo))
+		return bo;
+
+	/*
+	 * If we need to use IOMMU API to map the dma-buf into the internally managed
+	 * domain, map it first to the DRM device to get an sgt.
+	 */
+	if (tegra->domain) {
+		attach = dma_buf_attach(buf, drm->dev);
+		if (IS_ERR(attach)) {
+			err = PTR_ERR(attach);
+			goto free;
+		}
+
+		bo->sgt = dma_buf_map_attachment_unlocked(attach, DMA_TO_DEVICE);
+		if (IS_ERR(bo->sgt)) {
+			err = PTR_ERR(bo->sgt);
+			goto detach;
+		}
+
+		err = tegra_bo_iommu_map(tegra, bo);
+		if (err < 0)
+			goto detach;
+
+		bo->gem.import_attach = attach;
+	}
+
+	get_dma_buf(buf);
+	bo->dma_buf = buf;
+
+	return bo;
+
+detach:
+	if (!IS_ERR_OR_NULL(bo->sgt))
+		dma_buf_unmap_attachment_unlocked(attach, bo->sgt, DMA_TO_DEVICE);
+
+	dma_buf_detach(buf, attach);
+	dma_buf_put(buf);
+free:
+	drm_gem_object_release(&bo->gem);
+	kfree(bo);
+	return ERR_PTR(err);
+}
+
+void tegra_bo_free_object(struct drm_gem_object *gem)
+{
+	struct tegra_drm *tegra = gem->dev->dev_private;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+
+	host1x_bo_clear_cached_mappings(&bo->base);
+
+	if (tegra->domain) {
+		tegra_bo_iommu_unmap(tegra, bo);
+
+		if (drm_gem_is_imported(gem)) {
+			dma_buf_unmap_attachment_unlocked(gem->import_attach, bo->sgt,
+							  DMA_TO_DEVICE);
+			dma_buf_detach(gem->import_attach->dmabuf, gem->import_attach);
+		}
+	}
+
+	tegra_bo_free(gem->dev, bo);
+
+	if (bo->dma_buf)
+		dma_buf_put(bo->dma_buf);
+
+	drm_gem_object_release(gem);
+	kfree(bo);
+}
+
+int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
+			 struct drm_mode_create_dumb *args)
+{
+	struct tegra_drm *tegra = drm->dev_private;
+	struct tegra_bo *bo;
+	int ret;
+
+	ret = drm_mode_size_dumb(drm, args, tegra->pitch_align, 0);
+	if (ret)
+		return ret;
+
+	bo = tegra_bo_create_with_handle(file, drm, args->size, 0,
+					 &args->handle);
+	if (IS_ERR(bo))
+		return PTR_ERR(bo);
+
+	return 0;
+}
+
+static vm_fault_t tegra_bo_fault(struct vm_fault *vmf)
+{
+	struct vm_area_struct *vma = vmf->vma;
+	struct drm_gem_object *gem = vma->vm_private_data;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+	struct page *page;
+	pgoff_t offset;
+
+	if (!bo->pages)
+		return VM_FAULT_SIGBUS;
+
+	offset = linear_page_delta(vma, vmf->address);
+	page = bo->pages[offset];
+
+	return vmf_insert_page(vma, vmf->address, page);
+}
+
+const struct vm_operations_struct tegra_bo_vm_ops = {
+	.fault = tegra_bo_fault,
+	.open = drm_gem_vm_open,
+	.close = drm_gem_vm_close,
+};
+
+int __tegra_gem_mmap(struct drm_gem_object *gem, struct vm_area_struct *vma)
+{
+	struct tegra_bo *bo = to_tegra_bo(gem);
+
+	if (!bo->pages) {
+		unsigned long vm_pgoff = vma->vm_pgoff;
+		int err;
+
+		/*
+		 * Clear the VM_PFNMAP flag that was set by drm_gem_mmap(),
+		 * and set the vm_pgoff (used as a fake buffer offset by DRM)
+		 * to 0 as we want to map the whole buffer.
+		 */
+		vm_flags_clear(vma, VM_PFNMAP);
+		vma->vm_pgoff = 0;
+
+		err = dma_mmap_wc(gem->dev->dev, vma, bo->vaddr, bo->iova,
+				  gem->size);
+		if (err < 0) {
+			drm_gem_vm_close(vma);
+			return err;
+		}
+
+		vma->vm_pgoff = vm_pgoff;
+	} else {
+		pgprot_t prot = vma_get_page_prot(vma);
+
+		vm_flags_mod(vma, VM_MIXEDMAP, VM_PFNMAP);
+
+		vma->vm_page_prot = pgprot_writecombine(prot);
+	}
+
+	return 0;
+}
+
+int tegra_drm_mmap(struct file *file, struct vm_area_struct *vma)
+{
+	struct drm_gem_object *gem;
+	int err;
+
+	err = drm_gem_mmap(file, vma);
+	if (err < 0)
+		return err;
+
+	gem = vma->vm_private_data;
+
+	return __tegra_gem_mmap(gem, vma);
+}
+
+static struct sg_table *
+tegra_gem_prime_map_dma_buf(struct dma_buf_attachment *attach,
+			    enum dma_data_direction dir)
+{
+	struct drm_gem_object *gem = attach->dmabuf->priv;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+	struct sg_table *sgt;
+
+	sgt = kmalloc_obj(*sgt);
+	if (!sgt)
+		return NULL;
+
+	if (bo->pages) {
+		if (sg_alloc_table_from_pages(sgt, bo->pages, bo->num_pages,
+					      0, gem->size, GFP_KERNEL) < 0)
+			goto free;
+	} else {
+		if (dma_get_sgtable(attach->dev, sgt, bo->vaddr, bo->iova,
+				    gem->size) < 0)
+			goto free;
+	}
+
+	if (dma_map_sgtable(attach->dev, sgt, dir, 0))
+		goto free;
+
+	return sgt;
+
+free:
+	sg_free_table(sgt);
+	kfree(sgt);
+	return NULL;
+}
+
+static void tegra_gem_prime_unmap_dma_buf(struct dma_buf_attachment *attach,
+					  struct sg_table *sgt,
+					  enum dma_data_direction dir)
+{
+	struct drm_gem_object *gem = attach->dmabuf->priv;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+
+	if (bo->pages)
+		dma_unmap_sgtable(attach->dev, sgt, dir, 0);
+
+	sg_free_table(sgt);
+	kfree(sgt);
+}
+
+static void tegra_gem_prime_release(struct dma_buf *buf)
+{
+	drm_gem_dmabuf_release(buf);
+}
+
+static int tegra_gem_prime_begin_cpu_access(struct dma_buf *buf,
+					    enum dma_data_direction direction)
+{
+	struct drm_gem_object *gem = buf->priv;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+	struct drm_device *drm = gem->dev;
+
+	if (bo->pages)
+		dma_sync_sgtable_for_cpu(drm->dev, bo->sgt, DMA_FROM_DEVICE);
+
+	return 0;
+}
+
+static int tegra_gem_prime_end_cpu_access(struct dma_buf *buf,
+					  enum dma_data_direction direction)
+{
+	struct drm_gem_object *gem = buf->priv;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+	struct drm_device *drm = gem->dev;
+
+	if (bo->pages)
+		dma_sync_sgtable_for_device(drm->dev, bo->sgt, DMA_TO_DEVICE);
+
+	return 0;
+}
+
+static int tegra_gem_prime_mmap(struct dma_buf *buf, struct vm_area_struct *vma)
+{
+	struct drm_gem_object *gem = buf->priv;
+	int err;
+
+	err = drm_gem_mmap_obj(gem, gem->size, vma);
+	if (err < 0)
+		return err;
+
+	return __tegra_gem_mmap(gem, vma);
+}
+
+static int tegra_gem_prime_vmap(struct dma_buf *buf, struct iosys_map *map)
+{
+	struct drm_gem_object *gem = buf->priv;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+	void *vaddr;
+
+	vaddr = tegra_bo_mmap(&bo->base);
+	if (IS_ERR(vaddr))
+		return PTR_ERR(vaddr);
+
+	iosys_map_set_vaddr(map, vaddr);
+
+	return 0;
+}
+
+static void tegra_gem_prime_vunmap(struct dma_buf *buf, struct iosys_map *map)
+{
+	struct drm_gem_object *gem = buf->priv;
+	struct tegra_bo *bo = to_tegra_bo(gem);
+
+	tegra_bo_munmap(&bo->base, map->vaddr);
+}
+
+static const struct dma_buf_ops tegra_gem_prime_dmabuf_ops = {
+	.map_dma_buf = tegra_gem_prime_map_dma_buf,
+	.unmap_dma_buf = tegra_gem_prime_unmap_dma_buf,
+	.release = tegra_gem_prime_release,
+	.begin_cpu_access = tegra_gem_prime_begin_cpu_access,
+	.end_cpu_access = tegra_gem_prime_end_cpu_access,
+	.mmap = tegra_gem_prime_mmap,
+	.vmap = tegra_gem_prime_vmap,
+	.vunmap = tegra_gem_prime_vunmap,
+};
+
+struct dma_buf *tegra_gem_prime_export(struct drm_gem_object *gem,
+				       int flags)
+{
+	DEFINE_DMA_BUF_EXPORT_INFO(exp_info);
+
+	exp_info.exp_name = KBUILD_MODNAME;
+	exp_info.owner = gem->dev->driver->fops->owner;
+	exp_info.ops = &tegra_gem_prime_dmabuf_ops;
+	exp_info.size = gem->size;
+	exp_info.flags = flags;
+	exp_info.priv = gem;
+
+	return drm_gem_dmabuf_export(gem->dev, &exp_info);
+}
+
+struct drm_gem_object *tegra_gem_prime_import(struct drm_device *drm,
+					      struct dma_buf *buf)
+{
+	struct tegra_bo *bo;
+
+	if (buf->ops == &tegra_gem_prime_dmabuf_ops) {
+		struct drm_gem_object *gem = buf->priv;
+
+		if (gem->dev == drm) {
+			drm_gem_object_get(gem);
+			return gem;
+		}
+	}
+
+	bo = tegra_bo_import(drm, buf);
+	if (IS_ERR(bo))
+		return ERR_CAST(bo);
+
+	return &bo->gem;
+}
+
+struct host1x_bo *tegra_gem_lookup(struct drm_file *file, u32 handle)
+{
+	struct drm_gem_object *gem;
+	struct tegra_bo *bo;
+
+	gem = drm_gem_object_lookup(file, handle);
+	if (!gem)
+		return NULL;
+
+	bo = to_tegra_bo(gem);
+	return &bo->base;
+}
diff --git a/drivers/gpu/drm/tegra/gem.h b/drivers/gpu/drm/tegra/gem.h
new file mode 100644
index 000000000..bf2cbd48e
--- /dev/null
+++ b/drivers/gpu/drm/tegra/gem.h
@@ -0,0 +1,106 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Tegra host1x GEM implementation
+ *
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ */
+
+#ifndef __HOST1X_GEM_H
+#define __HOST1X_GEM_H
+
+#include <linux/host1x.h>
+
+#include <drm/drm.h>
+#include <drm/drm_gem.h>
+
+#define TEGRA_BO_BOTTOM_UP (1 << 0)
+
+enum tegra_bo_tiling_mode {
+	TEGRA_BO_TILING_MODE_PITCH,
+	TEGRA_BO_TILING_MODE_TILED,
+	TEGRA_BO_TILING_MODE_BLOCK,
+};
+
+enum tegra_bo_sector_layout {
+	TEGRA_BO_SECTOR_LAYOUT_TEGRA,
+	TEGRA_BO_SECTOR_LAYOUT_GPU,
+};
+
+struct tegra_bo_tiling {
+	enum tegra_bo_tiling_mode mode;
+	unsigned long value;
+	enum tegra_bo_sector_layout sector_layout;
+};
+
+/*
+ * How memory is referenced within a tegra_bo:
+ *
+ * Buffer source  | Mapping API(*)  | Fields
+ * ---------------+-----------------+---------------
+ * Allocated here | DMA API         | iova (IOVA mapped to drm->dev), vaddr (CPU VA)
+ *
+ * Allocated here | IOMMU API       | pages/num_pages (Phys. memory), sgt (Mapped to drm->dev),
+ *                                  | iova/size (Mapped to domain)
+ *
+ * Imported       | DMA API         | dma_buf (Imported dma_buf)
+ *
+ * Imported       | IOMMU API       | dma_buf (Imported dma_buf),
+ *                                  | gem->import_attach (Attachment on drm->dev),
+ *                                  | sgt (Mapped to drm->dev)
+ *                                  | iova/size (Mapped to domain)
+ *
+ * (*) If tegra->domain is set, i.e. TegraDRM IOMMU domain is directly managed through IOMMU API,
+ *     this is IOMMU API. Otherwise DMA API.
+ */
+struct tegra_bo {
+	struct drm_gem_object gem;
+	struct host1x_bo base;
+	unsigned long flags;
+	struct sg_table *sgt;
+	dma_addr_t iova;
+	void *vaddr;
+	struct dma_buf *dma_buf;
+
+	struct drm_mm_node *mm;
+	unsigned long num_pages;
+	struct page **pages;
+	/* size of IOMMU mapping */
+	size_t size;
+
+	struct tegra_bo_tiling tiling;
+};
+
+static inline struct tegra_bo *to_tegra_bo(struct drm_gem_object *gem)
+{
+	return container_of(gem, struct tegra_bo, gem);
+}
+
+static inline struct tegra_bo *host1x_to_tegra_bo(struct host1x_bo *bo)
+{
+	return container_of(bo, struct tegra_bo, base);
+}
+
+struct tegra_bo *tegra_bo_create(struct drm_device *drm, size_t size,
+				 unsigned long flags);
+struct tegra_bo *tegra_bo_create_with_handle(struct drm_file *file,
+					     struct drm_device *drm,
+					     size_t size,
+					     unsigned long flags,
+					     u32 *handle);
+void tegra_bo_free_object(struct drm_gem_object *gem);
+int tegra_bo_dumb_create(struct drm_file *file, struct drm_device *drm,
+			 struct drm_mode_create_dumb *args);
+
+extern const struct vm_operations_struct tegra_bo_vm_ops;
+
+int __tegra_gem_mmap(struct drm_gem_object *gem, struct vm_area_struct *vma);
+int tegra_drm_mmap(struct file *file, struct vm_area_struct *vma);
+
+struct dma_buf *tegra_gem_prime_export(struct drm_gem_object *gem,
+				       int flags);
+struct drm_gem_object *tegra_gem_prime_import(struct drm_device *drm,
+					      struct dma_buf *buf);
+
+struct host1x_bo *tegra_gem_lookup(struct drm_file *file, u32 handle);
+
+#endif
diff --git a/drivers/gpu/drm/tegra/gr2d.c b/drivers/gpu/drm/tegra/gr2d.c
new file mode 100644
index 000000000..892e3450b
--- /dev/null
+++ b/drivers/gpu/drm/tegra/gr2d.c
@@ -0,0 +1,397 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2012-2013, NVIDIA Corporation.
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/iommu.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/reset.h>
+
+#include <soc/tegra/common.h>
+
+#include "drm.h"
+#include "gem.h"
+#include "gr2d.h"
+
+enum {
+	RST_MC,
+	RST_GR2D,
+	RST_GR2D_MAX,
+};
+
+struct gr2d_soc {
+	unsigned int version;
+};
+
+struct gr2d {
+	struct tegra_drm_client client;
+	struct host1x_channel *channel;
+	struct clk *clk;
+
+	struct reset_control_bulk_data resets[RST_GR2D_MAX];
+	unsigned int nresets;
+
+	const struct gr2d_soc *soc;
+
+	DECLARE_BITMAP(addr_regs, GR2D_NUM_REGS);
+};
+
+static inline struct gr2d *to_gr2d(struct tegra_drm_client *client)
+{
+	return container_of(client, struct gr2d, client);
+}
+
+static int gr2d_init(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	unsigned long flags = HOST1X_SYNCPT_HAS_BASE;
+	struct gr2d *gr2d = to_gr2d(drm);
+	int err;
+
+	gr2d->channel = host1x_channel_request(client);
+	if (!gr2d->channel)
+		return -ENOMEM;
+
+	client->syncpts[0] = host1x_syncpt_request(client, flags);
+	if (!client->syncpts[0]) {
+		err = -ENOMEM;
+		dev_err(client->dev, "failed to request syncpoint: %d\n", err);
+		goto put;
+	}
+
+	err = host1x_client_iommu_attach(client);
+	if (err < 0) {
+		dev_err(client->dev, "failed to attach to domain: %d\n", err);
+		goto free;
+	}
+
+	err = tegra_drm_register_client(dev->dev_private, drm);
+	if (err < 0) {
+		dev_err(client->dev, "failed to register client: %d\n", err);
+		goto detach_iommu;
+	}
+
+	return 0;
+
+detach_iommu:
+	host1x_client_iommu_detach(client);
+free:
+	host1x_syncpt_put(client->syncpts[0]);
+put:
+	host1x_channel_put(gr2d->channel);
+	return err;
+}
+
+static int gr2d_exit(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = dev->dev_private;
+	struct gr2d *gr2d = to_gr2d(drm);
+	int err;
+
+	err = tegra_drm_unregister_client(tegra, drm);
+	if (err < 0)
+		return err;
+
+	host1x_client_iommu_detach(client);
+	host1x_syncpt_put(client->syncpts[0]);
+	host1x_channel_put(gr2d->channel);
+
+	gr2d->channel = NULL;
+
+	return 0;
+}
+
+static const struct host1x_client_ops gr2d_client_ops = {
+	.init = gr2d_init,
+	.exit = gr2d_exit,
+};
+
+static int gr2d_open_channel(struct tegra_drm_client *client,
+			     struct tegra_drm_context *context)
+{
+	struct gr2d *gr2d = to_gr2d(client);
+
+	context->channel = host1x_channel_get(gr2d->channel);
+	if (!context->channel)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static void gr2d_close_channel(struct tegra_drm_context *context)
+{
+	host1x_channel_put(context->channel);
+}
+
+static int gr2d_is_addr_reg(struct device *dev, u32 class, u32 offset)
+{
+	struct gr2d *gr2d = dev_get_drvdata(dev);
+
+	switch (class) {
+	case HOST1X_CLASS_HOST1X:
+		if (offset == 0x2b)
+			return 1;
+
+		break;
+
+	case HOST1X_CLASS_GR2D:
+	case HOST1X_CLASS_GR2D_SB:
+		if (offset >= GR2D_NUM_REGS)
+			break;
+
+		if (test_bit(offset, gr2d->addr_regs))
+			return 1;
+
+		break;
+	}
+
+	return 0;
+}
+
+static int gr2d_is_valid_class(u32 class)
+{
+	return (class == HOST1X_CLASS_GR2D ||
+		class == HOST1X_CLASS_GR2D_SB);
+}
+
+static const struct tegra_drm_client_ops gr2d_ops = {
+	.open_channel = gr2d_open_channel,
+	.close_channel = gr2d_close_channel,
+	.is_addr_reg = gr2d_is_addr_reg,
+	.is_valid_class = gr2d_is_valid_class,
+	.submit = tegra_drm_submit,
+};
+
+static const struct gr2d_soc tegra20_gr2d_soc = {
+	.version = 0x20,
+};
+
+static const struct gr2d_soc tegra30_gr2d_soc = {
+	.version = 0x30,
+};
+
+static const struct gr2d_soc tegra114_gr2d_soc = {
+	.version = 0x35,
+};
+
+static const struct of_device_id gr2d_match[] = {
+	{ .compatible = "nvidia,tegra114-gr2d", .data = &tegra114_gr2d_soc },
+	{ .compatible = "nvidia,tegra30-gr2d", .data = &tegra30_gr2d_soc },
+	{ .compatible = "nvidia,tegra20-gr2d", .data = &tegra20_gr2d_soc },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, gr2d_match);
+
+static const u32 gr2d_addr_regs[] = {
+	GR2D_UA_BASE_ADDR,
+	GR2D_VA_BASE_ADDR,
+	GR2D_PAT_BASE_ADDR,
+	GR2D_DSTA_BASE_ADDR,
+	GR2D_DSTB_BASE_ADDR,
+	GR2D_DSTC_BASE_ADDR,
+	GR2D_SRCA_BASE_ADDR,
+	GR2D_SRCB_BASE_ADDR,
+	GR2D_PATBASE_ADDR,
+	GR2D_SRC_BASE_ADDR_SB,
+	GR2D_DSTA_BASE_ADDR_SB,
+	GR2D_DSTB_BASE_ADDR_SB,
+	GR2D_UA_BASE_ADDR_SB,
+	GR2D_VA_BASE_ADDR_SB,
+};
+
+static int gr2d_get_resets(struct device *dev, struct gr2d *gr2d)
+{
+	int err;
+
+	gr2d->resets[RST_MC].id = "mc";
+	gr2d->resets[RST_GR2D].id = "2d";
+	gr2d->nresets = RST_GR2D_MAX;
+
+	err = devm_reset_control_bulk_get_optional_exclusive_released(
+				dev, gr2d->nresets, gr2d->resets);
+	if (err) {
+		dev_err(dev, "failed to get reset: %d\n", err);
+		return err;
+	}
+
+	if (WARN_ON(!gr2d->resets[RST_GR2D].rstc))
+		return -ENOENT;
+
+	return 0;
+}
+
+static int gr2d_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct host1x_syncpt **syncpts;
+	struct gr2d *gr2d;
+	unsigned int i;
+	int err;
+
+	gr2d = devm_kzalloc(dev, sizeof(*gr2d), GFP_KERNEL);
+	if (!gr2d)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, gr2d);
+
+	gr2d->soc = of_device_get_match_data(dev);
+
+	syncpts = devm_kzalloc(dev, sizeof(*syncpts), GFP_KERNEL);
+	if (!syncpts)
+		return -ENOMEM;
+
+	gr2d->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(gr2d->clk)) {
+		dev_err(dev, "cannot get clock\n");
+		return PTR_ERR(gr2d->clk);
+	}
+
+	err = gr2d_get_resets(dev, gr2d);
+	if (err)
+		return err;
+
+	INIT_LIST_HEAD(&gr2d->client.base.list);
+	gr2d->client.base.ops = &gr2d_client_ops;
+	gr2d->client.base.dev = dev;
+	gr2d->client.base.class = HOST1X_CLASS_GR2D;
+	gr2d->client.base.syncpts = syncpts;
+	gr2d->client.base.num_syncpts = 1;
+
+	INIT_LIST_HEAD(&gr2d->client.list);
+	gr2d->client.version = gr2d->soc->version;
+	gr2d->client.ops = &gr2d_ops;
+
+	err = devm_tegra_core_dev_init_opp_table_common(dev);
+	if (err)
+		return err;
+
+	/* initialize address register map */
+	for (i = 0; i < ARRAY_SIZE(gr2d_addr_regs); i++)
+		set_bit(gr2d_addr_regs[i], gr2d->addr_regs);
+
+	pm_runtime_enable(dev);
+
+	err = host1x_client_register(&gr2d->client.base);
+	if (err < 0) {
+		pm_runtime_disable(dev);
+		dev_err(dev, "failed to register host1x client: %d\n", err);
+		return err;
+	}
+
+	pm_runtime_use_autosuspend(dev);
+	pm_runtime_set_autosuspend_delay(dev, 500);
+
+	return 0;
+}
+
+static void gr2d_remove(struct platform_device *pdev)
+{
+	struct gr2d *gr2d = platform_get_drvdata(pdev);
+
+	pm_runtime_disable(&pdev->dev);
+	host1x_client_unregister(&gr2d->client.base);
+}
+
+static int __maybe_unused gr2d_runtime_suspend(struct device *dev)
+{
+	struct gr2d *gr2d = dev_get_drvdata(dev);
+	int err;
+
+	host1x_channel_stop(gr2d->channel);
+	reset_control_bulk_release(gr2d->nresets, gr2d->resets);
+
+	/*
+	 * GR2D module shouldn't be reset while hardware is idling, otherwise
+	 * host1x's cmdproc will stuck on trying to access any G2 register
+	 * after reset. GR2D module could be either hot-reset or reset after
+	 * power-gating of the HEG partition. Hence we will put in reset only
+	 * the memory client part of the module, the HEG GENPD will take care
+	 * of resetting GR2D module across power-gating.
+	 *
+	 * On Tegra20 there is no HEG partition, but it's okay to have
+	 * undetermined h/w state since userspace is expected to reprogram
+	 * the state on each job submission anyways.
+	 */
+	err = reset_control_acquire(gr2d->resets[RST_MC].rstc);
+	if (err) {
+		dev_err(dev, "failed to acquire MC reset: %d\n", err);
+		goto acquire_reset;
+	}
+
+	err = reset_control_assert(gr2d->resets[RST_MC].rstc);
+	reset_control_release(gr2d->resets[RST_MC].rstc);
+	if (err) {
+		dev_err(dev, "failed to assert MC reset: %d\n", err);
+		goto acquire_reset;
+	}
+
+	clk_disable_unprepare(gr2d->clk);
+
+	return 0;
+
+acquire_reset:
+	reset_control_bulk_acquire(gr2d->nresets, gr2d->resets);
+	reset_control_bulk_deassert(gr2d->nresets, gr2d->resets);
+
+	return err;
+}
+
+static int __maybe_unused gr2d_runtime_resume(struct device *dev)
+{
+	struct gr2d *gr2d = dev_get_drvdata(dev);
+	int err;
+
+	err = reset_control_bulk_acquire(gr2d->nresets, gr2d->resets);
+	if (err) {
+		dev_err(dev, "failed to acquire reset: %d\n", err);
+		return err;
+	}
+
+	err = clk_prepare_enable(gr2d->clk);
+	if (err) {
+		dev_err(dev, "failed to enable clock: %d\n", err);
+		goto release_reset;
+	}
+
+	usleep_range(2000, 4000);
+
+	/* this is a reset array which deasserts both 2D MC and 2D itself */
+	err = reset_control_bulk_deassert(gr2d->nresets, gr2d->resets);
+	if (err) {
+		dev_err(dev, "failed to deassert reset: %d\n", err);
+		goto disable_clk;
+	}
+
+	return 0;
+
+disable_clk:
+	clk_disable_unprepare(gr2d->clk);
+release_reset:
+	reset_control_bulk_release(gr2d->nresets, gr2d->resets);
+
+	return err;
+}
+
+static const struct dev_pm_ops tegra_gr2d_pm = {
+	SET_RUNTIME_PM_OPS(gr2d_runtime_suspend, gr2d_runtime_resume, NULL)
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+};
+
+struct platform_driver tegra_gr2d_driver = {
+	.driver = {
+		.name = "tegra-gr2d",
+		.of_match_table = gr2d_match,
+		.pm = &tegra_gr2d_pm,
+	},
+	.probe = gr2d_probe,
+	.remove = gr2d_remove,
+};
diff --git a/drivers/gpu/drm/tegra/gr2d.h b/drivers/gpu/drm/tegra/gr2d.h
new file mode 100644
index 000000000..9b7d66e15
--- /dev/null
+++ b/drivers/gpu/drm/tegra/gr2d.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#ifndef TEGRA_GR2D_H
+#define TEGRA_GR2D_H
+
+#define GR2D_UA_BASE_ADDR		0x1a
+#define GR2D_VA_BASE_ADDR		0x1b
+#define GR2D_PAT_BASE_ADDR		0x26
+#define GR2D_DSTA_BASE_ADDR		0x2b
+#define GR2D_DSTB_BASE_ADDR		0x2c
+#define GR2D_DSTC_BASE_ADDR		0x2d
+#define GR2D_SRCA_BASE_ADDR		0x31
+#define GR2D_SRCB_BASE_ADDR		0x32
+#define GR2D_PATBASE_ADDR		0x47
+#define GR2D_SRC_BASE_ADDR_SB		0x48
+#define GR2D_DSTA_BASE_ADDR_SB		0x49
+#define GR2D_DSTB_BASE_ADDR_SB		0x4a
+#define GR2D_UA_BASE_ADDR_SB		0x4b
+#define GR2D_VA_BASE_ADDR_SB		0x4c
+
+#define GR2D_NUM_REGS			0x4d
+
+#endif
diff --git a/drivers/gpu/drm/tegra/gr3d.c b/drivers/gpu/drm/tegra/gr3d.c
new file mode 100644
index 000000000..1ec6ced95
--- /dev/null
+++ b/drivers/gpu/drm/tegra/gr3d.c
@@ -0,0 +1,615 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2013 Avionic Design GmbH
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/host1x.h>
+#include <linux/iommu.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_domain.h>
+#include <linux/pm_opp.h>
+#include <linux/pm_runtime.h>
+#include <linux/reset.h>
+
+#include <soc/tegra/common.h>
+#include <soc/tegra/pmc.h>
+
+#include "drm.h"
+#include "gem.h"
+#include "gr3d.h"
+
+enum {
+	RST_MC,
+	RST_GR3D,
+	RST_MC2,
+	RST_GR3D2,
+	RST_GR3D_MAX,
+};
+
+struct gr3d_soc {
+	unsigned int version;
+	unsigned int num_clocks;
+	unsigned int num_resets;
+};
+
+struct gr3d {
+	struct tegra_drm_client client;
+	struct host1x_channel *channel;
+
+	const struct gr3d_soc *soc;
+	struct clk_bulk_data *clocks;
+	unsigned int nclocks;
+	struct reset_control_bulk_data resets[RST_GR3D_MAX];
+	unsigned int nresets;
+	struct tegra_pmc *pmc;
+	struct dev_pm_domain_list *pd_list;
+
+	DECLARE_BITMAP(addr_regs, GR3D_NUM_REGS);
+};
+
+static inline struct gr3d *to_gr3d(struct tegra_drm_client *client)
+{
+	return container_of(client, struct gr3d, client);
+}
+
+static int gr3d_init(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	unsigned long flags = HOST1X_SYNCPT_HAS_BASE;
+	struct gr3d *gr3d = to_gr3d(drm);
+	int err;
+
+	gr3d->channel = host1x_channel_request(client);
+	if (!gr3d->channel)
+		return -ENOMEM;
+
+	client->syncpts[0] = host1x_syncpt_request(client, flags);
+	if (!client->syncpts[0]) {
+		err = -ENOMEM;
+		dev_err(client->dev, "failed to request syncpoint: %d\n", err);
+		goto put;
+	}
+
+	err = host1x_client_iommu_attach(client);
+	if (err < 0) {
+		dev_err(client->dev, "failed to attach to domain: %d\n", err);
+		goto free;
+	}
+
+	err = tegra_drm_register_client(dev->dev_private, drm);
+	if (err < 0) {
+		dev_err(client->dev, "failed to register client: %d\n", err);
+		goto detach_iommu;
+	}
+
+	return 0;
+
+detach_iommu:
+	host1x_client_iommu_detach(client);
+free:
+	host1x_syncpt_put(client->syncpts[0]);
+put:
+	host1x_channel_put(gr3d->channel);
+	return err;
+}
+
+static int gr3d_exit(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	struct gr3d *gr3d = to_gr3d(drm);
+	int err;
+
+	err = tegra_drm_unregister_client(dev->dev_private, drm);
+	if (err < 0)
+		return err;
+
+	host1x_client_iommu_detach(client);
+	host1x_syncpt_put(client->syncpts[0]);
+	host1x_channel_put(gr3d->channel);
+
+	gr3d->channel = NULL;
+
+	return 0;
+}
+
+static const struct host1x_client_ops gr3d_client_ops = {
+	.init = gr3d_init,
+	.exit = gr3d_exit,
+};
+
+static int gr3d_open_channel(struct tegra_drm_client *client,
+			     struct tegra_drm_context *context)
+{
+	struct gr3d *gr3d = to_gr3d(client);
+
+	context->channel = host1x_channel_get(gr3d->channel);
+	if (!context->channel)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static void gr3d_close_channel(struct tegra_drm_context *context)
+{
+	host1x_channel_put(context->channel);
+}
+
+static int gr3d_is_addr_reg(struct device *dev, u32 class, u32 offset)
+{
+	struct gr3d *gr3d = dev_get_drvdata(dev);
+
+	switch (class) {
+	case HOST1X_CLASS_HOST1X:
+		if (offset == 0x2b)
+			return 1;
+
+		break;
+
+	case HOST1X_CLASS_GR3D:
+		if (offset >= GR3D_NUM_REGS)
+			break;
+
+		if (test_bit(offset, gr3d->addr_regs))
+			return 1;
+
+		break;
+	}
+
+	return 0;
+}
+
+static const struct tegra_drm_client_ops gr3d_ops = {
+	.open_channel = gr3d_open_channel,
+	.close_channel = gr3d_close_channel,
+	.is_addr_reg = gr3d_is_addr_reg,
+	.submit = tegra_drm_submit,
+};
+
+static const struct gr3d_soc tegra20_gr3d_soc = {
+	.version = 0x20,
+	.num_clocks = 1,
+	.num_resets = 2,
+};
+
+static const struct gr3d_soc tegra30_gr3d_soc = {
+	.version = 0x30,
+	.num_clocks = 2,
+	.num_resets = 4,
+};
+
+static const struct gr3d_soc tegra114_gr3d_soc = {
+	.version = 0x35,
+	.num_clocks = 1,
+	.num_resets = 2,
+};
+
+static const struct of_device_id tegra_gr3d_match[] = {
+	{ .compatible = "nvidia,tegra114-gr3d", .data = &tegra114_gr3d_soc },
+	{ .compatible = "nvidia,tegra30-gr3d", .data = &tegra30_gr3d_soc },
+	{ .compatible = "nvidia,tegra20-gr3d", .data = &tegra20_gr3d_soc },
+	{ }
+};
+MODULE_DEVICE_TABLE(of, tegra_gr3d_match);
+
+static const u32 gr3d_addr_regs[] = {
+	GR3D_IDX_ATTRIBUTE( 0),
+	GR3D_IDX_ATTRIBUTE( 1),
+	GR3D_IDX_ATTRIBUTE( 2),
+	GR3D_IDX_ATTRIBUTE( 3),
+	GR3D_IDX_ATTRIBUTE( 4),
+	GR3D_IDX_ATTRIBUTE( 5),
+	GR3D_IDX_ATTRIBUTE( 6),
+	GR3D_IDX_ATTRIBUTE( 7),
+	GR3D_IDX_ATTRIBUTE( 8),
+	GR3D_IDX_ATTRIBUTE( 9),
+	GR3D_IDX_ATTRIBUTE(10),
+	GR3D_IDX_ATTRIBUTE(11),
+	GR3D_IDX_ATTRIBUTE(12),
+	GR3D_IDX_ATTRIBUTE(13),
+	GR3D_IDX_ATTRIBUTE(14),
+	GR3D_IDX_ATTRIBUTE(15),
+	GR3D_IDX_INDEX_BASE,
+	GR3D_QR_ZTAG_ADDR,
+	GR3D_QR_CTAG_ADDR,
+	GR3D_QR_CZ_ADDR,
+	GR3D_TEX_TEX_ADDR( 0),
+	GR3D_TEX_TEX_ADDR( 1),
+	GR3D_TEX_TEX_ADDR( 2),
+	GR3D_TEX_TEX_ADDR( 3),
+	GR3D_TEX_TEX_ADDR( 4),
+	GR3D_TEX_TEX_ADDR( 5),
+	GR3D_TEX_TEX_ADDR( 6),
+	GR3D_TEX_TEX_ADDR( 7),
+	GR3D_TEX_TEX_ADDR( 8),
+	GR3D_TEX_TEX_ADDR( 9),
+	GR3D_TEX_TEX_ADDR(10),
+	GR3D_TEX_TEX_ADDR(11),
+	GR3D_TEX_TEX_ADDR(12),
+	GR3D_TEX_TEX_ADDR(13),
+	GR3D_TEX_TEX_ADDR(14),
+	GR3D_TEX_TEX_ADDR(15),
+	GR3D_DW_MEMORY_OUTPUT_ADDRESS,
+	GR3D_GLOBAL_SURFADDR( 0),
+	GR3D_GLOBAL_SURFADDR( 1),
+	GR3D_GLOBAL_SURFADDR( 2),
+	GR3D_GLOBAL_SURFADDR( 3),
+	GR3D_GLOBAL_SURFADDR( 4),
+	GR3D_GLOBAL_SURFADDR( 5),
+	GR3D_GLOBAL_SURFADDR( 6),
+	GR3D_GLOBAL_SURFADDR( 7),
+	GR3D_GLOBAL_SURFADDR( 8),
+	GR3D_GLOBAL_SURFADDR( 9),
+	GR3D_GLOBAL_SURFADDR(10),
+	GR3D_GLOBAL_SURFADDR(11),
+	GR3D_GLOBAL_SURFADDR(12),
+	GR3D_GLOBAL_SURFADDR(13),
+	GR3D_GLOBAL_SURFADDR(14),
+	GR3D_GLOBAL_SURFADDR(15),
+	GR3D_GLOBAL_SPILLSURFADDR,
+	GR3D_GLOBAL_SURFOVERADDR( 0),
+	GR3D_GLOBAL_SURFOVERADDR( 1),
+	GR3D_GLOBAL_SURFOVERADDR( 2),
+	GR3D_GLOBAL_SURFOVERADDR( 3),
+	GR3D_GLOBAL_SURFOVERADDR( 4),
+	GR3D_GLOBAL_SURFOVERADDR( 5),
+	GR3D_GLOBAL_SURFOVERADDR( 6),
+	GR3D_GLOBAL_SURFOVERADDR( 7),
+	GR3D_GLOBAL_SURFOVERADDR( 8),
+	GR3D_GLOBAL_SURFOVERADDR( 9),
+	GR3D_GLOBAL_SURFOVERADDR(10),
+	GR3D_GLOBAL_SURFOVERADDR(11),
+	GR3D_GLOBAL_SURFOVERADDR(12),
+	GR3D_GLOBAL_SURFOVERADDR(13),
+	GR3D_GLOBAL_SURFOVERADDR(14),
+	GR3D_GLOBAL_SURFOVERADDR(15),
+	GR3D_GLOBAL_SAMP01SURFADDR( 0),
+	GR3D_GLOBAL_SAMP01SURFADDR( 1),
+	GR3D_GLOBAL_SAMP01SURFADDR( 2),
+	GR3D_GLOBAL_SAMP01SURFADDR( 3),
+	GR3D_GLOBAL_SAMP01SURFADDR( 4),
+	GR3D_GLOBAL_SAMP01SURFADDR( 5),
+	GR3D_GLOBAL_SAMP01SURFADDR( 6),
+	GR3D_GLOBAL_SAMP01SURFADDR( 7),
+	GR3D_GLOBAL_SAMP01SURFADDR( 8),
+	GR3D_GLOBAL_SAMP01SURFADDR( 9),
+	GR3D_GLOBAL_SAMP01SURFADDR(10),
+	GR3D_GLOBAL_SAMP01SURFADDR(11),
+	GR3D_GLOBAL_SAMP01SURFADDR(12),
+	GR3D_GLOBAL_SAMP01SURFADDR(13),
+	GR3D_GLOBAL_SAMP01SURFADDR(14),
+	GR3D_GLOBAL_SAMP01SURFADDR(15),
+	GR3D_GLOBAL_SAMP23SURFADDR( 0),
+	GR3D_GLOBAL_SAMP23SURFADDR( 1),
+	GR3D_GLOBAL_SAMP23SURFADDR( 2),
+	GR3D_GLOBAL_SAMP23SURFADDR( 3),
+	GR3D_GLOBAL_SAMP23SURFADDR( 4),
+	GR3D_GLOBAL_SAMP23SURFADDR( 5),
+	GR3D_GLOBAL_SAMP23SURFADDR( 6),
+	GR3D_GLOBAL_SAMP23SURFADDR( 7),
+	GR3D_GLOBAL_SAMP23SURFADDR( 8),
+	GR3D_GLOBAL_SAMP23SURFADDR( 9),
+	GR3D_GLOBAL_SAMP23SURFADDR(10),
+	GR3D_GLOBAL_SAMP23SURFADDR(11),
+	GR3D_GLOBAL_SAMP23SURFADDR(12),
+	GR3D_GLOBAL_SAMP23SURFADDR(13),
+	GR3D_GLOBAL_SAMP23SURFADDR(14),
+	GR3D_GLOBAL_SAMP23SURFADDR(15),
+};
+
+static int gr3d_power_up_legacy_domain(struct device *dev, const char *name,
+				       unsigned int id)
+{
+	struct gr3d *gr3d = dev_get_drvdata(dev);
+	struct reset_control *reset;
+	struct clk *clk;
+	unsigned int i;
+	int err;
+
+	/*
+	 * Tegra20 device-tree doesn't specify 3d clock name and there is only
+	 * one clock for Tegra20. Tegra30+ device-trees always specified names
+	 * for the clocks.
+	 */
+	if (gr3d->nclocks == 1) {
+		if (id == TEGRA_POWERGATE_3D1)
+			return 0;
+
+		clk = gr3d->clocks[0].clk;
+	} else {
+		for (i = 0; i < gr3d->nclocks; i++) {
+			if (WARN_ON(!gr3d->clocks[i].id))
+				continue;
+
+			if (!strcmp(gr3d->clocks[i].id, name)) {
+				clk = gr3d->clocks[i].clk;
+				break;
+			}
+		}
+
+		if (WARN_ON(i == gr3d->nclocks))
+			return -EINVAL;
+	}
+
+	/*
+	 * We use array of resets, which includes MC resets, and MC
+	 * reset shouldn't be asserted while hardware is gated because
+	 * MC flushing will fail for gated hardware. Hence for legacy
+	 * PD we request the individual reset separately.
+	 */
+	reset = reset_control_get_exclusive_released(dev, name);
+	if (IS_ERR(reset))
+		return PTR_ERR(reset);
+
+	err = reset_control_acquire(reset);
+	if (err) {
+		dev_err(dev, "failed to acquire %s reset: %d\n", name, err);
+	} else {
+		err = tegra_pmc_powergate_sequence_power_up(gr3d->pmc, id,
+							    clk, reset);
+		reset_control_release(reset);
+	}
+
+	reset_control_put(reset);
+	if (err)
+		return err;
+
+	/*
+	 * tegra_powergate_sequence_power_up() leaves clocks enabled,
+	 * while GENPD not. Hence keep clock-enable balanced.
+	 */
+	clk_disable_unprepare(clk);
+
+	return 0;
+}
+
+static int gr3d_init_power(struct device *dev, struct gr3d *gr3d)
+{
+	struct dev_pm_domain_attach_data pd_data = {
+		.pd_names = (const char *[]) { "3d0", "3d1" },
+		.num_pd_names = 2,
+		.pd_flags = PD_FLAG_REQUIRED_OPP,
+	};
+	int err;
+
+	err = of_count_phandle_with_args(dev->of_node, "power-domains",
+					 "#power-domain-cells");
+	if (err < 0) {
+		if (err != -ENOENT)
+			return err;
+
+		gr3d->pmc = devm_tegra_pmc_get(dev);
+		if (IS_ERR(gr3d->pmc))
+			return dev_err_probe(dev, PTR_ERR(gr3d->pmc),
+					     "failed to get PMC\n");
+
+		/*
+		 * Older device-trees don't use GENPD. In this case we should
+		 * toggle power domain manually.
+		 */
+		err = gr3d_power_up_legacy_domain(dev, "3d",
+						  TEGRA_POWERGATE_3D);
+		if (err)
+			return err;
+
+		err = gr3d_power_up_legacy_domain(dev, "3d2",
+						  TEGRA_POWERGATE_3D1);
+		if (err)
+			return err;
+
+		return 0;
+	}
+
+	/*
+	 * The PM domain core automatically attaches a single power domain,
+	 * otherwise it skips attaching completely. We have a single domain
+	 * on Tegra20 and two domains on Tegra30+.
+	 */
+	if (dev->pm_domain)
+		return 0;
+
+	err = devm_pm_domain_attach_list(dev, &pd_data, &gr3d->pd_list);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static int gr3d_get_clocks(struct device *dev, struct gr3d *gr3d)
+{
+	int err;
+
+	err = devm_clk_bulk_get_all(dev, &gr3d->clocks);
+	if (err < 0) {
+		dev_err(dev, "failed to get clock: %d\n", err);
+		return err;
+	}
+	gr3d->nclocks = err;
+
+	if (gr3d->nclocks != gr3d->soc->num_clocks) {
+		dev_err(dev, "invalid number of clocks: %u\n", gr3d->nclocks);
+		return -ENOENT;
+	}
+
+	return 0;
+}
+
+static int gr3d_get_resets(struct device *dev, struct gr3d *gr3d)
+{
+	int err;
+
+	gr3d->resets[RST_MC].id = "mc";
+	gr3d->resets[RST_MC2].id = "mc2";
+	gr3d->resets[RST_GR3D].id = "3d";
+	gr3d->resets[RST_GR3D2].id = "3d2";
+	gr3d->nresets = gr3d->soc->num_resets;
+
+	err = devm_reset_control_bulk_get_optional_exclusive_released(
+				dev, gr3d->nresets, gr3d->resets);
+	if (err) {
+		dev_err(dev, "failed to get reset: %d\n", err);
+		return err;
+	}
+
+	if (WARN_ON(!gr3d->resets[RST_GR3D].rstc) ||
+	    WARN_ON(!gr3d->resets[RST_GR3D2].rstc && gr3d->nresets == 4))
+		return -ENOENT;
+
+	return 0;
+}
+
+static int gr3d_probe(struct platform_device *pdev)
+{
+	struct host1x_syncpt **syncpts;
+	struct gr3d *gr3d;
+	unsigned int i;
+	int err;
+
+	gr3d = devm_kzalloc(&pdev->dev, sizeof(*gr3d), GFP_KERNEL);
+	if (!gr3d)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, gr3d);
+
+	gr3d->soc = of_device_get_match_data(&pdev->dev);
+
+	syncpts = devm_kzalloc(&pdev->dev, sizeof(*syncpts), GFP_KERNEL);
+	if (!syncpts)
+		return -ENOMEM;
+
+	err = gr3d_get_clocks(&pdev->dev, gr3d);
+	if (err)
+		return err;
+
+	err = gr3d_get_resets(&pdev->dev, gr3d);
+	if (err)
+		return err;
+
+	err = gr3d_init_power(&pdev->dev, gr3d);
+	if (err)
+		return err;
+
+	INIT_LIST_HEAD(&gr3d->client.base.list);
+	gr3d->client.base.ops = &gr3d_client_ops;
+	gr3d->client.base.dev = &pdev->dev;
+	gr3d->client.base.class = HOST1X_CLASS_GR3D;
+	gr3d->client.base.syncpts = syncpts;
+	gr3d->client.base.num_syncpts = 1;
+
+	INIT_LIST_HEAD(&gr3d->client.list);
+	gr3d->client.version = gr3d->soc->version;
+	gr3d->client.ops = &gr3d_ops;
+
+	err = devm_tegra_core_dev_init_opp_table_common(&pdev->dev);
+	if (err)
+		return err;
+
+	/* initialize address register map */
+	for (i = 0; i < ARRAY_SIZE(gr3d_addr_regs); i++)
+		set_bit(gr3d_addr_regs[i], gr3d->addr_regs);
+
+	pm_runtime_enable(&pdev->dev);
+
+	err = host1x_client_register(&gr3d->client.base);
+	if (err < 0) {
+		pm_runtime_disable(&pdev->dev);
+		dev_err(&pdev->dev, "failed to register host1x client: %d\n",
+			err);
+		return err;
+	}
+
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 500);
+
+	return 0;
+}
+
+static void gr3d_remove(struct platform_device *pdev)
+{
+	struct gr3d *gr3d = platform_get_drvdata(pdev);
+
+	pm_runtime_disable(&pdev->dev);
+	host1x_client_unregister(&gr3d->client.base);
+}
+
+static int __maybe_unused gr3d_runtime_suspend(struct device *dev)
+{
+	struct gr3d *gr3d = dev_get_drvdata(dev);
+	int err;
+
+	host1x_channel_stop(gr3d->channel);
+
+	err = reset_control_bulk_assert(gr3d->nresets, gr3d->resets);
+	if (err) {
+		dev_err(dev, "failed to assert reset: %d\n", err);
+		return err;
+	}
+
+	usleep_range(10, 20);
+
+	/*
+	 * Older device-trees don't specify MC resets and power-gating can't
+	 * be done safely in that case. Hence we will keep the power ungated
+	 * for older DTBs. For newer DTBs, GENPD will perform the power-gating.
+	 */
+
+	clk_bulk_disable_unprepare(gr3d->nclocks, gr3d->clocks);
+	reset_control_bulk_release(gr3d->nresets, gr3d->resets);
+
+	return 0;
+}
+
+static int __maybe_unused gr3d_runtime_resume(struct device *dev)
+{
+	struct gr3d *gr3d = dev_get_drvdata(dev);
+	int err;
+
+	err = reset_control_bulk_acquire(gr3d->nresets, gr3d->resets);
+	if (err) {
+		dev_err(dev, "failed to acquire reset: %d\n", err);
+		return err;
+	}
+
+	err = clk_bulk_prepare_enable(gr3d->nclocks, gr3d->clocks);
+	if (err) {
+		dev_err(dev, "failed to enable clock: %d\n", err);
+		goto release_reset;
+	}
+
+	err = reset_control_bulk_deassert(gr3d->nresets, gr3d->resets);
+	if (err) {
+		dev_err(dev, "failed to deassert reset: %d\n", err);
+		goto disable_clk;
+	}
+
+	return 0;
+
+disable_clk:
+	clk_bulk_disable_unprepare(gr3d->nclocks, gr3d->clocks);
+release_reset:
+	reset_control_bulk_release(gr3d->nresets, gr3d->resets);
+
+	return err;
+}
+
+static const struct dev_pm_ops tegra_gr3d_pm = {
+	SET_RUNTIME_PM_OPS(gr3d_runtime_suspend, gr3d_runtime_resume, NULL)
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+};
+
+struct platform_driver tegra_gr3d_driver = {
+	.driver = {
+		.name = "tegra-gr3d",
+		.of_match_table = tegra_gr3d_match,
+		.pm = &tegra_gr3d_pm,
+	},
+	.probe = gr3d_probe,
+	.remove = gr3d_remove,
+};
diff --git a/drivers/gpu/drm/tegra/gr3d.h b/drivers/gpu/drm/tegra/gr3d.h
new file mode 100644
index 000000000..ca2921b68
--- /dev/null
+++ b/drivers/gpu/drm/tegra/gr3d.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#ifndef TEGRA_GR3D_H
+#define TEGRA_GR3D_H
+
+#define GR3D_IDX_ATTRIBUTE(x)		(0x100 + (x) * 2)
+#define GR3D_IDX_INDEX_BASE		0x121
+#define GR3D_QR_ZTAG_ADDR		0x415
+#define GR3D_QR_CTAG_ADDR		0x417
+#define GR3D_QR_CZ_ADDR			0x419
+#define GR3D_TEX_TEX_ADDR(x)		(0x710 + (x))
+#define GR3D_DW_MEMORY_OUTPUT_ADDRESS	0x904
+#define GR3D_GLOBAL_SURFADDR(x)		(0xe00 + (x))
+#define GR3D_GLOBAL_SPILLSURFADDR	0xe2a
+#define GR3D_GLOBAL_SURFOVERADDR(x)	(0xe30 + (x))
+#define GR3D_GLOBAL_SAMP01SURFADDR(x)	(0xe50 + (x))
+#define GR3D_GLOBAL_SAMP23SURFADDR(x)	(0xe60 + (x))
+
+#define GR3D_NUM_REGS			0xe88
+
+#endif
diff --git a/drivers/gpu/drm/tegra/hda.c b/drivers/gpu/drm/tegra/hda.c
new file mode 100644
index 000000000..94245a18a
--- /dev/null
+++ b/drivers/gpu/drm/tegra/hda.c
@@ -0,0 +1,63 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright (C) 2019 NVIDIA Corporation
+ */
+
+#include <linux/bug.h>
+
+#include <sound/hda_verbs.h>
+
+#include "hda.h"
+
+void tegra_hda_parse_format(unsigned int format, struct tegra_hda_format *fmt)
+{
+	unsigned int mul, div, bits, channels;
+
+	if (format & AC_FMT_TYPE_NON_PCM)
+		fmt->pcm = false;
+	else
+		fmt->pcm = true;
+
+	if (format & AC_FMT_BASE_44K)
+		fmt->sample_rate = 44100;
+	else
+		fmt->sample_rate = 48000;
+
+	mul = (format & AC_FMT_MULT_MASK) >> AC_FMT_MULT_SHIFT;
+	div = (format & AC_FMT_DIV_MASK) >> AC_FMT_DIV_SHIFT;
+
+	fmt->sample_rate *= (mul + 1) / (div + 1);
+
+	switch (format & AC_FMT_BITS_MASK) {
+	case AC_FMT_BITS_8:
+		fmt->bits = 8;
+		break;
+
+	case AC_FMT_BITS_16:
+		fmt->bits = 16;
+		break;
+
+	case AC_FMT_BITS_20:
+		fmt->bits = 20;
+		break;
+
+	case AC_FMT_BITS_24:
+		fmt->bits = 24;
+		break;
+
+	case AC_FMT_BITS_32:
+		fmt->bits = 32;
+		break;
+
+	default:
+		bits = (format & AC_FMT_BITS_MASK) >> AC_FMT_BITS_SHIFT;
+		WARN(1, "invalid number of bits: %#x\n", bits);
+		fmt->bits = 8;
+		break;
+	}
+
+	channels = (format & AC_FMT_CHAN_MASK) >> AC_FMT_CHAN_SHIFT;
+
+	/* channels are encoded as n - 1 */
+	fmt->channels = channels + 1;
+}
diff --git a/drivers/gpu/drm/tegra/hda.h b/drivers/gpu/drm/tegra/hda.h
new file mode 100644
index 000000000..77269955a
--- /dev/null
+++ b/drivers/gpu/drm/tegra/hda.h
@@ -0,0 +1,20 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright (C) 2019 NVIDIA Corporation
+ */
+
+#ifndef DRM_TEGRA_HDA_H
+#define DRM_TEGRA_HDA_H 1
+
+#include <linux/types.h>
+
+struct tegra_hda_format {
+	unsigned int sample_rate;
+	unsigned int channels;
+	unsigned int bits;
+	bool pcm;
+};
+
+void tegra_hda_parse_format(unsigned int format, struct tegra_hda_format *fmt);
+
+#endif
diff --git a/drivers/gpu/drm/tegra/hdmi.c b/drivers/gpu/drm/tegra/hdmi.c
new file mode 100644
index 000000000..0d2c0ab0a
--- /dev/null
+++ b/drivers/gpu/drm/tegra/hdmi.c
@@ -0,0 +1,1926 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#include <linux/clk.h>
+#include <linux/debugfs.h>
+#include <linux/delay.h>
+#include <linux/hdmi.h>
+#include <linux/math64.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_opp.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/reset.h>
+
+#include <soc/tegra/common.h>
+#include <sound/hdmi-codec.h>
+
+#include <drm/drm_bridge_connector.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_crtc.h>
+#include <drm/drm_debugfs.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_encoder.h>
+#include <drm/drm_eld.h>
+#include <drm/drm_file.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_print.h>
+#include <drm/drm_probe_helper.h>
+
+#include "hda.h"
+#include "hdmi.h"
+#include "drm.h"
+#include "dc.h"
+#include "trace.h"
+
+#define HDMI_ELD_BUFFER_SIZE 96
+
+struct tmds_config {
+	unsigned int pclk;
+	u32 pll0;
+	u32 pll1;
+	u32 pe_current;
+	u32 drive_current;
+	u32 peak_current;
+};
+
+struct tegra_hdmi_config {
+	const struct tmds_config *tmds;
+	unsigned int num_tmds;
+
+	unsigned long fuse_override_offset;
+	u32 fuse_override_value;
+
+	bool has_sor_io_peak_current;
+	bool has_hda;
+	bool has_hbr;
+};
+
+struct tegra_hdmi {
+	struct host1x_client client;
+	struct tegra_output output;
+	struct device *dev;
+
+	struct regulator *hdmi;
+	struct regulator *pll;
+	struct regulator *vdd;
+
+	void __iomem *regs;
+	unsigned int irq;
+
+	struct clk *clk_parent;
+	struct clk *clk;
+	struct reset_control *rst;
+
+	const struct tegra_hdmi_config *config;
+
+	unsigned int audio_source;
+	struct tegra_hda_format format;
+
+	unsigned int pixel_clock;
+	bool stereo;
+	bool dvi;
+
+	struct drm_info_list *debugfs_files;
+
+	struct platform_device *audio_pdev;
+	struct mutex audio_lock;
+};
+
+static inline struct tegra_hdmi *
+host1x_client_to_hdmi(struct host1x_client *client)
+{
+	return container_of(client, struct tegra_hdmi, client);
+}
+
+static inline struct tegra_hdmi *to_hdmi(struct tegra_output *output)
+{
+	return container_of(output, struct tegra_hdmi, output);
+}
+
+#define HDMI_AUDIOCLK_FREQ 216000000
+#define HDMI_REKEY_DEFAULT 56
+
+enum {
+	AUTO = 0,
+	SPDIF,
+	HDA,
+};
+
+static inline u32 tegra_hdmi_readl(struct tegra_hdmi *hdmi,
+				   unsigned int offset)
+{
+	u32 value = readl(hdmi->regs + (offset << 2));
+
+	trace_hdmi_readl(hdmi->dev, offset, value);
+
+	return value;
+}
+
+static inline void tegra_hdmi_writel(struct tegra_hdmi *hdmi, u32 value,
+				     unsigned int offset)
+{
+	trace_hdmi_writel(hdmi->dev, offset, value);
+	writel(value, hdmi->regs + (offset << 2));
+}
+
+struct tegra_hdmi_audio_config {
+	unsigned int n;
+	unsigned int cts;
+	unsigned int aval;
+};
+
+static const struct tmds_config tegra20_tmds_config[] = {
+	{ /* slow pixel clock modes */
+		.pclk = 27000000,
+		.pll0 = SOR_PLL_BG_V17_S(3) | SOR_PLL_ICHPMP(1) |
+			SOR_PLL_RESISTORSEL | SOR_PLL_VCOCAP(0) |
+			SOR_PLL_TX_REG_LOAD(3),
+		.pll1 = SOR_PLL_TMDS_TERM_ENABLE,
+		.pe_current = PE_CURRENT0(PE_CURRENT_0_0_mA) |
+			PE_CURRENT1(PE_CURRENT_0_0_mA) |
+			PE_CURRENT2(PE_CURRENT_0_0_mA) |
+			PE_CURRENT3(PE_CURRENT_0_0_mA),
+		.drive_current = DRIVE_CURRENT_LANE0(DRIVE_CURRENT_7_125_mA) |
+			DRIVE_CURRENT_LANE1(DRIVE_CURRENT_7_125_mA) |
+			DRIVE_CURRENT_LANE2(DRIVE_CURRENT_7_125_mA) |
+			DRIVE_CURRENT_LANE3(DRIVE_CURRENT_7_125_mA),
+	},
+	{ /* high pixel clock modes */
+		.pclk = UINT_MAX,
+		.pll0 = SOR_PLL_BG_V17_S(3) | SOR_PLL_ICHPMP(1) |
+			SOR_PLL_RESISTORSEL | SOR_PLL_VCOCAP(1) |
+			SOR_PLL_TX_REG_LOAD(3),
+		.pll1 = SOR_PLL_TMDS_TERM_ENABLE | SOR_PLL_PE_EN,
+		.pe_current = PE_CURRENT0(PE_CURRENT_6_0_mA) |
+			PE_CURRENT1(PE_CURRENT_6_0_mA) |
+			PE_CURRENT2(PE_CURRENT_6_0_mA) |
+			PE_CURRENT3(PE_CURRENT_6_0_mA),
+		.drive_current = DRIVE_CURRENT_LANE0(DRIVE_CURRENT_7_125_mA) |
+			DRIVE_CURRENT_LANE1(DRIVE_CURRENT_7_125_mA) |
+			DRIVE_CURRENT_LANE2(DRIVE_CURRENT_7_125_mA) |
+			DRIVE_CURRENT_LANE3(DRIVE_CURRENT_7_125_mA),
+	},
+};
+
+static const struct tmds_config tegra30_tmds_config[] = {
+	{ /* 480p modes */
+		.pclk = 27000000,
+		.pll0 = SOR_PLL_BG_V17_S(3) | SOR_PLL_ICHPMP(1) |
+			SOR_PLL_RESISTORSEL | SOR_PLL_VCOCAP(0) |
+			SOR_PLL_TX_REG_LOAD(0),
+		.pll1 = SOR_PLL_TMDS_TERM_ENABLE,
+		.pe_current = PE_CURRENT0(PE_CURRENT_0_0_mA) |
+			PE_CURRENT1(PE_CURRENT_0_0_mA) |
+			PE_CURRENT2(PE_CURRENT_0_0_mA) |
+			PE_CURRENT3(PE_CURRENT_0_0_mA),
+		.drive_current = DRIVE_CURRENT_LANE0(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE1(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE2(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE3(DRIVE_CURRENT_5_250_mA),
+	}, { /* 720p modes */
+		.pclk = 74250000,
+		.pll0 = SOR_PLL_BG_V17_S(3) | SOR_PLL_ICHPMP(1) |
+			SOR_PLL_RESISTORSEL | SOR_PLL_VCOCAP(1) |
+			SOR_PLL_TX_REG_LOAD(0),
+		.pll1 = SOR_PLL_TMDS_TERM_ENABLE | SOR_PLL_PE_EN,
+		.pe_current = PE_CURRENT0(PE_CURRENT_5_0_mA) |
+			PE_CURRENT1(PE_CURRENT_5_0_mA) |
+			PE_CURRENT2(PE_CURRENT_5_0_mA) |
+			PE_CURRENT3(PE_CURRENT_5_0_mA),
+		.drive_current = DRIVE_CURRENT_LANE0(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE1(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE2(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE3(DRIVE_CURRENT_5_250_mA),
+	}, { /* 1080p modes */
+		.pclk = UINT_MAX,
+		.pll0 = SOR_PLL_BG_V17_S(3) | SOR_PLL_ICHPMP(1) |
+			SOR_PLL_RESISTORSEL | SOR_PLL_VCOCAP(3) |
+			SOR_PLL_TX_REG_LOAD(0),
+		.pll1 = SOR_PLL_TMDS_TERM_ENABLE | SOR_PLL_PE_EN,
+		.pe_current = PE_CURRENT0(PE_CURRENT_5_0_mA) |
+			PE_CURRENT1(PE_CURRENT_5_0_mA) |
+			PE_CURRENT2(PE_CURRENT_5_0_mA) |
+			PE_CURRENT3(PE_CURRENT_5_0_mA),
+		.drive_current = DRIVE_CURRENT_LANE0(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE1(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE2(DRIVE_CURRENT_5_250_mA) |
+			DRIVE_CURRENT_LANE3(DRIVE_CURRENT_5_250_mA),
+	},
+};
+
+static const struct tmds_config tegra114_tmds_config[] = {
+	{ /* 480p/576p / 25.2MHz/27MHz modes */
+		.pclk = 27000000,
+		.pll0 = SOR_PLL_ICHPMP(1) | SOR_PLL_BG_V17_S(3) |
+			SOR_PLL_VCOCAP(0) | SOR_PLL_RESISTORSEL,
+		.pll1 = SOR_PLL_LOADADJ(3) | SOR_PLL_TMDS_TERMADJ(0),
+		.pe_current = PE_CURRENT0(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT1(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT2(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT3(PE_CURRENT_0_mA_T114),
+		.drive_current =
+			DRIVE_CURRENT_LANE0_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE1_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE2_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE3_T114(DRIVE_CURRENT_10_400_mA_T114),
+		.peak_current = PEAK_CURRENT_LANE0(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE1(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE2(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE3(PEAK_CURRENT_0_000_mA),
+	}, { /* 720p / 74.25MHz modes */
+		.pclk = 74250000,
+		.pll0 = SOR_PLL_ICHPMP(1) | SOR_PLL_BG_V17_S(3) |
+			SOR_PLL_VCOCAP(1) | SOR_PLL_RESISTORSEL,
+		.pll1 = SOR_PLL_PE_EN | SOR_PLL_LOADADJ(3) |
+			SOR_PLL_TMDS_TERMADJ(0),
+		.pe_current = PE_CURRENT0(PE_CURRENT_15_mA_T114) |
+			PE_CURRENT1(PE_CURRENT_15_mA_T114) |
+			PE_CURRENT2(PE_CURRENT_15_mA_T114) |
+			PE_CURRENT3(PE_CURRENT_15_mA_T114),
+		.drive_current =
+			DRIVE_CURRENT_LANE0_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE1_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE2_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE3_T114(DRIVE_CURRENT_10_400_mA_T114),
+		.peak_current = PEAK_CURRENT_LANE0(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE1(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE2(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE3(PEAK_CURRENT_0_000_mA),
+	}, { /* 1080p / 148.5MHz modes */
+		.pclk = 148500000,
+		.pll0 = SOR_PLL_ICHPMP(1) | SOR_PLL_BG_V17_S(3) |
+			SOR_PLL_VCOCAP(3) | SOR_PLL_RESISTORSEL,
+		.pll1 = SOR_PLL_PE_EN | SOR_PLL_LOADADJ(3) |
+			SOR_PLL_TMDS_TERMADJ(0),
+		.pe_current = PE_CURRENT0(PE_CURRENT_10_mA_T114) |
+			PE_CURRENT1(PE_CURRENT_10_mA_T114) |
+			PE_CURRENT2(PE_CURRENT_10_mA_T114) |
+			PE_CURRENT3(PE_CURRENT_10_mA_T114),
+		.drive_current =
+			DRIVE_CURRENT_LANE0_T114(DRIVE_CURRENT_12_400_mA_T114) |
+			DRIVE_CURRENT_LANE1_T114(DRIVE_CURRENT_12_400_mA_T114) |
+			DRIVE_CURRENT_LANE2_T114(DRIVE_CURRENT_12_400_mA_T114) |
+			DRIVE_CURRENT_LANE3_T114(DRIVE_CURRENT_12_400_mA_T114),
+		.peak_current = PEAK_CURRENT_LANE0(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE1(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE2(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE3(PEAK_CURRENT_0_000_mA),
+	}, { /* 225/297MHz modes */
+		.pclk = UINT_MAX,
+		.pll0 = SOR_PLL_ICHPMP(1) | SOR_PLL_BG_V17_S(3) |
+			SOR_PLL_VCOCAP(0xf) | SOR_PLL_RESISTORSEL,
+		.pll1 = SOR_PLL_LOADADJ(3) | SOR_PLL_TMDS_TERMADJ(7)
+			| SOR_PLL_TMDS_TERM_ENABLE,
+		.pe_current = PE_CURRENT0(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT1(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT2(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT3(PE_CURRENT_0_mA_T114),
+		.drive_current =
+			DRIVE_CURRENT_LANE0_T114(DRIVE_CURRENT_25_200_mA_T114) |
+			DRIVE_CURRENT_LANE1_T114(DRIVE_CURRENT_25_200_mA_T114) |
+			DRIVE_CURRENT_LANE2_T114(DRIVE_CURRENT_25_200_mA_T114) |
+			DRIVE_CURRENT_LANE3_T114(DRIVE_CURRENT_19_200_mA_T114),
+		.peak_current = PEAK_CURRENT_LANE0(PEAK_CURRENT_3_000_mA) |
+			PEAK_CURRENT_LANE1(PEAK_CURRENT_3_000_mA) |
+			PEAK_CURRENT_LANE2(PEAK_CURRENT_3_000_mA) |
+			PEAK_CURRENT_LANE3(PEAK_CURRENT_0_800_mA),
+	},
+};
+
+static const struct tmds_config tegra124_tmds_config[] = {
+	{ /* 480p/576p / 25.2MHz/27MHz modes */
+		.pclk = 27000000,
+		.pll0 = SOR_PLL_ICHPMP(1) | SOR_PLL_BG_V17_S(3) |
+			SOR_PLL_VCOCAP(0) | SOR_PLL_RESISTORSEL,
+		.pll1 = SOR_PLL_LOADADJ(3) | SOR_PLL_TMDS_TERMADJ(0),
+		.pe_current = PE_CURRENT0(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT1(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT2(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT3(PE_CURRENT_0_mA_T114),
+		.drive_current =
+			DRIVE_CURRENT_LANE0_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE1_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE2_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE3_T114(DRIVE_CURRENT_10_400_mA_T114),
+		.peak_current = PEAK_CURRENT_LANE0(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE1(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE2(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE3(PEAK_CURRENT_0_000_mA),
+	}, { /* 720p / 74.25MHz modes */
+		.pclk = 74250000,
+		.pll0 = SOR_PLL_ICHPMP(1) | SOR_PLL_BG_V17_S(3) |
+			SOR_PLL_VCOCAP(1) | SOR_PLL_RESISTORSEL,
+		.pll1 = SOR_PLL_PE_EN | SOR_PLL_LOADADJ(3) |
+			SOR_PLL_TMDS_TERMADJ(0),
+		.pe_current = PE_CURRENT0(PE_CURRENT_15_mA_T114) |
+			PE_CURRENT1(PE_CURRENT_15_mA_T114) |
+			PE_CURRENT2(PE_CURRENT_15_mA_T114) |
+			PE_CURRENT3(PE_CURRENT_15_mA_T114),
+		.drive_current =
+			DRIVE_CURRENT_LANE0_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE1_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE2_T114(DRIVE_CURRENT_10_400_mA_T114) |
+			DRIVE_CURRENT_LANE3_T114(DRIVE_CURRENT_10_400_mA_T114),
+		.peak_current = PEAK_CURRENT_LANE0(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE1(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE2(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE3(PEAK_CURRENT_0_000_mA),
+	}, { /* 1080p / 148.5MHz modes */
+		.pclk = 148500000,
+		.pll0 = SOR_PLL_ICHPMP(1) | SOR_PLL_BG_V17_S(3) |
+			SOR_PLL_VCOCAP(3) | SOR_PLL_RESISTORSEL,
+		.pll1 = SOR_PLL_PE_EN | SOR_PLL_LOADADJ(3) |
+			SOR_PLL_TMDS_TERMADJ(0),
+		.pe_current = PE_CURRENT0(PE_CURRENT_10_mA_T114) |
+			PE_CURRENT1(PE_CURRENT_10_mA_T114) |
+			PE_CURRENT2(PE_CURRENT_10_mA_T114) |
+			PE_CURRENT3(PE_CURRENT_10_mA_T114),
+		.drive_current =
+			DRIVE_CURRENT_LANE0_T114(DRIVE_CURRENT_12_400_mA_T114) |
+			DRIVE_CURRENT_LANE1_T114(DRIVE_CURRENT_12_400_mA_T114) |
+			DRIVE_CURRENT_LANE2_T114(DRIVE_CURRENT_12_400_mA_T114) |
+			DRIVE_CURRENT_LANE3_T114(DRIVE_CURRENT_12_400_mA_T114),
+		.peak_current = PEAK_CURRENT_LANE0(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE1(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE2(PEAK_CURRENT_0_000_mA) |
+			PEAK_CURRENT_LANE3(PEAK_CURRENT_0_000_mA),
+	}, { /* 225/297MHz modes */
+		.pclk = UINT_MAX,
+		.pll0 = SOR_PLL_ICHPMP(1) | SOR_PLL_BG_V17_S(3) |
+			SOR_PLL_VCOCAP(0xf) | SOR_PLL_RESISTORSEL,
+		.pll1 = SOR_PLL_LOADADJ(3) | SOR_PLL_TMDS_TERMADJ(7)
+			| SOR_PLL_TMDS_TERM_ENABLE,
+		.pe_current = PE_CURRENT0(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT1(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT2(PE_CURRENT_0_mA_T114) |
+			PE_CURRENT3(PE_CURRENT_0_mA_T114),
+		.drive_current =
+			DRIVE_CURRENT_LANE0_T114(DRIVE_CURRENT_25_200_mA_T114) |
+			DRIVE_CURRENT_LANE1_T114(DRIVE_CURRENT_25_200_mA_T114) |
+			DRIVE_CURRENT_LANE2_T114(DRIVE_CURRENT_25_200_mA_T114) |
+			DRIVE_CURRENT_LANE3_T114(DRIVE_CURRENT_19_200_mA_T114),
+		.peak_current = PEAK_CURRENT_LANE0(PEAK_CURRENT_3_000_mA) |
+			PEAK_CURRENT_LANE1(PEAK_CURRENT_3_000_mA) |
+			PEAK_CURRENT_LANE2(PEAK_CURRENT_3_000_mA) |
+			PEAK_CURRENT_LANE3(PEAK_CURRENT_0_800_mA),
+	},
+};
+
+static const struct drm_encoder_funcs tegra_hdmi_encoder_funcs = {
+	.destroy = drm_encoder_cleanup,
+};
+
+static void tegra_hdmi_audio_lock(struct tegra_hdmi *hdmi)
+{
+	mutex_lock(&hdmi->audio_lock);
+	disable_irq(hdmi->irq);
+}
+
+static void tegra_hdmi_audio_unlock(struct tegra_hdmi *hdmi)
+{
+	enable_irq(hdmi->irq);
+	mutex_unlock(&hdmi->audio_lock);
+}
+
+static int
+tegra_hdmi_get_audio_config(unsigned int audio_freq, unsigned int pix_clock,
+			    struct tegra_hdmi_audio_config *config)
+{
+	const unsigned int afreq = 128 * audio_freq;
+	const unsigned int min_n = afreq / 1500;
+	const unsigned int max_n = afreq / 300;
+	const unsigned int ideal_n = afreq / 1000;
+	int64_t min_err = (uint64_t)-1 >> 1;
+	unsigned int min_delta = -1;
+	int n;
+
+	memset(config, 0, sizeof(*config));
+	config->n = -1;
+
+	for (n = min_n; n <= max_n; n++) {
+		uint64_t cts_f, aval_f;
+		unsigned int delta;
+		int64_t cts, err;
+
+		/* compute aval in 48.16 fixed point */
+		aval_f = ((int64_t)24000000 << 16) * n;
+		do_div(aval_f, afreq);
+		/* It should round without any rest */
+		if (aval_f & 0xFFFF)
+			continue;
+
+		/* Compute cts in 48.16 fixed point */
+		cts_f = ((int64_t)pix_clock << 16) * n;
+		do_div(cts_f, afreq);
+		/* Round it to the nearest integer */
+		cts = (cts_f & ~0xFFFF) + ((cts_f & BIT(15)) << 1);
+
+		delta = abs(n - ideal_n);
+
+		/* Compute the absolute error */
+		err = abs((int64_t)cts_f - cts);
+		if (err < min_err || (err == min_err && delta < min_delta)) {
+			config->n = n;
+			config->cts = cts >> 16;
+			config->aval = aval_f >> 16;
+			min_delta = delta;
+			min_err = err;
+		}
+	}
+
+	return config->n != -1 ? 0 : -EINVAL;
+}
+
+static void tegra_hdmi_setup_audio_fs_tables(struct tegra_hdmi *hdmi)
+{
+	static const unsigned int freqs[] = {
+		32000, 44100, 48000, 88200, 96000, 176400, 192000
+	};
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(freqs); i++) {
+		unsigned int f = freqs[i];
+		unsigned int eight_half;
+		unsigned int delta;
+		u32 value;
+
+		if (f > 96000)
+			delta = 2;
+		else if (f > 48000)
+			delta = 6;
+		else
+			delta = 9;
+
+		eight_half = (8 * HDMI_AUDIOCLK_FREQ) / (f * 128);
+		value = AUDIO_FS_LOW(eight_half - delta) |
+			AUDIO_FS_HIGH(eight_half + delta);
+		tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_FS(i));
+	}
+}
+
+static void tegra_hdmi_write_aval(struct tegra_hdmi *hdmi, u32 value)
+{
+	static const struct {
+		unsigned int sample_rate;
+		unsigned int offset;
+	} regs[] = {
+		{  32000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0320 },
+		{  44100, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0441 },
+		{  48000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0480 },
+		{  88200, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0882 },
+		{  96000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_0960 },
+		{ 176400, HDMI_NV_PDISP_SOR_AUDIO_AVAL_1764 },
+		{ 192000, HDMI_NV_PDISP_SOR_AUDIO_AVAL_1920 },
+	};
+	unsigned int i;
+
+	for (i = 0; i < ARRAY_SIZE(regs); i++) {
+		if (regs[i].sample_rate == hdmi->format.sample_rate) {
+			tegra_hdmi_writel(hdmi, value, regs[i].offset);
+			break;
+		}
+	}
+}
+
+static int tegra_hdmi_setup_audio(struct tegra_hdmi *hdmi)
+{
+	struct tegra_hdmi_audio_config config;
+	u32 source, value;
+	int err;
+
+	switch (hdmi->audio_source) {
+	case HDA:
+		if (hdmi->config->has_hda)
+			source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_HDAL;
+		else
+			return -EINVAL;
+
+		break;
+
+	case SPDIF:
+		if (hdmi->config->has_hda)
+			source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_SPDIF;
+		else
+			source = AUDIO_CNTRL0_SOURCE_SELECT_SPDIF;
+		break;
+
+	default:
+		if (hdmi->config->has_hda)
+			source = SOR_AUDIO_CNTRL0_SOURCE_SELECT_AUTO;
+		else
+			source = AUDIO_CNTRL0_SOURCE_SELECT_AUTO;
+		break;
+	}
+
+	/*
+	 * Tegra30 and later use a slightly modified version of the register
+	 * layout to accomodate for changes related to supporting HDA as the
+	 * audio input source for HDMI. The source select field has moved to
+	 * the SOR_AUDIO_CNTRL0 register, but the error tolerance and frames
+	 * per block fields remain in the AUDIO_CNTRL0 register.
+	 */
+	if (hdmi->config->has_hda) {
+		/*
+		 * Inject null samples into the audio FIFO for every frame in
+		 * which the codec did not receive any samples. This applies
+		 * to stereo LPCM only.
+		 *
+		 * XXX: This seems to be a remnant of MCP days when this was
+		 * used to work around issues with monitors not being able to
+		 * play back system startup sounds early. It is possibly not
+		 * needed on Linux at all.
+		 */
+		if (hdmi->format.channels == 2)
+			value = SOR_AUDIO_CNTRL0_INJECT_NULLSMPL;
+		else
+			value = 0;
+
+		value |= source;
+
+		tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_CNTRL0);
+	}
+
+	/*
+	 * On Tegra20, HDA is not a supported audio source and the source
+	 * select field is part of the AUDIO_CNTRL0 register.
+	 */
+	value = AUDIO_CNTRL0_FRAMES_PER_BLOCK(0xc0) |
+		AUDIO_CNTRL0_ERROR_TOLERANCE(6);
+
+	if (!hdmi->config->has_hda)
+		value |= source;
+
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_CNTRL0);
+
+	/*
+	 * Advertise support for High Bit-Rate on Tegra114 and later.
+	 */
+	if (hdmi->config->has_hbr) {
+		value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_AUDIO_SPARE0);
+		value |= SOR_AUDIO_SPARE0_HBR_ENABLE;
+		tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_SPARE0);
+	}
+
+	err = tegra_hdmi_get_audio_config(hdmi->format.sample_rate,
+					  hdmi->pixel_clock, &config);
+	if (err < 0) {
+		dev_err(hdmi->dev,
+			"cannot set audio to %u Hz at %u Hz pixel clock\n",
+			hdmi->format.sample_rate, hdmi->pixel_clock);
+		return err;
+	}
+
+	dev_dbg(hdmi->dev, "audio: pixclk=%u, n=%u, cts=%u, aval=%u\n",
+		hdmi->pixel_clock, config.n, config.cts, config.aval);
+
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_HDMI_ACR_CTRL);
+
+	value = AUDIO_N_RESETF | AUDIO_N_GENERATE_ALTERNATE |
+		AUDIO_N_VALUE(config.n - 1);
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_N);
+
+	tegra_hdmi_writel(hdmi, ACR_SUBPACK_N(config.n) | ACR_ENABLE,
+			  HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_HIGH);
+
+	tegra_hdmi_writel(hdmi, ACR_SUBPACK_CTS(config.cts),
+			  HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_LOW);
+
+	value = SPARE_HW_CTS | SPARE_FORCE_SW_CTS | SPARE_CTS_RESET_VAL(1);
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_SPARE);
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_AUDIO_N);
+	value &= ~AUDIO_N_RESETF;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_AUDIO_N);
+
+	if (hdmi->config->has_hda)
+		tegra_hdmi_write_aval(hdmi, config.aval);
+
+	tegra_hdmi_setup_audio_fs_tables(hdmi);
+
+	return 0;
+}
+
+static void tegra_hdmi_disable_audio(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+	value &= ~GENERIC_CTRL_AUDIO;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+}
+
+static void tegra_hdmi_enable_audio(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+	value |= GENERIC_CTRL_AUDIO;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+}
+
+static void tegra_hdmi_write_eld(struct tegra_hdmi *hdmi)
+{
+	size_t length = drm_eld_size(hdmi->output.connector.eld), i;
+	u32 value;
+
+	for (i = 0; i < length; i++)
+		tegra_hdmi_writel(hdmi, i << 8 | hdmi->output.connector.eld[i],
+				  HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR);
+
+	/*
+	 * The HDA codec will always report an ELD buffer size of 96 bytes and
+	 * the HDA codec driver will check that each byte read from the buffer
+	 * is valid. Therefore every byte must be written, even if no 96 bytes
+	 * were parsed from EDID.
+	 */
+	for (i = length; i < HDMI_ELD_BUFFER_SIZE; i++)
+		tegra_hdmi_writel(hdmi, i << 8 | 0,
+				  HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR);
+
+	value = SOR_AUDIO_HDA_PRESENSE_VALID | SOR_AUDIO_HDA_PRESENSE_PRESENT;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE);
+}
+
+static inline u32 tegra_hdmi_subpack(const u8 *ptr, size_t size)
+{
+	u32 value = 0;
+	size_t i;
+
+	for (i = size; i > 0; i--)
+		value = (value << 8) | ptr[i - 1];
+
+	return value;
+}
+
+static void tegra_hdmi_write_infopack(struct tegra_hdmi *hdmi, const void *data,
+				      size_t size)
+{
+	const u8 *ptr = data;
+	unsigned long offset;
+	size_t i;
+	u32 value;
+
+	switch (ptr[0]) {
+	case HDMI_INFOFRAME_TYPE_AVI:
+		offset = HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_HEADER;
+		break;
+
+	case HDMI_INFOFRAME_TYPE_AUDIO:
+		offset = HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_HEADER;
+		break;
+
+	case HDMI_INFOFRAME_TYPE_VENDOR:
+		offset = HDMI_NV_PDISP_HDMI_GENERIC_HEADER;
+		break;
+
+	default:
+		dev_err(hdmi->dev, "unsupported infoframe type: %02x\n",
+			ptr[0]);
+		return;
+	}
+
+	value = INFOFRAME_HEADER_TYPE(ptr[0]) |
+		INFOFRAME_HEADER_VERSION(ptr[1]) |
+		INFOFRAME_HEADER_LEN(ptr[2]);
+	tegra_hdmi_writel(hdmi, value, offset);
+	offset++;
+
+	/*
+	 * Each subpack contains 7 bytes, divided into:
+	 * - subpack_low: bytes 0 - 3
+	 * - subpack_high: bytes 4 - 6 (with byte 7 padded to 0x00)
+	 */
+	for (i = 3; i < size; i += 7) {
+		size_t rem = size - i, num = min_t(size_t, rem, 4);
+
+		value = tegra_hdmi_subpack(&ptr[i], num);
+		tegra_hdmi_writel(hdmi, value, offset++);
+
+		num = min_t(size_t, rem - num, 3);
+
+		value = tegra_hdmi_subpack(&ptr[i + 4], num);
+		tegra_hdmi_writel(hdmi, value, offset++);
+	}
+}
+
+static void tegra_hdmi_setup_avi_infoframe(struct tegra_hdmi *hdmi,
+					   struct drm_display_mode *mode)
+{
+	struct hdmi_avi_infoframe frame;
+	u8 buffer[17];
+	ssize_t err;
+
+	err = drm_hdmi_avi_infoframe_from_display_mode(&frame,
+						       &hdmi->output.connector, mode);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to setup AVI infoframe: %zd\n", err);
+		return;
+	}
+
+	err = hdmi_avi_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to pack AVI infoframe: %zd\n", err);
+		return;
+	}
+
+	tegra_hdmi_write_infopack(hdmi, buffer, err);
+}
+
+static void tegra_hdmi_disable_avi_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
+	value &= ~INFOFRAME_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
+}
+
+static void tegra_hdmi_enable_avi_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
+	value |= INFOFRAME_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL);
+}
+
+static void tegra_hdmi_setup_audio_infoframe(struct tegra_hdmi *hdmi)
+{
+	struct hdmi_audio_infoframe frame;
+	u8 buffer[14];
+	ssize_t err;
+
+	err = hdmi_audio_infoframe_init(&frame);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to setup audio infoframe: %zd\n",
+			err);
+		return;
+	}
+
+	frame.channels = hdmi->format.channels;
+
+	err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to pack audio infoframe: %zd\n",
+			err);
+		return;
+	}
+
+	/*
+	 * The audio infoframe has only one set of subpack registers, so the
+	 * infoframe needs to be truncated. One set of subpack registers can
+	 * contain 7 bytes. Including the 3 byte header only the first 10
+	 * bytes can be programmed.
+	 */
+	tegra_hdmi_write_infopack(hdmi, buffer, min_t(size_t, 10, err));
+}
+
+static void tegra_hdmi_disable_audio_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
+	value &= ~INFOFRAME_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
+}
+
+static void tegra_hdmi_enable_audio_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
+	value |= INFOFRAME_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL);
+}
+
+static void tegra_hdmi_setup_stereo_infoframe(struct tegra_hdmi *hdmi)
+{
+	struct hdmi_vendor_infoframe frame;
+	u8 buffer[10];
+	ssize_t err;
+
+	hdmi_vendor_infoframe_init(&frame);
+	frame.s3d_struct = HDMI_3D_STRUCTURE_FRAME_PACKING;
+
+	err = hdmi_vendor_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to pack vendor infoframe: %zd\n",
+			err);
+		return;
+	}
+
+	tegra_hdmi_write_infopack(hdmi, buffer, err);
+}
+
+static void tegra_hdmi_disable_stereo_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+	value &= ~GENERIC_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+}
+
+static void tegra_hdmi_enable_stereo_infoframe(struct tegra_hdmi *hdmi)
+{
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+	value |= GENERIC_CTRL_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_GENERIC_CTRL);
+}
+
+static void tegra_hdmi_setup_tmds(struct tegra_hdmi *hdmi,
+				  const struct tmds_config *tmds)
+{
+	u32 value;
+
+	tegra_hdmi_writel(hdmi, tmds->pll0, HDMI_NV_PDISP_SOR_PLL0);
+	tegra_hdmi_writel(hdmi, tmds->pll1, HDMI_NV_PDISP_SOR_PLL1);
+	tegra_hdmi_writel(hdmi, tmds->pe_current, HDMI_NV_PDISP_PE_CURRENT);
+
+	tegra_hdmi_writel(hdmi, tmds->drive_current,
+			  HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT);
+
+	value = tegra_hdmi_readl(hdmi, hdmi->config->fuse_override_offset);
+	value |= hdmi->config->fuse_override_value;
+	tegra_hdmi_writel(hdmi, value, hdmi->config->fuse_override_offset);
+
+	if (hdmi->config->has_sor_io_peak_current)
+		tegra_hdmi_writel(hdmi, tmds->peak_current,
+				  HDMI_NV_PDISP_SOR_IO_PEAK_CURRENT);
+}
+
+static int tegra_hdmi_reconfigure_audio(struct tegra_hdmi *hdmi)
+{
+	int err;
+
+	err = tegra_hdmi_setup_audio(hdmi);
+	if (err < 0) {
+		tegra_hdmi_disable_audio_infoframe(hdmi);
+		tegra_hdmi_disable_audio(hdmi);
+	} else {
+		tegra_hdmi_setup_audio_infoframe(hdmi);
+		tegra_hdmi_enable_audio_infoframe(hdmi);
+		tegra_hdmi_enable_audio(hdmi);
+	}
+
+	return err;
+}
+
+static bool tegra_output_is_hdmi(struct tegra_output *output)
+{
+	return output->connector.display_info.is_hdmi;
+}
+
+static enum drm_connector_status
+tegra_hdmi_connector_detect(struct drm_connector *connector, bool force)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	struct tegra_hdmi *hdmi = to_hdmi(output);
+	enum drm_connector_status status;
+
+	status = tegra_output_connector_detect(connector, force);
+	if (status == connector_status_connected)
+		return status;
+
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE);
+	return status;
+}
+
+#define DEBUGFS_REG32(_name) { .name = #_name, .offset = _name }
+
+static const struct debugfs_reg32 tegra_hdmi_regs[] = {
+	DEBUGFS_REG32(HDMI_CTXSW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_STATE0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_STATE1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_STATE2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_AN_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_AN_LSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_CN_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_CN_LSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_AKSV_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_AKSV_LSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_BKSV_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_BKSV_LSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_CKSV_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_CKSV_LSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_DKSV_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_DKSV_LSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_CMODE),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_MPRIME_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_MPRIME_LSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_SPRIME_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_SPRIME_LSB2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_SPRIME_LSB1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_RI),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_CS_MSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_RG_HDCP_CS_LSB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_EMU0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_EMU_RDATA0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_EMU1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_EMU2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_STATUS),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_HEADER),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_SUBPACK0_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_SUBPACK0_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_STATUS),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_HEADER),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_SUBPACK0_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_SUBPACK0_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_SUBPACK1_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_SUBPACK1_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_STATUS),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_HEADER),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK0_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK0_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK1_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK1_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK2_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK2_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK3_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK3_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0320_SUBPACK_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0320_SUBPACK_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0882_SUBPACK_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0882_SUBPACK_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_1764_SUBPACK_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_1764_SUBPACK_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0480_SUBPACK_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0480_SUBPACK_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0960_SUBPACK_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_0960_SUBPACK_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_1920_SUBPACK_LOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_ACR_1920_SUBPACK_HIGH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_VSYNC_KEEPOUT),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_VSYNC_WINDOW),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GCP_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GCP_STATUS),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_GCP_SUBPACK),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_CHANNEL_STATUS1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_CHANNEL_STATUS2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_EMU0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_EMU1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_EMU1_RDATA),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_SPARE),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_SPDIF_CHN_STATUS1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_SPDIF_CHN_STATUS2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDMI_HDCPRIF_ROM_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_CAP),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_PWR),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_TEST),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_PLL0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_PLL1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_PLL2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_CSTM),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_LVDS),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_CRCA),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_CRCB),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_BLANK),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_CTL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(0)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(1)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(2)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(3)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(4)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(5)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(6)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(7)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(8)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(9)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(10)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(11)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(12)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(13)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(14)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_SEQ_INST(15)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_VCRCA0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_VCRCA1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_CCRCA0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_CCRCA1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_EDATAA0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_EDATAA1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_COUNTA0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_COUNTA1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_DEBUGA0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_DEBUGA1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_TRIG),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_MSCHECK),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_DEBUG0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_DEBUG1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_DEBUG2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_FS(0)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_FS(1)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_FS(2)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_FS(3)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_FS(4)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_FS(5)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_FS(6)),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_PULSE_WIDTH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_THRESHOLD),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_CNTRL0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_AUDIO_N),
+	DEBUGFS_REG32(HDMI_NV_PDISP_HDCPRIF_ROM_TIMING),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_REFCLK),
+	DEBUGFS_REG32(HDMI_NV_PDISP_CRC_CONTROL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_INPUT_CONTROL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SCRATCH),
+	DEBUGFS_REG32(HDMI_NV_PDISP_PE_CURRENT),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_CTRL),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_DEBUG0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_DEBUG1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_DEBUG2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_HDCP_KEY_0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_HDCP_KEY_1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_HDCP_KEY_2),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_HDCP_KEY_3),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_HDCP_KEY_TRIG),
+	DEBUGFS_REG32(HDMI_NV_PDISP_KEY_SKEY_INDEX),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_AUDIO_CNTRL0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_AUDIO_SPARE0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH1),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE),
+	DEBUGFS_REG32(HDMI_NV_PDISP_INT_STATUS),
+	DEBUGFS_REG32(HDMI_NV_PDISP_INT_MASK),
+	DEBUGFS_REG32(HDMI_NV_PDISP_INT_ENABLE),
+	DEBUGFS_REG32(HDMI_NV_PDISP_SOR_IO_PEAK_CURRENT),
+};
+
+static int tegra_hdmi_show_regs(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = s->private;
+	struct tegra_hdmi *hdmi = node->info_ent->data;
+	struct drm_crtc *crtc = hdmi->output.encoder.crtc;
+	struct drm_device *drm = node->minor->dev;
+	unsigned int i;
+	int err = 0;
+
+	drm_modeset_lock_all(drm);
+
+	if (!crtc || !crtc->state->active) {
+		err = -EBUSY;
+		goto unlock;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(tegra_hdmi_regs); i++) {
+		unsigned int offset = tegra_hdmi_regs[i].offset;
+
+		seq_printf(s, "%-56s %#05x %08x\n", tegra_hdmi_regs[i].name,
+			   offset, tegra_hdmi_readl(hdmi, offset));
+	}
+
+unlock:
+	drm_modeset_unlock_all(drm);
+	return err;
+}
+
+static struct drm_info_list debugfs_files[] = {
+	{ "regs", tegra_hdmi_show_regs, 0, NULL },
+};
+
+static int tegra_hdmi_late_register(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	unsigned int i, count = ARRAY_SIZE(debugfs_files);
+	struct drm_minor *minor = connector->dev->primary;
+	struct dentry *root = connector->debugfs_entry;
+	struct tegra_hdmi *hdmi = to_hdmi(output);
+
+	hdmi->debugfs_files = kmemdup(debugfs_files, sizeof(debugfs_files),
+				      GFP_KERNEL);
+	if (!hdmi->debugfs_files)
+		return -ENOMEM;
+
+	for (i = 0; i < count; i++)
+		hdmi->debugfs_files[i].data = hdmi;
+
+	drm_debugfs_create_files(hdmi->debugfs_files, count, root, minor);
+
+	return 0;
+}
+
+static void tegra_hdmi_early_unregister(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	struct drm_minor *minor = connector->dev->primary;
+	unsigned int count = ARRAY_SIZE(debugfs_files);
+	struct tegra_hdmi *hdmi = to_hdmi(output);
+
+	drm_debugfs_remove_files(hdmi->debugfs_files, count,
+				 connector->debugfs_entry, minor);
+	kfree(hdmi->debugfs_files);
+	hdmi->debugfs_files = NULL;
+}
+
+static const struct drm_connector_funcs tegra_hdmi_connector_funcs = {
+	.reset = drm_atomic_helper_connector_reset,
+	.detect = tegra_hdmi_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = tegra_output_connector_destroy,
+	.atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.late_register = tegra_hdmi_late_register,
+	.early_unregister = tegra_hdmi_early_unregister,
+};
+
+static enum drm_mode_status
+tegra_hdmi_connector_mode_valid(struct drm_connector *connector,
+				const struct drm_display_mode *mode)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	struct tegra_hdmi *hdmi = to_hdmi(output);
+	unsigned long pclk = mode->clock * 1000;
+	enum drm_mode_status status = MODE_OK;
+	struct clk *parent;
+	long err;
+
+	parent = clk_get_parent(hdmi->clk_parent);
+
+	err = clk_round_rate(parent, pclk * 4);
+	if (err <= 0)
+		status = MODE_NOCLOCK;
+
+	return status;
+}
+
+static const struct drm_connector_helper_funcs
+tegra_hdmi_connector_helper_funcs = {
+	.get_modes = tegra_output_connector_get_modes,
+	.mode_valid = tegra_hdmi_connector_mode_valid,
+};
+
+static void tegra_hdmi_encoder_disable(struct drm_encoder *encoder)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_hdmi *hdmi = to_hdmi(output);
+	u32 value;
+	int err;
+
+	tegra_hdmi_audio_lock(hdmi);
+
+	/*
+	 * The following accesses registers of the display controller, so make
+	 * sure it's only executed when the output is attached to one.
+	 */
+	if (dc) {
+		value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+		value &= ~HDMI_ENABLE;
+		tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+		tegra_dc_commit(dc);
+	}
+
+	if (!hdmi->dvi) {
+		if (hdmi->stereo)
+			tegra_hdmi_disable_stereo_infoframe(hdmi);
+
+		tegra_hdmi_disable_audio_infoframe(hdmi);
+		tegra_hdmi_disable_avi_infoframe(hdmi);
+		tegra_hdmi_disable_audio(hdmi);
+	}
+
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_INT_ENABLE);
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_INT_MASK);
+
+	hdmi->pixel_clock = 0;
+
+	tegra_hdmi_audio_unlock(hdmi);
+
+	err = host1x_client_suspend(&hdmi->client);
+	if (err < 0)
+		dev_err(hdmi->dev, "failed to suspend: %d\n", err);
+}
+
+static void tegra_hdmi_encoder_enable(struct drm_encoder *encoder)
+{
+	struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
+	unsigned int h_sync_width, h_front_porch, h_back_porch, i, rekey;
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_hdmi *hdmi = to_hdmi(output);
+	unsigned int pulse_start, div82;
+	int retries = 1000;
+	u32 value;
+	int err;
+
+	err = host1x_client_resume(&hdmi->client);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to resume: %d\n", err);
+		return;
+	}
+
+	tegra_hdmi_audio_lock(hdmi);
+
+	/*
+	 * Enable and unmask the HDA codec SCRATCH0 register interrupt. This
+	 * is used for interoperability between the HDA codec driver and the
+	 * HDMI driver.
+	 */
+	tegra_hdmi_writel(hdmi, INT_CODEC_SCRATCH0, HDMI_NV_PDISP_INT_ENABLE);
+	tegra_hdmi_writel(hdmi, INT_CODEC_SCRATCH0, HDMI_NV_PDISP_INT_MASK);
+
+	hdmi->pixel_clock = mode->clock * 1000;
+	h_sync_width = mode->hsync_end - mode->hsync_start;
+	h_back_porch = mode->htotal - mode->hsync_end;
+	h_front_porch = mode->hsync_start - mode->hdisplay;
+
+	err = dev_pm_opp_set_rate(hdmi->dev, hdmi->pixel_clock);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to set HDMI clock frequency: %d\n",
+			err);
+	}
+
+	DRM_DEBUG_KMS("HDMI clock rate: %lu Hz\n", clk_get_rate(hdmi->clk));
+
+	/* power up sequence */
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_PLL0);
+	value &= ~SOR_PLL_PDBG;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_PLL0);
+
+	usleep_range(10, 20);
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_PLL0);
+	value &= ~SOR_PLL_PWR;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_PLL0);
+
+	tegra_dc_writel(dc, VSYNC_H_POSITION(1),
+			DC_DISP_DISP_TIMING_OPTIONS);
+	tegra_dc_writel(dc, DITHER_CONTROL_DISABLE | BASE_COLOR_SIZE_888,
+			DC_DISP_DISP_COLOR_CONTROL);
+
+	/* video_preamble uses h_pulse2 */
+	pulse_start = 1 + h_sync_width + h_back_porch - 10;
+
+	tegra_dc_writel(dc, H_PULSE2_ENABLE, DC_DISP_DISP_SIGNAL_OPTIONS0);
+
+	value = PULSE_MODE_NORMAL | PULSE_POLARITY_HIGH | PULSE_QUAL_VACTIVE |
+		PULSE_LAST_END_A;
+	tegra_dc_writel(dc, value, DC_DISP_H_PULSE2_CONTROL);
+
+	value = PULSE_START(pulse_start) | PULSE_END(pulse_start + 8);
+	tegra_dc_writel(dc, value, DC_DISP_H_PULSE2_POSITION_A);
+
+	value = VSYNC_WINDOW_END(0x210) | VSYNC_WINDOW_START(0x200) |
+		VSYNC_WINDOW_ENABLE;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_VSYNC_WINDOW);
+
+	if (dc->pipe)
+		value = HDMI_SRC_DISPLAYB;
+	else
+		value = HDMI_SRC_DISPLAYA;
+
+	if ((mode->hdisplay == 720) && ((mode->vdisplay == 480) ||
+					(mode->vdisplay == 576)))
+		tegra_hdmi_writel(hdmi,
+				  value | ARM_VIDEO_RANGE_FULL,
+				  HDMI_NV_PDISP_INPUT_CONTROL);
+	else
+		tegra_hdmi_writel(hdmi,
+				  value | ARM_VIDEO_RANGE_LIMITED,
+				  HDMI_NV_PDISP_INPUT_CONTROL);
+
+	div82 = clk_get_rate(hdmi->clk) / 1000000 * 4;
+	value = SOR_REFCLK_DIV_INT(div82 >> 2) | SOR_REFCLK_DIV_FRAC(div82);
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_REFCLK);
+
+	hdmi->dvi = !tegra_output_is_hdmi(output);
+	if (!hdmi->dvi) {
+		/*
+		 * Make sure that the audio format has been configured before
+		 * enabling audio, otherwise we may try to divide by zero.
+		*/
+		if (hdmi->format.sample_rate > 0) {
+			err = tegra_hdmi_setup_audio(hdmi);
+			if (err < 0)
+				hdmi->dvi = true;
+		}
+	}
+
+	if (hdmi->config->has_hda)
+		tegra_hdmi_write_eld(hdmi);
+
+	rekey = HDMI_REKEY_DEFAULT;
+	value = HDMI_CTRL_REKEY(rekey);
+	value |= HDMI_CTRL_MAX_AC_PACKET((h_sync_width + h_back_porch +
+					  h_front_porch - rekey - 18) / 32);
+
+	if (!hdmi->dvi)
+		value |= HDMI_CTRL_ENABLE;
+
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_HDMI_CTRL);
+
+	if (!hdmi->dvi) {
+		tegra_hdmi_setup_avi_infoframe(hdmi, mode);
+		tegra_hdmi_setup_audio_infoframe(hdmi);
+
+		if (hdmi->stereo)
+			tegra_hdmi_setup_stereo_infoframe(hdmi);
+	}
+
+	/* TMDS CONFIG */
+	for (i = 0; i < hdmi->config->num_tmds; i++) {
+		if (hdmi->pixel_clock <= hdmi->config->tmds[i].pclk) {
+			tegra_hdmi_setup_tmds(hdmi, &hdmi->config->tmds[i]);
+			break;
+		}
+	}
+
+	tegra_hdmi_writel(hdmi,
+			  SOR_SEQ_PU_PC(0) |
+			  SOR_SEQ_PU_PC_ALT(0) |
+			  SOR_SEQ_PD_PC(8) |
+			  SOR_SEQ_PD_PC_ALT(8),
+			  HDMI_NV_PDISP_SOR_SEQ_CTL);
+
+	value = SOR_SEQ_INST_WAIT_TIME(1) |
+		SOR_SEQ_INST_WAIT_UNITS_VSYNC |
+		SOR_SEQ_INST_HALT |
+		SOR_SEQ_INST_PIN_A_LOW |
+		SOR_SEQ_INST_PIN_B_LOW |
+		SOR_SEQ_INST_DRIVE_PWM_OUT_LO;
+
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_SEQ_INST(0));
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_SEQ_INST(8));
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_CSTM);
+	value &= ~SOR_CSTM_ROTCLK(~0);
+	value |= SOR_CSTM_ROTCLK(2);
+	value |= SOR_CSTM_PLLDIV;
+	value &= ~SOR_CSTM_LVDS_ENABLE;
+	value &= ~SOR_CSTM_MODE_MASK;
+	value |= SOR_CSTM_MODE_TMDS;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_CSTM);
+
+	/* start SOR */
+	tegra_hdmi_writel(hdmi,
+			  SOR_PWR_NORMAL_STATE_PU |
+			  SOR_PWR_NORMAL_START_NORMAL |
+			  SOR_PWR_SAFE_STATE_PD |
+			  SOR_PWR_SETTING_NEW_TRIGGER,
+			  HDMI_NV_PDISP_SOR_PWR);
+	tegra_hdmi_writel(hdmi,
+			  SOR_PWR_NORMAL_STATE_PU |
+			  SOR_PWR_NORMAL_START_NORMAL |
+			  SOR_PWR_SAFE_STATE_PD |
+			  SOR_PWR_SETTING_NEW_DONE,
+			  HDMI_NV_PDISP_SOR_PWR);
+
+	do {
+		BUG_ON(--retries < 0);
+		value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_PWR);
+	} while (value & SOR_PWR_SETTING_NEW_PENDING);
+
+	value = SOR_STATE_ASY_CRCMODE_COMPLETE |
+		SOR_STATE_ASY_OWNER_HEAD0 |
+		SOR_STATE_ASY_SUBOWNER_BOTH |
+		SOR_STATE_ASY_PROTOCOL_SINGLE_TMDS_A |
+		SOR_STATE_ASY_DEPOL_POS;
+
+	/* setup sync polarities */
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		value |= SOR_STATE_ASY_HSYNCPOL_POS;
+
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		value |= SOR_STATE_ASY_HSYNCPOL_NEG;
+
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		value |= SOR_STATE_ASY_VSYNCPOL_POS;
+
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		value |= SOR_STATE_ASY_VSYNCPOL_NEG;
+
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_STATE2);
+
+	value = SOR_STATE_ASY_HEAD_OPMODE_AWAKE | SOR_STATE_ASY_ORMODE_NORMAL;
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_SOR_STATE1);
+
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_SOR_STATE0);
+	tegra_hdmi_writel(hdmi, SOR_STATE_UPDATE, HDMI_NV_PDISP_SOR_STATE0);
+	tegra_hdmi_writel(hdmi, value | SOR_STATE_ATTACHED,
+			  HDMI_NV_PDISP_SOR_STATE1);
+	tegra_hdmi_writel(hdmi, 0, HDMI_NV_PDISP_SOR_STATE0);
+
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+	value |= HDMI_ENABLE;
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+	tegra_dc_commit(dc);
+
+	if (!hdmi->dvi) {
+		tegra_hdmi_enable_avi_infoframe(hdmi);
+		tegra_hdmi_enable_audio_infoframe(hdmi);
+		tegra_hdmi_enable_audio(hdmi);
+
+		if (hdmi->stereo)
+			tegra_hdmi_enable_stereo_infoframe(hdmi);
+	}
+
+	/* TODO: add HDCP support */
+
+	tegra_hdmi_audio_unlock(hdmi);
+}
+
+static int
+tegra_hdmi_encoder_atomic_check(struct drm_encoder *encoder,
+				struct drm_crtc_state *crtc_state,
+				struct drm_connector_state *conn_state)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(conn_state->crtc);
+	unsigned long pclk = crtc_state->mode.clock * 1000;
+	struct tegra_hdmi *hdmi = to_hdmi(output);
+	int err;
+
+	err = tegra_dc_state_setup_clock(dc, crtc_state, hdmi->clk_parent,
+					 pclk, 0);
+	if (err < 0) {
+		dev_err(output->dev, "failed to setup CRTC state: %d\n", err);
+		return err;
+	}
+
+	return err;
+}
+
+static const struct drm_encoder_helper_funcs tegra_hdmi_encoder_helper_funcs = {
+	.disable = tegra_hdmi_encoder_disable,
+	.enable = tegra_hdmi_encoder_enable,
+	.atomic_check = tegra_hdmi_encoder_atomic_check,
+};
+
+static int tegra_hdmi_hw_params(struct device *dev, void *data,
+				struct hdmi_codec_daifmt *fmt,
+				struct hdmi_codec_params *hparms)
+{
+	struct tegra_hdmi *hdmi = data;
+	int ret = 0;
+
+	tegra_hdmi_audio_lock(hdmi);
+
+	hdmi->format.sample_rate = hparms->sample_rate;
+	hdmi->format.channels = hparms->channels;
+
+	if (hdmi->pixel_clock && !hdmi->dvi)
+		ret = tegra_hdmi_reconfigure_audio(hdmi);
+
+	tegra_hdmi_audio_unlock(hdmi);
+
+	return ret;
+}
+
+static int tegra_hdmi_audio_startup(struct device *dev, void *data)
+{
+	struct tegra_hdmi *hdmi = data;
+	int ret;
+
+	ret = host1x_client_resume(&hdmi->client);
+	if (ret < 0)
+		dev_err(hdmi->dev, "failed to resume: %d\n", ret);
+
+	return ret;
+}
+
+static void tegra_hdmi_audio_shutdown(struct device *dev, void *data)
+{
+	struct tegra_hdmi *hdmi = data;
+	int ret;
+
+	tegra_hdmi_audio_lock(hdmi);
+
+	hdmi->format.sample_rate = 0;
+	hdmi->format.channels = 0;
+
+	tegra_hdmi_audio_unlock(hdmi);
+
+	ret = host1x_client_suspend(&hdmi->client);
+	if (ret < 0)
+		dev_err(hdmi->dev, "failed to suspend: %d\n", ret);
+}
+
+static const struct hdmi_codec_ops tegra_hdmi_codec_ops = {
+	.hw_params = tegra_hdmi_hw_params,
+	.audio_startup = tegra_hdmi_audio_startup,
+	.audio_shutdown = tegra_hdmi_audio_shutdown,
+};
+
+static int tegra_hdmi_codec_register(struct tegra_hdmi *hdmi)
+{
+	struct hdmi_codec_pdata codec_data = {};
+
+	if (hdmi->config->has_hda)
+		return 0;
+
+	codec_data.ops = &tegra_hdmi_codec_ops;
+	codec_data.data = hdmi;
+	codec_data.spdif = 1;
+
+	hdmi->audio_pdev = platform_device_register_data(hdmi->dev,
+							 HDMI_CODEC_DRV_NAME,
+							 PLATFORM_DEVID_AUTO,
+							 &codec_data,
+							 sizeof(codec_data));
+	if (IS_ERR(hdmi->audio_pdev))
+		return PTR_ERR(hdmi->audio_pdev);
+
+	hdmi->format.channels = 2;
+
+	return 0;
+}
+
+static void tegra_hdmi_codec_unregister(struct tegra_hdmi *hdmi)
+{
+	if (hdmi->audio_pdev)
+		platform_device_unregister(hdmi->audio_pdev);
+}
+
+static int tegra_hdmi_init(struct host1x_client *client)
+{
+	struct tegra_hdmi *hdmi = host1x_client_to_hdmi(client);
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	struct drm_connector *connector;
+	int err;
+
+	hdmi->output.dev = client->dev;
+
+	drm_encoder_init(drm, &hdmi->output.encoder, &tegra_hdmi_encoder_funcs,
+			 DRM_MODE_ENCODER_TMDS, NULL);
+	drm_encoder_helper_add(&hdmi->output.encoder,
+			       &tegra_hdmi_encoder_helper_funcs);
+
+	if (hdmi->output.bridge) {
+		err = drm_bridge_attach(&hdmi->output.encoder, hdmi->output.bridge,
+					NULL, DRM_BRIDGE_ATTACH_NO_CONNECTOR);
+		if (err) {
+			dev_err(client->dev, "failed to attach bridge: %d\n",
+				err);
+			return err;
+		}
+
+		connector = drm_bridge_connector_init(drm, &hdmi->output.encoder);
+		if (IS_ERR(connector)) {
+			dev_err(client->dev,
+				"failed to initialize bridge connector: %pe\n",
+				connector);
+			return PTR_ERR(connector);
+		}
+	} else {
+		drm_connector_init_with_ddc(drm, &hdmi->output.connector,
+					    &tegra_hdmi_connector_funcs,
+					    DRM_MODE_CONNECTOR_HDMIA,
+					    hdmi->output.ddc);
+		drm_connector_helper_add(&hdmi->output.connector,
+					 &tegra_hdmi_connector_helper_funcs);
+		hdmi->output.connector.dpms = DRM_MODE_DPMS_OFF;
+
+		drm_connector_attach_encoder(&hdmi->output.connector,
+					     &hdmi->output.encoder);
+		drm_connector_register(&hdmi->output.connector);
+	}
+
+	err = tegra_output_init(drm, &hdmi->output);
+	if (err < 0) {
+		dev_err(client->dev, "failed to initialize output: %d\n", err);
+		return err;
+	}
+
+	hdmi->output.encoder.possible_crtcs = 0x3;
+
+	err = regulator_enable(hdmi->hdmi);
+	if (err < 0) {
+		dev_err(client->dev, "failed to enable HDMI regulator: %d\n",
+			err);
+		goto output_exit;
+	}
+
+	err = regulator_enable(hdmi->pll);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to enable PLL regulator: %d\n", err);
+		goto disable_hdmi;
+	}
+
+	err = regulator_enable(hdmi->vdd);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to enable VDD regulator: %d\n", err);
+		goto disable_pll;
+	}
+
+	err = tegra_hdmi_codec_register(hdmi);
+	if (err < 0) {
+		dev_err(hdmi->dev, "failed to register audio codec: %d\n", err);
+		goto disable_vdd;
+	}
+
+	return 0;
+
+disable_vdd:
+	regulator_disable(hdmi->vdd);
+disable_pll:
+	regulator_disable(hdmi->pll);
+disable_hdmi:
+	regulator_disable(hdmi->hdmi);
+output_exit:
+	tegra_output_exit(&hdmi->output);
+
+	return err;
+}
+
+static int tegra_hdmi_exit(struct host1x_client *client)
+{
+	struct tegra_hdmi *hdmi = host1x_client_to_hdmi(client);
+
+	tegra_hdmi_codec_unregister(hdmi);
+
+	tegra_output_exit(&hdmi->output);
+
+	regulator_disable(hdmi->vdd);
+	regulator_disable(hdmi->pll);
+	regulator_disable(hdmi->hdmi);
+
+	return 0;
+}
+
+static int tegra_hdmi_runtime_suspend(struct host1x_client *client)
+{
+	struct tegra_hdmi *hdmi = host1x_client_to_hdmi(client);
+	struct device *dev = client->dev;
+	int err;
+
+	err = reset_control_assert(hdmi->rst);
+	if (err < 0) {
+		dev_err(dev, "failed to assert reset: %d\n", err);
+		return err;
+	}
+
+	usleep_range(1000, 2000);
+
+	clk_disable_unprepare(hdmi->clk);
+	pm_runtime_put_sync(dev);
+
+	return 0;
+}
+
+static int tegra_hdmi_runtime_resume(struct host1x_client *client)
+{
+	struct tegra_hdmi *hdmi = host1x_client_to_hdmi(client);
+	struct device *dev = client->dev;
+	int err;
+
+	err = pm_runtime_resume_and_get(dev);
+	if (err < 0) {
+		dev_err(dev, "failed to get runtime PM: %d\n", err);
+		return err;
+	}
+
+	err = clk_prepare_enable(hdmi->clk);
+	if (err < 0) {
+		dev_err(dev, "failed to enable clock: %d\n", err);
+		goto put_rpm;
+	}
+
+	usleep_range(1000, 2000);
+
+	err = reset_control_deassert(hdmi->rst);
+	if (err < 0) {
+		dev_err(dev, "failed to deassert reset: %d\n", err);
+		goto disable_clk;
+	}
+
+	return 0;
+
+disable_clk:
+	clk_disable_unprepare(hdmi->clk);
+put_rpm:
+	pm_runtime_put_sync(dev);
+	return err;
+}
+
+static const struct host1x_client_ops hdmi_client_ops = {
+	.init = tegra_hdmi_init,
+	.exit = tegra_hdmi_exit,
+	.suspend = tegra_hdmi_runtime_suspend,
+	.resume = tegra_hdmi_runtime_resume,
+};
+
+static const struct tegra_hdmi_config tegra20_hdmi_config = {
+	.tmds = tegra20_tmds_config,
+	.num_tmds = ARRAY_SIZE(tegra20_tmds_config),
+	.fuse_override_offset = HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT,
+	.fuse_override_value = 1 << 31,
+	.has_sor_io_peak_current = false,
+	.has_hda = false,
+	.has_hbr = false,
+};
+
+static const struct tegra_hdmi_config tegra30_hdmi_config = {
+	.tmds = tegra30_tmds_config,
+	.num_tmds = ARRAY_SIZE(tegra30_tmds_config),
+	.fuse_override_offset = HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT,
+	.fuse_override_value = 1 << 31,
+	.has_sor_io_peak_current = false,
+	.has_hda = true,
+	.has_hbr = false,
+};
+
+static const struct tegra_hdmi_config tegra114_hdmi_config = {
+	.tmds = tegra114_tmds_config,
+	.num_tmds = ARRAY_SIZE(tegra114_tmds_config),
+	.fuse_override_offset = HDMI_NV_PDISP_SOR_PAD_CTLS0,
+	.fuse_override_value = 1 << 31,
+	.has_sor_io_peak_current = true,
+	.has_hda = true,
+	.has_hbr = true,
+};
+
+static const struct tegra_hdmi_config tegra124_hdmi_config = {
+	.tmds = tegra124_tmds_config,
+	.num_tmds = ARRAY_SIZE(tegra124_tmds_config),
+	.fuse_override_offset = HDMI_NV_PDISP_SOR_PAD_CTLS0,
+	.fuse_override_value = 1 << 31,
+	.has_sor_io_peak_current = true,
+	.has_hda = true,
+	.has_hbr = true,
+};
+
+static const struct of_device_id tegra_hdmi_of_match[] = {
+	{ .compatible = "nvidia,tegra124-hdmi", .data = &tegra124_hdmi_config },
+	{ .compatible = "nvidia,tegra114-hdmi", .data = &tegra114_hdmi_config },
+	{ .compatible = "nvidia,tegra30-hdmi", .data = &tegra30_hdmi_config },
+	{ .compatible = "nvidia,tegra20-hdmi", .data = &tegra20_hdmi_config },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, tegra_hdmi_of_match);
+
+static irqreturn_t tegra_hdmi_irq(int irq, void *data)
+{
+	struct tegra_hdmi *hdmi = data;
+	u32 value;
+
+	value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_INT_STATUS);
+	tegra_hdmi_writel(hdmi, value, HDMI_NV_PDISP_INT_STATUS);
+
+	if (value & INT_CODEC_SCRATCH0) {
+		unsigned int format;
+		u32 value;
+
+		value = tegra_hdmi_readl(hdmi, HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0);
+
+		if (value & SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID) {
+			format = value & SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK;
+
+			tegra_hda_parse_format(format, &hdmi->format);
+			tegra_hdmi_reconfigure_audio(hdmi);
+		} else {
+			tegra_hdmi_disable_audio_infoframe(hdmi);
+			tegra_hdmi_disable_audio(hdmi);
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int tegra_hdmi_probe(struct platform_device *pdev)
+{
+	struct tegra_hdmi *hdmi;
+	int err;
+
+	hdmi = devm_kzalloc(&pdev->dev, sizeof(*hdmi), GFP_KERNEL);
+	if (!hdmi)
+		return -ENOMEM;
+
+	hdmi->config = of_device_get_match_data(&pdev->dev);
+	hdmi->dev = &pdev->dev;
+
+	hdmi->audio_source = AUTO;
+	hdmi->stereo = false;
+	hdmi->dvi = false;
+
+	mutex_init(&hdmi->audio_lock);
+
+	hdmi->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(hdmi->clk)) {
+		dev_err(&pdev->dev, "failed to get clock\n");
+		return PTR_ERR(hdmi->clk);
+	}
+
+	hdmi->rst = devm_reset_control_get(&pdev->dev, "hdmi");
+	if (IS_ERR(hdmi->rst)) {
+		dev_err(&pdev->dev, "failed to get reset\n");
+		return PTR_ERR(hdmi->rst);
+	}
+
+	hdmi->clk_parent = devm_clk_get(&pdev->dev, "parent");
+	if (IS_ERR(hdmi->clk_parent))
+		return PTR_ERR(hdmi->clk_parent);
+
+	err = clk_set_parent(hdmi->clk, hdmi->clk_parent);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to setup clocks: %d\n", err);
+		return err;
+	}
+
+	hdmi->hdmi = devm_regulator_get(&pdev->dev, "hdmi");
+	err = PTR_ERR_OR_ZERO(hdmi->hdmi);
+	if (err)
+		return dev_err_probe(&pdev->dev, err,
+				     "failed to get HDMI regulator\n");
+
+	hdmi->pll = devm_regulator_get(&pdev->dev, "pll");
+	err = PTR_ERR_OR_ZERO(hdmi->pll);
+	if (err)
+		return dev_err_probe(&pdev->dev, err,
+				     "failed to get PLL regulator\n");
+
+	hdmi->vdd = devm_regulator_get(&pdev->dev, "vdd");
+	err = PTR_ERR_OR_ZERO(hdmi->vdd);
+	if (err)
+		return dev_err_probe(&pdev->dev, err,
+				     "failed to get VDD regulator\n");
+
+	hdmi->output.dev = &pdev->dev;
+
+	err = tegra_output_probe(&hdmi->output);
+	if (err < 0)
+		return err;
+
+	hdmi->regs = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(hdmi->regs)) {
+		err = PTR_ERR(hdmi->regs);
+		goto remove;
+	}
+
+	err = platform_get_irq(pdev, 0);
+	if (err < 0)
+		goto remove;
+
+	hdmi->irq = err;
+
+	err = devm_request_irq(hdmi->dev, hdmi->irq, tegra_hdmi_irq, 0,
+			       dev_name(hdmi->dev), hdmi);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to request IRQ#%u: %d\n",
+			hdmi->irq, err);
+		goto remove;
+	}
+
+	platform_set_drvdata(pdev, hdmi);
+
+	err = devm_pm_runtime_enable(&pdev->dev);
+	if (err)
+		goto remove;
+
+	err = devm_tegra_core_dev_init_opp_table_common(&pdev->dev);
+	if (err)
+		goto remove;
+
+	INIT_LIST_HEAD(&hdmi->client.list);
+	hdmi->client.ops = &hdmi_client_ops;
+	hdmi->client.dev = &pdev->dev;
+
+	err = host1x_client_register(&hdmi->client);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to register host1x client: %d\n",
+			err);
+		goto remove;
+	}
+
+	return 0;
+
+remove:
+	tegra_output_remove(&hdmi->output);
+	return err;
+}
+
+static void tegra_hdmi_remove(struct platform_device *pdev)
+{
+	struct tegra_hdmi *hdmi = platform_get_drvdata(pdev);
+
+	host1x_client_unregister(&hdmi->client);
+
+	tegra_output_remove(&hdmi->output);
+}
+
+struct platform_driver tegra_hdmi_driver = {
+	.driver = {
+		.name = "tegra-hdmi",
+		.of_match_table = tegra_hdmi_of_match,
+	},
+	.probe = tegra_hdmi_probe,
+	.remove = tegra_hdmi_remove,
+};
diff --git a/drivers/gpu/drm/tegra/hdmi.h b/drivers/gpu/drm/tegra/hdmi.h
new file mode 100644
index 000000000..8deb04223
--- /dev/null
+++ b/drivers/gpu/drm/tegra/hdmi.h
@@ -0,0 +1,557 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#ifndef TEGRA_HDMI_H
+#define TEGRA_HDMI_H 1
+
+/* register definitions */
+#define HDMI_CTXSW						0x00
+
+#define HDMI_NV_PDISP_SOR_STATE0				0x01
+#define SOR_STATE_UPDATE (1 << 0)
+
+#define HDMI_NV_PDISP_SOR_STATE1				0x02
+#define SOR_STATE_ASY_HEAD_OPMODE_AWAKE (2 << 0)
+#define SOR_STATE_ASY_ORMODE_NORMAL     (1 << 2)
+#define SOR_STATE_ATTACHED              (1 << 3)
+
+#define HDMI_NV_PDISP_SOR_STATE2				0x03
+#define SOR_STATE_ASY_OWNER_NONE         (0 <<  0)
+#define SOR_STATE_ASY_OWNER_HEAD0        (1 <<  0)
+#define SOR_STATE_ASY_SUBOWNER_NONE      (0 <<  4)
+#define SOR_STATE_ASY_SUBOWNER_SUBHEAD0  (1 <<  4)
+#define SOR_STATE_ASY_SUBOWNER_SUBHEAD1  (2 <<  4)
+#define SOR_STATE_ASY_SUBOWNER_BOTH      (3 <<  4)
+#define SOR_STATE_ASY_CRCMODE_ACTIVE     (0 <<  6)
+#define SOR_STATE_ASY_CRCMODE_COMPLETE   (1 <<  6)
+#define SOR_STATE_ASY_CRCMODE_NON_ACTIVE (2 <<  6)
+#define SOR_STATE_ASY_PROTOCOL_SINGLE_TMDS_A (1 << 8)
+#define SOR_STATE_ASY_PROTOCOL_CUSTOM        (15 << 8)
+#define SOR_STATE_ASY_HSYNCPOL_POS       (0 << 12)
+#define SOR_STATE_ASY_HSYNCPOL_NEG       (1 << 12)
+#define SOR_STATE_ASY_VSYNCPOL_POS       (0 << 13)
+#define SOR_STATE_ASY_VSYNCPOL_NEG       (1 << 13)
+#define SOR_STATE_ASY_DEPOL_POS          (0 << 14)
+#define SOR_STATE_ASY_DEPOL_NEG          (1 << 14)
+
+#define HDMI_NV_PDISP_RG_HDCP_AN_MSB				0x04
+#define HDMI_NV_PDISP_RG_HDCP_AN_LSB				0x05
+#define HDMI_NV_PDISP_RG_HDCP_CN_MSB				0x06
+#define HDMI_NV_PDISP_RG_HDCP_CN_LSB				0x07
+#define HDMI_NV_PDISP_RG_HDCP_AKSV_MSB				0x08
+#define HDMI_NV_PDISP_RG_HDCP_AKSV_LSB				0x09
+#define HDMI_NV_PDISP_RG_HDCP_BKSV_MSB				0x0a
+#define HDMI_NV_PDISP_RG_HDCP_BKSV_LSB				0x0b
+#define HDMI_NV_PDISP_RG_HDCP_CKSV_MSB				0x0c
+#define HDMI_NV_PDISP_RG_HDCP_CKSV_LSB				0x0d
+#define HDMI_NV_PDISP_RG_HDCP_DKSV_MSB				0x0e
+#define HDMI_NV_PDISP_RG_HDCP_DKSV_LSB				0x0f
+#define HDMI_NV_PDISP_RG_HDCP_CTRL				0x10
+#define HDMI_NV_PDISP_RG_HDCP_CMODE				0x11
+#define HDMI_NV_PDISP_RG_HDCP_MPRIME_MSB			0x12
+#define HDMI_NV_PDISP_RG_HDCP_MPRIME_LSB			0x13
+#define HDMI_NV_PDISP_RG_HDCP_SPRIME_MSB			0x14
+#define HDMI_NV_PDISP_RG_HDCP_SPRIME_LSB2			0x15
+#define HDMI_NV_PDISP_RG_HDCP_SPRIME_LSB1			0x16
+#define HDMI_NV_PDISP_RG_HDCP_RI				0x17
+#define HDMI_NV_PDISP_RG_HDCP_CS_MSB				0x18
+#define HDMI_NV_PDISP_RG_HDCP_CS_LSB				0x19
+#define HDMI_NV_PDISP_HDMI_AUDIO_EMU0				0x1a
+#define HDMI_NV_PDISP_HDMI_AUDIO_EMU_RDATA0			0x1b
+#define HDMI_NV_PDISP_HDMI_AUDIO_EMU1				0x1c
+#define HDMI_NV_PDISP_HDMI_AUDIO_EMU2				0x1d
+
+#define HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_CTRL			0x1e
+#define HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_STATUS		0x1f
+#define HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_HEADER		0x20
+#define HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_SUBPACK0_LOW		0x21
+#define HDMI_NV_PDISP_HDMI_AUDIO_INFOFRAME_SUBPACK0_HIGH	0x22
+#define HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_CTRL			0x23
+#define HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_STATUS			0x24
+#define HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_HEADER			0x25
+#define HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_SUBPACK0_LOW		0x26
+#define HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_SUBPACK0_HIGH		0x27
+#define HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_SUBPACK1_LOW		0x28
+#define HDMI_NV_PDISP_HDMI_AVI_INFOFRAME_SUBPACK1_HIGH		0x29
+
+#define INFOFRAME_CTRL_ENABLE (1 << 0)
+
+#define INFOFRAME_HEADER_TYPE(x)    (((x) & 0xff) <<  0)
+#define INFOFRAME_HEADER_VERSION(x) (((x) & 0xff) <<  8)
+#define INFOFRAME_HEADER_LEN(x)     (((x) & 0x0f) << 16)
+
+#define HDMI_NV_PDISP_HDMI_GENERIC_CTRL				0x2a
+#define GENERIC_CTRL_ENABLE (1 <<  0)
+#define GENERIC_CTRL_OTHER  (1 <<  4)
+#define GENERIC_CTRL_SINGLE (1 <<  8)
+#define GENERIC_CTRL_HBLANK (1 << 12)
+#define GENERIC_CTRL_AUDIO  (1 << 16)
+
+#define HDMI_NV_PDISP_HDMI_GENERIC_STATUS			0x2b
+#define HDMI_NV_PDISP_HDMI_GENERIC_HEADER			0x2c
+#define HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK0_LOW			0x2d
+#define HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK0_HIGH		0x2e
+#define HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK1_LOW			0x2f
+#define HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK1_HIGH		0x30
+#define HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK2_LOW			0x31
+#define HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK2_HIGH		0x32
+#define HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK3_LOW			0x33
+#define HDMI_NV_PDISP_HDMI_GENERIC_SUBPACK3_HIGH		0x34
+
+#define HDMI_NV_PDISP_HDMI_ACR_CTRL				0x35
+#define HDMI_NV_PDISP_HDMI_ACR_0320_SUBPACK_LOW			0x36
+#define HDMI_NV_PDISP_HDMI_ACR_0320_SUBPACK_HIGH		0x37
+#define HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_LOW			0x38
+#define HDMI_NV_PDISP_HDMI_ACR_0441_SUBPACK_HIGH		0x39
+#define HDMI_NV_PDISP_HDMI_ACR_0882_SUBPACK_LOW			0x3a
+#define HDMI_NV_PDISP_HDMI_ACR_0882_SUBPACK_HIGH		0x3b
+#define HDMI_NV_PDISP_HDMI_ACR_1764_SUBPACK_LOW			0x3c
+#define HDMI_NV_PDISP_HDMI_ACR_1764_SUBPACK_HIGH		0x3d
+#define HDMI_NV_PDISP_HDMI_ACR_0480_SUBPACK_LOW			0x3e
+#define HDMI_NV_PDISP_HDMI_ACR_0480_SUBPACK_HIGH		0x3f
+#define HDMI_NV_PDISP_HDMI_ACR_0960_SUBPACK_LOW			0x40
+#define HDMI_NV_PDISP_HDMI_ACR_0960_SUBPACK_HIGH		0x41
+#define HDMI_NV_PDISP_HDMI_ACR_1920_SUBPACK_LOW			0x42
+#define HDMI_NV_PDISP_HDMI_ACR_1920_SUBPACK_HIGH		0x43
+
+#define ACR_SUBPACK_CTS(x) (((x) & 0xffffff) << 8)
+#define ACR_SUBPACK_N(x)   (((x) & 0xffffff) << 0)
+#define ACR_ENABLE         (1 << 31)
+
+#define HDMI_NV_PDISP_HDMI_CTRL					0x44
+#define HDMI_CTRL_REKEY(x)         (((x) & 0x7f) <<  0)
+#define HDMI_CTRL_MAX_AC_PACKET(x) (((x) & 0x1f) << 16)
+#define HDMI_CTRL_ENABLE           (1 << 30)
+
+#define HDMI_NV_PDISP_HDMI_VSYNC_KEEPOUT			0x45
+#define HDMI_NV_PDISP_HDMI_VSYNC_WINDOW				0x46
+#define VSYNC_WINDOW_END(x)   (((x) & 0x3ff) <<  0)
+#define VSYNC_WINDOW_START(x) (((x) & 0x3ff) << 16)
+#define VSYNC_WINDOW_ENABLE   (1 << 31)
+
+#define HDMI_NV_PDISP_HDMI_GCP_CTRL				0x47
+#define HDMI_NV_PDISP_HDMI_GCP_STATUS				0x48
+#define HDMI_NV_PDISP_HDMI_GCP_SUBPACK				0x49
+#define HDMI_NV_PDISP_HDMI_CHANNEL_STATUS1			0x4a
+#define HDMI_NV_PDISP_HDMI_CHANNEL_STATUS2			0x4b
+#define HDMI_NV_PDISP_HDMI_EMU0					0x4c
+#define HDMI_NV_PDISP_HDMI_EMU1					0x4d
+#define HDMI_NV_PDISP_HDMI_EMU1_RDATA				0x4e
+
+#define HDMI_NV_PDISP_HDMI_SPARE				0x4f
+#define SPARE_HW_CTS           (1 << 0)
+#define SPARE_FORCE_SW_CTS     (1 << 1)
+#define SPARE_CTS_RESET_VAL(x) (((x) & 0x7) << 16)
+
+#define HDMI_NV_PDISP_HDMI_SPDIF_CHN_STATUS1			0x50
+#define HDMI_NV_PDISP_HDMI_SPDIF_CHN_STATUS2			0x51
+#define HDMI_NV_PDISP_HDMI_HDCPRIF_ROM_CTRL			0x53
+#define HDMI_NV_PDISP_SOR_CAP					0x54
+#define HDMI_NV_PDISP_SOR_PWR					0x55
+#define SOR_PWR_NORMAL_STATE_PD     (0 <<  0)
+#define SOR_PWR_NORMAL_STATE_PU     (1 <<  0)
+#define SOR_PWR_NORMAL_START_NORMAL (0 <<  1)
+#define SOR_PWR_NORMAL_START_ALT    (1 <<  1)
+#define SOR_PWR_SAFE_STATE_PD       (0 << 16)
+#define SOR_PWR_SAFE_STATE_PU       (1 << 16)
+#define SOR_PWR_SETTING_NEW_DONE    (0 << 31)
+#define SOR_PWR_SETTING_NEW_PENDING (1 << 31)
+#define SOR_PWR_SETTING_NEW_TRIGGER (1 << 31)
+
+#define HDMI_NV_PDISP_SOR_TEST					0x56
+#define HDMI_NV_PDISP_SOR_PLL0					0x57
+#define SOR_PLL_PWR            (1 << 0)
+#define SOR_PLL_PDBG           (1 << 1)
+#define SOR_PLL_VCAPD          (1 << 2)
+#define SOR_PLL_PDPORT         (1 << 3)
+#define SOR_PLL_RESISTORSEL    (1 << 4)
+#define SOR_PLL_PULLDOWN       (1 << 5)
+#define SOR_PLL_VCOCAP(x)      (((x) & 0xf) <<  8)
+#define SOR_PLL_BG_V17_S(x)    (((x) & 0xf) << 12)
+#define SOR_PLL_FILTER(x)      (((x) & 0xf) << 16)
+#define SOR_PLL_ICHPMP(x)      (((x) & 0xf) << 24)
+#define SOR_PLL_TX_REG_LOAD(x) (((x) & 0xf) << 28)
+
+#define HDMI_NV_PDISP_SOR_PLL1					0x58
+#define SOR_PLL_TMDS_TERM_ENABLE (1 << 8)
+#define SOR_PLL_TMDS_TERMADJ(x)  (((x) & 0xf) <<  9)
+#define SOR_PLL_LOADADJ(x)       (((x) & 0xf) << 20)
+#define SOR_PLL_PE_EN            (1 << 28)
+#define SOR_PLL_HALF_FULL_PE     (1 << 29)
+#define SOR_PLL_S_D_PIN_PE       (1 << 30)
+
+#define HDMI_NV_PDISP_SOR_PLL2					0x59
+
+#define HDMI_NV_PDISP_SOR_CSTM					0x5a
+#define SOR_CSTM_ROTCLK(x) (((x) & 0xf) << 24)
+#define SOR_CSTM_PLLDIV (1 << 21)
+#define SOR_CSTM_LVDS_ENABLE (1 << 16)
+#define SOR_CSTM_MODE_LVDS (0 << 12)
+#define SOR_CSTM_MODE_TMDS (1 << 12)
+#define SOR_CSTM_MODE_MASK (3 << 12)
+
+#define HDMI_NV_PDISP_SOR_LVDS					0x5b
+#define HDMI_NV_PDISP_SOR_CRCA					0x5c
+#define HDMI_NV_PDISP_SOR_CRCB					0x5d
+#define HDMI_NV_PDISP_SOR_BLANK					0x5e
+#define HDMI_NV_PDISP_SOR_SEQ_CTL				0x5f
+#define SOR_SEQ_PU_PC(x)     (((x) & 0xf) <<  0)
+#define SOR_SEQ_PU_PC_ALT(x) (((x) & 0xf) <<  4)
+#define SOR_SEQ_PD_PC(x)     (((x) & 0xf) <<  8)
+#define SOR_SEQ_PD_PC_ALT(x) (((x) & 0xf) << 12)
+#define SOR_SEQ_PC(x)        (((x) & 0xf) << 16)
+#define SOR_SEQ_STATUS       (1 << 28)
+#define SOR_SEQ_SWITCH       (1 << 30)
+
+#define HDMI_NV_PDISP_SOR_SEQ_INST(x)				(0x60 + (x))
+
+#define SOR_SEQ_INST_WAIT_TIME(x)     (((x) & 0x3ff) << 0)
+#define SOR_SEQ_INST_WAIT_UNITS_VSYNC (2 << 12)
+#define SOR_SEQ_INST_HALT             (1 << 15)
+#define SOR_SEQ_INST_PIN_A_LOW        (0 << 21)
+#define SOR_SEQ_INST_PIN_A_HIGH       (1 << 21)
+#define SOR_SEQ_INST_PIN_B_LOW        (0 << 22)
+#define SOR_SEQ_INST_PIN_B_HIGH       (1 << 22)
+#define SOR_SEQ_INST_DRIVE_PWM_OUT_LO (1 << 23)
+
+#define HDMI_NV_PDISP_SOR_VCRCA0				0x72
+#define HDMI_NV_PDISP_SOR_VCRCA1				0x73
+#define HDMI_NV_PDISP_SOR_CCRCA0				0x74
+#define HDMI_NV_PDISP_SOR_CCRCA1				0x75
+#define HDMI_NV_PDISP_SOR_EDATAA0				0x76
+#define HDMI_NV_PDISP_SOR_EDATAA1				0x77
+#define HDMI_NV_PDISP_SOR_COUNTA0				0x78
+#define HDMI_NV_PDISP_SOR_COUNTA1				0x79
+#define HDMI_NV_PDISP_SOR_DEBUGA0				0x7a
+#define HDMI_NV_PDISP_SOR_DEBUGA1				0x7b
+#define HDMI_NV_PDISP_SOR_TRIG					0x7c
+#define HDMI_NV_PDISP_SOR_MSCHECK				0x7d
+
+#define HDMI_NV_PDISP_SOR_LANE_DRIVE_CURRENT			0x7e
+#define DRIVE_CURRENT_LANE0(x)      (((x) & 0x3f) <<  0)
+#define DRIVE_CURRENT_LANE1(x)      (((x) & 0x3f) <<  8)
+#define DRIVE_CURRENT_LANE2(x)      (((x) & 0x3f) << 16)
+#define DRIVE_CURRENT_LANE3(x)      (((x) & 0x3f) << 24)
+#define DRIVE_CURRENT_LANE0_T114(x) (((x) & 0x7f) <<  0)
+#define DRIVE_CURRENT_LANE1_T114(x) (((x) & 0x7f) <<  8)
+#define DRIVE_CURRENT_LANE2_T114(x) (((x) & 0x7f) << 16)
+#define DRIVE_CURRENT_LANE3_T114(x) (((x) & 0x7f) << 24)
+
+#define DRIVE_CURRENT_1_500_mA  0x00
+#define DRIVE_CURRENT_1_875_mA  0x01
+#define DRIVE_CURRENT_2_250_mA  0x02
+#define DRIVE_CURRENT_2_625_mA  0x03
+#define DRIVE_CURRENT_3_000_mA  0x04
+#define DRIVE_CURRENT_3_375_mA  0x05
+#define DRIVE_CURRENT_3_750_mA  0x06
+#define DRIVE_CURRENT_4_125_mA  0x07
+#define DRIVE_CURRENT_4_500_mA  0x08
+#define DRIVE_CURRENT_4_875_mA  0x09
+#define DRIVE_CURRENT_5_250_mA  0x0a
+#define DRIVE_CURRENT_5_625_mA  0x0b
+#define DRIVE_CURRENT_6_000_mA  0x0c
+#define DRIVE_CURRENT_6_375_mA  0x0d
+#define DRIVE_CURRENT_6_750_mA  0x0e
+#define DRIVE_CURRENT_7_125_mA  0x0f
+#define DRIVE_CURRENT_7_500_mA  0x10
+#define DRIVE_CURRENT_7_875_mA  0x11
+#define DRIVE_CURRENT_8_250_mA  0x12
+#define DRIVE_CURRENT_8_625_mA  0x13
+#define DRIVE_CURRENT_9_000_mA  0x14
+#define DRIVE_CURRENT_9_375_mA  0x15
+#define DRIVE_CURRENT_9_750_mA  0x16
+#define DRIVE_CURRENT_10_125_mA 0x17
+#define DRIVE_CURRENT_10_500_mA 0x18
+#define DRIVE_CURRENT_10_875_mA 0x19
+#define DRIVE_CURRENT_11_250_mA 0x1a
+#define DRIVE_CURRENT_11_625_mA 0x1b
+#define DRIVE_CURRENT_12_000_mA 0x1c
+#define DRIVE_CURRENT_12_375_mA 0x1d
+#define DRIVE_CURRENT_12_750_mA 0x1e
+#define DRIVE_CURRENT_13_125_mA 0x1f
+#define DRIVE_CURRENT_13_500_mA 0x20
+#define DRIVE_CURRENT_13_875_mA 0x21
+#define DRIVE_CURRENT_14_250_mA 0x22
+#define DRIVE_CURRENT_14_625_mA 0x23
+#define DRIVE_CURRENT_15_000_mA 0x24
+#define DRIVE_CURRENT_15_375_mA 0x25
+#define DRIVE_CURRENT_15_750_mA 0x26
+#define DRIVE_CURRENT_16_125_mA 0x27
+#define DRIVE_CURRENT_16_500_mA 0x28
+#define DRIVE_CURRENT_16_875_mA 0x29
+#define DRIVE_CURRENT_17_250_mA 0x2a
+#define DRIVE_CURRENT_17_625_mA 0x2b
+#define DRIVE_CURRENT_18_000_mA 0x2c
+#define DRIVE_CURRENT_18_375_mA 0x2d
+#define DRIVE_CURRENT_18_750_mA 0x2e
+#define DRIVE_CURRENT_19_125_mA 0x2f
+#define DRIVE_CURRENT_19_500_mA 0x30
+#define DRIVE_CURRENT_19_875_mA 0x31
+#define DRIVE_CURRENT_20_250_mA 0x32
+#define DRIVE_CURRENT_20_625_mA 0x33
+#define DRIVE_CURRENT_21_000_mA 0x34
+#define DRIVE_CURRENT_21_375_mA 0x35
+#define DRIVE_CURRENT_21_750_mA 0x36
+#define DRIVE_CURRENT_22_125_mA 0x37
+#define DRIVE_CURRENT_22_500_mA 0x38
+#define DRIVE_CURRENT_22_875_mA 0x39
+#define DRIVE_CURRENT_23_250_mA 0x3a
+#define DRIVE_CURRENT_23_625_mA 0x3b
+#define DRIVE_CURRENT_24_000_mA 0x3c
+#define DRIVE_CURRENT_24_375_mA 0x3d
+#define DRIVE_CURRENT_24_750_mA 0x3e
+
+#define DRIVE_CURRENT_0_000_mA_T114 0x00
+#define DRIVE_CURRENT_0_400_mA_T114 0x01
+#define DRIVE_CURRENT_0_800_mA_T114 0x02
+#define DRIVE_CURRENT_1_200_mA_T114 0x03
+#define DRIVE_CURRENT_1_600_mA_T114 0x04
+#define DRIVE_CURRENT_2_000_mA_T114 0x05
+#define DRIVE_CURRENT_2_400_mA_T114 0x06
+#define DRIVE_CURRENT_2_800_mA_T114 0x07
+#define DRIVE_CURRENT_3_200_mA_T114 0x08
+#define DRIVE_CURRENT_3_600_mA_T114 0x09
+#define DRIVE_CURRENT_4_000_mA_T114 0x0a
+#define DRIVE_CURRENT_4_400_mA_T114 0x0b
+#define DRIVE_CURRENT_4_800_mA_T114 0x0c
+#define DRIVE_CURRENT_5_200_mA_T114 0x0d
+#define DRIVE_CURRENT_5_600_mA_T114 0x0e
+#define DRIVE_CURRENT_6_000_mA_T114 0x0f
+#define DRIVE_CURRENT_6_400_mA_T114 0x10
+#define DRIVE_CURRENT_6_800_mA_T114 0x11
+#define DRIVE_CURRENT_7_200_mA_T114 0x12
+#define DRIVE_CURRENT_7_600_mA_T114 0x13
+#define DRIVE_CURRENT_8_000_mA_T114 0x14
+#define DRIVE_CURRENT_8_400_mA_T114 0x15
+#define DRIVE_CURRENT_8_800_mA_T114 0x16
+#define DRIVE_CURRENT_9_200_mA_T114 0x17
+#define DRIVE_CURRENT_9_600_mA_T114 0x18
+#define DRIVE_CURRENT_10_000_mA_T114 0x19
+#define DRIVE_CURRENT_10_400_mA_T114 0x1a
+#define DRIVE_CURRENT_10_800_mA_T114 0x1b
+#define DRIVE_CURRENT_11_200_mA_T114 0x1c
+#define DRIVE_CURRENT_11_600_mA_T114 0x1d
+#define DRIVE_CURRENT_12_000_mA_T114 0x1e
+#define DRIVE_CURRENT_12_400_mA_T114 0x1f
+#define DRIVE_CURRENT_12_800_mA_T114 0x20
+#define DRIVE_CURRENT_13_200_mA_T114 0x21
+#define DRIVE_CURRENT_13_600_mA_T114 0x22
+#define DRIVE_CURRENT_14_000_mA_T114 0x23
+#define DRIVE_CURRENT_14_400_mA_T114 0x24
+#define DRIVE_CURRENT_14_800_mA_T114 0x25
+#define DRIVE_CURRENT_15_200_mA_T114 0x26
+#define DRIVE_CURRENT_15_600_mA_T114 0x27
+#define DRIVE_CURRENT_16_000_mA_T114 0x28
+#define DRIVE_CURRENT_16_400_mA_T114 0x29
+#define DRIVE_CURRENT_16_800_mA_T114 0x2a
+#define DRIVE_CURRENT_17_200_mA_T114 0x2b
+#define DRIVE_CURRENT_17_600_mA_T114 0x2c
+#define DRIVE_CURRENT_18_000_mA_T114 0x2d
+#define DRIVE_CURRENT_18_400_mA_T114 0x2e
+#define DRIVE_CURRENT_18_800_mA_T114 0x2f
+#define DRIVE_CURRENT_19_200_mA_T114 0x30
+#define DRIVE_CURRENT_19_600_mA_T114 0x31
+#define DRIVE_CURRENT_20_000_mA_T114 0x32
+#define DRIVE_CURRENT_20_400_mA_T114 0x33
+#define DRIVE_CURRENT_20_800_mA_T114 0x34
+#define DRIVE_CURRENT_21_200_mA_T114 0x35
+#define DRIVE_CURRENT_21_600_mA_T114 0x36
+#define DRIVE_CURRENT_22_000_mA_T114 0x37
+#define DRIVE_CURRENT_22_400_mA_T114 0x38
+#define DRIVE_CURRENT_22_800_mA_T114 0x39
+#define DRIVE_CURRENT_23_200_mA_T114 0x3a
+#define DRIVE_CURRENT_23_600_mA_T114 0x3b
+#define DRIVE_CURRENT_24_000_mA_T114 0x3c
+#define DRIVE_CURRENT_24_400_mA_T114 0x3d
+#define DRIVE_CURRENT_24_800_mA_T114 0x3e
+#define DRIVE_CURRENT_25_200_mA_T114 0x3f
+#define DRIVE_CURRENT_25_400_mA_T114 0x40
+#define DRIVE_CURRENT_25_800_mA_T114 0x41
+#define DRIVE_CURRENT_26_200_mA_T114 0x42
+#define DRIVE_CURRENT_26_600_mA_T114 0x43
+#define DRIVE_CURRENT_27_000_mA_T114 0x44
+#define DRIVE_CURRENT_27_400_mA_T114 0x45
+#define DRIVE_CURRENT_27_800_mA_T114 0x46
+#define DRIVE_CURRENT_28_200_mA_T114 0x47
+
+#define HDMI_NV_PDISP_AUDIO_DEBUG0				0x7f
+#define HDMI_NV_PDISP_AUDIO_DEBUG1				0x80
+#define HDMI_NV_PDISP_AUDIO_DEBUG2				0x81
+
+#define HDMI_NV_PDISP_AUDIO_FS(x)				(0x82 + (x))
+#define AUDIO_FS_LOW(x)  (((x) & 0xfff) <<  0)
+#define AUDIO_FS_HIGH(x) (((x) & 0xfff) << 16)
+
+#define HDMI_NV_PDISP_AUDIO_PULSE_WIDTH				0x89
+#define HDMI_NV_PDISP_AUDIO_THRESHOLD				0x8a
+#define HDMI_NV_PDISP_AUDIO_CNTRL0				0x8b
+#define AUDIO_CNTRL0_ERROR_TOLERANCE(x)  (((x) & 0xff) << 0)
+#define AUDIO_CNTRL0_SOURCE_SELECT_AUTO  (0 << 20)
+#define AUDIO_CNTRL0_SOURCE_SELECT_SPDIF (1 << 20)
+#define AUDIO_CNTRL0_SOURCE_SELECT_HDAL  (2 << 20)
+#define AUDIO_CNTRL0_FRAMES_PER_BLOCK(x) (((x) & 0xff) << 24)
+
+#define HDMI_NV_PDISP_AUDIO_N					0x8c
+#define AUDIO_N_VALUE(x)           (((x) & 0xfffff) << 0)
+#define AUDIO_N_RESETF             (1 << 20)
+#define AUDIO_N_GENERATE_NORMAL    (0 << 24)
+#define AUDIO_N_GENERATE_ALTERNATE (1 << 24)
+
+#define HDMI_NV_PDISP_HDCPRIF_ROM_TIMING			0x94
+#define HDMI_NV_PDISP_SOR_REFCLK				0x95
+#define SOR_REFCLK_DIV_INT(x)  (((x) & 0xff) << 8)
+#define SOR_REFCLK_DIV_FRAC(x) (((x) & 0x03) << 6)
+
+#define HDMI_NV_PDISP_CRC_CONTROL				0x96
+#define HDMI_NV_PDISP_INPUT_CONTROL				0x97
+#define HDMI_SRC_DISPLAYA       (0 << 0)
+#define HDMI_SRC_DISPLAYB       (1 << 0)
+#define ARM_VIDEO_RANGE_FULL    (0 << 1)
+#define ARM_VIDEO_RANGE_LIMITED (1 << 1)
+
+#define HDMI_NV_PDISP_SCRATCH					0x98
+#define HDMI_NV_PDISP_PE_CURRENT				0x99
+#define PE_CURRENT0(x) (((x) & 0xf) << 0)
+#define PE_CURRENT1(x) (((x) & 0xf) << 8)
+#define PE_CURRENT2(x) (((x) & 0xf) << 16)
+#define PE_CURRENT3(x) (((x) & 0xf) << 24)
+
+#define PE_CURRENT_0_0_mA 0x0
+#define PE_CURRENT_0_5_mA 0x1
+#define PE_CURRENT_1_0_mA 0x2
+#define PE_CURRENT_1_5_mA 0x3
+#define PE_CURRENT_2_0_mA 0x4
+#define PE_CURRENT_2_5_mA 0x5
+#define PE_CURRENT_3_0_mA 0x6
+#define PE_CURRENT_3_5_mA 0x7
+#define PE_CURRENT_4_0_mA 0x8
+#define PE_CURRENT_4_5_mA 0x9
+#define PE_CURRENT_5_0_mA 0xa
+#define PE_CURRENT_5_5_mA 0xb
+#define PE_CURRENT_6_0_mA 0xc
+#define PE_CURRENT_6_5_mA 0xd
+#define PE_CURRENT_7_0_mA 0xe
+#define PE_CURRENT_7_5_mA 0xf
+
+#define PE_CURRENT_0_mA_T114 0x0
+#define PE_CURRENT_1_mA_T114 0x1
+#define PE_CURRENT_2_mA_T114 0x2
+#define PE_CURRENT_3_mA_T114 0x3
+#define PE_CURRENT_4_mA_T114 0x4
+#define PE_CURRENT_5_mA_T114 0x5
+#define PE_CURRENT_6_mA_T114 0x6
+#define PE_CURRENT_7_mA_T114 0x7
+#define PE_CURRENT_8_mA_T114 0x8
+#define PE_CURRENT_9_mA_T114 0x9
+#define PE_CURRENT_10_mA_T114 0xa
+#define PE_CURRENT_11_mA_T114 0xb
+#define PE_CURRENT_12_mA_T114 0xc
+#define PE_CURRENT_13_mA_T114 0xd
+#define PE_CURRENT_14_mA_T114 0xe
+#define PE_CURRENT_15_mA_T114 0xf
+
+#define HDMI_NV_PDISP_KEY_CTRL					0x9a
+#define HDMI_NV_PDISP_KEY_DEBUG0				0x9b
+#define HDMI_NV_PDISP_KEY_DEBUG1				0x9c
+#define HDMI_NV_PDISP_KEY_DEBUG2				0x9d
+#define HDMI_NV_PDISP_KEY_HDCP_KEY_0				0x9e
+#define HDMI_NV_PDISP_KEY_HDCP_KEY_1				0x9f
+#define HDMI_NV_PDISP_KEY_HDCP_KEY_2				0xa0
+#define HDMI_NV_PDISP_KEY_HDCP_KEY_3				0xa1
+#define HDMI_NV_PDISP_KEY_HDCP_KEY_TRIG				0xa2
+#define HDMI_NV_PDISP_KEY_SKEY_INDEX				0xa3
+
+#define HDMI_NV_PDISP_SOR_AUDIO_CNTRL0				0xac
+#define  SOR_AUDIO_CNTRL0_SOURCE_SELECT_AUTO	(0 << 20)
+#define  SOR_AUDIO_CNTRL0_SOURCE_SELECT_SPDIF	(1 << 20)
+#define  SOR_AUDIO_CNTRL0_SOURCE_SELECT_HDAL	(2 << 20)
+#define  SOR_AUDIO_CNTRL0_INJECT_NULLSMPL	(1 << 29)
+#define HDMI_NV_PDISP_SOR_AUDIO_SPARE0				0xae
+#define  SOR_AUDIO_SPARE0_HBR_ENABLE		(1 << 27)
+#define HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH0		0xba
+#define  SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID	(1 << 30)
+#define  SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK	0xffff
+#define HDMI_NV_PDISP_SOR_AUDIO_HDA_CODEC_SCRATCH1		0xbb
+#define HDMI_NV_PDISP_SOR_AUDIO_HDA_ELD_BUFWR			0xbc
+#define HDMI_NV_PDISP_SOR_AUDIO_HDA_PRESENSE			0xbd
+#define  SOR_AUDIO_HDA_PRESENSE_VALID		(1 << 1)
+#define  SOR_AUDIO_HDA_PRESENSE_PRESENT		(1 << 0)
+
+#define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0320    0xbf
+#define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0441    0xc0
+#define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0882    0xc1
+#define HDMI_NV_PDISP_SOR_AUDIO_AVAL_1764    0xc2
+#define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0480    0xc3
+#define HDMI_NV_PDISP_SOR_AUDIO_AVAL_0960    0xc4
+#define HDMI_NV_PDISP_SOR_AUDIO_AVAL_1920    0xc5
+#define HDMI_NV_PDISP_SOR_AUDIO_AVAL_DEFAULT 0xc5
+
+#define HDMI_NV_PDISP_INT_STATUS			0xcc
+#define  INT_SCRATCH		(1 << 3)
+#define  INT_CP_REQUEST		(1 << 2)
+#define  INT_CODEC_SCRATCH1	(1 << 1)
+#define  INT_CODEC_SCRATCH0	(1 << 0)
+#define HDMI_NV_PDISP_INT_MASK				0xcd
+#define HDMI_NV_PDISP_INT_ENABLE			0xce
+
+#define HDMI_NV_PDISP_SOR_IO_PEAK_CURRENT		0xd1
+#define PEAK_CURRENT_LANE0(x) (((x) & 0x7f) <<  0)
+#define PEAK_CURRENT_LANE1(x) (((x) & 0x7f) <<  8)
+#define PEAK_CURRENT_LANE2(x) (((x) & 0x7f) << 16)
+#define PEAK_CURRENT_LANE3(x) (((x) & 0x7f) << 24)
+
+#define PEAK_CURRENT_0_000_mA 0x00
+#define PEAK_CURRENT_0_200_mA 0x01
+#define PEAK_CURRENT_0_400_mA 0x02
+#define PEAK_CURRENT_0_600_mA 0x03
+#define PEAK_CURRENT_0_800_mA 0x04
+#define PEAK_CURRENT_1_000_mA 0x05
+#define PEAK_CURRENT_1_200_mA 0x06
+#define PEAK_CURRENT_1_400_mA 0x07
+#define PEAK_CURRENT_1_600_mA 0x08
+#define PEAK_CURRENT_1_800_mA 0x09
+#define PEAK_CURRENT_2_000_mA 0x0a
+#define PEAK_CURRENT_2_200_mA 0x0b
+#define PEAK_CURRENT_2_400_mA 0x0c
+#define PEAK_CURRENT_2_600_mA 0x0d
+#define PEAK_CURRENT_2_800_mA 0x0e
+#define PEAK_CURRENT_3_000_mA 0x0f
+#define PEAK_CURRENT_3_200_mA 0x10
+#define PEAK_CURRENT_3_400_mA 0x11
+#define PEAK_CURRENT_3_600_mA 0x12
+#define PEAK_CURRENT_3_800_mA 0x13
+#define PEAK_CURRENT_4_000_mA 0x14
+#define PEAK_CURRENT_4_200_mA 0x15
+#define PEAK_CURRENT_4_400_mA 0x16
+#define PEAK_CURRENT_4_600_mA 0x17
+#define PEAK_CURRENT_4_800_mA 0x18
+#define PEAK_CURRENT_5_000_mA 0x19
+#define PEAK_CURRENT_5_200_mA 0x1a
+#define PEAK_CURRENT_5_400_mA 0x1b
+#define PEAK_CURRENT_5_600_mA 0x1c
+#define PEAK_CURRENT_5_800_mA 0x1d
+#define PEAK_CURRENT_6_000_mA 0x1e
+#define PEAK_CURRENT_6_200_mA 0x1f
+#define PEAK_CURRENT_6_400_mA 0x20
+#define PEAK_CURRENT_6_600_mA 0x21
+#define PEAK_CURRENT_6_800_mA 0x22
+#define PEAK_CURRENT_7_000_mA 0x23
+#define PEAK_CURRENT_7_200_mA 0x24
+#define PEAK_CURRENT_7_400_mA 0x25
+#define PEAK_CURRENT_7_600_mA 0x26
+#define PEAK_CURRENT_7_800_mA 0x27
+#define PEAK_CURRENT_8_000_mA 0x28
+#define PEAK_CURRENT_8_200_mA 0x29
+#define PEAK_CURRENT_8_400_mA 0x2a
+#define PEAK_CURRENT_8_600_mA 0x2b
+#define PEAK_CURRENT_8_800_mA 0x2c
+#define PEAK_CURRENT_9_000_mA 0x2d
+#define PEAK_CURRENT_9_200_mA 0x2e
+#define PEAK_CURRENT_9_400_mA 0x2f
+
+#define HDMI_NV_PDISP_SOR_PAD_CTLS0		0xd2
+
+#endif /* TEGRA_HDMI_H */
diff --git a/drivers/gpu/drm/tegra/hub.c b/drivers/gpu/drm/tegra/hub.c
new file mode 100644
index 000000000..bd442bfd4
--- /dev/null
+++ b/drivers/gpu/drm/tegra/hub.c
@@ -0,0 +1,1233 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2017 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/host1x.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_graph.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/reset.h>
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_blend.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_print.h>
+#include <drm/drm_probe_helper.h>
+
+#include "drm.h"
+#include "dc.h"
+#include "plane.h"
+
+#define NFB 24
+
+static const u32 tegra_shared_plane_formats[] = {
+	DRM_FORMAT_ARGB1555,
+	DRM_FORMAT_RGB565,
+	DRM_FORMAT_RGBA5551,
+	DRM_FORMAT_ARGB8888,
+	DRM_FORMAT_ABGR8888,
+	/* new on Tegra114 */
+	DRM_FORMAT_ABGR4444,
+	DRM_FORMAT_ABGR1555,
+	DRM_FORMAT_BGRA5551,
+	DRM_FORMAT_XRGB1555,
+	DRM_FORMAT_RGBX5551,
+	DRM_FORMAT_XBGR1555,
+	DRM_FORMAT_BGRX5551,
+	DRM_FORMAT_BGR565,
+	DRM_FORMAT_XRGB8888,
+	DRM_FORMAT_XBGR8888,
+	/* planar formats */
+	DRM_FORMAT_UYVY,
+	DRM_FORMAT_YUYV,
+	DRM_FORMAT_YUV420,
+	DRM_FORMAT_YUV422,
+};
+
+static const u64 tegra_shared_plane_modifiers[] = {
+	DRM_FORMAT_MOD_LINEAR,
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(0),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(1),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(2),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(3),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(4),
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(5),
+	/*
+	 * The GPU sector layout is only supported on Tegra194, but these will
+	 * be filtered out later on by ->format_mod_supported() on SoCs where
+	 * it isn't supported.
+	 */
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(0) | DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT,
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(1) | DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT,
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(2) | DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT,
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(3) | DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT,
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(4) | DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT,
+	DRM_FORMAT_MOD_NVIDIA_16BX2_BLOCK(5) | DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT,
+	/* sentinel */
+	DRM_FORMAT_MOD_INVALID
+};
+
+static inline unsigned int tegra_plane_offset(struct tegra_plane *plane,
+					      unsigned int offset)
+{
+	if (offset >= 0x500 && offset <= 0x581) {
+		offset = 0x000 + (offset - 0x500);
+		return plane->offset + offset;
+	}
+
+	if (offset >= 0x700 && offset <= 0x73c) {
+		offset = 0x180 + (offset - 0x700);
+		return plane->offset + offset;
+	}
+
+	if (offset >= 0x800 && offset <= 0x83e) {
+		offset = 0x1c0 + (offset - 0x800);
+		return plane->offset + offset;
+	}
+
+	dev_WARN(plane->dc->dev, "invalid offset: %x\n", offset);
+
+	return plane->offset + offset;
+}
+
+static inline u32 tegra_plane_readl(struct tegra_plane *plane,
+				    unsigned int offset)
+{
+	return tegra_dc_readl(plane->dc, tegra_plane_offset(plane, offset));
+}
+
+static inline void tegra_plane_writel(struct tegra_plane *plane, u32 value,
+				      unsigned int offset)
+{
+	tegra_dc_writel(plane->dc, value, tegra_plane_offset(plane, offset));
+}
+
+static int tegra_windowgroup_enable(struct tegra_windowgroup *wgrp)
+{
+	int err = 0;
+
+	mutex_lock(&wgrp->lock);
+
+	if (wgrp->usecount == 0) {
+		err = host1x_client_resume(wgrp->parent);
+		if (err < 0) {
+			dev_err(wgrp->parent->dev, "failed to resume: %d\n", err);
+			goto unlock;
+		}
+
+		reset_control_deassert(wgrp->rst);
+	}
+
+	wgrp->usecount++;
+
+unlock:
+	mutex_unlock(&wgrp->lock);
+	return err;
+}
+
+static void tegra_windowgroup_disable(struct tegra_windowgroup *wgrp)
+{
+	int err;
+
+	mutex_lock(&wgrp->lock);
+
+	if (wgrp->usecount == 1) {
+		err = reset_control_assert(wgrp->rst);
+		if (err < 0) {
+			pr_err("failed to assert reset for window group %u\n",
+			       wgrp->index);
+		}
+
+		host1x_client_suspend(wgrp->parent);
+	}
+
+	wgrp->usecount--;
+	mutex_unlock(&wgrp->lock);
+}
+
+int tegra_display_hub_prepare(struct tegra_display_hub *hub)
+{
+	unsigned int i;
+
+	/*
+	 * XXX Enabling/disabling windowgroups needs to happen when the owner
+	 * display controller is disabled. There's currently no good point at
+	 * which this could be executed, so unconditionally enable all window
+	 * groups for now.
+	 */
+	for (i = 0; i < hub->soc->num_wgrps; i++) {
+		struct tegra_windowgroup *wgrp = &hub->wgrps[i];
+
+		/* Skip orphaned window group whose parent DC is disabled */
+		if (wgrp->parent)
+			tegra_windowgroup_enable(wgrp);
+	}
+
+	return 0;
+}
+
+void tegra_display_hub_cleanup(struct tegra_display_hub *hub)
+{
+	unsigned int i;
+
+	/*
+	 * XXX Remove this once window groups can be more fine-grainedly
+	 * enabled and disabled.
+	 */
+	for (i = 0; i < hub->soc->num_wgrps; i++) {
+		struct tegra_windowgroup *wgrp = &hub->wgrps[i];
+
+		/* Skip orphaned window group whose parent DC is disabled */
+		if (wgrp->parent)
+			tegra_windowgroup_disable(wgrp);
+	}
+}
+
+static void tegra_shared_plane_update(struct tegra_plane *plane)
+{
+	struct tegra_dc *dc = plane->dc;
+	unsigned long timeout;
+	u32 mask, value;
+
+	mask = COMMON_UPDATE | WIN_A_UPDATE << plane->base.index;
+	tegra_dc_writel(dc, mask, DC_CMD_STATE_CONTROL);
+
+	timeout = jiffies + msecs_to_jiffies(1000);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_dc_readl(dc, DC_CMD_STATE_CONTROL);
+		if ((value & mask) == 0)
+			break;
+
+		usleep_range(100, 400);
+	}
+}
+
+static void tegra_shared_plane_activate(struct tegra_plane *plane)
+{
+	struct tegra_dc *dc = plane->dc;
+	unsigned long timeout;
+	u32 mask, value;
+
+	mask = COMMON_ACTREQ | WIN_A_ACT_REQ << plane->base.index;
+	tegra_dc_writel(dc, mask, DC_CMD_STATE_CONTROL);
+
+	timeout = jiffies + msecs_to_jiffies(1000);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_dc_readl(dc, DC_CMD_STATE_CONTROL);
+		if ((value & mask) == 0)
+			break;
+
+		usleep_range(100, 400);
+	}
+}
+
+static unsigned int
+tegra_shared_plane_get_owner(struct tegra_plane *plane, struct tegra_dc *dc)
+{
+	unsigned int offset =
+		tegra_plane_offset(plane, DC_WIN_CORE_WINDOWGROUP_SET_CONTROL);
+
+	return tegra_dc_readl(dc, offset) & OWNER_MASK;
+}
+
+static bool tegra_dc_owns_shared_plane(struct tegra_dc *dc,
+				       struct tegra_plane *plane)
+{
+	struct device *dev = dc->dev;
+
+	if (tegra_shared_plane_get_owner(plane, dc) == dc->pipe) {
+		if (plane->dc == dc)
+			return true;
+
+		dev_WARN(dev, "head %u owns window %u but is not attached\n",
+			 dc->pipe, plane->index);
+	}
+
+	return false;
+}
+
+static int tegra_shared_plane_set_owner(struct tegra_plane *plane,
+					struct tegra_dc *new)
+{
+	unsigned int offset =
+		tegra_plane_offset(plane, DC_WIN_CORE_WINDOWGROUP_SET_CONTROL);
+	struct tegra_dc *old = plane->dc, *dc = new ? new : old;
+	struct device *dev = new ? new->dev : old->dev;
+	unsigned int owner, index = plane->index;
+	u32 value;
+
+	value = tegra_dc_readl(dc, offset);
+	owner = value & OWNER_MASK;
+
+	if (new && (owner != OWNER_MASK && owner != new->pipe)) {
+		dev_WARN(dev, "window %u owned by head %u\n", index, owner);
+		return -EBUSY;
+	}
+
+	/*
+	 * This seems to happen whenever the head has been disabled with one
+	 * or more windows being active. This is harmless because we'll just
+	 * reassign the window to the new head anyway.
+	 */
+	if (old && owner == OWNER_MASK)
+		dev_dbg(dev, "window %u not owned by head %u but %u\n", index,
+			old->pipe, owner);
+
+	value &= ~OWNER_MASK;
+
+	if (new)
+		value |= OWNER(new->pipe);
+	else
+		value |= OWNER_MASK;
+
+	tegra_dc_writel(dc, value, offset);
+
+	plane->dc = new;
+
+	return 0;
+}
+
+static void tegra_shared_plane_setup_scaler(struct tegra_plane *plane)
+{
+	static const unsigned int coeffs[192] = {
+		0x00000000, 0x3c70e400, 0x3bb037e4, 0x0c51cc9c,
+		0x00100001, 0x3bf0dbfa, 0x3d00f406, 0x3fe003ff,
+		0x00300002, 0x3b80cbf5, 0x3da1040d, 0x3fb003fe,
+		0x00400002, 0x3b20bff1, 0x3e511015, 0x3f9003fc,
+		0x00500002, 0x3ad0b3ed, 0x3f21201d, 0x3f5003fb,
+		0x00500003, 0x3aa0a3e9, 0x3ff13026, 0x3f2007f9,
+		0x00500403, 0x3a7097e6, 0x00e1402f, 0x3ee007f7,
+		0x00500403, 0x3a608be4, 0x01d14c38, 0x3ea00bf6,
+		0x00500403, 0x3a507fe2, 0x02e15c42, 0x3e500ff4,
+		0x00500402, 0x3a6073e1, 0x03f16c4d, 0x3e000ff2,
+		0x00400402, 0x3a706be0, 0x05117858, 0x3db013f0,
+		0x00300402, 0x3a905fe0, 0x06318863, 0x3d6017ee,
+		0x00300402, 0x3ab057e0, 0x0771986e, 0x3d001beb,
+		0x00200001, 0x3af04fe1, 0x08a1a47a, 0x3cb023e9,
+		0x00100001, 0x3b2047e2, 0x09e1b485, 0x3c6027e7,
+		0x00100000, 0x3b703fe2, 0x0b11c091, 0x3c002fe6,
+		0x3f203800, 0x0391103f, 0x3ff0a014, 0x0811606c,
+		0x3f2037ff, 0x0351083c, 0x03e11842, 0x3f203c00,
+		0x3f302fff, 0x03010439, 0x04311c45, 0x3f104401,
+		0x3f302fff, 0x02c0fc35, 0x04812448, 0x3f104802,
+		0x3f4027ff, 0x0270f832, 0x04c1284b, 0x3f205003,
+		0x3f4023ff, 0x0230f030, 0x0511304e, 0x3f205403,
+		0x3f601fff, 0x01f0e82d, 0x05613451, 0x3f205c04,
+		0x3f701bfe, 0x01b0e02a, 0x05a13c54, 0x3f306006,
+		0x3f7017fe, 0x0170d827, 0x05f14057, 0x3f406807,
+		0x3f8017ff, 0x0140d424, 0x0641445a, 0x3f406c08,
+		0x3fa013ff, 0x0100cc22, 0x0681485d, 0x3f507409,
+		0x3fa00fff, 0x00d0c41f, 0x06d14c60, 0x3f607c0b,
+		0x3fc00fff, 0x0090bc1c, 0x07115063, 0x3f80840c,
+		0x3fd00bff, 0x0070b41a, 0x07515465, 0x3f908c0e,
+		0x3fe007ff, 0x0040b018, 0x07915868, 0x3fb0900f,
+		0x3ff00400, 0x0010a816, 0x07d15c6a, 0x3fd09811,
+		0x00a04c0e, 0x0460f442, 0x0240a827, 0x05c15859,
+		0x0090440d, 0x0440f040, 0x0480fc43, 0x00b05010,
+		0x0080400c, 0x0410ec3e, 0x04910044, 0x00d05411,
+		0x0070380b, 0x03f0e83d, 0x04b10846, 0x00e05812,
+		0x0060340a, 0x03d0e43b, 0x04d10c48, 0x00f06013,
+		0x00503009, 0x03b0e039, 0x04e11449, 0x01106415,
+		0x00402c08, 0x0390d838, 0x05011c4b, 0x01206c16,
+		0x00302807, 0x0370d436, 0x0511204c, 0x01407018,
+		0x00302406, 0x0340d034, 0x0531244e, 0x01507419,
+		0x00202005, 0x0320cc32, 0x05412c50, 0x01707c1b,
+		0x00101c04, 0x0300c431, 0x05613451, 0x0180801d,
+		0x00101803, 0x02e0c02f, 0x05713853, 0x01a0881e,
+		0x00101002, 0x02b0bc2d, 0x05814054, 0x01c08c20,
+		0x00000c02, 0x02a0b82c, 0x05914455, 0x01e09421,
+		0x00000801, 0x0280b02a, 0x05a14c57, 0x02009c23,
+		0x00000400, 0x0260ac28, 0x05b15458, 0x0220a025,
+	};
+	unsigned int ratio, row, column;
+
+	for (ratio = 0; ratio <= 2; ratio++) {
+		for (row = 0; row <= 15; row++) {
+			for (column = 0; column <= 3; column++) {
+				unsigned int index = (ratio << 6) + (row << 2) + column;
+				u32 value;
+
+				value = COEFF_INDEX(index) | COEFF_DATA(coeffs[index]);
+				tegra_plane_writel(plane, value,
+						   DC_WIN_WINDOWGROUP_SET_INPUT_SCALER_COEFF);
+			}
+		}
+	}
+}
+
+static void tegra_dc_assign_shared_plane(struct tegra_dc *dc,
+					 struct tegra_plane *plane)
+{
+	u32 value;
+	int err;
+
+	if (!tegra_dc_owns_shared_plane(dc, plane)) {
+		err = tegra_shared_plane_set_owner(plane, dc);
+		if (err < 0)
+			return;
+	}
+
+	value = tegra_plane_readl(plane, DC_WIN_CORE_IHUB_LINEBUF_CONFIG);
+	value |= MODE_FOUR_LINES;
+	tegra_plane_writel(plane, value, DC_WIN_CORE_IHUB_LINEBUF_CONFIG);
+
+	value = tegra_plane_readl(plane, DC_WIN_CORE_IHUB_WGRP_FETCH_METER);
+	value = SLOTS(1);
+	tegra_plane_writel(plane, value, DC_WIN_CORE_IHUB_WGRP_FETCH_METER);
+
+	/* disable watermark */
+	value = tegra_plane_readl(plane, DC_WIN_CORE_IHUB_WGRP_LATENCY_CTLA);
+	value &= ~LATENCY_CTL_MODE_ENABLE;
+	tegra_plane_writel(plane, value, DC_WIN_CORE_IHUB_WGRP_LATENCY_CTLA);
+
+	value = tegra_plane_readl(plane, DC_WIN_CORE_IHUB_WGRP_LATENCY_CTLB);
+	value |= WATERMARK_MASK;
+	tegra_plane_writel(plane, value, DC_WIN_CORE_IHUB_WGRP_LATENCY_CTLB);
+
+	/* pipe meter */
+	value = tegra_plane_readl(plane, DC_WIN_CORE_PRECOMP_WGRP_PIPE_METER);
+	value = PIPE_METER_INT(0) | PIPE_METER_FRAC(0);
+	tegra_plane_writel(plane, value, DC_WIN_CORE_PRECOMP_WGRP_PIPE_METER);
+
+	/* mempool entries */
+	value = tegra_plane_readl(plane, DC_WIN_CORE_IHUB_WGRP_POOL_CONFIG);
+	value = MEMPOOL_ENTRIES(0x331);
+	tegra_plane_writel(plane, value, DC_WIN_CORE_IHUB_WGRP_POOL_CONFIG);
+
+	value = tegra_plane_readl(plane, DC_WIN_CORE_IHUB_THREAD_GROUP);
+	value &= ~THREAD_NUM_MASK;
+	value |= THREAD_NUM(plane->base.index);
+	value |= THREAD_GROUP_ENABLE;
+	tegra_plane_writel(plane, value, DC_WIN_CORE_IHUB_THREAD_GROUP);
+
+	tegra_shared_plane_setup_scaler(plane);
+
+	tegra_shared_plane_update(plane);
+	tegra_shared_plane_activate(plane);
+}
+
+static void tegra_dc_remove_shared_plane(struct tegra_dc *dc,
+					 struct tegra_plane *plane)
+{
+	tegra_shared_plane_set_owner(plane, NULL);
+}
+
+static int tegra_shared_plane_atomic_check(struct drm_plane *plane,
+					   struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *new_plane_state = drm_atomic_get_new_plane_state(state,
+										 plane);
+	struct tegra_plane_state *plane_state = to_tegra_plane_state(new_plane_state);
+	struct tegra_shared_plane *tegra = to_tegra_shared_plane(plane);
+	struct tegra_bo_tiling *tiling = &plane_state->tiling;
+	struct tegra_dc *dc = to_tegra_dc(new_plane_state->crtc);
+	int err;
+
+	/* no need for further checks if the plane is being disabled */
+	if (!new_plane_state->crtc || !new_plane_state->fb)
+		return 0;
+
+	err = tegra_plane_format(new_plane_state->fb->format->format,
+				 &plane_state->format,
+				 &plane_state->swap);
+	if (err < 0)
+		return err;
+
+	err = tegra_fb_get_tiling(new_plane_state->fb, tiling);
+	if (err < 0)
+		return err;
+
+	if (tiling->mode == TEGRA_BO_TILING_MODE_BLOCK &&
+	    !dc->soc->supports_block_linear) {
+		DRM_ERROR("hardware doesn't support block linear mode\n");
+		return -EINVAL;
+	}
+
+	if (tiling->sector_layout == TEGRA_BO_SECTOR_LAYOUT_GPU &&
+	    !dc->soc->supports_sector_layout) {
+		DRM_ERROR("hardware doesn't support GPU sector layout\n");
+		return -EINVAL;
+	}
+
+	/*
+	 * Tegra doesn't support different strides for U and V planes so we
+	 * error out if the user tries to display a framebuffer with such a
+	 * configuration.
+	 */
+	if (new_plane_state->fb->format->num_planes > 2) {
+		if (new_plane_state->fb->pitches[2] != new_plane_state->fb->pitches[1]) {
+			DRM_ERROR("unsupported UV-plane configuration\n");
+			return -EINVAL;
+		}
+	}
+
+	/* XXX scaling is not yet supported, add a check here */
+
+	err = tegra_plane_state_add(&tegra->base, new_plane_state);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static void tegra_shared_plane_atomic_disable(struct drm_plane *plane,
+					      struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *old_state = drm_atomic_get_old_plane_state(state,
+									   plane);
+	struct tegra_plane *p = to_tegra_plane(plane);
+	struct tegra_dc *dc;
+	u32 value;
+	int err;
+
+	/* rien ne va plus */
+	if (!old_state || !old_state->crtc)
+		return;
+
+	dc = to_tegra_dc(old_state->crtc);
+
+	err = host1x_client_resume(&dc->client);
+	if (err < 0) {
+		dev_err(dc->dev, "failed to resume: %d\n", err);
+		return;
+	}
+
+	/*
+	 * XXX Legacy helpers seem to sometimes call ->atomic_disable() even
+	 * on planes that are already disabled. Make sure we fallback to the
+	 * head for this particular state instead of crashing.
+	 */
+	if (WARN_ON(p->dc == NULL))
+		p->dc = dc;
+
+	value = tegra_plane_readl(p, DC_WIN_WIN_OPTIONS);
+	value &= ~WIN_ENABLE;
+	tegra_plane_writel(p, value, DC_WIN_WIN_OPTIONS);
+
+	tegra_dc_remove_shared_plane(dc, p);
+
+	host1x_client_suspend(&dc->client);
+}
+
+static inline u32 compute_phase_incr(fixed20_12 in, unsigned int out)
+{
+	u64 tmp, tmp1;
+
+	tmp = (u64)dfixed_trunc(in);
+	tmp1 = (tmp << NFB) + ((u64)out >> 1);
+	do_div(tmp1, out);
+
+	return lower_32_bits(tmp1);
+}
+
+static void tegra_shared_plane_atomic_update(struct drm_plane *plane,
+					     struct drm_atomic_commit *state)
+{
+	struct drm_plane_state *new_state = drm_atomic_get_new_plane_state(state,
+									   plane);
+	struct tegra_plane_state *tegra_plane_state = to_tegra_plane_state(new_state);
+	struct tegra_dc *dc = to_tegra_dc(new_state->crtc);
+	unsigned int zpos = new_state->normalized_zpos;
+	struct drm_framebuffer *fb = new_state->fb;
+	struct tegra_plane *p = to_tegra_plane(plane);
+	u32 value, min_width, bypass = 0;
+	dma_addr_t base, addr_flag = 0;
+	unsigned int bpc, planes;
+	bool yuv;
+	int err;
+
+	/* rien ne va plus */
+	if (!new_state->crtc || !new_state->fb)
+		return;
+
+	if (!new_state->visible) {
+		tegra_shared_plane_atomic_disable(plane, state);
+		return;
+	}
+
+	err = host1x_client_resume(&dc->client);
+	if (err < 0) {
+		dev_err(dc->dev, "failed to resume: %d\n", err);
+		return;
+	}
+
+	yuv = tegra_plane_format_is_yuv(tegra_plane_state->format, &planes, &bpc);
+
+	tegra_dc_assign_shared_plane(dc, p);
+
+	tegra_plane_writel(p, VCOUNTER, DC_WIN_CORE_ACT_CONTROL);
+
+	/* blending */
+	value = BLEND_FACTOR_DST_ALPHA_ZERO | BLEND_FACTOR_SRC_ALPHA_K2 |
+		BLEND_FACTOR_DST_COLOR_NEG_K1_TIMES_SRC |
+		BLEND_FACTOR_SRC_COLOR_K1_TIMES_SRC;
+	tegra_plane_writel(p, value, DC_WIN_BLEND_MATCH_SELECT);
+
+	value = BLEND_FACTOR_DST_ALPHA_ZERO | BLEND_FACTOR_SRC_ALPHA_K2 |
+		BLEND_FACTOR_DST_COLOR_NEG_K1_TIMES_SRC |
+		BLEND_FACTOR_SRC_COLOR_K1_TIMES_SRC;
+	tegra_plane_writel(p, value, DC_WIN_BLEND_NOMATCH_SELECT);
+
+	value = K2(255) | K1(255) | WINDOW_LAYER_DEPTH(255 - zpos);
+	tegra_plane_writel(p, value, DC_WIN_BLEND_LAYER_CONTROL);
+
+	/* scaling */
+	min_width = min(new_state->src_w >> 16, new_state->crtc_w);
+
+	value = tegra_plane_readl(p, DC_WINC_PRECOMP_WGRP_PIPE_CAPC);
+
+	if (min_width < MAX_PIXELS_5TAP444(value)) {
+		value = HORIZONTAL_TAPS_5 | VERTICAL_TAPS_5;
+	} else {
+		value = tegra_plane_readl(p, DC_WINC_PRECOMP_WGRP_PIPE_CAPE);
+
+		if (min_width < MAX_PIXELS_2TAP444(value))
+			value = HORIZONTAL_TAPS_2 | VERTICAL_TAPS_2;
+		else
+			dev_err(dc->dev, "invalid minimum width: %u\n", min_width);
+	}
+
+	value = HORIZONTAL_TAPS_5 | VERTICAL_TAPS_5;
+	tegra_plane_writel(p, value, DC_WIN_WINDOWGROUP_SET_CONTROL_INPUT_SCALER);
+
+	if (new_state->src_w != new_state->crtc_w << 16) {
+		fixed20_12 width = dfixed_init(new_state->src_w >> 16);
+		u32 incr = compute_phase_incr(width, new_state->crtc_w) & ~0x1;
+		u32 init = (1 << (NFB - 1)) + (incr >> 1);
+
+		tegra_plane_writel(p, incr, DC_WIN_SET_INPUT_SCALER_HPHASE_INCR);
+		tegra_plane_writel(p, init, DC_WIN_SET_INPUT_SCALER_H_START_PHASE);
+	} else {
+		bypass |= INPUT_SCALER_HBYPASS;
+	}
+
+	if (new_state->src_h != new_state->crtc_h << 16) {
+		fixed20_12 height = dfixed_init(new_state->src_h >> 16);
+		u32 incr = compute_phase_incr(height, new_state->crtc_h) & ~0x1;
+		u32 init = (1 << (NFB - 1)) + (incr >> 1);
+
+		tegra_plane_writel(p, incr, DC_WIN_SET_INPUT_SCALER_VPHASE_INCR);
+		tegra_plane_writel(p, init, DC_WIN_SET_INPUT_SCALER_V_START_PHASE);
+	} else {
+		bypass |= INPUT_SCALER_VBYPASS;
+	}
+
+	tegra_plane_writel(p, bypass, DC_WIN_WINDOWGROUP_SET_INPUT_SCALER_USAGE);
+
+	/* disable compression */
+	tegra_plane_writel(p, 0, DC_WINBUF_CDE_CONTROL);
+
+#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
+	/*
+	 * Physical address bit 39 in Tegra194 is used as a switch for special
+	 * logic that swizzles the memory using either the legacy Tegra or the
+	 * dGPU sector layout.
+	 */
+	if (tegra_plane_state->tiling.sector_layout == TEGRA_BO_SECTOR_LAYOUT_GPU)
+		addr_flag = BIT_ULL(39);
+#endif
+
+	base = tegra_plane_state->iova[0] + fb->offsets[0];
+	base |= addr_flag;
+
+	tegra_plane_writel(p, tegra_plane_state->format, DC_WIN_COLOR_DEPTH);
+	tegra_plane_writel(p, 0, DC_WIN_PRECOMP_WGRP_PARAMS);
+
+	value = V_POSITION(new_state->crtc_y) |
+		H_POSITION(new_state->crtc_x);
+	tegra_plane_writel(p, value, DC_WIN_POSITION);
+
+	value = V_SIZE(new_state->crtc_h) | H_SIZE(new_state->crtc_w);
+	tegra_plane_writel(p, value, DC_WIN_SIZE);
+
+	value = WIN_ENABLE | COLOR_EXPAND;
+	tegra_plane_writel(p, value, DC_WIN_WIN_OPTIONS);
+
+	value = V_SIZE(new_state->src_h >> 16) | H_SIZE(new_state->src_w >> 16);
+	tegra_plane_writel(p, value, DC_WIN_CROPPED_SIZE);
+
+	tegra_plane_writel(p, upper_32_bits(base), DC_WINBUF_START_ADDR_HI);
+	tegra_plane_writel(p, lower_32_bits(base), DC_WINBUF_START_ADDR);
+
+	value = PITCH(fb->pitches[0]);
+	tegra_plane_writel(p, value, DC_WIN_PLANAR_STORAGE);
+
+	if (yuv && planes > 1) {
+		base = tegra_plane_state->iova[1] + fb->offsets[1];
+		base |= addr_flag;
+
+		tegra_plane_writel(p, upper_32_bits(base), DC_WINBUF_START_ADDR_HI_U);
+		tegra_plane_writel(p, lower_32_bits(base), DC_WINBUF_START_ADDR_U);
+
+		if (planes > 2) {
+			base = tegra_plane_state->iova[2] + fb->offsets[2];
+			base |= addr_flag;
+
+			tegra_plane_writel(p, upper_32_bits(base), DC_WINBUF_START_ADDR_HI_V);
+			tegra_plane_writel(p, lower_32_bits(base), DC_WINBUF_START_ADDR_V);
+		}
+
+		value = PITCH_U(fb->pitches[1]);
+
+		if (planes > 2)
+			value |= PITCH_V(fb->pitches[2]);
+
+		tegra_plane_writel(p, value, DC_WIN_PLANAR_STORAGE_UV);
+	} else {
+		tegra_plane_writel(p, 0, DC_WINBUF_START_ADDR_U);
+		tegra_plane_writel(p, 0, DC_WINBUF_START_ADDR_HI_U);
+		tegra_plane_writel(p, 0, DC_WINBUF_START_ADDR_V);
+		tegra_plane_writel(p, 0, DC_WINBUF_START_ADDR_HI_V);
+		tegra_plane_writel(p, 0, DC_WIN_PLANAR_STORAGE_UV);
+	}
+
+	value = CLAMP_BEFORE_BLEND | INPUT_RANGE_FULL;
+
+	if (yuv) {
+		if (bpc < 12)
+			value |= DEGAMMA_YUV8_10;
+		else
+			value |= DEGAMMA_YUV12;
+
+		/* XXX parameterize */
+		value |= COLOR_SPACE_YUV_2020;
+	} else {
+		if (!tegra_plane_format_is_indexed(tegra_plane_state->format))
+			value |= DEGAMMA_SRGB;
+	}
+
+	tegra_plane_writel(p, value, DC_WIN_SET_PARAMS);
+
+	value = OFFSET_X(new_state->src_y >> 16) |
+		OFFSET_Y(new_state->src_x >> 16);
+	tegra_plane_writel(p, value, DC_WINBUF_CROPPED_POINT);
+
+	if (dc->soc->supports_block_linear) {
+		unsigned long height = tegra_plane_state->tiling.value;
+
+		/* XXX */
+		switch (tegra_plane_state->tiling.mode) {
+		case TEGRA_BO_TILING_MODE_PITCH:
+			value = DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(0) |
+				DC_WINBUF_SURFACE_KIND_PITCH;
+			break;
+
+		/* XXX not supported on Tegra186 and later */
+		case TEGRA_BO_TILING_MODE_TILED:
+			value = DC_WINBUF_SURFACE_KIND_TILED;
+			break;
+
+		case TEGRA_BO_TILING_MODE_BLOCK:
+			value = DC_WINBUF_SURFACE_KIND_BLOCK_HEIGHT(height) |
+				DC_WINBUF_SURFACE_KIND_BLOCK;
+			break;
+		}
+
+		tegra_plane_writel(p, value, DC_WINBUF_SURFACE_KIND);
+	}
+
+	/* disable gamut CSC */
+	value = tegra_plane_readl(p, DC_WIN_WINDOW_SET_CONTROL);
+	value &= ~CONTROL_CSC_ENABLE;
+	tegra_plane_writel(p, value, DC_WIN_WINDOW_SET_CONTROL);
+
+	host1x_client_suspend(&dc->client);
+}
+
+static const struct drm_plane_helper_funcs tegra_shared_plane_helper_funcs = {
+	.prepare_fb = tegra_plane_prepare_fb,
+	.cleanup_fb = tegra_plane_cleanup_fb,
+	.atomic_check = tegra_shared_plane_atomic_check,
+	.atomic_update = tegra_shared_plane_atomic_update,
+	.atomic_disable = tegra_shared_plane_atomic_disable,
+};
+
+struct drm_plane *tegra_shared_plane_create(struct drm_device *drm,
+					    struct tegra_dc *dc,
+					    unsigned int wgrp,
+					    unsigned int index,
+					    enum drm_plane_type type)
+{
+	struct tegra_drm *tegra = drm->dev_private;
+	struct tegra_display_hub *hub = tegra->hub;
+	struct tegra_shared_plane *plane;
+	unsigned int possible_crtcs;
+	unsigned int num_formats;
+	const u64 *modifiers;
+	struct drm_plane *p;
+	const u32 *formats;
+	int err;
+
+	plane = kzalloc_obj(*plane);
+	if (!plane)
+		return ERR_PTR(-ENOMEM);
+
+	plane->base.offset = 0x0a00 + 0x0300 * index;
+	plane->base.index = index;
+
+	plane->wgrp = &hub->wgrps[wgrp];
+	plane->wgrp->parent = &dc->client;
+
+	p = &plane->base.base;
+
+	/* planes can be assigned to arbitrary CRTCs */
+	possible_crtcs = BIT(tegra->num_crtcs) - 1;
+
+	num_formats = ARRAY_SIZE(tegra_shared_plane_formats);
+	formats = tegra_shared_plane_formats;
+	modifiers = tegra_shared_plane_modifiers;
+
+	err = drm_universal_plane_init(drm, p, possible_crtcs,
+				       &tegra_plane_funcs, formats,
+				       num_formats, modifiers, type, NULL);
+	if (err < 0) {
+		kfree(plane);
+		return ERR_PTR(err);
+	}
+
+	drm_plane_helper_add(p, &tegra_shared_plane_helper_funcs);
+	drm_plane_create_zpos_property(p, 0, 0, 255);
+
+	return p;
+}
+
+static struct drm_private_state *
+tegra_display_hub_duplicate_state(struct drm_private_obj *obj)
+{
+	struct tegra_display_hub_state *state;
+
+	state = kmemdup(obj->state, sizeof(*state), GFP_KERNEL);
+	if (!state)
+		return NULL;
+
+	__drm_atomic_helper_private_obj_duplicate_state(obj, &state->base);
+
+	return &state->base;
+}
+
+static void tegra_display_hub_destroy_state(struct drm_private_obj *obj,
+					    struct drm_private_state *state)
+{
+	struct tegra_display_hub_state *hub_state =
+		to_tegra_display_hub_state(state);
+
+	kfree(hub_state);
+}
+
+static struct drm_private_state *
+tegra_display_hub_create_state(struct drm_private_obj *obj)
+{
+	struct tegra_display_hub_state *hub_state;
+
+	hub_state = kzalloc_obj(*hub_state);
+	if (!hub_state)
+		return ERR_PTR(-ENOMEM);
+
+	__drm_atomic_helper_private_obj_create_state(obj, &hub_state->base);
+
+	return &hub_state->base;
+}
+
+static const struct drm_private_state_funcs tegra_display_hub_state_funcs = {
+	.atomic_create_state = tegra_display_hub_create_state,
+	.atomic_duplicate_state = tegra_display_hub_duplicate_state,
+	.atomic_destroy_state = tegra_display_hub_destroy_state,
+};
+
+static struct tegra_display_hub_state *
+tegra_display_hub_get_state(struct tegra_display_hub *hub,
+			    struct drm_atomic_commit *state)
+{
+	struct drm_private_state *priv;
+
+	priv = drm_atomic_get_private_obj_state(state, &hub->base);
+	if (IS_ERR(priv))
+		return ERR_CAST(priv);
+
+	return to_tegra_display_hub_state(priv);
+}
+
+int tegra_display_hub_atomic_check(struct drm_device *drm,
+				   struct drm_atomic_commit *state)
+{
+	struct tegra_drm *tegra = drm->dev_private;
+	struct tegra_display_hub_state *hub_state;
+	struct drm_crtc_state *old, *new;
+	struct drm_crtc *crtc;
+	unsigned int i;
+
+	if (!tegra->hub)
+		return 0;
+
+	hub_state = tegra_display_hub_get_state(tegra->hub, state);
+	if (IS_ERR(hub_state))
+		return PTR_ERR(hub_state);
+
+	/*
+	 * The display hub display clock needs to be fed by the display clock
+	 * with the highest frequency to ensure proper functioning of all the
+	 * displays.
+	 *
+	 * Note that this isn't used before Tegra186, but it doesn't hurt and
+	 * conditionalizing it would make the code less clean.
+	 */
+	for_each_oldnew_crtc_in_state(state, crtc, old, new, i) {
+		struct tegra_dc_state *dc = to_dc_state(new);
+
+		if (new->active) {
+			if (!hub_state->clk || dc->pclk > hub_state->rate) {
+				hub_state->dc = to_tegra_dc(dc->base.crtc);
+				hub_state->clk = hub_state->dc->clk;
+				hub_state->rate = dc->pclk;
+			}
+		}
+	}
+
+	return 0;
+}
+
+static void tegra_display_hub_update(struct tegra_dc *dc)
+{
+	u32 value;
+	int err;
+
+	err = host1x_client_resume(&dc->client);
+	if (err < 0) {
+		dev_err(dc->dev, "failed to resume: %d\n", err);
+		return;
+	}
+
+	value = tegra_dc_readl(dc, DC_CMD_IHUB_COMMON_MISC_CTL);
+	value &= ~LATENCY_EVENT;
+	tegra_dc_writel(dc, value, DC_CMD_IHUB_COMMON_MISC_CTL);
+
+	value = tegra_dc_readl(dc, DC_DISP_IHUB_COMMON_DISPLAY_FETCH_METER);
+	value = CURS_SLOTS(1) | WGRP_SLOTS(1);
+	tegra_dc_writel(dc, value, DC_DISP_IHUB_COMMON_DISPLAY_FETCH_METER);
+
+	tegra_dc_writel(dc, COMMON_UPDATE, DC_CMD_STATE_CONTROL);
+	tegra_dc_readl(dc, DC_CMD_STATE_CONTROL);
+	tegra_dc_writel(dc, COMMON_ACTREQ, DC_CMD_STATE_CONTROL);
+	tegra_dc_readl(dc, DC_CMD_STATE_CONTROL);
+
+	host1x_client_suspend(&dc->client);
+}
+
+void tegra_display_hub_atomic_commit(struct drm_device *drm,
+				     struct drm_atomic_commit *state)
+{
+	struct tegra_drm *tegra = drm->dev_private;
+	struct tegra_display_hub *hub = tegra->hub;
+	struct tegra_display_hub_state *hub_state;
+	struct device *dev = hub->client.dev;
+	int err;
+
+	hub_state = to_tegra_display_hub_state(hub->base.state);
+
+	if (hub_state->clk) {
+		err = clk_set_rate(hub_state->clk, hub_state->rate);
+		if (err < 0)
+			dev_err(dev, "failed to set rate of %pC to %lu Hz\n",
+				hub_state->clk, hub_state->rate);
+
+		err = clk_set_parent(hub->clk_disp, hub_state->clk);
+		if (err < 0)
+			dev_err(dev, "failed to set parent of %pC to %pC: %d\n",
+				hub->clk_disp, hub_state->clk, err);
+	}
+
+	if (hub_state->dc)
+		tegra_display_hub_update(hub_state->dc);
+}
+
+static int tegra_display_hub_init(struct host1x_client *client)
+{
+	struct tegra_display_hub *hub = to_tegra_display_hub(client);
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = drm->dev_private;
+
+	drm_atomic_private_obj_init(drm, &hub->base,
+				    &tegra_display_hub_state_funcs);
+
+	tegra->hub = hub;
+
+	return 0;
+}
+
+static int tegra_display_hub_exit(struct host1x_client *client)
+{
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = drm->dev_private;
+
+	drm_atomic_private_obj_fini(&tegra->hub->base);
+	tegra->hub = NULL;
+
+	return 0;
+}
+
+static int tegra_display_hub_runtime_suspend(struct host1x_client *client)
+{
+	struct tegra_display_hub *hub = to_tegra_display_hub(client);
+	struct device *dev = client->dev;
+	unsigned int i = hub->num_heads;
+	int err;
+
+	err = reset_control_assert(hub->rst);
+	if (err < 0)
+		return err;
+
+	while (i--)
+		clk_disable_unprepare(hub->clk_heads[i]);
+
+	clk_disable_unprepare(hub->clk_hub);
+	clk_disable_unprepare(hub->clk_dsc);
+	clk_disable_unprepare(hub->clk_disp);
+
+	pm_runtime_put_sync(dev);
+
+	return 0;
+}
+
+static int tegra_display_hub_runtime_resume(struct host1x_client *client)
+{
+	struct tegra_display_hub *hub = to_tegra_display_hub(client);
+	struct device *dev = client->dev;
+	unsigned int i;
+	int err;
+
+	err = pm_runtime_resume_and_get(dev);
+	if (err < 0) {
+		dev_err(dev, "failed to get runtime PM: %d\n", err);
+		return err;
+	}
+
+	err = clk_prepare_enable(hub->clk_disp);
+	if (err < 0)
+		goto put_rpm;
+
+	err = clk_prepare_enable(hub->clk_dsc);
+	if (err < 0)
+		goto disable_disp;
+
+	err = clk_prepare_enable(hub->clk_hub);
+	if (err < 0)
+		goto disable_dsc;
+
+	for (i = 0; i < hub->num_heads; i++) {
+		err = clk_prepare_enable(hub->clk_heads[i]);
+		if (err < 0)
+			goto disable_heads;
+	}
+
+	err = reset_control_deassert(hub->rst);
+	if (err < 0)
+		goto disable_heads;
+
+	return 0;
+
+disable_heads:
+	while (i--)
+		clk_disable_unprepare(hub->clk_heads[i]);
+
+	clk_disable_unprepare(hub->clk_hub);
+disable_dsc:
+	clk_disable_unprepare(hub->clk_dsc);
+disable_disp:
+	clk_disable_unprepare(hub->clk_disp);
+put_rpm:
+	pm_runtime_put_sync(dev);
+	return err;
+}
+
+static const struct host1x_client_ops tegra_display_hub_ops = {
+	.init = tegra_display_hub_init,
+	.exit = tegra_display_hub_exit,
+	.suspend = tegra_display_hub_runtime_suspend,
+	.resume = tegra_display_hub_runtime_resume,
+};
+
+static int tegra_display_hub_probe(struct platform_device *pdev)
+{
+	u64 dma_mask = dma_get_mask(pdev->dev.parent);
+	struct device_node *child = NULL;
+	struct tegra_display_hub *hub;
+	struct clk *clk;
+	unsigned int i;
+	int err;
+
+	err = dma_coerce_mask_and_coherent(&pdev->dev, dma_mask);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set DMA mask: %d\n", err);
+		return err;
+	}
+
+	hub = devm_kzalloc(&pdev->dev, sizeof(*hub), GFP_KERNEL);
+	if (!hub)
+		return -ENOMEM;
+
+	hub->soc = of_device_get_match_data(&pdev->dev);
+
+	hub->clk_disp = devm_clk_get(&pdev->dev, "disp");
+	if (IS_ERR(hub->clk_disp)) {
+		err = PTR_ERR(hub->clk_disp);
+		return err;
+	}
+
+	if (hub->soc->supports_dsc) {
+		hub->clk_dsc = devm_clk_get(&pdev->dev, "dsc");
+		if (IS_ERR(hub->clk_dsc)) {
+			err = PTR_ERR(hub->clk_dsc);
+			return err;
+		}
+	}
+
+	hub->clk_hub = devm_clk_get(&pdev->dev, "hub");
+	if (IS_ERR(hub->clk_hub)) {
+		err = PTR_ERR(hub->clk_hub);
+		return err;
+	}
+
+	hub->rst = devm_reset_control_get(&pdev->dev, "misc");
+	if (IS_ERR(hub->rst)) {
+		err = PTR_ERR(hub->rst);
+		return err;
+	}
+
+	hub->wgrps = devm_kcalloc(&pdev->dev, hub->soc->num_wgrps,
+				  sizeof(*hub->wgrps), GFP_KERNEL);
+	if (!hub->wgrps)
+		return -ENOMEM;
+
+	for (i = 0; i < hub->soc->num_wgrps; i++) {
+		struct tegra_windowgroup *wgrp = &hub->wgrps[i];
+		char id[16];
+
+		snprintf(id, sizeof(id), "wgrp%u", i);
+		mutex_init(&wgrp->lock);
+		wgrp->usecount = 0;
+		wgrp->index = i;
+
+		wgrp->rst = devm_reset_control_get(&pdev->dev, id);
+		if (IS_ERR(wgrp->rst))
+			return PTR_ERR(wgrp->rst);
+
+		err = reset_control_assert(wgrp->rst);
+		if (err < 0)
+			return err;
+	}
+
+	hub->num_heads = of_get_child_count(pdev->dev.of_node);
+
+	hub->clk_heads = devm_kcalloc(&pdev->dev, hub->num_heads, sizeof(clk),
+				      GFP_KERNEL);
+	if (!hub->clk_heads)
+		return -ENOMEM;
+
+	for (i = 0; i < hub->num_heads; i++) {
+		child = of_get_next_child(pdev->dev.of_node, child);
+		if (!child) {
+			dev_err(&pdev->dev, "failed to find node for head %u\n",
+				i);
+			return -ENODEV;
+		}
+
+		clk = devm_get_clk_from_child(&pdev->dev, child, "dc");
+		if (IS_ERR(clk)) {
+			dev_err(&pdev->dev, "failed to get clock for head %u\n",
+				i);
+			of_node_put(child);
+			return PTR_ERR(clk);
+		}
+
+		hub->clk_heads[i] = clk;
+	}
+
+	of_node_put(child);
+
+	/* XXX: enable clock across reset? */
+	err = reset_control_assert(hub->rst);
+	if (err < 0)
+		return err;
+
+	platform_set_drvdata(pdev, hub);
+	pm_runtime_enable(&pdev->dev);
+
+	INIT_LIST_HEAD(&hub->client.list);
+	hub->client.ops = &tegra_display_hub_ops;
+	hub->client.dev = &pdev->dev;
+
+	err = host1x_client_register(&hub->client);
+	if (err < 0)
+		dev_err(&pdev->dev, "failed to register host1x client: %d\n",
+			err);
+
+	err = devm_of_platform_populate(&pdev->dev);
+	if (err < 0)
+		goto unregister;
+
+	return err;
+
+unregister:
+	host1x_client_unregister(&hub->client);
+	pm_runtime_disable(&pdev->dev);
+	return err;
+}
+
+static void tegra_display_hub_remove(struct platform_device *pdev)
+{
+	struct tegra_display_hub *hub = platform_get_drvdata(pdev);
+	unsigned int i;
+
+	host1x_client_unregister(&hub->client);
+
+	for (i = 0; i < hub->soc->num_wgrps; i++) {
+		struct tegra_windowgroup *wgrp = &hub->wgrps[i];
+
+		mutex_destroy(&wgrp->lock);
+	}
+
+	pm_runtime_disable(&pdev->dev);
+}
+
+static const struct tegra_display_hub_soc tegra186_display_hub = {
+	.num_wgrps = 6,
+	.supports_dsc = true,
+};
+
+static const struct tegra_display_hub_soc tegra194_display_hub = {
+	.num_wgrps = 6,
+	.supports_dsc = false,
+};
+
+static const struct of_device_id tegra_display_hub_of_match[] = {
+	{
+		.compatible = "nvidia,tegra194-display",
+		.data = &tegra194_display_hub
+	}, {
+		.compatible = "nvidia,tegra186-display",
+		.data = &tegra186_display_hub
+	}, {
+		/* sentinel */
+	}
+};
+MODULE_DEVICE_TABLE(of, tegra_display_hub_of_match);
+
+struct platform_driver tegra_display_hub_driver = {
+	.driver = {
+		.name = "tegra-display-hub",
+		.of_match_table = tegra_display_hub_of_match,
+	},
+	.probe = tegra_display_hub_probe,
+	.remove = tegra_display_hub_remove,
+};
diff --git a/drivers/gpu/drm/tegra/hub.h b/drivers/gpu/drm/tegra/hub.h
new file mode 100644
index 000000000..9f3a108f6
--- /dev/null
+++ b/drivers/gpu/drm/tegra/hub.h
@@ -0,0 +1,98 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2017 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#ifndef TEGRA_HUB_H
+#define TEGRA_HUB_H 1
+
+#include <drm/drm_plane.h>
+
+#include "plane.h"
+
+struct tegra_dc;
+
+struct tegra_windowgroup {
+	unsigned int usecount;
+	struct mutex lock;
+
+	unsigned int index;
+	struct host1x_client *parent;
+	struct reset_control *rst;
+};
+
+struct tegra_shared_plane {
+	struct tegra_plane base;
+	struct tegra_windowgroup *wgrp;
+};
+
+static inline struct tegra_shared_plane *
+to_tegra_shared_plane(struct drm_plane *plane)
+{
+	return container_of(plane, struct tegra_shared_plane, base.base);
+}
+
+struct tegra_display_hub_soc {
+	unsigned int num_wgrps;
+	bool supports_dsc;
+};
+
+struct tegra_display_hub {
+	struct drm_private_obj base;
+	struct host1x_client client;
+	struct clk *clk_disp;
+	struct clk *clk_dsc;
+	struct clk *clk_hub;
+	struct reset_control *rst;
+
+	unsigned int num_heads;
+	struct clk **clk_heads;
+
+	const struct tegra_display_hub_soc *soc;
+	struct tegra_windowgroup *wgrps;
+};
+
+static inline struct tegra_display_hub *
+to_tegra_display_hub(struct host1x_client *client)
+{
+	return container_of(client, struct tegra_display_hub, client);
+}
+
+struct tegra_display_hub_state {
+	struct drm_private_state base;
+
+	struct tegra_dc *dc;
+	unsigned long rate;
+	struct clk *clk;
+};
+
+static inline struct tegra_display_hub_state *
+to_tegra_display_hub_state(struct drm_private_state *priv)
+{
+	return container_of(priv, struct tegra_display_hub_state, base);
+}
+
+struct tegra_plane;
+
+int tegra_display_hub_prepare(struct tegra_display_hub *hub);
+void tegra_display_hub_cleanup(struct tegra_display_hub *hub);
+
+struct drm_plane *tegra_shared_plane_create(struct drm_device *drm,
+					    struct tegra_dc *dc,
+					    unsigned int wgrp,
+					    unsigned int index,
+					    enum drm_plane_type type);
+
+int tegra_display_hub_atomic_check(struct drm_device *drm,
+				   struct drm_atomic_commit *state);
+void tegra_display_hub_atomic_commit(struct drm_device *drm,
+				     struct drm_atomic_commit *state);
+
+#define DC_CMD_IHUB_COMMON_MISC_CTL 0x068
+#define  LATENCY_EVENT (1 << 3)
+
+#define DC_DISP_IHUB_COMMON_DISPLAY_FETCH_METER 0x451
+#define  CURS_SLOTS(x) (((x) & 0xff) << 8)
+#define  WGRP_SLOTS(x) (((x) & 0xff) << 0)
+
+#endif /* TEGRA_HUB_H */
diff --git a/drivers/gpu/drm/tegra/mipi-phy.c b/drivers/gpu/drm/tegra/mipi-phy.c
new file mode 100644
index 000000000..4fe05f2df
--- /dev/null
+++ b/drivers/gpu/drm/tegra/mipi-phy.c
@@ -0,0 +1,134 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#include <linux/errno.h>
+#include <linux/kernel.h>
+
+#include "mipi-phy.h"
+
+/*
+ * Default D-PHY timings based on MIPI D-PHY specification. Derived from the
+ * valid ranges specified in Section 6.9, Table 14, Page 40 of the D-PHY
+ * specification (v1.2) with minor adjustments.
+ */
+int mipi_dphy_timing_get_default(struct mipi_dphy_timing *timing,
+				 unsigned long period)
+{
+	timing->clkmiss = 0;
+	timing->clkpost = 70 + 52 * period;
+	timing->clkpre = 8;
+	timing->clkprepare = 65;
+	timing->clksettle = 95;
+	timing->clktermen = 0;
+	timing->clktrail = 80;
+	timing->clkzero = 260;
+	timing->dtermen = 0;
+	timing->eot = 0;
+	timing->hsexit = 120;
+	timing->hsprepare = 65 + 5 * period;
+	timing->hszero = 145 + 5 * period;
+	timing->hssettle = 85 + 6 * period;
+	timing->hsskip = 40;
+
+	/*
+	 * The MIPI D-PHY specification (Section 6.9, v1.2, Table 14, Page 40)
+	 * contains this formula as:
+	 *
+	 *     T_HS-TRAIL = max(n * 8 * period, 60 + n * 4 * period)
+	 *
+	 * where n = 1 for forward-direction HS mode and n = 4 for reverse-
+	 * direction HS mode. There's only one setting and this function does
+	 * not parameterize on anything other that period, so this code will
+	 * assumes that reverse-direction HS mode is supported and uses n = 4.
+	 */
+	timing->hstrail = max(4 * 8 * period, 60 + 4 * 4 * period);
+
+	timing->init = 100000;
+	timing->lpx = 60;
+	timing->taget = 5 * timing->lpx;
+	timing->tago = 4 * timing->lpx;
+	timing->tasure = 2 * timing->lpx;
+	timing->wakeup = 1000000;
+
+	return 0;
+}
+
+/*
+ * Validate D-PHY timing according to MIPI D-PHY specification (v1.2, Section
+ * Section 6.9 "Global Operation Timing Parameters").
+ */
+int mipi_dphy_timing_validate(struct mipi_dphy_timing *timing,
+			      unsigned long period)
+{
+	if (timing->clkmiss > 60)
+		return -EINVAL;
+
+	if (timing->clkpost < (60 + 52 * period))
+		return -EINVAL;
+
+	if (timing->clkpre < 8)
+		return -EINVAL;
+
+	if (timing->clkprepare < 38 || timing->clkprepare > 95)
+		return -EINVAL;
+
+	if (timing->clksettle < 95 || timing->clksettle > 300)
+		return -EINVAL;
+
+	if (timing->clktermen > 38)
+		return -EINVAL;
+
+	if (timing->clktrail < 60)
+		return -EINVAL;
+
+	if (timing->clkprepare + timing->clkzero < 300)
+		return -EINVAL;
+
+	if (timing->dtermen > 35 + 4 * period)
+		return -EINVAL;
+
+	if (timing->eot > 105 + 12 * period)
+		return -EINVAL;
+
+	if (timing->hsexit < 100)
+		return -EINVAL;
+
+	if (timing->hsprepare < 40 + 4 * period ||
+	    timing->hsprepare > 85 + 6 * period)
+		return -EINVAL;
+
+	if (timing->hsprepare + timing->hszero < 145 + 10 * period)
+		return -EINVAL;
+
+	if ((timing->hssettle < 85 + 6 * period) ||
+	    (timing->hssettle > 145 + 10 * period))
+		return -EINVAL;
+
+	if (timing->hsskip < 40 || timing->hsskip > 55 + 4 * period)
+		return -EINVAL;
+
+	if (timing->hstrail < max(8 * period, 60 + 4 * period))
+		return -EINVAL;
+
+	if (timing->init < 100000)
+		return -EINVAL;
+
+	if (timing->lpx < 50)
+		return -EINVAL;
+
+	if (timing->taget != 5 * timing->lpx)
+		return -EINVAL;
+
+	if (timing->tago != 4 * timing->lpx)
+		return -EINVAL;
+
+	if (timing->tasure < timing->lpx || timing->tasure > 2 * timing->lpx)
+		return -EINVAL;
+
+	if (timing->wakeup < 1000000)
+		return -EINVAL;
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/tegra/mipi-phy.h b/drivers/gpu/drm/tegra/mipi-phy.h
new file mode 100644
index 000000000..41889a750
--- /dev/null
+++ b/drivers/gpu/drm/tegra/mipi-phy.h
@@ -0,0 +1,48 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#ifndef DRM_TEGRA_MIPI_PHY_H
+#define DRM_TEGRA_MIPI_PHY_H
+
+/*
+ * D-PHY timing parameters
+ *
+ * A detailed description of these parameters can be found in the  MIPI
+ * Alliance Specification for D-PHY, Section 5.9 "Global Operation Timing
+ * Parameters".
+ *
+ * All parameters are specified in nanoseconds.
+ */
+struct mipi_dphy_timing {
+	unsigned int clkmiss;
+	unsigned int clkpost;
+	unsigned int clkpre;
+	unsigned int clkprepare;
+	unsigned int clksettle;
+	unsigned int clktermen;
+	unsigned int clktrail;
+	unsigned int clkzero;
+	unsigned int dtermen;
+	unsigned int eot;
+	unsigned int hsexit;
+	unsigned int hsprepare;
+	unsigned int hszero;
+	unsigned int hssettle;
+	unsigned int hsskip;
+	unsigned int hstrail;
+	unsigned int init;
+	unsigned int lpx;
+	unsigned int taget;
+	unsigned int tago;
+	unsigned int tasure;
+	unsigned int wakeup;
+};
+
+int mipi_dphy_timing_get_default(struct mipi_dphy_timing *timing,
+				 unsigned long period);
+int mipi_dphy_timing_validate(struct mipi_dphy_timing *timing,
+			      unsigned long period);
+
+#endif
diff --git a/drivers/gpu/drm/tegra/nvdec.c b/drivers/gpu/drm/tegra/nvdec.c
new file mode 100644
index 000000000..7a38664e8
--- /dev/null
+++ b/drivers/gpu/drm/tegra/nvdec.c
@@ -0,0 +1,578 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2015-2022, NVIDIA Corporation.
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/host1x.h>
+#include <linux/iommu.h>
+#include <linux/iopoll.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/reset.h>
+
+#include <soc/tegra/mc.h>
+
+#include "drm.h"
+#include "falcon.h"
+#include "riscv.h"
+#include "vic.h"
+
+#define NVDEC_FALCON_DEBUGINFO			0x1094
+#define NVDEC_TFBIF_TRANSCFG			0x2c44
+
+struct nvdec_config {
+	const char *firmware;
+	unsigned int version;
+	bool supports_sid;
+	bool has_riscv;
+	bool has_extra_clocks;
+};
+
+struct nvdec {
+	struct falcon falcon;
+
+	void __iomem *regs;
+	struct tegra_drm_client client;
+	struct host1x_channel *channel;
+	struct device *dev;
+	struct clk_bulk_data clks[3];
+	unsigned int num_clks;
+	struct reset_control *reset;
+
+	/* Platform configuration */
+	const struct nvdec_config *config;
+
+	/* RISC-V specific data */
+	struct tegra_drm_riscv riscv;
+	phys_addr_t carveout_base;
+};
+
+static inline struct nvdec *to_nvdec(struct tegra_drm_client *client)
+{
+	return container_of(client, struct nvdec, client);
+}
+
+static inline void nvdec_writel(struct nvdec *nvdec, u32 value,
+				unsigned int offset)
+{
+	writel(value, nvdec->regs + offset);
+}
+
+static int nvdec_boot_falcon(struct nvdec *nvdec)
+{
+	u32 stream_id;
+	int err;
+
+	if (nvdec->config->supports_sid && tegra_dev_iommu_get_stream_id(nvdec->dev, &stream_id)) {
+		u32 value;
+
+		value = TRANSCFG_ATT(1, TRANSCFG_SID_FALCON) | TRANSCFG_ATT(0, TRANSCFG_SID_HW);
+		nvdec_writel(nvdec, value, NVDEC_TFBIF_TRANSCFG);
+
+		nvdec_writel(nvdec, stream_id, VIC_THI_STREAMID0);
+		nvdec_writel(nvdec, stream_id, VIC_THI_STREAMID1);
+	}
+
+	err = falcon_boot(&nvdec->falcon);
+	if (err < 0)
+		return err;
+
+	err = falcon_wait_idle(&nvdec->falcon);
+	if (err < 0) {
+		dev_err(nvdec->dev, "falcon boot timed out\n");
+		return err;
+	}
+
+	return 0;
+}
+
+static int nvdec_wait_debuginfo(struct nvdec *nvdec, const char *phase)
+{
+	int err;
+	u32 val;
+
+	err = readl_poll_timeout(nvdec->regs + NVDEC_FALCON_DEBUGINFO, val, val == 0x0, 10, 100000);
+	if (err) {
+		dev_err(nvdec->dev, "failed to boot %s, debuginfo=0x%x\n", phase, val);
+		return err;
+	}
+
+	return 0;
+}
+
+static int nvdec_boot_riscv(struct nvdec *nvdec)
+{
+	int err;
+
+	err = reset_control_acquire(nvdec->reset);
+	if (err)
+		return err;
+
+	nvdec_writel(nvdec, 0xabcd1234, NVDEC_FALCON_DEBUGINFO);
+
+	err = tegra_drm_riscv_boot_bootrom(&nvdec->riscv, nvdec->carveout_base, 1,
+					   &nvdec->riscv.bl_desc);
+	if (err) {
+		dev_err(nvdec->dev, "failed to execute bootloader\n");
+		goto release_reset;
+	}
+
+	err = nvdec_wait_debuginfo(nvdec, "bootloader");
+	if (err)
+		goto release_reset;
+
+	err = reset_control_reset(nvdec->reset);
+	if (err)
+		goto release_reset;
+
+	nvdec_writel(nvdec, 0xabcd1234, NVDEC_FALCON_DEBUGINFO);
+
+	err = tegra_drm_riscv_boot_bootrom(&nvdec->riscv, nvdec->carveout_base, 1,
+					   &nvdec->riscv.os_desc);
+	if (err) {
+		dev_err(nvdec->dev, "failed to execute firmware\n");
+		goto release_reset;
+	}
+
+	err = nvdec_wait_debuginfo(nvdec, "firmware");
+	if (err)
+		goto release_reset;
+
+release_reset:
+	reset_control_release(nvdec->reset);
+
+	return err;
+}
+
+static int nvdec_init(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = dev->dev_private;
+	struct nvdec *nvdec = to_nvdec(drm);
+	int err;
+
+	err = host1x_client_iommu_attach(client);
+	if (err < 0 && err != -ENODEV) {
+		dev_err(nvdec->dev, "failed to attach to domain: %d\n", err);
+		return err;
+	}
+
+	nvdec->channel = host1x_channel_request(client);
+	if (!nvdec->channel) {
+		err = -ENOMEM;
+		goto detach;
+	}
+
+	client->syncpts[0] = host1x_syncpt_request(client, 0);
+	if (!client->syncpts[0]) {
+		err = -ENOMEM;
+		goto free_channel;
+	}
+
+	err = tegra_drm_register_client(tegra, drm);
+	if (err < 0)
+		goto free_syncpt;
+
+	/*
+	 * Inherit the DMA parameters (such as maximum segment size) from the
+	 * parent host1x device.
+	 */
+	client->dev->dma_parms = client->host->dma_parms;
+
+	return 0;
+
+free_syncpt:
+	host1x_syncpt_put(client->syncpts[0]);
+free_channel:
+	host1x_channel_put(nvdec->channel);
+detach:
+	host1x_client_iommu_detach(client);
+
+	return err;
+}
+
+static int nvdec_exit(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = dev->dev_private;
+	struct nvdec *nvdec = to_nvdec(drm);
+	int err;
+
+	/* avoid a dangling pointer just in case this disappears */
+	client->dev->dma_parms = NULL;
+
+	err = tegra_drm_unregister_client(tegra, drm);
+	if (err < 0)
+		return err;
+
+	pm_runtime_dont_use_autosuspend(client->dev);
+	pm_runtime_force_suspend(client->dev);
+
+	host1x_syncpt_put(client->syncpts[0]);
+	host1x_channel_put(nvdec->channel);
+	host1x_client_iommu_detach(client);
+
+	nvdec->channel = NULL;
+
+	if (client->group) {
+		dma_unmap_single(nvdec->dev, nvdec->falcon.firmware.phys,
+				 nvdec->falcon.firmware.size, DMA_TO_DEVICE);
+		tegra_drm_free(tegra, nvdec->falcon.firmware.size,
+			       nvdec->falcon.firmware.virt,
+			       nvdec->falcon.firmware.iova);
+	} else {
+		dma_free_coherent(nvdec->dev, nvdec->falcon.firmware.size,
+				  nvdec->falcon.firmware.virt,
+				  nvdec->falcon.firmware.iova);
+	}
+
+	return 0;
+}
+
+static const struct host1x_client_ops nvdec_client_ops = {
+	.init = nvdec_init,
+	.exit = nvdec_exit,
+};
+
+static int nvdec_load_falcon_firmware(struct nvdec *nvdec)
+{
+	struct host1x_client *client = &nvdec->client.base;
+	struct tegra_drm *tegra = nvdec->client.drm;
+	dma_addr_t iova;
+	size_t size;
+	void *virt;
+	int err;
+
+	if (nvdec->falcon.firmware.virt)
+		return 0;
+
+	err = falcon_read_firmware(&nvdec->falcon, nvdec->config->firmware);
+	if (err < 0)
+		return err;
+
+	size = nvdec->falcon.firmware.size;
+
+	if (!client->group) {
+		virt = dma_alloc_coherent(nvdec->dev, size, &iova, GFP_KERNEL);
+		if (!virt)
+			return -ENOMEM;
+	} else {
+		virt = tegra_drm_alloc(tegra, size, &iova);
+		if (IS_ERR(virt))
+			return PTR_ERR(virt);
+	}
+
+	nvdec->falcon.firmware.virt = virt;
+	nvdec->falcon.firmware.iova = iova;
+
+	err = falcon_load_firmware(&nvdec->falcon);
+	if (err < 0)
+		goto cleanup;
+
+	/*
+	 * In this case we have received an IOVA from the shared domain, so we
+	 * need to make sure to get the physical address so that the DMA API
+	 * knows what memory pages to flush the cache for.
+	 */
+	if (client->group) {
+		dma_addr_t phys;
+
+		phys = dma_map_single(nvdec->dev, virt, size, DMA_TO_DEVICE);
+
+		err = dma_mapping_error(nvdec->dev, phys);
+		if (err < 0)
+			goto cleanup;
+
+		nvdec->falcon.firmware.phys = phys;
+	}
+
+	return 0;
+
+cleanup:
+	if (!client->group)
+		dma_free_coherent(nvdec->dev, size, virt, iova);
+	else
+		tegra_drm_free(tegra, size, virt, iova);
+
+	return err;
+}
+
+static __maybe_unused int nvdec_runtime_resume(struct device *dev)
+{
+	struct nvdec *nvdec = dev_get_drvdata(dev);
+	int err;
+
+	err = clk_bulk_prepare_enable(nvdec->num_clks, nvdec->clks);
+	if (err < 0)
+		return err;
+
+	usleep_range(10, 20);
+
+	if (nvdec->config->has_riscv) {
+		err = nvdec_boot_riscv(nvdec);
+		if (err < 0)
+			goto disable;
+	} else {
+		err = nvdec_load_falcon_firmware(nvdec);
+		if (err < 0)
+			goto disable;
+
+		err = nvdec_boot_falcon(nvdec);
+		if (err < 0)
+			goto disable;
+	}
+
+	return 0;
+
+disable:
+	clk_bulk_disable_unprepare(nvdec->num_clks, nvdec->clks);
+	return err;
+}
+
+static __maybe_unused int nvdec_runtime_suspend(struct device *dev)
+{
+	struct nvdec *nvdec = dev_get_drvdata(dev);
+
+	host1x_channel_stop(nvdec->channel);
+
+	clk_bulk_disable_unprepare(nvdec->num_clks, nvdec->clks);
+
+	return 0;
+}
+
+static int nvdec_open_channel(struct tegra_drm_client *client,
+			    struct tegra_drm_context *context)
+{
+	struct nvdec *nvdec = to_nvdec(client);
+
+	context->channel = host1x_channel_get(nvdec->channel);
+	if (!context->channel)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static void nvdec_close_channel(struct tegra_drm_context *context)
+{
+	host1x_channel_put(context->channel);
+}
+
+static int nvdec_can_use_memory_ctx(struct tegra_drm_client *client, bool *supported)
+{
+	*supported = true;
+
+	return 0;
+}
+
+static const struct tegra_drm_client_ops nvdec_ops = {
+	.open_channel = nvdec_open_channel,
+	.close_channel = nvdec_close_channel,
+	.submit = tegra_drm_submit,
+	.get_streamid_offset = tegra_drm_get_streamid_offset_thi,
+	.can_use_memory_ctx = nvdec_can_use_memory_ctx,
+};
+
+#define NVIDIA_TEGRA_210_NVDEC_FIRMWARE "nvidia/tegra210/nvdec.bin"
+
+static const struct nvdec_config nvdec_t210_config = {
+	.firmware = NVIDIA_TEGRA_210_NVDEC_FIRMWARE,
+	.version = 0x21,
+	.supports_sid = false,
+};
+
+#define NVIDIA_TEGRA_186_NVDEC_FIRMWARE "nvidia/tegra186/nvdec.bin"
+
+static const struct nvdec_config nvdec_t186_config = {
+	.firmware = NVIDIA_TEGRA_186_NVDEC_FIRMWARE,
+	.version = 0x18,
+	.supports_sid = true,
+};
+
+#define NVIDIA_TEGRA_194_NVDEC_FIRMWARE "nvidia/tegra194/nvdec.bin"
+
+static const struct nvdec_config nvdec_t194_config = {
+	.firmware = NVIDIA_TEGRA_194_NVDEC_FIRMWARE,
+	.version = 0x19,
+	.supports_sid = true,
+};
+
+static const struct nvdec_config nvdec_t234_config = {
+	.version = 0x23,
+	.supports_sid = true,
+	.has_riscv = true,
+	.has_extra_clocks = true,
+};
+
+static const struct of_device_id tegra_nvdec_of_match[] = {
+	{ .compatible = "nvidia,tegra210-nvdec", .data = &nvdec_t210_config },
+	{ .compatible = "nvidia,tegra186-nvdec", .data = &nvdec_t186_config },
+	{ .compatible = "nvidia,tegra194-nvdec", .data = &nvdec_t194_config },
+	{ .compatible = "nvidia,tegra234-nvdec", .data = &nvdec_t234_config },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, tegra_nvdec_of_match);
+
+static int nvdec_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct host1x_syncpt **syncpts;
+	struct nvdec *nvdec;
+	u32 host_class;
+	int err;
+
+	/* inherit DMA mask from host1x parent */
+	err = dma_coerce_mask_and_coherent(dev, *dev->parent->dma_mask);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set DMA mask: %d\n", err);
+		return err;
+	}
+
+	nvdec = devm_kzalloc(dev, sizeof(*nvdec), GFP_KERNEL);
+	if (!nvdec)
+		return -ENOMEM;
+
+	nvdec->config = of_device_get_match_data(dev);
+
+	syncpts = devm_kzalloc(dev, sizeof(*syncpts), GFP_KERNEL);
+	if (!syncpts)
+		return -ENOMEM;
+
+	nvdec->regs = devm_platform_get_and_ioremap_resource(pdev, 0, NULL);
+	if (IS_ERR(nvdec->regs))
+		return PTR_ERR(nvdec->regs);
+
+	nvdec->clks[0].id = "nvdec";
+	nvdec->num_clks = 1;
+
+	if (nvdec->config->has_extra_clocks) {
+		nvdec->num_clks = 3;
+		nvdec->clks[1].id = "fuse";
+		nvdec->clks[2].id = "tsec_pka";
+	}
+
+	err = devm_clk_bulk_get(dev, nvdec->num_clks, nvdec->clks);
+	if (err) {
+		dev_err(&pdev->dev, "failed to get clock(s)\n");
+		return err;
+	}
+
+	err = clk_set_rate(nvdec->clks[0].clk, ULONG_MAX);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set clock rate\n");
+		return err;
+	}
+
+	err = of_property_read_u32(dev->of_node, "nvidia,host1x-class", &host_class);
+	if (err < 0)
+		host_class = HOST1X_CLASS_NVDEC;
+
+	if (nvdec->config->has_riscv) {
+		struct tegra_mc *mc;
+
+		mc = devm_tegra_memory_controller_get(dev);
+		if (IS_ERR(mc)) {
+			dev_err_probe(dev, PTR_ERR(mc),
+				"failed to get memory controller handle\n");
+			return PTR_ERR(mc);
+		}
+
+		err = tegra_mc_get_carveout_info(mc, 1, &nvdec->carveout_base, NULL);
+		if (err) {
+			dev_err(dev, "failed to get carveout info: %d\n", err);
+			return err;
+		}
+
+		nvdec->reset = devm_reset_control_get_exclusive_released(dev, "nvdec");
+		if (IS_ERR(nvdec->reset)) {
+			dev_err_probe(dev, PTR_ERR(nvdec->reset), "failed to get reset\n");
+			return PTR_ERR(nvdec->reset);
+		}
+
+		nvdec->riscv.dev = dev;
+		nvdec->riscv.regs = nvdec->regs;
+
+		err = tegra_drm_riscv_read_descriptors(&nvdec->riscv);
+		if (err < 0)
+			return err;
+	} else {
+		nvdec->falcon.dev = dev;
+		nvdec->falcon.regs = nvdec->regs;
+
+		err = falcon_init(&nvdec->falcon);
+		if (err < 0)
+			return err;
+	}
+
+	platform_set_drvdata(pdev, nvdec);
+
+	INIT_LIST_HEAD(&nvdec->client.base.list);
+	nvdec->client.base.ops = &nvdec_client_ops;
+	nvdec->client.base.dev = dev;
+	nvdec->client.base.class = host_class;
+	nvdec->client.base.syncpts = syncpts;
+	nvdec->client.base.num_syncpts = 1;
+	nvdec->dev = dev;
+
+	INIT_LIST_HEAD(&nvdec->client.list);
+	nvdec->client.version = nvdec->config->version;
+	nvdec->client.ops = &nvdec_ops;
+
+	err = host1x_client_register(&nvdec->client.base);
+	if (err < 0) {
+		dev_err(dev, "failed to register host1x client: %d\n", err);
+		goto exit_falcon;
+	}
+
+	pm_runtime_enable(dev);
+	pm_runtime_use_autosuspend(dev);
+	pm_runtime_set_autosuspend_delay(dev, 500);
+
+	return 0;
+
+exit_falcon:
+	falcon_exit(&nvdec->falcon);
+
+	return err;
+}
+
+static void nvdec_remove(struct platform_device *pdev)
+{
+	struct nvdec *nvdec = platform_get_drvdata(pdev);
+
+	pm_runtime_disable(&pdev->dev);
+	host1x_client_unregister(&nvdec->client.base);
+	falcon_exit(&nvdec->falcon);
+}
+
+static const struct dev_pm_ops nvdec_pm_ops = {
+	SET_RUNTIME_PM_OPS(nvdec_runtime_suspend, nvdec_runtime_resume, NULL)
+	SET_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend,
+				pm_runtime_force_resume)
+};
+
+struct platform_driver tegra_nvdec_driver = {
+	.driver = {
+		.name = "tegra-nvdec",
+		.of_match_table = tegra_nvdec_of_match,
+		.pm = &nvdec_pm_ops
+	},
+	.probe = nvdec_probe,
+	.remove = nvdec_remove,
+};
+
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_210_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_210_NVDEC_FIRMWARE);
+#endif
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_186_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_186_NVDEC_FIRMWARE);
+#endif
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_194_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_194_NVDEC_FIRMWARE);
+#endif
diff --git a/drivers/gpu/drm/tegra/nvjpg.c b/drivers/gpu/drm/tegra/nvjpg.c
new file mode 100644
index 000000000..94503fd0d
--- /dev/null
+++ b/drivers/gpu/drm/tegra/nvjpg.c
@@ -0,0 +1,330 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/host1x.h>
+#include <linux/iommu.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+#include "drm.h"
+#include "falcon.h"
+
+struct nvjpg_config {
+	const char *firmware;
+	unsigned int version;
+};
+
+struct nvjpg {
+	struct falcon falcon;
+
+	void __iomem *regs;
+	struct tegra_drm_client client;
+	struct device *dev;
+	struct clk *clk;
+
+	/* Platform configuration */
+	const struct nvjpg_config *config;
+};
+
+static inline struct nvjpg *to_nvjpg(struct tegra_drm_client *client)
+{
+	return container_of(client, struct nvjpg, client);
+}
+
+static int nvjpg_init(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = dev->dev_private;
+	struct nvjpg *nvjpg = to_nvjpg(drm);
+	int err;
+
+	err = host1x_client_iommu_attach(client);
+	if (err < 0 && err != -ENODEV) {
+		dev_err(nvjpg->dev, "failed to attach to domain: %d\n", err);
+		return err;
+	}
+
+	err = tegra_drm_register_client(tegra, drm);
+	if (err < 0)
+		goto detach;
+
+	/*
+	 * Inherit the DMA parameters (such as maximum segment size) from the
+	 * parent host1x device.
+	 */
+	client->dev->dma_parms = client->host->dma_parms;
+
+	return 0;
+
+detach:
+	host1x_client_iommu_detach(client);
+
+	return err;
+}
+
+static int nvjpg_exit(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = dev->dev_private;
+	struct nvjpg *nvjpg = to_nvjpg(drm);
+	int err;
+
+	/* avoid a dangling pointer just in case this disappears */
+	client->dev->dma_parms = NULL;
+
+	err = tegra_drm_unregister_client(tegra, drm);
+	if (err < 0)
+		return err;
+
+	pm_runtime_dont_use_autosuspend(client->dev);
+	pm_runtime_force_suspend(client->dev);
+
+	host1x_client_iommu_detach(client);
+
+	if (client->group) {
+		dma_unmap_single(nvjpg->dev, nvjpg->falcon.firmware.phys,
+				 nvjpg->falcon.firmware.size, DMA_TO_DEVICE);
+		tegra_drm_free(tegra, nvjpg->falcon.firmware.size,
+			       nvjpg->falcon.firmware.virt,
+			       nvjpg->falcon.firmware.iova);
+	} else {
+		dma_free_coherent(nvjpg->dev, nvjpg->falcon.firmware.size,
+				  nvjpg->falcon.firmware.virt,
+				  nvjpg->falcon.firmware.iova);
+	}
+
+	return 0;
+}
+
+static const struct host1x_client_ops nvjpg_client_ops = {
+	.init = nvjpg_init,
+	.exit = nvjpg_exit,
+};
+
+static int nvjpg_load_falcon_firmware(struct nvjpg *nvjpg)
+{
+	struct host1x_client *client = &nvjpg->client.base;
+	struct tegra_drm *tegra = nvjpg->client.drm;
+	dma_addr_t iova;
+	size_t size;
+	void *virt;
+	int err;
+
+	if (nvjpg->falcon.firmware.virt)
+		return 0;
+
+	err = falcon_read_firmware(&nvjpg->falcon, nvjpg->config->firmware);
+	if (err < 0)
+		return err;
+
+	size = nvjpg->falcon.firmware.size;
+
+	if (!client->group) {
+		virt = dma_alloc_coherent(nvjpg->dev, size, &iova, GFP_KERNEL);
+		if (!virt)
+			return -ENOMEM;
+	} else {
+		virt = tegra_drm_alloc(tegra, size, &iova);
+		if (IS_ERR(virt))
+			return PTR_ERR(virt);
+	}
+
+	nvjpg->falcon.firmware.virt = virt;
+	nvjpg->falcon.firmware.iova = iova;
+
+	err = falcon_load_firmware(&nvjpg->falcon);
+	if (err < 0)
+		goto cleanup;
+
+	/*
+	 * In this case we have received an IOVA from the shared domain, so we
+	 * need to make sure to get the physical address so that the DMA API
+	 * knows what memory pages to flush the cache for.
+	 */
+	if (client->group) {
+		dma_addr_t phys;
+
+		phys = dma_map_single(nvjpg->dev, virt, size, DMA_TO_DEVICE);
+
+		err = dma_mapping_error(nvjpg->dev, phys);
+		if (err < 0)
+			goto cleanup;
+
+		nvjpg->falcon.firmware.phys = phys;
+	}
+
+	return 0;
+
+cleanup:
+	if (!client->group)
+		dma_free_coherent(nvjpg->dev, size, virt, iova);
+	else
+		tegra_drm_free(tegra, size, virt, iova);
+
+	return err;
+}
+
+static __maybe_unused int nvjpg_runtime_resume(struct device *dev)
+{
+	struct nvjpg *nvjpg = dev_get_drvdata(dev);
+	int err;
+
+	err = clk_prepare_enable(nvjpg->clk);
+	if (err < 0)
+		return err;
+
+	usleep_range(20, 30);
+
+	err = nvjpg_load_falcon_firmware(nvjpg);
+	if (err < 0)
+		goto disable_clk;
+
+	err = falcon_boot(&nvjpg->falcon);
+	if (err < 0)
+		goto disable_clk;
+
+	return 0;
+
+disable_clk:
+	clk_disable_unprepare(nvjpg->clk);
+	return err;
+}
+
+static __maybe_unused int nvjpg_runtime_suspend(struct device *dev)
+{
+	struct nvjpg *nvjpg = dev_get_drvdata(dev);
+
+	clk_disable_unprepare(nvjpg->clk);
+
+	return 0;
+}
+
+static int nvjpg_can_use_memory_ctx(struct tegra_drm_client *client, bool *supported)
+{
+	*supported = false;
+
+	return 0;
+}
+
+static const struct tegra_drm_client_ops nvjpg_ops = {
+	.get_streamid_offset = NULL,
+	.can_use_memory_ctx = nvjpg_can_use_memory_ctx,
+};
+
+#define NVIDIA_TEGRA_210_NVJPG_FIRMWARE "nvidia/tegra210/nvjpg.bin"
+
+static const struct nvjpg_config tegra210_nvjpg_config = {
+	.firmware = NVIDIA_TEGRA_210_NVJPG_FIRMWARE,
+	.version = 0x21,
+};
+
+static const struct of_device_id tegra_nvjpg_of_match[] = {
+	{ .compatible = "nvidia,tegra210-nvjpg", .data = &tegra210_nvjpg_config },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, tegra_nvjpg_of_match);
+
+static int nvjpg_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct nvjpg *nvjpg;
+	int err;
+
+	/* inherit DMA mask from host1x parent */
+	err = dma_coerce_mask_and_coherent(dev, *dev->parent->dma_mask);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set DMA mask: %d\n", err);
+		return err;
+	}
+
+	nvjpg = devm_kzalloc(dev, sizeof(*nvjpg), GFP_KERNEL);
+	if (!nvjpg)
+		return -ENOMEM;
+
+	nvjpg->config = of_device_get_match_data(dev);
+
+	nvjpg->regs = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(nvjpg->regs))
+		return PTR_ERR(nvjpg->regs);
+
+	nvjpg->clk = devm_clk_get(dev, "nvjpg");
+	if (IS_ERR(nvjpg->clk)) {
+		dev_err(&pdev->dev, "failed to get clock\n");
+		return PTR_ERR(nvjpg->clk);
+	}
+
+	err = clk_set_rate(nvjpg->clk, ULONG_MAX);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set clock rate\n");
+		return err;
+	}
+
+	nvjpg->falcon.dev = dev;
+	nvjpg->falcon.regs = nvjpg->regs;
+
+	err = falcon_init(&nvjpg->falcon);
+	if (err < 0)
+		return err;
+
+	platform_set_drvdata(pdev, nvjpg);
+
+	INIT_LIST_HEAD(&nvjpg->client.base.list);
+	nvjpg->client.base.ops = &nvjpg_client_ops;
+	nvjpg->client.base.dev = dev;
+	nvjpg->client.base.class = HOST1X_CLASS_NVJPG;
+	nvjpg->dev = dev;
+
+	INIT_LIST_HEAD(&nvjpg->client.list);
+	nvjpg->client.version = nvjpg->config->version;
+	nvjpg->client.ops = &nvjpg_ops;
+
+	err = host1x_client_register(&nvjpg->client.base);
+	if (err < 0) {
+		dev_err(dev, "failed to register host1x client: %d\n", err);
+		goto exit_falcon;
+	}
+
+	pm_runtime_use_autosuspend(dev);
+	pm_runtime_set_autosuspend_delay(dev, 500);
+	devm_pm_runtime_enable(dev);
+
+	return 0;
+
+exit_falcon:
+	falcon_exit(&nvjpg->falcon);
+
+	return err;
+}
+
+static void nvjpg_remove(struct platform_device *pdev)
+{
+	struct nvjpg *nvjpg = platform_get_drvdata(pdev);
+
+	host1x_client_unregister(&nvjpg->client.base);
+	falcon_exit(&nvjpg->falcon);
+}
+
+static const struct dev_pm_ops nvjpg_pm_ops = {
+	RUNTIME_PM_OPS(nvjpg_runtime_suspend, nvjpg_runtime_resume, NULL)
+	SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, pm_runtime_force_resume)
+};
+
+struct platform_driver tegra_nvjpg_driver = {
+	.driver = {
+		.name = "tegra-nvjpg",
+		.of_match_table = tegra_nvjpg_of_match,
+		.pm = &nvjpg_pm_ops
+	},
+	.probe = nvjpg_probe,
+	.remove = nvjpg_remove,
+};
+
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_210_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_210_NVJPG_FIRMWARE);
+#endif
diff --git a/drivers/gpu/drm/tegra/output.c b/drivers/gpu/drm/tegra/output.c
new file mode 100644
index 000000000..d0ba3bd0a
--- /dev/null
+++ b/drivers/gpu/drm/tegra/output.c
@@ -0,0 +1,300 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#include <linux/i2c.h>
+#include <linux/of.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_of.h>
+#include <drm/drm_panel.h>
+#include <drm/drm_simple_kms_helper.h>
+
+#include "drm.h"
+#include "dc.h"
+
+#include <media/cec-notifier.h>
+
+int tegra_output_connector_get_modes(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	const struct drm_edid *drm_edid = NULL;
+	int err = 0;
+
+	/*
+	 * If the panel provides one or more modes, use them exclusively and
+	 * ignore any other means of obtaining a mode.
+	 */
+	if (output->panel) {
+		err = drm_panel_get_modes(output->panel, connector);
+		if (err > 0)
+			return err;
+	}
+
+	if (output->drm_edid)
+		drm_edid = drm_edid_dup(output->drm_edid);
+	else if (output->ddc)
+		drm_edid = drm_edid_read_ddc(connector, output->ddc);
+
+	drm_edid_connector_update(connector, drm_edid);
+	cec_notifier_set_phys_addr(output->cec,
+				   connector->display_info.source_physical_address);
+
+	err = drm_edid_connector_add_modes(connector);
+	drm_edid_free(drm_edid);
+
+	return err;
+}
+
+enum drm_connector_status
+tegra_output_connector_detect(struct drm_connector *connector, bool force)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	enum drm_connector_status status = connector_status_unknown;
+
+	if (output->hpd_gpio) {
+		if (gpiod_get_value(output->hpd_gpio) == 0)
+			status = connector_status_disconnected;
+		else
+			status = connector_status_connected;
+	} else {
+		if (!output->panel)
+			status = connector_status_disconnected;
+		else
+			status = connector_status_connected;
+	}
+
+	if (status != connector_status_connected)
+		cec_notifier_phys_addr_invalidate(output->cec);
+
+	return status;
+}
+
+void tegra_output_connector_destroy(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+
+	if (output->cec)
+		cec_notifier_conn_unregister(output->cec);
+
+	drm_connector_unregister(connector);
+	drm_connector_cleanup(connector);
+}
+
+static irqreturn_t hpd_irq(int irq, void *data)
+{
+	struct tegra_output *output = data;
+
+	if (output->connector.dev)
+		drm_helper_hpd_irq_event(output->connector.dev);
+
+	return IRQ_HANDLED;
+}
+
+int tegra_output_probe(struct tegra_output *output)
+{
+	struct device_node *ddc, *panel;
+	const void *edid;
+	unsigned long flags;
+	int err, size;
+
+	if (!output->of_node)
+		output->of_node = output->dev->of_node;
+
+	err = drm_of_find_panel_or_bridge(output->of_node, -1, -1,
+					  &output->panel, &output->bridge);
+	if (err && err != -ENODEV)
+		return err;
+
+	panel = of_parse_phandle(output->of_node, "nvidia,panel", 0);
+	if (panel) {
+		/*
+		 * Don't mix nvidia,panel phandle with the graph in a
+		 * device-tree.
+		 */
+		WARN_ON(output->panel || output->bridge);
+
+		if (output->panel) {
+			drm_panel_put(output->panel);
+			output->panel = NULL;
+		}
+
+		output->panel = of_drm_find_panel(panel);
+		of_node_put(panel);
+
+		if (IS_ERR(output->panel)) {
+			err = PTR_ERR(output->panel);
+			output->panel = NULL;
+			return err;
+		}
+	}
+
+	ddc = of_parse_phandle(output->of_node, "nvidia,ddc-i2c-bus", 0);
+	if (ddc) {
+		output->ddc = of_get_i2c_adapter_by_node(ddc);
+		of_node_put(ddc);
+
+		if (!output->ddc) {
+			err = -EPROBE_DEFER;
+			goto put_i2c;
+		}
+	}
+
+	edid = of_get_property(output->of_node, "nvidia,edid", &size);
+	output->drm_edid = drm_edid_alloc(edid, size);
+
+	output->hpd_gpio = devm_fwnode_gpiod_get(output->dev,
+					of_fwnode_handle(output->of_node),
+					"nvidia,hpd",
+					GPIOD_IN,
+					"HDMI hotplug detect");
+	if (IS_ERR(output->hpd_gpio)) {
+		if (PTR_ERR(output->hpd_gpio) != -ENOENT) {
+			err = PTR_ERR(output->hpd_gpio);
+			goto put_i2c;
+		}
+
+		output->hpd_gpio = NULL;
+	}
+
+	if (output->hpd_gpio) {
+		err = gpiod_to_irq(output->hpd_gpio);
+		if (err < 0) {
+			dev_err(output->dev, "gpiod_to_irq(): %d\n", err);
+			goto put_i2c;
+		}
+
+		output->hpd_irq = err;
+
+		flags = IRQF_TRIGGER_RISING | IRQF_TRIGGER_FALLING |
+			IRQF_ONESHOT;
+
+		err = request_threaded_irq(output->hpd_irq, NULL, hpd_irq,
+					   flags, "hpd", output);
+		if (err < 0) {
+			dev_err(output->dev, "failed to request IRQ#%u: %d\n",
+				output->hpd_irq, err);
+			goto put_i2c;
+		}
+
+		output->connector.polled = DRM_CONNECTOR_POLL_HPD;
+
+		/*
+		 * Disable the interrupt until the connector has been
+		 * initialized to avoid a race in the hotplug interrupt
+		 * handler.
+		 */
+		disable_irq(output->hpd_irq);
+	}
+
+	return 0;
+
+put_i2c:
+	if (output->panel) {
+		drm_panel_put(output->panel);
+		output->panel = NULL;
+	}
+
+	if (output->ddc)
+		i2c_put_adapter(output->ddc);
+
+	drm_edid_free(output->drm_edid);
+
+	return err;
+}
+
+void tegra_output_remove(struct tegra_output *output)
+{
+	if (output->panel) {
+		drm_panel_put(output->panel);
+		output->panel = NULL;
+	}
+
+	if (output->hpd_gpio)
+		free_irq(output->hpd_irq, output);
+
+	if (output->ddc)
+		i2c_put_adapter(output->ddc);
+
+	drm_edid_free(output->drm_edid);
+}
+
+int tegra_output_init(struct drm_device *drm, struct tegra_output *output)
+{
+	int connector_type;
+
+	/*
+	 * The connector is now registered and ready to receive hotplug events
+	 * so the hotplug interrupt can be enabled.
+	 */
+	if (output->hpd_gpio)
+		enable_irq(output->hpd_irq);
+
+	connector_type = output->connector.connector_type;
+	/*
+	 * Create a CEC notifier for HDMI connector.
+	 */
+	if (connector_type == DRM_MODE_CONNECTOR_HDMIA ||
+	    connector_type == DRM_MODE_CONNECTOR_HDMIB) {
+		struct cec_connector_info conn_info;
+
+		cec_fill_conn_info_from_drm(&conn_info, &output->connector);
+		output->cec = cec_notifier_conn_register(output->dev, NULL,
+							 &conn_info);
+		if (!output->cec)
+			return -ENOMEM;
+	}
+
+	return 0;
+}
+
+void tegra_output_exit(struct tegra_output *output)
+{
+	/*
+	 * The connector is going away, so the interrupt must be disabled to
+	 * prevent the hotplug interrupt handler from potentially crashing.
+	 */
+	if (output->hpd_gpio)
+		disable_irq(output->hpd_irq);
+}
+
+void tegra_output_find_possible_crtcs(struct tegra_output *output,
+				      struct drm_device *drm)
+{
+	struct device *dev = output->dev;
+	struct drm_crtc *crtc;
+	unsigned int mask = 0;
+
+	drm_for_each_crtc(crtc, drm) {
+		struct tegra_dc *dc = to_tegra_dc(crtc);
+
+		if (tegra_dc_has_output(dc, dev))
+			mask |= drm_crtc_mask(crtc);
+	}
+
+	if (mask == 0) {
+		dev_warn(dev, "missing output definition for heads in DT\n");
+		mask = 0x3;
+	}
+
+	output->encoder.possible_crtcs = mask;
+}
+
+int tegra_output_suspend(struct tegra_output *output)
+{
+	if (output->hpd_irq)
+		disable_irq(output->hpd_irq);
+
+	return 0;
+}
+
+int tegra_output_resume(struct tegra_output *output)
+{
+	if (output->hpd_irq)
+		enable_irq(output->hpd_irq);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/tegra/plane.c b/drivers/gpu/drm/tegra/plane.c
new file mode 100644
index 000000000..0cb309107
--- /dev/null
+++ b/drivers/gpu/drm/tegra/plane.c
@@ -0,0 +1,794 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2017 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#include <linux/dma-mapping.h>
+#include <linux/iommu.h>
+#include <linux/interconnect.h>
+
+#include <drm/drm_atomic.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_fourcc.h>
+#include <drm/drm_framebuffer.h>
+#include <drm/drm_gem_atomic_helper.h>
+
+#include "dc.h"
+#include "plane.h"
+
+static void tegra_plane_destroy(struct drm_plane *plane)
+{
+	struct tegra_plane *p = to_tegra_plane(plane);
+
+	drm_plane_cleanup(plane);
+	kfree(p);
+}
+
+static void tegra_plane_reset(struct drm_plane *plane)
+{
+	struct tegra_plane *p = to_tegra_plane(plane);
+	struct tegra_plane_state *state;
+	unsigned int i;
+
+	if (plane->state)
+		__drm_atomic_helper_plane_destroy_state(plane->state);
+
+	kfree(plane->state);
+	plane->state = NULL;
+
+	state = kzalloc_obj(*state);
+	if (state) {
+		plane->state = &state->base;
+		plane->state->plane = plane;
+		plane->state->zpos = p->index;
+		plane->state->normalized_zpos = p->index;
+
+		for (i = 0; i < 3; i++)
+			state->iova[i] = DMA_MAPPING_ERROR;
+	}
+}
+
+static struct drm_plane_state *
+tegra_plane_atomic_duplicate_state(struct drm_plane *plane)
+{
+	struct tegra_plane_state *state = to_tegra_plane_state(plane->state);
+	struct tegra_plane_state *copy;
+	unsigned int i;
+
+	copy = kmalloc_obj(*copy);
+	if (!copy)
+		return NULL;
+
+	__drm_atomic_helper_plane_duplicate_state(plane, &copy->base);
+	copy->tiling = state->tiling;
+	copy->format = state->format;
+	copy->swap = state->swap;
+	copy->reflect_x = state->reflect_x;
+	copy->reflect_y = state->reflect_y;
+	copy->opaque = state->opaque;
+	copy->total_peak_memory_bandwidth = state->total_peak_memory_bandwidth;
+	copy->peak_memory_bandwidth = state->peak_memory_bandwidth;
+	copy->avg_memory_bandwidth = state->avg_memory_bandwidth;
+
+	for (i = 0; i < 2; i++)
+		copy->blending[i] = state->blending[i];
+
+	for (i = 0; i < 3; i++) {
+		copy->iova[i] = DMA_MAPPING_ERROR;
+		copy->map[i] = NULL;
+	}
+
+	return &copy->base;
+}
+
+static void tegra_plane_atomic_destroy_state(struct drm_plane *plane,
+					     struct drm_plane_state *state)
+{
+	__drm_atomic_helper_plane_destroy_state(state);
+	kfree(state);
+}
+
+static bool tegra_plane_supports_sector_layout(struct drm_plane *plane)
+{
+	struct drm_crtc *crtc;
+
+	drm_for_each_crtc(crtc, plane->dev) {
+		if (plane->possible_crtcs & drm_crtc_mask(crtc)) {
+			struct tegra_dc *dc = to_tegra_dc(crtc);
+
+			if (!dc->soc->supports_sector_layout)
+				return false;
+		}
+	}
+
+	return true;
+}
+
+static bool tegra_plane_format_mod_supported(struct drm_plane *plane,
+					     uint32_t format,
+					     uint64_t modifier)
+{
+	const struct drm_format_info *info = drm_format_info(format);
+
+	if (modifier == DRM_FORMAT_MOD_LINEAR)
+		return true;
+
+	/* check for the sector layout bit */
+	if (fourcc_mod_is_vendor(modifier, NVIDIA)) {
+		if (modifier & DRM_FORMAT_MOD_NVIDIA_SECTOR_LAYOUT) {
+			if (!tegra_plane_supports_sector_layout(plane))
+				return false;
+		}
+	}
+
+	if (info->num_planes == 1)
+		return true;
+
+	return false;
+}
+
+const struct drm_plane_funcs tegra_plane_funcs = {
+	.update_plane = drm_atomic_helper_update_plane,
+	.disable_plane = drm_atomic_helper_disable_plane,
+	.destroy = tegra_plane_destroy,
+	.reset = tegra_plane_reset,
+	.atomic_duplicate_state = tegra_plane_atomic_duplicate_state,
+	.atomic_destroy_state = tegra_plane_atomic_destroy_state,
+	.format_mod_supported = tegra_plane_format_mod_supported,
+};
+
+static int tegra_dc_pin(struct tegra_dc *dc, struct tegra_plane_state *state)
+{
+	unsigned int i;
+	int err;
+
+	for (i = 0; i < state->base.fb->format->num_planes; i++) {
+		struct tegra_bo *bo = tegra_fb_get_plane(state->base.fb, i);
+		struct host1x_bo_mapping *map;
+
+		map = host1x_bo_pin(dc->dev, &bo->base, DMA_TO_DEVICE, &dc->client.cache);
+		if (IS_ERR(map)) {
+			err = PTR_ERR(map);
+			goto unpin;
+		}
+
+		if (!dc->client.group) {
+			/*
+			 * The display controller needs contiguous memory, so
+			 * fail if the buffer is discontiguous and we fail to
+			 * map its SG table to a single contiguous chunk of
+			 * I/O virtual memory.
+			 */
+			if (map->chunks > 1) {
+				err = -EINVAL;
+				goto unpin;
+			}
+
+			state->iova[i] = map->phys;
+		} else {
+			state->iova[i] = bo->iova;
+		}
+
+		state->map[i] = map;
+	}
+
+	return 0;
+
+unpin:
+	dev_err(dc->dev, "failed to map plane %u: %d\n", i, err);
+
+	while (i--) {
+		host1x_bo_unpin(state->map[i]);
+		state->iova[i] = DMA_MAPPING_ERROR;
+		state->map[i] = NULL;
+	}
+
+	return err;
+}
+
+static void tegra_dc_unpin(struct tegra_dc *dc, struct tegra_plane_state *state)
+{
+	unsigned int i;
+
+	for (i = 0; i < state->base.fb->format->num_planes; i++) {
+		host1x_bo_unpin(state->map[i]);
+		state->iova[i] = DMA_MAPPING_ERROR;
+		state->map[i] = NULL;
+	}
+}
+
+int tegra_plane_prepare_fb(struct drm_plane *plane,
+			   struct drm_plane_state *state)
+{
+	struct tegra_dc *dc = to_tegra_dc(state->crtc);
+	int err;
+
+	if (!state->fb)
+		return 0;
+
+	err = drm_gem_plane_helper_prepare_fb(plane, state);
+	if (err < 0)
+		return err;
+
+	return tegra_dc_pin(dc, to_tegra_plane_state(state));
+}
+
+void tegra_plane_cleanup_fb(struct drm_plane *plane,
+			    struct drm_plane_state *state)
+{
+	struct tegra_dc *dc = to_tegra_dc(state->crtc);
+
+	if (dc)
+		tegra_dc_unpin(dc, to_tegra_plane_state(state));
+}
+
+static int tegra_plane_calculate_memory_bandwidth(struct drm_plane_state *state)
+{
+	struct tegra_plane_state *tegra_state = to_tegra_plane_state(state);
+	unsigned int i, bpp, dst_w, dst_h, src_w, src_h, mul;
+	const struct tegra_dc_soc_info *soc;
+	const struct drm_format_info *fmt;
+	struct drm_crtc_state *crtc_state;
+	u64 avg_bandwidth, peak_bandwidth;
+
+	if (!state->visible)
+		return 0;
+
+	crtc_state = drm_atomic_get_new_crtc_state(state->state, state->crtc);
+	if (!crtc_state)
+		return -EINVAL;
+
+	src_w = drm_rect_width(&state->src) >> 16;
+	src_h = drm_rect_height(&state->src) >> 16;
+	dst_w = drm_rect_width(&state->dst);
+	dst_h = drm_rect_height(&state->dst);
+
+	fmt = state->fb->format;
+	soc = to_tegra_dc(state->crtc)->soc;
+
+	/*
+	 * Note that real memory bandwidth vary depending on format and
+	 * memory layout, we are not taking that into account because small
+	 * estimation error isn't important since bandwidth is rounded up
+	 * anyway.
+	 */
+	for (i = 0, bpp = 0; i < fmt->num_planes; i++) {
+		unsigned int bpp_plane = fmt->cpp[i] * 8;
+
+		/*
+		 * Sub-sampling is relevant for chroma planes only and vertical
+		 * readouts are not cached, hence only horizontal sub-sampling
+		 * matters.
+		 */
+		if (i > 0)
+			bpp_plane /= fmt->hsub;
+
+		bpp += bpp_plane;
+	}
+
+	/* average bandwidth in kbytes/sec */
+	avg_bandwidth  = min(src_w, dst_w) * min(src_h, dst_h);
+	avg_bandwidth *= drm_mode_vrefresh(&crtc_state->adjusted_mode);
+	avg_bandwidth  = DIV_ROUND_UP(avg_bandwidth * bpp, 8) + 999;
+	do_div(avg_bandwidth, 1000);
+
+	/* mode.clock in kHz, peak bandwidth in kbytes/sec */
+	peak_bandwidth = DIV_ROUND_UP(crtc_state->adjusted_mode.clock * bpp, 8);
+
+	/*
+	 * Tegra30/114 Memory Controller can't interleave DC memory requests
+	 * for the tiled windows because DC uses 16-bytes atom, while DDR3
+	 * uses 32-bytes atom.  Hence there is x2 memory overfetch for tiled
+	 * framebuffer and DDR3 on these SoCs.
+	 */
+	if (soc->plane_tiled_memory_bandwidth_x2 &&
+	    tegra_state->tiling.mode == TEGRA_BO_TILING_MODE_TILED)
+		mul = 2;
+	else
+		mul = 1;
+
+	/* ICC bandwidth in kbytes/sec */
+	tegra_state->peak_memory_bandwidth = kBps_to_icc(peak_bandwidth) * mul;
+	tegra_state->avg_memory_bandwidth  = kBps_to_icc(avg_bandwidth)  * mul;
+
+	return 0;
+}
+
+int tegra_plane_state_add(struct tegra_plane *plane,
+			  struct drm_plane_state *state)
+{
+	struct drm_crtc_state *crtc_state;
+	struct tegra_dc_state *tegra;
+	int err;
+
+	/* Propagate errors from allocation or locking failures. */
+	crtc_state = drm_atomic_get_crtc_state(state->state, state->crtc);
+	if (IS_ERR(crtc_state))
+		return PTR_ERR(crtc_state);
+
+	/* Check plane state for visibility and calculate clipping bounds */
+	err = drm_atomic_helper_check_plane_state(state, crtc_state,
+						  0, INT_MAX, true, true);
+	if (err < 0)
+		return err;
+
+	err = tegra_plane_calculate_memory_bandwidth(state);
+	if (err < 0)
+		return err;
+
+	tegra = to_dc_state(crtc_state);
+
+	tegra->planes |= WIN_A_ACT_REQ << plane->index;
+
+	return 0;
+}
+
+int tegra_plane_format(u32 fourcc, u32 *format, u32 *swap)
+{
+	/* assume no swapping of fetched data */
+	if (swap)
+		*swap = BYTE_SWAP_NOSWAP;
+
+	switch (fourcc) {
+	case DRM_FORMAT_ARGB4444:
+		*format = WIN_COLOR_DEPTH_B4G4R4A4;
+		break;
+
+	case DRM_FORMAT_ARGB1555:
+		*format = WIN_COLOR_DEPTH_B5G5R5A1;
+		break;
+
+	case DRM_FORMAT_RGB565:
+		*format = WIN_COLOR_DEPTH_B5G6R5;
+		break;
+
+	case DRM_FORMAT_RGBA5551:
+		*format = WIN_COLOR_DEPTH_A1B5G5R5;
+		break;
+
+	case DRM_FORMAT_ARGB8888:
+		*format = WIN_COLOR_DEPTH_B8G8R8A8;
+		break;
+
+	case DRM_FORMAT_ABGR8888:
+		*format = WIN_COLOR_DEPTH_R8G8B8A8;
+		break;
+
+	case DRM_FORMAT_ABGR4444:
+		*format = WIN_COLOR_DEPTH_R4G4B4A4;
+		break;
+
+	case DRM_FORMAT_ABGR1555:
+		*format = WIN_COLOR_DEPTH_R5G5B5A;
+		break;
+
+	case DRM_FORMAT_BGRA5551:
+		*format = WIN_COLOR_DEPTH_AR5G5B5;
+		break;
+
+	case DRM_FORMAT_XRGB1555:
+		*format = WIN_COLOR_DEPTH_B5G5R5X1;
+		break;
+
+	case DRM_FORMAT_RGBX5551:
+		*format = WIN_COLOR_DEPTH_X1B5G5R5;
+		break;
+
+	case DRM_FORMAT_XBGR1555:
+		*format = WIN_COLOR_DEPTH_R5G5B5X1;
+		break;
+
+	case DRM_FORMAT_BGRX5551:
+		*format = WIN_COLOR_DEPTH_X1R5G5B5;
+		break;
+
+	case DRM_FORMAT_BGR565:
+		*format = WIN_COLOR_DEPTH_R5G6B5;
+		break;
+
+	case DRM_FORMAT_BGRA8888:
+		*format = WIN_COLOR_DEPTH_A8R8G8B8;
+		break;
+
+	case DRM_FORMAT_RGBA8888:
+		*format = WIN_COLOR_DEPTH_A8B8G8R8;
+		break;
+
+	case DRM_FORMAT_XRGB8888:
+		*format = WIN_COLOR_DEPTH_B8G8R8X8;
+		break;
+
+	case DRM_FORMAT_XBGR8888:
+		*format = WIN_COLOR_DEPTH_R8G8B8X8;
+		break;
+
+	case DRM_FORMAT_UYVY:
+		*format = WIN_COLOR_DEPTH_YCbCr422;
+		break;
+
+	case DRM_FORMAT_YUYV:
+		if (!swap)
+			return -EINVAL;
+
+		*format = WIN_COLOR_DEPTH_YCbCr422;
+		*swap = BYTE_SWAP_SWAP2;
+		break;
+
+	case DRM_FORMAT_YVYU:
+		if (!swap)
+			return -EINVAL;
+
+		*format = WIN_COLOR_DEPTH_YCbCr422;
+		*swap = BYTE_SWAP_SWAP4;
+		break;
+
+	case DRM_FORMAT_VYUY:
+		if (!swap)
+			return -EINVAL;
+
+		*format = WIN_COLOR_DEPTH_YCbCr422;
+		*swap = BYTE_SWAP_SWAP4HW;
+		break;
+
+	case DRM_FORMAT_YUV420:
+		*format = WIN_COLOR_DEPTH_YCbCr420P;
+		break;
+
+	case DRM_FORMAT_YUV422:
+		*format = WIN_COLOR_DEPTH_YCbCr422P;
+		break;
+
+	case DRM_FORMAT_YUV444:
+		*format = WIN_COLOR_DEPTH_YCbCr444P;
+		break;
+
+	case DRM_FORMAT_NV12:
+		*format = WIN_COLOR_DEPTH_YCbCr420SP;
+		break;
+
+	case DRM_FORMAT_NV21:
+		*format = WIN_COLOR_DEPTH_YCrCb420SP;
+		break;
+
+	case DRM_FORMAT_NV16:
+		*format = WIN_COLOR_DEPTH_YCbCr422SP;
+		break;
+
+	case DRM_FORMAT_NV61:
+		*format = WIN_COLOR_DEPTH_YCrCb422SP;
+		break;
+
+	case DRM_FORMAT_NV24:
+		*format = WIN_COLOR_DEPTH_YCbCr444SP;
+		break;
+
+	case DRM_FORMAT_NV42:
+		*format = WIN_COLOR_DEPTH_YCrCb444SP;
+		break;
+
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+bool tegra_plane_format_is_indexed(unsigned int format)
+{
+	switch (format) {
+	case WIN_COLOR_DEPTH_P1:
+	case WIN_COLOR_DEPTH_P2:
+	case WIN_COLOR_DEPTH_P4:
+	case WIN_COLOR_DEPTH_P8:
+		return true;
+	}
+
+	return false;
+}
+
+bool tegra_plane_format_is_yuv(unsigned int format, unsigned int *planes, unsigned int *bpc)
+{
+	switch (format) {
+	case WIN_COLOR_DEPTH_YCbCr422:
+	case WIN_COLOR_DEPTH_YUV422:
+		if (planes)
+			*planes = 1;
+
+		if (bpc)
+			*bpc = 8;
+
+		return true;
+
+	case WIN_COLOR_DEPTH_YCbCr420P:
+	case WIN_COLOR_DEPTH_YUV420P:
+	case WIN_COLOR_DEPTH_YCbCr422P:
+	case WIN_COLOR_DEPTH_YUV422P:
+	case WIN_COLOR_DEPTH_YCbCr422R:
+	case WIN_COLOR_DEPTH_YUV422R:
+	case WIN_COLOR_DEPTH_YCbCr422RA:
+	case WIN_COLOR_DEPTH_YUV422RA:
+	case WIN_COLOR_DEPTH_YCbCr444P:
+		if (planes)
+			*planes = 3;
+
+		if (bpc)
+			*bpc = 8;
+
+		return true;
+
+	case WIN_COLOR_DEPTH_YCrCb420SP:
+	case WIN_COLOR_DEPTH_YCbCr420SP:
+	case WIN_COLOR_DEPTH_YCrCb422SP:
+	case WIN_COLOR_DEPTH_YCbCr422SP:
+	case WIN_COLOR_DEPTH_YCrCb444SP:
+	case WIN_COLOR_DEPTH_YCbCr444SP:
+		if (planes)
+			*planes = 2;
+
+		if (bpc)
+			*bpc = 8;
+
+		return true;
+	}
+
+	if (planes)
+		*planes = 1;
+
+	return false;
+}
+
+static bool __drm_format_has_alpha(u32 format)
+{
+	switch (format) {
+	case DRM_FORMAT_ARGB1555:
+	case DRM_FORMAT_RGBA5551:
+	case DRM_FORMAT_ABGR8888:
+	case DRM_FORMAT_ARGB8888:
+		return true;
+	}
+
+	return false;
+}
+
+static int tegra_plane_format_get_alpha(unsigned int opaque,
+					unsigned int *alpha)
+{
+	if (tegra_plane_format_is_yuv(opaque, NULL, NULL)) {
+		*alpha = opaque;
+		return 0;
+	}
+
+	switch (opaque) {
+	case WIN_COLOR_DEPTH_B5G5R5X1:
+		*alpha = WIN_COLOR_DEPTH_B5G5R5A1;
+		return 0;
+
+	case WIN_COLOR_DEPTH_X1B5G5R5:
+		*alpha = WIN_COLOR_DEPTH_A1B5G5R5;
+		return 0;
+
+	case WIN_COLOR_DEPTH_R8G8B8X8:
+		*alpha = WIN_COLOR_DEPTH_R8G8B8A8;
+		return 0;
+
+	case WIN_COLOR_DEPTH_B8G8R8X8:
+		*alpha = WIN_COLOR_DEPTH_B8G8R8A8;
+		return 0;
+
+	case WIN_COLOR_DEPTH_B5G6R5:
+		*alpha = opaque;
+		return 0;
+	}
+
+	return -EINVAL;
+}
+
+/*
+ * This is applicable to Tegra20 and Tegra30 only where the opaque formats can
+ * be emulated using the alpha formats and alpha blending disabled.
+ */
+static int tegra_plane_setup_opacity(struct tegra_plane *tegra,
+				     struct tegra_plane_state *state)
+{
+	unsigned int format;
+	int err;
+
+	switch (state->format) {
+	case WIN_COLOR_DEPTH_B5G5R5A1:
+	case WIN_COLOR_DEPTH_A1B5G5R5:
+	case WIN_COLOR_DEPTH_R8G8B8A8:
+	case WIN_COLOR_DEPTH_B8G8R8A8:
+		state->opaque = false;
+		break;
+
+	default:
+		err = tegra_plane_format_get_alpha(state->format, &format);
+		if (err < 0)
+			return err;
+
+		state->format = format;
+		state->opaque = true;
+		break;
+	}
+
+	return 0;
+}
+
+static int tegra_plane_check_transparency(struct tegra_plane *tegra,
+					  struct tegra_plane_state *state)
+{
+	struct drm_plane_state *old, *plane_state;
+	struct drm_plane *plane;
+
+	old = drm_atomic_get_old_plane_state(state->base.state, &tegra->base);
+
+	/* check if zpos / transparency changed */
+	if (old->normalized_zpos == state->base.normalized_zpos &&
+	    to_tegra_plane_state(old)->opaque == state->opaque)
+		return 0;
+
+	/* include all sibling planes into this commit */
+	drm_for_each_plane(plane, tegra->base.dev) {
+		struct tegra_plane *p = to_tegra_plane(plane);
+
+		/* skip this plane and planes on different CRTCs */
+		if (p == tegra || p->dc != tegra->dc)
+			continue;
+
+		plane_state = drm_atomic_get_plane_state(state->base.state,
+							 plane);
+		if (IS_ERR(plane_state))
+			return PTR_ERR(plane_state);
+	}
+
+	return 1;
+}
+
+static unsigned int tegra_plane_get_overlap_index(struct tegra_plane *plane,
+						  struct tegra_plane *other)
+{
+	unsigned int index = 0, i;
+
+	WARN_ON(plane == other);
+
+	for (i = 0; i < 3; i++) {
+		if (i == plane->index)
+			continue;
+
+		if (i == other->index)
+			break;
+
+		index++;
+	}
+
+	return index;
+}
+
+static void tegra_plane_update_transparency(struct tegra_plane *tegra,
+					    struct tegra_plane_state *state)
+{
+	struct drm_plane_state *new;
+	struct drm_plane *plane;
+	unsigned int i;
+
+	for_each_new_plane_in_state(state->base.state, plane, new, i) {
+		struct tegra_plane *p = to_tegra_plane(plane);
+		unsigned index;
+
+		/* skip this plane and planes on different CRTCs */
+		if (p == tegra || p->dc != tegra->dc)
+			continue;
+
+		index = tegra_plane_get_overlap_index(tegra, p);
+
+		if (new->fb && __drm_format_has_alpha(new->fb->format->format))
+			state->blending[index].alpha = true;
+		else
+			state->blending[index].alpha = false;
+
+		if (new->normalized_zpos > state->base.normalized_zpos)
+			state->blending[index].top = true;
+		else
+			state->blending[index].top = false;
+
+		/*
+		 * Missing framebuffer means that plane is disabled, in this
+		 * case mark B / C window as top to be able to differentiate
+		 * windows indices order in regards to zPos for the middle
+		 * window X / Y registers programming.
+		 */
+		if (!new->fb)
+			state->blending[index].top = (index == 1);
+	}
+}
+
+static int tegra_plane_setup_transparency(struct tegra_plane *tegra,
+					  struct tegra_plane_state *state)
+{
+	struct tegra_plane_state *tegra_state;
+	struct drm_plane_state *new;
+	struct drm_plane *plane;
+	int err;
+
+	/*
+	 * If planes zpos / transparency changed, sibling planes blending
+	 * state may require adjustment and in this case they will be included
+	 * into this atom commit, otherwise blending state is unchanged.
+	 */
+	err = tegra_plane_check_transparency(tegra, state);
+	if (err <= 0)
+		return err;
+
+	/*
+	 * All planes are now in the atomic state, walk them up and update
+	 * transparency state for each plane.
+	 */
+	drm_for_each_plane(plane, tegra->base.dev) {
+		struct tegra_plane *p = to_tegra_plane(plane);
+
+		/* skip planes on different CRTCs */
+		if (p->dc != tegra->dc)
+			continue;
+
+		new = drm_atomic_get_new_plane_state(state->base.state, plane);
+		tegra_state = to_tegra_plane_state(new);
+
+		/*
+		 * There is no need to update blending state for the disabled
+		 * plane.
+		 */
+		if (new->fb)
+			tegra_plane_update_transparency(p, tegra_state);
+	}
+
+	return 0;
+}
+
+int tegra_plane_setup_legacy_state(struct tegra_plane *tegra,
+				   struct tegra_plane_state *state)
+{
+	int err;
+
+	err = tegra_plane_setup_opacity(tegra, state);
+	if (err < 0)
+		return err;
+
+	err = tegra_plane_setup_transparency(tegra, state);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static const char * const tegra_plane_icc_names[TEGRA_DC_LEGACY_PLANES_NUM] = {
+	"wina", "winb", "winc", NULL, NULL, NULL, "cursor",
+};
+
+int tegra_plane_interconnect_init(struct tegra_plane *plane)
+{
+	const char *icc_name = tegra_plane_icc_names[plane->index];
+	struct device *dev = plane->dc->dev;
+	struct tegra_dc *dc = plane->dc;
+	int err;
+
+	if (WARN_ON(plane->index >= TEGRA_DC_LEGACY_PLANES_NUM) ||
+	    WARN_ON(!tegra_plane_icc_names[plane->index]))
+		return -EINVAL;
+
+	plane->icc_mem = devm_of_icc_get(dev, icc_name);
+	err = PTR_ERR_OR_ZERO(plane->icc_mem);
+	if (err)
+		return dev_err_probe(dev, err, "failed to get %s interconnect\n",
+				     icc_name);
+
+	/* plane B on T20/30 has a dedicated memory client for a 6-tap vertical filter */
+	if (plane->index == 1 && dc->soc->has_win_b_vfilter_mem_client) {
+		plane->icc_mem_vfilter = devm_of_icc_get(dev, "winb-vfilter");
+		err = PTR_ERR_OR_ZERO(plane->icc_mem_vfilter);
+		if (err)
+			return dev_err_probe(dev, err, "failed to get %s interconnect\n",
+					     "winb-vfilter");
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/tegra/plane.h b/drivers/gpu/drm/tegra/plane.h
new file mode 100644
index 000000000..e33a581e6
--- /dev/null
+++ b/drivers/gpu/drm/tegra/plane.h
@@ -0,0 +1,98 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2017 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#ifndef TEGRA_PLANE_H
+#define TEGRA_PLANE_H 1
+
+#include <drm/drm_plane.h>
+
+struct icc_path;
+struct tegra_bo;
+struct tegra_dc;
+
+struct tegra_plane {
+	struct drm_plane base;
+	struct tegra_dc *dc;
+	unsigned int offset;
+	unsigned int index;
+
+	struct icc_path *icc_mem;
+	struct icc_path *icc_mem_vfilter;
+};
+
+struct tegra_cursor {
+	struct tegra_plane base;
+
+	struct tegra_bo *bo;
+	unsigned int width;
+	unsigned int height;
+};
+
+static inline struct tegra_plane *to_tegra_plane(struct drm_plane *plane)
+{
+	return container_of(plane, struct tegra_plane, base);
+}
+
+struct tegra_plane_legacy_blending_state {
+	bool alpha;
+	bool top;
+};
+
+struct tegra_plane_state {
+	struct drm_plane_state base;
+
+	struct host1x_bo_mapping *map[3];
+	dma_addr_t iova[3];
+
+	struct tegra_bo_tiling tiling;
+	u32 format;
+	u32 swap;
+
+	bool reflect_x;
+	bool reflect_y;
+
+	/* used for legacy blending support only */
+	struct tegra_plane_legacy_blending_state blending[2];
+	bool opaque;
+
+	/* bandwidths are in ICC units, i.e. kbytes/sec */
+	u32 total_peak_memory_bandwidth;
+	u32 peak_memory_bandwidth;
+	u32 avg_memory_bandwidth;
+};
+
+static inline struct tegra_plane_state *
+to_tegra_plane_state(struct drm_plane_state *state)
+{
+	if (state)
+		return container_of(state, struct tegra_plane_state, base);
+
+	return NULL;
+}
+
+static inline const struct tegra_plane_state *
+to_const_tegra_plane_state(const struct drm_plane_state *state)
+{
+	return to_tegra_plane_state((struct drm_plane_state *)state);
+}
+
+extern const struct drm_plane_funcs tegra_plane_funcs;
+
+int tegra_plane_prepare_fb(struct drm_plane *plane,
+			   struct drm_plane_state *state);
+void tegra_plane_cleanup_fb(struct drm_plane *plane,
+			    struct drm_plane_state *state);
+
+int tegra_plane_state_add(struct tegra_plane *plane,
+			  struct drm_plane_state *state);
+
+int tegra_plane_format(u32 fourcc, u32 *format, u32 *swap);
+bool tegra_plane_format_is_indexed(unsigned int format);
+bool tegra_plane_format_is_yuv(unsigned int format, unsigned int *planes, unsigned int *bpc);
+int tegra_plane_setup_legacy_state(struct tegra_plane *tegra,
+				   struct tegra_plane_state *state);
+int tegra_plane_interconnect_init(struct tegra_plane *plane);
+
+#endif /* TEGRA_PLANE_H */
diff --git a/drivers/gpu/drm/tegra/rgb.c b/drivers/gpu/drm/tegra/rgb.c
new file mode 100644
index 000000000..e67fbb236
--- /dev/null
+++ b/drivers/gpu/drm/tegra/rgb.c
@@ -0,0 +1,382 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2012 Avionic Design GmbH
+ * Copyright (C) 2012 NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#include <linux/clk.h>
+#include <linux/of.h>
+
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_bridge_connector.h>
+#include <drm/drm_simple_kms_helper.h>
+
+#include "drm.h"
+#include "dc.h"
+
+struct tegra_rgb {
+	struct tegra_output output;
+	struct tegra_dc *dc;
+
+	struct clk *pll_d_out0;
+	struct clk *pll_d2_out0;
+	struct clk *clk_parent;
+	struct clk *clk;
+};
+
+static inline struct tegra_rgb *to_rgb(struct tegra_output *output)
+{
+	return container_of(output, struct tegra_rgb, output);
+}
+
+struct reg_entry {
+	unsigned long offset;
+	unsigned long value;
+};
+
+static const struct reg_entry rgb_enable[] = {
+	{ DC_COM_PIN_OUTPUT_ENABLE(0),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_ENABLE(1),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_ENABLE(2),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_ENABLE(3),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_POLARITY(0), 0x00000000 },
+	{ DC_COM_PIN_OUTPUT_POLARITY(1), 0x01000000 },
+	{ DC_COM_PIN_OUTPUT_POLARITY(2), 0x00000000 },
+	{ DC_COM_PIN_OUTPUT_POLARITY(3), 0x00000000 },
+	{ DC_COM_PIN_OUTPUT_DATA(0),     0x00000000 },
+	{ DC_COM_PIN_OUTPUT_DATA(1),     0x00000000 },
+	{ DC_COM_PIN_OUTPUT_DATA(2),     0x00000000 },
+	{ DC_COM_PIN_OUTPUT_DATA(3),     0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(0),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(1),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(2),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(3),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(4),   0x00210222 },
+	{ DC_COM_PIN_OUTPUT_SELECT(5),   0x00002200 },
+	{ DC_COM_PIN_OUTPUT_SELECT(6),   0x00020000 },
+};
+
+static const struct reg_entry rgb_disable[] = {
+	{ DC_COM_PIN_OUTPUT_SELECT(6),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(5),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(4),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(3),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(2),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(1),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_SELECT(0),   0x00000000 },
+	{ DC_COM_PIN_OUTPUT_DATA(3),     0xaaaaaaaa },
+	{ DC_COM_PIN_OUTPUT_DATA(2),     0xaaaaaaaa },
+	{ DC_COM_PIN_OUTPUT_DATA(1),     0xaaaaaaaa },
+	{ DC_COM_PIN_OUTPUT_DATA(0),     0xaaaaaaaa },
+	{ DC_COM_PIN_OUTPUT_POLARITY(3), 0x00000000 },
+	{ DC_COM_PIN_OUTPUT_POLARITY(2), 0x00000000 },
+	{ DC_COM_PIN_OUTPUT_POLARITY(1), 0x00000000 },
+	{ DC_COM_PIN_OUTPUT_POLARITY(0), 0x00000000 },
+	{ DC_COM_PIN_OUTPUT_ENABLE(3),   0x55555555 },
+	{ DC_COM_PIN_OUTPUT_ENABLE(2),   0x55555555 },
+	{ DC_COM_PIN_OUTPUT_ENABLE(1),   0x55150005 },
+	{ DC_COM_PIN_OUTPUT_ENABLE(0),   0x55555555 },
+};
+
+static void tegra_dc_write_regs(struct tegra_dc *dc,
+				const struct reg_entry *table,
+				unsigned int num)
+{
+	unsigned int i;
+
+	for (i = 0; i < num; i++)
+		tegra_dc_writel(dc, table[i].value, table[i].offset);
+}
+
+static void tegra_rgb_encoder_disable(struct drm_encoder *encoder)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_rgb *rgb = to_rgb(output);
+
+	tegra_dc_write_regs(rgb->dc, rgb_disable, ARRAY_SIZE(rgb_disable));
+	tegra_dc_commit(rgb->dc);
+}
+
+static void tegra_rgb_encoder_enable(struct drm_encoder *encoder)
+{
+	struct drm_display_mode *mode = &encoder->crtc->state->adjusted_mode;
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_rgb *rgb = to_rgb(output);
+	u32 value;
+
+	tegra_dc_write_regs(rgb->dc, rgb_enable, ARRAY_SIZE(rgb_enable));
+
+	value = DE_SELECT_ACTIVE | DE_CONTROL_NORMAL;
+	tegra_dc_writel(rgb->dc, value, DC_DISP_DATA_ENABLE_OPTIONS);
+
+	/* configure H- and V-sync signal polarities */
+	value = tegra_dc_readl(rgb->dc, DC_COM_PIN_OUTPUT_POLARITY(1));
+
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		value |= LHS_OUTPUT_POLARITY_LOW;
+	else
+		value &= ~LHS_OUTPUT_POLARITY_LOW;
+
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		value |= LVS_OUTPUT_POLARITY_LOW;
+	else
+		value &= ~LVS_OUTPUT_POLARITY_LOW;
+
+	tegra_dc_writel(rgb->dc, value, DC_COM_PIN_OUTPUT_POLARITY(1));
+
+	/* XXX: parameterize? */
+	value = DISP_DATA_FORMAT_DF1P1C | DISP_ALIGNMENT_MSB |
+		DISP_ORDER_RED_BLUE;
+	tegra_dc_writel(rgb->dc, value, DC_DISP_DISP_INTERFACE_CONTROL);
+
+	tegra_dc_commit(rgb->dc);
+}
+
+static bool tegra_rgb_pll_rate_change_allowed(struct tegra_rgb *rgb)
+{
+	if (!rgb->pll_d2_out0)
+		return false;
+
+	if (!clk_is_match(rgb->clk_parent, rgb->pll_d_out0) &&
+	    !clk_is_match(rgb->clk_parent, rgb->pll_d2_out0))
+		return false;
+
+	return true;
+}
+
+static int
+tegra_rgb_encoder_atomic_check(struct drm_encoder *encoder,
+			       struct drm_crtc_state *crtc_state,
+			       struct drm_connector_state *conn_state)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(conn_state->crtc);
+	unsigned long pclk = crtc_state->mode.clock * 1000;
+	struct tegra_rgb *rgb = to_rgb(output);
+	unsigned int div;
+	int err;
+
+	/*
+	 * We may not want to change the frequency of the parent clock, since
+	 * it may be a parent for other peripherals. This is due to the fact
+	 * that on Tegra20 there's only a single clock dedicated to display
+	 * (pll_d_out0), whereas later generations have a second one that can
+	 * be used to independently drive a second output (pll_d2_out0).
+	 *
+	 * As a way to support multiple outputs on Tegra20 as well, pll_p is
+	 * typically used as the parent clock for the display controllers.
+	 * But this comes at a cost: pll_p is the parent of several other
+	 * peripherals, so its frequency shouldn't change out of the blue.
+	 *
+	 * The best we can do at this point is to use the shift clock divider
+	 * and hope that the desired frequency can be matched (or at least
+	 * matched sufficiently close that the panel will still work).
+	 */
+	if (tegra_rgb_pll_rate_change_allowed(rgb)) {
+		/*
+		 * Set display controller clock to x2 of PCLK in order to
+		 * produce higher resolution pulse positions.
+		 */
+		div = 2;
+		pclk *= 2;
+	} else {
+		div = ((clk_get_rate(rgb->clk) * 2) / pclk) - 2;
+		pclk = 0;
+	}
+
+	err = tegra_dc_state_setup_clock(dc, crtc_state, rgb->clk_parent,
+					 pclk, div);
+	if (err < 0) {
+		dev_err(output->dev, "failed to setup CRTC state: %d\n", err);
+		return err;
+	}
+
+	return err;
+}
+
+static const struct drm_encoder_helper_funcs tegra_rgb_encoder_helper_funcs = {
+	.disable = tegra_rgb_encoder_disable,
+	.enable = tegra_rgb_encoder_enable,
+	.atomic_check = tegra_rgb_encoder_atomic_check,
+};
+
+static void tegra_dc_of_node_put(void *data)
+{
+	of_node_put(data);
+}
+
+int tegra_dc_rgb_probe(struct tegra_dc *dc)
+{
+	struct device_node *np;
+	struct tegra_rgb *rgb;
+	int err;
+
+	np = of_get_child_by_name(dc->dev->of_node, "rgb");
+	if (!np)
+		return -ENODEV;
+
+	err = devm_add_action_or_reset(dc->dev, tegra_dc_of_node_put, np);
+	if (err < 0)
+		return err;
+
+	if (!of_device_is_available(np))
+		return -ENODEV;
+
+	rgb = devm_kzalloc(dc->dev, sizeof(*rgb), GFP_KERNEL);
+	if (!rgb)
+		return -ENOMEM;
+
+	rgb->output.dev = dc->dev;
+	rgb->output.of_node = np;
+	rgb->dc = dc;
+
+	err = tegra_output_probe(&rgb->output);
+	if (err < 0)
+		return err;
+
+	rgb->clk = devm_clk_get(dc->dev, NULL);
+	if (IS_ERR(rgb->clk)) {
+		dev_err(dc->dev, "failed to get clock\n");
+		err = PTR_ERR(rgb->clk);
+		goto remove;
+	}
+
+	rgb->clk_parent = devm_clk_get(dc->dev, "parent");
+	if (IS_ERR(rgb->clk_parent)) {
+		dev_err(dc->dev, "failed to get parent clock\n");
+		err = PTR_ERR(rgb->clk_parent);
+		goto remove;
+	}
+
+	err = clk_set_parent(rgb->clk, rgb->clk_parent);
+	if (err < 0) {
+		dev_err(dc->dev, "failed to set parent clock: %d\n", err);
+		goto remove;
+	}
+
+	rgb->pll_d_out0 = clk_get_sys(NULL, "pll_d_out0");
+	if (IS_ERR(rgb->pll_d_out0)) {
+		err = PTR_ERR(rgb->pll_d_out0);
+		dev_err(dc->dev, "failed to get pll_d_out0: %d\n", err);
+		goto remove;
+	}
+
+	if (dc->soc->has_pll_d2_out0) {
+		rgb->pll_d2_out0 = clk_get_sys(NULL, "pll_d2_out0");
+		if (IS_ERR(rgb->pll_d2_out0)) {
+			err = PTR_ERR(rgb->pll_d2_out0);
+			dev_err(dc->dev, "failed to get pll_d2_out0: %d\n", err);
+			goto put_pll;
+		}
+	}
+
+	dc->rgb = &rgb->output;
+
+	return 0;
+
+put_pll:
+	clk_put(rgb->pll_d_out0);
+remove:
+	tegra_output_remove(&rgb->output);
+	return err;
+}
+
+void tegra_dc_rgb_remove(struct tegra_dc *dc)
+{
+	struct tegra_rgb *rgb;
+
+	if (!dc->rgb)
+		return;
+
+	rgb = to_rgb(dc->rgb);
+	clk_put(rgb->pll_d2_out0);
+	clk_put(rgb->pll_d_out0);
+
+	tegra_output_remove(dc->rgb);
+	dc->rgb = NULL;
+}
+
+int tegra_dc_rgb_init(struct drm_device *drm, struct tegra_dc *dc)
+{
+	struct tegra_output *output = dc->rgb;
+	struct drm_connector *connector;
+	int err;
+
+	if (!dc->rgb)
+		return -ENODEV;
+
+	drm_simple_encoder_init(drm, &output->encoder, DRM_MODE_ENCODER_LVDS);
+	drm_encoder_helper_add(&output->encoder,
+			       &tegra_rgb_encoder_helper_funcs);
+
+	/*
+	 * Wrap directly-connected panel into DRM bridge in order to let
+	 * DRM core to handle panel for us.
+	 */
+	if (output->panel) {
+		output->bridge = devm_drm_panel_bridge_add(output->dev,
+							   output->panel);
+		if (IS_ERR(output->bridge)) {
+			dev_err(output->dev,
+				"failed to wrap panel into bridge: %pe\n",
+				output->bridge);
+			return PTR_ERR(output->bridge);
+		}
+
+		output->panel = NULL;
+	}
+
+	/*
+	 * Tegra devices that have LVDS panel utilize LVDS encoder bridge
+	 * for converting up to 28 LCD LVTTL lanes into 5/4 LVDS lanes that
+	 * go to display panel's receiver.
+	 *
+	 * Encoder usually have a power-down control which needs to be enabled
+	 * in order to transmit data to the panel.  Historically devices that
+	 * use an older device-tree version didn't model the bridge, assuming
+	 * that encoder is turned ON by default, while today's DRM allows us
+	 * to model LVDS encoder properly.
+	 *
+	 * Newer device-trees utilize LVDS encoder bridge, which provides
+	 * us with a connector and handles the display panel.
+	 *
+	 * For older device-trees we wrapped panel into the panel-bridge.
+	 */
+	if (output->bridge) {
+		err = drm_bridge_attach(&output->encoder, output->bridge,
+					NULL, DRM_BRIDGE_ATTACH_NO_CONNECTOR);
+		if (err)
+			return err;
+
+		connector = drm_bridge_connector_init(drm, &output->encoder);
+		if (IS_ERR(connector)) {
+			dev_err(output->dev,
+				"failed to initialize bridge connector: %pe\n",
+				connector);
+			return PTR_ERR(connector);
+		}
+	}
+
+	err = tegra_output_init(drm, output);
+	if (err < 0) {
+		dev_err(output->dev, "failed to initialize output: %d\n", err);
+		return err;
+	}
+
+	/*
+	 * Other outputs can be attached to either display controller. The RGB
+	 * outputs are an exception and work only with their parent display
+	 * controller.
+	 */
+	output->encoder.possible_crtcs = drm_crtc_mask(&dc->base);
+
+	return 0;
+}
+
+int tegra_dc_rgb_exit(struct tegra_dc *dc)
+{
+	if (dc->rgb)
+		tegra_output_exit(dc->rgb);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/tegra/riscv.c b/drivers/gpu/drm/tegra/riscv.c
new file mode 100644
index 000000000..658041640
--- /dev/null
+++ b/drivers/gpu/drm/tegra/riscv.c
@@ -0,0 +1,106 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2022, NVIDIA Corporation.
+ */
+
+#include <linux/dev_printk.h>
+#include <linux/device.h>
+#include <linux/iopoll.h>
+#include <linux/of.h>
+
+#include "riscv.h"
+
+#define RISCV_CPUCTL					0x4388
+#define RISCV_CPUCTL_STARTCPU_TRUE			(1 << 0)
+#define RISCV_BR_RETCODE				0x465c
+#define RISCV_BR_RETCODE_RESULT_V(x)			((x) & 0x3)
+#define RISCV_BR_RETCODE_RESULT_PASS_V			3
+#define RISCV_BCR_CTRL					0x4668
+#define RISCV_BCR_CTRL_CORE_SELECT_RISCV		(1 << 4)
+#define RISCV_BCR_DMACFG				0x466c
+#define RISCV_BCR_DMACFG_TARGET_LOCAL_FB		(0 << 0)
+#define RISCV_BCR_DMACFG_LOCK_LOCKED			(1 << 31)
+#define RISCV_BCR_DMAADDR_PKCPARAM_LO			0x4670
+#define RISCV_BCR_DMAADDR_PKCPARAM_HI			0x4674
+#define RISCV_BCR_DMAADDR_FMCCODE_LO			0x4678
+#define RISCV_BCR_DMAADDR_FMCCODE_HI			0x467c
+#define RISCV_BCR_DMAADDR_FMCDATA_LO			0x4680
+#define RISCV_BCR_DMAADDR_FMCDATA_HI			0x4684
+#define RISCV_BCR_DMACFG_SEC				0x4694
+#define RISCV_BCR_DMACFG_SEC_GSCID(v)			((v) << 16)
+
+static void riscv_writel(struct tegra_drm_riscv *riscv, u32 value, u32 offset)
+{
+	writel(value, riscv->regs + offset);
+}
+
+int tegra_drm_riscv_read_descriptors(struct tegra_drm_riscv *riscv)
+{
+	struct tegra_drm_riscv_descriptor *bl = &riscv->bl_desc;
+	struct tegra_drm_riscv_descriptor *os = &riscv->os_desc;
+	const struct device_node *np = riscv->dev->of_node;
+	int err;
+
+#define READ_PROP(name, location) \
+	err = of_property_read_u32(np, name, location); \
+	if (err) { \
+		dev_err(riscv->dev, "failed to read " name ": %d\n", err); \
+		return err; \
+	}
+
+	READ_PROP("nvidia,bl-manifest-offset", &bl->manifest_offset);
+	READ_PROP("nvidia,bl-code-offset", &bl->code_offset);
+	READ_PROP("nvidia,bl-data-offset", &bl->data_offset);
+	READ_PROP("nvidia,os-manifest-offset", &os->manifest_offset);
+	READ_PROP("nvidia,os-code-offset", &os->code_offset);
+	READ_PROP("nvidia,os-data-offset", &os->data_offset);
+#undef READ_PROP
+
+	if (bl->manifest_offset == 0 && bl->code_offset == 0 &&
+	    bl->data_offset == 0 && os->manifest_offset == 0 &&
+	    os->code_offset == 0 && os->data_offset == 0) {
+		dev_err(riscv->dev, "descriptors not available\n");
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int tegra_drm_riscv_boot_bootrom(struct tegra_drm_riscv *riscv, phys_addr_t image_address,
+				 u32 gscid, const struct tegra_drm_riscv_descriptor *desc)
+{
+	phys_addr_t addr;
+	int err;
+	u32 val;
+
+	riscv_writel(riscv, RISCV_BCR_CTRL_CORE_SELECT_RISCV, RISCV_BCR_CTRL);
+
+	addr = image_address + desc->manifest_offset;
+	riscv_writel(riscv, lower_32_bits(addr >> 8), RISCV_BCR_DMAADDR_PKCPARAM_LO);
+	riscv_writel(riscv, upper_32_bits(addr >> 8), RISCV_BCR_DMAADDR_PKCPARAM_HI);
+
+	addr = image_address + desc->code_offset;
+	riscv_writel(riscv, lower_32_bits(addr >> 8), RISCV_BCR_DMAADDR_FMCCODE_LO);
+	riscv_writel(riscv, upper_32_bits(addr >> 8), RISCV_BCR_DMAADDR_FMCCODE_HI);
+
+	addr = image_address + desc->data_offset;
+	riscv_writel(riscv, lower_32_bits(addr >> 8), RISCV_BCR_DMAADDR_FMCDATA_LO);
+	riscv_writel(riscv, upper_32_bits(addr >> 8), RISCV_BCR_DMAADDR_FMCDATA_HI);
+
+	riscv_writel(riscv, RISCV_BCR_DMACFG_SEC_GSCID(gscid), RISCV_BCR_DMACFG_SEC);
+	riscv_writel(riscv,
+		RISCV_BCR_DMACFG_TARGET_LOCAL_FB | RISCV_BCR_DMACFG_LOCK_LOCKED, RISCV_BCR_DMACFG);
+
+	riscv_writel(riscv, RISCV_CPUCTL_STARTCPU_TRUE, RISCV_CPUCTL);
+
+	err = readl_poll_timeout(
+		riscv->regs + RISCV_BR_RETCODE, val,
+		RISCV_BR_RETCODE_RESULT_V(val) == RISCV_BR_RETCODE_RESULT_PASS_V,
+		10, 100000);
+	if (err) {
+		dev_err(riscv->dev, "error during bootrom execution. BR_RETCODE=%d\n", val);
+		return err;
+	}
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/tegra/riscv.h b/drivers/gpu/drm/tegra/riscv.h
new file mode 100644
index 000000000..bbeb2db07
--- /dev/null
+++ b/drivers/gpu/drm/tegra/riscv.h
@@ -0,0 +1,30 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2022, NVIDIA Corporation.
+ */
+
+#ifndef DRM_TEGRA_RISCV_H
+#define DRM_TEGRA_RISCV_H
+
+struct tegra_drm_riscv_descriptor {
+	u32 manifest_offset;
+	u32 code_offset;
+	u32 code_size;
+	u32 data_offset;
+	u32 data_size;
+};
+
+struct tegra_drm_riscv {
+	/* User initializes */
+	struct device *dev;
+	void __iomem *regs;
+
+	struct tegra_drm_riscv_descriptor bl_desc;
+	struct tegra_drm_riscv_descriptor os_desc;
+};
+
+int tegra_drm_riscv_read_descriptors(struct tegra_drm_riscv *riscv);
+int tegra_drm_riscv_boot_bootrom(struct tegra_drm_riscv *riscv, phys_addr_t image_address,
+				 u32 gscid, const struct tegra_drm_riscv_descriptor *desc);
+
+#endif
diff --git a/drivers/gpu/drm/tegra/sor.c b/drivers/gpu/drm/tegra/sor.c
new file mode 100644
index 000000000..e384cbd0c
--- /dev/null
+++ b/drivers/gpu/drm/tegra/sor.c
@@ -0,0 +1,4082 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#include <linux/clk.h>
+#include <linux/clk-provider.h>
+#include <linux/debugfs.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regulator/consumer.h>
+#include <linux/reset.h>
+#include <linux/string_choices.h>
+
+#include <soc/tegra/pmc.h>
+
+#include <drm/display/drm_dp_helper.h>
+#include <drm/display/drm_scdc_helper.h>
+#include <drm/drm_atomic_helper.h>
+#include <drm/drm_debugfs.h>
+#include <drm/drm_edid.h>
+#include <drm/drm_eld.h>
+#include <drm/drm_encoder.h>
+#include <drm/drm_file.h>
+#include <drm/drm_panel.h>
+#include <drm/drm_print.h>
+
+#include "dc.h"
+#include "dp.h"
+#include "drm.h"
+#include "hda.h"
+#include "sor.h"
+#include "trace.h"
+
+#define SOR_REKEY 0x38
+
+struct tegra_sor_hdmi_settings {
+	unsigned long frequency;
+
+	u8 vcocap;
+	u8 filter;
+	u8 ichpmp;
+	u8 loadadj;
+	u8 tmds_termadj;
+	u8 tx_pu_value;
+	u8 bg_temp_coef;
+	u8 bg_vref_level;
+	u8 avdd10_level;
+	u8 avdd14_level;
+	u8 sparepll;
+
+	u8 drive_current[4];
+	u8 preemphasis[4];
+};
+
+#if 1
+static const struct tegra_sor_hdmi_settings tegra210_sor_hdmi_defaults[] = {
+	{
+		.frequency = 54000000,
+		.vcocap = 0x0,
+		.filter = 0x0,
+		.ichpmp = 0x1,
+		.loadadj = 0x3,
+		.tmds_termadj = 0x9,
+		.tx_pu_value = 0x10,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0x8,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x33, 0x3a, 0x3a, 0x3a },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 75000000,
+		.vcocap = 0x3,
+		.filter = 0x0,
+		.ichpmp = 0x1,
+		.loadadj = 0x3,
+		.tmds_termadj = 0x9,
+		.tx_pu_value = 0x40,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0x8,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x33, 0x3a, 0x3a, 0x3a },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 150000000,
+		.vcocap = 0x3,
+		.filter = 0x0,
+		.ichpmp = 0x1,
+		.loadadj = 0x3,
+		.tmds_termadj = 0x9,
+		.tx_pu_value = 0x66,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0x8,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x33, 0x3a, 0x3a, 0x3a },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 300000000,
+		.vcocap = 0x3,
+		.filter = 0x0,
+		.ichpmp = 0x1,
+		.loadadj = 0x3,
+		.tmds_termadj = 0x9,
+		.tx_pu_value = 0x66,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0xa,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x33, 0x3f, 0x3f, 0x3f },
+		.preemphasis = { 0x00, 0x17, 0x17, 0x17 },
+	}, {
+		.frequency = 600000000,
+		.vcocap = 0x3,
+		.filter = 0x0,
+		.ichpmp = 0x1,
+		.loadadj = 0x3,
+		.tmds_termadj = 0x9,
+		.tx_pu_value = 0x66,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0x8,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x33, 0x3f, 0x3f, 0x3f },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	},
+};
+#else
+static const struct tegra_sor_hdmi_settings tegra210_sor_hdmi_defaults[] = {
+	{
+		.frequency = 75000000,
+		.vcocap = 0x3,
+		.filter = 0x0,
+		.ichpmp = 0x1,
+		.loadadj = 0x3,
+		.tmds_termadj = 0x9,
+		.tx_pu_value = 0x40,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0x8,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x29, 0x29, 0x29, 0x29 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 150000000,
+		.vcocap = 0x3,
+		.filter = 0x0,
+		.ichpmp = 0x1,
+		.loadadj = 0x3,
+		.tmds_termadj = 0x9,
+		.tx_pu_value = 0x66,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0x8,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x30, 0x37, 0x37, 0x37 },
+		.preemphasis = { 0x01, 0x02, 0x02, 0x02 },
+	}, {
+		.frequency = 300000000,
+		.vcocap = 0x3,
+		.filter = 0x0,
+		.ichpmp = 0x6,
+		.loadadj = 0x3,
+		.tmds_termadj = 0x9,
+		.tx_pu_value = 0x66,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0xf,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x30, 0x37, 0x37, 0x37 },
+		.preemphasis = { 0x10, 0x3e, 0x3e, 0x3e },
+	}, {
+		.frequency = 600000000,
+		.vcocap = 0x3,
+		.filter = 0x0,
+		.ichpmp = 0xa,
+		.loadadj = 0x3,
+		.tmds_termadj = 0xb,
+		.tx_pu_value = 0x66,
+		.bg_temp_coef = 0x3,
+		.bg_vref_level = 0xe,
+		.avdd10_level = 0x4,
+		.avdd14_level = 0x4,
+		.sparepll = 0x0,
+		.drive_current = { 0x35, 0x3e, 0x3e, 0x3e },
+		.preemphasis = { 0x02, 0x3f, 0x3f, 0x3f },
+	},
+};
+#endif
+
+static const struct tegra_sor_hdmi_settings tegra186_sor_hdmi_defaults[] = {
+	{
+		.frequency = 54000000,
+		.vcocap = 0,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 0xf,
+		.tx_pu_value = 0,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x54,
+		.drive_current = { 0x3a, 0x3a, 0x3a, 0x33 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 75000000,
+		.vcocap = 1,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 0xf,
+		.tx_pu_value = 0,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x44,
+		.drive_current = { 0x3a, 0x3a, 0x3a, 0x33 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 150000000,
+		.vcocap = 3,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 15,
+		.tx_pu_value = 0x66 /* 0 */,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x00, /* 0x34 */
+		.drive_current = { 0x3a, 0x3a, 0x3a, 0x37 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 300000000,
+		.vcocap = 3,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 15,
+		.tx_pu_value = 64,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x34,
+		.drive_current = { 0x3d, 0x3d, 0x3d, 0x33 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 600000000,
+		.vcocap = 3,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 12,
+		.tx_pu_value = 96,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x34,
+		.drive_current = { 0x3d, 0x3d, 0x3d, 0x33 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}
+};
+
+static const struct tegra_sor_hdmi_settings tegra194_sor_hdmi_defaults[] = {
+	{
+		.frequency = 54000000,
+		.vcocap = 0,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 0xf,
+		.tx_pu_value = 0,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x54,
+		.drive_current = { 0x3a, 0x3a, 0x3a, 0x33 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 75000000,
+		.vcocap = 1,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 0xf,
+		.tx_pu_value = 0,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x44,
+		.drive_current = { 0x3a, 0x3a, 0x3a, 0x33 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 150000000,
+		.vcocap = 3,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 15,
+		.tx_pu_value = 0x66 /* 0 */,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x00, /* 0x34 */
+		.drive_current = { 0x3a, 0x3a, 0x3a, 0x37 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 300000000,
+		.vcocap = 3,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 15,
+		.tx_pu_value = 64,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x34,
+		.drive_current = { 0x3d, 0x3d, 0x3d, 0x33 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}, {
+		.frequency = 600000000,
+		.vcocap = 3,
+		.filter = 5,
+		.ichpmp = 5,
+		.loadadj = 3,
+		.tmds_termadj = 12,
+		.tx_pu_value = 96,
+		.bg_temp_coef = 3,
+		.bg_vref_level = 8,
+		.avdd10_level = 4,
+		.avdd14_level = 4,
+		.sparepll = 0x34,
+		.drive_current = { 0x3d, 0x3d, 0x3d, 0x33 },
+		.preemphasis = { 0x00, 0x00, 0x00, 0x00 },
+	}
+};
+
+struct tegra_sor_regs {
+	unsigned int head_state0;
+	unsigned int head_state1;
+	unsigned int head_state2;
+	unsigned int head_state3;
+	unsigned int head_state4;
+	unsigned int head_state5;
+	unsigned int pll0;
+	unsigned int pll1;
+	unsigned int pll2;
+	unsigned int pll3;
+	unsigned int dp_padctl0;
+	unsigned int dp_padctl2;
+};
+
+struct tegra_sor_soc {
+	bool supports_lvds;
+	bool supports_hdmi;
+	bool supports_dp;
+	bool supports_audio;
+	bool supports_hdcp;
+
+	const struct tegra_sor_regs *regs;
+	bool has_nvdisplay;
+
+	const struct tegra_sor_hdmi_settings *settings;
+	unsigned int num_settings;
+
+	const u8 *xbar_cfg;
+	const u8 *lane_map;
+
+	const u8 (*voltage_swing)[4][4];
+	const u8 (*pre_emphasis)[4][4];
+	const u8 (*post_cursor)[4][4];
+	const u8 (*tx_pu)[4][4];
+};
+
+struct tegra_sor;
+
+struct tegra_sor_ops {
+	const char *name;
+	int (*probe)(struct tegra_sor *sor);
+	void (*audio_enable)(struct tegra_sor *sor);
+	void (*audio_disable)(struct tegra_sor *sor);
+};
+
+struct tegra_sor {
+	struct host1x_client client;
+	struct tegra_output output;
+	struct device *dev;
+
+	const struct tegra_sor_soc *soc;
+	void __iomem *regs;
+	unsigned int index;
+	unsigned int irq;
+
+	struct reset_control *rst;
+	struct clk *clk_parent;
+	struct clk *clk_safe;
+	struct clk *clk_out;
+	struct clk *clk_pad;
+	struct clk *clk_dp;
+	struct clk *clk;
+
+	struct tegra_pmc *pmc;
+
+	u8 xbar_cfg[5];
+
+	struct drm_dp_link link;
+	struct drm_dp_aux *aux;
+
+	struct drm_info_list *debugfs_files;
+
+	const struct tegra_sor_ops *ops;
+	enum tegra_io_pad pad;
+
+	/* for HDMI 2.0 */
+	struct tegra_sor_hdmi_settings *settings;
+	unsigned int num_settings;
+
+	struct regulator *avdd_io_supply;
+	struct regulator *vdd_pll_supply;
+	struct regulator *hdmi_supply;
+
+	struct delayed_work scdc;
+	bool scdc_enabled;
+
+	struct tegra_hda_format format;
+};
+
+struct tegra_sor_state {
+	struct drm_connector_state base;
+
+	unsigned int link_speed;
+	unsigned long pclk;
+	unsigned int bpc;
+};
+
+static inline struct tegra_sor_state *
+to_sor_state(struct drm_connector_state *state)
+{
+	return container_of(state, struct tegra_sor_state, base);
+}
+
+struct tegra_sor_config {
+	u32 bits_per_pixel;
+
+	u32 active_polarity;
+	u32 active_count;
+	u32 tu_size;
+	u32 active_frac;
+	u32 watermark;
+
+	u32 hblank_symbols;
+	u32 vblank_symbols;
+};
+
+static inline struct tegra_sor *
+host1x_client_to_sor(struct host1x_client *client)
+{
+	return container_of(client, struct tegra_sor, client);
+}
+
+static inline struct tegra_sor *to_sor(struct tegra_output *output)
+{
+	return container_of(output, struct tegra_sor, output);
+}
+
+static inline u32 tegra_sor_readl(struct tegra_sor *sor, unsigned int offset)
+{
+	u32 value = readl(sor->regs + (offset << 2));
+
+	trace_sor_readl(sor->dev, offset, value);
+
+	return value;
+}
+
+static inline void tegra_sor_writel(struct tegra_sor *sor, u32 value,
+				    unsigned int offset)
+{
+	trace_sor_writel(sor->dev, offset, value);
+	writel(value, sor->regs + (offset << 2));
+}
+
+static int tegra_sor_set_parent_clock(struct tegra_sor *sor, struct clk *parent)
+{
+	int err;
+
+	clk_disable_unprepare(sor->clk);
+
+	err = clk_set_parent(sor->clk_out, parent);
+	if (err < 0)
+		return err;
+
+	err = clk_prepare_enable(sor->clk);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+struct tegra_clk_sor_pad {
+	struct clk_hw hw;
+	struct tegra_sor *sor;
+};
+
+static inline struct tegra_clk_sor_pad *to_pad(struct clk_hw *hw)
+{
+	return container_of(hw, struct tegra_clk_sor_pad, hw);
+}
+
+static const char * const tegra_clk_sor_pad_parents[2][2] = {
+	{ "pll_d_out0", "pll_dp" },
+	{ "pll_d2_out0", "pll_dp" },
+};
+
+/*
+ * Implementing ->set_parent() here isn't really required because the parent
+ * will be explicitly selected in the driver code via the DP_CLK_SEL mux in
+ * the SOR_CLK_CNTRL register. This is primarily for compatibility with the
+ * Tegra186 and later SoC generations where the BPMP implements this clock
+ * and doesn't expose the mux via the common clock framework.
+ */
+
+static int tegra_clk_sor_pad_set_parent(struct clk_hw *hw, u8 index)
+{
+	struct tegra_clk_sor_pad *pad = to_pad(hw);
+	struct tegra_sor *sor = pad->sor;
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_CLK_CNTRL);
+	value &= ~SOR_CLK_CNTRL_DP_CLK_SEL_MASK;
+
+	switch (index) {
+	case 0:
+		value |= SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_PCLK;
+		break;
+
+	case 1:
+		value |= SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_DPCLK;
+		break;
+	}
+
+	tegra_sor_writel(sor, value, SOR_CLK_CNTRL);
+
+	return 0;
+}
+
+static u8 tegra_clk_sor_pad_get_parent(struct clk_hw *hw)
+{
+	struct tegra_clk_sor_pad *pad = to_pad(hw);
+	struct tegra_sor *sor = pad->sor;
+	u8 parent = U8_MAX;
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_CLK_CNTRL);
+
+	switch (value & SOR_CLK_CNTRL_DP_CLK_SEL_MASK) {
+	case SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_PCLK:
+	case SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_PCLK:
+		parent = 0;
+		break;
+
+	case SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_DPCLK:
+	case SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_DPCLK:
+		parent = 1;
+		break;
+	}
+
+	return parent;
+}
+
+static const struct clk_ops tegra_clk_sor_pad_ops = {
+	.determine_rate = clk_hw_determine_rate_no_reparent,
+	.set_parent = tegra_clk_sor_pad_set_parent,
+	.get_parent = tegra_clk_sor_pad_get_parent,
+};
+
+static struct clk *tegra_clk_sor_pad_register(struct tegra_sor *sor,
+					      const char *name)
+{
+	struct tegra_clk_sor_pad *pad;
+	struct clk_init_data init;
+	struct clk *clk;
+
+	pad = devm_kzalloc(sor->dev, sizeof(*pad), GFP_KERNEL);
+	if (!pad)
+		return ERR_PTR(-ENOMEM);
+
+	pad->sor = sor;
+
+	init.name = name;
+	init.flags = 0;
+	init.parent_names = tegra_clk_sor_pad_parents[sor->index];
+	init.num_parents = ARRAY_SIZE(tegra_clk_sor_pad_parents[sor->index]);
+	init.ops = &tegra_clk_sor_pad_ops;
+
+	pad->hw.init = &init;
+
+	clk = devm_clk_register(sor->dev, &pad->hw);
+
+	return clk;
+}
+
+static void tegra_sor_filter_rates(struct tegra_sor *sor)
+{
+	struct drm_dp_link *link = &sor->link;
+	unsigned int i;
+
+	/* Tegra only supports RBR, HBR and HBR2 */
+	for (i = 0; i < link->num_rates; i++) {
+		switch (link->rates[i]) {
+		case 1620000:
+		case 2700000:
+		case 5400000:
+			break;
+
+		default:
+			DRM_DEBUG_KMS("link rate %lu kHz not supported\n",
+				      link->rates[i]);
+			link->rates[i] = 0;
+			break;
+		}
+	}
+
+	drm_dp_link_update_rates(link);
+}
+
+static int tegra_sor_power_up_lanes(struct tegra_sor *sor, unsigned int lanes)
+{
+	unsigned long timeout;
+	u32 value;
+
+	/*
+	 * Clear or set the PD_TXD bit corresponding to each lane, depending
+	 * on whether it is used or not.
+	 */
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+
+	if (lanes <= 2)
+		value &= ~(SOR_DP_PADCTL_PD_TXD(sor->soc->lane_map[3]) |
+			   SOR_DP_PADCTL_PD_TXD(sor->soc->lane_map[2]));
+	else
+		value |= SOR_DP_PADCTL_PD_TXD(sor->soc->lane_map[3]) |
+			 SOR_DP_PADCTL_PD_TXD(sor->soc->lane_map[2]);
+
+	if (lanes <= 1)
+		value &= ~SOR_DP_PADCTL_PD_TXD(sor->soc->lane_map[1]);
+	else
+		value |= SOR_DP_PADCTL_PD_TXD(sor->soc->lane_map[1]);
+
+	if (lanes == 0)
+		value &= ~SOR_DP_PADCTL_PD_TXD(sor->soc->lane_map[0]);
+	else
+		value |= SOR_DP_PADCTL_PD_TXD(sor->soc->lane_map[0]);
+
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	/* start lane sequencer */
+	value = SOR_LANE_SEQ_CTL_TRIGGER | SOR_LANE_SEQ_CTL_SEQUENCE_DOWN |
+		SOR_LANE_SEQ_CTL_POWER_STATE_UP;
+	tegra_sor_writel(sor, value, SOR_LANE_SEQ_CTL);
+
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_LANE_SEQ_CTL);
+		if ((value & SOR_LANE_SEQ_CTL_TRIGGER) == 0)
+			break;
+
+		usleep_range(250, 1000);
+	}
+
+	if ((value & SOR_LANE_SEQ_CTL_TRIGGER) != 0)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+static int tegra_sor_power_down_lanes(struct tegra_sor *sor)
+{
+	unsigned long timeout;
+	u32 value;
+
+	/* power down all lanes */
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value &= ~(SOR_DP_PADCTL_PD_TXD_3 | SOR_DP_PADCTL_PD_TXD_0 |
+		   SOR_DP_PADCTL_PD_TXD_1 | SOR_DP_PADCTL_PD_TXD_2);
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	/* start lane sequencer */
+	value = SOR_LANE_SEQ_CTL_TRIGGER | SOR_LANE_SEQ_CTL_SEQUENCE_UP |
+		SOR_LANE_SEQ_CTL_POWER_STATE_DOWN;
+	tegra_sor_writel(sor, value, SOR_LANE_SEQ_CTL);
+
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_LANE_SEQ_CTL);
+		if ((value & SOR_LANE_SEQ_CTL_TRIGGER) == 0)
+			break;
+
+		usleep_range(25, 100);
+	}
+
+	if ((value & SOR_LANE_SEQ_CTL_TRIGGER) != 0)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+static void tegra_sor_dp_precharge(struct tegra_sor *sor, unsigned int lanes)
+{
+	u32 value;
+
+	/* pre-charge all used lanes */
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+
+	if (lanes <= 2)
+		value &= ~(SOR_DP_PADCTL_CM_TXD(sor->soc->lane_map[3]) |
+			   SOR_DP_PADCTL_CM_TXD(sor->soc->lane_map[2]));
+	else
+		value |= SOR_DP_PADCTL_CM_TXD(sor->soc->lane_map[3]) |
+			 SOR_DP_PADCTL_CM_TXD(sor->soc->lane_map[2]);
+
+	if (lanes <= 1)
+		value &= ~SOR_DP_PADCTL_CM_TXD(sor->soc->lane_map[1]);
+	else
+		value |= SOR_DP_PADCTL_CM_TXD(sor->soc->lane_map[1]);
+
+	if (lanes == 0)
+		value &= ~SOR_DP_PADCTL_CM_TXD(sor->soc->lane_map[0]);
+	else
+		value |= SOR_DP_PADCTL_CM_TXD(sor->soc->lane_map[0]);
+
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	usleep_range(15, 100);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value &= ~(SOR_DP_PADCTL_CM_TXD_3 | SOR_DP_PADCTL_CM_TXD_2 |
+		   SOR_DP_PADCTL_CM_TXD_1 | SOR_DP_PADCTL_CM_TXD_0);
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+}
+
+static void tegra_sor_dp_term_calibrate(struct tegra_sor *sor)
+{
+	u32 mask = 0x08, adj = 0, value;
+
+	/* enable pad calibration logic */
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value &= ~SOR_DP_PADCTL_PAD_CAL_PD;
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll1);
+	value |= SOR_PLL1_TMDS_TERM;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll1);
+
+	while (mask) {
+		adj |= mask;
+
+		value = tegra_sor_readl(sor, sor->soc->regs->pll1);
+		value &= ~SOR_PLL1_TMDS_TERMADJ_MASK;
+		value |= SOR_PLL1_TMDS_TERMADJ(adj);
+		tegra_sor_writel(sor, value, sor->soc->regs->pll1);
+
+		usleep_range(100, 200);
+
+		value = tegra_sor_readl(sor, sor->soc->regs->pll1);
+		if (value & SOR_PLL1_TERM_COMPOUT)
+			adj &= ~mask;
+
+		mask >>= 1;
+	}
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll1);
+	value &= ~SOR_PLL1_TMDS_TERMADJ_MASK;
+	value |= SOR_PLL1_TMDS_TERMADJ(adj);
+	tegra_sor_writel(sor, value, sor->soc->regs->pll1);
+
+	/* disable pad calibration logic */
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value |= SOR_DP_PADCTL_PAD_CAL_PD;
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+}
+
+static int tegra_sor_dp_link_apply_training(struct drm_dp_link *link)
+{
+	struct tegra_sor *sor = container_of(link, struct tegra_sor, link);
+	u32 voltage_swing = 0, pre_emphasis = 0, post_cursor = 0;
+	const struct tegra_sor_soc *soc = sor->soc;
+	u32 pattern = 0, tx_pu = 0, value;
+	unsigned int i;
+
+	for (value = 0, i = 0; i < link->lanes; i++) {
+		u8 vs = link->train.request.voltage_swing[i];
+		u8 pe = link->train.request.pre_emphasis[i];
+		u8 pc = link->train.request.post_cursor[i];
+		u8 shift = sor->soc->lane_map[i] << 3;
+
+		voltage_swing |= soc->voltage_swing[pc][vs][pe] << shift;
+		pre_emphasis |= soc->pre_emphasis[pc][vs][pe] << shift;
+		post_cursor |= soc->post_cursor[pc][vs][pe] << shift;
+
+		if (sor->soc->tx_pu[pc][vs][pe] > tx_pu)
+			tx_pu = sor->soc->tx_pu[pc][vs][pe];
+
+		switch (link->train.pattern) {
+		case DP_TRAINING_PATTERN_DISABLE:
+			value = SOR_DP_TPG_SCRAMBLER_GALIOS |
+				SOR_DP_TPG_PATTERN_NONE;
+			break;
+
+		case DP_TRAINING_PATTERN_1:
+			value = SOR_DP_TPG_SCRAMBLER_NONE |
+				SOR_DP_TPG_PATTERN_TRAIN1;
+			break;
+
+		case DP_TRAINING_PATTERN_2:
+			value = SOR_DP_TPG_SCRAMBLER_NONE |
+				SOR_DP_TPG_PATTERN_TRAIN2;
+			break;
+
+		case DP_TRAINING_PATTERN_3:
+			value = SOR_DP_TPG_SCRAMBLER_NONE |
+				SOR_DP_TPG_PATTERN_TRAIN3;
+			break;
+
+		default:
+			return -EINVAL;
+		}
+
+		if (link->caps.channel_coding)
+			value |= SOR_DP_TPG_CHANNEL_CODING;
+
+		pattern = pattern << 8 | value;
+	}
+
+	tegra_sor_writel(sor, voltage_swing, SOR_LANE_DRIVE_CURRENT0);
+	tegra_sor_writel(sor, pre_emphasis, SOR_LANE_PREEMPHASIS0);
+
+	if (link->caps.tps3_supported)
+		tegra_sor_writel(sor, post_cursor, SOR_LANE_POSTCURSOR0);
+
+	tegra_sor_writel(sor, pattern, SOR_DP_TPG);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value &= ~SOR_DP_PADCTL_TX_PU_MASK;
+	value |= SOR_DP_PADCTL_TX_PU_ENABLE;
+	value |= SOR_DP_PADCTL_TX_PU(tx_pu);
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	usleep_range(20, 100);
+
+	return 0;
+}
+
+static int tegra_sor_dp_link_configure(struct drm_dp_link *link)
+{
+	struct tegra_sor *sor = container_of(link, struct tegra_sor, link);
+	unsigned int rate, lanes;
+	u32 value;
+	int err;
+
+	rate = drm_dp_link_rate_to_bw_code(link->rate);
+	lanes = link->lanes;
+
+	/* configure link speed and lane count */
+	value = tegra_sor_readl(sor, SOR_CLK_CNTRL);
+	value &= ~SOR_CLK_CNTRL_DP_LINK_SPEED_MASK;
+	value |= SOR_CLK_CNTRL_DP_LINK_SPEED(rate);
+	tegra_sor_writel(sor, value, SOR_CLK_CNTRL);
+
+	value = tegra_sor_readl(sor, SOR_DP_LINKCTL0);
+	value &= ~SOR_DP_LINKCTL_LANE_COUNT_MASK;
+	value |= SOR_DP_LINKCTL_LANE_COUNT(lanes);
+
+	if (link->caps.enhanced_framing)
+		value |= SOR_DP_LINKCTL_ENHANCED_FRAME;
+
+	tegra_sor_writel(sor, value, SOR_DP_LINKCTL0);
+
+	usleep_range(400, 1000);
+
+	/* configure load pulse position adjustment */
+	value = tegra_sor_readl(sor, sor->soc->regs->pll1);
+	value &= ~SOR_PLL1_LOADADJ_MASK;
+
+	switch (rate) {
+	case DP_LINK_BW_1_62:
+		value |= SOR_PLL1_LOADADJ(0x3);
+		break;
+
+	case DP_LINK_BW_2_7:
+		value |= SOR_PLL1_LOADADJ(0x4);
+		break;
+
+	case DP_LINK_BW_5_4:
+		value |= SOR_PLL1_LOADADJ(0x6);
+		break;
+	}
+
+	tegra_sor_writel(sor, value, sor->soc->regs->pll1);
+
+	/* use alternate scrambler reset for eDP */
+	value = tegra_sor_readl(sor, SOR_DP_SPARE0);
+
+	if (link->edp == 0)
+		value &= ~SOR_DP_SPARE_PANEL_INTERNAL;
+	else
+		value |= SOR_DP_SPARE_PANEL_INTERNAL;
+
+	tegra_sor_writel(sor, value, SOR_DP_SPARE0);
+
+	err = tegra_sor_power_down_lanes(sor);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to power down lanes: %d\n", err);
+		return err;
+	}
+
+	/* power up and pre-charge lanes */
+	err = tegra_sor_power_up_lanes(sor, lanes);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to power up %u lane%s: %d\n",
+			lanes, str_plural(lanes), err);
+		return err;
+	}
+
+	tegra_sor_dp_precharge(sor, lanes);
+
+	return 0;
+}
+
+static const struct drm_dp_link_ops tegra_sor_dp_link_ops = {
+	.apply_training = tegra_sor_dp_link_apply_training,
+	.configure = tegra_sor_dp_link_configure,
+};
+
+static void tegra_sor_super_update(struct tegra_sor *sor)
+{
+	tegra_sor_writel(sor, 0, SOR_SUPER_STATE0);
+	tegra_sor_writel(sor, 1, SOR_SUPER_STATE0);
+	tegra_sor_writel(sor, 0, SOR_SUPER_STATE0);
+}
+
+static void tegra_sor_update(struct tegra_sor *sor)
+{
+	tegra_sor_writel(sor, 0, SOR_STATE0);
+	tegra_sor_writel(sor, 1, SOR_STATE0);
+	tegra_sor_writel(sor, 0, SOR_STATE0);
+}
+
+static int tegra_sor_setup_pwm(struct tegra_sor *sor, unsigned long timeout)
+{
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_PWM_DIV);
+	value &= ~SOR_PWM_DIV_MASK;
+	value |= 0x400; /* period */
+	tegra_sor_writel(sor, value, SOR_PWM_DIV);
+
+	value = tegra_sor_readl(sor, SOR_PWM_CTL);
+	value &= ~SOR_PWM_CTL_DUTY_CYCLE_MASK;
+	value |= 0x400; /* duty cycle */
+	value &= ~SOR_PWM_CTL_CLK_SEL; /* clock source: PCLK */
+	value |= SOR_PWM_CTL_TRIGGER;
+	tegra_sor_writel(sor, value, SOR_PWM_CTL);
+
+	timeout = jiffies + msecs_to_jiffies(timeout);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_PWM_CTL);
+		if ((value & SOR_PWM_CTL_TRIGGER) == 0)
+			return 0;
+
+		usleep_range(25, 100);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static int tegra_sor_attach(struct tegra_sor *sor)
+{
+	unsigned long value, timeout;
+
+	/* wake up in normal mode */
+	value = tegra_sor_readl(sor, SOR_SUPER_STATE1);
+	value |= SOR_SUPER_STATE_HEAD_MODE_AWAKE;
+	value |= SOR_SUPER_STATE_MODE_NORMAL;
+	tegra_sor_writel(sor, value, SOR_SUPER_STATE1);
+	tegra_sor_super_update(sor);
+
+	/* attach */
+	value = tegra_sor_readl(sor, SOR_SUPER_STATE1);
+	value |= SOR_SUPER_STATE_ATTACHED;
+	tegra_sor_writel(sor, value, SOR_SUPER_STATE1);
+	tegra_sor_super_update(sor);
+
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_TEST);
+		if ((value & SOR_TEST_ATTACHED) != 0)
+			return 0;
+
+		usleep_range(25, 100);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static int tegra_sor_wakeup(struct tegra_sor *sor)
+{
+	unsigned long value, timeout;
+
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	/* wait for head to wake up */
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_TEST);
+		value &= SOR_TEST_HEAD_MODE_MASK;
+
+		if (value == SOR_TEST_HEAD_MODE_AWAKE)
+			return 0;
+
+		usleep_range(25, 100);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static int tegra_sor_power_up(struct tegra_sor *sor, unsigned long timeout)
+{
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_PWR);
+	value |= SOR_PWR_TRIGGER | SOR_PWR_NORMAL_STATE_PU;
+	tegra_sor_writel(sor, value, SOR_PWR);
+
+	timeout = jiffies + msecs_to_jiffies(timeout);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_PWR);
+		if ((value & SOR_PWR_TRIGGER) == 0)
+			return 0;
+
+		usleep_range(25, 100);
+	}
+
+	return -ETIMEDOUT;
+}
+
+struct tegra_sor_params {
+	/* number of link clocks per line */
+	unsigned int num_clocks;
+	/* ratio between input and output */
+	u64 ratio;
+	/* precision factor */
+	u64 precision;
+
+	unsigned int active_polarity;
+	unsigned int active_count;
+	unsigned int active_frac;
+	unsigned int tu_size;
+	unsigned int error;
+};
+
+static int tegra_sor_compute_params(struct tegra_sor *sor,
+				    struct tegra_sor_params *params,
+				    unsigned int tu_size)
+{
+	u64 active_sym, active_count, frac, approx;
+	u32 active_polarity, active_frac = 0;
+	const u64 f = params->precision;
+	s64 error;
+
+	active_sym = params->ratio * tu_size;
+	active_count = div_u64(active_sym, f) * f;
+	frac = active_sym - active_count;
+
+	/* fraction < 0.5 */
+	if (frac >= (f / 2)) {
+		active_polarity = 1;
+		frac = f - frac;
+	} else {
+		active_polarity = 0;
+	}
+
+	if (frac != 0) {
+		frac = div_u64(f * f,  frac); /* 1/fraction */
+		if (frac <= (15 * f)) {
+			active_frac = div_u64(frac, f);
+
+			/* round up */
+			if (active_polarity)
+				active_frac++;
+		} else {
+			active_frac = active_polarity ? 1 : 15;
+		}
+	}
+
+	if (active_frac == 1)
+		active_polarity = 0;
+
+	if (active_polarity == 1) {
+		if (active_frac) {
+			approx = active_count + (active_frac * (f - 1)) * f;
+			approx = div_u64(approx, active_frac * f);
+		} else {
+			approx = active_count + f;
+		}
+	} else {
+		if (active_frac)
+			approx = active_count + div_u64(f, active_frac);
+		else
+			approx = active_count;
+	}
+
+	error = div_s64(active_sym - approx, tu_size);
+	error *= params->num_clocks;
+
+	if (error <= 0 && abs(error) < params->error) {
+		params->active_count = div_u64(active_count, f);
+		params->active_polarity = active_polarity;
+		params->active_frac = active_frac;
+		params->error = abs(error);
+		params->tu_size = tu_size;
+
+		if (error == 0)
+			return true;
+	}
+
+	return false;
+}
+
+static int tegra_sor_compute_config(struct tegra_sor *sor,
+				    const struct drm_display_mode *mode,
+				    struct tegra_sor_config *config,
+				    struct drm_dp_link *link)
+{
+	const u64 f = 100000, link_rate = link->rate * 1000;
+	const u64 pclk = (u64)mode->clock * 1000;
+	u64 input, output, watermark, num;
+	struct tegra_sor_params params;
+	u32 num_syms_per_line;
+	unsigned int i;
+
+	if (!link_rate || !link->lanes || !pclk || !config->bits_per_pixel)
+		return -EINVAL;
+
+	input = pclk * config->bits_per_pixel;
+	output = link_rate * 8 * link->lanes;
+
+	if (input >= output)
+		return -ERANGE;
+
+	memset(&params, 0, sizeof(params));
+	params.ratio = div64_u64(input * f, output);
+	params.num_clocks = div_u64(link_rate * mode->hdisplay, pclk);
+	params.precision = f;
+	params.error = 64 * f;
+	params.tu_size = 64;
+
+	for (i = params.tu_size; i >= 32; i--)
+		if (tegra_sor_compute_params(sor, &params, i))
+			break;
+
+	if (params.active_frac == 0) {
+		config->active_polarity = 0;
+		config->active_count = params.active_count;
+
+		if (!params.active_polarity)
+			config->active_count--;
+
+		config->tu_size = params.tu_size;
+		config->active_frac = 1;
+	} else {
+		config->active_polarity = params.active_polarity;
+		config->active_count = params.active_count;
+		config->active_frac = params.active_frac;
+		config->tu_size = params.tu_size;
+	}
+
+	dev_dbg(sor->dev,
+		"polarity: %d active count: %d tu size: %d active frac: %d\n",
+		config->active_polarity, config->active_count,
+		config->tu_size, config->active_frac);
+
+	watermark = params.ratio * config->tu_size * (f - params.ratio);
+	watermark = div_u64(watermark, f);
+
+	watermark = div_u64(watermark + params.error, f);
+	config->watermark = watermark + (config->bits_per_pixel / 8) + 2;
+	num_syms_per_line = (mode->hdisplay * config->bits_per_pixel) *
+			    (link->lanes * 8);
+
+	if (config->watermark > 30) {
+		config->watermark = 30;
+		dev_err(sor->dev,
+			"unable to compute TU size, forcing watermark to %u\n",
+			config->watermark);
+	} else if (config->watermark > num_syms_per_line) {
+		config->watermark = num_syms_per_line;
+		dev_err(sor->dev, "watermark too high, forcing to %u\n",
+			config->watermark);
+	}
+
+	/* compute the number of symbols per horizontal blanking interval */
+	num = ((mode->htotal - mode->hdisplay) - 7) * link_rate;
+	config->hblank_symbols = div_u64(num, pclk);
+
+	if (link->caps.enhanced_framing)
+		config->hblank_symbols -= 3;
+
+	config->hblank_symbols -= 12 / link->lanes;
+
+	/* compute the number of symbols per vertical blanking interval */
+	num = (mode->hdisplay - 25) * link_rate;
+	config->vblank_symbols = div_u64(num, pclk);
+	config->vblank_symbols -= 36 / link->lanes + 4;
+
+	dev_dbg(sor->dev, "blank symbols: H:%u V:%u\n", config->hblank_symbols,
+		config->vblank_symbols);
+
+	return 0;
+}
+
+static void tegra_sor_apply_config(struct tegra_sor *sor,
+				   const struct tegra_sor_config *config)
+{
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_DP_LINKCTL0);
+	value &= ~SOR_DP_LINKCTL_TU_SIZE_MASK;
+	value |= SOR_DP_LINKCTL_TU_SIZE(config->tu_size);
+	tegra_sor_writel(sor, value, SOR_DP_LINKCTL0);
+
+	value = tegra_sor_readl(sor, SOR_DP_CONFIG0);
+	value &= ~SOR_DP_CONFIG_WATERMARK_MASK;
+	value |= SOR_DP_CONFIG_WATERMARK(config->watermark);
+
+	value &= ~SOR_DP_CONFIG_ACTIVE_SYM_COUNT_MASK;
+	value |= SOR_DP_CONFIG_ACTIVE_SYM_COUNT(config->active_count);
+
+	value &= ~SOR_DP_CONFIG_ACTIVE_SYM_FRAC_MASK;
+	value |= SOR_DP_CONFIG_ACTIVE_SYM_FRAC(config->active_frac);
+
+	if (config->active_polarity)
+		value |= SOR_DP_CONFIG_ACTIVE_SYM_POLARITY;
+	else
+		value &= ~SOR_DP_CONFIG_ACTIVE_SYM_POLARITY;
+
+	value |= SOR_DP_CONFIG_ACTIVE_SYM_ENABLE;
+	value |= SOR_DP_CONFIG_DISPARITY_NEGATIVE;
+	tegra_sor_writel(sor, value, SOR_DP_CONFIG0);
+
+	value = tegra_sor_readl(sor, SOR_DP_AUDIO_HBLANK_SYMBOLS);
+	value &= ~SOR_DP_AUDIO_HBLANK_SYMBOLS_MASK;
+	value |= config->hblank_symbols & 0xffff;
+	tegra_sor_writel(sor, value, SOR_DP_AUDIO_HBLANK_SYMBOLS);
+
+	value = tegra_sor_readl(sor, SOR_DP_AUDIO_VBLANK_SYMBOLS);
+	value &= ~SOR_DP_AUDIO_VBLANK_SYMBOLS_MASK;
+	value |= config->vblank_symbols & 0xffff;
+	tegra_sor_writel(sor, value, SOR_DP_AUDIO_VBLANK_SYMBOLS);
+}
+
+static void tegra_sor_mode_set(struct tegra_sor *sor,
+			       const struct drm_display_mode *mode,
+			       struct tegra_sor_state *state)
+{
+	struct tegra_dc *dc = to_tegra_dc(sor->output.encoder.crtc);
+	unsigned int vbe, vse, hbe, hse, vbs, hbs;
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_STATE1);
+	value &= ~SOR_STATE_ASY_PIXELDEPTH_MASK;
+	value &= ~SOR_STATE_ASY_CRC_MODE_MASK;
+	value &= ~SOR_STATE_ASY_OWNER_MASK;
+
+	value |= SOR_STATE_ASY_CRC_MODE_COMPLETE |
+		 SOR_STATE_ASY_OWNER(dc->pipe + 1);
+
+	if (mode->flags & DRM_MODE_FLAG_PHSYNC)
+		value &= ~SOR_STATE_ASY_HSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+		value |= SOR_STATE_ASY_HSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_PVSYNC)
+		value &= ~SOR_STATE_ASY_VSYNCPOL;
+
+	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+		value |= SOR_STATE_ASY_VSYNCPOL;
+
+	switch (state->bpc) {
+	case 16:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_48_444;
+		break;
+
+	case 12:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_36_444;
+		break;
+
+	case 10:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_30_444;
+		break;
+
+	case 8:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444;
+		break;
+
+	case 6:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_18_444;
+		break;
+
+	default:
+		value |= SOR_STATE_ASY_PIXELDEPTH_BPP_24_444;
+		break;
+	}
+
+	tegra_sor_writel(sor, value, SOR_STATE1);
+
+	/*
+	 * TODO: The video timing programming below doesn't seem to match the
+	 * register definitions.
+	 */
+
+	value = ((mode->vtotal & 0x7fff) << 16) | (mode->htotal & 0x7fff);
+	tegra_sor_writel(sor, value, sor->soc->regs->head_state1 + dc->pipe);
+
+	/* sync end = sync width - 1 */
+	vse = mode->vsync_end - mode->vsync_start - 1;
+	hse = mode->hsync_end - mode->hsync_start - 1;
+
+	value = ((vse & 0x7fff) << 16) | (hse & 0x7fff);
+	tegra_sor_writel(sor, value, sor->soc->regs->head_state2 + dc->pipe);
+
+	/* blank end = sync end + back porch */
+	vbe = vse + (mode->vtotal - mode->vsync_end);
+	hbe = hse + (mode->htotal - mode->hsync_end);
+
+	value = ((vbe & 0x7fff) << 16) | (hbe & 0x7fff);
+	tegra_sor_writel(sor, value, sor->soc->regs->head_state3 + dc->pipe);
+
+	/* blank start = blank end + active */
+	vbs = vbe + mode->vdisplay;
+	hbs = hbe + mode->hdisplay;
+
+	value = ((vbs & 0x7fff) << 16) | (hbs & 0x7fff);
+	tegra_sor_writel(sor, value, sor->soc->regs->head_state4 + dc->pipe);
+
+	/* XXX interlacing support */
+	tegra_sor_writel(sor, 0x001, sor->soc->regs->head_state5 + dc->pipe);
+}
+
+static int tegra_sor_detach(struct tegra_sor *sor)
+{
+	unsigned long value, timeout;
+
+	/* switch to safe mode */
+	value = tegra_sor_readl(sor, SOR_SUPER_STATE1);
+	value &= ~SOR_SUPER_STATE_MODE_NORMAL;
+	tegra_sor_writel(sor, value, SOR_SUPER_STATE1);
+	tegra_sor_super_update(sor);
+
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_PWR);
+		if (value & SOR_PWR_MODE_SAFE)
+			break;
+	}
+
+	if ((value & SOR_PWR_MODE_SAFE) == 0)
+		return -ETIMEDOUT;
+
+	/* go to sleep */
+	value = tegra_sor_readl(sor, SOR_SUPER_STATE1);
+	value &= ~SOR_SUPER_STATE_HEAD_MODE_MASK;
+	tegra_sor_writel(sor, value, SOR_SUPER_STATE1);
+	tegra_sor_super_update(sor);
+
+	/* detach */
+	value = tegra_sor_readl(sor, SOR_SUPER_STATE1);
+	value &= ~SOR_SUPER_STATE_ATTACHED;
+	tegra_sor_writel(sor, value, SOR_SUPER_STATE1);
+	tegra_sor_super_update(sor);
+
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_TEST);
+		if ((value & SOR_TEST_ATTACHED) == 0)
+			break;
+
+		usleep_range(25, 100);
+	}
+
+	if ((value & SOR_TEST_ATTACHED) != 0)
+		return -ETIMEDOUT;
+
+	return 0;
+}
+
+static int tegra_sor_power_down(struct tegra_sor *sor)
+{
+	unsigned long value, timeout;
+	int err;
+
+	value = tegra_sor_readl(sor, SOR_PWR);
+	value &= ~SOR_PWR_NORMAL_STATE_PU;
+	value |= SOR_PWR_TRIGGER;
+	tegra_sor_writel(sor, value, SOR_PWR);
+
+	timeout = jiffies + msecs_to_jiffies(250);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_PWR);
+		if ((value & SOR_PWR_TRIGGER) == 0)
+			return 0;
+
+		usleep_range(25, 100);
+	}
+
+	if ((value & SOR_PWR_TRIGGER) != 0)
+		return -ETIMEDOUT;
+
+	/* switch to safe parent clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_safe);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to set safe parent clock: %d\n", err);
+		return err;
+	}
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll2);
+	value |= SOR_PLL2_PORT_POWERDOWN;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll2);
+
+	usleep_range(20, 100);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll0);
+	value |= SOR_PLL0_VCOPD | SOR_PLL0_PWR;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll0);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll2);
+	value |= SOR_PLL2_SEQ_PLLCAPPD;
+	value |= SOR_PLL2_SEQ_PLLCAPPD_ENFORCE;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll2);
+
+	usleep_range(20, 100);
+
+	return 0;
+}
+
+static int tegra_sor_crc_wait(struct tegra_sor *sor, unsigned long timeout)
+{
+	u32 value;
+
+	timeout = jiffies + msecs_to_jiffies(timeout);
+
+	while (time_before(jiffies, timeout)) {
+		value = tegra_sor_readl(sor, SOR_CRCA);
+		if (value & SOR_CRCA_VALID)
+			return 0;
+
+		usleep_range(100, 200);
+	}
+
+	return -ETIMEDOUT;
+}
+
+static int tegra_sor_show_crc(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = s->private;
+	struct tegra_sor *sor = node->info_ent->data;
+	struct drm_crtc *crtc = sor->output.encoder.crtc;
+	struct drm_device *drm = node->minor->dev;
+	int err = 0;
+	u32 value;
+
+	drm_modeset_lock_all(drm);
+
+	if (!crtc || !crtc->state->active) {
+		err = -EBUSY;
+		goto unlock;
+	}
+
+	value = tegra_sor_readl(sor, SOR_STATE1);
+	value &= ~SOR_STATE_ASY_CRC_MODE_MASK;
+	tegra_sor_writel(sor, value, SOR_STATE1);
+
+	value = tegra_sor_readl(sor, SOR_CRC_CNTRL);
+	value |= SOR_CRC_CNTRL_ENABLE;
+	tegra_sor_writel(sor, value, SOR_CRC_CNTRL);
+
+	value = tegra_sor_readl(sor, SOR_TEST);
+	value &= ~SOR_TEST_CRC_POST_SERIALIZE;
+	tegra_sor_writel(sor, value, SOR_TEST);
+
+	err = tegra_sor_crc_wait(sor, 100);
+	if (err < 0)
+		goto unlock;
+
+	tegra_sor_writel(sor, SOR_CRCA_RESET, SOR_CRCA);
+	value = tegra_sor_readl(sor, SOR_CRCB);
+
+	seq_printf(s, "%08x\n", value);
+
+unlock:
+	drm_modeset_unlock_all(drm);
+	return err;
+}
+
+#define DEBUGFS_REG32(_name) { .name = #_name, .offset = _name }
+
+static const struct debugfs_reg32 tegra_sor_regs[] = {
+	DEBUGFS_REG32(SOR_CTXSW),
+	DEBUGFS_REG32(SOR_SUPER_STATE0),
+	DEBUGFS_REG32(SOR_SUPER_STATE1),
+	DEBUGFS_REG32(SOR_STATE0),
+	DEBUGFS_REG32(SOR_STATE1),
+	DEBUGFS_REG32(SOR_HEAD_STATE0(0)),
+	DEBUGFS_REG32(SOR_HEAD_STATE0(1)),
+	DEBUGFS_REG32(SOR_HEAD_STATE1(0)),
+	DEBUGFS_REG32(SOR_HEAD_STATE1(1)),
+	DEBUGFS_REG32(SOR_HEAD_STATE2(0)),
+	DEBUGFS_REG32(SOR_HEAD_STATE2(1)),
+	DEBUGFS_REG32(SOR_HEAD_STATE3(0)),
+	DEBUGFS_REG32(SOR_HEAD_STATE3(1)),
+	DEBUGFS_REG32(SOR_HEAD_STATE4(0)),
+	DEBUGFS_REG32(SOR_HEAD_STATE4(1)),
+	DEBUGFS_REG32(SOR_HEAD_STATE5(0)),
+	DEBUGFS_REG32(SOR_HEAD_STATE5(1)),
+	DEBUGFS_REG32(SOR_CRC_CNTRL),
+	DEBUGFS_REG32(SOR_DP_DEBUG_MVID),
+	DEBUGFS_REG32(SOR_CLK_CNTRL),
+	DEBUGFS_REG32(SOR_CAP),
+	DEBUGFS_REG32(SOR_PWR),
+	DEBUGFS_REG32(SOR_TEST),
+	DEBUGFS_REG32(SOR_PLL0),
+	DEBUGFS_REG32(SOR_PLL1),
+	DEBUGFS_REG32(SOR_PLL2),
+	DEBUGFS_REG32(SOR_PLL3),
+	DEBUGFS_REG32(SOR_CSTM),
+	DEBUGFS_REG32(SOR_LVDS),
+	DEBUGFS_REG32(SOR_CRCA),
+	DEBUGFS_REG32(SOR_CRCB),
+	DEBUGFS_REG32(SOR_BLANK),
+	DEBUGFS_REG32(SOR_SEQ_CTL),
+	DEBUGFS_REG32(SOR_LANE_SEQ_CTL),
+	DEBUGFS_REG32(SOR_SEQ_INST(0)),
+	DEBUGFS_REG32(SOR_SEQ_INST(1)),
+	DEBUGFS_REG32(SOR_SEQ_INST(2)),
+	DEBUGFS_REG32(SOR_SEQ_INST(3)),
+	DEBUGFS_REG32(SOR_SEQ_INST(4)),
+	DEBUGFS_REG32(SOR_SEQ_INST(5)),
+	DEBUGFS_REG32(SOR_SEQ_INST(6)),
+	DEBUGFS_REG32(SOR_SEQ_INST(7)),
+	DEBUGFS_REG32(SOR_SEQ_INST(8)),
+	DEBUGFS_REG32(SOR_SEQ_INST(9)),
+	DEBUGFS_REG32(SOR_SEQ_INST(10)),
+	DEBUGFS_REG32(SOR_SEQ_INST(11)),
+	DEBUGFS_REG32(SOR_SEQ_INST(12)),
+	DEBUGFS_REG32(SOR_SEQ_INST(13)),
+	DEBUGFS_REG32(SOR_SEQ_INST(14)),
+	DEBUGFS_REG32(SOR_SEQ_INST(15)),
+	DEBUGFS_REG32(SOR_PWM_DIV),
+	DEBUGFS_REG32(SOR_PWM_CTL),
+	DEBUGFS_REG32(SOR_VCRC_A0),
+	DEBUGFS_REG32(SOR_VCRC_A1),
+	DEBUGFS_REG32(SOR_VCRC_B0),
+	DEBUGFS_REG32(SOR_VCRC_B1),
+	DEBUGFS_REG32(SOR_CCRC_A0),
+	DEBUGFS_REG32(SOR_CCRC_A1),
+	DEBUGFS_REG32(SOR_CCRC_B0),
+	DEBUGFS_REG32(SOR_CCRC_B1),
+	DEBUGFS_REG32(SOR_EDATA_A0),
+	DEBUGFS_REG32(SOR_EDATA_A1),
+	DEBUGFS_REG32(SOR_EDATA_B0),
+	DEBUGFS_REG32(SOR_EDATA_B1),
+	DEBUGFS_REG32(SOR_COUNT_A0),
+	DEBUGFS_REG32(SOR_COUNT_A1),
+	DEBUGFS_REG32(SOR_COUNT_B0),
+	DEBUGFS_REG32(SOR_COUNT_B1),
+	DEBUGFS_REG32(SOR_DEBUG_A0),
+	DEBUGFS_REG32(SOR_DEBUG_A1),
+	DEBUGFS_REG32(SOR_DEBUG_B0),
+	DEBUGFS_REG32(SOR_DEBUG_B1),
+	DEBUGFS_REG32(SOR_TRIG),
+	DEBUGFS_REG32(SOR_MSCHECK),
+	DEBUGFS_REG32(SOR_XBAR_CTRL),
+	DEBUGFS_REG32(SOR_XBAR_POL),
+	DEBUGFS_REG32(SOR_DP_LINKCTL0),
+	DEBUGFS_REG32(SOR_DP_LINKCTL1),
+	DEBUGFS_REG32(SOR_LANE_DRIVE_CURRENT0),
+	DEBUGFS_REG32(SOR_LANE_DRIVE_CURRENT1),
+	DEBUGFS_REG32(SOR_LANE4_DRIVE_CURRENT0),
+	DEBUGFS_REG32(SOR_LANE4_DRIVE_CURRENT1),
+	DEBUGFS_REG32(SOR_LANE_PREEMPHASIS0),
+	DEBUGFS_REG32(SOR_LANE_PREEMPHASIS1),
+	DEBUGFS_REG32(SOR_LANE4_PREEMPHASIS0),
+	DEBUGFS_REG32(SOR_LANE4_PREEMPHASIS1),
+	DEBUGFS_REG32(SOR_LANE_POSTCURSOR0),
+	DEBUGFS_REG32(SOR_LANE_POSTCURSOR1),
+	DEBUGFS_REG32(SOR_DP_CONFIG0),
+	DEBUGFS_REG32(SOR_DP_CONFIG1),
+	DEBUGFS_REG32(SOR_DP_MN0),
+	DEBUGFS_REG32(SOR_DP_MN1),
+	DEBUGFS_REG32(SOR_DP_PADCTL0),
+	DEBUGFS_REG32(SOR_DP_PADCTL1),
+	DEBUGFS_REG32(SOR_DP_PADCTL2),
+	DEBUGFS_REG32(SOR_DP_DEBUG0),
+	DEBUGFS_REG32(SOR_DP_DEBUG1),
+	DEBUGFS_REG32(SOR_DP_SPARE0),
+	DEBUGFS_REG32(SOR_DP_SPARE1),
+	DEBUGFS_REG32(SOR_DP_AUDIO_CTRL),
+	DEBUGFS_REG32(SOR_DP_AUDIO_HBLANK_SYMBOLS),
+	DEBUGFS_REG32(SOR_DP_AUDIO_VBLANK_SYMBOLS),
+	DEBUGFS_REG32(SOR_DP_GENERIC_INFOFRAME_HEADER),
+	DEBUGFS_REG32(SOR_DP_GENERIC_INFOFRAME_SUBPACK0),
+	DEBUGFS_REG32(SOR_DP_GENERIC_INFOFRAME_SUBPACK1),
+	DEBUGFS_REG32(SOR_DP_GENERIC_INFOFRAME_SUBPACK2),
+	DEBUGFS_REG32(SOR_DP_GENERIC_INFOFRAME_SUBPACK3),
+	DEBUGFS_REG32(SOR_DP_GENERIC_INFOFRAME_SUBPACK4),
+	DEBUGFS_REG32(SOR_DP_GENERIC_INFOFRAME_SUBPACK5),
+	DEBUGFS_REG32(SOR_DP_GENERIC_INFOFRAME_SUBPACK6),
+	DEBUGFS_REG32(SOR_DP_TPG),
+	DEBUGFS_REG32(SOR_DP_TPG_CONFIG),
+	DEBUGFS_REG32(SOR_DP_LQ_CSTM0),
+	DEBUGFS_REG32(SOR_DP_LQ_CSTM1),
+	DEBUGFS_REG32(SOR_DP_LQ_CSTM2),
+};
+
+static int tegra_sor_show_regs(struct seq_file *s, void *data)
+{
+	struct drm_info_node *node = s->private;
+	struct tegra_sor *sor = node->info_ent->data;
+	struct drm_crtc *crtc = sor->output.encoder.crtc;
+	struct drm_device *drm = node->minor->dev;
+	unsigned int i;
+	int err = 0;
+
+	drm_modeset_lock_all(drm);
+
+	if (!crtc || !crtc->state->active) {
+		err = -EBUSY;
+		goto unlock;
+	}
+
+	for (i = 0; i < ARRAY_SIZE(tegra_sor_regs); i++) {
+		unsigned int offset = tegra_sor_regs[i].offset;
+
+		seq_printf(s, "%-38s %#05x %08x\n", tegra_sor_regs[i].name,
+			   offset, tegra_sor_readl(sor, offset));
+	}
+
+unlock:
+	drm_modeset_unlock_all(drm);
+	return err;
+}
+
+static const struct drm_info_list debugfs_files[] = {
+	{ "crc", tegra_sor_show_crc, 0, NULL },
+	{ "regs", tegra_sor_show_regs, 0, NULL },
+};
+
+static int tegra_sor_late_register(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	unsigned int i, count = ARRAY_SIZE(debugfs_files);
+	struct drm_minor *minor = connector->dev->primary;
+	struct dentry *root = connector->debugfs_entry;
+	struct tegra_sor *sor = to_sor(output);
+
+	sor->debugfs_files = kmemdup(debugfs_files, sizeof(debugfs_files),
+				     GFP_KERNEL);
+	if (!sor->debugfs_files)
+		return -ENOMEM;
+
+	for (i = 0; i < count; i++)
+		sor->debugfs_files[i].data = sor;
+
+	drm_debugfs_create_files(sor->debugfs_files, count, root, minor);
+
+	return 0;
+}
+
+static void tegra_sor_early_unregister(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	unsigned int count = ARRAY_SIZE(debugfs_files);
+	struct tegra_sor *sor = to_sor(output);
+
+	drm_debugfs_remove_files(sor->debugfs_files, count,
+				 connector->debugfs_entry,
+				 connector->dev->primary);
+	kfree(sor->debugfs_files);
+	sor->debugfs_files = NULL;
+}
+
+static void tegra_sor_connector_reset(struct drm_connector *connector)
+{
+	struct tegra_sor_state *state;
+
+	state = kzalloc_obj(*state);
+	if (!state)
+		return;
+
+	if (connector->state) {
+		__drm_atomic_helper_connector_destroy_state(connector->state);
+		kfree(connector->state);
+	}
+
+	__drm_atomic_helper_connector_reset(connector, &state->base);
+}
+
+static enum drm_connector_status
+tegra_sor_connector_detect(struct drm_connector *connector, bool force)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	struct tegra_sor *sor = to_sor(output);
+
+	if (sor->aux)
+		return drm_dp_aux_detect(sor->aux);
+
+	return tegra_output_connector_detect(connector, force);
+}
+
+static struct drm_connector_state *
+tegra_sor_connector_duplicate_state(struct drm_connector *connector)
+{
+	struct tegra_sor_state *state = to_sor_state(connector->state);
+	struct tegra_sor_state *copy;
+
+	copy = kmemdup(state, sizeof(*state), GFP_KERNEL);
+	if (!copy)
+		return NULL;
+
+	__drm_atomic_helper_connector_duplicate_state(connector, &copy->base);
+
+	return &copy->base;
+}
+
+static const struct drm_connector_funcs tegra_sor_connector_funcs = {
+	.reset = tegra_sor_connector_reset,
+	.detect = tegra_sor_connector_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.destroy = tegra_output_connector_destroy,
+	.atomic_duplicate_state = tegra_sor_connector_duplicate_state,
+	.atomic_destroy_state = drm_atomic_helper_connector_destroy_state,
+	.late_register = tegra_sor_late_register,
+	.early_unregister = tegra_sor_early_unregister,
+};
+
+static int tegra_sor_connector_get_modes(struct drm_connector *connector)
+{
+	struct tegra_output *output = connector_to_output(connector);
+	struct tegra_sor *sor = to_sor(output);
+	int err;
+
+	if (sor->aux)
+		drm_dp_aux_enable(sor->aux);
+
+	err = tegra_output_connector_get_modes(connector);
+
+	if (sor->aux)
+		drm_dp_aux_disable(sor->aux);
+
+	return err;
+}
+
+static enum drm_mode_status
+tegra_sor_connector_mode_valid(struct drm_connector *connector,
+			       const struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static const struct drm_connector_helper_funcs tegra_sor_connector_helper_funcs = {
+	.get_modes = tegra_sor_connector_get_modes,
+	.mode_valid = tegra_sor_connector_mode_valid,
+};
+
+static int
+tegra_sor_encoder_atomic_check(struct drm_encoder *encoder,
+			       struct drm_crtc_state *crtc_state,
+			       struct drm_connector_state *conn_state)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_sor_state *state = to_sor_state(conn_state);
+	struct tegra_dc *dc = to_tegra_dc(conn_state->crtc);
+	unsigned long pclk = crtc_state->mode.clock * 1000;
+	struct tegra_sor *sor = to_sor(output);
+	struct drm_display_info *info;
+	int err;
+
+	info = &output->connector.display_info;
+
+	/*
+	 * For HBR2 modes, the SOR brick needs to use the x20 multiplier, so
+	 * the pixel clock must be corrected accordingly.
+	 */
+	if (pclk >= 340000000) {
+		state->link_speed = 20;
+		state->pclk = pclk / 2;
+	} else {
+		state->link_speed = 10;
+		state->pclk = pclk;
+	}
+
+	err = tegra_dc_state_setup_clock(dc, crtc_state, sor->clk_parent,
+					 pclk, 0);
+	if (err < 0) {
+		dev_err(output->dev, "failed to setup CRTC state: %d\n", err);
+		return err;
+	}
+
+	switch (info->bpc) {
+	case 8:
+	case 6:
+		state->bpc = info->bpc;
+		break;
+
+	default:
+		DRM_DEBUG_KMS("%u bits-per-color not supported\n", info->bpc);
+		state->bpc = 8;
+		break;
+	}
+
+	return 0;
+}
+
+static inline u32 tegra_sor_hdmi_subpack(const u8 *ptr, size_t size)
+{
+	u32 value = 0;
+	size_t i;
+
+	for (i = size; i > 0; i--)
+		value = (value << 8) | ptr[i - 1];
+
+	return value;
+}
+
+static void tegra_sor_hdmi_write_infopack(struct tegra_sor *sor,
+					  const void *data, size_t size)
+{
+	const u8 *ptr = data;
+	unsigned long offset;
+	size_t i;
+	u32 value;
+
+	switch (ptr[0]) {
+	case HDMI_INFOFRAME_TYPE_AVI:
+		offset = SOR_HDMI_AVI_INFOFRAME_HEADER;
+		break;
+
+	case HDMI_INFOFRAME_TYPE_AUDIO:
+		offset = SOR_HDMI_AUDIO_INFOFRAME_HEADER;
+		break;
+
+	case HDMI_INFOFRAME_TYPE_VENDOR:
+		offset = SOR_HDMI_VSI_INFOFRAME_HEADER;
+		break;
+
+	default:
+		dev_err(sor->dev, "unsupported infoframe type: %02x\n",
+			ptr[0]);
+		return;
+	}
+
+	value = INFOFRAME_HEADER_TYPE(ptr[0]) |
+		INFOFRAME_HEADER_VERSION(ptr[1]) |
+		INFOFRAME_HEADER_LEN(ptr[2]);
+	tegra_sor_writel(sor, value, offset);
+	offset++;
+
+	/*
+	 * Each subpack contains 7 bytes, divided into:
+	 * - subpack_low: bytes 0 - 3
+	 * - subpack_high: bytes 4 - 6 (with byte 7 padded to 0x00)
+	 */
+	for (i = 3; i < size; i += 7) {
+		size_t rem = size - i, num = min_t(size_t, rem, 4);
+
+		value = tegra_sor_hdmi_subpack(&ptr[i], num);
+		tegra_sor_writel(sor, value, offset++);
+
+		num = min_t(size_t, rem - num, 3);
+
+		value = tegra_sor_hdmi_subpack(&ptr[i + 4], num);
+		tegra_sor_writel(sor, value, offset++);
+	}
+}
+
+static int
+tegra_sor_hdmi_setup_avi_infoframe(struct tegra_sor *sor,
+				   const struct drm_display_mode *mode)
+{
+	u8 buffer[HDMI_INFOFRAME_SIZE(AVI)];
+	struct hdmi_avi_infoframe frame;
+	u32 value;
+	int err;
+
+	/* disable AVI infoframe */
+	value = tegra_sor_readl(sor, SOR_HDMI_AVI_INFOFRAME_CTRL);
+	value &= ~INFOFRAME_CTRL_SINGLE;
+	value &= ~INFOFRAME_CTRL_OTHER;
+	value &= ~INFOFRAME_CTRL_ENABLE;
+	tegra_sor_writel(sor, value, SOR_HDMI_AVI_INFOFRAME_CTRL);
+
+	err = drm_hdmi_avi_infoframe_from_display_mode(&frame,
+						       &sor->output.connector, mode);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to setup AVI infoframe: %d\n", err);
+		return err;
+	}
+
+	err = hdmi_avi_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		dev_err(sor->dev, "failed to pack AVI infoframe: %d\n", err);
+		return err;
+	}
+
+	tegra_sor_hdmi_write_infopack(sor, buffer, err);
+
+	/* enable AVI infoframe */
+	value = tegra_sor_readl(sor, SOR_HDMI_AVI_INFOFRAME_CTRL);
+	value |= INFOFRAME_CTRL_CHECKSUM_ENABLE;
+	value |= INFOFRAME_CTRL_ENABLE;
+	tegra_sor_writel(sor, value, SOR_HDMI_AVI_INFOFRAME_CTRL);
+
+	return 0;
+}
+
+static void tegra_sor_write_eld(struct tegra_sor *sor)
+{
+	size_t length = drm_eld_size(sor->output.connector.eld), i;
+
+	for (i = 0; i < length; i++)
+		tegra_sor_writel(sor, i << 8 | sor->output.connector.eld[i],
+				 SOR_AUDIO_HDA_ELD_BUFWR);
+
+	/*
+	 * The HDA codec will always report an ELD buffer size of 96 bytes and
+	 * the HDA codec driver will check that each byte read from the buffer
+	 * is valid. Therefore every byte must be written, even if no 96 bytes
+	 * were parsed from EDID.
+	 */
+	for (i = length; i < 96; i++)
+		tegra_sor_writel(sor, i << 8 | 0, SOR_AUDIO_HDA_ELD_BUFWR);
+}
+
+static void tegra_sor_audio_prepare(struct tegra_sor *sor)
+{
+	u32 value;
+
+	/*
+	 * Enable and unmask the HDA codec SCRATCH0 register interrupt. This
+	 * is used for interoperability between the HDA codec driver and the
+	 * HDMI/DP driver.
+	 */
+	value = SOR_INT_CODEC_SCRATCH1 | SOR_INT_CODEC_SCRATCH0;
+	tegra_sor_writel(sor, value, SOR_INT_ENABLE);
+	tegra_sor_writel(sor, value, SOR_INT_MASK);
+
+	tegra_sor_write_eld(sor);
+
+	value = SOR_AUDIO_HDA_PRESENSE_ELDV | SOR_AUDIO_HDA_PRESENSE_PD;
+	tegra_sor_writel(sor, value, SOR_AUDIO_HDA_PRESENSE);
+}
+
+static void tegra_sor_audio_unprepare(struct tegra_sor *sor)
+{
+	tegra_sor_writel(sor, 0, SOR_AUDIO_HDA_PRESENSE);
+	tegra_sor_writel(sor, 0, SOR_INT_MASK);
+	tegra_sor_writel(sor, 0, SOR_INT_ENABLE);
+}
+
+static void tegra_sor_audio_enable(struct tegra_sor *sor)
+{
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_AUDIO_CNTRL);
+
+	/* select HDA audio input */
+	value &= ~SOR_AUDIO_CNTRL_SOURCE_SELECT(SOURCE_SELECT_MASK);
+	value |= SOR_AUDIO_CNTRL_SOURCE_SELECT(SOURCE_SELECT_HDA);
+
+	/* inject null samples */
+	if (sor->format.channels != 2)
+		value &= ~SOR_AUDIO_CNTRL_INJECT_NULLSMPL;
+	else
+		value |= SOR_AUDIO_CNTRL_INJECT_NULLSMPL;
+
+	value |= SOR_AUDIO_CNTRL_AFIFO_FLUSH;
+
+	tegra_sor_writel(sor, value, SOR_AUDIO_CNTRL);
+
+	/* enable advertising HBR capability */
+	tegra_sor_writel(sor, SOR_AUDIO_SPARE_HBR_ENABLE, SOR_AUDIO_SPARE);
+}
+
+static int tegra_sor_hdmi_enable_audio_infoframe(struct tegra_sor *sor)
+{
+	u8 buffer[HDMI_INFOFRAME_SIZE(AUDIO)];
+	struct hdmi_audio_infoframe frame;
+	u32 value;
+	int err;
+
+	err = hdmi_audio_infoframe_init(&frame);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to setup audio infoframe: %d\n", err);
+		return err;
+	}
+
+	frame.channels = sor->format.channels;
+
+	err = hdmi_audio_infoframe_pack(&frame, buffer, sizeof(buffer));
+	if (err < 0) {
+		dev_err(sor->dev, "failed to pack audio infoframe: %d\n", err);
+		return err;
+	}
+
+	tegra_sor_hdmi_write_infopack(sor, buffer, err);
+
+	value = tegra_sor_readl(sor, SOR_HDMI_AUDIO_INFOFRAME_CTRL);
+	value |= INFOFRAME_CTRL_CHECKSUM_ENABLE;
+	value |= INFOFRAME_CTRL_ENABLE;
+	tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_INFOFRAME_CTRL);
+
+	return 0;
+}
+
+static void tegra_sor_hdmi_audio_enable(struct tegra_sor *sor)
+{
+	u32 value;
+
+	tegra_sor_audio_enable(sor);
+
+	tegra_sor_writel(sor, 0, SOR_HDMI_ACR_CTRL);
+
+	value = SOR_HDMI_SPARE_ACR_PRIORITY_HIGH |
+		SOR_HDMI_SPARE_CTS_RESET(1) |
+		SOR_HDMI_SPARE_HW_CTS_ENABLE;
+	tegra_sor_writel(sor, value, SOR_HDMI_SPARE);
+
+	/* enable HW CTS */
+	value = SOR_HDMI_ACR_SUBPACK_LOW_SB1(0);
+	tegra_sor_writel(sor, value, SOR_HDMI_ACR_0441_SUBPACK_LOW);
+
+	/* allow packet to be sent */
+	value = SOR_HDMI_ACR_SUBPACK_HIGH_ENABLE;
+	tegra_sor_writel(sor, value, SOR_HDMI_ACR_0441_SUBPACK_HIGH);
+
+	/* reset N counter and enable lookup */
+	value = SOR_HDMI_AUDIO_N_RESET | SOR_HDMI_AUDIO_N_LOOKUP;
+	tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_N);
+
+	value = (24000 * 4096) / (128 * sor->format.sample_rate / 1000);
+	tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0320);
+	tegra_sor_writel(sor, 4096, SOR_AUDIO_NVAL_0320);
+
+	tegra_sor_writel(sor, 20000, SOR_AUDIO_AVAL_0441);
+	tegra_sor_writel(sor, 4704, SOR_AUDIO_NVAL_0441);
+
+	tegra_sor_writel(sor, 20000, SOR_AUDIO_AVAL_0882);
+	tegra_sor_writel(sor, 9408, SOR_AUDIO_NVAL_0882);
+
+	tegra_sor_writel(sor, 20000, SOR_AUDIO_AVAL_1764);
+	tegra_sor_writel(sor, 18816, SOR_AUDIO_NVAL_1764);
+
+	value = (24000 * 6144) / (128 * sor->format.sample_rate / 1000);
+	tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0480);
+	tegra_sor_writel(sor, 6144, SOR_AUDIO_NVAL_0480);
+
+	value = (24000 * 12288) / (128 * sor->format.sample_rate / 1000);
+	tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_0960);
+	tegra_sor_writel(sor, 12288, SOR_AUDIO_NVAL_0960);
+
+	value = (24000 * 24576) / (128 * sor->format.sample_rate / 1000);
+	tegra_sor_writel(sor, value, SOR_AUDIO_AVAL_1920);
+	tegra_sor_writel(sor, 24576, SOR_AUDIO_NVAL_1920);
+
+	value = tegra_sor_readl(sor, SOR_HDMI_AUDIO_N);
+	value &= ~SOR_HDMI_AUDIO_N_RESET;
+	tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_N);
+
+	tegra_sor_hdmi_enable_audio_infoframe(sor);
+}
+
+static void tegra_sor_hdmi_disable_audio_infoframe(struct tegra_sor *sor)
+{
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_HDMI_AUDIO_INFOFRAME_CTRL);
+	value &= ~INFOFRAME_CTRL_ENABLE;
+	tegra_sor_writel(sor, value, SOR_HDMI_AUDIO_INFOFRAME_CTRL);
+}
+
+static void tegra_sor_hdmi_audio_disable(struct tegra_sor *sor)
+{
+	tegra_sor_hdmi_disable_audio_infoframe(sor);
+}
+
+static struct tegra_sor_hdmi_settings *
+tegra_sor_hdmi_find_settings(struct tegra_sor *sor, unsigned long frequency)
+{
+	unsigned int i;
+
+	for (i = 0; i < sor->num_settings; i++)
+		if (frequency <= sor->settings[i].frequency)
+			return &sor->settings[i];
+
+	return NULL;
+}
+
+static void tegra_sor_hdmi_disable_scrambling(struct tegra_sor *sor)
+{
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_HDMI2_CTRL);
+	value &= ~SOR_HDMI2_CTRL_CLOCK_MODE_DIV_BY_4;
+	value &= ~SOR_HDMI2_CTRL_SCRAMBLE;
+	tegra_sor_writel(sor, value, SOR_HDMI2_CTRL);
+}
+
+static void tegra_sor_hdmi_scdc_disable(struct tegra_sor *sor)
+{
+	drm_scdc_set_high_tmds_clock_ratio(&sor->output.connector, false);
+	drm_scdc_set_scrambling(&sor->output.connector, false);
+
+	tegra_sor_hdmi_disable_scrambling(sor);
+}
+
+static void tegra_sor_hdmi_scdc_stop(struct tegra_sor *sor)
+{
+	if (sor->scdc_enabled) {
+		cancel_delayed_work_sync(&sor->scdc);
+		tegra_sor_hdmi_scdc_disable(sor);
+	}
+}
+
+static void tegra_sor_hdmi_enable_scrambling(struct tegra_sor *sor)
+{
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_HDMI2_CTRL);
+	value |= SOR_HDMI2_CTRL_CLOCK_MODE_DIV_BY_4;
+	value |= SOR_HDMI2_CTRL_SCRAMBLE;
+	tegra_sor_writel(sor, value, SOR_HDMI2_CTRL);
+}
+
+static void tegra_sor_hdmi_scdc_enable(struct tegra_sor *sor)
+{
+	drm_scdc_set_high_tmds_clock_ratio(&sor->output.connector, true);
+	drm_scdc_set_scrambling(&sor->output.connector, true);
+
+	tegra_sor_hdmi_enable_scrambling(sor);
+}
+
+static void tegra_sor_hdmi_scdc_work(struct work_struct *work)
+{
+	struct tegra_sor *sor = container_of(work, struct tegra_sor, scdc.work);
+
+	if (!drm_scdc_get_scrambling_status(&sor->output.connector)) {
+		DRM_DEBUG_KMS("SCDC not scrambled\n");
+		tegra_sor_hdmi_scdc_enable(sor);
+	}
+
+	schedule_delayed_work(&sor->scdc, msecs_to_jiffies(5000));
+}
+
+static void tegra_sor_hdmi_scdc_start(struct tegra_sor *sor)
+{
+	struct drm_scdc *scdc = &sor->output.connector.display_info.hdmi.scdc;
+	struct drm_display_mode *mode;
+
+	mode = &sor->output.encoder.crtc->state->adjusted_mode;
+
+	if (mode->clock >= 340000 && scdc->supported) {
+		schedule_delayed_work(&sor->scdc, msecs_to_jiffies(5000));
+		tegra_sor_hdmi_scdc_enable(sor);
+		sor->scdc_enabled = true;
+	}
+}
+
+static void tegra_sor_hdmi_disable(struct drm_encoder *encoder)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_sor *sor = to_sor(output);
+	u32 value;
+	int err;
+
+	tegra_sor_audio_unprepare(sor);
+	tegra_sor_hdmi_scdc_stop(sor);
+
+	err = tegra_sor_detach(sor);
+	if (err < 0)
+		dev_err(sor->dev, "failed to detach SOR: %d\n", err);
+
+	tegra_sor_writel(sor, 0, SOR_STATE1);
+	tegra_sor_update(sor);
+
+	/* disable display to SOR clock */
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+
+	if (!sor->soc->has_nvdisplay)
+		value &= ~SOR1_TIMING_CYA;
+
+	value &= ~SOR_ENABLE(sor->index);
+
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+	tegra_dc_commit(dc);
+
+	err = tegra_sor_power_down(sor);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power down SOR: %d\n", err);
+
+	err = tegra_pmc_io_pad_power_disable(sor->pmc, sor->pad);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power off I/O pad: %d\n", err);
+
+	host1x_client_suspend(&sor->client);
+}
+
+static void tegra_sor_hdmi_enable(struct drm_encoder *encoder)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	unsigned int h_ref_to_sync = 1, pulse_start, max_ac;
+	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_sor_hdmi_settings *settings;
+	struct tegra_sor *sor = to_sor(output);
+	struct tegra_sor_state *state;
+	struct drm_display_mode *mode;
+	unsigned long rate, pclk;
+	unsigned int div, i;
+	u32 value;
+	int err;
+
+	state = to_sor_state(output->connector.state);
+	mode = &encoder->crtc->state->adjusted_mode;
+	pclk = mode->clock * 1000;
+
+	err = host1x_client_resume(&sor->client);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to resume: %d\n", err);
+		return;
+	}
+
+	/* switch to safe parent clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_safe);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to set safe parent clock: %d\n", err);
+		return;
+	}
+
+	div = clk_get_rate(sor->clk) / 1000000 * 4;
+
+	err = tegra_pmc_io_pad_power_enable(sor->pmc, sor->pad);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power on I/O pad: %d\n", err);
+
+	usleep_range(20, 100);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll2);
+	value &= ~SOR_PLL2_BANDGAP_POWERDOWN;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll2);
+
+	usleep_range(20, 100);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll3);
+	value &= ~SOR_PLL3_PLL_VDD_MODE_3V3;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll3);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll0);
+	value &= ~SOR_PLL0_VCOPD;
+	value &= ~SOR_PLL0_PWR;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll0);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll2);
+	value &= ~SOR_PLL2_SEQ_PLLCAPPD_ENFORCE;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll2);
+
+	usleep_range(200, 400);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll2);
+	value &= ~SOR_PLL2_POWERDOWN_OVERRIDE;
+	value &= ~SOR_PLL2_PORT_POWERDOWN;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll2);
+
+	usleep_range(20, 100);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value |= SOR_DP_PADCTL_PD_TXD_3 | SOR_DP_PADCTL_PD_TXD_0 |
+		 SOR_DP_PADCTL_PD_TXD_1 | SOR_DP_PADCTL_PD_TXD_2;
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	while (true) {
+		value = tegra_sor_readl(sor, SOR_LANE_SEQ_CTL);
+		if ((value & SOR_LANE_SEQ_CTL_STATE_BUSY) == 0)
+			break;
+
+		usleep_range(250, 1000);
+	}
+
+	value = SOR_LANE_SEQ_CTL_TRIGGER | SOR_LANE_SEQ_CTL_SEQUENCE_DOWN |
+		SOR_LANE_SEQ_CTL_POWER_STATE_UP | SOR_LANE_SEQ_CTL_DELAY(5);
+	tegra_sor_writel(sor, value, SOR_LANE_SEQ_CTL);
+
+	while (true) {
+		value = tegra_sor_readl(sor, SOR_LANE_SEQ_CTL);
+		if ((value & SOR_LANE_SEQ_CTL_TRIGGER) == 0)
+			break;
+
+		usleep_range(250, 1000);
+	}
+
+	value = tegra_sor_readl(sor, SOR_CLK_CNTRL);
+	value &= ~SOR_CLK_CNTRL_DP_LINK_SPEED_MASK;
+	value &= ~SOR_CLK_CNTRL_DP_CLK_SEL_MASK;
+
+	if (mode->clock < 340000) {
+		DRM_DEBUG_KMS("setting 2.7 GHz link speed\n");
+		value |= SOR_CLK_CNTRL_DP_LINK_SPEED_G2_70;
+	} else {
+		DRM_DEBUG_KMS("setting 5.4 GHz link speed\n");
+		value |= SOR_CLK_CNTRL_DP_LINK_SPEED_G5_40;
+	}
+
+	value |= SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_PCLK;
+	tegra_sor_writel(sor, value, SOR_CLK_CNTRL);
+
+	/* SOR pad PLL stabilization time */
+	usleep_range(250, 1000);
+
+	value = tegra_sor_readl(sor, SOR_DP_LINKCTL0);
+	value &= ~SOR_DP_LINKCTL_LANE_COUNT_MASK;
+	value |= SOR_DP_LINKCTL_LANE_COUNT(4);
+	tegra_sor_writel(sor, value, SOR_DP_LINKCTL0);
+
+	value = tegra_sor_readl(sor, SOR_DP_SPARE0);
+	value &= ~SOR_DP_SPARE_DISP_VIDEO_PREAMBLE;
+	value &= ~SOR_DP_SPARE_PANEL_INTERNAL;
+	value &= ~SOR_DP_SPARE_SEQ_ENABLE;
+	value &= ~SOR_DP_SPARE_MACRO_SOR_CLK;
+	tegra_sor_writel(sor, value, SOR_DP_SPARE0);
+
+	value = SOR_SEQ_CTL_PU_PC(0) | SOR_SEQ_CTL_PU_PC_ALT(0) |
+		SOR_SEQ_CTL_PD_PC(8) | SOR_SEQ_CTL_PD_PC_ALT(8);
+	tegra_sor_writel(sor, value, SOR_SEQ_CTL);
+
+	value = SOR_SEQ_INST_DRIVE_PWM_OUT_LO | SOR_SEQ_INST_HALT |
+		SOR_SEQ_INST_WAIT_VSYNC | SOR_SEQ_INST_WAIT(1);
+	tegra_sor_writel(sor, value, SOR_SEQ_INST(0));
+	tegra_sor_writel(sor, value, SOR_SEQ_INST(8));
+
+	if (!sor->soc->has_nvdisplay) {
+		/* program the reference clock */
+		value = SOR_REFCLK_DIV_INT(div) | SOR_REFCLK_DIV_FRAC(div);
+		tegra_sor_writel(sor, value, SOR_REFCLK);
+	}
+
+	/* XXX not in TRM */
+	for (value = 0, i = 0; i < 5; i++)
+		value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->xbar_cfg[i]) |
+			 SOR_XBAR_CTRL_LINK1_XSEL(i, i);
+
+	tegra_sor_writel(sor, 0x00000000, SOR_XBAR_POL);
+	tegra_sor_writel(sor, value, SOR_XBAR_CTRL);
+
+	/*
+	 * Switch the pad clock to the DP clock. Note that we cannot actually
+	 * do this because Tegra186 and later don't support clk_set_parent()
+	 * on the sorX_pad_clkout clocks. We already do the equivalent above
+	 * using the DP_CLK_SEL mux of the SOR_CLK_CNTRL register.
+	 */
+#if 0
+	err = clk_set_parent(sor->clk_pad, sor->clk_dp);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to select pad parent clock: %d\n",
+			err);
+		return;
+	}
+#endif
+
+	/* switch the SOR clock to the pad clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_pad);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to select SOR parent clock: %d\n",
+			err);
+		return;
+	}
+
+	/* switch the output clock to the parent pixel clock */
+	err = clk_set_parent(sor->clk, sor->clk_parent);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to select output parent clock: %d\n",
+			err);
+		return;
+	}
+
+	/* adjust clock rate for HDMI 2.0 modes */
+	rate = clk_get_rate(sor->clk_parent);
+
+	if (mode->clock >= 340000)
+		rate /= 2;
+
+	DRM_DEBUG_KMS("setting clock to %lu Hz, mode: %lu Hz\n", rate, pclk);
+
+	clk_set_rate(sor->clk, rate);
+
+	if (!sor->soc->has_nvdisplay) {
+		value = SOR_INPUT_CONTROL_HDMI_SRC_SELECT(dc->pipe);
+
+		/* XXX is this the proper check? */
+		if (mode->clock < 75000)
+			value |= SOR_INPUT_CONTROL_ARM_VIDEO_RANGE_LIMITED;
+
+		tegra_sor_writel(sor, value, SOR_INPUT_CONTROL);
+	}
+
+	max_ac = ((mode->htotal - mode->hdisplay) - SOR_REKEY - 18) / 32;
+
+	value = SOR_HDMI_CTRL_ENABLE | SOR_HDMI_CTRL_MAX_AC_PACKET(max_ac) |
+		SOR_HDMI_CTRL_AUDIO_LAYOUT | SOR_HDMI_CTRL_REKEY(SOR_REKEY);
+	tegra_sor_writel(sor, value, SOR_HDMI_CTRL);
+
+	if (!dc->soc->has_nvdisplay) {
+		/* H_PULSE2 setup */
+		pulse_start = h_ref_to_sync +
+			      (mode->hsync_end - mode->hsync_start) +
+			      (mode->htotal - mode->hsync_end) - 10;
+
+		value = PULSE_LAST_END_A | PULSE_QUAL_VACTIVE |
+			PULSE_POLARITY_HIGH | PULSE_MODE_NORMAL;
+		tegra_dc_writel(dc, value, DC_DISP_H_PULSE2_CONTROL);
+
+		value = PULSE_END(pulse_start + 8) | PULSE_START(pulse_start);
+		tegra_dc_writel(dc, value, DC_DISP_H_PULSE2_POSITION_A);
+
+		value = tegra_dc_readl(dc, DC_DISP_DISP_SIGNAL_OPTIONS0);
+		value |= H_PULSE2_ENABLE;
+		tegra_dc_writel(dc, value, DC_DISP_DISP_SIGNAL_OPTIONS0);
+	}
+
+	/* infoframe setup */
+	err = tegra_sor_hdmi_setup_avi_infoframe(sor, mode);
+	if (err < 0)
+		dev_err(sor->dev, "failed to setup AVI infoframe: %d\n", err);
+
+	/* XXX HDMI audio support not implemented yet */
+	tegra_sor_hdmi_disable_audio_infoframe(sor);
+
+	/* use single TMDS protocol */
+	value = tegra_sor_readl(sor, SOR_STATE1);
+	value &= ~SOR_STATE_ASY_PROTOCOL_MASK;
+	value |= SOR_STATE_ASY_PROTOCOL_SINGLE_TMDS_A;
+	tegra_sor_writel(sor, value, SOR_STATE1);
+
+	/* power up pad calibration */
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value &= ~SOR_DP_PADCTL_PAD_CAL_PD;
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	/* production settings */
+	settings = tegra_sor_hdmi_find_settings(sor, mode->clock * 1000);
+	if (!settings) {
+		dev_err(sor->dev, "no settings for pixel clock %d Hz\n",
+			mode->clock * 1000);
+		return;
+	}
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll0);
+	value &= ~SOR_PLL0_ICHPMP_MASK;
+	value &= ~SOR_PLL0_FILTER_MASK;
+	value &= ~SOR_PLL0_VCOCAP_MASK;
+	value |= SOR_PLL0_ICHPMP(settings->ichpmp);
+	value |= SOR_PLL0_FILTER(settings->filter);
+	value |= SOR_PLL0_VCOCAP(settings->vcocap);
+	tegra_sor_writel(sor, value, sor->soc->regs->pll0);
+
+	/* XXX not in TRM */
+	value = tegra_sor_readl(sor, sor->soc->regs->pll1);
+	value &= ~SOR_PLL1_LOADADJ_MASK;
+	value &= ~SOR_PLL1_TMDS_TERMADJ_MASK;
+	value |= SOR_PLL1_LOADADJ(settings->loadadj);
+	value |= SOR_PLL1_TMDS_TERMADJ(settings->tmds_termadj);
+	value |= SOR_PLL1_TMDS_TERM;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll1);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll3);
+	value &= ~SOR_PLL3_BG_TEMP_COEF_MASK;
+	value &= ~SOR_PLL3_BG_VREF_LEVEL_MASK;
+	value &= ~SOR_PLL3_AVDD10_LEVEL_MASK;
+	value &= ~SOR_PLL3_AVDD14_LEVEL_MASK;
+	value |= SOR_PLL3_BG_TEMP_COEF(settings->bg_temp_coef);
+	value |= SOR_PLL3_BG_VREF_LEVEL(settings->bg_vref_level);
+	value |= SOR_PLL3_AVDD10_LEVEL(settings->avdd10_level);
+	value |= SOR_PLL3_AVDD14_LEVEL(settings->avdd14_level);
+	tegra_sor_writel(sor, value, sor->soc->regs->pll3);
+
+	value = settings->drive_current[3] << 24 |
+		settings->drive_current[2] << 16 |
+		settings->drive_current[1] <<  8 |
+		settings->drive_current[0] <<  0;
+	tegra_sor_writel(sor, value, SOR_LANE_DRIVE_CURRENT0);
+
+	value = settings->preemphasis[3] << 24 |
+		settings->preemphasis[2] << 16 |
+		settings->preemphasis[1] <<  8 |
+		settings->preemphasis[0] <<  0;
+	tegra_sor_writel(sor, value, SOR_LANE_PREEMPHASIS0);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value &= ~SOR_DP_PADCTL_TX_PU_MASK;
+	value |= SOR_DP_PADCTL_TX_PU_ENABLE;
+	value |= SOR_DP_PADCTL_TX_PU(settings->tx_pu_value);
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl2);
+	value &= ~SOR_DP_PADCTL_SPAREPLL_MASK;
+	value |= SOR_DP_PADCTL_SPAREPLL(settings->sparepll);
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl2);
+
+	/* power down pad calibration */
+	value = tegra_sor_readl(sor, sor->soc->regs->dp_padctl0);
+	value |= SOR_DP_PADCTL_PAD_CAL_PD;
+	tegra_sor_writel(sor, value, sor->soc->regs->dp_padctl0);
+
+	if (!dc->soc->has_nvdisplay) {
+		/* miscellaneous display controller settings */
+		value = VSYNC_H_POSITION(1);
+		tegra_dc_writel(dc, value, DC_DISP_DISP_TIMING_OPTIONS);
+	}
+
+	value = tegra_dc_readl(dc, DC_DISP_DISP_COLOR_CONTROL);
+	value &= ~DITHER_CONTROL_MASK;
+	value &= ~BASE_COLOR_SIZE_MASK;
+	if (dc->soc->has_nvdisplay) {
+		tegra_dc_writel(dc, lower_32_bits(dc->cmu_output_lut_phys),
+				DC_DISP_COREPVT_HEAD_SET_OUTPUT_LUT_BASE);
+		tegra_dc_writel(dc, upper_32_bits(dc->cmu_output_lut_phys),
+				DC_DISP_COREPVT_HEAD_SET_OUTPUT_LUT_BASE_HI);
+
+		tegra_dc_writel(dc, OUTPUT_LUT_MODE_INTERPOLATE | OUTPUT_LUT_SIZE_SIZE_1025,
+				DC_DISP_CORE_HEAD_SET_CONTROL_OUTPUT_LUT);
+
+		value |= CMU_ENABLE_ENABLE;
+	}
+
+	switch (state->bpc) {
+	case 6:
+		value |= BASE_COLOR_SIZE_666;
+		break;
+
+	case 8:
+		value |= BASE_COLOR_SIZE_888;
+		break;
+
+	case 10:
+		value |= BASE_COLOR_SIZE_101010;
+		break;
+
+	case 12:
+		value |= BASE_COLOR_SIZE_121212;
+		break;
+
+	default:
+		WARN(1, "%u bits-per-color not supported\n", state->bpc);
+		value |= BASE_COLOR_SIZE_888;
+		break;
+	}
+
+	tegra_dc_writel(dc, value, DC_DISP_DISP_COLOR_CONTROL);
+
+	/* XXX set display head owner */
+	value = tegra_sor_readl(sor, SOR_STATE1);
+	value &= ~SOR_STATE_ASY_OWNER_MASK;
+	value |= SOR_STATE_ASY_OWNER(1 + dc->pipe);
+	tegra_sor_writel(sor, value, SOR_STATE1);
+
+	err = tegra_sor_power_up(sor, 250);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power up SOR: %d\n", err);
+
+	/* configure dynamic range of output */
+	value = tegra_sor_readl(sor, sor->soc->regs->head_state0 + dc->pipe);
+	value &= ~SOR_HEAD_STATE_RANGECOMPRESS_MASK;
+	value &= ~SOR_HEAD_STATE_DYNRANGE_MASK;
+	tegra_sor_writel(sor, value, sor->soc->regs->head_state0 + dc->pipe);
+
+	/* configure colorspace */
+	value = tegra_sor_readl(sor, sor->soc->regs->head_state0 + dc->pipe);
+	value &= ~SOR_HEAD_STATE_COLORSPACE_MASK;
+	value |= SOR_HEAD_STATE_COLORSPACE_RGB;
+	tegra_sor_writel(sor, value, sor->soc->regs->head_state0 + dc->pipe);
+
+	tegra_sor_mode_set(sor, mode, state);
+
+	tegra_sor_update(sor);
+
+	/* program preamble timing in SOR (XXX) */
+	value = tegra_sor_readl(sor, SOR_DP_SPARE0);
+	value &= ~SOR_DP_SPARE_DISP_VIDEO_PREAMBLE;
+	tegra_sor_writel(sor, value, SOR_DP_SPARE0);
+
+	err = tegra_sor_attach(sor);
+	if (err < 0)
+		dev_err(sor->dev, "failed to attach SOR: %d\n", err);
+
+	/* enable display to SOR clock and generate HDMI preamble */
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+
+	if (!sor->soc->has_nvdisplay)
+		value |= SOR1_TIMING_CYA;
+
+	value |= SOR_ENABLE(sor->index);
+
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+	if (dc->soc->has_nvdisplay) {
+		value = tegra_dc_readl(dc, DC_DISP_CORE_SOR_SET_CONTROL(sor->index));
+		value &= ~PROTOCOL_MASK;
+		value |= PROTOCOL_SINGLE_TMDS_A;
+		tegra_dc_writel(dc, value, DC_DISP_CORE_SOR_SET_CONTROL(sor->index));
+	}
+
+	tegra_dc_commit(dc);
+
+	err = tegra_sor_wakeup(sor);
+	if (err < 0)
+		dev_err(sor->dev, "failed to wakeup SOR: %d\n", err);
+
+	tegra_sor_hdmi_scdc_start(sor);
+	tegra_sor_audio_prepare(sor);
+}
+
+static const struct drm_encoder_helper_funcs tegra_sor_hdmi_helpers = {
+	.disable = tegra_sor_hdmi_disable,
+	.enable = tegra_sor_hdmi_enable,
+	.atomic_check = tegra_sor_encoder_atomic_check,
+};
+
+static void tegra_sor_dp_disable(struct drm_encoder *encoder)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_sor *sor = to_sor(output);
+	u32 value;
+	int err;
+
+	if (output->panel)
+		drm_panel_disable(output->panel);
+
+	/*
+	 * Do not attempt to power down a DP link if we're not connected since
+	 * the AUX transactions would just be timing out.
+	 */
+	if (output->connector.status != connector_status_disconnected) {
+		err = drm_dp_link_power_down(sor->aux, sor->link.revision);
+		if (err < 0)
+			dev_err(sor->dev, "failed to power down link: %d\n",
+				err);
+	}
+
+	err = tegra_sor_detach(sor);
+	if (err < 0)
+		dev_err(sor->dev, "failed to detach SOR: %d\n", err);
+
+	tegra_sor_writel(sor, 0, SOR_STATE1);
+	tegra_sor_update(sor);
+
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+	value &= ~SOR_ENABLE(sor->index);
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+	tegra_dc_commit(dc);
+
+	value = tegra_sor_readl(sor, SOR_STATE1);
+	value &= ~SOR_STATE_ASY_PROTOCOL_MASK;
+	value &= ~SOR_STATE_ASY_SUBOWNER_MASK;
+	value &= ~SOR_STATE_ASY_OWNER_MASK;
+	tegra_sor_writel(sor, value, SOR_STATE1);
+	tegra_sor_update(sor);
+
+	/* switch to safe parent clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_safe);
+	if (err < 0)
+		dev_err(sor->dev, "failed to set safe clock: %d\n", err);
+
+	err = tegra_sor_power_down(sor);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power down SOR: %d\n", err);
+
+	err = tegra_pmc_io_pad_power_disable(sor->pmc, sor->pad);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power off I/O pad: %d\n", err);
+
+	err = drm_dp_aux_disable(sor->aux);
+	if (err < 0)
+		dev_err(sor->dev, "failed disable DPAUX: %d\n", err);
+
+	if (output->panel)
+		drm_panel_unprepare(output->panel);
+
+	host1x_client_suspend(&sor->client);
+}
+
+static void tegra_sor_dp_enable(struct drm_encoder *encoder)
+{
+	struct tegra_output *output = encoder_to_output(encoder);
+	struct tegra_dc *dc = to_tegra_dc(encoder->crtc);
+	struct tegra_sor *sor = to_sor(output);
+	struct tegra_sor_config config;
+	struct tegra_sor_state *state;
+	struct drm_display_mode *mode;
+	struct drm_display_info *info;
+	unsigned int i;
+	u32 value;
+	int err;
+
+	state = to_sor_state(output->connector.state);
+	mode = &encoder->crtc->state->adjusted_mode;
+	info = &output->connector.display_info;
+
+	err = host1x_client_resume(&sor->client);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to resume: %d\n", err);
+		return;
+	}
+
+	/* switch to safe parent clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_safe);
+	if (err < 0)
+		dev_err(sor->dev, "failed to set safe parent clock: %d\n", err);
+
+	err = tegra_pmc_io_pad_power_enable(sor->pmc, sor->pad);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power on LVDS rail: %d\n", err);
+
+	usleep_range(20, 100);
+
+	err = drm_dp_aux_enable(sor->aux);
+	if (err < 0)
+		dev_err(sor->dev, "failed to enable DPAUX: %d\n", err);
+
+	err = drm_dp_link_probe(sor->aux, &sor->link);
+	if (err < 0)
+		dev_err(sor->dev, "failed to probe DP link: %d\n", err);
+
+	tegra_sor_filter_rates(sor);
+
+	err = drm_dp_link_choose(&sor->link, mode, info);
+	if (err < 0)
+		dev_err(sor->dev, "failed to choose link: %d\n", err);
+
+	if (output->panel)
+		drm_panel_prepare(output->panel);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll2);
+	value &= ~SOR_PLL2_BANDGAP_POWERDOWN;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll2);
+
+	usleep_range(20, 40);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll3);
+	value |= SOR_PLL3_PLL_VDD_MODE_3V3;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll3);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll0);
+	value &= ~(SOR_PLL0_VCOPD | SOR_PLL0_PWR);
+	tegra_sor_writel(sor, value, sor->soc->regs->pll0);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll2);
+	value &= ~SOR_PLL2_SEQ_PLLCAPPD_ENFORCE;
+	value |= SOR_PLL2_SEQ_PLLCAPPD;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll2);
+
+	usleep_range(200, 400);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll2);
+	value &= ~SOR_PLL2_POWERDOWN_OVERRIDE;
+	value &= ~SOR_PLL2_PORT_POWERDOWN;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll2);
+
+	value = tegra_sor_readl(sor, SOR_CLK_CNTRL);
+	value &= ~SOR_CLK_CNTRL_DP_CLK_SEL_MASK;
+
+	if (output->panel)
+		value |= SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_DPCLK;
+	else
+		value |= SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_DPCLK;
+
+	tegra_sor_writel(sor, value, SOR_CLK_CNTRL);
+
+	usleep_range(200, 400);
+
+	value = tegra_sor_readl(sor, SOR_DP_SPARE0);
+	/* XXX not in TRM */
+	if (output->panel)
+		value |= SOR_DP_SPARE_PANEL_INTERNAL;
+	else
+		value &= ~SOR_DP_SPARE_PANEL_INTERNAL;
+
+	value |= SOR_DP_SPARE_SEQ_ENABLE;
+	tegra_sor_writel(sor, value, SOR_DP_SPARE0);
+
+	/* XXX not in TRM */
+	tegra_sor_writel(sor, 0, SOR_LVDS);
+
+	value = tegra_sor_readl(sor, sor->soc->regs->pll0);
+	value &= ~SOR_PLL0_ICHPMP_MASK;
+	value &= ~SOR_PLL0_VCOCAP_MASK;
+	value |= SOR_PLL0_ICHPMP(0x1);
+	value |= SOR_PLL0_VCOCAP(0x3);
+	value |= SOR_PLL0_RESISTOR_EXT;
+	tegra_sor_writel(sor, value, sor->soc->regs->pll0);
+
+	/* XXX not in TRM */
+	for (value = 0, i = 0; i < 5; i++)
+		value |= SOR_XBAR_CTRL_LINK0_XSEL(i, sor->soc->xbar_cfg[i]) |
+			 SOR_XBAR_CTRL_LINK1_XSEL(i, i);
+
+	tegra_sor_writel(sor, 0x00000000, SOR_XBAR_POL);
+	tegra_sor_writel(sor, value, SOR_XBAR_CTRL);
+
+	/*
+	 * Switch the pad clock to the DP clock. Note that we cannot actually
+	 * do this because Tegra186 and later don't support clk_set_parent()
+	 * on the sorX_pad_clkout clocks. We already do the equivalent above
+	 * using the DP_CLK_SEL mux of the SOR_CLK_CNTRL register.
+	 */
+#if 0
+	err = clk_set_parent(sor->clk_pad, sor->clk_parent);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to select pad parent clock: %d\n",
+			err);
+		return;
+	}
+#endif
+
+	/* switch the SOR clock to the pad clock */
+	err = tegra_sor_set_parent_clock(sor, sor->clk_pad);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to select SOR parent clock: %d\n",
+			err);
+		return;
+	}
+
+	/* switch the output clock to the parent pixel clock */
+	err = clk_set_parent(sor->clk, sor->clk_parent);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to select output parent clock: %d\n",
+			err);
+		return;
+	}
+
+	/* use DP-A protocol */
+	value = tegra_sor_readl(sor, SOR_STATE1);
+	value &= ~SOR_STATE_ASY_PROTOCOL_MASK;
+	value |= SOR_STATE_ASY_PROTOCOL_DP_A;
+	tegra_sor_writel(sor, value, SOR_STATE1);
+
+	/* enable port */
+	value = tegra_sor_readl(sor, SOR_DP_LINKCTL0);
+	value |= SOR_DP_LINKCTL_ENABLE;
+	tegra_sor_writel(sor, value, SOR_DP_LINKCTL0);
+
+	tegra_sor_dp_term_calibrate(sor);
+
+	err = drm_dp_link_train(&sor->link);
+	if (err < 0)
+		dev_err(sor->dev, "link training failed: %d\n", err);
+	else
+		dev_dbg(sor->dev, "link training succeeded\n");
+
+	err = drm_dp_link_power_up(sor->aux, sor->link.revision);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power up DP link: %d\n", err);
+
+	/* compute configuration */
+	memset(&config, 0, sizeof(config));
+	config.bits_per_pixel = state->bpc * 3;
+
+	err = tegra_sor_compute_config(sor, mode, &config, &sor->link);
+	if (err < 0)
+		dev_err(sor->dev, "failed to compute configuration: %d\n", err);
+
+	tegra_sor_apply_config(sor, &config);
+	tegra_sor_mode_set(sor, mode, state);
+
+	if (output->panel) {
+		/* CSTM (LVDS, link A/B, upper) */
+		value = SOR_CSTM_LVDS | SOR_CSTM_LINK_ACT_A | SOR_CSTM_LINK_ACT_B |
+			SOR_CSTM_UPPER;
+		tegra_sor_writel(sor, value, SOR_CSTM);
+
+		/* PWM setup */
+		err = tegra_sor_setup_pwm(sor, 250);
+		if (err < 0)
+			dev_err(sor->dev, "failed to setup PWM: %d\n", err);
+	}
+
+	tegra_sor_update(sor);
+
+	err = tegra_sor_power_up(sor, 250);
+	if (err < 0)
+		dev_err(sor->dev, "failed to power up SOR: %d\n", err);
+
+	/* attach and wake up */
+	err = tegra_sor_attach(sor);
+	if (err < 0)
+		dev_err(sor->dev, "failed to attach SOR: %d\n", err);
+
+	if (dc->soc->has_nvdisplay) {
+		value = tegra_dc_readl(dc, DC_DISP_DISP_COLOR_CONTROL);
+		tegra_dc_writel(dc, lower_32_bits(dc->cmu_output_lut_phys),
+				DC_DISP_COREPVT_HEAD_SET_OUTPUT_LUT_BASE);
+		tegra_dc_writel(dc, upper_32_bits(dc->cmu_output_lut_phys),
+				DC_DISP_COREPVT_HEAD_SET_OUTPUT_LUT_BASE_HI);
+
+		tegra_dc_writel(dc, OUTPUT_LUT_MODE_INTERPOLATE | OUTPUT_LUT_SIZE_SIZE_1025,
+				DC_DISP_CORE_HEAD_SET_CONTROL_OUTPUT_LUT);
+
+		value |= CMU_ENABLE_ENABLE;
+		tegra_dc_writel(dc, value, DC_DISP_DISP_COLOR_CONTROL);
+	}
+
+	value = tegra_dc_readl(dc, DC_DISP_DISP_WIN_OPTIONS);
+	value |= SOR_ENABLE(sor->index);
+	tegra_dc_writel(dc, value, DC_DISP_DISP_WIN_OPTIONS);
+
+	tegra_dc_commit(dc);
+
+	err = tegra_sor_wakeup(sor);
+	if (err < 0)
+		dev_err(sor->dev, "failed to wakeup SOR: %d\n", err);
+
+	if (output->panel)
+		drm_panel_enable(output->panel);
+}
+
+static const struct drm_encoder_helper_funcs tegra_sor_dp_helpers = {
+	.disable = tegra_sor_dp_disable,
+	.enable = tegra_sor_dp_enable,
+	.atomic_check = tegra_sor_encoder_atomic_check,
+};
+
+static void tegra_sor_disable_regulator(void *data)
+{
+	struct regulator *reg = data;
+
+	regulator_disable(reg);
+}
+
+static int tegra_sor_enable_regulator(struct tegra_sor *sor, struct regulator *reg)
+{
+	int err;
+
+	err = regulator_enable(reg);
+	if (err)
+		return err;
+
+	return devm_add_action_or_reset(sor->dev, tegra_sor_disable_regulator, reg);
+}
+
+static int tegra_sor_hdmi_probe(struct tegra_sor *sor)
+{
+	int err;
+
+	sor->avdd_io_supply = devm_regulator_get(sor->dev, "avdd-io-hdmi-dp");
+	if (IS_ERR(sor->avdd_io_supply))
+		return dev_err_probe(sor->dev, PTR_ERR(sor->avdd_io_supply),
+				     "cannot get AVDD I/O supply\n");
+
+	err = tegra_sor_enable_regulator(sor, sor->avdd_io_supply);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to enable AVDD I/O supply: %d\n",
+			err);
+		return err;
+	}
+
+	sor->vdd_pll_supply = devm_regulator_get(sor->dev, "vdd-hdmi-dp-pll");
+	if (IS_ERR(sor->vdd_pll_supply))
+		return dev_err_probe(sor->dev, PTR_ERR(sor->vdd_pll_supply),
+				     "cannot get VDD PLL supply\n");
+
+	err = tegra_sor_enable_regulator(sor, sor->vdd_pll_supply);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to enable VDD PLL supply: %d\n",
+			err);
+		return err;
+	}
+
+	sor->hdmi_supply = devm_regulator_get(sor->dev, "hdmi");
+	if (IS_ERR(sor->hdmi_supply))
+		return dev_err_probe(sor->dev, PTR_ERR(sor->hdmi_supply),
+				     "cannot get HDMI supply\n");
+
+	err = tegra_sor_enable_regulator(sor, sor->hdmi_supply);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to enable HDMI supply: %d\n", err);
+		return err;
+	}
+
+	INIT_DELAYED_WORK(&sor->scdc, tegra_sor_hdmi_scdc_work);
+
+	return 0;
+}
+
+static const struct tegra_sor_ops tegra_sor_hdmi_ops = {
+	.name = "HDMI",
+	.probe = tegra_sor_hdmi_probe,
+	.audio_enable = tegra_sor_hdmi_audio_enable,
+	.audio_disable = tegra_sor_hdmi_audio_disable,
+};
+
+static int tegra_sor_dp_probe(struct tegra_sor *sor)
+{
+	int err;
+
+	sor->avdd_io_supply = devm_regulator_get(sor->dev, "avdd-io-hdmi-dp");
+	if (IS_ERR(sor->avdd_io_supply))
+		return PTR_ERR(sor->avdd_io_supply);
+
+	err = tegra_sor_enable_regulator(sor, sor->avdd_io_supply);
+	if (err < 0)
+		return err;
+
+	sor->vdd_pll_supply = devm_regulator_get(sor->dev, "vdd-hdmi-dp-pll");
+	if (IS_ERR(sor->vdd_pll_supply))
+		return PTR_ERR(sor->vdd_pll_supply);
+
+	err = tegra_sor_enable_regulator(sor, sor->vdd_pll_supply);
+	if (err < 0)
+		return err;
+
+	return 0;
+}
+
+static const struct tegra_sor_ops tegra_sor_dp_ops = {
+	.name = "DP",
+	.probe = tegra_sor_dp_probe,
+};
+
+static const struct drm_encoder_funcs tegra_sor_encoder_funcs_cleanup = {
+	.destroy = drm_encoder_cleanup,
+};
+
+static int tegra_sor_init(struct host1x_client *client)
+{
+	struct drm_device *drm = dev_get_drvdata(client->host);
+	const struct drm_encoder_helper_funcs *helpers = NULL;
+	struct tegra_sor *sor = host1x_client_to_sor(client);
+	int connector = DRM_MODE_CONNECTOR_Unknown;
+	int encoder = DRM_MODE_ENCODER_NONE;
+	int err;
+
+	if (!sor->aux) {
+		if (sor->ops == &tegra_sor_hdmi_ops) {
+			connector = DRM_MODE_CONNECTOR_HDMIA;
+			encoder = DRM_MODE_ENCODER_TMDS;
+			helpers = &tegra_sor_hdmi_helpers;
+		} else if (sor->soc->supports_lvds) {
+			connector = DRM_MODE_CONNECTOR_LVDS;
+			encoder = DRM_MODE_ENCODER_LVDS;
+		}
+	} else {
+		if (sor->output.panel) {
+			connector = DRM_MODE_CONNECTOR_eDP;
+			encoder = DRM_MODE_ENCODER_TMDS;
+			helpers = &tegra_sor_dp_helpers;
+		} else {
+			connector = DRM_MODE_CONNECTOR_DisplayPort;
+			encoder = DRM_MODE_ENCODER_TMDS;
+			helpers = &tegra_sor_dp_helpers;
+		}
+
+		sor->link.ops = &tegra_sor_dp_link_ops;
+		sor->link.aux = sor->aux;
+	}
+
+	sor->output.dev = sor->dev;
+
+	drm_connector_init_with_ddc(drm, &sor->output.connector,
+				    &tegra_sor_connector_funcs,
+				    connector,
+				    sor->output.ddc);
+	drm_connector_helper_add(&sor->output.connector,
+				 &tegra_sor_connector_helper_funcs);
+	sor->output.connector.dpms = DRM_MODE_DPMS_OFF;
+
+	drm_encoder_init(drm, &sor->output.encoder,
+			 &tegra_sor_encoder_funcs_cleanup, encoder, NULL);
+	drm_encoder_helper_add(&sor->output.encoder, helpers);
+
+	drm_connector_attach_encoder(&sor->output.connector,
+					  &sor->output.encoder);
+	drm_connector_register(&sor->output.connector);
+
+	err = tegra_output_init(drm, &sor->output);
+	if (err < 0) {
+		dev_err(client->dev, "failed to initialize output: %d\n", err);
+		return err;
+	}
+
+	tegra_output_find_possible_crtcs(&sor->output, drm);
+
+	if (sor->aux) {
+		err = drm_dp_aux_attach(sor->aux, &sor->output);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to attach DP: %d\n", err);
+			return err;
+		}
+	}
+
+	/*
+	 * XXX: Remove this reset once proper hand-over from firmware to
+	 * kernel is possible.
+	 */
+	if (sor->rst) {
+		err = pm_runtime_resume_and_get(sor->dev);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to get runtime PM: %d\n", err);
+			return err;
+		}
+
+		err = reset_control_acquire(sor->rst);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to acquire SOR reset: %d\n",
+				err);
+			goto rpm_put;
+		}
+
+		err = reset_control_assert(sor->rst);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to assert SOR reset: %d\n",
+				err);
+			goto rpm_put;
+		}
+	}
+
+	err = clk_prepare_enable(sor->clk);
+	if (err < 0) {
+		dev_err(sor->dev, "failed to enable clock: %d\n", err);
+		goto rpm_put;
+	}
+
+	usleep_range(1000, 3000);
+
+	if (sor->rst) {
+		err = reset_control_deassert(sor->rst);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to deassert SOR reset: %d\n",
+				err);
+			clk_disable_unprepare(sor->clk);
+			goto rpm_put;
+		}
+
+		reset_control_release(sor->rst);
+		pm_runtime_put(sor->dev);
+	}
+
+	err = clk_prepare_enable(sor->clk_safe);
+	if (err < 0) {
+		clk_disable_unprepare(sor->clk);
+		return err;
+	}
+
+	err = clk_prepare_enable(sor->clk_dp);
+	if (err < 0) {
+		clk_disable_unprepare(sor->clk_safe);
+		clk_disable_unprepare(sor->clk);
+		return err;
+	}
+
+	return 0;
+
+rpm_put:
+	if (sor->rst)
+		pm_runtime_put(sor->dev);
+
+	return err;
+}
+
+static int tegra_sor_exit(struct host1x_client *client)
+{
+	struct tegra_sor *sor = host1x_client_to_sor(client);
+	int err;
+
+	tegra_output_exit(&sor->output);
+
+	if (sor->aux) {
+		err = drm_dp_aux_detach(sor->aux);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to detach DP: %d\n", err);
+			return err;
+		}
+	}
+
+	clk_disable_unprepare(sor->clk_safe);
+	clk_disable_unprepare(sor->clk_dp);
+	clk_disable_unprepare(sor->clk);
+
+	return 0;
+}
+
+static int tegra_sor_runtime_suspend(struct host1x_client *client)
+{
+	struct tegra_sor *sor = host1x_client_to_sor(client);
+	struct device *dev = client->dev;
+	int err;
+
+	if (sor->rst) {
+		err = reset_control_assert(sor->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to assert reset: %d\n", err);
+			return err;
+		}
+
+		reset_control_release(sor->rst);
+	}
+
+	usleep_range(1000, 2000);
+
+	clk_disable_unprepare(sor->clk);
+	pm_runtime_put_sync(dev);
+
+	return 0;
+}
+
+static int tegra_sor_runtime_resume(struct host1x_client *client)
+{
+	struct tegra_sor *sor = host1x_client_to_sor(client);
+	struct device *dev = client->dev;
+	int err;
+
+	err = pm_runtime_resume_and_get(dev);
+	if (err < 0) {
+		dev_err(dev, "failed to get runtime PM: %d\n", err);
+		return err;
+	}
+
+	err = clk_prepare_enable(sor->clk);
+	if (err < 0) {
+		dev_err(dev, "failed to enable clock: %d\n", err);
+		goto put_rpm;
+	}
+
+	usleep_range(1000, 2000);
+
+	if (sor->rst) {
+		err = reset_control_acquire(sor->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to acquire reset: %d\n", err);
+			goto disable_clk;
+		}
+
+		err = reset_control_deassert(sor->rst);
+		if (err < 0) {
+			dev_err(dev, "failed to deassert reset: %d\n", err);
+			goto release_reset;
+		}
+	}
+
+	return 0;
+
+release_reset:
+	reset_control_release(sor->rst);
+disable_clk:
+	clk_disable_unprepare(sor->clk);
+put_rpm:
+	pm_runtime_put_sync(dev);
+	return err;
+}
+
+static const struct host1x_client_ops sor_client_ops = {
+	.init = tegra_sor_init,
+	.exit = tegra_sor_exit,
+	.suspend = tegra_sor_runtime_suspend,
+	.resume = tegra_sor_runtime_resume,
+};
+
+static const u8 tegra124_sor_xbar_cfg[5] = {
+	0, 1, 2, 3, 4
+};
+
+static const struct tegra_sor_regs tegra124_sor_regs = {
+	.head_state0 = 0x05,
+	.head_state1 = 0x07,
+	.head_state2 = 0x09,
+	.head_state3 = 0x0b,
+	.head_state4 = 0x0d,
+	.head_state5 = 0x0f,
+	.pll0 = 0x17,
+	.pll1 = 0x18,
+	.pll2 = 0x19,
+	.pll3 = 0x1a,
+	.dp_padctl0 = 0x5c,
+	.dp_padctl2 = 0x73,
+};
+
+/* Tegra124 and Tegra132 have lanes 0 and 2 swapped. */
+static const u8 tegra124_sor_lane_map[4] = {
+	2, 1, 0, 3,
+};
+
+static const u8 tegra124_sor_voltage_swing[4][4][4] = {
+	{
+		{ 0x13, 0x19, 0x1e, 0x28 },
+		{ 0x1e, 0x25, 0x2d, },
+		{ 0x28, 0x32, },
+		{ 0x3c, },
+	}, {
+		{ 0x12, 0x17, 0x1b, 0x25 },
+		{ 0x1c, 0x23, 0x2a, },
+		{ 0x25, 0x2f, },
+		{ 0x39, }
+	}, {
+		{ 0x12, 0x16, 0x1a, 0x22 },
+		{ 0x1b, 0x20, 0x27, },
+		{ 0x24, 0x2d, },
+		{ 0x36, },
+	}, {
+		{ 0x11, 0x14, 0x17, 0x1f },
+		{ 0x19, 0x1e, 0x24, },
+		{ 0x22, 0x2a, },
+		{ 0x32, },
+	},
+};
+
+static const u8 tegra124_sor_pre_emphasis[4][4][4] = {
+	{
+		{ 0x00, 0x09, 0x13, 0x25 },
+		{ 0x00, 0x0f, 0x1e, },
+		{ 0x00, 0x14, },
+		{ 0x00, },
+	}, {
+		{ 0x00, 0x0a, 0x14, 0x28 },
+		{ 0x00, 0x0f, 0x1e, },
+		{ 0x00, 0x14, },
+		{ 0x00 },
+	}, {
+		{ 0x00, 0x0a, 0x14, 0x28 },
+		{ 0x00, 0x0f, 0x1e, },
+		{ 0x00, 0x14, },
+		{ 0x00, },
+	}, {
+		{ 0x00, 0x0a, 0x14, 0x28 },
+		{ 0x00, 0x0f, 0x1e, },
+		{ 0x00, 0x14, },
+		{ 0x00, },
+	},
+};
+
+static const u8 tegra124_sor_post_cursor[4][4][4] = {
+	{
+		{ 0x00, 0x00, 0x00, 0x00 },
+		{ 0x00, 0x00, 0x00, },
+		{ 0x00, 0x00, },
+		{ 0x00, },
+	}, {
+		{ 0x02, 0x02, 0x04, 0x05 },
+		{ 0x02, 0x04, 0x05, },
+		{ 0x04, 0x05, },
+		{ 0x05, },
+	}, {
+		{ 0x04, 0x05, 0x08, 0x0b },
+		{ 0x05, 0x09, 0x0b, },
+		{ 0x08, 0x0a, },
+		{ 0x0b, },
+	}, {
+		{ 0x05, 0x09, 0x0b, 0x12 },
+		{ 0x09, 0x0d, 0x12, },
+		{ 0x0b, 0x0f, },
+		{ 0x12, },
+	},
+};
+
+static const u8 tegra124_sor_tx_pu[4][4][4] = {
+	{
+		{ 0x20, 0x30, 0x40, 0x60 },
+		{ 0x30, 0x40, 0x60, },
+		{ 0x40, 0x60, },
+		{ 0x60, },
+	}, {
+		{ 0x20, 0x20, 0x30, 0x50 },
+		{ 0x30, 0x40, 0x50, },
+		{ 0x40, 0x50, },
+		{ 0x60, },
+	}, {
+		{ 0x20, 0x20, 0x30, 0x40, },
+		{ 0x30, 0x30, 0x40, },
+		{ 0x40, 0x50, },
+		{ 0x60, },
+	}, {
+		{ 0x20, 0x20, 0x20, 0x40, },
+		{ 0x30, 0x30, 0x40, },
+		{ 0x40, 0x40, },
+		{ 0x60, },
+	},
+};
+
+static const struct tegra_sor_soc tegra124_sor = {
+	.supports_lvds = true,
+	.supports_hdmi = false,
+	.supports_dp = true,
+	.supports_audio = false,
+	.supports_hdcp = false,
+	.regs = &tegra124_sor_regs,
+	.has_nvdisplay = false,
+	.xbar_cfg = tegra124_sor_xbar_cfg,
+	.lane_map = tegra124_sor_lane_map,
+	.voltage_swing = tegra124_sor_voltage_swing,
+	.pre_emphasis = tegra124_sor_pre_emphasis,
+	.post_cursor = tegra124_sor_post_cursor,
+	.tx_pu = tegra124_sor_tx_pu,
+};
+
+static const u8 tegra132_sor_pre_emphasis[4][4][4] = {
+	{
+		{ 0x00, 0x08, 0x12, 0x24 },
+		{ 0x01, 0x0e, 0x1d, },
+		{ 0x01, 0x13, },
+		{ 0x00, },
+	}, {
+		{ 0x00, 0x08, 0x12, 0x24 },
+		{ 0x00, 0x0e, 0x1d, },
+		{ 0x00, 0x13, },
+		{ 0x00 },
+	}, {
+		{ 0x00, 0x08, 0x12, 0x24 },
+		{ 0x00, 0x0e, 0x1d, },
+		{ 0x00, 0x13, },
+		{ 0x00, },
+	}, {
+		{ 0x00, 0x08, 0x12, 0x24 },
+		{ 0x00, 0x0e, 0x1d, },
+		{ 0x00, 0x13, },
+		{ 0x00, },
+	},
+};
+
+static const struct tegra_sor_soc tegra132_sor = {
+	.supports_lvds = true,
+	.supports_hdmi = false,
+	.supports_dp = true,
+	.supports_audio = false,
+	.supports_hdcp = false,
+	.regs = &tegra124_sor_regs,
+	.has_nvdisplay = false,
+	.xbar_cfg = tegra124_sor_xbar_cfg,
+	.lane_map = tegra124_sor_lane_map,
+	.voltage_swing = tegra124_sor_voltage_swing,
+	.pre_emphasis = tegra132_sor_pre_emphasis,
+	.post_cursor = tegra124_sor_post_cursor,
+	.tx_pu = tegra124_sor_tx_pu,
+};
+
+static const struct tegra_sor_regs tegra210_sor_regs = {
+	.head_state0 = 0x05,
+	.head_state1 = 0x07,
+	.head_state2 = 0x09,
+	.head_state3 = 0x0b,
+	.head_state4 = 0x0d,
+	.head_state5 = 0x0f,
+	.pll0 = 0x17,
+	.pll1 = 0x18,
+	.pll2 = 0x19,
+	.pll3 = 0x1a,
+	.dp_padctl0 = 0x5c,
+	.dp_padctl2 = 0x73,
+};
+
+static const u8 tegra210_sor_xbar_cfg[5] = {
+	2, 1, 0, 3, 4
+};
+
+static const u8 tegra210_sor_lane_map[4] = {
+	0, 1, 2, 3,
+};
+
+static const struct tegra_sor_soc tegra210_sor = {
+	.supports_lvds = false,
+	.supports_hdmi = false,
+	.supports_dp = true,
+	.supports_audio = false,
+	.supports_hdcp = false,
+
+	.regs = &tegra210_sor_regs,
+	.has_nvdisplay = false,
+
+	.xbar_cfg = tegra210_sor_xbar_cfg,
+	.lane_map = tegra210_sor_lane_map,
+	.voltage_swing = tegra124_sor_voltage_swing,
+	.pre_emphasis = tegra124_sor_pre_emphasis,
+	.post_cursor = tegra124_sor_post_cursor,
+	.tx_pu = tegra124_sor_tx_pu,
+};
+
+static const struct tegra_sor_soc tegra210_sor1 = {
+	.supports_lvds = false,
+	.supports_hdmi = true,
+	.supports_dp = true,
+	.supports_audio = true,
+	.supports_hdcp = true,
+
+	.regs = &tegra210_sor_regs,
+	.has_nvdisplay = false,
+
+	.num_settings = ARRAY_SIZE(tegra210_sor_hdmi_defaults),
+	.settings = tegra210_sor_hdmi_defaults,
+	.xbar_cfg = tegra210_sor_xbar_cfg,
+	.lane_map = tegra210_sor_lane_map,
+	.voltage_swing = tegra124_sor_voltage_swing,
+	.pre_emphasis = tegra124_sor_pre_emphasis,
+	.post_cursor = tegra124_sor_post_cursor,
+	.tx_pu = tegra124_sor_tx_pu,
+};
+
+static const struct tegra_sor_regs tegra186_sor_regs = {
+	.head_state0 = 0x151,
+	.head_state1 = 0x154,
+	.head_state2 = 0x157,
+	.head_state3 = 0x15a,
+	.head_state4 = 0x15d,
+	.head_state5 = 0x160,
+	.pll0 = 0x163,
+	.pll1 = 0x164,
+	.pll2 = 0x165,
+	.pll3 = 0x166,
+	.dp_padctl0 = 0x168,
+	.dp_padctl2 = 0x16a,
+};
+
+static const u8 tegra186_sor_voltage_swing[4][4][4] = {
+	{
+		{ 0x13, 0x19, 0x1e, 0x28 },
+		{ 0x1e, 0x25, 0x2d, },
+		{ 0x28, 0x32, },
+		{ 0x39, },
+	}, {
+		{ 0x12, 0x16, 0x1b, 0x25 },
+		{ 0x1c, 0x23, 0x2a, },
+		{ 0x25, 0x2f, },
+		{ 0x37, }
+	}, {
+		{ 0x12, 0x16, 0x1a, 0x22 },
+		{ 0x1b, 0x20, 0x27, },
+		{ 0x24, 0x2d, },
+		{ 0x35, },
+	}, {
+		{ 0x11, 0x14, 0x17, 0x1f },
+		{ 0x19, 0x1e, 0x24, },
+		{ 0x22, 0x2a, },
+		{ 0x32, },
+	},
+};
+
+static const u8 tegra186_sor_pre_emphasis[4][4][4] = {
+	{
+		{ 0x00, 0x08, 0x12, 0x24 },
+		{ 0x01, 0x0e, 0x1d, },
+		{ 0x01, 0x13, },
+		{ 0x00, },
+	}, {
+		{ 0x00, 0x08, 0x12, 0x24 },
+		{ 0x00, 0x0e, 0x1d, },
+		{ 0x00, 0x13, },
+		{ 0x00 },
+	}, {
+		{ 0x00, 0x08, 0x14, 0x24 },
+		{ 0x00, 0x0e, 0x1d, },
+		{ 0x00, 0x13, },
+		{ 0x00, },
+	}, {
+		{ 0x00, 0x08, 0x12, 0x24 },
+		{ 0x00, 0x0e, 0x1d, },
+		{ 0x00, 0x13, },
+		{ 0x00, },
+	},
+};
+
+static const struct tegra_sor_soc tegra186_sor = {
+	.supports_lvds = false,
+	.supports_hdmi = true,
+	.supports_dp = true,
+	.supports_audio = true,
+	.supports_hdcp = true,
+
+	.regs = &tegra186_sor_regs,
+	.has_nvdisplay = true,
+
+	.num_settings = ARRAY_SIZE(tegra186_sor_hdmi_defaults),
+	.settings = tegra186_sor_hdmi_defaults,
+	.xbar_cfg = tegra124_sor_xbar_cfg,
+	.lane_map = tegra124_sor_lane_map,
+	.voltage_swing = tegra186_sor_voltage_swing,
+	.pre_emphasis = tegra186_sor_pre_emphasis,
+	.post_cursor = tegra124_sor_post_cursor,
+	.tx_pu = tegra124_sor_tx_pu,
+};
+
+static const struct tegra_sor_regs tegra194_sor_regs = {
+	.head_state0 = 0x151,
+	.head_state1 = 0x155,
+	.head_state2 = 0x159,
+	.head_state3 = 0x15d,
+	.head_state4 = 0x161,
+	.head_state5 = 0x165,
+	.pll0 = 0x169,
+	.pll1 = 0x16a,
+	.pll2 = 0x16b,
+	.pll3 = 0x16c,
+	.dp_padctl0 = 0x16e,
+	.dp_padctl2 = 0x16f,
+};
+
+static const struct tegra_sor_soc tegra194_sor = {
+	.supports_lvds = false,
+	.supports_hdmi = true,
+	.supports_dp = true,
+	.supports_audio = true,
+	.supports_hdcp = true,
+
+	.regs = &tegra194_sor_regs,
+	.has_nvdisplay = true,
+
+	.num_settings = ARRAY_SIZE(tegra194_sor_hdmi_defaults),
+	.settings = tegra194_sor_hdmi_defaults,
+
+	.xbar_cfg = tegra210_sor_xbar_cfg,
+	.lane_map = tegra124_sor_lane_map,
+	.voltage_swing = tegra186_sor_voltage_swing,
+	.pre_emphasis = tegra186_sor_pre_emphasis,
+	.post_cursor = tegra124_sor_post_cursor,
+	.tx_pu = tegra124_sor_tx_pu,
+};
+
+static const struct of_device_id tegra_sor_of_match[] = {
+	{ .compatible = "nvidia,tegra194-sor", .data = &tegra194_sor },
+	{ .compatible = "nvidia,tegra186-sor", .data = &tegra186_sor },
+	{ .compatible = "nvidia,tegra210-sor1", .data = &tegra210_sor1 },
+	{ .compatible = "nvidia,tegra210-sor", .data = &tegra210_sor },
+	{ .compatible = "nvidia,tegra132-sor", .data = &tegra132_sor },
+	{ .compatible = "nvidia,tegra124-sor", .data = &tegra124_sor },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, tegra_sor_of_match);
+
+static int tegra_sor_parse_dt(struct tegra_sor *sor)
+{
+	struct device_node *np = sor->dev->of_node;
+	u32 xbar_cfg[5];
+	unsigned int i;
+	u32 value;
+	int err;
+
+	if (sor->soc->has_nvdisplay) {
+		err = of_property_read_u32(np, "nvidia,interface", &value);
+		if (err < 0)
+			return err;
+
+		sor->index = value;
+
+		/*
+		 * override the default that we already set for Tegra210 and
+		 * earlier
+		 */
+		sor->pad = TEGRA_IO_PAD_HDMI_DP0 + sor->index;
+	} else {
+		if (!sor->soc->supports_audio)
+			sor->index = 0;
+		else
+			sor->index = 1;
+	}
+
+	err = of_property_read_u32_array(np, "nvidia,xbar-cfg", xbar_cfg, 5);
+	if (err < 0) {
+		/* fall back to default per-SoC XBAR configuration */
+		for (i = 0; i < 5; i++)
+			sor->xbar_cfg[i] = sor->soc->xbar_cfg[i];
+	} else {
+		/* copy cells to SOR XBAR configuration */
+		for (i = 0; i < 5; i++)
+			sor->xbar_cfg[i] = xbar_cfg[i];
+	}
+
+	return 0;
+}
+
+static irqreturn_t tegra_sor_irq(int irq, void *data)
+{
+	struct tegra_sor *sor = data;
+	u32 value;
+
+	value = tegra_sor_readl(sor, SOR_INT_STATUS);
+	tegra_sor_writel(sor, value, SOR_INT_STATUS);
+
+	if (value & SOR_INT_CODEC_SCRATCH0) {
+		value = tegra_sor_readl(sor, SOR_AUDIO_HDA_CODEC_SCRATCH0);
+
+		if (value & SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID) {
+			unsigned int format;
+
+			format = value & SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK;
+
+			tegra_hda_parse_format(format, &sor->format);
+
+			if (sor->ops->audio_enable)
+				sor->ops->audio_enable(sor);
+		} else {
+			if (sor->ops->audio_disable)
+				sor->ops->audio_disable(sor);
+		}
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int tegra_sor_probe(struct platform_device *pdev)
+{
+	struct device_node *np;
+	struct tegra_sor *sor;
+	int err;
+
+	sor = devm_kzalloc(&pdev->dev, sizeof(*sor), GFP_KERNEL);
+	if (!sor)
+		return -ENOMEM;
+
+	sor->soc = of_device_get_match_data(&pdev->dev);
+	sor->output.dev = sor->dev = &pdev->dev;
+
+	sor->settings = devm_kmemdup(&pdev->dev, sor->soc->settings,
+				     sor->soc->num_settings *
+					sizeof(*sor->settings),
+				     GFP_KERNEL);
+	if (!sor->settings)
+		return -ENOMEM;
+
+	sor->num_settings = sor->soc->num_settings;
+
+	sor->pmc = devm_tegra_pmc_get(&pdev->dev);
+	if (IS_ERR(sor->pmc))
+		return PTR_ERR(sor->pmc);
+
+	np = of_parse_phandle(pdev->dev.of_node, "nvidia,dpaux", 0);
+	if (np) {
+		sor->aux = drm_dp_aux_find_by_of_node(np);
+		of_node_put(np);
+
+		if (!sor->aux)
+			return -EPROBE_DEFER;
+
+		if (get_device(sor->aux->dev))
+			sor->output.ddc = &sor->aux->ddc;
+	}
+
+	if (!sor->aux) {
+		if (sor->soc->supports_hdmi) {
+			sor->ops = &tegra_sor_hdmi_ops;
+			sor->pad = TEGRA_IO_PAD_HDMI;
+		} else if (sor->soc->supports_lvds) {
+			dev_err(&pdev->dev, "LVDS not supported yet\n");
+			return -ENODEV;
+		} else {
+			dev_err(&pdev->dev, "unknown (non-DP) support\n");
+			return -ENODEV;
+		}
+	} else {
+		np = of_parse_phandle(pdev->dev.of_node, "nvidia,panel", 0);
+		/*
+		 * No need to keep this around since we only use it as a check
+		 * to see if a panel is connected (eDP) or not (DP).
+		 */
+		of_node_put(np);
+
+		sor->ops = &tegra_sor_dp_ops;
+		sor->pad = TEGRA_IO_PAD_LVDS;
+	}
+
+	err = tegra_sor_parse_dt(sor);
+	if (err < 0)
+		goto put_aux;
+
+	err = tegra_output_probe(&sor->output);
+	if (err < 0) {
+		dev_err_probe(&pdev->dev, err, "failed to probe output\n");
+		goto put_aux;
+	}
+
+	if (sor->ops && sor->ops->probe) {
+		err = sor->ops->probe(sor);
+		if (err < 0) {
+			dev_err(&pdev->dev, "failed to probe %s: %d\n",
+				sor->ops->name, err);
+			goto remove;
+		}
+	}
+
+	sor->regs = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(sor->regs)) {
+		err = PTR_ERR(sor->regs);
+		goto remove;
+	}
+
+	err = platform_get_irq(pdev, 0);
+	if (err < 0)
+		goto remove;
+
+	sor->irq = err;
+
+	err = devm_request_irq(sor->dev, sor->irq, tegra_sor_irq, 0,
+			       dev_name(sor->dev), sor);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to request IRQ: %d\n", err);
+		goto remove;
+	}
+
+	sor->rst = devm_reset_control_get_exclusive_released(&pdev->dev, "sor");
+	if (IS_ERR(sor->rst)) {
+		err = PTR_ERR(sor->rst);
+
+		if (err != -EBUSY || WARN_ON(!pdev->dev.pm_domain)) {
+			dev_err(&pdev->dev, "failed to get reset control: %d\n",
+				err);
+			goto remove;
+		}
+
+		/*
+		 * At this point, the reset control is most likely being used
+		 * by the generic power domain implementation. With any luck
+		 * the power domain will have taken care of resetting the SOR
+		 * and we don't have to do anything.
+		 */
+		sor->rst = NULL;
+	}
+
+	sor->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(sor->clk)) {
+		err = PTR_ERR(sor->clk);
+		dev_err(&pdev->dev, "failed to get module clock: %d\n", err);
+		goto remove;
+	}
+
+	if (sor->soc->supports_hdmi || sor->soc->supports_dp) {
+		struct device_node *np = pdev->dev.of_node;
+		const char *name;
+
+		/*
+		 * For backwards compatibility with Tegra210 device trees,
+		 * fall back to the old clock name "source" if the new "out"
+		 * clock is not available.
+		 */
+		if (of_property_match_string(np, "clock-names", "out") < 0)
+			name = "source";
+		else
+			name = "out";
+
+		sor->clk_out = devm_clk_get(&pdev->dev, name);
+		if (IS_ERR(sor->clk_out)) {
+			err = PTR_ERR(sor->clk_out);
+			dev_err(sor->dev, "failed to get %s clock: %d\n",
+				name, err);
+			goto remove;
+		}
+	} else {
+		/* fall back to the module clock on SOR0 (eDP/LVDS only) */
+		sor->clk_out = sor->clk;
+	}
+
+	sor->clk_parent = devm_clk_get(&pdev->dev, "parent");
+	if (IS_ERR(sor->clk_parent)) {
+		err = PTR_ERR(sor->clk_parent);
+		dev_err(&pdev->dev, "failed to get parent clock: %d\n", err);
+		goto remove;
+	}
+
+	sor->clk_safe = devm_clk_get(&pdev->dev, "safe");
+	if (IS_ERR(sor->clk_safe)) {
+		err = PTR_ERR(sor->clk_safe);
+		dev_err(&pdev->dev, "failed to get safe clock: %d\n", err);
+		goto remove;
+	}
+
+	sor->clk_dp = devm_clk_get(&pdev->dev, "dp");
+	if (IS_ERR(sor->clk_dp)) {
+		err = PTR_ERR(sor->clk_dp);
+		dev_err(&pdev->dev, "failed to get DP clock: %d\n", err);
+		goto remove;
+	}
+
+	/*
+	 * Starting with Tegra186, the BPMP provides an implementation for
+	 * the pad output clock, so we have to look it up from device tree.
+	 */
+	sor->clk_pad = devm_clk_get(&pdev->dev, "pad");
+	if (IS_ERR(sor->clk_pad)) {
+		if (sor->clk_pad != ERR_PTR(-ENOENT)) {
+			err = PTR_ERR(sor->clk_pad);
+			goto remove;
+		}
+
+		/*
+		 * If the pad output clock is not available, then we assume
+		 * we're on Tegra210 or earlier and have to provide our own
+		 * implementation.
+		 */
+		sor->clk_pad = NULL;
+	}
+
+	/*
+	 * The bootloader may have set up the SOR such that it's module clock
+	 * is sourced by one of the display PLLs. However, that doesn't work
+	 * without properly having set up other bits of the SOR.
+	 */
+	err = clk_set_parent(sor->clk_out, sor->clk_safe);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to use safe clock: %d\n", err);
+		goto remove;
+	}
+
+	platform_set_drvdata(pdev, sor);
+	pm_runtime_enable(&pdev->dev);
+
+	host1x_client_init(&sor->client);
+	sor->client.ops = &sor_client_ops;
+	sor->client.dev = &pdev->dev;
+
+	/*
+	 * On Tegra210 and earlier, provide our own implementation for the
+	 * pad output clock.
+	 */
+	if (!sor->clk_pad) {
+		char *name;
+
+		name = devm_kasprintf(sor->dev, GFP_KERNEL, "sor%u_pad_clkout",
+				      sor->index);
+		if (!name) {
+			err = -ENOMEM;
+			goto uninit;
+		}
+
+		err = host1x_client_resume(&sor->client);
+		if (err < 0) {
+			dev_err(sor->dev, "failed to resume: %d\n", err);
+			goto uninit;
+		}
+
+		sor->clk_pad = tegra_clk_sor_pad_register(sor, name);
+		host1x_client_suspend(&sor->client);
+	}
+
+	if (IS_ERR(sor->clk_pad)) {
+		err = PTR_ERR(sor->clk_pad);
+		dev_err(sor->dev, "failed to register SOR pad clock: %d\n",
+			err);
+		goto uninit;
+	}
+
+	err = __host1x_client_register(&sor->client);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to register host1x client: %d\n",
+			err);
+		goto uninit;
+	}
+
+	return 0;
+
+uninit:
+	host1x_client_exit(&sor->client);
+	pm_runtime_disable(&pdev->dev);
+remove:
+	if (sor->aux)
+		sor->output.ddc = NULL;
+
+	tegra_output_remove(&sor->output);
+put_aux:
+	if (sor->aux)
+		put_device(sor->aux->dev);
+
+	return err;
+}
+
+static void tegra_sor_remove(struct platform_device *pdev)
+{
+	struct tegra_sor *sor = platform_get_drvdata(pdev);
+
+	host1x_client_unregister(&sor->client);
+
+	pm_runtime_disable(&pdev->dev);
+
+	if (sor->aux) {
+		put_device(sor->aux->dev);
+		sor->output.ddc = NULL;
+	}
+
+	tegra_output_remove(&sor->output);
+}
+
+static int __maybe_unused tegra_sor_suspend(struct device *dev)
+{
+	struct tegra_sor *sor = dev_get_drvdata(dev);
+	int err;
+
+	err = tegra_output_suspend(&sor->output);
+	if (err < 0) {
+		dev_err(dev, "failed to suspend output: %d\n", err);
+		return err;
+	}
+
+	if (sor->hdmi_supply) {
+		err = regulator_disable(sor->hdmi_supply);
+		if (err < 0) {
+			tegra_output_resume(&sor->output);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static int __maybe_unused tegra_sor_resume(struct device *dev)
+{
+	struct tegra_sor *sor = dev_get_drvdata(dev);
+	int err;
+
+	if (sor->hdmi_supply) {
+		err = regulator_enable(sor->hdmi_supply);
+		if (err < 0)
+			return err;
+	}
+
+	err = tegra_output_resume(&sor->output);
+	if (err < 0) {
+		dev_err(dev, "failed to resume output: %d\n", err);
+
+		if (sor->hdmi_supply)
+			regulator_disable(sor->hdmi_supply);
+
+		return err;
+	}
+
+	return 0;
+}
+
+static const struct dev_pm_ops tegra_sor_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(tegra_sor_suspend, tegra_sor_resume)
+};
+
+struct platform_driver tegra_sor_driver = {
+	.driver = {
+		.name = "tegra-sor",
+		.of_match_table = tegra_sor_of_match,
+		.pm = &tegra_sor_pm_ops,
+	},
+	.probe = tegra_sor_probe,
+	.remove = tegra_sor_remove,
+};
diff --git a/drivers/gpu/drm/tegra/sor.h b/drivers/gpu/drm/tegra/sor.h
new file mode 100644
index 000000000..00e09d5dc
--- /dev/null
+++ b/drivers/gpu/drm/tegra/sor.h
@@ -0,0 +1,457 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (C) 2013 NVIDIA Corporation
+ */
+
+#ifndef DRM_TEGRA_SOR_H
+#define DRM_TEGRA_SOR_H
+
+#define SOR_CTXSW 0x00
+
+#define SOR_SUPER_STATE0 0x01
+
+#define SOR_SUPER_STATE1 0x02
+#define  SOR_SUPER_STATE_ATTACHED		(1 << 3)
+#define  SOR_SUPER_STATE_MODE_NORMAL		(1 << 2)
+#define  SOR_SUPER_STATE_HEAD_MODE_MASK		(3 << 0)
+#define  SOR_SUPER_STATE_HEAD_MODE_AWAKE	(2 << 0)
+#define  SOR_SUPER_STATE_HEAD_MODE_SNOOZE	(1 << 0)
+#define  SOR_SUPER_STATE_HEAD_MODE_SLEEP	(0 << 0)
+
+#define SOR_STATE0 0x03
+
+#define SOR_STATE1 0x04
+#define  SOR_STATE_ASY_PIXELDEPTH_MASK		(0xf << 17)
+#define  SOR_STATE_ASY_PIXELDEPTH_BPP_18_444	(0x2 << 17)
+#define  SOR_STATE_ASY_PIXELDEPTH_BPP_24_444	(0x5 << 17)
+#define  SOR_STATE_ASY_PIXELDEPTH_BPP_30_444	(0x6 << 17)
+#define  SOR_STATE_ASY_PIXELDEPTH_BPP_36_444	(0x8 << 17)
+#define  SOR_STATE_ASY_PIXELDEPTH_BPP_48_444	(0x9 << 17)
+#define  SOR_STATE_ASY_VSYNCPOL			(1 << 13)
+#define  SOR_STATE_ASY_HSYNCPOL			(1 << 12)
+#define  SOR_STATE_ASY_PROTOCOL_MASK		(0xf << 8)
+#define  SOR_STATE_ASY_PROTOCOL_CUSTOM		(0xf << 8)
+#define  SOR_STATE_ASY_PROTOCOL_DP_A		(0x8 << 8)
+#define  SOR_STATE_ASY_PROTOCOL_DP_B		(0x9 << 8)
+#define  SOR_STATE_ASY_PROTOCOL_SINGLE_TMDS_A	(0x1 << 8)
+#define  SOR_STATE_ASY_PROTOCOL_LVDS		(0x0 << 8)
+#define  SOR_STATE_ASY_CRC_MODE_MASK		(0x3 << 6)
+#define  SOR_STATE_ASY_CRC_MODE_NON_ACTIVE	(0x2 << 6)
+#define  SOR_STATE_ASY_CRC_MODE_COMPLETE	(0x1 << 6)
+#define  SOR_STATE_ASY_CRC_MODE_ACTIVE		(0x0 << 6)
+#define  SOR_STATE_ASY_SUBOWNER_MASK		(0x3 << 4)
+#define  SOR_STATE_ASY_OWNER_MASK		0xf
+#define  SOR_STATE_ASY_OWNER(x)			(((x) & 0xf) << 0)
+
+#define SOR_HEAD_STATE0(x) (0x05 + (x))
+#define  SOR_HEAD_STATE_RANGECOMPRESS_MASK (0x1 << 3)
+#define  SOR_HEAD_STATE_DYNRANGE_MASK (0x1 << 2)
+#define  SOR_HEAD_STATE_DYNRANGE_VESA (0 << 2)
+#define  SOR_HEAD_STATE_DYNRANGE_CEA (1 << 2)
+#define  SOR_HEAD_STATE_COLORSPACE_MASK (0x3 << 0)
+#define  SOR_HEAD_STATE_COLORSPACE_RGB (0 << 0)
+#define SOR_HEAD_STATE1(x) (0x07 + (x))
+#define SOR_HEAD_STATE2(x) (0x09 + (x))
+#define SOR_HEAD_STATE3(x) (0x0b + (x))
+#define SOR_HEAD_STATE4(x) (0x0d + (x))
+#define SOR_HEAD_STATE5(x) (0x0f + (x))
+#define SOR_CRC_CNTRL 0x11
+#define  SOR_CRC_CNTRL_ENABLE			(1 << 0)
+#define SOR_DP_DEBUG_MVID 0x12
+
+#define SOR_CLK_CNTRL 0x13
+#define  SOR_CLK_CNTRL_DP_LINK_SPEED_MASK	(0x1f << 2)
+#define  SOR_CLK_CNTRL_DP_LINK_SPEED(x)		(((x) & 0x1f) << 2)
+#define  SOR_CLK_CNTRL_DP_LINK_SPEED_G1_62	(0x06 << 2)
+#define  SOR_CLK_CNTRL_DP_LINK_SPEED_G2_70	(0x0a << 2)
+#define  SOR_CLK_CNTRL_DP_LINK_SPEED_G5_40	(0x14 << 2)
+#define  SOR_CLK_CNTRL_DP_CLK_SEL_MASK		(3 << 0)
+#define  SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_PCLK	(0 << 0)
+#define  SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_PCLK	(1 << 0)
+#define  SOR_CLK_CNTRL_DP_CLK_SEL_SINGLE_DPCLK	(2 << 0)
+#define  SOR_CLK_CNTRL_DP_CLK_SEL_DIFF_DPCLK	(3 << 0)
+
+#define SOR_CAP 0x14
+
+#define SOR_PWR 0x15
+#define  SOR_PWR_TRIGGER			(1 << 31)
+#define  SOR_PWR_MODE_SAFE			(1 << 28)
+#define  SOR_PWR_NORMAL_STATE_PU		(1 << 0)
+
+#define SOR_TEST 0x16
+#define  SOR_TEST_CRC_POST_SERIALIZE		(1 << 23)
+#define  SOR_TEST_ATTACHED			(1 << 10)
+#define  SOR_TEST_HEAD_MODE_MASK		(3 << 8)
+#define  SOR_TEST_HEAD_MODE_AWAKE		(2 << 8)
+
+#define SOR_PLL0 0x17
+#define  SOR_PLL0_ICHPMP_MASK			(0xf << 24)
+#define  SOR_PLL0_ICHPMP(x)			(((x) & 0xf) << 24)
+#define  SOR_PLL0_FILTER_MASK			(0xf << 16)
+#define  SOR_PLL0_FILTER(x)			(((x) & 0xf) << 16)
+#define  SOR_PLL0_VCOCAP_MASK			(0xf << 8)
+#define  SOR_PLL0_VCOCAP(x)			(((x) & 0xf) << 8)
+#define  SOR_PLL0_VCOCAP_RST			SOR_PLL0_VCOCAP(3)
+#define  SOR_PLL0_PLLREG_MASK			(0x3 << 6)
+#define  SOR_PLL0_PLLREG_LEVEL(x)		(((x) & 0x3) << 6)
+#define  SOR_PLL0_PLLREG_LEVEL_V25		SOR_PLL0_PLLREG_LEVEL(0)
+#define  SOR_PLL0_PLLREG_LEVEL_V15		SOR_PLL0_PLLREG_LEVEL(1)
+#define  SOR_PLL0_PLLREG_LEVEL_V35		SOR_PLL0_PLLREG_LEVEL(2)
+#define  SOR_PLL0_PLLREG_LEVEL_V45		SOR_PLL0_PLLREG_LEVEL(3)
+#define  SOR_PLL0_PULLDOWN			(1 << 5)
+#define  SOR_PLL0_RESISTOR_EXT			(1 << 4)
+#define  SOR_PLL0_VCOPD				(1 << 2)
+#define  SOR_PLL0_PWR				(1 << 0)
+
+#define SOR_PLL1 0x18
+/* XXX: read-only bit? */
+#define  SOR_PLL1_LOADADJ_MASK			(0xf << 20)
+#define  SOR_PLL1_LOADADJ(x)			(((x) & 0xf) << 20)
+#define  SOR_PLL1_TERM_COMPOUT			(1 << 15)
+#define  SOR_PLL1_TMDS_TERMADJ_MASK		(0xf << 9)
+#define  SOR_PLL1_TMDS_TERMADJ(x)		(((x) & 0xf) << 9)
+#define  SOR_PLL1_TMDS_TERM			(1 << 8)
+
+#define SOR_PLL2 0x19
+#define  SOR_PLL2_LVDS_ENABLE			(1 << 25)
+#define  SOR_PLL2_SEQ_PLLCAPPD_ENFORCE		(1 << 24)
+#define  SOR_PLL2_PORT_POWERDOWN		(1 << 23)
+#define  SOR_PLL2_BANDGAP_POWERDOWN		(1 << 22)
+#define  SOR_PLL2_POWERDOWN_OVERRIDE		(1 << 18)
+#define  SOR_PLL2_SEQ_PLLCAPPD			(1 << 17)
+#define  SOR_PLL2_SEQ_PLL_PULLDOWN		(1 << 16)
+
+#define SOR_PLL3 0x1a
+#define  SOR_PLL3_BG_TEMP_COEF_MASK		(0xf << 28)
+#define  SOR_PLL3_BG_TEMP_COEF(x)		(((x) & 0xf) << 28)
+#define  SOR_PLL3_BG_VREF_LEVEL_MASK		(0xf << 24)
+#define  SOR_PLL3_BG_VREF_LEVEL(x)		(((x) & 0xf) << 24)
+#define  SOR_PLL3_PLL_VDD_MODE_1V8		(0 << 13)
+#define  SOR_PLL3_PLL_VDD_MODE_3V3		(1 << 13)
+#define  SOR_PLL3_AVDD10_LEVEL_MASK		(0xf << 8)
+#define  SOR_PLL3_AVDD10_LEVEL(x)		(((x) & 0xf) << 8)
+#define  SOR_PLL3_AVDD14_LEVEL_MASK		(0xf << 4)
+#define  SOR_PLL3_AVDD14_LEVEL(x)		(((x) & 0xf) << 4)
+
+#define SOR_CSTM 0x1b
+#define  SOR_CSTM_ROTCLK_MASK			(0xf << 24)
+#define  SOR_CSTM_ROTCLK(x)			(((x) & 0xf) << 24)
+#define  SOR_CSTM_LVDS				(1 << 16)
+#define  SOR_CSTM_LINK_ACT_B			(1 << 15)
+#define  SOR_CSTM_LINK_ACT_A			(1 << 14)
+#define  SOR_CSTM_UPPER				(1 << 11)
+
+#define SOR_LVDS 0x1c
+#define SOR_CRCA 0x1d
+#define  SOR_CRCA_VALID			(1 << 0)
+#define  SOR_CRCA_RESET			(1 << 0)
+#define SOR_CRCB 0x1e
+#define SOR_BLANK 0x1f
+#define SOR_SEQ_CTL 0x20
+#define  SOR_SEQ_CTL_PD_PC_ALT(x)	(((x) & 0xf) << 12)
+#define  SOR_SEQ_CTL_PD_PC(x)		(((x) & 0xf) <<  8)
+#define  SOR_SEQ_CTL_PU_PC_ALT(x)	(((x) & 0xf) <<  4)
+#define  SOR_SEQ_CTL_PU_PC(x)		(((x) & 0xf) <<  0)
+
+#define SOR_LANE_SEQ_CTL 0x21
+#define  SOR_LANE_SEQ_CTL_TRIGGER		(1 << 31)
+#define  SOR_LANE_SEQ_CTL_STATE_BUSY		(1 << 28)
+#define  SOR_LANE_SEQ_CTL_SEQUENCE_UP		(0 << 20)
+#define  SOR_LANE_SEQ_CTL_SEQUENCE_DOWN		(1 << 20)
+#define  SOR_LANE_SEQ_CTL_POWER_STATE_UP	(0 << 16)
+#define  SOR_LANE_SEQ_CTL_POWER_STATE_DOWN	(1 << 16)
+#define  SOR_LANE_SEQ_CTL_DELAY(x)		(((x) & 0xf) << 12)
+
+#define SOR_SEQ_INST(x) (0x22 + (x))
+#define  SOR_SEQ_INST_PLL_PULLDOWN (1 << 31)
+#define  SOR_SEQ_INST_POWERDOWN_MACRO (1 << 30)
+#define  SOR_SEQ_INST_ASSERT_PLL_RESET (1 << 29)
+#define  SOR_SEQ_INST_BLANK_V (1 << 28)
+#define  SOR_SEQ_INST_BLANK_H (1 << 27)
+#define  SOR_SEQ_INST_BLANK_DE (1 << 26)
+#define  SOR_SEQ_INST_BLACK_DATA (1 << 25)
+#define  SOR_SEQ_INST_TRISTATE_IOS (1 << 24)
+#define  SOR_SEQ_INST_DRIVE_PWM_OUT_LO (1 << 23)
+#define  SOR_SEQ_INST_PIN_B_LOW (0 << 22)
+#define  SOR_SEQ_INST_PIN_B_HIGH (1 << 22)
+#define  SOR_SEQ_INST_PIN_A_LOW (0 << 21)
+#define  SOR_SEQ_INST_PIN_A_HIGH (1 << 21)
+#define  SOR_SEQ_INST_SEQUENCE_UP (0 << 19)
+#define  SOR_SEQ_INST_SEQUENCE_DOWN (1 << 19)
+#define  SOR_SEQ_INST_LANE_SEQ_STOP (0 << 18)
+#define  SOR_SEQ_INST_LANE_SEQ_RUN (1 << 18)
+#define  SOR_SEQ_INST_PORT_POWERDOWN (1 << 17)
+#define  SOR_SEQ_INST_PLL_POWERDOWN (1 << 16)
+#define  SOR_SEQ_INST_HALT (1 << 15)
+#define  SOR_SEQ_INST_WAIT_US (0 << 12)
+#define  SOR_SEQ_INST_WAIT_MS (1 << 12)
+#define  SOR_SEQ_INST_WAIT_VSYNC (2 << 12)
+#define  SOR_SEQ_INST_WAIT(x) (((x) & 0x3ff) << 0)
+
+#define SOR_PWM_DIV 0x32
+#define  SOR_PWM_DIV_MASK			0xffffff
+
+#define SOR_PWM_CTL 0x33
+#define  SOR_PWM_CTL_TRIGGER			(1 << 31)
+#define  SOR_PWM_CTL_CLK_SEL			(1 << 30)
+#define  SOR_PWM_CTL_DUTY_CYCLE_MASK		0xffffff
+
+#define SOR_VCRC_A0 0x34
+#define SOR_VCRC_A1 0x35
+#define SOR_VCRC_B0 0x36
+#define SOR_VCRC_B1 0x37
+#define SOR_CCRC_A0 0x38
+#define SOR_CCRC_A1 0x39
+#define SOR_CCRC_B0 0x3a
+#define SOR_CCRC_B1 0x3b
+#define SOR_EDATA_A0 0x3c
+#define SOR_EDATA_A1 0x3d
+#define SOR_EDATA_B0 0x3e
+#define SOR_EDATA_B1 0x3f
+#define SOR_COUNT_A0 0x40
+#define SOR_COUNT_A1 0x41
+#define SOR_COUNT_B0 0x42
+#define SOR_COUNT_B1 0x43
+#define SOR_DEBUG_A0 0x44
+#define SOR_DEBUG_A1 0x45
+#define SOR_DEBUG_B0 0x46
+#define SOR_DEBUG_B1 0x47
+#define SOR_TRIG 0x48
+#define SOR_MSCHECK 0x49
+#define SOR_XBAR_CTRL 0x4a
+#define  SOR_XBAR_CTRL_LINK1_XSEL(channel, value) ((((value) & 0x7) << ((channel) * 3)) << 17)
+#define  SOR_XBAR_CTRL_LINK0_XSEL(channel, value) ((((value) & 0x7) << ((channel) * 3)) <<  2)
+#define  SOR_XBAR_CTRL_LINK_SWAP (1 << 1)
+#define  SOR_XBAR_CTRL_BYPASS (1 << 0)
+#define SOR_XBAR_POL 0x4b
+
+#define SOR_DP_LINKCTL0 0x4c
+#define  SOR_DP_LINKCTL_LANE_COUNT_MASK		(0x1f << 16)
+#define  SOR_DP_LINKCTL_LANE_COUNT(x)		(((1 << (x)) - 1) << 16)
+#define  SOR_DP_LINKCTL_ENHANCED_FRAME		(1 << 14)
+#define  SOR_DP_LINKCTL_TU_SIZE_MASK		(0x7f << 2)
+#define  SOR_DP_LINKCTL_TU_SIZE(x)		(((x) & 0x7f) << 2)
+#define  SOR_DP_LINKCTL_ENABLE			(1 << 0)
+
+#define SOR_DP_LINKCTL1 0x4d
+
+#define SOR_LANE_DRIVE_CURRENT0 0x4e
+#define SOR_LANE_DRIVE_CURRENT1 0x4f
+#define SOR_LANE4_DRIVE_CURRENT0 0x50
+#define SOR_LANE4_DRIVE_CURRENT1 0x51
+#define  SOR_LANE_DRIVE_CURRENT_LANE3(x) (((x) & 0xff) << 24)
+#define  SOR_LANE_DRIVE_CURRENT_LANE2(x) (((x) & 0xff) << 16)
+#define  SOR_LANE_DRIVE_CURRENT_LANE1(x) (((x) & 0xff) << 8)
+#define  SOR_LANE_DRIVE_CURRENT_LANE0(x) (((x) & 0xff) << 0)
+
+#define SOR_LANE_PREEMPHASIS0 0x52
+#define SOR_LANE_PREEMPHASIS1 0x53
+#define SOR_LANE4_PREEMPHASIS0 0x54
+#define SOR_LANE4_PREEMPHASIS1 0x55
+#define  SOR_LANE_PREEMPHASIS_LANE3(x) (((x) & 0xff) << 24)
+#define  SOR_LANE_PREEMPHASIS_LANE2(x) (((x) & 0xff) << 16)
+#define  SOR_LANE_PREEMPHASIS_LANE1(x) (((x) & 0xff) << 8)
+#define  SOR_LANE_PREEMPHASIS_LANE0(x) (((x) & 0xff) << 0)
+
+#define SOR_LANE_POSTCURSOR0 0x56
+#define SOR_LANE_POSTCURSOR1 0x57
+#define  SOR_LANE_POSTCURSOR_LANE3(x) (((x) & 0xff) << 24)
+#define  SOR_LANE_POSTCURSOR_LANE2(x) (((x) & 0xff) << 16)
+#define  SOR_LANE_POSTCURSOR_LANE1(x) (((x) & 0xff) << 8)
+#define  SOR_LANE_POSTCURSOR_LANE0(x) (((x) & 0xff) << 0)
+
+#define SOR_DP_CONFIG0 0x58
+#define SOR_DP_CONFIG_DISPARITY_NEGATIVE	(1 << 31)
+#define SOR_DP_CONFIG_ACTIVE_SYM_ENABLE		(1 << 26)
+#define SOR_DP_CONFIG_ACTIVE_SYM_POLARITY	(1 << 24)
+#define SOR_DP_CONFIG_ACTIVE_SYM_FRAC_MASK	(0xf << 16)
+#define SOR_DP_CONFIG_ACTIVE_SYM_FRAC(x)	(((x) & 0xf) << 16)
+#define SOR_DP_CONFIG_ACTIVE_SYM_COUNT_MASK	(0x7f << 8)
+#define SOR_DP_CONFIG_ACTIVE_SYM_COUNT(x)	(((x) & 0x7f) << 8)
+#define SOR_DP_CONFIG_WATERMARK_MASK	(0x3f << 0)
+#define SOR_DP_CONFIG_WATERMARK(x)	(((x) & 0x3f) << 0)
+
+#define SOR_DP_CONFIG1 0x59
+#define SOR_DP_MN0 0x5a
+#define SOR_DP_MN1 0x5b
+
+#define SOR_DP_PADCTL0 0x5c
+#define  SOR_DP_PADCTL_PAD_CAL_PD	(1 << 23)
+#define  SOR_DP_PADCTL_TX_PU_ENABLE	(1 << 22)
+#define  SOR_DP_PADCTL_TX_PU_MASK	(0xff << 8)
+#define  SOR_DP_PADCTL_TX_PU(x)		(((x) & 0xff) << 8)
+#define  SOR_DP_PADCTL_CM_TXD_3		(1 << 7)
+#define  SOR_DP_PADCTL_CM_TXD_2		(1 << 6)
+#define  SOR_DP_PADCTL_CM_TXD_1		(1 << 5)
+#define  SOR_DP_PADCTL_CM_TXD_0		(1 << 4)
+#define  SOR_DP_PADCTL_CM_TXD(x)	(1 << (4 + (x)))
+#define  SOR_DP_PADCTL_PD_TXD_3		(1 << 3)
+#define  SOR_DP_PADCTL_PD_TXD_0		(1 << 2)
+#define  SOR_DP_PADCTL_PD_TXD_1		(1 << 1)
+#define  SOR_DP_PADCTL_PD_TXD_2		(1 << 0)
+#define  SOR_DP_PADCTL_PD_TXD(x)	(1 << (0 + (x)))
+
+#define SOR_DP_PADCTL1 0x5d
+
+#define SOR_DP_DEBUG0 0x5e
+#define SOR_DP_DEBUG1 0x5f
+
+#define SOR_DP_SPARE0 0x60
+#define  SOR_DP_SPARE_DISP_VIDEO_PREAMBLE	(1 << 3)
+#define  SOR_DP_SPARE_MACRO_SOR_CLK		(1 << 2)
+#define  SOR_DP_SPARE_PANEL_INTERNAL		(1 << 1)
+#define  SOR_DP_SPARE_SEQ_ENABLE		(1 << 0)
+
+#define SOR_DP_SPARE1 0x61
+#define SOR_DP_AUDIO_CTRL 0x62
+
+#define SOR_DP_AUDIO_HBLANK_SYMBOLS 0x63
+#define SOR_DP_AUDIO_HBLANK_SYMBOLS_MASK (0x01ffff << 0)
+
+#define SOR_DP_AUDIO_VBLANK_SYMBOLS 0x64
+#define SOR_DP_AUDIO_VBLANK_SYMBOLS_MASK (0x1fffff << 0)
+
+#define SOR_DP_GENERIC_INFOFRAME_HEADER 0x65
+#define SOR_DP_GENERIC_INFOFRAME_SUBPACK0 0x66
+#define SOR_DP_GENERIC_INFOFRAME_SUBPACK1 0x67
+#define SOR_DP_GENERIC_INFOFRAME_SUBPACK2 0x68
+#define SOR_DP_GENERIC_INFOFRAME_SUBPACK3 0x69
+#define SOR_DP_GENERIC_INFOFRAME_SUBPACK4 0x6a
+#define SOR_DP_GENERIC_INFOFRAME_SUBPACK5 0x6b
+#define SOR_DP_GENERIC_INFOFRAME_SUBPACK6 0x6c
+
+#define SOR_DP_TPG 0x6d
+#define  SOR_DP_TPG_CHANNEL_CODING	(1 << 6)
+#define  SOR_DP_TPG_SCRAMBLER_MASK	(3 << 4)
+#define  SOR_DP_TPG_SCRAMBLER_FIBONACCI	(2 << 4)
+#define  SOR_DP_TPG_SCRAMBLER_GALIOS	(1 << 4)
+#define  SOR_DP_TPG_SCRAMBLER_NONE	(0 << 4)
+#define  SOR_DP_TPG_PATTERN_MASK	(0xf << 0)
+#define  SOR_DP_TPG_PATTERN_HBR2	(0x8 << 0)
+#define  SOR_DP_TPG_PATTERN_CSTM	(0x7 << 0)
+#define  SOR_DP_TPG_PATTERN_PRBS7	(0x6 << 0)
+#define  SOR_DP_TPG_PATTERN_SBLERRRATE	(0x5 << 0)
+#define  SOR_DP_TPG_PATTERN_D102	(0x4 << 0)
+#define  SOR_DP_TPG_PATTERN_TRAIN3	(0x3 << 0)
+#define  SOR_DP_TPG_PATTERN_TRAIN2	(0x2 << 0)
+#define  SOR_DP_TPG_PATTERN_TRAIN1	(0x1 << 0)
+#define  SOR_DP_TPG_PATTERN_NONE	(0x0 << 0)
+
+#define SOR_DP_TPG_CONFIG 0x6e
+#define SOR_DP_LQ_CSTM0 0x6f
+#define SOR_DP_LQ_CSTM1 0x70
+#define SOR_DP_LQ_CSTM2 0x71
+
+#define SOR_DP_PADCTL2 0x73
+#define  SOR_DP_PADCTL_SPAREPLL_MASK (0xff << 24)
+#define  SOR_DP_PADCTL_SPAREPLL(x) (((x) & 0xff) << 24)
+
+#define SOR_HDMI_AUDIO_INFOFRAME_CTRL 0x9a
+#define SOR_HDMI_AUDIO_INFOFRAME_STATUS 0x9b
+#define SOR_HDMI_AUDIO_INFOFRAME_HEADER 0x9c
+
+#define SOR_HDMI_AVI_INFOFRAME_CTRL 0x9f
+#define  INFOFRAME_CTRL_CHECKSUM_ENABLE	(1 << 9)
+#define  INFOFRAME_CTRL_SINGLE		(1 << 8)
+#define  INFOFRAME_CTRL_OTHER		(1 << 4)
+#define  INFOFRAME_CTRL_ENABLE		(1 << 0)
+
+#define SOR_HDMI_AVI_INFOFRAME_STATUS 0xa0
+#define  INFOFRAME_STATUS_DONE		(1 << 0)
+
+#define SOR_HDMI_AVI_INFOFRAME_HEADER 0xa1
+#define  INFOFRAME_HEADER_LEN(x) (((x) & 0xff) << 16)
+#define  INFOFRAME_HEADER_VERSION(x) (((x) & 0xff) << 8)
+#define  INFOFRAME_HEADER_TYPE(x) (((x) & 0xff) << 0)
+
+#define SOR_HDMI_ACR_CTRL 0xb1
+
+#define SOR_HDMI_ACR_0320_SUBPACK_LOW 0xb2
+#define  SOR_HDMI_ACR_SUBPACK_LOW_SB1(x) (((x) & 0xff) << 24)
+
+#define SOR_HDMI_ACR_0320_SUBPACK_HIGH 0xb3
+#define  SOR_HDMI_ACR_SUBPACK_HIGH_ENABLE (1 << 31)
+
+#define SOR_HDMI_ACR_0441_SUBPACK_LOW 0xb4
+#define SOR_HDMI_ACR_0441_SUBPACK_HIGH 0xb5
+
+#define SOR_HDMI_CTRL 0xc0
+#define  SOR_HDMI_CTRL_ENABLE (1 << 30)
+#define  SOR_HDMI_CTRL_MAX_AC_PACKET(x) (((x) & 0x1f) << 16)
+#define  SOR_HDMI_CTRL_AUDIO_LAYOUT (1 << 10)
+#define  SOR_HDMI_CTRL_REKEY(x) (((x) & 0x7f) << 0)
+
+#define SOR_HDMI_SPARE 0xcb
+#define  SOR_HDMI_SPARE_ACR_PRIORITY_HIGH (1 << 31)
+#define  SOR_HDMI_SPARE_CTS_RESET(x) (((x) & 0x7) << 16)
+#define  SOR_HDMI_SPARE_HW_CTS_ENABLE (1 << 0)
+
+#define SOR_REFCLK 0xe6
+#define  SOR_REFCLK_DIV_INT(x) ((((x) >> 2) & 0xff) << 8)
+#define  SOR_REFCLK_DIV_FRAC(x) (((x) & 0x3) << 6)
+
+#define SOR_INPUT_CONTROL 0xe8
+#define  SOR_INPUT_CONTROL_ARM_VIDEO_RANGE_LIMITED (1 << 1)
+#define  SOR_INPUT_CONTROL_HDMI_SRC_SELECT(x) (((x) & 0x1) << 0)
+
+#define SOR_AUDIO_CNTRL 0xfc
+#define  SOR_AUDIO_CNTRL_INJECT_NULLSMPL (1 << 29)
+#define  SOR_AUDIO_CNTRL_SOURCE_SELECT(x) (((x) & 0x3) << 20)
+#define   SOURCE_SELECT_MASK 0x3
+#define   SOURCE_SELECT_HDA 0x2
+#define   SOURCE_SELECT_SPDIF 0x1
+#define   SOURCE_SELECT_AUTO 0x0
+#define  SOR_AUDIO_CNTRL_AFIFO_FLUSH (1 << 12)
+
+#define SOR_AUDIO_SPARE 0xfe
+#define  SOR_AUDIO_SPARE_HBR_ENABLE (1 << 27)
+
+#define SOR_AUDIO_NVAL_0320 0xff
+#define SOR_AUDIO_NVAL_0441 0x100
+#define SOR_AUDIO_NVAL_0882 0x101
+#define SOR_AUDIO_NVAL_1764 0x102
+#define SOR_AUDIO_NVAL_0480 0x103
+#define SOR_AUDIO_NVAL_0960 0x104
+#define SOR_AUDIO_NVAL_1920 0x105
+
+#define SOR_AUDIO_HDA_CODEC_SCRATCH0 0x10a
+#define  SOR_AUDIO_HDA_CODEC_SCRATCH0_VALID (1 << 30)
+#define  SOR_AUDIO_HDA_CODEC_SCRATCH0_FMT_MASK 0xffff
+
+#define SOR_AUDIO_HDA_ELD_BUFWR 0x10c
+#define  SOR_AUDIO_HDA_ELD_BUFWR_INDEX(x) (((x) & 0xff) << 8)
+#define  SOR_AUDIO_HDA_ELD_BUFWR_DATA(x) (((x) & 0xff) << 0)
+
+#define SOR_AUDIO_HDA_PRESENSE 0x10d
+#define  SOR_AUDIO_HDA_PRESENSE_ELDV (1 << 1)
+#define  SOR_AUDIO_HDA_PRESENSE_PD (1 << 0)
+
+#define SOR_AUDIO_AVAL_0320 0x10f
+#define SOR_AUDIO_AVAL_0441 0x110
+#define SOR_AUDIO_AVAL_0882 0x111
+#define SOR_AUDIO_AVAL_1764 0x112
+#define SOR_AUDIO_AVAL_0480 0x113
+#define SOR_AUDIO_AVAL_0960 0x114
+#define SOR_AUDIO_AVAL_1920 0x115
+
+#define SOR_INT_STATUS 0x11c
+#define  SOR_INT_CODEC_CP_REQUEST (1 << 2)
+#define  SOR_INT_CODEC_SCRATCH1 (1 << 1)
+#define  SOR_INT_CODEC_SCRATCH0 (1 << 0)
+
+#define SOR_INT_MASK 0x11d
+#define SOR_INT_ENABLE 0x11e
+
+#define SOR_HDMI_VSI_INFOFRAME_CTRL 0x123
+#define SOR_HDMI_VSI_INFOFRAME_STATUS 0x124
+#define SOR_HDMI_VSI_INFOFRAME_HEADER 0x125
+
+#define SOR_HDMI_AUDIO_N 0x13c
+#define SOR_HDMI_AUDIO_N_LOOKUP (1 << 28)
+#define SOR_HDMI_AUDIO_N_RESET (1 << 20)
+
+#define SOR_HDMI2_CTRL 0x13e
+#define  SOR_HDMI2_CTRL_CLOCK_MODE_DIV_BY_4 (1 << 1)
+#define  SOR_HDMI2_CTRL_SCRAMBLE (1 << 0)
+
+#endif
diff --git a/drivers/gpu/drm/tegra/submit.c b/drivers/gpu/drm/tegra/submit.c
new file mode 100644
index 000000000..e5841857c
--- /dev/null
+++ b/drivers/gpu/drm/tegra/submit.c
@@ -0,0 +1,683 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (c) 2020 NVIDIA Corporation */
+
+#include <linux/dma-fence-array.h>
+#include <linux/dma-mapping.h>
+#include <linux/file.h>
+#include <linux/host1x.h>
+#include <linux/iommu.h>
+#include <linux/kref.h>
+#include <linux/list.h>
+#include <linux/nospec.h>
+#include <linux/pm_runtime.h>
+#include <linux/scatterlist.h>
+#include <linux/slab.h>
+#include <linux/sync_file.h>
+
+#include <drm/drm_drv.h>
+#include <drm/drm_file.h>
+#include <drm/drm_syncobj.h>
+
+#include "drm.h"
+#include "gem.h"
+#include "submit.h"
+#include "uapi.h"
+
+#define SUBMIT_ERR(context, fmt, ...) \
+	dev_err_ratelimited(context->client->base.dev, \
+		"%s: job submission failed: " fmt "\n", \
+		current->comm, ##__VA_ARGS__)
+
+struct gather_bo {
+	struct host1x_bo base;
+
+	struct kref ref;
+
+	struct device *dev;
+	u32 *gather_data;
+	dma_addr_t gather_data_dma;
+	size_t gather_data_words;
+};
+
+static struct host1x_bo *gather_bo_get(struct host1x_bo *host_bo)
+{
+	struct gather_bo *bo = container_of(host_bo, struct gather_bo, base);
+
+	kref_get(&bo->ref);
+
+	return host_bo;
+}
+
+static void gather_bo_release(struct kref *ref)
+{
+	struct gather_bo *bo = container_of(ref, struct gather_bo, ref);
+
+	dma_free_attrs(bo->dev, bo->gather_data_words * 4, bo->gather_data, bo->gather_data_dma,
+		       0);
+	kfree(bo);
+}
+
+static void gather_bo_put(struct host1x_bo *host_bo)
+{
+	struct gather_bo *bo = container_of(host_bo, struct gather_bo, base);
+
+	kref_put(&bo->ref, gather_bo_release);
+}
+
+static struct host1x_bo_mapping *
+gather_bo_pin(struct device *dev, struct host1x_bo *bo, enum dma_data_direction direction)
+{
+	struct gather_bo *gather = container_of(bo, struct gather_bo, base);
+	struct host1x_bo_mapping *map;
+	int err;
+
+	map = kzalloc_obj(*map);
+	if (!map)
+		return ERR_PTR(-ENOMEM);
+
+	kref_init(&map->ref);
+	map->bo = bo;
+	map->direction = direction;
+	map->dev = dev;
+
+	map->sgt = kzalloc_obj(*map->sgt);
+	if (!map->sgt) {
+		err = -ENOMEM;
+		goto free;
+	}
+
+	err = dma_get_sgtable(gather->dev, map->sgt, gather->gather_data, gather->gather_data_dma,
+			      gather->gather_data_words * 4);
+	if (err)
+		goto free_sgt;
+
+	err = dma_map_sgtable(dev, map->sgt, direction, 0);
+	if (err)
+		goto free_sgt;
+
+	map->phys = sg_dma_address(map->sgt->sgl);
+	map->size = gather->gather_data_words * 4;
+	map->chunks = err;
+
+	return map;
+
+free_sgt:
+	sg_free_table(map->sgt);
+	kfree(map->sgt);
+free:
+	kfree(map);
+	return ERR_PTR(err);
+}
+
+static void gather_bo_unpin(struct host1x_bo_mapping *map)
+{
+	if (!map)
+		return;
+
+	dma_unmap_sgtable(map->dev, map->sgt, map->direction, 0);
+	sg_free_table(map->sgt);
+	kfree(map->sgt);
+
+	kfree(map);
+}
+
+static void *gather_bo_mmap(struct host1x_bo *host_bo)
+{
+	struct gather_bo *bo = container_of(host_bo, struct gather_bo, base);
+
+	return bo->gather_data;
+}
+
+static void gather_bo_munmap(struct host1x_bo *host_bo, void *addr)
+{
+}
+
+static const struct host1x_bo_ops gather_bo_ops = {
+	.get = gather_bo_get,
+	.put = gather_bo_put,
+	.pin = gather_bo_pin,
+	.unpin = gather_bo_unpin,
+	.mmap = gather_bo_mmap,
+	.munmap = gather_bo_munmap,
+};
+
+static struct tegra_drm_mapping *
+tegra_drm_mapping_get(struct tegra_drm_context *context, u32 id)
+{
+	struct tegra_drm_mapping *mapping;
+
+	xa_lock(&context->mappings);
+
+	mapping = xa_load(&context->mappings, id);
+	if (mapping)
+		kref_get(&mapping->ref);
+
+	xa_unlock(&context->mappings);
+
+	return mapping;
+}
+
+static void *alloc_copy_user_array(void __user *from, size_t count, size_t size)
+{
+	size_t copy_len;
+	void *data;
+
+	if (check_mul_overflow(count, size, &copy_len))
+		return ERR_PTR(-EINVAL);
+
+	if (copy_len > 0x4000)
+		return ERR_PTR(-E2BIG);
+
+	data = vmemdup_user(from, copy_len);
+	if (IS_ERR(data))
+		return ERR_CAST(data);
+
+	return data;
+}
+
+static int submit_copy_gather_data(struct gather_bo **pbo, struct device *dev,
+				   struct tegra_drm_context *context,
+				   struct drm_tegra_channel_submit *args)
+{
+	struct gather_bo *bo;
+	size_t copy_len;
+
+	if (args->gather_data_words == 0) {
+		SUBMIT_ERR(context, "gather_data_words cannot be zero");
+		return -EINVAL;
+	}
+
+	if (check_mul_overflow((size_t)args->gather_data_words, (size_t)4, &copy_len)) {
+		SUBMIT_ERR(context, "gather_data_words is too large");
+		return -EINVAL;
+	}
+
+	bo = kzalloc_obj(*bo);
+	if (!bo) {
+		SUBMIT_ERR(context, "failed to allocate memory for bo info");
+		return -ENOMEM;
+	}
+
+	host1x_bo_init(&bo->base, &gather_bo_ops);
+	kref_init(&bo->ref);
+	bo->dev = dev;
+
+	bo->gather_data = dma_alloc_attrs(dev, copy_len, &bo->gather_data_dma,
+					  GFP_KERNEL | __GFP_NOWARN, 0);
+	if (!bo->gather_data) {
+		SUBMIT_ERR(context, "failed to allocate memory for gather data");
+		kfree(bo);
+		return -ENOMEM;
+	}
+
+	if (copy_from_user(bo->gather_data, u64_to_user_ptr(args->gather_data_ptr), copy_len)) {
+		SUBMIT_ERR(context, "failed to copy gather data from userspace");
+		dma_free_attrs(dev, copy_len, bo->gather_data, bo->gather_data_dma, 0);
+		kfree(bo);
+		return -EFAULT;
+	}
+
+	bo->gather_data_words = args->gather_data_words;
+
+	*pbo = bo;
+
+	return 0;
+}
+
+static int submit_write_reloc(struct tegra_drm_context *context, struct gather_bo *bo,
+			      struct drm_tegra_submit_buf *buf, struct tegra_drm_mapping *mapping)
+{
+	/* TODO check that target_offset is within bounds */
+	dma_addr_t iova = mapping->iova + buf->reloc.target_offset;
+	u32 written_ptr;
+
+#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
+	if (buf->flags & DRM_TEGRA_SUBMIT_RELOC_SECTOR_LAYOUT)
+		iova |= BIT_ULL(39);
+#endif
+
+	written_ptr = iova >> buf->reloc.shift;
+
+	if (buf->reloc.gather_offset_words >= bo->gather_data_words) {
+		SUBMIT_ERR(context,
+			   "relocation has too large gather offset (%u vs gather length %zu)",
+			   buf->reloc.gather_offset_words, bo->gather_data_words);
+		return -EINVAL;
+	}
+
+	buf->reloc.gather_offset_words = array_index_nospec(buf->reloc.gather_offset_words,
+							    bo->gather_data_words);
+
+	bo->gather_data[buf->reloc.gather_offset_words] = written_ptr;
+
+	return 0;
+}
+
+static int submit_process_bufs(struct tegra_drm_context *context, struct gather_bo *bo,
+			       struct drm_tegra_channel_submit *args,
+			       struct tegra_drm_submit_data *job_data)
+{
+	struct tegra_drm_used_mapping *mappings;
+	struct drm_tegra_submit_buf *bufs;
+	int err;
+	u32 i;
+
+	bufs = alloc_copy_user_array(u64_to_user_ptr(args->bufs_ptr), args->num_bufs,
+				     sizeof(*bufs));
+	if (IS_ERR(bufs)) {
+		SUBMIT_ERR(context, "failed to copy bufs array from userspace");
+		return PTR_ERR(bufs);
+	}
+
+	mappings = kzalloc_objs(*mappings, args->num_bufs);
+	if (!mappings) {
+		SUBMIT_ERR(context, "failed to allocate memory for mapping info");
+		err = -ENOMEM;
+		goto done;
+	}
+
+	for (i = 0; i < args->num_bufs; i++) {
+		struct drm_tegra_submit_buf *buf = &bufs[i];
+		struct tegra_drm_mapping *mapping;
+
+		if (buf->flags & ~DRM_TEGRA_SUBMIT_RELOC_SECTOR_LAYOUT) {
+			SUBMIT_ERR(context, "invalid flag specified for buffer");
+			err = -EINVAL;
+			goto drop_refs;
+		}
+
+		mapping = tegra_drm_mapping_get(context, buf->mapping);
+		if (!mapping) {
+			SUBMIT_ERR(context, "invalid mapping ID '%u' for buffer", buf->mapping);
+			err = -EINVAL;
+			goto drop_refs;
+		}
+
+		err = submit_write_reloc(context, bo, buf, mapping);
+		if (err) {
+			tegra_drm_mapping_put(mapping);
+			goto drop_refs;
+		}
+
+		mappings[i].mapping = mapping;
+		mappings[i].flags = buf->flags;
+	}
+
+	job_data->used_mappings = mappings;
+	job_data->num_used_mappings = i;
+
+	err = 0;
+
+	goto done;
+
+drop_refs:
+	while (i--)
+		tegra_drm_mapping_put(mappings[i].mapping);
+
+	kfree(mappings);
+	job_data->used_mappings = NULL;
+
+done:
+	kvfree(bufs);
+
+	return err;
+}
+
+static int submit_get_syncpt(struct tegra_drm_context *context, struct host1x_job *job,
+			     struct xarray *syncpoints, struct drm_tegra_channel_submit *args)
+{
+	struct host1x_syncpt *sp;
+
+	if (args->syncpt.flags) {
+		SUBMIT_ERR(context, "invalid flag specified for syncpt");
+		return -EINVAL;
+	}
+
+	/* Syncpt ref will be dropped on job release */
+	sp = xa_load(syncpoints, args->syncpt.id);
+	if (!sp) {
+		SUBMIT_ERR(context, "syncpoint specified in syncpt was not allocated");
+		return -EINVAL;
+	}
+
+	job->syncpt = host1x_syncpt_get(sp);
+	job->syncpt_incrs = args->syncpt.increments;
+
+	return 0;
+}
+
+static int submit_job_add_gather(struct host1x_job *job, struct tegra_drm_context *context,
+				 struct drm_tegra_submit_cmd_gather_uptr *cmd,
+				 struct gather_bo *bo, u32 *offset,
+				 struct tegra_drm_submit_data *job_data,
+				 u32 *class)
+{
+	u32 next_offset;
+
+	if (cmd->reserved[0] || cmd->reserved[1] || cmd->reserved[2]) {
+		SUBMIT_ERR(context, "non-zero reserved field in GATHER_UPTR command");
+		return -EINVAL;
+	}
+
+	/* Check for maximum gather size */
+	if (cmd->words > 16383) {
+		SUBMIT_ERR(context, "too many words in GATHER_UPTR command");
+		return -EINVAL;
+	}
+
+	if (check_add_overflow(*offset, cmd->words, &next_offset)) {
+		SUBMIT_ERR(context, "too many total words in job");
+		return -EINVAL;
+	}
+
+	if (next_offset > bo->gather_data_words) {
+		SUBMIT_ERR(context, "GATHER_UPTR command overflows gather data");
+		return -EINVAL;
+	}
+
+	if (tegra_drm_fw_validate(context->client, bo->gather_data, *offset,
+				  cmd->words, job_data, class)) {
+		SUBMIT_ERR(context, "job was rejected by firewall");
+		return -EINVAL;
+	}
+
+	host1x_job_add_gather(job, &bo->base, cmd->words, *offset * 4);
+
+	*offset = next_offset;
+
+	return 0;
+}
+
+static struct host1x_job *
+submit_create_job(struct tegra_drm_context *context, struct gather_bo *bo,
+		  struct drm_tegra_channel_submit *args, struct tegra_drm_submit_data *job_data,
+		  struct xarray *syncpoints)
+{
+	struct drm_tegra_submit_cmd *cmds;
+	u32 i, gather_offset = 0, class;
+	struct host1x_job *job;
+	int err;
+
+	/* Set initial class for firewall. */
+	class = context->client->base.class;
+
+	cmds = alloc_copy_user_array(u64_to_user_ptr(args->cmds_ptr), args->num_cmds,
+				     sizeof(*cmds));
+	if (IS_ERR(cmds)) {
+		SUBMIT_ERR(context, "failed to copy cmds array from userspace");
+		return ERR_CAST(cmds);
+	}
+
+	job = host1x_job_alloc(context->channel, args->num_cmds, 0, true);
+	if (!job) {
+		SUBMIT_ERR(context, "failed to allocate memory for job");
+		job = ERR_PTR(-ENOMEM);
+		goto done;
+	}
+
+	err = submit_get_syncpt(context, job, syncpoints, args);
+	if (err < 0)
+		goto free_job;
+
+	job->client = &context->client->base;
+	job->class = context->client->base.class;
+	job->serialize = true;
+
+	for (i = 0; i < args->num_cmds; i++) {
+		struct drm_tegra_submit_cmd *cmd = &cmds[i];
+
+		if (cmd->flags) {
+			SUBMIT_ERR(context, "unknown flags given for cmd");
+			err = -EINVAL;
+			goto free_job;
+		}
+
+		if (cmd->type == DRM_TEGRA_SUBMIT_CMD_GATHER_UPTR) {
+			err = submit_job_add_gather(job, context, &cmd->gather_uptr, bo,
+						    &gather_offset, job_data, &class);
+			if (err)
+				goto free_job;
+		} else if (cmd->type == DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT) {
+			if (cmd->wait_syncpt.reserved[0] || cmd->wait_syncpt.reserved[1]) {
+				SUBMIT_ERR(context, "non-zero reserved value");
+				err = -EINVAL;
+				goto free_job;
+			}
+
+			host1x_job_add_wait(job, cmd->wait_syncpt.id, cmd->wait_syncpt.value,
+					    false, class);
+		} else if (cmd->type == DRM_TEGRA_SUBMIT_CMD_WAIT_SYNCPT_RELATIVE) {
+			if (cmd->wait_syncpt.reserved[0] || cmd->wait_syncpt.reserved[1]) {
+				SUBMIT_ERR(context, "non-zero reserved value");
+				err = -EINVAL;
+				goto free_job;
+			}
+
+			if (cmd->wait_syncpt.id != args->syncpt.id) {
+				SUBMIT_ERR(context, "syncpoint ID in CMD_WAIT_SYNCPT_RELATIVE is not used by the job");
+				err = -EINVAL;
+				goto free_job;
+			}
+
+			host1x_job_add_wait(job, cmd->wait_syncpt.id, cmd->wait_syncpt.value,
+					    true, class);
+		} else {
+			SUBMIT_ERR(context, "unknown cmd type");
+			err = -EINVAL;
+			goto free_job;
+		}
+	}
+
+	if (gather_offset == 0) {
+		SUBMIT_ERR(context, "job must have at least one gather");
+		err = -EINVAL;
+		goto free_job;
+	}
+
+	goto done;
+
+free_job:
+	host1x_job_put(job);
+	job = ERR_PTR(err);
+
+done:
+	kvfree(cmds);
+
+	return job;
+}
+
+static void release_job(struct host1x_job *job)
+{
+	struct tegra_drm_client *client = container_of(job->client, struct tegra_drm_client, base);
+	struct tegra_drm_submit_data *job_data = job->user_data;
+	u32 i;
+
+	if (job->memory_context)
+		host1x_memory_context_put(job->memory_context);
+
+	for (i = 0; i < job_data->num_used_mappings; i++)
+		tegra_drm_mapping_put(job_data->used_mappings[i].mapping);
+
+	kfree(job_data->used_mappings);
+	kfree(job_data);
+
+	pm_runtime_mark_last_busy(client->base.dev);
+	pm_runtime_put_autosuspend(client->base.dev);
+}
+
+int tegra_drm_ioctl_channel_submit(struct drm_device *drm, void *data,
+				   struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_channel_submit *args = data;
+	struct tegra_drm_submit_data *job_data;
+	struct drm_syncobj *syncobj = NULL;
+	struct tegra_drm_context *context;
+	struct host1x_job *job;
+	struct gather_bo *bo;
+	u32 i;
+	int err;
+
+	mutex_lock(&fpriv->lock);
+
+	context = xa_load(&fpriv->contexts, args->context);
+	if (!context) {
+		mutex_unlock(&fpriv->lock);
+		pr_err_ratelimited("%s: %s: invalid channel context '%#x'", __func__,
+				   current->comm, args->context);
+		return -EINVAL;
+	}
+
+	if (args->syncobj_in) {
+		struct dma_fence *fence;
+
+		err = drm_syncobj_find_fence(file, args->syncobj_in, 0, 0, &fence);
+		if (err) {
+			SUBMIT_ERR(context, "invalid syncobj_in '%#x'", args->syncobj_in);
+			goto unlock;
+		}
+
+		err = dma_fence_wait_timeout(fence, true, msecs_to_jiffies(10000));
+		dma_fence_put(fence);
+		if (err) {
+			SUBMIT_ERR(context, "wait for syncobj_in timed out");
+			goto unlock;
+		}
+	}
+
+	if (args->syncobj_out) {
+		syncobj = drm_syncobj_find(file, args->syncobj_out);
+		if (!syncobj) {
+			SUBMIT_ERR(context, "invalid syncobj_out '%#x'", args->syncobj_out);
+			err = -ENOENT;
+			goto unlock;
+		}
+	}
+
+	/* Allocate gather BO and copy gather words in. */
+	err = submit_copy_gather_data(&bo, drm->dev, context, args);
+	if (err)
+		goto unlock;
+
+	job_data = kzalloc_obj(*job_data);
+	if (!job_data) {
+		SUBMIT_ERR(context, "failed to allocate memory for job data");
+		err = -ENOMEM;
+		goto put_bo;
+	}
+
+	/* Get data buffer mappings and do relocation patching. */
+	err = submit_process_bufs(context, bo, args, job_data);
+	if (err)
+		goto free_job_data;
+
+	/* Allocate host1x_job and add gathers and waits to it. */
+	job = submit_create_job(context, bo, args, job_data, &fpriv->syncpoints);
+	if (IS_ERR(job)) {
+		err = PTR_ERR(job);
+		goto free_job_data;
+	}
+
+	/* Map gather data for Host1x. */
+	err = host1x_job_pin(job, context->client->base.dev);
+	if (err) {
+		SUBMIT_ERR(context, "failed to pin job: %d", err);
+		goto put_job;
+	}
+
+	if (context->client->ops->get_streamid_offset) {
+		err = context->client->ops->get_streamid_offset(
+			context->client, &job->engine_streamid_offset);
+		if (err) {
+			SUBMIT_ERR(context, "failed to get streamid offset: %d", err);
+			goto unpin_job;
+		}
+	}
+
+	if (context->memory_context && context->client->ops->can_use_memory_ctx) {
+		bool supported;
+
+		err = context->client->ops->can_use_memory_ctx(context->client, &supported);
+		if (err) {
+			SUBMIT_ERR(context, "failed to detect if engine can use memory context: %d", err);
+			goto unpin_job;
+		}
+
+		if (supported) {
+			job->memory_context = context->memory_context;
+			host1x_memory_context_get(job->memory_context);
+		}
+	} else if (context->client->ops->get_streamid_offset) {
+		/*
+		 * Job submission will need to temporarily change stream ID,
+		 * so need to tell it what to change it back to.
+		 */
+		if (!tegra_dev_iommu_get_stream_id(context->client->base.dev,
+						   &job->engine_fallback_streamid))
+			job->engine_fallback_streamid = TEGRA_STREAM_ID_BYPASS;
+	}
+
+	/* Boot engine. */
+	err = pm_runtime_resume_and_get(context->client->base.dev);
+	if (err < 0) {
+		SUBMIT_ERR(context, "could not power up engine: %d", err);
+		goto put_memory_context;
+	}
+
+	job->user_data = job_data;
+	job->release = release_job;
+	job->timeout = 10000;
+
+	/*
+	 * job_data is now part of job reference counting, so don't release
+	 * it from here.
+	 */
+	job_data = NULL;
+
+	/* Submit job to hardware. */
+	err = host1x_job_submit(job);
+	if (err) {
+		SUBMIT_ERR(context, "host1x job submission failed: %d", err);
+		goto unpin_job;
+	}
+
+	/* Return postfences to userspace and add fences to DMA reservations. */
+	args->syncpt.value = job->syncpt_end;
+
+	if (syncobj) {
+		struct dma_fence *fence = host1x_fence_create(job->syncpt, job->syncpt_end, true);
+		if (IS_ERR(fence)) {
+			err = PTR_ERR(fence);
+			SUBMIT_ERR(context, "failed to create postfence: %d", err);
+		}
+
+		drm_syncobj_replace_fence(syncobj, fence);
+	}
+
+	goto put_job;
+
+put_memory_context:
+	if (job->memory_context)
+		host1x_memory_context_put(job->memory_context);
+unpin_job:
+	host1x_job_unpin(job);
+put_job:
+	host1x_job_put(job);
+free_job_data:
+	if (job_data && job_data->used_mappings) {
+		for (i = 0; i < job_data->num_used_mappings; i++)
+			tegra_drm_mapping_put(job_data->used_mappings[i].mapping);
+
+		kfree(job_data->used_mappings);
+	}
+
+	kfree(job_data);
+put_bo:
+	gather_bo_put(&bo->base);
+unlock:
+	if (syncobj)
+		drm_syncobj_put(syncobj);
+
+	mutex_unlock(&fpriv->lock);
+	return err;
+}
diff --git a/drivers/gpu/drm/tegra/submit.h b/drivers/gpu/drm/tegra/submit.h
new file mode 100644
index 000000000..cf6a2f0a2
--- /dev/null
+++ b/drivers/gpu/drm/tegra/submit.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/* Copyright (c) 2020 NVIDIA Corporation */
+
+#ifndef _TEGRA_DRM_UAPI_SUBMIT_H
+#define _TEGRA_DRM_UAPI_SUBMIT_H
+
+struct tegra_drm_used_mapping {
+	struct tegra_drm_mapping *mapping;
+	u32 flags;
+};
+
+struct tegra_drm_submit_data {
+	struct tegra_drm_used_mapping *used_mappings;
+	u32 num_used_mappings;
+};
+
+int tegra_drm_fw_validate(struct tegra_drm_client *client, u32 *data, u32 start,
+			  u32 words, struct tegra_drm_submit_data *submit,
+			  u32 *job_class);
+
+#endif
diff --git a/drivers/gpu/drm/tegra/trace.c b/drivers/gpu/drm/tegra/trace.c
new file mode 100644
index 000000000..006f65c72
--- /dev/null
+++ b/drivers/gpu/drm/tegra/trace.c
@@ -0,0 +1,2 @@
+#define CREATE_TRACE_POINTS
+#include "trace.h"
diff --git a/drivers/gpu/drm/tegra/trace.h b/drivers/gpu/drm/tegra/trace.h
new file mode 100644
index 000000000..5a1ab4046
--- /dev/null
+++ b/drivers/gpu/drm/tegra/trace.h
@@ -0,0 +1,68 @@
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM tegra
+
+#if !defined(DRM_TEGRA_TRACE_H) || defined(TRACE_HEADER_MULTI_READ)
+#define DRM_TEGRA_TRACE_H 1
+
+#include <linux/device.h>
+#include <linux/tracepoint.h>
+
+DECLARE_EVENT_CLASS(register_access,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value),
+	TP_STRUCT__entry(
+		__field(struct device *, dev)
+		__field(unsigned int, offset)
+		__field(u32, value)
+	),
+	TP_fast_assign(
+		__entry->dev = dev;
+		__entry->offset = offset;
+		__entry->value = value;
+	),
+	TP_printk("%s %04x %08x", dev_name(__entry->dev), __entry->offset,
+		  __entry->value)
+);
+
+DEFINE_EVENT(register_access, dc_writel,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+DEFINE_EVENT(register_access, dc_readl,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+
+DEFINE_EVENT(register_access, hdmi_writel,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+DEFINE_EVENT(register_access, hdmi_readl,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+
+DEFINE_EVENT(register_access, dsi_writel,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+DEFINE_EVENT(register_access, dsi_readl,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+
+DEFINE_EVENT(register_access, dpaux_writel,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+DEFINE_EVENT(register_access, dpaux_readl,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+
+DEFINE_EVENT(register_access, sor_writel,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+DEFINE_EVENT(register_access, sor_readl,
+	TP_PROTO(struct device *dev, unsigned int offset, u32 value),
+	TP_ARGS(dev, offset, value));
+
+#endif /* DRM_TEGRA_TRACE_H */
+
+/* This part must be outside protection */
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH ../../drivers/gpu/drm/tegra
+#define TRACE_INCLUDE_FILE trace
+#include <trace/define_trace.h>
diff --git a/drivers/gpu/drm/tegra/uapi.c b/drivers/gpu/drm/tegra/uapi.c
new file mode 100644
index 000000000..c0ac6b45f
--- /dev/null
+++ b/drivers/gpu/drm/tegra/uapi.c
@@ -0,0 +1,362 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Copyright (c) 2020 NVIDIA Corporation */
+
+#include <linux/host1x.h>
+#include <linux/iommu.h>
+#include <linux/list.h>
+
+#include <drm/drm_drv.h>
+#include <drm/drm_file.h>
+#include <drm/drm_utils.h>
+
+#include "drm.h"
+#include "uapi.h"
+
+static void tegra_drm_mapping_release(struct kref *ref)
+{
+	struct tegra_drm_mapping *mapping =
+		container_of(ref, struct tegra_drm_mapping, ref);
+
+	host1x_bo_unpin(mapping->map);
+	host1x_bo_put(mapping->bo);
+
+	kfree(mapping);
+}
+
+void tegra_drm_mapping_put(struct tegra_drm_mapping *mapping)
+{
+	kref_put(&mapping->ref, tegra_drm_mapping_release);
+}
+
+static void tegra_drm_channel_context_close(struct tegra_drm_context *context)
+{
+	struct tegra_drm_mapping *mapping;
+	unsigned long id;
+
+	if (context->memory_context)
+		host1x_memory_context_put(context->memory_context);
+
+	xa_for_each(&context->mappings, id, mapping)
+		tegra_drm_mapping_put(mapping);
+
+	xa_destroy(&context->mappings);
+
+	host1x_channel_put(context->channel);
+
+	kfree(context);
+}
+
+void tegra_drm_uapi_close_file(struct tegra_drm_file *file)
+{
+	struct tegra_drm_context *context;
+	struct host1x_syncpt *sp;
+	unsigned long id;
+
+	xa_for_each(&file->contexts, id, context)
+		tegra_drm_channel_context_close(context);
+
+	xa_for_each(&file->syncpoints, id, sp)
+		host1x_syncpt_put(sp);
+
+	xa_destroy(&file->contexts);
+	xa_destroy(&file->syncpoints);
+}
+
+static struct tegra_drm_client *tegra_drm_find_client(struct tegra_drm *tegra, u32 class)
+{
+	struct tegra_drm_client *client;
+
+	list_for_each_entry(client, &tegra->clients, list)
+		if (client->base.class == class)
+			return client;
+
+	return NULL;
+}
+
+int tegra_drm_ioctl_channel_open(struct drm_device *drm, void *data, struct drm_file *file)
+{
+	struct host1x *host = tegra_drm_to_host1x(drm->dev_private);
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct tegra_drm *tegra = drm->dev_private;
+	struct drm_tegra_channel_open *args = data;
+	struct tegra_drm_client *client = NULL;
+	struct tegra_drm_context *context;
+	int err;
+
+	if (args->flags)
+		return -EINVAL;
+
+	context = kzalloc_obj(*context);
+	if (!context)
+		return -ENOMEM;
+
+	client = tegra_drm_find_client(tegra, args->host1x_class);
+	if (!client) {
+		err = -ENODEV;
+		goto free;
+	}
+
+	if (client->shared_channel) {
+		context->channel = host1x_channel_get(client->shared_channel);
+	} else {
+		context->channel = host1x_channel_request(&client->base);
+		if (!context->channel) {
+			err = -EBUSY;
+			goto free;
+		}
+	}
+
+	/* Only allocate context if the engine supports context isolation. */
+	if (device_iommu_mapped(client->base.dev) && client->ops->can_use_memory_ctx) {
+		bool supported;
+
+		err = client->ops->can_use_memory_ctx(client, &supported);
+		if (err)
+			goto put_channel;
+
+		if (supported) {
+			struct pid *pid = get_task_pid(current, PIDTYPE_TGID);
+			context->memory_context = host1x_memory_context_alloc(
+				host, client->base.dev, pid);
+			put_pid(pid);
+		}
+
+		if (IS_ERR(context->memory_context)) {
+			if (PTR_ERR(context->memory_context) != -EOPNOTSUPP) {
+				err = PTR_ERR(context->memory_context);
+				goto put_channel;
+			} else {
+				/*
+				 * OK, HW does not support contexts or contexts
+				 * are disabled.
+				 */
+				context->memory_context = NULL;
+			}
+		}
+	}
+
+	err = xa_alloc(&fpriv->contexts, &args->context, context, XA_LIMIT(1, U32_MAX),
+		       GFP_KERNEL);
+	if (err < 0)
+		goto put_memctx;
+
+	context->client = client;
+	xa_init_flags(&context->mappings, XA_FLAGS_ALLOC1);
+
+	args->version = client->version;
+	args->capabilities = 0;
+
+	if (device_get_dma_attr(client->base.dev) == DEV_DMA_COHERENT)
+		args->capabilities |= DRM_TEGRA_CHANNEL_CAP_CACHE_COHERENT;
+
+	return 0;
+
+put_memctx:
+	if (context->memory_context)
+		host1x_memory_context_put(context->memory_context);
+put_channel:
+	host1x_channel_put(context->channel);
+free:
+	kfree(context);
+
+	return err;
+}
+
+int tegra_drm_ioctl_channel_close(struct drm_device *drm, void *data, struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_channel_close *args = data;
+	struct tegra_drm_context *context;
+
+	mutex_lock(&fpriv->lock);
+
+	context = xa_load(&fpriv->contexts, args->context);
+	if (!context) {
+		mutex_unlock(&fpriv->lock);
+		return -EINVAL;
+	}
+
+	xa_erase(&fpriv->contexts, args->context);
+
+	mutex_unlock(&fpriv->lock);
+
+	tegra_drm_channel_context_close(context);
+
+	return 0;
+}
+
+int tegra_drm_ioctl_channel_map(struct drm_device *drm, void *data, struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_channel_map *args = data;
+	struct tegra_drm_mapping *mapping;
+	struct tegra_drm_context *context;
+	enum dma_data_direction direction;
+	struct device *mapping_dev;
+	int err = 0;
+
+	if (args->flags & ~DRM_TEGRA_CHANNEL_MAP_READ_WRITE)
+		return -EINVAL;
+
+	mutex_lock(&fpriv->lock);
+
+	context = xa_load(&fpriv->contexts, args->context);
+	if (!context) {
+		mutex_unlock(&fpriv->lock);
+		return -EINVAL;
+	}
+
+	mapping = kzalloc_obj(*mapping);
+	if (!mapping) {
+		err = -ENOMEM;
+		goto unlock;
+	}
+
+	kref_init(&mapping->ref);
+
+	if (context->memory_context)
+		mapping_dev = &context->memory_context->dev;
+	else
+		mapping_dev = context->client->base.dev;
+
+	mapping->bo = tegra_gem_lookup(file, args->handle);
+	if (!mapping->bo) {
+		err = -EINVAL;
+		goto free;
+	}
+
+	switch (args->flags & DRM_TEGRA_CHANNEL_MAP_READ_WRITE) {
+	case DRM_TEGRA_CHANNEL_MAP_READ_WRITE:
+		direction = DMA_BIDIRECTIONAL;
+		break;
+
+	case DRM_TEGRA_CHANNEL_MAP_WRITE:
+		direction = DMA_FROM_DEVICE;
+		break;
+
+	case DRM_TEGRA_CHANNEL_MAP_READ:
+		direction = DMA_TO_DEVICE;
+		break;
+
+	default:
+		err = -EINVAL;
+		goto put_gem;
+	}
+
+	mapping->map = host1x_bo_pin(mapping_dev, mapping->bo, direction, NULL);
+	if (IS_ERR(mapping->map)) {
+		err = PTR_ERR(mapping->map);
+		goto put_gem;
+	}
+
+	mapping->iova = mapping->map->phys;
+	mapping->iova_end = mapping->iova + host1x_to_tegra_bo(mapping->bo)->gem.size;
+
+	err = xa_alloc(&context->mappings, &args->mapping, mapping, XA_LIMIT(1, U32_MAX),
+		       GFP_KERNEL);
+	if (err < 0)
+		goto unpin;
+
+	mutex_unlock(&fpriv->lock);
+
+	return 0;
+
+unpin:
+	host1x_bo_unpin(mapping->map);
+put_gem:
+	host1x_bo_put(mapping->bo);
+free:
+	kfree(mapping);
+unlock:
+	mutex_unlock(&fpriv->lock);
+	return err;
+}
+
+int tegra_drm_ioctl_channel_unmap(struct drm_device *drm, void *data, struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_channel_unmap *args = data;
+	struct tegra_drm_mapping *mapping;
+	struct tegra_drm_context *context;
+
+	mutex_lock(&fpriv->lock);
+
+	context = xa_load(&fpriv->contexts, args->context);
+	if (!context) {
+		mutex_unlock(&fpriv->lock);
+		return -EINVAL;
+	}
+
+	mapping = xa_erase(&context->mappings, args->mapping);
+
+	mutex_unlock(&fpriv->lock);
+
+	if (!mapping)
+		return -EINVAL;
+
+	tegra_drm_mapping_put(mapping);
+	return 0;
+}
+
+int tegra_drm_ioctl_syncpoint_allocate(struct drm_device *drm, void *data, struct drm_file *file)
+{
+	struct host1x *host1x = tegra_drm_to_host1x(drm->dev_private);
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_syncpoint_allocate *args = data;
+	struct host1x_syncpt *sp;
+	int err;
+
+	if (args->id)
+		return -EINVAL;
+
+	sp = host1x_syncpt_alloc(host1x, HOST1X_SYNCPT_CLIENT_MANAGED, current->comm);
+	if (!sp)
+		return -EBUSY;
+
+	args->id = host1x_syncpt_id(sp);
+
+	err = xa_insert(&fpriv->syncpoints, args->id, sp, GFP_KERNEL);
+	if (err) {
+		host1x_syncpt_put(sp);
+		return err;
+	}
+
+	return 0;
+}
+
+int tegra_drm_ioctl_syncpoint_free(struct drm_device *drm, void *data, struct drm_file *file)
+{
+	struct tegra_drm_file *fpriv = file->driver_priv;
+	struct drm_tegra_syncpoint_allocate *args = data;
+	struct host1x_syncpt *sp;
+
+	mutex_lock(&fpriv->lock);
+	sp = xa_erase(&fpriv->syncpoints, args->id);
+	mutex_unlock(&fpriv->lock);
+
+	if (!sp)
+		return -EINVAL;
+
+	host1x_syncpt_put(sp);
+
+	return 0;
+}
+
+int tegra_drm_ioctl_syncpoint_wait(struct drm_device *drm, void *data, struct drm_file *file)
+{
+	struct host1x *host1x = tegra_drm_to_host1x(drm->dev_private);
+	struct drm_tegra_syncpoint_wait *args = data;
+	signed long timeout_jiffies;
+	struct host1x_syncpt *sp;
+
+	if (args->padding != 0)
+		return -EINVAL;
+
+	sp = host1x_syncpt_get_by_id_noref(host1x, args->id);
+	if (!sp)
+		return -EINVAL;
+
+	timeout_jiffies = drm_timeout_abs_to_jiffies(args->timeout_ns);
+
+	return host1x_syncpt_wait(sp, args->threshold, timeout_jiffies, &args->value);
+}
diff --git a/drivers/gpu/drm/tegra/uapi.h b/drivers/gpu/drm/tegra/uapi.h
new file mode 100644
index 000000000..92ff1e44f
--- /dev/null
+++ b/drivers/gpu/drm/tegra/uapi.h
@@ -0,0 +1,57 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/* Copyright (c) 2020 NVIDIA Corporation */
+
+#ifndef _TEGRA_DRM_UAPI_H
+#define _TEGRA_DRM_UAPI_H
+
+#include <linux/dma-mapping.h>
+#include <linux/idr.h>
+#include <linux/kref.h>
+#include <linux/xarray.h>
+
+#include <drm/drm.h>
+
+struct drm_file;
+struct drm_device;
+
+struct tegra_drm_file {
+	/* Legacy UAPI state */
+	struct idr legacy_contexts;
+	struct mutex lock;
+
+	/* New UAPI state */
+	struct xarray contexts;
+	struct xarray syncpoints;
+};
+
+struct tegra_drm_mapping {
+	struct kref ref;
+
+	struct host1x_bo_mapping *map;
+	struct host1x_bo *bo;
+
+	dma_addr_t iova;
+	dma_addr_t iova_end;
+};
+
+int tegra_drm_ioctl_channel_open(struct drm_device *drm, void *data,
+				 struct drm_file *file);
+int tegra_drm_ioctl_channel_close(struct drm_device *drm, void *data,
+				  struct drm_file *file);
+int tegra_drm_ioctl_channel_map(struct drm_device *drm, void *data,
+				struct drm_file *file);
+int tegra_drm_ioctl_channel_unmap(struct drm_device *drm, void *data,
+				  struct drm_file *file);
+int tegra_drm_ioctl_channel_submit(struct drm_device *drm, void *data,
+				   struct drm_file *file);
+int tegra_drm_ioctl_syncpoint_allocate(struct drm_device *drm, void *data,
+				       struct drm_file *file);
+int tegra_drm_ioctl_syncpoint_free(struct drm_device *drm, void *data,
+				   struct drm_file *file);
+int tegra_drm_ioctl_syncpoint_wait(struct drm_device *drm, void *data,
+				   struct drm_file *file);
+
+void tegra_drm_uapi_close_file(struct tegra_drm_file *file);
+void tegra_drm_mapping_put(struct tegra_drm_mapping *mapping);
+
+#endif
diff --git a/drivers/gpu/drm/tegra/vic.c b/drivers/gpu/drm/tegra/vic.c
new file mode 100644
index 000000000..fe097e967
--- /dev/null
+++ b/drivers/gpu/drm/tegra/vic.c
@@ -0,0 +1,621 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2015, NVIDIA Corporation.
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/dma-mapping.h>
+#include <linux/host1x.h>
+#include <linux/iommu.h>
+#include <linux/iopoll.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/reset.h>
+
+#include <soc/tegra/pmc.h>
+
+#include "drm.h"
+#include "falcon.h"
+#include "vic.h"
+
+#define VIC_FALCON_DEBUGINFO			0x1094
+#define VIC_DEBUGINFO_DUMMY			0xabcd1234
+#define VIC_DEBUGINFO_CLEAR			0x0
+
+struct vic_config {
+	const char *firmware;
+	unsigned int version;
+	bool supports_sid;
+	bool has_riscv;
+	unsigned int transcfg_offset;
+};
+
+struct vic {
+	struct falcon falcon;
+
+	void __iomem *regs;
+	struct tegra_drm_client client;
+	struct host1x_channel *channel;
+	struct device *dev;
+	struct clk *clk;
+	struct reset_control *rst;
+
+	bool can_use_context;
+
+	/* Platform configuration */
+	const struct vic_config *config;
+};
+
+static inline struct vic *to_vic(struct tegra_drm_client *client)
+{
+	return container_of(client, struct vic, client);
+}
+
+static void vic_writel(struct vic *vic, u32 value, unsigned int offset)
+{
+	writel(value, vic->regs + offset);
+}
+
+static int vic_boot(struct vic *vic)
+{
+	u32 stream_id;
+	u32 val;
+	int err = 0;
+
+	if (vic->config->supports_sid && tegra_dev_iommu_get_stream_id(vic->dev, &stream_id)) {
+		u32 value;
+
+		value = TRANSCFG_ATT(1, TRANSCFG_SID_FALCON) |
+			TRANSCFG_ATT(0, TRANSCFG_SID_HW);
+		vic_writel(vic, value, vic->config->transcfg_offset);
+
+		/*
+		 * STREAMID0 is used for input/output buffers. Initialize it to SID_VIC in case
+		 * context isolation is not enabled, and SID_VIC is used for both firmware and
+		 * data buffers.
+		 *
+		 * If context isolation is enabled, it will be overridden by the SETSTREAMID
+		 * opcode as part of each job.
+		 */
+		vic_writel(vic, stream_id, VIC_THI_STREAMID0);
+
+		/* STREAMID1 is used for firmware loading. */
+		vic_writel(vic, stream_id, VIC_THI_STREAMID1);
+	}
+
+	/* setup clockgating registers */
+	vic_writel(vic, CG_IDLE_CG_DLY_CNT(4) |
+			CG_IDLE_CG_EN |
+			CG_WAKEUP_DLY_CNT(4),
+		   NV_PVIC_MISC_PRI_VIC_CG);
+
+	if (vic->config->has_riscv) {
+		/* Write a known pattern into DEBUGINFO register */
+		vic_writel(vic, VIC_DEBUGINFO_DUMMY, VIC_FALCON_DEBUGINFO);
+	}
+
+	err = falcon_boot(&vic->falcon);
+	if (err < 0)
+		return err;
+
+	if (vic->config->has_riscv) {
+		/* Check VIC has reached a proper initialized state */
+		err = readl_poll_timeout(vic->regs + VIC_FALCON_DEBUGINFO, val,
+					 val == VIC_DEBUGINFO_CLEAR,
+					 1000, 2000000);
+		if (err) {
+			dev_err(vic->dev, "VIC not initialized, timeout, val=0x%x\n", val);
+			return err;
+		}
+	} else {
+		u32 fce_ucode_size, fce_bin_data_offset;
+		dma_addr_t iova;
+		void *hdr;
+
+		iova = vic->falcon.firmware.iova;
+		hdr = vic->falcon.firmware.virt;
+		fce_bin_data_offset = *(u32 *)(hdr + VIC_UCODE_FCE_DATA_OFFSET);
+
+		/* Old VIC firmware needs kernel help with setting up FCE microcode. */
+		if (fce_bin_data_offset != 0x0 && fce_bin_data_offset != 0xa5a5a5a5) {
+			hdr = vic->falcon.firmware.virt +
+				*(u32 *)(hdr + VIC_UCODE_FCE_HEADER_OFFSET);
+			fce_ucode_size = *(u32 *)(hdr + FCE_UCODE_SIZE_OFFSET);
+
+			falcon_execute_method(&vic->falcon, VIC_SET_FCE_UCODE_SIZE,
+					      fce_ucode_size);
+			falcon_execute_method(&vic->falcon, VIC_SET_FCE_UCODE_OFFSET,
+					      (iova + fce_bin_data_offset) >> 8);
+		}
+
+		err = falcon_wait_idle(&vic->falcon);
+		if (err < 0) {
+			dev_err(vic->dev,
+				"failed to set application ID and FCE base\n");
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+static int vic_init(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = dev->dev_private;
+	struct vic *vic = to_vic(drm);
+	int err;
+
+	err = host1x_client_iommu_attach(client);
+	if (err < 0 && err != -ENODEV) {
+		dev_err(vic->dev, "failed to attach to domain: %d\n", err);
+		return err;
+	}
+
+	vic->channel = host1x_channel_request(client);
+	if (!vic->channel) {
+		err = -ENOMEM;
+		goto detach;
+	}
+
+	client->syncpts[0] = host1x_syncpt_request(client, 0);
+	if (!client->syncpts[0]) {
+		err = -ENOMEM;
+		goto free_channel;
+	}
+
+	err = tegra_drm_register_client(tegra, drm);
+	if (err < 0)
+		goto free_syncpt;
+
+	/*
+	 * Inherit the DMA parameters (such as maximum segment size) from the
+	 * parent host1x device.
+	 */
+	client->dev->dma_parms = client->host->dma_parms;
+
+	return 0;
+
+free_syncpt:
+	host1x_syncpt_put(client->syncpts[0]);
+free_channel:
+	host1x_channel_put(vic->channel);
+detach:
+	host1x_client_iommu_detach(client);
+
+	return err;
+}
+
+static int vic_exit(struct host1x_client *client)
+{
+	struct tegra_drm_client *drm = host1x_to_drm_client(client);
+	struct drm_device *dev = dev_get_drvdata(client->host);
+	struct tegra_drm *tegra = dev->dev_private;
+	struct vic *vic = to_vic(drm);
+	int err;
+
+	/* avoid a dangling pointer just in case this disappears */
+	client->dev->dma_parms = NULL;
+
+	err = tegra_drm_unregister_client(tegra, drm);
+	if (err < 0)
+		return err;
+
+	pm_runtime_dont_use_autosuspend(client->dev);
+	pm_runtime_force_suspend(client->dev);
+
+	host1x_syncpt_put(client->syncpts[0]);
+	host1x_channel_put(vic->channel);
+	host1x_client_iommu_detach(client);
+
+	vic->channel = NULL;
+
+	if (client->group) {
+		dma_unmap_single(vic->dev, vic->falcon.firmware.phys,
+				 vic->falcon.firmware.size, DMA_TO_DEVICE);
+		tegra_drm_free(tegra, vic->falcon.firmware.size,
+			       vic->falcon.firmware.virt,
+			       vic->falcon.firmware.iova);
+	} else {
+		dma_free_coherent(vic->dev, vic->falcon.firmware.size,
+				  vic->falcon.firmware.virt,
+				  vic->falcon.firmware.iova);
+	}
+
+	return 0;
+}
+
+static const struct host1x_client_ops vic_client_ops = {
+	.init = vic_init,
+	.exit = vic_exit,
+};
+
+static int vic_load_firmware(struct vic *vic)
+{
+	struct host1x_client *client = &vic->client.base;
+	struct tegra_drm *tegra = vic->client.drm;
+	static DEFINE_MUTEX(lock);
+	u32 fce_bin_data_offset;
+	dma_addr_t iova;
+	size_t size;
+	void *virt;
+	int err;
+
+	mutex_lock(&lock);
+
+	if (vic->falcon.firmware.virt) {
+		err = 0;
+		goto unlock;
+	}
+
+	err = falcon_read_firmware(&vic->falcon, vic->config->firmware);
+	if (err < 0)
+		goto unlock;
+
+	size = vic->falcon.firmware.size;
+
+	if (!client->group) {
+		virt = dma_alloc_coherent(vic->dev, size, &iova, GFP_KERNEL);
+		if (!virt) {
+			err = -ENOMEM;
+			goto unlock;
+		}
+	} else {
+		virt = tegra_drm_alloc(tegra, size, &iova);
+		if (IS_ERR(virt)) {
+			err = PTR_ERR(virt);
+			goto unlock;
+		}
+	}
+
+	vic->falcon.firmware.virt = virt;
+	vic->falcon.firmware.iova = iova;
+
+	err = falcon_load_firmware(&vic->falcon);
+	if (err < 0)
+		goto cleanup;
+
+	/*
+	 * In this case we have received an IOVA from the shared domain, so we
+	 * need to make sure to get the physical address so that the DMA API
+	 * knows what memory pages to flush the cache for.
+	 */
+	if (client->group) {
+		dma_addr_t phys;
+
+		phys = dma_map_single(vic->dev, virt, size, DMA_TO_DEVICE);
+
+		err = dma_mapping_error(vic->dev, phys);
+		if (err < 0)
+			goto cleanup;
+
+		vic->falcon.firmware.phys = phys;
+	}
+
+	/*
+	 * Check if firmware is new enough to not require mapping firmware
+	 * to data buffer domains.
+	 */
+	fce_bin_data_offset = *(u32 *)(virt + VIC_UCODE_FCE_DATA_OFFSET);
+
+	if (!vic->config->supports_sid) {
+		vic->can_use_context = false;
+	} else if (vic->config->has_riscv) {
+		vic->can_use_context = true;
+	} else if (fce_bin_data_offset != 0x0 && fce_bin_data_offset != 0xa5a5a5a5) {
+		/*
+		 * Firmware will access FCE through STREAMID0, so context
+		 * isolation cannot be used.
+		 */
+		vic->can_use_context = false;
+		dev_warn_once(vic->dev, "context isolation disabled due to old firmware\n");
+	} else {
+		vic->can_use_context = true;
+	}
+
+unlock:
+	mutex_unlock(&lock);
+	return err;
+
+cleanup:
+	if (!client->group)
+		dma_free_coherent(vic->dev, size, virt, iova);
+	else
+		tegra_drm_free(tegra, size, virt, iova);
+
+	mutex_unlock(&lock);
+	return err;
+}
+
+static int __maybe_unused vic_runtime_resume(struct device *dev)
+{
+	struct vic *vic = dev_get_drvdata(dev);
+	int err;
+
+	err = clk_prepare_enable(vic->clk);
+	if (err < 0)
+		return err;
+
+	usleep_range(10, 20);
+
+	err = reset_control_deassert(vic->rst);
+	if (err < 0)
+		goto disable;
+
+	usleep_range(10, 20);
+
+	err = vic_load_firmware(vic);
+	if (err < 0)
+		goto assert;
+
+	err = vic_boot(vic);
+	if (err < 0)
+		goto assert;
+
+	return 0;
+
+assert:
+	reset_control_assert(vic->rst);
+disable:
+	clk_disable_unprepare(vic->clk);
+	return err;
+}
+
+static int __maybe_unused vic_runtime_suspend(struct device *dev)
+{
+	struct vic *vic = dev_get_drvdata(dev);
+	int err;
+
+	host1x_channel_stop(vic->channel);
+
+	err = reset_control_assert(vic->rst);
+	if (err < 0)
+		return err;
+
+	usleep_range(2000, 4000);
+
+	clk_disable_unprepare(vic->clk);
+
+	return 0;
+}
+
+static int vic_open_channel(struct tegra_drm_client *client,
+			    struct tegra_drm_context *context)
+{
+	struct vic *vic = to_vic(client);
+
+	context->channel = host1x_channel_get(vic->channel);
+	if (!context->channel)
+		return -ENOMEM;
+
+	return 0;
+}
+
+static void vic_close_channel(struct tegra_drm_context *context)
+{
+	host1x_channel_put(context->channel);
+}
+
+static int vic_can_use_memory_ctx(struct tegra_drm_client *client, bool *supported)
+{
+	struct vic *vic = to_vic(client);
+	int err;
+
+	/* This doesn't access HW so it's safe to call without powering up. */
+	err = vic_load_firmware(vic);
+	if (err < 0)
+		return err;
+
+	*supported = vic->can_use_context;
+
+	return 0;
+}
+
+static const struct tegra_drm_client_ops vic_ops = {
+	.open_channel = vic_open_channel,
+	.close_channel = vic_close_channel,
+	.submit = tegra_drm_submit,
+	.get_streamid_offset = tegra_drm_get_streamid_offset_thi,
+	.can_use_memory_ctx = vic_can_use_memory_ctx,
+};
+
+#define NVIDIA_TEGRA_124_VIC_FIRMWARE "nvidia/tegra124/vic03_ucode.bin"
+
+static const struct vic_config vic_t124_config = {
+	.firmware = NVIDIA_TEGRA_124_VIC_FIRMWARE,
+	.version = 0x40,
+	.supports_sid = false,
+};
+
+#define NVIDIA_TEGRA_210_VIC_FIRMWARE "nvidia/tegra210/vic04_ucode.bin"
+
+static const struct vic_config vic_t210_config = {
+	.firmware = NVIDIA_TEGRA_210_VIC_FIRMWARE,
+	.version = 0x21,
+	.supports_sid = false,
+};
+
+#define NVIDIA_TEGRA_186_VIC_FIRMWARE "nvidia/tegra186/vic04_ucode.bin"
+
+static const struct vic_config vic_t186_config = {
+	.firmware = NVIDIA_TEGRA_186_VIC_FIRMWARE,
+	.version = 0x18,
+	.supports_sid = true,
+	.transcfg_offset = 0x2044,
+};
+
+#define NVIDIA_TEGRA_194_VIC_FIRMWARE "nvidia/tegra194/vic.bin"
+
+static const struct vic_config vic_t194_config = {
+	.firmware = NVIDIA_TEGRA_194_VIC_FIRMWARE,
+	.version = 0x19,
+	.supports_sid = true,
+	.transcfg_offset = 0x2044,
+};
+
+#define NVIDIA_TEGRA_234_VIC_FIRMWARE "nvidia/tegra234/vic.bin"
+
+static const struct vic_config vic_t234_config = {
+	.firmware = NVIDIA_TEGRA_234_VIC_FIRMWARE,
+	.version = 0x23,
+	.supports_sid = true,
+	.transcfg_offset = 0x2044,
+};
+
+#define NVIDIA_TEGRA_264_VIC_FIRMWARE "nvidia/tegra264/vic.bin"
+#define NVIDIA_TEGRA_264_VIC_DESC "nvidia/tegra264/vic.bin.desc"
+
+static const struct vic_config vic_t264_config = {
+	.firmware = NVIDIA_TEGRA_264_VIC_FIRMWARE,
+	.version = 0x264,
+	.supports_sid = true,
+	.has_riscv = true,
+	.transcfg_offset = 0x2244,
+};
+
+static const struct of_device_id tegra_vic_of_match[] = {
+	{ .compatible = "nvidia,tegra124-vic", .data = &vic_t124_config },
+	{ .compatible = "nvidia,tegra210-vic", .data = &vic_t210_config },
+	{ .compatible = "nvidia,tegra186-vic", .data = &vic_t186_config },
+	{ .compatible = "nvidia,tegra194-vic", .data = &vic_t194_config },
+	{ .compatible = "nvidia,tegra234-vic", .data = &vic_t234_config },
+	{ .compatible = "nvidia,tegra264-vic", .data = &vic_t264_config },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, tegra_vic_of_match);
+
+static int vic_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	struct host1x_syncpt **syncpts;
+	struct vic *vic;
+	int err;
+
+	/* inherit DMA mask from host1x parent */
+	err = dma_coerce_mask_and_coherent(dev, *dev->parent->dma_mask);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set DMA mask: %d\n", err);
+		return err;
+	}
+
+	vic = devm_kzalloc(dev, sizeof(*vic), GFP_KERNEL);
+	if (!vic)
+		return -ENOMEM;
+
+	vic->config = of_device_get_match_data(dev);
+
+	syncpts = devm_kzalloc(dev, sizeof(*syncpts), GFP_KERNEL);
+	if (!syncpts)
+		return -ENOMEM;
+
+	vic->regs = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(vic->regs))
+		return PTR_ERR(vic->regs);
+
+	vic->clk = devm_clk_get(dev, NULL);
+	if (IS_ERR(vic->clk)) {
+		dev_err(&pdev->dev, "failed to get clock\n");
+		return PTR_ERR(vic->clk);
+	}
+
+	err = clk_set_rate(vic->clk, ULONG_MAX);
+	if (err < 0) {
+		dev_err(&pdev->dev, "failed to set clock rate\n");
+		return err;
+	}
+
+	if (!dev->pm_domain) {
+		vic->rst = devm_reset_control_get(dev, "vic");
+		if (IS_ERR(vic->rst)) {
+			dev_err(&pdev->dev, "failed to get reset\n");
+			return PTR_ERR(vic->rst);
+		}
+	}
+
+	vic->falcon.dev = dev;
+	vic->falcon.regs = vic->regs;
+	vic->falcon.riscv = vic->config->has_riscv;
+
+	err = falcon_init(&vic->falcon);
+	if (err < 0)
+		return err;
+
+	platform_set_drvdata(pdev, vic);
+
+	INIT_LIST_HEAD(&vic->client.base.list);
+	vic->client.base.ops = &vic_client_ops;
+	vic->client.base.dev = dev;
+	vic->client.base.class = HOST1X_CLASS_VIC;
+	vic->client.base.syncpts = syncpts;
+	vic->client.base.num_syncpts = 1;
+	vic->dev = dev;
+
+	INIT_LIST_HEAD(&vic->client.list);
+	vic->client.version = vic->config->version;
+	vic->client.ops = &vic_ops;
+
+	err = host1x_client_register(&vic->client.base);
+	if (err < 0) {
+		dev_err(dev, "failed to register host1x client: %d\n", err);
+		goto exit_falcon;
+	}
+
+	pm_runtime_enable(dev);
+	pm_runtime_use_autosuspend(dev);
+	pm_runtime_set_autosuspend_delay(dev, 500);
+
+	return 0;
+
+exit_falcon:
+	falcon_exit(&vic->falcon);
+
+	return err;
+}
+
+static void vic_remove(struct platform_device *pdev)
+{
+	struct vic *vic = platform_get_drvdata(pdev);
+
+	pm_runtime_disable(&pdev->dev);
+	host1x_client_unregister(&vic->client.base);
+	falcon_exit(&vic->falcon);
+}
+
+static const struct dev_pm_ops vic_pm_ops = {
+	RUNTIME_PM_OPS(vic_runtime_suspend, vic_runtime_resume, NULL)
+	SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, pm_runtime_force_resume)
+};
+
+struct platform_driver tegra_vic_driver = {
+	.driver = {
+		.name = "tegra-vic",
+		.of_match_table = tegra_vic_of_match,
+		.pm = &vic_pm_ops
+	},
+	.probe = vic_probe,
+	.remove = vic_remove,
+};
+
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_124_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_124_VIC_FIRMWARE);
+#endif
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_210_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_210_VIC_FIRMWARE);
+#endif
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_186_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_186_VIC_FIRMWARE);
+#endif
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_194_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_194_VIC_FIRMWARE);
+#endif
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_234_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_234_VIC_FIRMWARE);
+#endif
+#if IS_ENABLED(CONFIG_ARCH_TEGRA_264_SOC)
+MODULE_FIRMWARE(NVIDIA_TEGRA_264_VIC_FIRMWARE);
+MODULE_FIRMWARE(NVIDIA_TEGRA_264_VIC_DESC);
+#endif
diff --git a/drivers/gpu/drm/tegra/vic.h b/drivers/gpu/drm/tegra/vic.h
new file mode 100644
index 000000000..aca98a09c
--- /dev/null
+++ b/drivers/gpu/drm/tegra/vic.h
@@ -0,0 +1,35 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright (c) 2015, NVIDIA Corporation.
+ */
+
+#ifndef TEGRA_VIC_H
+#define TEGRA_VIC_H
+
+/* VIC methods */
+
+#define VIC_SET_FCE_UCODE_SIZE			0x0000071C
+#define VIC_SET_FCE_UCODE_OFFSET		0x0000072C
+
+/* VIC registers */
+
+#define VIC_THI_STREAMID0	0x00000030
+#define VIC_THI_STREAMID1	0x00000034
+
+#define NV_PVIC_MISC_PRI_VIC_CG			0x000016d0
+#define CG_IDLE_CG_DLY_CNT(val)			((val & 0x3f) << 0)
+#define CG_IDLE_CG_EN				(1 << 6)
+#define CG_WAKEUP_DLY_CNT(val)			((val & 0xf) << 16)
+
+#define TRANSCFG_ATT(i, v)	(((v) & 0x3) << ((i) * 4))
+#define TRANSCFG_SID_HW		0
+#define TRANSCFG_SID_PHY	1
+#define TRANSCFG_SID_FALCON	2
+
+/* Firmware offsets */
+
+#define VIC_UCODE_FCE_HEADER_OFFSET		(6*4)
+#define VIC_UCODE_FCE_DATA_OFFSET		(7*4)
+#define FCE_UCODE_SIZE_OFFSET			(2*4)
+
+#endif /* TEGRA_VIC_H */