Merge tag 'drm-for-v4.9' of git://people.freedesktop.org/~airlied/linux

Pull drm updates from Dave Airlie:
 "Core:
   - Fence destaging work
   - DRIVER_LEGACY to split off legacy drm drivers
   - drm_mm refactoring
   - Splitting drm_crtc.c into chunks and documenting better
   - Display info fixes
   - rbtree support for prime buffer lookup
   - Simple VGA DAC driver

  Panel:
   - Add Nexus 7 panel
   - More simple panels

  i915:
   - Refactoring GEM naming
   - Refactored vma/active tracking
   - Lockless request lookups
   - Better stolen memory support
   - FBC fixes
   - SKL watermark fixes
   - VGPU improvements
   - dma-buf fencing support
   - Better DP dongle support

  amdgpu:
   - Powerplay for Iceland asics
   - Improved GPU reset support
   - UVD/VEC powergating support for CZ/ST
   - Preinitialised VRAM buffer support
   - Virtual display support
   - Initial SI support
   - GTT rework
   - PCI shutdown callback support
   - HPD IRQ storm fixes

  amdkfd:
   - bugfixes

  tilcdc:
   - Atomic modesetting support

  mediatek:
   - AAL + GAMMA engine support
   - Hook up gamma LUT
   - Temporal dithering support

  imx:
   - Pixel clock from devicetree
   - drm bridge support for LVDS bridges
   - active plane reconfiguration
   - VDIC deinterlacer support
   - Frame synchronisation unit support
   - Color space conversion support

  analogix:
   - PSR support
   - Better panel on/off support

  rockchip:
   - rk3399 vop/crtc support
   - PSR support

  vc4:
   - Interlaced vblank timing
   - 3D rendering CPU overhead reduction
   - HDMI output fixes

  tda998x:
   - HDMI audio ASoC support

  sunxi:
   - Allwinner A33 support
   - better TCON support

  msm:
   - DT binding cleanups
   - Explicit fence-fd support

  sti:
   - remove sti415/416 support

  etnaviv:
   - MMUv2 refactoring
   - GC3000 support

  exynos:
   - Refactoring HDMI DCC/PHY
   - G2D pm regression fix
   - Page fault issues with wait for vblank

  There is no nouveau work in this tree, as Ben didn't get a pull
  request in, and he was fighting moving to atomic and adding mst
  support, so maybe best it waits for a cycle"

* tag 'drm-for-v4.9' of git://people.freedesktop.org/~airlied/linux: (1412 commits)
  drm/crtc: constify drm_crtc_index parameter
  drm/i915: Fix conflict resolution from backmerge of v4.8-rc8 to drm-next
  drm/i915/guc: Unwind GuC workqueue reservation if request construction fails
  drm/i915: Reset the breadcrumbs IRQ more carefully
  drm/i915: Force relocations via cpu if we run out of idle aperture
  drm/i915: Distinguish last emitted request from last submitted request
  drm/i915: Allow DP to work w/o EDID
  drm/i915: Move long hpd handling into the hotplug work
  drm/i915/execlists: Reinitialise context image after GPU hang
  drm/i915: Use correct index for backtracking HUNG semaphores
  drm/i915: Unalias obj->phys_handle and obj->userptr
  drm/i915: Just clear the mmiodebug before a register access
  drm/i915/gen9: only add the planes actually affected by ddb changes
  drm/i915: Allow PCH DPLL sharing regardless of DPLL_SDVO_HIGH_SPEED
  drm/i915/bxt: Fix HDMI DPLL configuration
  drm/i915/gen9: fix the watermark res_blocks value
  drm/i915/gen9: fix plane_blocks_per_line on watermarks calculations
  drm/i915/gen9: minimum scanlines for Y tile is not always 4
  drm/i915/gen9: fix the WaWmMemoryReadLatency implementation
  drm/i915/kbl: KBL also needs to run the SAGV code
  ...

Documentation/devicetree/bindings/display/bridge/dumb-vga-dac.txt

@@ -0,0 +1,48 @@
+Dumb RGB to VGA DAC bridge
+---------------------------
+
+This binding is aimed for dumb RGB to VGA DAC based bridges that do not require
+any configuration.
+
+Required properties:
+
+- compatible: Must be "dumb-vga-dac"
+
+Required nodes:
+
+This device has two video ports. Their connections are modelled using the OF
+graph bindings specified in Documentation/devicetree/bindings/graph.txt.
+
+- Video port 0 for RGB input
+- Video port 1 for VGA output
+
+
+Example
+-------
+
+bridge {
+	compatible = "dumb-vga-dac";
+	#address-cells = <1>;
+	#size-cells = <0>;
+
+	ports {
+		#address-cells = <1>;
+		#size-cells = <0>;
+
+		port@0 {
+			reg = <0>;
+
+			vga_bridge_in: endpoint {
+				remote-endpoint = <&tcon0_out_vga>;
+			};
+		};
+
+		port@1 {
+			reg = <1>;
+
+			vga_bridge_out: endpoint {
+				remote-endpoint = <&vga_con_in>;
+			};
+		};
+	};
+};

Documentation/devicetree/bindings/display/bridge/tda998x.txt

@@ -21,8 +21,19 @@ Optional properties:
   - video-ports: 24 bits value which defines how the video controller
 	output is wired to the TDA998x input - default: <0x230145>
 
+  - audio-ports: array of 8-bit values, 2 values per one DAI[1].
+	The first value defines the DAI type: TDA998x_SPDIF or TDA998x_I2S[2].
+	The second value defines the tda998x AP_ENA reg content when the DAI
+	in question is used. The implementation allows one or two DAIs. If two
+	DAIs are defined, they must be of different type.
+
+[1] Documentation/sound/alsa/soc/DAI.txt
+[2] include/dt-bindings/display/tda998x.h
+
 Example:
 
+#include <dt-bindings/display/tda998x.h>
+
 	tda998x: hdmi-encoder {
 		compatible = "nxp,tda998x";
 		reg = <0x70>;
@@ -30,4 +41,11 @@ Example:
 		interrupts = <27 2>;		/* falling edge */
 		pinctrl-0 = <&pmx_camera>;
 		pinctrl-names = "default";
+		video-ports = <0x230145>;
+
+		#sound-dai-cells = <2>;
+			     /*	DAI-format	AP_ENA reg value */
+		audio-ports = <	TDA998x_SPDIF	0x04
+				TDA998x_I2S	0x03>;
+
 	};

Documentation/devicetree/bindings/display/msm/hdmi.txt

@@ -14,17 +14,16 @@ Required properties:
 - power-domains: Should be <&mmcc MDSS_GDSC>.
 - clocks: device clocks
   See ../clocks/clock-bindings.txt for details.
-- qcom,hdmi-tx-ddc-clk-gpio: ddc clk pin
-- qcom,hdmi-tx-ddc-data-gpio: ddc data pin
-- qcom,hdmi-tx-hpd-gpio: hpd pin
 - core-vdda-supply: phandle to supply regulator
 - hdmi-mux-supply: phandle to mux regulator
 - phys: the phandle for the HDMI PHY device
 - phy-names: the name of the corresponding PHY device
 
 Optional properties:
-- qcom,hdmi-tx-mux-en-gpio: hdmi mux enable pin
-- qcom,hdmi-tx-mux-sel-gpio: hdmi mux select pin
+- hpd-gpios: hpd pin
+- qcom,hdmi-tx-mux-en-gpios: hdmi mux enable pin
+- qcom,hdmi-tx-mux-sel-gpios: hdmi mux select pin
+- qcom,hdmi-tx-mux-lpm-gpios: hdmi mux lpm pin
 - power-domains: reference to the power domain(s), if available.
 - pinctrl-names: the pin control state names; should contain "default"
 - pinctrl-0: the default pinctrl state (active)

Documentation/devicetree/bindings/display/panel/innolux,g101ice-l01.txt

@@ -0,0 +1,7 @@
+Innolux Corporation 10.1" G101ICE-L01 WXGA (1280x800) LVDS panel
+
+Required properties:
+- compatible: should be "innolux,g101ice-l01"
+
+This binding is compatible with the simple-panel binding, which is specified
+in simple-panel.txt in this directory.

Documentation/devicetree/bindings/display/panel/jdi,lt070me05000.txt

@@ -0,0 +1,31 @@
+JDI model LT070ME05000 1200x1920 7" DSI Panel
+
+Required properties:
+- compatible: should be "jdi,lt070me05000"
+- vddp-supply: phandle of the regulator that provides the supply voltage
+  Power IC supply (3-5V)
+- iovcc-supply: phandle of the regulator that provides the supply voltage
+  IOVCC , power supply for LCM (1.8V)
+- enable-gpios: phandle of gpio for enable line
+  LED_EN, LED backlight enable, High active
+- reset-gpios: phandle of gpio for reset line
+  This should be 8mA, gpio can be configured using mux, pinctrl, pinctrl-names
+  XRES, Reset, Low active
+- dcdc-en-gpios: phandle of the gpio for power ic line
+  Power IC supply enable, High active
+
+Example:
+
+	dsi0: qcom,mdss_dsi@4700000 {
+		panel@0 {
+			compatible = "jdi,lt070me05000";
+			reg = <0>;
+
+			vddp-supply = <&pm8921_l17>;
+			iovcc-supply = <&pm8921_lvs7>;
+
+			enable-gpios = <&pm8921_gpio 36 GPIO_ACTIVE_HIGH>;
+			reset-gpios = <&tlmm_pinmux 54 GPIO_ACTIVE_LOW>;
+			dcdc-en-gpios = <&pm8921_gpio 23 GPIO_ACTIVE_HIGH>;
+		};
+	};

Documentation/devicetree/bindings/display/rockchip/rockchip-vop.txt

@@ -6,8 +6,10 @@ buffer to an external LCD interface.
 
 Required properties:
 - compatible: value should be one of the following
-		"rockchip,rk3288-vop";
 		"rockchip,rk3036-vop";
+		"rockchip,rk3288-vop";
+		"rockchip,rk3399-vop-big";
+		"rockchip,rk3399-vop-lit";
 
 - interrupts: should contain a list of all VOP IP block interrupts in the
 		 order: VSYNC, LCD_SYSTEM. The interrupt specifier

Documentation/devicetree/bindings/display/sunxi/sun4i-drm.txt

@@ -26,13 +26,14 @@ TCON
 The TCON acts as a timing controller for RGB, LVDS and TV interfaces.
 
 Required properties:
- - compatible: value should be "allwinner,sun5i-a13-tcon".
+ - compatible: value must be either:
+   * allwinner,sun5i-a13-tcon
+   * allwinner,sun8i-a33-tcon
  - reg: base address and size of memory-mapped region
  - interrupts: interrupt associated to this IP
  - clocks: phandles to the clocks feeding the TCON. Three are needed:
    - 'ahb': the interface clocks
    - 'tcon-ch0': The clock driving the TCON channel 0
-   - 'tcon-ch1': The clock driving the TCON channel 1
  - resets: phandles to the reset controllers driving the encoder
    - "lcd": the reset line for the TCON channel 0
 
@@ -49,6 +50,33 @@ Required properties:
   second the block connected to the TCON channel 1 (usually the TV
   encoder)
 
+On the A13, there is one more clock required:
+   - 'tcon-ch1': The clock driving the TCON channel 1
+
+DRC
+---
+
+The DRC (Dynamic Range Controller), found in the latest Allwinner SoCs
+(A31, A23, A33), allows to dynamically adjust pixel
+brightness/contrast based on histogram measurements for LCD content
+adaptive backlight control.
+
+
+Required properties:
+  - compatible: value must be one of:
+    * allwinner,sun8i-a33-drc
+  - reg: base address and size of the memory-mapped region.
+  - interrupts: interrupt associated to this IP
+  - clocks: phandles to the clocks feeding the DRC
+    * ahb: the DRC interface clock
+    * mod: the DRC module clock
+    * ram: the DRC DRAM clock
+  - clock-names: the clock names mentioned above
+  - resets: phandles to the reset line driving the DRC
+
+- ports: A ports node with endpoint definitions as defined in
+  Documentation/devicetree/bindings/media/video-interfaces.txt. The
+  first port should be the input endpoints, the second one the outputs
 
 Display Engine Backend
 ----------------------
@@ -59,6 +87,7 @@ system.
 Required properties:
   - compatible: value must be one of:
     * allwinner,sun5i-a13-display-backend
+    * allwinner,sun8i-a33-display-backend
   - reg: base address and size of the memory-mapped region.
   - clocks: phandles to the clocks feeding the frontend and backend
     * ahb: the backend interface clock
@@ -71,6 +100,14 @@ Required properties:
   Documentation/devicetree/bindings/media/video-interfaces.txt. The
   first port should be the input endpoints, the second one the output
 
+On the A33, some additional properties are required:
+  - reg needs to have an additional region corresponding to the SAT
+  - reg-names need to be set, with "be" and "sat"
+  - clocks and clock-names need to have a phandle to the SAT bus
+    clocks, whose name will be "sat"
+  - resets and reset-names need to have a phandle to the SAT bus
+    resets, whose name will be "sat"
+
 Display Engine Frontend
 -----------------------
 
@@ -80,6 +117,7 @@ deinterlacing and color space conversion.
 Required properties:
   - compatible: value must be one of:
     * allwinner,sun5i-a13-display-frontend
+    * allwinner,sun8i-a33-display-frontend
   - reg: base address and size of the memory-mapped region.
   - interrupts: interrupt associated to this IP
   - clocks: phandles to the clocks feeding the frontend and backend
@@ -104,6 +142,7 @@ extra node.
 Required properties:
   - compatible: value must be one of:
     * allwinner,sun5i-a13-display-engine
+    * allwinner,sun8i-a33-display-engine
 
   - allwinner,pipelines: list of phandle to the display engine
     frontends available.

Documentation/devicetree/bindings/display/tilcdc/tilcdc.txt

@@ -17,6 +17,18 @@ Optional properties:
    the lcd controller.
  - max-pixelclock: The maximum pixel clock that can be supported
    by the lcd controller in KHz.
+ - blue-and-red-wiring: Recognized values "straight" or "crossed".
+   This property deals with the LCDC revision 2 (found on AM335x)
+   color errata [1].
+    - "straight" indicates normal wiring that supports RGB565,
+      BGR888, and XBGR8888 color formats.
+    - "crossed" indicates wiring that has blue and red wires
+      crossed. This setup supports BGR565, RGB888 and XRGB8888
+      formats.
+    - If the property is not present or its value is not recognized
+      the legacy mode is assumed. This configuration supports RGB565,
+      RGB888 and XRGB8888 formats. However, depending on wiring, the red
+      and blue colors are swapped in either 16 or 24-bit color modes.
 
 Optional nodes:
 
@@ -24,6 +36,18 @@ Optional nodes:
    binding follows Documentation/devicetree/bindings/graph.txt and
    suppors a single port with a single endpoint.
 
+ - See also Documentation/devicetree/bindings/display/tilcdc/panel.txt and
+   Documentation/devicetree/bindings/display/tilcdc/tfp410.txt for connecting
+   tfp410 DVI encoder or lcd panel to lcdc
+
+[1] There is an errata about AM335x color wiring. For 16-bit color mode
+    the wires work as they should (LCD_DATA[0:4] is for Blue[3:7]),
+    but for 24 bit color modes the wiring of blue and red components is
+    crossed and LCD_DATA[0:4] is for Red[3:7] and LCD_DATA[11:15] is
+    for Blue[3-7]. For more details see section 3.1.1 in AM335x
+    Silicon Errata:
+    http://www.ti.com/general/docs/lit/getliterature.tsp?baseLiteratureNumber=sprz360
+
 Example:
 
 	fb: fb@4830e000 {
@@ -33,6 +57,8 @@ Example:
 		interrupts = <36>;
 		ti,hwmods = "lcdc";
 
+		blue-and-red-wiring = "crossed";
+
 		port {
 			lcdc_0: endpoint@0 {
 				remote-endpoint = <&hdmi_0>;

Documentation/gpu/drm-internals.rst

@@ -53,9 +53,12 @@ u32 driver_features;
 DRIVER_USE_AGP
     Driver uses AGP interface, the DRM core will manage AGP resources.
 
-DRIVER_REQUIRE_AGP
-    Driver needs AGP interface to function. AGP initialization failure
-    will become a fatal error.
+DRIVER_LEGACY
+    Denote a legacy driver using shadow attach. Don't use.
+
+DRIVER_KMS_LEGACY_CONTEXT
+    Used only by nouveau for backwards compatibility with existing userspace.
+    Don't use.
 
 DRIVER_PCI_DMA
     Driver is capable of PCI DMA, mapping of PCI DMA buffers to

Documentation/gpu/drm-kms-helpers.rst

@@ -2,38 +2,45 @@
 Mode Setting Helper Functions
 =============================
 
-The plane, CRTC, encoder and connector functions provided by the drivers
-implement the DRM API. They're called by the DRM core and ioctl handlers
-to handle device state changes and configuration request. As
-implementing those functions often requires logic not specific to
-drivers, mid-layer helper functions are available to avoid duplicating
-boilerplate code.
-
-The DRM core contains one mid-layer implementation. The mid-layer
-provides implementations of several plane, CRTC, encoder and connector
-functions (called from the top of the mid-layer) that pre-process
-requests and call lower-level functions provided by the driver (at the
-bottom of the mid-layer). For instance, the
-:c:func:`drm_crtc_helper_set_config()` function can be used to
-fill the :c:type:`struct drm_crtc_funcs <drm_crtc_funcs>`
-set_config field. When called, it will split the set_config operation
-in smaller, simpler operations and call the driver to handle them.
-
-To use the mid-layer, drivers call
-:c:func:`drm_crtc_helper_add()`,
-:c:func:`drm_encoder_helper_add()` and
-:c:func:`drm_connector_helper_add()` functions to install their
-mid-layer bottom operations handlers, and fill the :c:type:`struct
-drm_crtc_funcs <drm_crtc_funcs>`, :c:type:`struct
-drm_encoder_funcs <drm_encoder_funcs>` and :c:type:`struct
-drm_connector_funcs <drm_connector_funcs>` structures with
-pointers to the mid-layer top API functions. Installing the mid-layer
-bottom operation handlers is best done right after registering the
-corresponding KMS object.
-
-The mid-layer is not split between CRTC, encoder and connector
-operations. To use it, a driver must provide bottom functions for all of
-the three KMS entities.
+The DRM subsystem aims for a strong separation between core code and helper
+libraries. Core code takes care of general setup and teardown and decoding
+userspace requests to kernel internal objects. Everything else is handled by a
+large set of helper libraries, which can be combined freely to pick and choose
+for each driver what fits, and avoid shared code where special behaviour is
+needed.
+
+This distinction between core code and helpers is especially strong in the
+modesetting code, where there's a shared userspace ABI for all drivers. This is
+in contrast to the render side, where pretty much everything (with very few
+exceptions) can be considered optional helper code.
+
+There are a few areas these helpers can grouped into:
+
+* Helpers to implement modesetting. The important ones here are the atomic
+  helpers. Old drivers still often use the legacy CRTC helpers. They both share
+  the same set of common helper vtables. For really simple drivers (anything
+  that would have been a great fit in the deprecated fbdev subsystem) there's
+  also the simple display pipe helpers.
+
+* There's a big pile of helpers for handling outputs. First the generic bridge
+  helpers for handling encoder and transcoder IP blocks. Second the panel helpers
+  for handling panel-related information and logic. Plus then a big set of
+  helpers for the various sink standards (DisplayPort, HDMI, MIPI DSI). Finally
+  there's also generic helpers for handling output probing, and for dealing with
+  EDIDs.
+
+* The last group of helpers concerns itself with the frontend side of a display
+  pipeline: Planes, handling rectangles for visibility checking and scissoring,
+  flip queues and assorted bits.
+
+Modeset Helper Reference for Common Vtables
+===========================================
+
+.. kernel-doc:: include/drm/drm_modeset_helper_vtables.h
+   :internal:
+
+.. kernel-doc:: include/drm/drm_modeset_helper_vtables.h
+   :doc: overview
 
 Atomic Modeset Helper Functions Reference
 =========================================
@@ -62,33 +69,27 @@ Atomic State Reset and Initialization
 .. kernel-doc:: drivers/gpu/drm/drm_atomic_helper.c
    :export:
 
-Modeset Helper Reference for Common Vtables
-===========================================
-
-.. kernel-doc:: include/drm/drm_modeset_helper_vtables.h
-   :internal:
-
-.. kernel-doc:: include/drm/drm_modeset_helper_vtables.h
-   :doc: overview
-
 Legacy CRTC/Modeset Helper Functions Reference
 ==============================================
 
 .. kernel-doc:: drivers/gpu/drm/drm_crtc_helper.c
-   :export:
+   :doc: overview
 
 .. kernel-doc:: drivers/gpu/drm/drm_crtc_helper.c
-   :doc: overview
+   :export:
 
-Output Probing Helper Functions Reference
-=========================================
+Simple KMS Helper Reference
+===========================
 
-.. kernel-doc:: drivers/gpu/drm/drm_probe_helper.c
-   :doc: output probing helper overview
+.. kernel-doc:: include/drm/drm_simple_kms_helper.h
+   :internal:
 
-.. kernel-doc:: drivers/gpu/drm/drm_probe_helper.c
+.. kernel-doc:: drivers/gpu/drm/drm_simple_kms_helper.c
    :export:
 
+.. kernel-doc:: drivers/gpu/drm/drm_simple_kms_helper.c
+   :doc: overview
+
 fbdev Helper Functions Reference
 ================================
 
@@ -110,6 +111,43 @@ Framebuffer CMA Helper Functions Reference
 .. kernel-doc:: drivers/gpu/drm/drm_fb_cma_helper.c
    :export:
 
+Bridges
+=======
+
+Overview
+--------
+
+.. kernel-doc:: drivers/gpu/drm/drm_bridge.c
+   :doc: overview
+
+Default bridge callback sequence
+--------------------------------
+
+.. kernel-doc:: drivers/gpu/drm/drm_bridge.c
+   :doc: bridge callbacks
+
+
+Bridge Helper Reference
+-------------------------
+
+.. kernel-doc:: include/drm/drm_bridge.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_bridge.c
+   :export:
+
+Panel Helper Reference
+======================
+
+.. kernel-doc:: include/drm/drm_panel.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_panel.c
+   :export:
+
+.. kernel-doc:: drivers/gpu/drm/drm_panel.c
+   :doc: drm panel
+
 Display Port Helper Functions Reference
 =======================================
 
@@ -158,9 +196,21 @@ MIPI DSI Helper Functions Reference
 .. kernel-doc:: drivers/gpu/drm/drm_mipi_dsi.c
    :export:
 
+Output Probing Helper Functions Reference
+=========================================
+
+.. kernel-doc:: drivers/gpu/drm/drm_probe_helper.c
+   :doc: output probing helper overview
+
+.. kernel-doc:: drivers/gpu/drm/drm_probe_helper.c
+   :export:
+
 EDID Helper Functions Reference
 ===============================
 
+.. kernel-doc:: include/drm/drm_edid.h
+   :internal:
+
 .. kernel-doc:: drivers/gpu/drm/drm_edid.c
    :export:
 
@@ -176,18 +226,6 @@ Rectangle Utilities Reference
 .. kernel-doc:: drivers/gpu/drm/drm_rect.c
    :export:
 
-Flip-work Helper Reference
-==========================
-
-.. kernel-doc:: include/drm/drm_flip_work.h
-   :doc: flip utils
-
-.. kernel-doc:: include/drm/drm_flip_work.h
-   :internal:
-
-.. kernel-doc:: drivers/gpu/drm/drm_flip_work.c
-   :export:
-
 HDMI Infoframes Helper Reference
 ================================
 
@@ -202,59 +240,40 @@ libraries and hence is also included here.
 .. kernel-doc:: drivers/video/hdmi.c
    :export:
 
-Plane Helper Reference
-======================
-
-.. kernel-doc:: drivers/gpu/drm/drm_plane_helper.c
-   :export:
-
-.. kernel-doc:: drivers/gpu/drm/drm_plane_helper.c
-   :doc: overview
+Flip-work Helper Reference
+==========================
 
-Tile group
-----------
+.. kernel-doc:: include/drm/drm_flip_work.h
+   :doc: flip utils
 
-.. kernel-doc:: drivers/gpu/drm/drm_crtc.c
-   :doc: Tile group
+.. kernel-doc:: include/drm/drm_flip_work.h
+   :internal:
 
-Bridges
-=======
+.. kernel-doc:: drivers/gpu/drm/drm_flip_work.c
+   :export:
 
-Overview
---------
+Plane Helper Reference
+======================
 
-.. kernel-doc:: drivers/gpu/drm/drm_bridge.c
+.. kernel-doc:: drivers/gpu/drm/drm_plane_helper.c
    :doc: overview
 
-Default bridge callback sequence
---------------------------------
-
-.. kernel-doc:: drivers/gpu/drm/drm_bridge.c
-   :doc: bridge callbacks
-
-.. kernel-doc:: drivers/gpu/drm/drm_bridge.c
+.. kernel-doc:: drivers/gpu/drm/drm_plane_helper.c
    :export:
 
-Panel Helper Reference
-======================
-
-.. kernel-doc:: include/drm/drm_panel.h
-   :internal:
+Tile group
+==========
 
-.. kernel-doc:: drivers/gpu/drm/drm_panel.c
-   :export:
+# FIXME: This should probably be moved into a property documentation section
 
-.. kernel-doc:: drivers/gpu/drm/drm_panel.c
-   :doc: drm panel
+.. kernel-doc:: drivers/gpu/drm/drm_crtc.c
+   :doc: Tile group
 
-Simple KMS Helper Reference
-===========================
+Auxiliary Modeset Helpers
+=========================
 
-.. kernel-doc:: include/drm/drm_simple_kms_helper.h
-   :internal:
+.. kernel-doc:: drivers/gpu/drm/drm_modeset_helper.c
+   :doc: aux kms helpers
 
-.. kernel-doc:: drivers/gpu/drm/drm_simple_kms_helper.c
+.. kernel-doc:: drivers/gpu/drm/drm_modeset_helper.c
    :export:
-
-.. kernel-doc:: drivers/gpu/drm/drm_simple_kms_helper.c
-   :doc: overview

Documentation/gpu/drm-kms.rst

@@ -2,9 +2,6 @@
 Kernel Mode Setting (KMS)
 =========================
 
-Mode Setting
-============
-
 Drivers must initialize the mode setting core by calling
 :c:func:`drm_mode_config_init()` on the DRM device. The function
 initializes the :c:type:`struct drm_device <drm_device>`
@@ -18,60 +15,59 @@ be setup by initializing the following fields.
 -  struct drm_mode_config_funcs \*funcs;
    Mode setting functions.
 
-Display Modes Function Reference
---------------------------------
+Modeset Base Object Abstraction
+===============================
 
-.. kernel-doc:: include/drm/drm_modes.h
+.. kernel-doc:: include/drm/drm_mode_object.h
    :internal:
 
-.. kernel-doc:: drivers/gpu/drm/drm_modes.c
+.. kernel-doc:: drivers/gpu/drm/drm_mode_object.c
+   :export:
+
+KMS Data Structures
+===================
+
+.. kernel-doc:: include/drm/drm_crtc.h
+   :internal:
+
+KMS API Functions
+=================
+
+.. kernel-doc:: drivers/gpu/drm/drm_crtc.c
    :export:
 
 Atomic Mode Setting Function Reference
---------------------------------------
+======================================
 
 .. kernel-doc:: drivers/gpu/drm/drm_atomic.c
    :export:
 
-.. kernel-doc:: drivers/gpu/drm/drm_atomic.c
+.. kernel-doc:: include/drm/drm_atomic.h
    :internal:
 
 Frame Buffer Abstraction
-------------------------
-
-Frame buffers are abstract memory objects that provide a source of
-pixels to scanout to a CRTC. Applications explicitly request the
-creation of frame buffers through the DRM_IOCTL_MODE_ADDFB(2) ioctls
-and receive an opaque handle that can be passed to the KMS CRTC control,
-plane configuration and page flip functions.
-
-Frame buffers rely on the underneath memory manager for low-level memory
-operations. When creating a frame buffer applications pass a memory
-handle (or a list of memory handles for multi-planar formats) through
-the ``drm_mode_fb_cmd2`` argument. For drivers using GEM as their
-userspace buffer management interface this would be a GEM handle.
-Drivers are however free to use their own backing storage object
-handles, e.g. vmwgfx directly exposes special TTM handles to userspace
-and so expects TTM handles in the create ioctl and not GEM handles.
-
-The lifetime of a drm framebuffer is controlled with a reference count,
-drivers can grab additional references with
-:c:func:`drm_framebuffer_reference()`and drop them again with
-:c:func:`drm_framebuffer_unreference()`. For driver-private
-framebuffers for which the last reference is never dropped (e.g. for the
-fbdev framebuffer when the struct :c:type:`struct drm_framebuffer
-<drm_framebuffer>` is embedded into the fbdev helper struct)
-drivers can manually clean up a framebuffer at module unload time with
-:c:func:`drm_framebuffer_unregister_private()`.
+========================
+
+.. kernel-doc:: drivers/gpu/drm/drm_framebuffer.c
+   :doc: overview
+
+Frame Buffer Functions Reference
+--------------------------------
+
+.. kernel-doc:: drivers/gpu/drm/drm_framebuffer.c
+   :export:
+
+.. kernel-doc:: include/drm/drm_framebuffer.h
+   :internal:
 
 DRM Format Handling
--------------------
+===================
 
 .. kernel-doc:: drivers/gpu/drm/drm_fourcc.c
    :export:
 
 Dumb Buffer Objects
--------------------
+===================
 
 The KMS API doesn't standardize backing storage object creation and
 leaves it to driver-specific ioctls. Furthermore actually creating a
@@ -114,14 +110,59 @@ Note that dumb objects may not be used for gpu acceleration, as has been
 attempted on some ARM embedded platforms. Such drivers really must have
 a hardware-specific ioctl to allocate suitable buffer objects.
 
-Output Polling
---------------
+Plane Abstraction
+=================
+
+.. kernel-doc:: drivers/gpu/drm/drm_plane.c
+   :doc: overview
+
+Plane Functions Reference
+-------------------------
+
+.. kernel-doc:: include/drm/drm_plane.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_plane.c
+   :export:
+
+Display Modes Function Reference
+================================
+
+.. kernel-doc:: include/drm/drm_modes.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_modes.c
+   :export:
 
-void (\*output_poll_changed)(struct drm_device \*dev);
-This operation notifies the driver that the status of one or more
-connectors has changed. Drivers that use the fb helper can just call the
-:c:func:`drm_fb_helper_hotplug_event()` function to handle this
-operation.
+Connector Abstraction
+=====================
+
+.. kernel-doc:: drivers/gpu/drm/drm_connector.c
+   :doc: overview
+
+Connector Functions Reference
+-----------------------------
+
+.. kernel-doc:: include/drm/drm_connector.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_connector.c
+   :export:
+
+Encoder Abstraction
+===================
+
+.. kernel-doc:: drivers/gpu/drm/drm_encoder.c
+   :doc: overview
+
+Encoder Functions Reference
+---------------------------
+
+.. kernel-doc:: include/drm/drm_encoder.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_encoder.c
+   :export:
 
 KMS Initialization and Cleanup
 ==============================
@@ -151,250 +192,6 @@ allocated and zeroed by the driver, possibly as part of a larger
 structure, and registered with a call to :c:func:`drm_crtc_init()`
 with a pointer to CRTC functions.
 
-Planes (:c:type:`struct drm_plane <drm_plane>`)
------------------------------------------------
-
-A plane represents an image source that can be blended with or overlayed
-on top of a CRTC during the scanout process. Planes are associated with
-a frame buffer to crop a portion of the image memory (source) and
-optionally scale it to a destination size. The result is then blended
-with or overlayed on top of a CRTC.
-
-The DRM core recognizes three types of planes:
-
--  DRM_PLANE_TYPE_PRIMARY represents a "main" plane for a CRTC.
-   Primary planes are the planes operated upon by CRTC modesetting and
-   flipping operations described in the page_flip hook in
-   :c:type:`struct drm_crtc_funcs <drm_crtc_funcs>`.
--  DRM_PLANE_TYPE_CURSOR represents a "cursor" plane for a CRTC.
-   Cursor planes are the planes operated upon by the
-   DRM_IOCTL_MODE_CURSOR and DRM_IOCTL_MODE_CURSOR2 ioctls.
--  DRM_PLANE_TYPE_OVERLAY represents all non-primary, non-cursor
-   planes. Some drivers refer to these types of planes as "sprites"
-   internally.
-
-For compatibility with legacy userspace, only overlay planes are made
-available to userspace by default. Userspace clients may set the
-DRM_CLIENT_CAP_UNIVERSAL_PLANES client capability bit to indicate
-that they wish to receive a universal plane list containing all plane
-types.
-
-Plane Initialization
-~~~~~~~~~~~~~~~~~~~~
-
-To create a plane, a KMS drivers allocates and zeroes an instances of
-:c:type:`struct drm_plane <drm_plane>` (possibly as part of a
-larger structure) and registers it with a call to
-:c:func:`drm_universal_plane_init()`. The function takes a
-bitmask of the CRTCs that can be associated with the plane, a pointer to
-the plane functions, a list of format supported formats, and the type of
-plane (primary, cursor, or overlay) being initialized.
-
-Cursor and overlay planes are optional. All drivers should provide one
-primary plane per CRTC (although this requirement may change in the
-future); drivers that do not wish to provide special handling for
-primary planes may make use of the helper functions described in ? to
-create and register a primary plane with standard capabilities.
-
-Encoders (:c:type:`struct drm_encoder <drm_encoder>`)
------------------------------------------------------
-
-An encoder takes pixel data from a CRTC and converts it to a format
-suitable for any attached connectors. On some devices, it may be
-possible to have a CRTC send data to more than one encoder. In that
-case, both encoders would receive data from the same scanout buffer,
-resulting in a "cloned" display configuration across the connectors
-attached to each encoder.
-
-Encoder Initialization
-~~~~~~~~~~~~~~~~~~~~~~
-
-As for CRTCs, a KMS driver must create, initialize and register at least
-one :c:type:`struct drm_encoder <drm_encoder>` instance. The
-instance is allocated and zeroed by the driver, possibly as part of a
-larger structure.
-
-Drivers must initialize the :c:type:`struct drm_encoder
-<drm_encoder>` possible_crtcs and possible_clones fields before
-registering the encoder. Both fields are bitmasks of respectively the
-CRTCs that the encoder can be connected to, and sibling encoders
-candidate for cloning.
-
-After being initialized, the encoder must be registered with a call to
-:c:func:`drm_encoder_init()`. The function takes a pointer to the
-encoder functions and an encoder type. Supported types are
-
--  DRM_MODE_ENCODER_DAC for VGA and analog on DVI-I/DVI-A
--  DRM_MODE_ENCODER_TMDS for DVI, HDMI and (embedded) DisplayPort
--  DRM_MODE_ENCODER_LVDS for display panels
--  DRM_MODE_ENCODER_TVDAC for TV output (Composite, S-Video,
-   Component, SCART)
--  DRM_MODE_ENCODER_VIRTUAL for virtual machine displays
-
-Encoders must be attached to a CRTC to be used. DRM drivers leave
-encoders unattached at initialization time. Applications (or the fbdev
-compatibility layer when implemented) are responsible for attaching the
-encoders they want to use to a CRTC.
-
-Connectors (:c:type:`struct drm_connector <drm_connector>`)
------------------------------------------------------------
-
-A connector is the final destination for pixel data on a device, and
-usually connects directly to an external display device like a monitor
-or laptop panel. A connector can only be attached to one encoder at a
-time. The connector is also the structure where information about the
-attached display is kept, so it contains fields for display data, EDID
-data, DPMS & connection status, and information about modes supported on
-the attached displays.
-
-Connector Initialization
-~~~~~~~~~~~~~~~~~~~~~~~~
-
-Finally a KMS driver must create, initialize, register and attach at
-least one :c:type:`struct drm_connector <drm_connector>`
-instance. The instance is created as other KMS objects and initialized
-by setting the following fields.
-
-interlace_allowed
-    Whether the connector can handle interlaced modes.
-
-doublescan_allowed
-    Whether the connector can handle doublescan.
-
-display_info
-    Display information is filled from EDID information when a display
-    is detected. For non hot-pluggable displays such as flat panels in
-    embedded systems, the driver should initialize the
-    display_info.width_mm and display_info.height_mm fields with the
-    physical size of the display.
-
-polled
-    Connector polling mode, a combination of
-
-    DRM_CONNECTOR_POLL_HPD
-        The connector generates hotplug events and doesn't need to be
-        periodically polled. The CONNECT and DISCONNECT flags must not
-        be set together with the HPD flag.
-
-    DRM_CONNECTOR_POLL_CONNECT
-        Periodically poll the connector for connection.
-
-    DRM_CONNECTOR_POLL_DISCONNECT
-        Periodically poll the connector for disconnection.
-
-    Set to 0 for connectors that don't support connection status
-    discovery.
-
-The connector is then registered with a call to
-:c:func:`drm_connector_init()` with a pointer to the connector
-functions and a connector type, and exposed through sysfs with a call to
-:c:func:`drm_connector_register()`.
-
-Supported connector types are
-
--  DRM_MODE_CONNECTOR_VGA
--  DRM_MODE_CONNECTOR_DVII
--  DRM_MODE_CONNECTOR_DVID
--  DRM_MODE_CONNECTOR_DVIA
--  DRM_MODE_CONNECTOR_Composite
--  DRM_MODE_CONNECTOR_SVIDEO
--  DRM_MODE_CONNECTOR_LVDS
--  DRM_MODE_CONNECTOR_Component
--  DRM_MODE_CONNECTOR_9PinDIN
--  DRM_MODE_CONNECTOR_DisplayPort
--  DRM_MODE_CONNECTOR_HDMIA
--  DRM_MODE_CONNECTOR_HDMIB
--  DRM_MODE_CONNECTOR_TV
--  DRM_MODE_CONNECTOR_eDP
--  DRM_MODE_CONNECTOR_VIRTUAL
-
-Connectors must be attached to an encoder to be used. For devices that
-map connectors to encoders 1:1, the connector should be attached at
-initialization time with a call to
-:c:func:`drm_mode_connector_attach_encoder()`. The driver must
-also set the :c:type:`struct drm_connector <drm_connector>`
-encoder field to point to the attached encoder.
-
-Finally, drivers must initialize the connectors state change detection
-with a call to :c:func:`drm_kms_helper_poll_init()`. If at least
-one connector is pollable but can't generate hotplug interrupts
-(indicated by the DRM_CONNECTOR_POLL_CONNECT and
-DRM_CONNECTOR_POLL_DISCONNECT connector flags), a delayed work will
-automatically be queued to periodically poll for changes. Connectors
-that can generate hotplug interrupts must be marked with the
-DRM_CONNECTOR_POLL_HPD flag instead, and their interrupt handler must
-call :c:func:`drm_helper_hpd_irq_event()`. The function will
-queue a delayed work to check the state of all connectors, but no
-periodic polling will be done.
-
-Connector Operations
-~~~~~~~~~~~~~~~~~~~~
-
-    **Note**
-
-    Unless otherwise state, all operations are mandatory.
-
-DPMS
-''''
-
-void (\*dpms)(struct drm_connector \*connector, int mode);
-The DPMS operation sets the power state of a connector. The mode
-argument is one of
-
--  DRM_MODE_DPMS_ON
-
--  DRM_MODE_DPMS_STANDBY
-
--  DRM_MODE_DPMS_SUSPEND
-
--  DRM_MODE_DPMS_OFF
-
-In all but DPMS_ON mode the encoder to which the connector is attached
-should put the display in low-power mode by driving its signals
-appropriately. If more than one connector is attached to the encoder
-care should be taken not to change the power state of other displays as
-a side effect. Low-power mode should be propagated to the encoders and
-CRTCs when all related connectors are put in low-power mode.
-
-Modes
-'''''
-
-int (\*fill_modes)(struct drm_connector \*connector, uint32_t
-max_width, uint32_t max_height);
-Fill the mode list with all supported modes for the connector. If the
-``max_width`` and ``max_height`` arguments are non-zero, the
-implementation must ignore all modes wider than ``max_width`` or higher
-than ``max_height``.
-
-The connector must also fill in this operation its display_info
-width_mm and height_mm fields with the connected display physical size
-in millimeters. The fields should be set to 0 if the value isn't known
-or is not applicable (for instance for projector devices).
-
-Connection Status
-'''''''''''''''''
-
-The connection status is updated through polling or hotplug events when
-supported (see ?). The status value is reported to userspace through
-ioctls and must not be used inside the driver, as it only gets
-initialized by a call to :c:func:`drm_mode_getconnector()` from
-userspace.
-
-enum drm_connector_status (\*detect)(struct drm_connector
-\*connector, bool force);
-Check to see if anything is attached to the connector. The ``force``
-parameter is set to false whilst polling or to true when checking the
-connector due to user request. ``force`` can be used by the driver to
-avoid expensive, destructive operations during automated probing.
-
-Return connector_status_connected if something is connected to the
-connector, connector_status_disconnected if nothing is connected and
-connector_status_unknown if the connection state isn't known.
-
-Drivers should only return connector_status_connected if the
-connection status has really been probed as connected. Connectors that
-can't detect the connection status, or failed connection status probes,
-should return connector_status_unknown.
 
 Cleanup
 -------
@@ -463,20 +260,8 @@ created for fetching EDID data and performing monitor detection. Once
 the process is complete, the new connector is registered with sysfs to
 make its properties available to applications.
 
-KMS API Functions
------------------
-
-.. kernel-doc:: drivers/gpu/drm/drm_crtc.c
-   :export:
-
-KMS Data Structures
--------------------
-
-.. kernel-doc:: include/drm/drm_crtc.h
-   :internal:
-
 KMS Locking
------------
+===========
 
 .. kernel-doc:: drivers/gpu/drm/drm_modeset_lock.c
    :doc: kms locking
@@ -490,90 +275,38 @@ KMS Locking
 KMS Properties
 ==============
 
-Drivers may need to expose additional parameters to applications than
-those described in the previous sections. KMS supports attaching
-properties to CRTCs, connectors and planes and offers a userspace API to
-list, get and set the property values.
-
-Properties are identified by a name that uniquely defines the property
-purpose, and store an associated value. For all property types except
-blob properties the value is a 64-bit unsigned integer.
-
-KMS differentiates between properties and property instances. Drivers
-first create properties and then create and associate individual
-instances of those properties to objects. A property can be instantiated
-multiple times and associated with different objects. Values are stored
-in property instances, and all other property information are stored in
-the property and shared between all instances of the property.
-
-Every property is created with a type that influences how the KMS core
-handles the property. Supported property types are
-
-DRM_MODE_PROP_RANGE
-    Range properties report their minimum and maximum admissible values.
-    The KMS core verifies that values set by application fit in that
-    range.
-
-DRM_MODE_PROP_ENUM
-    Enumerated properties take a numerical value that ranges from 0 to
-    the number of enumerated values defined by the property minus one,
-    and associate a free-formed string name to each value. Applications
-    can retrieve the list of defined value-name pairs and use the
-    numerical value to get and set property instance values.
-
-DRM_MODE_PROP_BITMASK
-    Bitmask properties are enumeration properties that additionally
-    restrict all enumerated values to the 0..63 range. Bitmask property
-    instance values combine one or more of the enumerated bits defined
-    by the property.
-
-DRM_MODE_PROP_BLOB
-    Blob properties store a binary blob without any format restriction.
-    The binary blobs are created as KMS standalone objects, and blob
-    property instance values store the ID of their associated blob
-    object.
-
-    Blob properties are only used for the connector EDID property and
-    cannot be created by drivers.
-
-To create a property drivers call one of the following functions
-depending on the property type. All property creation functions take
-property flags and name, as well as type-specific arguments.
-
--  struct drm_property \*drm_property_create_range(struct
-   drm_device \*dev, int flags, const char \*name, uint64_t min,
-   uint64_t max);
-   Create a range property with the given minimum and maximum values.
-
--  struct drm_property \*drm_property_create_enum(struct drm_device
-   \*dev, int flags, const char \*name, const struct
-   drm_prop_enum_list \*props, int num_values);
-   Create an enumerated property. The ``props`` argument points to an
-   array of ``num_values`` value-name pairs.
-
--  struct drm_property \*drm_property_create_bitmask(struct
-   drm_device \*dev, int flags, const char \*name, const struct
-   drm_prop_enum_list \*props, int num_values);
-   Create a bitmask property. The ``props`` argument points to an array
-   of ``num_values`` value-name pairs.
-
-Properties can additionally be created as immutable, in which case they
-will be read-only for applications but can be modified by the driver. To
-create an immutable property drivers must set the
-DRM_MODE_PROP_IMMUTABLE flag at property creation time.
-
-When no array of value-name pairs is readily available at property
-creation time for enumerated or range properties, drivers can create the
-property using the :c:func:`drm_property_create()` function and
-manually add enumeration value-name pairs by calling the
-:c:func:`drm_property_add_enum()` function. Care must be taken to
-properly specify the property type through the ``flags`` argument.
-
-After creating properties drivers can attach property instances to CRTC,
-connector and plane objects by calling the
-:c:func:`drm_object_attach_property()`. The function takes a
-pointer to the target object, a pointer to the previously created
-property and an initial instance value.
+Property Types and Blob Property Support
+----------------------------------------
+
+.. kernel-doc:: drivers/gpu/drm/drm_property.c
+   :doc: overview
+
+.. kernel-doc:: include/drm/drm_property.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_property.c
+   :export:
+
+Plane Composition Properties
+----------------------------
+
+.. kernel-doc:: drivers/gpu/drm/drm_blend.c
+   :doc: overview
+
+.. kernel-doc:: drivers/gpu/drm/drm_blend.c
+   :export:
+
+Color Management Properties
+---------------------------
+
+.. kernel-doc:: drivers/gpu/drm/drm_color_mgmt.c
+   :doc: overview
+
+.. kernel-doc:: include/drm/drm_color_mgmt.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_color_mgmt.c
+   :export:
 
 Existing KMS Properties
 -----------------------

Documentation/gpu/drm-mm.rst

@@ -26,12 +26,12 @@ TTM, but has no video RAM management capabilities and is thus limited to
 UMA devices.
 
 The Translation Table Manager (TTM)
------------------------------------
+===================================
 
 TTM design background and information belongs here.
 
 TTM initialization
-~~~~~~~~~~~~~~~~~~
+------------------
 
     **Warning**
 
@@ -77,7 +77,7 @@ object, ttm_global_item_ref() is used to create an initial reference
 count for the TTM, which will call your initialization function.
 
 The Graphics Execution Manager (GEM)
-------------------------------------
+====================================
 
 The GEM design approach has resulted in a memory manager that doesn't
 provide full coverage of all (or even all common) use cases in its
@@ -114,7 +114,7 @@ read & write, mapping, and domain ownership transfers are left to
 driver-specific ioctls.
 
 GEM Initialization
-~~~~~~~~~~~~~~~~~~
+------------------
 
 Drivers that use GEM must set the DRIVER_GEM bit in the struct
 :c:type:`struct drm_driver <drm_driver>` driver_features
@@ -132,7 +132,7 @@ typically not managed by GEM, and must be initialized separately into
 its own DRM MM object.
 
 GEM Objects Creation
-~~~~~~~~~~~~~~~~~~~~
+--------------------
 
 GEM splits creation of GEM objects and allocation of the memory that
 backs them in two distinct operations.
@@ -173,7 +173,7 @@ a call to :c:func:`drm_gem_private_object_init()` instead of
 must be managed by drivers.
 
 GEM Objects Lifetime
-~~~~~~~~~~~~~~~~~~~~
+--------------------
 
 All GEM objects are reference-counted by the GEM core. References can be
 acquired and release by :c:func:`calling
@@ -196,7 +196,7 @@ resources created by the GEM core, which need to be released with
 :c:func:`drm_gem_object_release()`.
 
 GEM Objects Naming
-~~~~~~~~~~~~~~~~~~
+------------------
 
 Communication between userspace and the kernel refers to GEM objects
 using local handles, global names or, more recently, file descriptors.
@@ -245,7 +245,7 @@ Furthermore PRIME also allows cross-device buffer sharing since it is
 based on dma-bufs.
 
 GEM Objects Mapping
-~~~~~~~~~~~~~~~~~~~
+-------------------
 
 Because mapping operations are fairly heavyweight GEM favours
 read/write-like access to buffers, implemented through driver-specific
@@ -304,7 +304,7 @@ Drivers that want to map the GEM object upfront instead of handling page
 faults can implement their own mmap file operation handler.
 
 Memory Coherency
-~~~~~~~~~~~~~~~~
+----------------
 
 When mapped to the device or used in a command buffer, backing pages for
 an object are flushed to memory and marked write combined so as to be
@@ -320,7 +320,7 @@ blocks the client and waits for rendering to complete before performing
 any necessary flushing operations).
 
 Command Execution
-~~~~~~~~~~~~~~~~~
+-----------------
 
 Perhaps the most important GEM function for GPU devices is providing a
 command execution interface to clients. Client programs construct
@@ -348,8 +348,20 @@ GEM Function Reference
 .. kernel-doc:: include/drm/drm_gem.h
    :internal:
 
+GEM CMA Helper Functions Reference
+----------------------------------
+
+.. kernel-doc:: drivers/gpu/drm/drm_gem_cma_helper.c
+   :doc: cma helpers
+
+.. kernel-doc:: drivers/gpu/drm/drm_gem_cma_helper.c
+   :export:
+
+.. kernel-doc:: include/drm/drm_gem_cma_helper.h
+   :internal:
+
 VMA Offset Manager
-------------------
+==================
 
 .. kernel-doc:: drivers/gpu/drm/drm_vma_manager.c
    :doc: vma offset manager
@@ -361,14 +373,14 @@ VMA Offset Manager
    :internal:
 
 PRIME Buffer Sharing
---------------------
+====================
 
 PRIME is the cross device buffer sharing framework in drm, originally
 created for the OPTIMUS range of multi-gpu platforms. To userspace PRIME
 buffers are dma-buf based file descriptors.
 
 Overview and Driver Interface
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+-----------------------------
 
 Similar to GEM global names, PRIME file descriptors are also used to
 share buffer objects across processes. They offer additional security:
@@ -406,7 +418,7 @@ struct drm_gem_object \*obj, int flags); struct drm_gem_object \*
 support PRIME.
 
 PRIME Helper Functions
-~~~~~~~~~~~~~~~~~~~~~~
+----------------------
 
 .. kernel-doc:: drivers/gpu/drm/drm_prime.c
    :doc: PRIME Helpers
@@ -418,16 +430,16 @@ PRIME Function References
    :export:
 
 DRM MM Range Allocator
-----------------------
+======================
 
 Overview
-~~~~~~~~
+--------
 
 .. kernel-doc:: drivers/gpu/drm/drm_mm.c
    :doc: Overview
 
 LRU Scan/Eviction Support
-~~~~~~~~~~~~~~~~~~~~~~~~~
+-------------------------
 
 .. kernel-doc:: drivers/gpu/drm/drm_mm.c
    :doc: lru scan roaster
@@ -440,15 +452,3 @@ DRM MM Range Allocator Function References
 
 .. kernel-doc:: include/drm/drm_mm.h
    :internal:
-
-CMA Helper Functions Reference
-------------------------------
-
-.. kernel-doc:: drivers/gpu/drm/drm_gem_cma_helper.c
-   :doc: cma helpers
-
-.. kernel-doc:: drivers/gpu/drm/drm_gem_cma_helper.c
-   :export:
-
-.. kernel-doc:: include/drm/drm_gem_cma_helper.h
-   :internal:

Documentation/gpu/drm-uapi.rst

@@ -33,6 +33,76 @@ Primary Nodes, DRM Master and Authentication
 .. kernel-doc:: include/drm/drm_auth.h
    :internal:
 
+Open-Source Userspace Requirements
+==================================
+
+The DRM subsystem has stricter requirements than most other kernel subsystems on
+what the userspace side for new uAPI needs to look like. This section here
+explains what exactly those requirements are, and why they exist.
+
+The short summary is that any addition of DRM uAPI requires corresponding
+open-sourced userspace patches, and those patches must be reviewed and ready for
+merging into a suitable and canonical upstream project.
+
+GFX devices (both display and render/GPU side) are really complex bits of
+hardware, with userspace and kernel by necessity having to work together really
+closely.  The interfaces, for rendering and modesetting, must be extremely wide
+and flexible, and therefore it is almost always impossible to precisely define
+them for every possible corner case. This in turn makes it really practically
+infeasible to differentiate between behaviour that's required by userspace, and
+which must not be changed to avoid regressions, and behaviour which is only an
+accidental artifact of the current implementation.
+
+Without access to the full source code of all userspace users that means it
+becomes impossible to change the implementation details, since userspace could
+depend upon the accidental behaviour of the current implementation in minute
+details. And debugging such regressions without access to source code is pretty
+much impossible. As a consequence this means:
+
+- The Linux kernel's "no regression" policy holds in practice only for
+  open-source userspace of the DRM subsystem. DRM developers are perfectly fine
+  if closed-source blob drivers in userspace use the same uAPI as the open
+  drivers, but they must do so in the exact same way as the open drivers.
+  Creative (ab)use of the interfaces will, and in the past routinely has, lead
+  to breakage.
+
+- Any new userspace interface must have an open-source implementation as
+  demonstration vehicle.
+
+The other reason for requiring open-source userspace is uAPI review. Since the
+kernel and userspace parts of a GFX stack must work together so closely, code
+review can only assess whether a new interface achieves its goals by looking at
+both sides. Making sure that the interface indeed covers the use-case fully
+leads to a few additional requirements:
+
+- The open-source userspace must not be a toy/test application, but the real
+  thing. Specifically it needs to handle all the usual error and corner cases.
+  These are often the places where new uAPI falls apart and hence essential to
+  assess the fitness of a proposed interface.
+
+- The userspace side must be fully reviewed and tested to the standards of that
+  userspace project. For e.g. mesa this means piglit testcases and review on the
+  mailing list. This is again to ensure that the new interface actually gets the
+  job done.
+
+- The userspace patches must be against the canonical upstream, not some vendor
+  fork. This is to make sure that no one cheats on the review and testing
+  requirements by doing a quick fork.
+
+- The kernel patch can only be merged after all the above requirements are met,
+  but it **must** be merged **before** the userspace patches land. uAPI always flows
+  from the kernel, doing things the other way round risks divergence of the uAPI
+  definitions and header files.
+
+These are fairly steep requirements, but have grown out from years of shared
+pain and experience with uAPI added hastily, and almost always regretted about
+just as fast. GFX devices change really fast, requiring a paradigm shift and
+entire new set of uAPI interfaces every few years at least. Together with the
+Linux kernel's guarantee to keep existing userspace running for 10+ years this
+is already rather painful for the DRM subsystem, with multiple different uAPIs
+for the same thing co-existing. If we add a few more complete mistakes into the
+mix every year it would be entirely unmanageable.
+
 Render nodes
 ============
 
@@ -86,6 +156,43 @@ other hand, a driver requires shared state between clients which is
 visible to user-space and accessible beyond open-file boundaries, they
 cannot support render nodes.
 
+Validating changes with IGT
+===========================
+
+There's a collection of tests that aims to cover the whole functionality of
+DRM drivers and that can be used to check that changes to DRM drivers or the
+core don't regress existing functionality. This test suite is called IGT and
+its code can be found in https://cgit.freedesktop.org/drm/igt-gpu-tools/.
+
+To build IGT, start by installing its build dependencies. In Debian-based
+systems::
+
+	# apt-get build-dep intel-gpu-tools
+
+And in Fedora-based systems::
+
+	# dnf builddep intel-gpu-tools
+
+Then clone the repository::
+
+	$ git clone git://anongit.freedesktop.org/drm/igt-gpu-tools
+
+Configure the build system and start the build::
+
+	$ cd igt-gpu-tools && ./autogen.sh && make -j6
+
+Download the piglit dependency::
+
+	$ ./scripts/run-tests.sh -d
+
+And run the tests::
+
+	$ ./scripts/run-tests.sh -t kms -t core -s
+
+run-tests.sh is a wrapper around piglit that will execute the tests matching
+the -t options. A report in HTML format will be available in
+./results/html/index.html. Results can be compared with piglit.
+
 VBlank event handling
 =====================
 

Documentation/gpu/i915.rst

@@ -70,6 +70,9 @@ Frontbuffer Tracking
 .. kernel-doc:: drivers/gpu/drm/i915/intel_frontbuffer.c
    :doc: frontbuffer tracking
 
+.. kernel-doc:: drivers/gpu/drm/i915/intel_frontbuffer.h
+   :internal:
+
 .. kernel-doc:: drivers/gpu/drm/i915/intel_frontbuffer.c
    :internal:
 

Documentation/gpu/index.rst

@@ -12,6 +12,7 @@ Linux GPU Driver Developer's Guide
    drm-uapi
    i915
    vga-switcheroo
+   vgaarbiter
 
 .. only::  subproject
 

Documentation/gpu/kms-properties.csv

@@ -1,23 +1,10 @@
 Owner Module/Drivers,Group,Property Name,Type,Property Values,Object attached,Description/Restrictions
-DRM,Generic,“rotation”,BITMASK,"{ 0, ""rotate-0"" }, { 1, ""rotate-90"" }, { 2, ""rotate-180"" }, { 3, ""rotate-270"" }, { 4, ""reflect-x"" }, { 5, ""reflect-y"" }","CRTC, Plane",rotate-(degrees) rotates the image by the specified amount in degrees in counter clockwise direction. reflect-x and reflect-y reflects the image along the specified axis prior to rotation
 ,,“scaling mode”,ENUM,"{ ""None"", ""Full"", ""Center"", ""Full aspect"" }",Connector,"Supported by: amdgpu, gma500, i915, nouveau and radeon."
 ,Connector,“EDID”,BLOB | IMMUTABLE,0,Connector,Contains id of edid blob ptr object.
 ,,“DPMS”,ENUM,"{ “On”, “Standby”, “Suspend”, “Off” }",Connector,Contains DPMS operation mode value.
 ,,“PATH”,BLOB | IMMUTABLE,0,Connector,Contains topology path to a connector.
 ,,“TILE”,BLOB | IMMUTABLE,0,Connector,Contains tiling information for a connector.
 ,,“CRTC_ID”,OBJECT,DRM_MODE_OBJECT_CRTC,Connector,CRTC that connector is attached to (atomic)
-,Plane,“type”,ENUM | IMMUTABLE,"{ ""Overlay"", ""Primary"", ""Cursor"" }",Plane,Plane type
-,,“SRC_X”,RANGE,"Min=0, Max=UINT_MAX",Plane,Scanout source x coordinate in 16.16 fixed point (atomic)
-,,“SRC_Y”,RANGE,"Min=0, Max=UINT_MAX",Plane,Scanout source y coordinate in 16.16 fixed point (atomic)
-,,“SRC_W”,RANGE,"Min=0, Max=UINT_MAX",Plane,Scanout source width in 16.16 fixed point (atomic)
-,,“SRC_H”,RANGE,"Min=0, Max=UINT_MAX",Plane,Scanout source height in 16.16 fixed point (atomic)
-,,“CRTC_X”,SIGNED_RANGE,"Min=INT_MIN, Max=INT_MAX",Plane,Scanout CRTC (destination) x coordinate (atomic)
-,,“CRTC_Y”,SIGNED_RANGE,"Min=INT_MIN, Max=INT_MAX",Plane,Scanout CRTC (destination) y coordinate (atomic)
-,,“CRTC_W”,RANGE,"Min=0, Max=UINT_MAX",Plane,Scanout CRTC (destination) width (atomic)
-,,“CRTC_H”,RANGE,"Min=0, Max=UINT_MAX",Plane,Scanout CRTC (destination) height (atomic)
-,,“FB_ID”,OBJECT,DRM_MODE_OBJECT_FB,Plane,Scanout framebuffer (atomic)
-,,“CRTC_ID”,OBJECT,DRM_MODE_OBJECT_CRTC,Plane,CRTC that plane is attached to (atomic)
-,,“zpos”,RANGE,"Min=0, Max=UINT_MAX","Plane,Z-order of the plane.Planes with higher Z-order values are displayed on top, planes with identical Z-order values are display in an undefined order"
 ,DVI-I,“subconnector”,ENUM,"{ “Unknown”, “DVI-D”, “DVI-A” }",Connector,TBD
 ,,“select subconnector”,ENUM,"{ “Automatic”, “DVI-D”, “DVI-A” }",Connector,TBD
 ,TV,“subconnector”,ENUM,"{ ""Unknown"", ""Composite"", ""SVIDEO"", ""Component"", ""SCART"" }",Connector,TBD
@@ -36,12 +23,6 @@ DRM,Generic,“rotation”,BITMASK,"{ 0, ""rotate-0"" }, { 1, ""rotate-90"" }, {
 ,Virtual GPU,“suggested X”,RANGE,"Min=0, Max=0xffffffff",Connector,property to suggest an X offset for a connector
 ,,“suggested Y”,RANGE,"Min=0, Max=0xffffffff",Connector,property to suggest an Y offset for a connector
 ,Optional,"""aspect ratio""",ENUM,"{ ""None"", ""4:3"", ""16:9"" }",Connector,TDB
-,,“dirty”,ENUM | IMMUTABLE,"{ ""Off"", ""On"", ""Annotate"" }",Connector,TBD
-,,“DEGAMMA_LUT”,BLOB,0,CRTC,DRM property to set the degamma lookup table (LUT) mapping pixel data from the framebuffer before it is given to the transformation matrix. The data is an interpreted as an array of struct drm_color_lut elements. Hardware might choose not to use the full precision of the LUT elements nor use all the elements of the LUT (for example the hardware might choose to interpolate between LUT[0] and LUT[4]).
-,,“DEGAMMA_LUT_SIZE”,RANGE | IMMUTABLE,"Min=0, Max=UINT_MAX",CRTC,DRM property to gives the size of the lookup table to be set on the DEGAMMA_LUT property (the size depends on the underlying hardware).
-,,“CTM”,BLOB,0,CRTC,DRM property to set the current transformation matrix (CTM) apply to pixel data after the lookup through the degamma LUT and before the lookup through the gamma LUT. The data is an interpreted as a struct drm_color_ctm.
-,,“GAMMA_LUT”,BLOB,0,CRTC,DRM property to set the gamma lookup table (LUT) mapping pixel data after to the transformation matrix to data sent to the connector. The data is an interpreted as an array of struct drm_color_lut elements. Hardware might choose not to use the full precision of the LUT elements nor use all the elements of the LUT (for example the hardware might choose to interpolate between LUT[0] and LUT[4]).
-,,“GAMMA_LUT_SIZE”,RANGE | IMMUTABLE,"Min=0, Max=UINT_MAX",CRTC,DRM property to gives the size of the lookup table to be set on the GAMMA_LUT property (the size depends on the underlying hardware).
 i915,Generic,"""Broadcast RGB""",ENUM,"{ ""Automatic"", ""Full"", ""Limited 16:235"" }",Connector,"When this property is set to Limited 16:235 and CTM is set, the hardware will be programmed with the result of the multiplication of CTM by the limited range matrix to ensure the pixels normaly in the range 0..1.0 are remapped to the range 16/255..235/255."
 ,,“audio”,ENUM,"{ ""force-dvi"", ""off"", ""auto"", ""on"" }",Connector,TBD
 ,SDVO-TV,“mode”,ENUM,"{ ""NTSC_M"", ""NTSC_J"", ""NTSC_443"", ""PAL_B"" } etc.",Connector,TBD
@@ -95,7 +76,6 @@ armada,CRTC,"""CSC_YUV""",ENUM,"{ ""Auto"" , ""CCIR601"", ""CCIR709"" }",CRTC,TB
 ,,"""contrast""",RANGE,"Min=0, Max=0x7fff",Plane,TBD
 ,,"""saturation""",RANGE,"Min=0, Max=0x7fff",Plane,TBD
 exynos,CRTC,“mode”,ENUM,"{ ""normal"", ""blank"" }",CRTC,TBD
-,Overlay,“zpos”,RANGE,"Min=0, Max=MAX_PLANE-1",Plane,TBD
 i2c/ch7006_drv,Generic,“scale”,RANGE,"Min=0, Max=2",Connector,TBD
 ,TV,“mode”,ENUM,"{ ""PAL"", ""PAL-M"",""PAL-N""}, ”PAL-Nc"" , ""PAL-60"", ""NTSC-M"", ""NTSC-J"" }",Connector,TBD
 nouveau,NV10 Overlay,"""colorkey""",RANGE,"Min=0, Max=0x01ffffff",Plane,TBD
@@ -126,4 +106,3 @@ radeon,DVI-I,“coherent”,RANGE,"Min=0, Max=1",Connector,TBD
 ,FMT Dithering,“dither”,ENUM,"{ ""off"", ""on"" }",Connector,TBD
 rcar-du,Generic,"""alpha""",RANGE,"Min=0, Max=255",Plane,TBD
 ,,"""colorkey""",RANGE,"Min=0, Max=0x01ffffff",Plane,TBD
-,,"""zpos""",RANGE,"Min=1, Max=7",Plane,TBD

Documentation/vgaarbiter.txt → Documentation/gpu/vgaarbiter.rst

@@ -1,4 +1,4 @@
-
+===========
 VGA Arbiter
 ===========
 
@@ -19,21 +19,8 @@ control bus resources. Therefore an arbitration scheme outside of the X server
 is needed to control the sharing of these resources. This document introduces
 the operation of the VGA arbiter implemented for the Linux kernel.
 
-----------------------------------------------------------------------------
-
-I.  Details and Theory of Operation
-        I.1 vgaarb
-        I.2 libpciaccess
-        I.3 xf86VGAArbiter (X server implementation)
-II. Credits
-III.References
-
-
-I. Details and Theory of Operation
-==================================
-
-I.1 vgaarb
-----------
+vgaarb kernel/userspace ABI
+---------------------------
 
 The vgaarb is a module of the Linux Kernel. When it is initially loaded, it
 scans all PCI devices and adds the VGA ones inside the arbitration. The
@@ -44,42 +31,52 @@ explicitly tell it by calling vga_set_legacy_decoding().
 The kernel exports a char device interface (/dev/vga_arbiter) to the clients,
 which has the following semantics:
 
- open       : open user instance of the arbiter. By default, it's attached to
-              the default VGA device of the system.
-
- close      : close user instance. Release locks made by the user
-
- read       : return a string indicating the status of the target like:
-
-              "<card_ID>,decodes=<io_state>,owns=<io_state>,locks=<io_state> (ic,mc)"
-
-              An IO state string is of the form {io,mem,io+mem,none}, mc and
-              ic are respectively mem and io lock counts (for debugging/
-              diagnostic only). "decodes" indicate what the card currently
-              decodes, "owns" indicates what is currently enabled on it, and
-              "locks" indicates what is locked by this card. If the card is
-              unplugged, we get "invalid" then for card_ID and an -ENODEV
-              error is returned for any command until a new card is targeted.
-
-
- write       : write a command to the arbiter. List of commands:
-
-  target <card_ID>   : switch target to card <card_ID> (see below)
-  lock <io_state>    : acquires locks on target ("none" is an invalid io_state)
-  trylock <io_state> : non-blocking acquire locks on target (returns EBUSY if
-                       unsuccessful)
-  unlock <io_state>  : release locks on target
-  unlock all         : release all locks on target held by this user (not
-                       implemented yet)
-  decodes <io_state> : set the legacy decoding attributes for the card
-
-  poll               : event if something changes on any card (not just the
-                       target)
-
-  card_ID is of the form "PCI:domain:bus:dev.fn". It can be set to "default"
-  to go back to the system default card (TODO: not implemented yet). Currently,
-  only PCI is supported as a prefix, but the userland API may support other bus
-  types in the future, even if the current kernel implementation doesn't.
+open
+        Opens a user instance of the arbiter. By default, it's attached to the
+        default VGA device of the system.
+
+close
+        Close a user instance. Release locks made by the user
+
+read
+        Return a string indicating the status of the target like:
+
+        "<card_ID>,decodes=<io_state>,owns=<io_state>,locks=<io_state> (ic,mc)"
+
+        An IO state string is of the form {io,mem,io+mem,none}, mc and
+        ic are respectively mem and io lock counts (for debugging/
+        diagnostic only). "decodes" indicate what the card currently
+        decodes, "owns" indicates what is currently enabled on it, and
+        "locks" indicates what is locked by this card. If the card is
+        unplugged, we get "invalid" then for card_ID and an -ENODEV
+        error is returned for any command until a new card is targeted.
+
+
+write
+        Write a command to the arbiter. List of commands:
+
+        target <card_ID>
+                switch target to card <card_ID> (see below)
+        lock <io_state>
+                acquires locks on target ("none" is an invalid io_state)
+        trylock <io_state>
+                non-blocking acquire locks on target (returns EBUSY if
+                unsuccessful)
+        unlock <io_state>
+                release locks on target
+        unlock all
+                release all locks on target held by this user (not implemented
+                yet)
+        decodes <io_state>
+                set the legacy decoding attributes for the card
+
+        poll
+                event if something changes on any card (not just the target)
+
+        card_ID is of the form "PCI:domain:bus:dev.fn". It can be set to "default"
+        to go back to the system default card (TODO: not implemented yet). Currently,
+        only PCI is supported as a prefix, but the userland API may support other bus
+        types in the future, even if the current kernel implementation doesn't.
 
 Note about locks:
 
@@ -97,29 +94,35 @@ in the arbiter.
 There is also an in-kernel API of the arbiter in case DRM, vgacon, or other
 drivers want to use it.
 
+In-kernel interface
+-------------------
+
+.. kernel-doc:: include/linux/vgaarb.h
+   :internal:
 
-I.2 libpciaccess
-----------------
+.. kernel-doc:: drivers/gpu/vga/vgaarb.c
+   :export:
+
+libpciaccess
+------------
 
 To use the vga arbiter char device it was implemented an API inside the
 libpciaccess library. One field was added to struct pci_device (each device
-on the system):
+on the system)::
 
     /* the type of resource decoded by the device */
     int vgaarb_rsrc;
 
-Besides it, in pci_system were added:
+Besides it, in pci_system were added::
 
     int vgaarb_fd;
     int vga_count;
     struct pci_device *vga_target;
     struct pci_device *vga_default_dev;
 
-
 The vga_count is used to track how many cards are being arbitrated, so for
 instance, if there is only one card, then it can completely escape arbitration.
 
-
 These functions below acquire VGA resources for the given card and mark those
 resources as locked. If the resources requested are "normal" (and not legacy)
 resources, the arbiter will first check whether the card is doing legacy
@@ -136,44 +139,44 @@ VGA memory and IO afaik). If the card already owns the resources, the function
 succeeds.  vga_arb_trylock() will return (-EBUSY) instead of blocking. Nested
 calls are supported (a per-resource counter is maintained).
 
+Set the target device of this client. ::
 
-Set the target device of this client.
     int  pci_device_vgaarb_set_target   (struct pci_device *dev);
 
-
 For instance, in x86 if two devices on the same bus want to lock different
 resources, both will succeed (lock). If devices are in different buses and
-trying to lock different resources, only the first who tried succeeds.
+trying to lock different resources, only the first who tried succeeds. ::
+
     int  pci_device_vgaarb_lock         (void);
     int  pci_device_vgaarb_trylock      (void);
 
-Unlock resources of device.
+Unlock resources of device. ::
+
     int  pci_device_vgaarb_unlock       (void);
 
 Indicates to the arbiter if the card decodes legacy VGA IOs, legacy VGA
 Memory, both, or none. All cards default to both, the card driver (fbdev for
 example) should tell the arbiter if it has disabled legacy decoding, so the
 card can be left out of the arbitration process (and can be safe to take
-interrupts at any time.
+interrupts at any time. ::
+
     int  pci_device_vgaarb_decodes      (int new_vgaarb_rsrc);
 
-Connects to the arbiter device, allocates the struct
-    int  pci_device_vgaarb_init         (void);
+Connects to the arbiter device, allocates the struct ::
 
-Close the connection
-    void pci_device_vgaarb_fini         (void);
+    int  pci_device_vgaarb_init         (void);
 
+Close the connection ::
 
-I.3 xf86VGAArbiter (X server implementation)
---------------------------------------------
+    void pci_device_vgaarb_fini         (void);
 
-(TODO)
+xf86VGAArbiter (X server implementation)
+----------------------------------------
 
 X server basically wraps all the functions that touch VGA registers somehow.
 
-
-II. Credits
-===========
+References
+----------
 
 Benjamin Herrenschmidt (IBM?) started this work when he discussed such design
 with the Xorg community in 2005 [1, 2]. In the end of 2007, Paulo Zanoni and
@@ -182,11 +185,7 @@ enhancing the kernel code to adapt as a kernel module and also did the
 implementation of the user space side [3]. Now (2009) Tiago Vignatti and Dave
 Airlie finally put this work in shape and queued to Jesse Barnes' PCI tree.
 
-
-III. References
-==============
-
-[0] http://cgit.freedesktop.org/xorg/xserver/commit/?id=4b42448a2388d40f257774fbffdccaea87bd0347
-[1] http://lists.freedesktop.org/archives/xorg/2005-March/006663.html
-[2] http://lists.freedesktop.org/archives/xorg/2005-March/006745.html
-[3] http://lists.freedesktop.org/archives/xorg/2007-October/029507.html
+0) http://cgit.freedesktop.org/xorg/xserver/commit/?id=4b42448a2388d40f257774fbffdccaea87bd0347
+1) http://lists.freedesktop.org/archives/xorg/2005-March/006663.html
+2) http://lists.freedesktop.org/archives/xorg/2005-March/006745.html
+3) http://lists.freedesktop.org/archives/xorg/2007-October/029507.html

Documentation/sync_file.txt

@@ -64,6 +64,20 @@ The sync_file fd now can be sent to userspace.
 If the creation process fail, or the sync_file needs to be released by any
 other reason fput(sync_file->file) should be used.
 
+Receiving Sync Files from Userspace
+-----------------------------------
+
+When userspace needs to send an in-fence to the driver it passes file descriptor
+of the Sync File to the kernel. The kernel can then retrieve the fences
+from it.
+
+Interface:
+	struct fence *sync_file_get_fence(int fd);
+
+
+The returned reference is owned by the caller and must be disposed of
+afterwards using fence_put(). In case of error, a NULL is returned instead.
+
 References:
 [1] struct sync_file in include/linux/sync_file.h
 [2] All interfaces mentioned above defined in include/linux/sync_file.h

MAINTAINERS

@@ -4122,6 +4122,14 @@ S:	Orphan / Obsolete
 F:	drivers/gpu/drm/i810/
 F:	include/uapi/drm/i810_drm.h
 
+DRM DRIVERS FOR MEDIATEK
+M:	CK Hu <ck.hu@mediatek.com>
+M:	Philipp Zabel <p.zabel@pengutronix.de>
+L:	dri-devel@lists.freedesktop.org
+S:	Supported
+F:	drivers/gpu/drm/mediatek/
+F:	Documentation/devicetree/bindings/display/mediatek/
+
 DRM DRIVER FOR MSM ADRENO GPU
 M:	Rob Clark <robdclark@gmail.com>
 L:	linux-arm-msm@vger.kernel.org

arch/arm/boot/dts/am335x-boneblack.dts

@@ -9,6 +9,7 @@
 
 #include "am33xx.dtsi"
 #include "am335x-bone-common.dtsi"
+#include <dt-bindings/display/tda998x.h>
 
 / {
 	model = "TI AM335x BeagleBone Black";
@@ -64,6 +65,16 @@
 			AM33XX_IOPAD(0x9b0, PIN_OUTPUT_PULLDOWN | MUX_MODE3)	/* xdma_event_intr0 */
 		>;
 	};
+
+	mcasp0_pins: mcasp0_pins {
+		pinctrl-single,pins = <
+			AM33XX_IOPAD(0x9ac, PIN_INPUT_PULLUP | MUX_MODE0) /* mcasp0_ahcklx.mcasp0_ahclkx */
+			AM33XX_IOPAD(0x99c, PIN_OUTPUT_PULLDOWN | MUX_MODE2) /* mcasp0_ahclkr.mcasp0_axr2*/
+			AM33XX_IOPAD(0x994, PIN_OUTPUT_PULLUP | MUX_MODE0) /* mcasp0_fsx.mcasp0_fsx */
+			AM33XX_IOPAD(0x990, PIN_OUTPUT_PULLDOWN | MUX_MODE0) /* mcasp0_aclkx.mcasp0_aclkx */
+			AM33XX_IOPAD(0x86c, PIN_OUTPUT_PULLDOWN | MUX_MODE7) /* gpmc_a11.GPIO1_27 */
+		>;
+	};
 };
 
 &lcdc {
@@ -76,16 +87,22 @@
 };
 
 &i2c0 {
-	tda19988 {
+	tda19988: tda19988 {
 		compatible = "nxp,tda998x";
 		reg = <0x70>;
+
 		pinctrl-names = "default", "off";
 		pinctrl-0 = <&nxp_hdmi_bonelt_pins>;
 		pinctrl-1 = <&nxp_hdmi_bonelt_off_pins>;
 
-		port {
-			hdmi_0: endpoint@0 {
-				remote-endpoint = <&lcdc_0>;
+		#sound-dai-cells = <0>;
+		audio-ports = <	TDA998x_I2S	0x03>;
+
+		ports {
+			port@0 {
+				hdmi_0: endpoint@0 {
+					remote-endpoint = <&lcdc_0>;
+				};
 			};
 		};
 	};
@@ -94,3 +111,49 @@
 &rtc {
 	system-power-controller;
 };
+
+&mcasp0	{
+	#sound-dai-cells = <0>;
+	pinctrl-names = "default";
+	pinctrl-0 = <&mcasp0_pins>;
+	status = "okay";
+	op-mode = <0>;	/* MCASP_IIS_MODE */
+	tdm-slots = <2>;
+	serial-dir = <	/* 0: INACTIVE, 1: TX, 2: RX */
+			0 0 1 0
+		>;
+	tx-num-evt = <32>;
+	rx-num-evt = <32>;
+};
+
+/ {
+	clk_mcasp0_fixed: clk_mcasp0_fixed {
+		#clock-cells = <0>;
+		compatible = "fixed-clock";
+		clock-frequency = <24576000>;
+	};
+
+	clk_mcasp0: clk_mcasp0 {
+		#clock-cells = <0>;
+		compatible = "gpio-gate-clock";
+		clocks = <&clk_mcasp0_fixed>;
+		enable-gpios = <&gpio1 27 0>; /* BeagleBone Black Clk enable on GPIO1_27 */
+	};
+
+	sound {
+		compatible = "simple-audio-card";
+		simple-audio-card,name = "TI BeagleBone Black";
+		simple-audio-card,format = "i2s";
+		simple-audio-card,bitclock-master = <&dailink0_master>;
+		simple-audio-card,frame-master = <&dailink0_master>;
+
+		dailink0_master: simple-audio-card,cpu {
+			sound-dai = <&mcasp0>;
+			clocks = <&clk_mcasp0>;
+		};
+
+		simple-audio-card,codec {
+			sound-dai = <&tda19988>;
+		};
+	};
+};

arch/x86/kernel/early-quirks.c

@@ -317,16 +317,11 @@ static phys_addr_t __init i85x_stolen_base(int num, int slot, int func,
 static phys_addr_t __init i865_stolen_base(int num, int slot, int func,
 					   size_t stolen_size)
 {
-	u16 toud;
+	u16 toud = 0;
 
-	/*
-	 * FIXME is the graphics stolen memory region
-	 * always at TOUD? Ie. is it always the last
-	 * one to be allocated by the BIOS?
-	 */
 	toud = read_pci_config_16(0, 0, 0, I865_TOUD);
 
-	return (phys_addr_t)toud << 16;
+	return (phys_addr_t)(toud << 16) + i845_tseg_size();
 }
 
 static phys_addr_t __init gen3_stolen_base(int num, int slot, int func,
@@ -512,8 +507,7 @@ static const struct pci_device_id intel_early_ids[] __initconst = {
 	INTEL_I915GM_IDS(&gen3_early_ops),
 	INTEL_I945G_IDS(&gen3_early_ops),
 	INTEL_I945GM_IDS(&gen3_early_ops),
-	INTEL_VLV_M_IDS(&gen6_early_ops),
-	INTEL_VLV_D_IDS(&gen6_early_ops),
+	INTEL_VLV_IDS(&gen6_early_ops),
 	INTEL_PINEVIEW_IDS(&gen3_early_ops),
 	INTEL_I965G_IDS(&gen3_early_ops),
 	INTEL_G33_IDS(&gen3_early_ops),
@@ -526,10 +520,8 @@ static const struct pci_device_id intel_early_ids[] __initconst = {
 	INTEL_SNB_M_IDS(&gen6_early_ops),
 	INTEL_IVB_M_IDS(&gen6_early_ops),
 	INTEL_IVB_D_IDS(&gen6_early_ops),
-	INTEL_HSW_D_IDS(&gen6_early_ops),
-	INTEL_HSW_M_IDS(&gen6_early_ops),
-	INTEL_BDW_M_IDS(&gen8_early_ops),
-	INTEL_BDW_D_IDS(&gen8_early_ops),
+	INTEL_HSW_IDS(&gen6_early_ops),
+	INTEL_BDW_IDS(&gen8_early_ops),
 	INTEL_CHV_IDS(&chv_early_ops),
 	INTEL_SKL_IDS(&gen9_early_ops),
 	INTEL_BXT_IDS(&gen9_early_ops),

drivers/char/agp/intel-gtt.c

@@ -845,6 +845,8 @@ void intel_gtt_insert_page(dma_addr_t addr,
 			   unsigned int flags)
 {
 	intel_private.driver->write_entry(addr, pg, flags);
+	if (intel_private.driver->chipset_flush)
+		intel_private.driver->chipset_flush();
 }
 EXPORT_SYMBOL(intel_gtt_insert_page);
 

drivers/dma-buf/dma-buf.c

@@ -586,6 +586,22 @@ void dma_buf_unmap_attachment(struct dma_buf_attachment *attach,
 }
 EXPORT_SYMBOL_GPL(dma_buf_unmap_attachment);
 
+static int __dma_buf_begin_cpu_access(struct dma_buf *dmabuf,
+				      enum dma_data_direction direction)
+{
+	bool write = (direction == DMA_BIDIRECTIONAL ||
+		      direction == DMA_TO_DEVICE);
+	struct reservation_object *resv = dmabuf->resv;
+	long ret;
+
+	/* Wait on any implicit rendering fences */
+	ret = reservation_object_wait_timeout_rcu(resv, write, true,
+						  MAX_SCHEDULE_TIMEOUT);
+	if (ret < 0)
+		return ret;
+
+	return 0;
+}
 
 /**
  * dma_buf_begin_cpu_access - Must be called before accessing a dma_buf from the
@@ -608,6 +624,13 @@ int dma_buf_begin_cpu_access(struct dma_buf *dmabuf,
 	if (dmabuf->ops->begin_cpu_access)
 		ret = dmabuf->ops->begin_cpu_access(dmabuf, direction);
 
+	/* Ensure that all fences are waited upon - but we first allow
+	 * the native handler the chance to do so more efficiently if it
+	 * chooses. A double invocation here will be reasonably cheap no-op.
+	 */
+	if (ret == 0)
+		ret = __dma_buf_begin_cpu_access(dmabuf, direction);
+
 	return ret;
 }
 EXPORT_SYMBOL_GPL(dma_buf_begin_cpu_access);

drivers/dma-buf/fence-array.c

@@ -99,6 +99,7 @@ const struct fence_ops fence_array_ops = {
 	.wait = fence_default_wait,
 	.release = fence_array_release,
 };
+EXPORT_SYMBOL(fence_array_ops);
 
 /**
  * fence_array_create - Create a custom fence array
@@ -106,14 +107,14 @@ const struct fence_ops fence_array_ops = {
  * @fences:		[in]	array containing the fences
  * @context:		[in]	fence context to use
  * @seqno:		[in]	sequence number to use
- * @signal_on_any	[in]	signal on any fence in the array
+ * @signal_on_any:	[in]	signal on any fence in the array
  *
  * Allocate a fence_array object and initialize the base fence with fence_init().
  * In case of error it returns NULL.
  *
- * The caller should allocte the fences array with num_fences size
+ * The caller should allocate the fences array with num_fences size
  * and fill it with the fences it wants to add to the object. Ownership of this
- * array is take and fence_put() is used on each fence on release.
+ * array is taken and fence_put() is used on each fence on release.
  *
  * If @signal_on_any is true the fence array signals if any fence in the array
  * signals, otherwise it signals when all fences in the array signal.

drivers/dma-buf/reservation.c

@@ -205,7 +205,7 @@ done:
  * @fence: the shared fence to add
  *
  * Add a fence to a shared slot, obj->lock must be held, and
- * reservation_object_reserve_shared_fence has been called.
+ * reservation_object_reserve_shared() has been called.
  */
 void reservation_object_add_shared_fence(struct reservation_object *obj,
 					 struct fence *fence)

drivers/dma-buf/sync_debug.c

@@ -135,10 +135,16 @@ static void sync_print_sync_file(struct seq_file *s,
 	int i;
 
 	seq_printf(s, "[%p] %s: %s\n", sync_file, sync_file->name,
-		   sync_status_str(atomic_read(&sync_file->status)));
+		   sync_status_str(!fence_is_signaled(sync_file->fence)));
 
-	for (i = 0; i < sync_file->num_fences; ++i)
-		sync_print_fence(s, sync_file->cbs[i].fence, true);
+	if (fence_is_array(sync_file->fence)) {
+		struct fence_array *array = to_fence_array(sync_file->fence);
+
+		for (i = 0; i < array->num_fences; ++i)
+			sync_print_fence(s, array->fences[i], true);
+	} else {
+		sync_print_fence(s, sync_file->fence, true);
+	}
 }
 
 static int sync_debugfs_show(struct seq_file *s, void *unused)

drivers/dma-buf/sync_file.c

@@ -28,11 +28,11 @@
 
 static const struct file_operations sync_file_fops;
 
-static struct sync_file *sync_file_alloc(int size)
+static struct sync_file *sync_file_alloc(void)
 {
 	struct sync_file *sync_file;
 
-	sync_file = kzalloc(size, GFP_KERNEL);
+	sync_file = kzalloc(sizeof(*sync_file), GFP_KERNEL);
 	if (!sync_file)
 		return NULL;
 
@@ -45,6 +45,8 @@ static struct sync_file *sync_file_alloc(int size)
 
 	init_waitqueue_head(&sync_file->wq);
 
+	INIT_LIST_HEAD(&sync_file->cb.node);
+
 	return sync_file;
 
 err:
@@ -54,14 +56,11 @@ err:
 
 static void fence_check_cb_func(struct fence *f, struct fence_cb *cb)
 {
-	struct sync_file_cb *check;
 	struct sync_file *sync_file;
 
-	check = container_of(cb, struct sync_file_cb, cb);
-	sync_file = check->sync_file;
+	sync_file = container_of(cb, struct sync_file, cb);
 
-	if (atomic_dec_and_test(&sync_file->status))
-		wake_up_all(&sync_file->wq);
+	wake_up_all(&sync_file->wq);
 }
 
 /**
@@ -76,23 +75,17 @@ struct sync_file *sync_file_create(struct fence *fence)
 {
 	struct sync_file *sync_file;
 
-	sync_file = sync_file_alloc(offsetof(struct sync_file, cbs[1]));
+	sync_file = sync_file_alloc();
 	if (!sync_file)
 		return NULL;
 
-	sync_file->num_fences = 1;
-	atomic_set(&sync_file->status, 1);
+	sync_file->fence = fence;
+
 	snprintf(sync_file->name, sizeof(sync_file->name), "%s-%s%llu-%d",
 		 fence->ops->get_driver_name(fence),
 		 fence->ops->get_timeline_name(fence), fence->context,
 		 fence->seqno);
 
-	sync_file->cbs[0].fence = fence;
-	sync_file->cbs[0].sync_file = sync_file;
-	if (fence_add_callback(fence, &sync_file->cbs[0].cb,
-			       fence_check_cb_func))
-		atomic_dec(&sync_file->status);
-
 	return sync_file;
 }
 EXPORT_SYMBOL(sync_file_create);
@@ -121,14 +114,73 @@ err:
 	return NULL;
 }
 
-static void sync_file_add_pt(struct sync_file *sync_file, int *i,
-			     struct fence *fence)
+/**
+ * sync_file_get_fence - get the fence related to the sync_file fd
+ * @fd:		sync_file fd to get the fence from
+ *
+ * Ensures @fd references a valid sync_file and returns a fence that
+ * represents all fence in the sync_file. On error NULL is returned.
+ */
+struct fence *sync_file_get_fence(int fd)
+{
+	struct sync_file *sync_file;
+	struct fence *fence;
+
+	sync_file = sync_file_fdget(fd);
+	if (!sync_file)
+		return NULL;
+
+	fence = fence_get(sync_file->fence);
+	fput(sync_file->file);
+
+	return fence;
+}
+EXPORT_SYMBOL(sync_file_get_fence);
+
+static int sync_file_set_fence(struct sync_file *sync_file,
+			       struct fence **fences, int num_fences)
+{
+	struct fence_array *array;
+
+	/*
+	 * The reference for the fences in the new sync_file and held
+	 * in add_fence() during the merge procedure, so for num_fences == 1
+	 * we already own a new reference to the fence. For num_fence > 1
+	 * we own the reference of the fence_array creation.
+	 */
+	if (num_fences == 1) {
+		sync_file->fence = fences[0];
+		kfree(fences);
+	} else {
+		array = fence_array_create(num_fences, fences,
+					   fence_context_alloc(1), 1, false);
+		if (!array)
+			return -ENOMEM;
+
+		sync_file->fence = &array->base;
+	}
+
+	return 0;
+}
+
+static struct fence **get_fences(struct sync_file *sync_file, int *num_fences)
+{
+	if (fence_is_array(sync_file->fence)) {
+		struct fence_array *array = to_fence_array(sync_file->fence);
+
+		*num_fences = array->num_fences;
+		return array->fences;
+	}
+
+	*num_fences = 1;
+	return &sync_file->fence;
+}
+
+static void add_fence(struct fence **fences, int *i, struct fence *fence)
 {
-	sync_file->cbs[*i].fence = fence;
-	sync_file->cbs[*i].sync_file = sync_file;
+	fences[*i] = fence;
 
-	if (!fence_add_callback(fence, &sync_file->cbs[*i].cb,
-				fence_check_cb_func)) {
+	if (!fence_is_signaled(fence)) {
 		fence_get(fence);
 		(*i)++;
 	}
@@ -147,16 +199,24 @@ static void sync_file_add_pt(struct sync_file *sync_file, int *i,
 static struct sync_file *sync_file_merge(const char *name, struct sync_file *a,
 					 struct sync_file *b)
 {
-	int num_fences = a->num_fences + b->num_fences;
 	struct sync_file *sync_file;
-	int i, i_a, i_b;
-	unsigned long size = offsetof(struct sync_file, cbs[num_fences]);
+	struct fence **fences, **nfences, **a_fences, **b_fences;
+	int i, i_a, i_b, num_fences, a_num_fences, b_num_fences;
 
-	sync_file = sync_file_alloc(size);
+	sync_file = sync_file_alloc();
 	if (!sync_file)
 		return NULL;
 
-	atomic_set(&sync_file->status, num_fences);
+	a_fences = get_fences(a, &a_num_fences);
+	b_fences = get_fences(b, &b_num_fences);
+	if (a_num_fences > INT_MAX - b_num_fences)
+		return NULL;
+
+	num_fences = a_num_fences + b_num_fences;
+
+	fences = kcalloc(num_fences, sizeof(*fences), GFP_KERNEL);
+	if (!fences)
+		goto err;
 
 	/*
 	 * Assume sync_file a and b are both ordered and have no
@@ -165,55 +225,69 @@ static struct sync_file *sync_file_merge(const char *name, struct sync_file *a,
 	 * If a sync_file can only be created with sync_file_merge
 	 * and sync_file_create, this is a reasonable assumption.
 	 */
-	for (i = i_a = i_b = 0; i_a < a->num_fences && i_b < b->num_fences; ) {
-		struct fence *pt_a = a->cbs[i_a].fence;
-		struct fence *pt_b = b->cbs[i_b].fence;
+	for (i = i_a = i_b = 0; i_a < a_num_fences && i_b < b_num_fences; ) {
+		struct fence *pt_a = a_fences[i_a];
+		struct fence *pt_b = b_fences[i_b];
 
 		if (pt_a->context < pt_b->context) {
-			sync_file_add_pt(sync_file, &i, pt_a);
+			add_fence(fences, &i, pt_a);
 
 			i_a++;
 		} else if (pt_a->context > pt_b->context) {
-			sync_file_add_pt(sync_file, &i, pt_b);
+			add_fence(fences, &i, pt_b);
 
 			i_b++;
 		} else {
 			if (pt_a->seqno - pt_b->seqno <= INT_MAX)
-				sync_file_add_pt(sync_file, &i, pt_a);
+				add_fence(fences, &i, pt_a);
 			else
-				sync_file_add_pt(sync_file, &i, pt_b);
+				add_fence(fences, &i, pt_b);
 
 			i_a++;
 			i_b++;
 		}
 	}
 
-	for (; i_a < a->num_fences; i_a++)
-		sync_file_add_pt(sync_file, &i, a->cbs[i_a].fence);
+	for (; i_a < a_num_fences; i_a++)
+		add_fence(fences, &i, a_fences[i_a]);
+
+	for (; i_b < b_num_fences; i_b++)
+		add_fence(fences, &i, b_fences[i_b]);
 
-	for (; i_b < b->num_fences; i_b++)
-		sync_file_add_pt(sync_file, &i, b->cbs[i_b].fence);
+	if (i == 0)
+		fences[i++] = fence_get(a_fences[0]);
 
-	if (num_fences > i)
-		atomic_sub(num_fences - i, &sync_file->status);
-	sync_file->num_fences = i;
+	if (num_fences > i) {
+		nfences = krealloc(fences, i * sizeof(*fences),
+				  GFP_KERNEL);
+		if (!nfences)
+			goto err;
+
+		fences = nfences;
+	}
+
+	if (sync_file_set_fence(sync_file, fences, i) < 0) {
+		kfree(fences);
+		goto err;
+	}
 
 	strlcpy(sync_file->name, name, sizeof(sync_file->name));
 	return sync_file;
+
+err:
+	fput(sync_file->file);
+	return NULL;
+
 }
 
 static void sync_file_free(struct kref *kref)
 {
 	struct sync_file *sync_file = container_of(kref, struct sync_file,
 						     kref);
-	int i;
-
-	for (i = 0; i < sync_file->num_fences; ++i) {
-		fence_remove_callback(sync_file->cbs[i].fence,
-				      &sync_file->cbs[i].cb);
-		fence_put(sync_file->cbs[i].fence);
-	}
 
+	if (test_bit(POLL_ENABLED, &sync_file->fence->flags))
+		fence_remove_callback(sync_file->fence, &sync_file->cb);
+	fence_put(sync_file->fence);
 	kfree(sync_file);
 }
 
@@ -228,17 +302,17 @@ static int sync_file_release(struct inode *inode, struct file *file)
 static unsigned int sync_file_poll(struct file *file, poll_table *wait)
 {
 	struct sync_file *sync_file = file->private_data;
-	int status;
 
 	poll_wait(file, &sync_file->wq, wait);
 
-	status = atomic_read(&sync_file->status);
+	if (!poll_does_not_wait(wait) &&
+	    !test_and_set_bit(POLL_ENABLED, &sync_file->fence->flags)) {
+		if (fence_add_callback(sync_file->fence, &sync_file->cb,
+				       fence_check_cb_func) < 0)
+			wake_up_all(&sync_file->wq);
+	}
 
-	if (!status)
-		return POLLIN;
-	if (status < 0)
-		return POLLERR;
-	return 0;
+	return fence_is_signaled(sync_file->fence) ? POLLIN : 0;
 }
 
 static long sync_file_ioctl_merge(struct sync_file *sync_file,
@@ -315,8 +389,9 @@ static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
 {
 	struct sync_file_info info;
 	struct sync_fence_info *fence_info = NULL;
+	struct fence **fences;
 	__u32 size;
-	int ret, i;
+	int num_fences, ret, i;
 
 	if (copy_from_user(&info, (void __user *)arg, sizeof(info)))
 		return -EFAULT;
@@ -324,6 +399,8 @@ static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
 	if (info.flags || info.pad)
 		return -EINVAL;
 
+	fences = get_fences(sync_file, &num_fences);
+
 	/*
 	 * Passing num_fences = 0 means that userspace doesn't want to
 	 * retrieve any sync_fence_info. If num_fences = 0 we skip filling
@@ -333,16 +410,16 @@ static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
 	if (!info.num_fences)
 		goto no_fences;
 
-	if (info.num_fences < sync_file->num_fences)
+	if (info.num_fences < num_fences)
 		return -EINVAL;
 
-	size = sync_file->num_fences * sizeof(*fence_info);
+	size = num_fences * sizeof(*fence_info);
 	fence_info = kzalloc(size, GFP_KERNEL);
 	if (!fence_info)
 		return -ENOMEM;
 
-	for (i = 0; i < sync_file->num_fences; ++i)
-		sync_fill_fence_info(sync_file->cbs[i].fence, &fence_info[i]);
+	for (i = 0; i < num_fences; i++)
+		sync_fill_fence_info(fences[i], &fence_info[i]);
 
 	if (copy_to_user(u64_to_user_ptr(info.sync_fence_info), fence_info,
 			 size)) {
@@ -352,11 +429,8 @@ static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
 
 no_fences:
 	strlcpy(info.name, sync_file->name, sizeof(info.name));
-	info.status = atomic_read(&sync_file->status);
-	if (info.status >= 0)
-		info.status = !info.status;
-
-	info.num_fences = sync_file->num_fences;
+	info.status = fence_is_signaled(sync_file->fence);
+	info.num_fences = num_fences;
 
 	if (copy_to_user((void __user *)arg, &info, sizeof(info)))
 		ret = -EFAULT;

drivers/gpu/drm/Kconfig

@@ -108,33 +108,13 @@ config DRM_KMS_CMA_HELPER
 
 source "drivers/gpu/drm/i2c/Kconfig"
 
-config DRM_TDFX
-	tristate "3dfx Banshee/Voodoo3+"
-	depends on DRM && PCI
-	help
-	  Choose this option if you have a 3dfx Banshee or Voodoo3 (or later),
-	  graphics card.  If M is selected, the module will be called tdfx.
-
 source "drivers/gpu/drm/arm/Kconfig"
 
-config DRM_R128
-	tristate "ATI Rage 128"
-	depends on DRM && PCI
-	select FW_LOADER
-	help
-	  Choose this option if you have an ATI Rage 128 graphics card.  If M
-	  is selected, the module will be called r128.  AGP support for
-	  this card is strongly suggested (unless you have a PCI version).
-
 config DRM_RADEON
 	tristate "ATI Radeon"
 	depends on DRM && PCI
-	select FB_CFB_FILLRECT
-	select FB_CFB_COPYAREA
-	select FB_CFB_IMAGEBLIT
 	select FW_LOADER
         select DRM_KMS_HELPER
-	select DRM_KMS_FB_HELPER
         select DRM_TTM
 	select POWER_SUPPLY
 	select HWMON
@@ -153,12 +133,8 @@ source "drivers/gpu/drm/radeon/Kconfig"
 config DRM_AMDGPU
 	tristate "AMD GPU"
 	depends on DRM && PCI
-	select FB_CFB_FILLRECT
-	select FB_CFB_COPYAREA
-	select FB_CFB_IMAGEBLIT
 	select FW_LOADER
         select DRM_KMS_HELPER
-	select DRM_KMS_FB_HELPER
         select DRM_TTM
 	select POWER_SUPPLY
 	select HWMON
@@ -171,55 +147,11 @@ config DRM_AMDGPU
 	  If M is selected, the module will be called amdgpu.
 
 source "drivers/gpu/drm/amd/amdgpu/Kconfig"
-source "drivers/gpu/drm/amd/powerplay/Kconfig"
-
-source "drivers/gpu/drm/amd/acp/Kconfig"
 
 source "drivers/gpu/drm/nouveau/Kconfig"
 
-config DRM_I810
-	tristate "Intel I810"
-	# !PREEMPT because of missing ioctl locking
-	depends on DRM && AGP && AGP_INTEL && (!PREEMPT || BROKEN)
-	help
-	  Choose this option if you have an Intel I810 graphics card.  If M is
-	  selected, the module will be called i810.  AGP support is required
-	  for this driver to work.
-
 source "drivers/gpu/drm/i915/Kconfig"
 
-config DRM_MGA
-	tristate "Matrox g200/g400"
-	depends on DRM && PCI
-	select FW_LOADER
-	help
-	  Choose this option if you have a Matrox G200, G400 or G450 graphics
-	  card.  If M is selected, the module will be called mga.  AGP
-	  support is required for this driver to work.
-
-config DRM_SIS
-	tristate "SiS video cards"
-	depends on DRM && AGP
-	depends on FB_SIS || FB_SIS=n
-	help
-	  Choose this option if you have a SiS 630 or compatible video
-          chipset. If M is selected the module will be called sis. AGP
-          support is required for this driver to work.
-
-config DRM_VIA
-	tristate "Via unichrome video cards"
-	depends on DRM && PCI
-	help
-	  Choose this option if you have a Via unichrome or compatible video
-	  chipset. If M is selected the module will be called via.
-
-config DRM_SAVAGE
-	tristate "Savage video cards"
-	depends on DRM && PCI
-	help
-	  Choose this option if you have a Savage3D/4/SuperSavage/Pro/Twister
-	  chipset. If M is selected the module will be called savage.
-
 config DRM_VGEM
 	tristate "Virtual GEM provider"
 	depends on DRM
@@ -290,3 +222,81 @@ source "drivers/gpu/drm/arc/Kconfig"
 source "drivers/gpu/drm/hisilicon/Kconfig"
 
 source "drivers/gpu/drm/mediatek/Kconfig"
+
+# Keep legacy drivers last
+
+menuconfig DRM_LEGACY
+	bool "Enable legacy drivers (DANGEROUS)"
+	depends on DRM
+	help
+	  Enable legacy DRI1 drivers. Those drivers expose unsafe and dangerous
+	  APIs to user-space, which can be used to circumvent access
+	  restrictions and other security measures. For backwards compatibility
+	  those drivers are still available, but their use is highly
+	  inadvisable and might harm your system.
+
+	  You are recommended to use the safe modeset-only drivers instead, and
+	  perform 3D emulation in user-space.
+
+	  Unless you have strong reasons to go rogue, say "N".
+
+if DRM_LEGACY
+
+config DRM_TDFX
+	tristate "3dfx Banshee/Voodoo3+"
+	depends on DRM && PCI
+	help
+	  Choose this option if you have a 3dfx Banshee or Voodoo3 (or later),
+	  graphics card.  If M is selected, the module will be called tdfx.
+
+config DRM_R128
+	tristate "ATI Rage 128"
+	depends on DRM && PCI
+	select FW_LOADER
+	help
+	  Choose this option if you have an ATI Rage 128 graphics card.  If M
+	  is selected, the module will be called r128.  AGP support for
+	  this card is strongly suggested (unless you have a PCI version).
+
+config DRM_I810
+	tristate "Intel I810"
+	# !PREEMPT because of missing ioctl locking
+	depends on DRM && AGP && AGP_INTEL && (!PREEMPT || BROKEN)
+	help
+	  Choose this option if you have an Intel I810 graphics card.  If M is
+	  selected, the module will be called i810.  AGP support is required
+	  for this driver to work.
+
+config DRM_MGA
+	tristate "Matrox g200/g400"
+	depends on DRM && PCI
+	select FW_LOADER
+	help
+	  Choose this option if you have a Matrox G200, G400 or G450 graphics
+	  card.  If M is selected, the module will be called mga.  AGP
+	  support is required for this driver to work.
+
+config DRM_SIS
+	tristate "SiS video cards"
+	depends on DRM && AGP
+	depends on FB_SIS || FB_SIS=n
+	help
+	  Choose this option if you have a SiS 630 or compatible video
+	  chipset. If M is selected the module will be called sis. AGP
+	  support is required for this driver to work.
+
+config DRM_VIA
+	tristate "Via unichrome video cards"
+	depends on DRM && PCI
+	help
+	  Choose this option if you have a Via unichrome or compatible video
+	  chipset. If M is selected the module will be called via.
+
+config DRM_SAVAGE
+	tristate "Savage video cards"
+	depends on DRM && PCI
+	help
+	  Choose this option if you have a Savage3D/4/SuperSavage/Pro/Twister
+	  chipset. If M is selected the module will be called savage.
+
+endif # DRM_LEGACY

drivers/gpu/drm/Makefile

@@ -12,7 +12,10 @@ drm-y       :=	drm_auth.o drm_bufs.o drm_cache.o \
 		drm_info.o drm_debugfs.o drm_encoder_slave.o \
 		drm_trace_points.o drm_global.o drm_prime.o \
 		drm_rect.o drm_vma_manager.o drm_flip_work.o \
-		drm_modeset_lock.o drm_atomic.o drm_bridge.o
+		drm_modeset_lock.o drm_atomic.o drm_bridge.o \
+		drm_framebuffer.o drm_connector.o drm_blend.o \
+		drm_encoder.o drm_mode_object.o drm_property.o \
+		drm_plane.o drm_color_mgmt.o
 
 drm-$(CONFIG_COMPAT) += drm_ioc32.o
 drm-$(CONFIG_DRM_GEM_CMA_HELPER) += drm_gem_cma_helper.o
@@ -24,7 +27,7 @@ drm-$(CONFIG_AGP) += drm_agpsupport.o
 drm_kms_helper-y := drm_crtc_helper.o drm_dp_helper.o drm_probe_helper.o \
 		drm_plane_helper.o drm_dp_mst_topology.o drm_atomic_helper.o \
 		drm_kms_helper_common.o drm_dp_dual_mode_helper.o \
-		drm_simple_kms_helper.o drm_blend.o
+		drm_simple_kms_helper.o drm_modeset_helper.o
 
 drm_kms_helper-$(CONFIG_DRM_LOAD_EDID_FIRMWARE) += drm_edid_load.o
 drm_kms_helper-$(CONFIG_DRM_FBDEV_EMULATION) += drm_fb_helper.o
@@ -46,7 +49,7 @@ obj-$(CONFIG_DRM_RADEON)+= radeon/
 obj-$(CONFIG_DRM_AMDGPU)+= amd/amdgpu/
 obj-$(CONFIG_DRM_MGA)	+= mga/
 obj-$(CONFIG_DRM_I810)	+= i810/
-obj-$(CONFIG_DRM_I915)  += i915/
+obj-$(CONFIG_DRM_I915)	+= i915/
 obj-$(CONFIG_DRM_MGAG200) += mgag200/
 obj-$(CONFIG_DRM_VC4)  += vc4/
 obj-$(CONFIG_DRM_CIRRUS_QEMU) += cirrus/

drivers/gpu/drm/amd/amdgpu/Kconfig

@@ -1,3 +1,10 @@
+config DRM_AMDGPU_SI
+	bool "Enable amdgpu support for SI parts"
+	depends on DRM_AMDGPU
+	help
+	  Choose this option if you want to enable experimental support
+	  for SI asics.
+
 config DRM_AMDGPU_CIK
 	bool "Enable amdgpu support for CIK parts"
 	depends on DRM_AMDGPU
@@ -25,3 +32,4 @@ config DRM_AMDGPU_GART_DEBUGFS
 	  Selecting this option creates a debugfs file to inspect the mapped
 	  pages. Uses more memory for housekeeping, enable only for debugging.
 
+source "drivers/gpu/drm/amd/acp/Kconfig"

drivers/gpu/drm/amd/amdgpu/Makefile

@@ -23,13 +23,16 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \
 	amdgpu_pm.o atombios_dp.o amdgpu_afmt.o amdgpu_trace_points.o \
 	atombios_encoders.o amdgpu_sa.o atombios_i2c.o \
 	amdgpu_prime.o amdgpu_vm.o amdgpu_ib.o amdgpu_pll.o \
-	amdgpu_ucode.o amdgpu_bo_list.o amdgpu_ctx.o amdgpu_sync.o
+	amdgpu_ucode.o amdgpu_bo_list.o amdgpu_ctx.o amdgpu_sync.o \
+	amdgpu_gtt_mgr.o
 
 # add asic specific block
 amdgpu-$(CONFIG_DRM_AMDGPU_CIK)+= cik.o cik_ih.o kv_smc.o kv_dpm.o \
 	ci_smc.o ci_dpm.o dce_v8_0.o gfx_v7_0.o cik_sdma.o uvd_v4_2.o vce_v2_0.o \
 	amdgpu_amdkfd_gfx_v7.o
 
+amdgpu-$(CONFIG_DRM_AMDGPU_SI)+= si.o gmc_v6_0.o gfx_v6_0.o si_ih.o si_dma.o dce_v6_0.o si_dpm.o si_smc.o
+
 amdgpu-y += \
 	vi.o
 
@@ -50,15 +53,13 @@ amdgpu-y += \
 amdgpu-y += \
 	amdgpu_dpm.o \
 	amdgpu_powerplay.o \
-	cz_smc.o cz_dpm.o \
-	tonga_smc.o tonga_dpm.o \
-	fiji_smc.o fiji_dpm.o \
-	iceland_smc.o iceland_dpm.o
+	cz_smc.o cz_dpm.o
 
 # add DCE block
 amdgpu-y += \
 	dce_v10_0.o \
-	dce_v11_0.o
+	dce_v11_0.o \
+	dce_virtual.o
 
 # add GFX block
 amdgpu-y += \
@@ -110,14 +111,10 @@ amdgpu-$(CONFIG_VGA_SWITCHEROO) += amdgpu_atpx_handler.o
 amdgpu-$(CONFIG_ACPI) += amdgpu_acpi.o
 amdgpu-$(CONFIG_MMU_NOTIFIER) += amdgpu_mn.o
 
-ifneq ($(CONFIG_DRM_AMD_POWERPLAY),)
-
 include $(FULL_AMD_PATH)/powerplay/Makefile
 
 amdgpu-y += $(AMD_POWERPLAY_FILES)
 
-endif
-
 obj-$(CONFIG_DRM_AMDGPU)+= amdgpu.o
 
 CFLAGS_amdgpu_trace_points.o := -I$(src)

drivers/gpu/drm/amd/amdgpu/ObjectID.h

@@ -90,6 +90,7 @@
 #define ENCODER_OBJECT_ID_INTERNAL_VCE            0x24
 #define ENCODER_OBJECT_ID_INTERNAL_UNIPHY3        0x25
 #define ENCODER_OBJECT_ID_INTERNAL_AMCLK          0x27
+#define ENCODER_OBJECT_ID_VIRTUAL                 0x28
 
 #define ENCODER_OBJECT_ID_GENERAL_EXTERNAL_DVO    0xFF
 
@@ -119,6 +120,7 @@
 #define CONNECTOR_OBJECT_ID_eDP                   0x14
 #define CONNECTOR_OBJECT_ID_MXM                   0x15
 #define CONNECTOR_OBJECT_ID_LVDS_eDP              0x16
+#define CONNECTOR_OBJECT_ID_VIRTUAL               0x17
 
 /* deleted */
 
@@ -147,6 +149,7 @@
 #define GRAPH_OBJECT_ENUM_ID5                     0x05
 #define GRAPH_OBJECT_ENUM_ID6                     0x06
 #define GRAPH_OBJECT_ENUM_ID7                     0x07
+#define GRAPH_OBJECT_ENUM_VIRTUAL                 0x08
 
 /****************************************************/
 /* Graphics Object ID Bit definition                */
@@ -408,6 +411,10 @@
                                                   GRAPH_OBJECT_ENUM_ID1 << ENUM_ID_SHIFT |\
                                                   ENCODER_OBJECT_ID_HDMI_ANX9805 << OBJECT_ID_SHIFT)
 
+#define ENCODER_VIRTUAL_ENUM_VIRTUAL            ( GRAPH_OBJECT_TYPE_ENCODER << OBJECT_TYPE_SHIFT |\
+                                                  GRAPH_OBJECT_ENUM_VIRTUAL << ENUM_ID_SHIFT |\
+                                                  ENCODER_OBJECT_ID_VIRTUAL << OBJECT_ID_SHIFT)
+
 /****************************************************/
 /* Connector Object ID definition - Shared with BIOS */
 /****************************************************/

drivers/gpu/drm/amd/amdgpu/amdgpu.h

@@ -51,11 +51,13 @@
 #include "amdgpu_ih.h"
 #include "amdgpu_irq.h"
 #include "amdgpu_ucode.h"
+#include "amdgpu_ttm.h"
 #include "amdgpu_gds.h"
 #include "amd_powerplay.h"
 #include "amdgpu_acp.h"
 
 #include "gpu_scheduler.h"
+#include "amdgpu_virt.h"
 
 /*
  * Modules parameters.
@@ -63,6 +65,7 @@
 extern int amdgpu_modeset;
 extern int amdgpu_vram_limit;
 extern int amdgpu_gart_size;
+extern int amdgpu_moverate;
 extern int amdgpu_benchmarking;
 extern int amdgpu_testing;
 extern int amdgpu_audio;
@@ -91,6 +94,9 @@ extern unsigned amdgpu_pcie_lane_cap;
 extern unsigned amdgpu_cg_mask;
 extern unsigned amdgpu_pg_mask;
 extern char *amdgpu_disable_cu;
+extern int amdgpu_sclk_deep_sleep_en;
+extern char *amdgpu_virtual_display;
+extern unsigned amdgpu_pp_feature_mask;
 
 #define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS	        3000
 #define AMDGPU_MAX_USEC_TIMEOUT			100000	/* 100 ms */
@@ -105,7 +111,7 @@ extern char *amdgpu_disable_cu;
 #define AMDGPU_MAX_RINGS			16
 #define AMDGPU_MAX_GFX_RINGS			1
 #define AMDGPU_MAX_COMPUTE_RINGS		8
-#define AMDGPU_MAX_VCE_RINGS			2
+#define AMDGPU_MAX_VCE_RINGS			3
 
 /* max number of IP instances */
 #define AMDGPU_MAX_SDMA_INSTANCES		2
@@ -248,10 +254,9 @@ struct amdgpu_vm_pte_funcs {
 			 uint64_t pe, uint64_t src,
 			 unsigned count);
 	/* write pte one entry at a time with addr mapping */
-	void (*write_pte)(struct amdgpu_ib *ib,
-			  const dma_addr_t *pages_addr, uint64_t pe,
-			  uint64_t addr, unsigned count,
-			  uint32_t incr, uint32_t flags);
+	void (*write_pte)(struct amdgpu_ib *ib, uint64_t pe,
+			  uint64_t value, unsigned count,
+			  uint32_t incr);
 	/* for linear pte/pde updates without addr mapping */
 	void (*set_pte_pde)(struct amdgpu_ib *ib,
 			    uint64_t pe,
@@ -316,6 +321,10 @@ struct amdgpu_ring_funcs {
 	/* note usage for clock and power gating */
 	void (*begin_use)(struct amdgpu_ring *ring);
 	void (*end_use)(struct amdgpu_ring *ring);
+	void (*emit_switch_buffer) (struct amdgpu_ring *ring);
+	void (*emit_cntxcntl) (struct amdgpu_ring *ring, uint32_t flags);
+	unsigned (*get_emit_ib_size) (struct amdgpu_ring *ring);
+	unsigned (*get_dma_frame_size) (struct amdgpu_ring *ring);
 };
 
 /*
@@ -396,48 +405,8 @@ int amdgpu_fence_wait_empty(struct amdgpu_ring *ring);
 unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring);
 
 /*
- * TTM.
+ * BO.
  */
-
-#define AMDGPU_TTM_LRU_SIZE	20
-
-struct amdgpu_mman_lru {
-	struct list_head		*lru[TTM_NUM_MEM_TYPES];
-	struct list_head		*swap_lru;
-};
-
-struct amdgpu_mman {
-	struct ttm_bo_global_ref        bo_global_ref;
-	struct drm_global_reference	mem_global_ref;
-	struct ttm_bo_device		bdev;
-	bool				mem_global_referenced;
-	bool				initialized;
-
-#if defined(CONFIG_DEBUG_FS)
-	struct dentry			*vram;
-	struct dentry			*gtt;
-#endif
-
-	/* buffer handling */
-	const struct amdgpu_buffer_funcs	*buffer_funcs;
-	struct amdgpu_ring			*buffer_funcs_ring;
-	/* Scheduler entity for buffer moves */
-	struct amd_sched_entity			entity;
-
-	/* custom LRU management */
-	struct amdgpu_mman_lru			log2_size[AMDGPU_TTM_LRU_SIZE];
-	/* guard for log2_size array, don't add anything in between */
-	struct amdgpu_mman_lru			guard;
-};
-
-int amdgpu_copy_buffer(struct amdgpu_ring *ring,
-		       uint64_t src_offset,
-		       uint64_t dst_offset,
-		       uint32_t byte_count,
-		       struct reservation_object *resv,
-		       struct fence **fence);
-int amdgpu_mmap(struct file *filp, struct vm_area_struct *vma);
-
 struct amdgpu_bo_list_entry {
 	struct amdgpu_bo		*robj;
 	struct ttm_validate_buffer	tv;
@@ -476,8 +445,6 @@ struct amdgpu_bo_va {
 #define AMDGPU_GEM_DOMAIN_MAX		0x3
 
 struct amdgpu_bo {
-	/* Protected by gem.mutex */
-	struct list_head		list;
 	/* Protected by tbo.reserved */
 	u32				prefered_domains;
 	u32				allowed_domains;
@@ -500,10 +467,12 @@ struct amdgpu_bo {
 	struct amdgpu_device		*adev;
 	struct drm_gem_object		gem_base;
 	struct amdgpu_bo		*parent;
+	struct amdgpu_bo		*shadow;
 
 	struct ttm_bo_kmap_obj		dma_buf_vmap;
 	struct amdgpu_mn		*mn;
 	struct list_head		mn_list;
+	struct list_head		shadow_list;
 };
 #define gem_to_amdgpu_bo(gobj) container_of((gobj), struct amdgpu_bo, gem_base)
 
@@ -653,6 +622,7 @@ void amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
 int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
 		     int pages, struct page **pagelist,
 		     dma_addr_t *dma_addr, uint32_t flags);
+int amdgpu_ttm_recover_gart(struct amdgpu_device *adev);
 
 /*
  * GPU MC structures, functions & helpers
@@ -679,6 +649,8 @@ struct amdgpu_mc {
 	uint32_t                fw_version;
 	struct amdgpu_irq_src	vm_fault;
 	uint32_t		vram_type;
+	uint32_t                srbm_soft_reset;
+	struct amdgpu_mode_mc_save save;
 };
 
 /*
@@ -723,13 +695,14 @@ void amdgpu_doorbell_get_kfd_info(struct amdgpu_device *adev,
  */
 
 struct amdgpu_flip_work {
-	struct work_struct		flip_work;
+	struct delayed_work		flip_work;
 	struct work_struct		unpin_work;
 	struct amdgpu_device		*adev;
 	int				crtc_id;
+	u32				target_vblank;
 	uint64_t			base;
 	struct drm_pending_vblank_event *event;
-	struct amdgpu_bo		*old_rbo;
+	struct amdgpu_bo		*old_abo;
 	struct fence			*excl;
 	unsigned			shared_count;
 	struct fence			**shared;
@@ -817,13 +790,17 @@ struct amdgpu_ring {
 /* maximum number of VMIDs */
 #define AMDGPU_NUM_VM	16
 
+/* Maximum number of PTEs the hardware can write with one command */
+#define AMDGPU_VM_MAX_UPDATE_SIZE	0x3FFFF
+
 /* number of entries in page table */
 #define AMDGPU_VM_PTE_COUNT (1 << amdgpu_vm_block_size)
 
 /* PTBs (Page Table Blocks) need to be aligned to 32K */
 #define AMDGPU_VM_PTB_ALIGN_SIZE   32768
-#define AMDGPU_VM_PTB_ALIGN_MASK (AMDGPU_VM_PTB_ALIGN_SIZE - 1)
-#define AMDGPU_VM_PTB_ALIGN(a) (((a) + AMDGPU_VM_PTB_ALIGN_MASK) & ~AMDGPU_VM_PTB_ALIGN_MASK)
+
+/* LOG2 number of continuous pages for the fragment field */
+#define AMDGPU_LOG2_PAGES_PER_FRAG 4
 
 #define AMDGPU_PTE_VALID	(1 << 0)
 #define AMDGPU_PTE_SYSTEM	(1 << 1)
@@ -835,10 +812,7 @@ struct amdgpu_ring {
 #define AMDGPU_PTE_READABLE	(1 << 5)
 #define AMDGPU_PTE_WRITEABLE	(1 << 6)
 
-/* PTE (Page Table Entry) fragment field for different page sizes */
-#define AMDGPU_PTE_FRAG_4KB	(0 << 7)
-#define AMDGPU_PTE_FRAG_64KB	(4 << 7)
-#define AMDGPU_LOG2_PAGES_PER_FRAG 4
+#define AMDGPU_PTE_FRAG(x)	((x & 0x1f) << 7)
 
 /* How to programm VM fault handling */
 #define AMDGPU_VM_FAULT_STOP_NEVER	0
@@ -848,6 +822,7 @@ struct amdgpu_ring {
 struct amdgpu_vm_pt {
 	struct amdgpu_bo_list_entry	entry;
 	uint64_t			addr;
+	uint64_t			shadow_addr;
 };
 
 struct amdgpu_vm {
@@ -950,7 +925,6 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 		      struct amdgpu_job *job);
 int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job);
 void amdgpu_vm_reset_id(struct amdgpu_device *adev, unsigned vm_id);
-uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr);
 int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
 				    struct amdgpu_vm *vm);
 int amdgpu_vm_clear_freed(struct amdgpu_device *adev,
@@ -959,7 +933,7 @@ int amdgpu_vm_clear_invalids(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 			     struct amdgpu_sync *sync);
 int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 			struct amdgpu_bo_va *bo_va,
-			struct ttm_mem_reg *mem);
+			bool clear);
 void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev,
 			     struct amdgpu_bo *bo);
 struct amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm,
@@ -994,6 +968,7 @@ struct amdgpu_ctx {
 	spinlock_t		ring_lock;
 	struct fence            **fences;
 	struct amdgpu_ctx_ring	rings[AMDGPU_MAX_RINGS];
+	bool preamble_presented;
 };
 
 struct amdgpu_ctx_mgr {
@@ -1197,6 +1172,10 @@ struct amdgpu_gfx {
 	unsigned			ce_ram_size;
 	struct amdgpu_cu_info		cu_info;
 	const struct amdgpu_gfx_funcs	*funcs;
+
+	/* reset mask */
+	uint32_t                        grbm_soft_reset;
+	uint32_t                        srbm_soft_reset;
 };
 
 int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm,
@@ -1249,11 +1228,16 @@ struct amdgpu_cs_parser {
 	struct fence			*fence;
 	uint64_t			bytes_moved_threshold;
 	uint64_t			bytes_moved;
+	struct amdgpu_bo_list_entry	*evictable;
 
 	/* user fence */
 	struct amdgpu_bo_list_entry	uf_entry;
 };
 
+#define AMDGPU_PREAMBLE_IB_PRESENT          (1 << 0) /* bit set means command submit involves a preamble IB */
+#define AMDGPU_PREAMBLE_IB_PRESENT_FIRST    (1 << 1) /* bit set means preamble IB is first presented in belonging context */
+#define AMDGPU_HAVE_CTX_SWITCH              (1 << 2) /* bit set means context switch occured */
+
 struct amdgpu_job {
 	struct amd_sched_job    base;
 	struct amdgpu_device	*adev;
@@ -1262,9 +1246,10 @@ struct amdgpu_job {
 	struct amdgpu_sync	sync;
 	struct amdgpu_ib	*ibs;
 	struct fence		*fence; /* the hw fence */
+	uint32_t		preamble_status;
 	uint32_t		num_ibs;
 	void			*owner;
-	uint64_t		ctx;
+	uint64_t		fence_ctx; /* the fence_context this job uses */
 	bool                    vm_needs_flush;
 	unsigned		vm_id;
 	uint64_t		vm_pd_addr;
@@ -1685,6 +1670,7 @@ struct amdgpu_uvd {
 	bool			address_64_bit;
 	bool			use_ctx_buf;
 	struct amd_sched_entity entity;
+	uint32_t                srbm_soft_reset;
 };
 
 /*
@@ -1711,6 +1697,8 @@ struct amdgpu_vce {
 	struct amdgpu_irq_src	irq;
 	unsigned		harvest_config;
 	struct amd_sched_entity	entity;
+	uint32_t                srbm_soft_reset;
+	unsigned		num_rings;
 };
 
 /*
@@ -1728,9 +1716,14 @@ struct amdgpu_sdma_instance {
 
 struct amdgpu_sdma {
 	struct amdgpu_sdma_instance instance[AMDGPU_MAX_SDMA_INSTANCES];
+#ifdef CONFIG_DRM_AMDGPU_SI
+	//SI DMA has a difference trap irq number for the second engine
+	struct amdgpu_irq_src	trap_irq_1;
+#endif
 	struct amdgpu_irq_src	trap_irq;
 	struct amdgpu_irq_src	illegal_inst_irq;
 	int			num_instances;
+	uint32_t                    srbm_soft_reset;
 };
 
 /*
@@ -1832,6 +1825,7 @@ struct amdgpu_asic_funcs {
 	bool (*read_disabled_bios)(struct amdgpu_device *adev);
 	bool (*read_bios_from_rom)(struct amdgpu_device *adev,
 				   u8 *bios, u32 length_bytes);
+	void (*detect_hw_virtualization) (struct amdgpu_device *adev);
 	int (*read_register)(struct amdgpu_device *adev, u32 se_num,
 			     u32 sh_num, u32 reg_offset, u32 *value);
 	void (*set_vga_state)(struct amdgpu_device *adev, bool state);
@@ -1841,8 +1835,9 @@ struct amdgpu_asic_funcs {
 	/* MM block clocks */
 	int (*set_uvd_clocks)(struct amdgpu_device *adev, u32 vclk, u32 dclk);
 	int (*set_vce_clocks)(struct amdgpu_device *adev, u32 evclk, u32 ecclk);
-	/* query virtual capabilities */
-	u32 (*get_virtual_caps)(struct amdgpu_device *adev);
+	/* static power management */
+	int (*get_pcie_lanes)(struct amdgpu_device *adev);
+	void (*set_pcie_lanes)(struct amdgpu_device *adev, int lanes);
 };
 
 /*
@@ -1935,16 +1930,6 @@ struct amdgpu_atcs {
 struct cgs_device *amdgpu_cgs_create_device(struct amdgpu_device *adev);
 void amdgpu_cgs_destroy_device(struct cgs_device *cgs_device);
 
-
-/* GPU virtualization */
-#define AMDGPU_VIRT_CAPS_SRIOV_EN       (1 << 0)
-#define AMDGPU_VIRT_CAPS_IS_VF          (1 << 1)
-struct amdgpu_virtualization {
-	bool supports_sr_iov;
-	bool is_virtual;
-	u32 caps;
-};
-
 /*
  * Core structure, functions and helpers.
  */
@@ -1958,6 +1943,8 @@ struct amdgpu_ip_block_status {
 	bool valid;
 	bool sw;
 	bool hw;
+	bool late_initialized;
+	bool hang;
 };
 
 struct amdgpu_device {
@@ -2016,6 +2003,8 @@ struct amdgpu_device {
 	spinlock_t pcie_idx_lock;
 	amdgpu_rreg_t			pcie_rreg;
 	amdgpu_wreg_t			pcie_wreg;
+	amdgpu_rreg_t			pciep_rreg;
+	amdgpu_wreg_t			pciep_wreg;
 	/* protects concurrent UVD register access */
 	spinlock_t uvd_ctx_idx_lock;
 	amdgpu_rreg_t			uvd_ctx_rreg;
@@ -2056,7 +2045,16 @@ struct amdgpu_device {
 	atomic64_t			num_evictions;
 	atomic_t			gpu_reset_counter;
 
+	/* data for buffer migration throttling */
+	struct {
+		spinlock_t		lock;
+		s64			last_update_us;
+		s64			accum_us; /* accumulated microseconds */
+		u32			log2_max_MBps;
+	} mm_stats;
+
 	/* display */
+	bool				enable_virtual_display;
 	struct amdgpu_mode_info		mode_info;
 	struct work_struct		hotplug_work;
 	struct amdgpu_irq_src		crtc_irq;
@@ -2119,6 +2117,14 @@ struct amdgpu_device {
 	struct kfd_dev          *kfd;
 
 	struct amdgpu_virtualization virtualization;
+
+	/* link all shadow bo */
+	struct list_head                shadow_list;
+	struct mutex                    shadow_list_lock;
+	/* link all gtt */
+	spinlock_t			gtt_list_lock;
+	struct list_head                gtt_list;
+
 };
 
 bool amdgpu_device_is_px(struct drm_device *dev);
@@ -2151,6 +2157,8 @@ void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v);
 #define REG_GET(FIELD, v) (((v) << FIELD##_SHIFT) & FIELD##_MASK)
 #define RREG32_PCIE(reg) adev->pcie_rreg(adev, (reg))
 #define WREG32_PCIE(reg, v) adev->pcie_wreg(adev, (reg), (v))
+#define RREG32_PCIE_PORT(reg) adev->pciep_rreg(adev, (reg))
+#define WREG32_PCIE_PORT(reg, v) adev->pciep_wreg(adev, (reg), (v))
 #define RREG32_SMC(reg) adev->smc_rreg(adev, (reg))
 #define WREG32_SMC(reg, v) adev->smc_wreg(adev, (reg), (v))
 #define RREG32_UVD_CTX(reg) adev->uvd_ctx_rreg(adev, (reg))
@@ -2194,6 +2202,9 @@ void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v);
 #define REG_GET_FIELD(value, reg, field)				\
 	(((value) & REG_FIELD_MASK(reg, field)) >> REG_FIELD_SHIFT(reg, field))
 
+#define WREG32_FIELD(reg, field, val)	\
+	WREG32(mm##reg, (RREG32(mm##reg) & ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
+
 /*
  * BIOS helpers.
  */
@@ -2237,14 +2248,17 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_asic_get_xclk(adev) (adev)->asic_funcs->get_xclk((adev))
 #define amdgpu_asic_set_uvd_clocks(adev, v, d) (adev)->asic_funcs->set_uvd_clocks((adev), (v), (d))
 #define amdgpu_asic_set_vce_clocks(adev, ev, ec) (adev)->asic_funcs->set_vce_clocks((adev), (ev), (ec))
-#define amdgpu_asic_get_virtual_caps(adev) ((adev)->asic_funcs->get_virtual_caps((adev)))
+#define amdgpu_get_pcie_lanes(adev) (adev)->asic_funcs->get_pcie_lanes((adev))
+#define amdgpu_set_pcie_lanes(adev, l) (adev)->asic_funcs->set_pcie_lanes((adev), (l))
+#define amdgpu_asic_get_gpu_clock_counter(adev) (adev)->asic_funcs->get_gpu_clock_counter((adev))
 #define amdgpu_asic_read_disabled_bios(adev) (adev)->asic_funcs->read_disabled_bios((adev))
 #define amdgpu_asic_read_bios_from_rom(adev, b, l) (adev)->asic_funcs->read_bios_from_rom((adev), (b), (l))
+#define amdgpu_asic_detect_hw_virtualization(adev) (adev)->asic_funcs->detect_hw_virtualization((adev))
 #define amdgpu_asic_read_register(adev, se, sh, offset, v)((adev)->asic_funcs->read_register((adev), (se), (sh), (offset), (v)))
 #define amdgpu_gart_flush_gpu_tlb(adev, vmid) (adev)->gart.gart_funcs->flush_gpu_tlb((adev), (vmid))
 #define amdgpu_gart_set_pte_pde(adev, pt, idx, addr, flags) (adev)->gart.gart_funcs->set_pte_pde((adev), (pt), (idx), (addr), (flags))
 #define amdgpu_vm_copy_pte(adev, ib, pe, src, count) ((adev)->vm_manager.vm_pte_funcs->copy_pte((ib), (pe), (src), (count)))
-#define amdgpu_vm_write_pte(adev, ib, pa, pe, addr, count, incr, flags) ((adev)->vm_manager.vm_pte_funcs->write_pte((ib), (pa), (pe), (addr), (count), (incr), (flags)))
+#define amdgpu_vm_write_pte(adev, ib, pe, value, count, incr) ((adev)->vm_manager.vm_pte_funcs->write_pte((ib), (pe), (value), (count), (incr)))
 #define amdgpu_vm_set_pte_pde(adev, ib, pe, addr, count, incr, flags) ((adev)->vm_manager.vm_pte_funcs->set_pte_pde((ib), (pe), (addr), (count), (incr), (flags)))
 #define amdgpu_ring_parse_cs(r, p, ib) ((r)->funcs->parse_cs((p), (ib)))
 #define amdgpu_ring_test_ring(r) (r)->funcs->test_ring((r))
@@ -2259,9 +2273,13 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_ring_emit_gds_switch(r, v, db, ds, wb, ws, ab, as) (r)->funcs->emit_gds_switch((r), (v), (db), (ds), (wb), (ws), (ab), (as))
 #define amdgpu_ring_emit_hdp_flush(r) (r)->funcs->emit_hdp_flush((r))
 #define amdgpu_ring_emit_hdp_invalidate(r) (r)->funcs->emit_hdp_invalidate((r))
+#define amdgpu_ring_emit_switch_buffer(r) (r)->funcs->emit_switch_buffer((r))
+#define amdgpu_ring_emit_cntxcntl(r, d) (r)->funcs->emit_cntxcntl((r), (d))
 #define amdgpu_ring_pad_ib(r, ib) ((r)->funcs->pad_ib((r), (ib)))
 #define amdgpu_ring_init_cond_exec(r) (r)->funcs->init_cond_exec((r))
 #define amdgpu_ring_patch_cond_exec(r,o) (r)->funcs->patch_cond_exec((r),(o))
+#define amdgpu_ring_get_emit_ib_size(r) (r)->funcs->get_emit_ib_size((r))
+#define amdgpu_ring_get_dma_frame_size(r) (r)->funcs->get_dma_frame_size((r))
 #define amdgpu_ih_get_wptr(adev) (adev)->irq.ih_funcs->get_wptr((adev))
 #define amdgpu_ih_decode_iv(adev, iv) (adev)->irq.ih_funcs->decode_iv((adev), (iv))
 #define amdgpu_ih_set_rptr(adev) (adev)->irq.ih_funcs->set_rptr((adev))
@@ -2293,6 +2311,11 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 #define amdgpu_gfx_get_gpu_clock_counter(adev) (adev)->gfx.funcs->get_gpu_clock_counter((adev))
 #define amdgpu_gfx_select_se_sh(adev, se, sh, instance) (adev)->gfx.funcs->select_se_sh((adev), (se), (sh), (instance))
 
+#define amdgpu_dpm_read_sensor(adev, idx, value) \
+	((adev)->pp_enabled ? \
+		(adev)->powerplay.pp_funcs->read_sensor(adev->powerplay.pp_handle, (idx), (value)) : \
+		-EINVAL)
+
 #define amdgpu_dpm_get_temperature(adev) \
 	((adev)->pp_enabled ?						\
 	      (adev)->powerplay.pp_funcs->get_temperature((adev)->powerplay.pp_handle) : \
@@ -2344,11 +2367,6 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 	      (adev)->powerplay.pp_funcs->powergate_vce((adev)->powerplay.pp_handle, (g)) : \
 	      (adev)->pm.funcs->powergate_vce((adev), (g)))
 
-#define amdgpu_dpm_debugfs_print_current_performance_level(adev, m) \
-	((adev)->pp_enabled ?						\
-	      (adev)->powerplay.pp_funcs->print_current_performance_level((adev)->powerplay.pp_handle, (m)) : \
-	      (adev)->pm.funcs->debugfs_print_current_performance_level((adev), (m)))
-
 #define amdgpu_dpm_get_current_power_state(adev) \
 	(adev)->powerplay.pp_funcs->get_current_power_state((adev)->powerplay.pp_handle)
 
@@ -2389,6 +2407,7 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring)
 
 /* Common functions */
 int amdgpu_gpu_reset(struct amdgpu_device *adev);
+bool amdgpu_need_backup(struct amdgpu_device *adev);
 void amdgpu_pci_config_reset(struct amdgpu_device *adev);
 bool amdgpu_card_posted(struct amdgpu_device *adev);
 void amdgpu_update_display_priority(struct amdgpu_device *adev);
@@ -2397,7 +2416,7 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data);
 int amdgpu_cs_get_ring(struct amdgpu_device *adev, u32 ip_type,
 		       u32 ip_instance, u32 ring,
 		       struct amdgpu_ring **out_ring);
-void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *rbo, u32 domain);
+void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *abo, u32 domain);
 bool amdgpu_ttm_bo_is_amdgpu_bo(struct ttm_buffer_object *bo);
 int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages);
 int amdgpu_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr,
@@ -2414,6 +2433,10 @@ uint32_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm,
 void amdgpu_vram_location(struct amdgpu_device *adev, struct amdgpu_mc *mc, u64 base);
 void amdgpu_gtt_location(struct amdgpu_device *adev, struct amdgpu_mc *mc);
 void amdgpu_ttm_set_active_vram_size(struct amdgpu_device *adev, u64 size);
+u64 amdgpu_ttm_get_gtt_mem_size(struct amdgpu_device *adev);
+int amdgpu_ttm_global_init(struct amdgpu_device *adev);
+int amdgpu_ttm_init(struct amdgpu_device *adev);
+void amdgpu_ttm_fini(struct amdgpu_device *adev);
 void amdgpu_program_register_sequence(struct amdgpu_device *adev,
 					     const u32 *registers,
 					     const u32 array_size);
@@ -2425,11 +2448,13 @@ void amdgpu_register_atpx_handler(void);
 void amdgpu_unregister_atpx_handler(void);
 bool amdgpu_has_atpx_dgpu_power_cntl(void);
 bool amdgpu_is_atpx_hybrid(void);
+bool amdgpu_atpx_dgpu_req_power_for_displays(void);
 #else
 static inline void amdgpu_register_atpx_handler(void) {}
 static inline void amdgpu_unregister_atpx_handler(void) {}
 static inline bool amdgpu_has_atpx_dgpu_power_cntl(void) { return false; }
 static inline bool amdgpu_is_atpx_hybrid(void) { return false; }
+static inline bool amdgpu_atpx_dgpu_req_power_for_displays(void) { return false; }
 #endif
 
 /*
@@ -2446,8 +2471,8 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
 				 struct drm_file *file_priv);
 void amdgpu_driver_preclose_kms(struct drm_device *dev,
 				struct drm_file *file_priv);
-int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon);
-int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon);
+int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon);
+int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon);
 u32 amdgpu_get_vblank_counter_kms(struct drm_device *dev, unsigned int pipe);
 int amdgpu_enable_vblank_kms(struct drm_device *dev, unsigned int pipe);
 void amdgpu_disable_vblank_kms(struct drm_device *dev, unsigned int pipe);
@@ -2493,6 +2518,7 @@ static inline void amdgpu_acpi_fini(struct amdgpu_device *adev) { }
 struct amdgpu_bo_va_mapping *
 amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
 		       uint64_t addr, struct amdgpu_bo **bo);
+int amdgpu_cs_sysvm_access_required(struct amdgpu_cs_parser *parser);
 
 #include "amdgpu_object.h"
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_acpi.c

@@ -25,6 +25,7 @@
 #include <linux/acpi.h>
 #include <linux/slab.h>
 #include <linux/power_supply.h>
+#include <linux/pm_runtime.h>
 #include <acpi/video.h>
 #include <drm/drmP.h>
 #include <drm/drm_crtc_helper.h>
@@ -333,6 +334,16 @@ int amdgpu_atif_handler(struct amdgpu_device *adev,
 #endif
 		}
 	}
+	if (req.pending & ATIF_DGPU_DISPLAY_EVENT) {
+		if ((adev->flags & AMD_IS_PX) &&
+		    amdgpu_atpx_dgpu_req_power_for_displays()) {
+			pm_runtime_get_sync(adev->ddev->dev);
+			/* Just fire off a uevent and let userspace tell us what to do */
+			drm_helper_hpd_irq_event(adev->ddev);
+			pm_runtime_mark_last_busy(adev->ddev->dev);
+			pm_runtime_put_autosuspend(adev->ddev->dev);
+		}
+	}
 	/* TODO: check other events */
 
 	/* We've handled the event, stop the notifier chain. The ACPI interface

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c

@@ -143,14 +143,6 @@ int amdgpu_amdkfd_resume(struct amdgpu_device *rdev)
 	return r;
 }
 
-u32 pool_to_domain(enum kgd_memory_pool p)
-{
-	switch (p) {
-	case KGD_POOL_FRAMEBUFFER: return AMDGPU_GEM_DOMAIN_VRAM;
-	default: return AMDGPU_GEM_DOMAIN_GTT;
-	}
-}
-
 int alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
 			void **mem_obj, uint64_t *gpu_addr,
 			void **cpu_ptr)

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v7.c

@@ -103,11 +103,11 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 				uint32_t pipe_id, uint32_t queue_id);
 
 static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
-				unsigned int timeout, uint32_t pipe_id,
+				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id);
 static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd);
 static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
-				unsigned int timeout);
+				unsigned int utimeout);
 static int kgd_address_watch_disable(struct kgd_dev *kgd);
 static int kgd_address_watch_execute(struct kgd_dev *kgd,
 					unsigned int watch_point_id,
@@ -437,11 +437,12 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 }
 
 static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
-				unsigned int timeout, uint32_t pipe_id,
+				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t temp;
+	int timeout = utimeout;
 
 	acquire_queue(kgd, pipe_id, queue_id);
 	WREG32(mmCP_HQD_PQ_DOORBELL_CONTROL, 0);
@@ -452,9 +453,8 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
 		temp = RREG32(mmCP_HQD_ACTIVE);
 		if (temp & CP_HQD_ACTIVE__ACTIVE_MASK)
 			break;
-		if (timeout == 0) {
-			pr_err("kfd: cp queue preemption time out (%dms)\n",
-				temp);
+		if (timeout <= 0) {
+			pr_err("kfd: cp queue preemption time out.\n");
 			release_queue(kgd);
 			return -ETIME;
 		}
@@ -467,12 +467,13 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
 }
 
 static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
-				unsigned int timeout)
+				unsigned int utimeout)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_sdma_rlc_registers *m;
 	uint32_t sdma_base_addr;
 	uint32_t temp;
+	int timeout = utimeout;
 
 	m = get_sdma_mqd(mqd);
 	sdma_base_addr = get_sdma_base_addr(m);
@@ -485,7 +486,7 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 		temp = RREG32(sdma_base_addr + mmSDMA0_RLC0_CONTEXT_STATUS);
 		if (temp & SDMA0_STATUS_REG__RB_CMD_IDLE__SHIFT)
 			break;
-		if (timeout == 0)
+		if (timeout <= 0)
 			return -ETIME;
 		msleep(20);
 		timeout -= 20;

drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v8.c

@@ -62,10 +62,10 @@ static bool kgd_hqd_is_occupied(struct kgd_dev *kgd, uint64_t queue_address,
 		uint32_t pipe_id, uint32_t queue_id);
 static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd);
 static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
-				unsigned int timeout, uint32_t pipe_id,
+				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id);
 static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
-				unsigned int timeout);
+				unsigned int utimeout);
 static void write_vmid_invalidate_request(struct kgd_dev *kgd, uint8_t vmid);
 static int kgd_address_watch_disable(struct kgd_dev *kgd);
 static int kgd_address_watch_execute(struct kgd_dev *kgd,
@@ -349,11 +349,12 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 }
 
 static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
-				unsigned int timeout, uint32_t pipe_id,
+				unsigned int utimeout, uint32_t pipe_id,
 				uint32_t queue_id)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	uint32_t temp;
+	int timeout = utimeout;
 
 	acquire_queue(kgd, pipe_id, queue_id);
 
@@ -363,9 +364,8 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
 		temp = RREG32(mmCP_HQD_ACTIVE);
 		if (temp & CP_HQD_ACTIVE__ACTIVE_MASK)
 			break;
-		if (timeout == 0) {
-			pr_err("kfd: cp queue preemption time out (%dms)\n",
-				temp);
+		if (timeout <= 0) {
+			pr_err("kfd: cp queue preemption time out.\n");
 			release_queue(kgd);
 			return -ETIME;
 		}
@@ -378,12 +378,13 @@ static int kgd_hqd_destroy(struct kgd_dev *kgd, uint32_t reset_type,
 }
 
 static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
-				unsigned int timeout)
+				unsigned int utimeout)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct cik_sdma_rlc_registers *m;
 	uint32_t sdma_base_addr;
 	uint32_t temp;
+	int timeout = utimeout;
 
 	m = get_sdma_mqd(mqd);
 	sdma_base_addr = get_sdma_base_addr(m);
@@ -396,7 +397,7 @@ static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 		temp = RREG32(sdma_base_addr + mmSDMA0_RLC0_CONTEXT_STATUS);
 		if (temp & SDMA0_STATUS_REG__RB_CMD_IDLE__SHIFT)
 			break;
-		if (timeout == 0)
+		if (timeout <= 0)
 			return -ETIME;
 		msleep(20);
 		timeout -= 20;

drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c

@@ -259,6 +259,33 @@ static const int object_connector_convert[] = {
 	DRM_MODE_CONNECTOR_Unknown
 };
 
+bool amdgpu_atombios_has_dce_engine_info(struct amdgpu_device *adev)
+{
+	struct amdgpu_mode_info *mode_info = &adev->mode_info;
+	struct atom_context *ctx = mode_info->atom_context;
+	int index = GetIndexIntoMasterTable(DATA, Object_Header);
+	u16 size, data_offset;
+	u8 frev, crev;
+	ATOM_DISPLAY_OBJECT_PATH_TABLE *path_obj;
+	ATOM_OBJECT_HEADER *obj_header;
+
+	if (!amdgpu_atom_parse_data_header(ctx, index, &size, &frev, &crev, &data_offset))
+		return false;
+
+	if (crev < 2)
+		return false;
+
+	obj_header = (ATOM_OBJECT_HEADER *) (ctx->bios + data_offset);
+	path_obj = (ATOM_DISPLAY_OBJECT_PATH_TABLE *)
+	    (ctx->bios + data_offset +
+	     le16_to_cpu(obj_header->usDisplayPathTableOffset));
+
+	if (path_obj->ucNumOfDispPath)
+		return true;
+	else
+		return false;
+}
+
 bool amdgpu_atombios_get_connector_info_from_object_table(struct amdgpu_device *adev)
 {
 	struct amdgpu_mode_info *mode_info = &adev->mode_info;
@@ -964,6 +991,48 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 		return -EINVAL;
 
 	switch (crev) {
+	case 2:
+	case 3:
+	case 5:
+		/* r6xx, r7xx, evergreen, ni, si.
+		 * TODO: add support for asic_type <= CHIP_RV770*/
+		if (clock_type == COMPUTE_ENGINE_PLL_PARAM) {
+			args.v3.ulClockParams = cpu_to_le32((clock_type << 24) | clock);
+
+			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+
+			dividers->post_div = args.v3.ucPostDiv;
+			dividers->enable_post_div = (args.v3.ucCntlFlag &
+						     ATOM_PLL_CNTL_FLAG_PLL_POST_DIV_EN) ? true : false;
+			dividers->enable_dithen = (args.v3.ucCntlFlag &
+						   ATOM_PLL_CNTL_FLAG_FRACTION_DISABLE) ? false : true;
+			dividers->whole_fb_div = le16_to_cpu(args.v3.ulFbDiv.usFbDiv);
+			dividers->frac_fb_div = le16_to_cpu(args.v3.ulFbDiv.usFbDivFrac);
+			dividers->ref_div = args.v3.ucRefDiv;
+			dividers->vco_mode = (args.v3.ucCntlFlag &
+					      ATOM_PLL_CNTL_FLAG_MPLL_VCO_MODE) ? 1 : 0;
+		} else {
+			/* for SI we use ComputeMemoryClockParam for memory plls */
+			if (adev->asic_type >= CHIP_TAHITI)
+				return -EINVAL;
+			args.v5.ulClockParams = cpu_to_le32((clock_type << 24) | clock);
+			if (strobe_mode)
+				args.v5.ucInputFlag = ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN;
+
+			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+
+			dividers->post_div = args.v5.ucPostDiv;
+			dividers->enable_post_div = (args.v5.ucCntlFlag &
+						     ATOM_PLL_CNTL_FLAG_PLL_POST_DIV_EN) ? true : false;
+			dividers->enable_dithen = (args.v5.ucCntlFlag &
+						   ATOM_PLL_CNTL_FLAG_FRACTION_DISABLE) ? false : true;
+			dividers->whole_fb_div = le16_to_cpu(args.v5.ulFbDiv.usFbDiv);
+			dividers->frac_fb_div = le16_to_cpu(args.v5.ulFbDiv.usFbDivFrac);
+			dividers->ref_div = args.v5.ucRefDiv;
+			dividers->vco_mode = (args.v5.ucCntlFlag &
+					      ATOM_PLL_CNTL_FLAG_MPLL_VCO_MODE) ? 1 : 0;
+		}
+		break;
 	case 4:
 		/* fusion */
 		args.v4.ulClock = cpu_to_le32(clock);	/* 10 khz */
@@ -1108,6 +1177,32 @@ void amdgpu_atombios_set_engine_dram_timings(struct amdgpu_device *adev,
 	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
 }
 
+void amdgpu_atombios_get_default_voltages(struct amdgpu_device *adev,
+					  u16 *vddc, u16 *vddci, u16 *mvdd)
+{
+	struct amdgpu_mode_info *mode_info = &adev->mode_info;
+	int index = GetIndexIntoMasterTable(DATA, FirmwareInfo);
+	u8 frev, crev;
+	u16 data_offset;
+	union firmware_info *firmware_info;
+
+	*vddc = 0;
+	*vddci = 0;
+	*mvdd = 0;
+
+	if (amdgpu_atom_parse_data_header(mode_info->atom_context, index, NULL,
+				   &frev, &crev, &data_offset)) {
+		firmware_info =
+			(union firmware_info *)(mode_info->atom_context->bios +
+						data_offset);
+		*vddc = le16_to_cpu(firmware_info->info_14.usBootUpVDDCVoltage);
+		if ((frev == 2) && (crev >= 2)) {
+			*vddci = le16_to_cpu(firmware_info->info_22.usBootUpVDDCIVoltage);
+			*mvdd = le16_to_cpu(firmware_info->info_22.usBootUpMVDDCVoltage);
+		}
+	}
+}
+
 union set_voltage {
 	struct _SET_VOLTAGE_PS_ALLOCATION alloc;
 	struct _SET_VOLTAGE_PARAMETERS v1;
@@ -1115,6 +1210,52 @@ union set_voltage {
 	struct _SET_VOLTAGE_PARAMETERS_V1_3 v3;
 };
 
+int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage)
+{
+	union set_voltage args;
+	int index = GetIndexIntoMasterTable(COMMAND, SetVoltage);
+	u8 frev, crev;
+
+	if (!amdgpu_atom_parse_cmd_header(adev->mode_info.atom_context, index, &frev, &crev))
+		return -EINVAL;
+
+	switch (crev) {
+	case 1:
+		return -EINVAL;
+	case 2:
+		args.v2.ucVoltageType = SET_VOLTAGE_GET_MAX_VOLTAGE;
+		args.v2.ucVoltageMode = 0;
+		args.v2.usVoltageLevel = 0;
+
+		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		*voltage = le16_to_cpu(args.v2.usVoltageLevel);
+		break;
+	case 3:
+		args.v3.ucVoltageType = voltage_type;
+		args.v3.ucVoltageMode = ATOM_GET_VOLTAGE_LEVEL;
+		args.v3.usVoltageLevel = cpu_to_le16(voltage_id);
+
+		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		*voltage = le16_to_cpu(args.v3.usVoltageLevel);
+		break;
+	default:
+		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int amdgpu_atombios_get_leakage_vddc_based_on_leakage_idx(struct amdgpu_device *adev,
+						      u16 *voltage,
+						      u16 leakage_idx)
+{
+	return amdgpu_atombios_get_max_vddc(adev, VOLTAGE_TYPE_VDDC, leakage_idx, voltage);
+}
+
 void amdgpu_atombios_set_voltage(struct amdgpu_device *adev,
 				 u16 voltage_level,
 				 u8 voltage_type)
@@ -1335,6 +1476,50 @@ static ATOM_VOLTAGE_OBJECT_V3 *amdgpu_atombios_lookup_voltage_object_v3(ATOM_VOL
 	return NULL;
 }
 
+int amdgpu_atombios_get_svi2_info(struct amdgpu_device *adev,
+			      u8 voltage_type,
+			      u8 *svd_gpio_id, u8 *svc_gpio_id)
+{
+	int index = GetIndexIntoMasterTable(DATA, VoltageObjectInfo);
+	u8 frev, crev;
+	u16 data_offset, size;
+	union voltage_object_info *voltage_info;
+	union voltage_object *voltage_object = NULL;
+
+	if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context, index, &size,
+				   &frev, &crev, &data_offset)) {
+		voltage_info = (union voltage_object_info *)
+			(adev->mode_info.atom_context->bios + data_offset);
+
+		switch (frev) {
+		case 3:
+			switch (crev) {
+			case 1:
+				voltage_object = (union voltage_object *)
+					amdgpu_atombios_lookup_voltage_object_v3(&voltage_info->v3,
+								      voltage_type,
+								      VOLTAGE_OBJ_SVID2);
+				if (voltage_object) {
+					*svd_gpio_id = voltage_object->v3.asSVID2Obj.ucSVDGpioId;
+					*svc_gpio_id = voltage_object->v3.asSVID2Obj.ucSVCGpioId;
+				} else {
+					return -EINVAL;
+				}
+				break;
+			default:
+				DRM_ERROR("unknown voltage object table\n");
+				return -EINVAL;
+			}
+			break;
+		default:
+			DRM_ERROR("unknown voltage object table\n");
+			return -EINVAL;
+		}
+
+	}
+	return 0;
+}
+
 bool
 amdgpu_atombios_is_voltage_gpio(struct amdgpu_device *adev,
 				u8 voltage_type, u8 voltage_mode)

drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h

@@ -140,6 +140,8 @@ struct amdgpu_i2c_bus_rec amdgpu_atombios_lookup_i2c_gpio(struct amdgpu_device *
 							  uint8_t id);
 void amdgpu_atombios_i2c_init(struct amdgpu_device *adev);
 
+bool amdgpu_atombios_has_dce_engine_info(struct amdgpu_device *adev);
+
 bool amdgpu_atombios_get_connector_info_from_object_table(struct amdgpu_device *adev);
 
 int amdgpu_atombios_get_clock_info(struct amdgpu_device *adev);
@@ -206,5 +208,19 @@ void amdgpu_atombios_scratch_regs_save(struct amdgpu_device *adev);
 void amdgpu_atombios_scratch_regs_restore(struct amdgpu_device *adev);
 
 void amdgpu_atombios_copy_swap(u8 *dst, u8 *src, u8 num_bytes, bool to_le);
-
+int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,
+			     u16 voltage_id, u16 *voltage);
+int amdgpu_atombios_get_leakage_vddc_based_on_leakage_idx(struct amdgpu_device *adev,
+						      u16 *voltage,
+						      u16 leakage_idx);
+void amdgpu_atombios_get_default_voltages(struct amdgpu_device *adev,
+					  u16 *vddc, u16 *vddci, u16 *mvdd);
+int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
+				       u8 clock_type,
+				       u32 clock,
+				       bool strobe_mode,
+				       struct atom_clock_dividers *dividers);
+int amdgpu_atombios_get_svi2_info(struct amdgpu_device *adev,
+			      u8 voltage_type,
+			      u8 *svd_gpio_id, u8 *svc_gpio_id);
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c

@@ -29,6 +29,7 @@ struct amdgpu_atpx {
 	acpi_handle handle;
 	struct amdgpu_atpx_functions functions;
 	bool is_hybrid;
+	bool dgpu_req_power_for_displays;
 };
 
 static struct amdgpu_atpx_priv {
@@ -73,6 +74,10 @@ bool amdgpu_is_atpx_hybrid(void) {
 	return amdgpu_atpx_priv.atpx.is_hybrid;
 }
 
+bool amdgpu_atpx_dgpu_req_power_for_displays(void) {
+	return amdgpu_atpx_priv.atpx.dgpu_req_power_for_displays;
+}
+
 /**
  * amdgpu_atpx_call - call an ATPX method
  *
@@ -204,6 +209,10 @@ static int amdgpu_atpx_validate(struct amdgpu_atpx *atpx)
 		atpx->is_hybrid = true;
 	}
 
+	atpx->dgpu_req_power_for_displays = false;
+	if (valid_bits & ATPX_DGPU_REQ_POWER_FOR_DISPLAYS)
+		atpx->dgpu_req_power_for_displays = true;
+
 	return 0;
 }
 

drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c

@@ -39,7 +39,8 @@ static int amdgpu_benchmark_do_move(struct amdgpu_device *adev, unsigned size,
 	start_jiffies = jiffies;
 	for (i = 0; i < n; i++) {
 		struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-		r = amdgpu_copy_buffer(ring, saddr, daddr, size, NULL, &fence);
+		r = amdgpu_copy_buffer(ring, saddr, daddr, size, NULL, &fence,
+				       false);
 		if (r)
 			goto exit_do_move;
 		r = fence_wait(fence, false);

drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c

@@ -616,7 +616,7 @@ static int amdgpu_cgs_irq_put(struct cgs_device *cgs_device, unsigned src_id, un
 	return amdgpu_irq_put(adev, adev->irq.sources[src_id], type);
 }
 
-int amdgpu_cgs_set_clockgating_state(struct cgs_device *cgs_device,
+static int amdgpu_cgs_set_clockgating_state(struct cgs_device *cgs_device,
 				  enum amd_ip_block_type block_type,
 				  enum amd_clockgating_state state)
 {
@@ -637,7 +637,7 @@ int amdgpu_cgs_set_clockgating_state(struct cgs_device *cgs_device,
 	return r;
 }
 
-int amdgpu_cgs_set_powergating_state(struct cgs_device *cgs_device,
+static int amdgpu_cgs_set_powergating_state(struct cgs_device *cgs_device,
 				  enum amd_ip_block_type block_type,
 				  enum amd_powergating_state state)
 {
@@ -711,6 +711,47 @@ static int amdgpu_cgs_rel_firmware(struct cgs_device *cgs_device, enum cgs_ucode
 	return -EINVAL;
 }
 
+static uint16_t amdgpu_get_firmware_version(struct cgs_device *cgs_device,
+					enum cgs_ucode_id type)
+{
+	CGS_FUNC_ADEV;
+	uint16_t fw_version;
+
+	switch (type) {
+		case CGS_UCODE_ID_SDMA0:
+			fw_version = adev->sdma.instance[0].fw_version;
+			break;
+		case CGS_UCODE_ID_SDMA1:
+			fw_version = adev->sdma.instance[1].fw_version;
+			break;
+		case CGS_UCODE_ID_CP_CE:
+			fw_version = adev->gfx.ce_fw_version;
+			break;
+		case CGS_UCODE_ID_CP_PFP:
+			fw_version = adev->gfx.pfp_fw_version;
+			break;
+		case CGS_UCODE_ID_CP_ME:
+			fw_version = adev->gfx.me_fw_version;
+			break;
+		case CGS_UCODE_ID_CP_MEC:
+			fw_version = adev->gfx.mec_fw_version;
+			break;
+		case CGS_UCODE_ID_CP_MEC_JT1:
+			fw_version = adev->gfx.mec_fw_version;
+			break;
+		case CGS_UCODE_ID_CP_MEC_JT2:
+			fw_version = adev->gfx.mec_fw_version;
+			break;
+		case CGS_UCODE_ID_RLC_G:
+			fw_version = adev->gfx.rlc_fw_version;
+			break;
+		default:
+			DRM_ERROR("firmware type %d do not have version\n", type);
+			fw_version = 0;
+	}
+	return fw_version;
+}
+
 static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 					enum cgs_ucode_id type,
 					struct cgs_firmware_info *info)
@@ -741,6 +782,7 @@ static int amdgpu_cgs_get_firmware_info(struct cgs_device *cgs_device,
 		info->mc_addr = gpu_addr;
 		info->image_size = data_size;
 		info->version = (uint16_t)le32_to_cpu(header->header.ucode_version);
+		info->fw_version = amdgpu_get_firmware_version(cgs_device, type);
 		info->feature_version = (uint16_t)le32_to_cpu(header->ucode_feature_version);
 	} else {
 		char fw_name[30] = {0};
@@ -848,6 +890,12 @@ static int amdgpu_cgs_query_system_info(struct cgs_device *cgs_device,
 	case CGS_SYSTEM_INFO_GFX_SE_INFO:
 		sys_info->value = adev->gfx.config.max_shader_engines;
 		break;
+	case CGS_SYSTEM_INFO_PCIE_SUB_SYS_ID:
+		sys_info->value = adev->pdev->subsystem_device;
+		break;
+	case CGS_SYSTEM_INFO_PCIE_SUB_SYS_VENDOR_ID:
+		sys_info->value = adev->pdev->subsystem_vendor;
+		break;
 	default:
 		return -ENODEV;
 	}

drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c

@@ -168,12 +168,12 @@ int amdgpu_connector_get_monitor_bpc(struct drm_connector *connector)
 		}
 
 		/* Any defined maximum tmds clock limit we must not exceed? */
-		if (connector->max_tmds_clock > 0) {
+		if (connector->display_info.max_tmds_clock > 0) {
 			/* mode_clock is clock in kHz for mode to be modeset on this connector */
 			mode_clock = amdgpu_connector->pixelclock_for_modeset;
 
 			/* Maximum allowable input clock in kHz */
-			max_tmds_clock = connector->max_tmds_clock * 1000;
+			max_tmds_clock = connector->display_info.max_tmds_clock;
 
 			DRM_DEBUG("%s: hdmi mode dotclock %d kHz, max tmds input clock %d kHz.\n",
 				  connector->name, mode_clock, max_tmds_clock);
@@ -769,8 +769,10 @@ static void amdgpu_connector_destroy(struct drm_connector *connector)
 {
 	struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
 
-	if (amdgpu_connector->ddc_bus->has_aux)
+	if (amdgpu_connector->ddc_bus->has_aux) {
 		drm_dp_aux_unregister(&amdgpu_connector->ddc_bus->aux);
+		amdgpu_connector->ddc_bus->has_aux = false;
+	}
 	amdgpu_connector_free_edid(connector);
 	kfree(amdgpu_connector->con_priv);
 	drm_connector_unregister(connector);
@@ -1504,6 +1506,88 @@ static const struct drm_connector_funcs amdgpu_connector_edp_funcs = {
 	.force = amdgpu_connector_dvi_force,
 };
 
+static struct drm_encoder *
+amdgpu_connector_virtual_encoder(struct drm_connector *connector)
+{
+	int enc_id = connector->encoder_ids[0];
+	struct drm_encoder *encoder;
+	int i;
+	for (i = 0; i < DRM_CONNECTOR_MAX_ENCODER; i++) {
+		if (connector->encoder_ids[i] == 0)
+			break;
+
+		encoder = drm_encoder_find(connector->dev, connector->encoder_ids[i]);
+		if (!encoder)
+			continue;
+
+		if (encoder->encoder_type == DRM_MODE_ENCODER_VIRTUAL)
+			return encoder;
+	}
+
+	/* pick the first one */
+	if (enc_id)
+		return drm_encoder_find(connector->dev, enc_id);
+	return NULL;
+}
+
+static int amdgpu_connector_virtual_get_modes(struct drm_connector *connector)
+{
+	struct drm_encoder *encoder = amdgpu_connector_best_single_encoder(connector);
+
+	if (encoder) {
+		amdgpu_connector_add_common_modes(encoder, connector);
+	}
+
+	return 0;
+}
+
+static int amdgpu_connector_virtual_mode_valid(struct drm_connector *connector,
+					   struct drm_display_mode *mode)
+{
+	return MODE_OK;
+}
+
+static int
+amdgpu_connector_virtual_dpms(struct drm_connector *connector, int mode)
+{
+	return 0;
+}
+
+static enum drm_connector_status
+
+amdgpu_connector_virtual_detect(struct drm_connector *connector, bool force)
+{
+	return connector_status_connected;
+}
+
+static int
+amdgpu_connector_virtual_set_property(struct drm_connector *connector,
+				  struct drm_property *property,
+				  uint64_t val)
+{
+	return 0;
+}
+
+static void amdgpu_connector_virtual_force(struct drm_connector *connector)
+{
+	return;
+}
+
+static const struct drm_connector_helper_funcs amdgpu_connector_virtual_helper_funcs = {
+	.get_modes = amdgpu_connector_virtual_get_modes,
+	.mode_valid = amdgpu_connector_virtual_mode_valid,
+	.best_encoder = amdgpu_connector_virtual_encoder,
+};
+
+static const struct drm_connector_funcs amdgpu_connector_virtual_funcs = {
+	.dpms = amdgpu_connector_virtual_dpms,
+	.detect = amdgpu_connector_virtual_detect,
+	.fill_modes = drm_helper_probe_single_connector_modes,
+	.set_property = amdgpu_connector_virtual_set_property,
+	.destroy = amdgpu_connector_destroy,
+	.force = amdgpu_connector_virtual_force,
+};
+
 void
 amdgpu_connector_add(struct amdgpu_device *adev,
 		      uint32_t connector_id,
@@ -1888,6 +1972,17 @@ amdgpu_connector_add(struct amdgpu_device *adev,
 			connector->interlace_allowed = false;
 			connector->doublescan_allowed = false;
 			break;
+		case DRM_MODE_CONNECTOR_VIRTUAL:
+			amdgpu_dig_connector = kzalloc(sizeof(struct amdgpu_connector_atom_dig), GFP_KERNEL);
+			if (!amdgpu_dig_connector)
+				goto failed;
+			amdgpu_connector->con_priv = amdgpu_dig_connector;
+			drm_connector_init(dev, &amdgpu_connector->base, &amdgpu_connector_virtual_funcs, connector_type);
+			drm_connector_helper_add(&amdgpu_connector->base, &amdgpu_connector_virtual_helper_funcs);
+			subpixel_order = SubPixelHorizontalRGB;
+			connector->interlace_allowed = false;
+			connector->doublescan_allowed = false;
+			break;
 		}
 	}
 

drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c

@@ -91,6 +91,7 @@ static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p,
 				      uint32_t *offset)
 {
 	struct drm_gem_object *gobj;
+	unsigned long size;
 
 	gobj = drm_gem_object_lookup(p->filp, data->handle);
 	if (gobj == NULL)
@@ -101,6 +102,11 @@ static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p,
 	p->uf_entry.tv.bo = &p->uf_entry.robj->tbo;
 	p->uf_entry.tv.shared = true;
 	p->uf_entry.user_pages = NULL;
+
+	size = amdgpu_bo_size(p->uf_entry.robj);
+	if (size != PAGE_SIZE || (data->offset + 8) > size)
+		return -EINVAL;
+
 	*offset = data->offset;
 
 	drm_gem_object_unreference_unlocked(gobj);
@@ -235,70 +241,212 @@ free_chunk:
 	return ret;
 }
 
-/* Returns how many bytes TTM can move per IB.
+/* Convert microseconds to bytes. */
+static u64 us_to_bytes(struct amdgpu_device *adev, s64 us)
+{
+	if (us <= 0 || !adev->mm_stats.log2_max_MBps)
+		return 0;
+
+	/* Since accum_us is incremented by a million per second, just
+	 * multiply it by the number of MB/s to get the number of bytes.
+	 */
+	return us << adev->mm_stats.log2_max_MBps;
+}
+
+static s64 bytes_to_us(struct amdgpu_device *adev, u64 bytes)
+{
+	if (!adev->mm_stats.log2_max_MBps)
+		return 0;
+
+	return bytes >> adev->mm_stats.log2_max_MBps;
+}
+
+/* Returns how many bytes TTM can move right now. If no bytes can be moved,
+ * it returns 0. If it returns non-zero, it's OK to move at least one buffer,
+ * which means it can go over the threshold once. If that happens, the driver
+ * will be in debt and no other buffer migrations can be done until that debt
+ * is repaid.
+ *
+ * This approach allows moving a buffer of any size (it's important to allow
+ * that).
+ *
+ * The currency is simply time in microseconds and it increases as the clock
+ * ticks. The accumulated microseconds (us) are converted to bytes and
+ * returned.
  */
 static u64 amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev)
 {
-	u64 real_vram_size = adev->mc.real_vram_size;
-	u64 vram_usage = atomic64_read(&adev->vram_usage);
+	s64 time_us, increment_us;
+	u64 max_bytes;
+	u64 free_vram, total_vram, used_vram;
 
-	/* This function is based on the current VRAM usage.
+	/* Allow a maximum of 200 accumulated ms. This is basically per-IB
+	 * throttling.
 	 *
-	 * - If all of VRAM is free, allow relocating the number of bytes that
-	 *   is equal to 1/4 of the size of VRAM for this IB.
+	 * It means that in order to get full max MBps, at least 5 IBs per
+	 * second must be submitted and not more than 200ms apart from each
+	 * other.
+	 */
+	const s64 us_upper_bound = 200000;
 
-	 * - If more than one half of VRAM is occupied, only allow relocating
-	 *   1 MB of data for this IB.
-	 *
-	 * - From 0 to one half of used VRAM, the threshold decreases
-	 *   linearly.
-	 *         __________________
-	 * 1/4 of -|\               |
-	 * VRAM    | \              |
-	 *         |  \             |
-	 *         |   \            |
-	 *         |    \           |
-	 *         |     \          |
-	 *         |      \         |
-	 *         |       \________|1 MB
-	 *         |----------------|
-	 *    VRAM 0 %             100 %
-	 *         used            used
-	 *
-	 * Note: It's a threshold, not a limit. The threshold must be crossed
-	 * for buffer relocations to stop, so any buffer of an arbitrary size
-	 * can be moved as long as the threshold isn't crossed before
-	 * the relocation takes place. We don't want to disable buffer
-	 * relocations completely.
+	if (!adev->mm_stats.log2_max_MBps)
+		return 0;
+
+	total_vram = adev->mc.real_vram_size - adev->vram_pin_size;
+	used_vram = atomic64_read(&adev->vram_usage);
+	free_vram = used_vram >= total_vram ? 0 : total_vram - used_vram;
+
+	spin_lock(&adev->mm_stats.lock);
+
+	/* Increase the amount of accumulated us. */
+	time_us = ktime_to_us(ktime_get());
+	increment_us = time_us - adev->mm_stats.last_update_us;
+	adev->mm_stats.last_update_us = time_us;
+	adev->mm_stats.accum_us = min(adev->mm_stats.accum_us + increment_us,
+                                      us_upper_bound);
+
+	/* This prevents the short period of low performance when the VRAM
+	 * usage is low and the driver is in debt or doesn't have enough
+	 * accumulated us to fill VRAM quickly.
 	 *
-	 * The idea is that buffers should be placed in VRAM at creation time
-	 * and TTM should only do a minimum number of relocations during
-	 * command submission. In practice, you need to submit at least
-	 * a dozen IBs to move all buffers to VRAM if they are in GTT.
+	 * The situation can occur in these cases:
+	 * - a lot of VRAM is freed by userspace
+	 * - the presence of a big buffer causes a lot of evictions
+	 *   (solution: split buffers into smaller ones)
 	 *
-	 * Also, things can get pretty crazy under memory pressure and actual
-	 * VRAM usage can change a lot, so playing safe even at 50% does
-	 * consistently increase performance.
+	 * If 128 MB or 1/8th of VRAM is free, start filling it now by setting
+	 * accum_us to a positive number.
+	 */
+	if (free_vram >= 128 * 1024 * 1024 || free_vram >= total_vram / 8) {
+		s64 min_us;
+
+		/* Be more aggresive on dGPUs. Try to fill a portion of free
+		 * VRAM now.
+		 */
+		if (!(adev->flags & AMD_IS_APU))
+			min_us = bytes_to_us(adev, free_vram / 4);
+		else
+			min_us = 0; /* Reset accum_us on APUs. */
+
+		adev->mm_stats.accum_us = max(min_us, adev->mm_stats.accum_us);
+	}
+
+	/* This returns 0 if the driver is in debt to disallow (optional)
+	 * buffer moves.
+	 */
+	max_bytes = us_to_bytes(adev, adev->mm_stats.accum_us);
+
+	spin_unlock(&adev->mm_stats.lock);
+	return max_bytes;
+}
+
+/* Report how many bytes have really been moved for the last command
+ * submission. This can result in a debt that can stop buffer migrations
+ * temporarily.
+ */
+static void amdgpu_cs_report_moved_bytes(struct amdgpu_device *adev,
+					 u64 num_bytes)
+{
+	spin_lock(&adev->mm_stats.lock);
+	adev->mm_stats.accum_us -= bytes_to_us(adev, num_bytes);
+	spin_unlock(&adev->mm_stats.lock);
+}
+
+static int amdgpu_cs_bo_validate(struct amdgpu_cs_parser *p,
+				 struct amdgpu_bo *bo)
+{
+	u64 initial_bytes_moved;
+	uint32_t domain;
+	int r;
+
+	if (bo->pin_count)
+		return 0;
+
+	/* Don't move this buffer if we have depleted our allowance
+	 * to move it. Don't move anything if the threshold is zero.
 	 */
+	if (p->bytes_moved < p->bytes_moved_threshold)
+		domain = bo->prefered_domains;
+	else
+		domain = bo->allowed_domains;
+
+retry:
+	amdgpu_ttm_placement_from_domain(bo, domain);
+	initial_bytes_moved = atomic64_read(&bo->adev->num_bytes_moved);
+	r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+	p->bytes_moved += atomic64_read(&bo->adev->num_bytes_moved) -
+		initial_bytes_moved;
+
+	if (unlikely(r == -ENOMEM) && domain != bo->allowed_domains) {
+		domain = bo->allowed_domains;
+		goto retry;
+	}
 
-	u64 half_vram = real_vram_size >> 1;
-	u64 half_free_vram = vram_usage >= half_vram ? 0 : half_vram - vram_usage;
-	u64 bytes_moved_threshold = half_free_vram >> 1;
-	return max(bytes_moved_threshold, 1024*1024ull);
+	return r;
 }
 
-int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
+/* Last resort, try to evict something from the current working set */
+static bool amdgpu_cs_try_evict(struct amdgpu_cs_parser *p,
+				struct amdgpu_bo_list_entry *lobj)
+{
+	uint32_t domain = lobj->robj->allowed_domains;
+	int r;
+
+	if (!p->evictable)
+		return false;
+
+	for (;&p->evictable->tv.head != &p->validated;
+	     p->evictable = list_prev_entry(p->evictable, tv.head)) {
+
+		struct amdgpu_bo_list_entry *candidate = p->evictable;
+		struct amdgpu_bo *bo = candidate->robj;
+		u64 initial_bytes_moved;
+		uint32_t other;
+
+		/* If we reached our current BO we can forget it */
+		if (candidate == lobj)
+			break;
+
+		other = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type);
+
+		/* Check if this BO is in one of the domains we need space for */
+		if (!(other & domain))
+			continue;
+
+		/* Check if we can move this BO somewhere else */
+		other = bo->allowed_domains & ~domain;
+		if (!other)
+			continue;
+
+		/* Good we can try to move this BO somewhere else */
+		amdgpu_ttm_placement_from_domain(bo, other);
+		initial_bytes_moved = atomic64_read(&bo->adev->num_bytes_moved);
+		r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
+		p->bytes_moved += atomic64_read(&bo->adev->num_bytes_moved) -
+			initial_bytes_moved;
+
+		if (unlikely(r))
+			break;
+
+		p->evictable = list_prev_entry(p->evictable, tv.head);
+		list_move(&candidate->tv.head, &p->validated);
+
+		return true;
+	}
+
+	return false;
+}
+
+static int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
 			    struct list_head *validated)
 {
 	struct amdgpu_bo_list_entry *lobj;
-	u64 initial_bytes_moved;
 	int r;
 
 	list_for_each_entry(lobj, validated, tv.head) {
 		struct amdgpu_bo *bo = lobj->robj;
 		bool binding_userptr = false;
 		struct mm_struct *usermm;
-		uint32_t domain;
 
 		usermm = amdgpu_ttm_tt_get_usermm(bo->tbo.ttm);
 		if (usermm && usermm != current->mm)
@@ -313,35 +461,19 @@ int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p,
 			binding_userptr = true;
 		}
 
-		if (bo->pin_count)
-			continue;
-
-		/* Avoid moving this one if we have moved too many buffers
-		 * for this IB already.
-		 *
-		 * Note that this allows moving at least one buffer of
-		 * any size, because it doesn't take the current "bo"
-		 * into account. We don't want to disallow buffer moves
-		 * completely.
-		 */
-		if (p->bytes_moved <= p->bytes_moved_threshold)
-			domain = bo->prefered_domains;
-		else
-			domain = bo->allowed_domains;
-
-	retry:
-		amdgpu_ttm_placement_from_domain(bo, domain);
-		initial_bytes_moved = atomic64_read(&bo->adev->num_bytes_moved);
-		r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false);
-		p->bytes_moved += atomic64_read(&bo->adev->num_bytes_moved) -
-			       initial_bytes_moved;
+		if (p->evictable == lobj)
+			p->evictable = NULL;
 
-		if (unlikely(r)) {
-			if (r != -ERESTARTSYS && domain != bo->allowed_domains) {
-				domain = bo->allowed_domains;
-				goto retry;
-			}
+		do {
+			r = amdgpu_cs_bo_validate(p, bo);
+		} while (r == -ENOMEM && amdgpu_cs_try_evict(p, lobj));
+		if (r)
 			return r;
+
+		if (bo->shadow) {
+			r = amdgpu_cs_bo_validate(p, bo);
+			if (r)
+				return r;
 		}
 
 		if (binding_userptr) {
@@ -386,8 +518,10 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 
 		r = ttm_eu_reserve_buffers(&p->ticket, &p->validated, true,
 					   &duplicates);
-		if (unlikely(r != 0))
+		if (unlikely(r != 0)) {
+			DRM_ERROR("ttm_eu_reserve_buffers failed.\n");
 			goto error_free_pages;
+		}
 
 		/* Without a BO list we don't have userptr BOs */
 		if (!p->bo_list)
@@ -427,9 +561,10 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 		/* Unreserve everything again. */
 		ttm_eu_backoff_reservation(&p->ticket, &p->validated);
 
-		/* We tried to often, just abort */
+		/* We tried too many times, just abort */
 		if (!--tries) {
 			r = -EDEADLK;
+			DRM_ERROR("deadlock in %s\n", __func__);
 			goto error_free_pages;
 		}
 
@@ -441,11 +576,13 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 							 sizeof(struct page*));
 			if (!e->user_pages) {
 				r = -ENOMEM;
+				DRM_ERROR("calloc failure in %s\n", __func__);
 				goto error_free_pages;
 			}
 
 			r = amdgpu_ttm_tt_get_user_pages(ttm, e->user_pages);
 			if (r) {
+				DRM_ERROR("amdgpu_ttm_tt_get_user_pages failed.\n");
 				drm_free_large(e->user_pages);
 				e->user_pages = NULL;
 				goto error_free_pages;
@@ -460,14 +597,23 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 
 	p->bytes_moved_threshold = amdgpu_cs_get_threshold_for_moves(p->adev);
 	p->bytes_moved = 0;
+	p->evictable = list_last_entry(&p->validated,
+				       struct amdgpu_bo_list_entry,
+				       tv.head);
 
 	r = amdgpu_cs_list_validate(p, &duplicates);
-	if (r)
+	if (r) {
+		DRM_ERROR("amdgpu_cs_list_validate(duplicates) failed.\n");
 		goto error_validate;
+	}
 
 	r = amdgpu_cs_list_validate(p, &p->validated);
-	if (r)
+	if (r) {
+		DRM_ERROR("amdgpu_cs_list_validate(validated) failed.\n");
 		goto error_validate;
+	}
+
+	amdgpu_cs_report_moved_bytes(p->adev, p->bytes_moved);
 
 	fpriv->vm.last_eviction_counter =
 		atomic64_read(&p->adev->num_evictions);
@@ -499,8 +645,12 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 		}
 	}
 
-	if (p->uf_entry.robj)
-		p->job->uf_addr += amdgpu_bo_gpu_offset(p->uf_entry.robj);
+	if (!r && p->uf_entry.robj) {
+		struct amdgpu_bo *uf = p->uf_entry.robj;
+
+		r = amdgpu_ttm_bind(&uf->tbo, &uf->tbo.mem);
+		p->job->uf_addr += amdgpu_bo_gpu_offset(uf);
+	}
 
 error_validate:
 	if (r) {
@@ -617,7 +767,7 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p,
 			if (bo_va == NULL)
 				continue;
 
-			r = amdgpu_vm_bo_update(adev, bo_va, &bo->tbo.mem);
+			r = amdgpu_vm_bo_update(adev, bo_va, false);
 			if (r)
 				return r;
 
@@ -710,6 +860,14 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev,
 		if (r)
 			return r;
 
+		if (ib->flags & AMDGPU_IB_FLAG_PREAMBLE) {
+			parser->job->preamble_status |= AMDGPU_PREAMBLE_IB_PRESENT;
+			if (!parser->ctx->preamble_presented) {
+				parser->job->preamble_status |= AMDGPU_PREAMBLE_IB_PRESENT_FIRST;
+				parser->ctx->preamble_presented = true;
+			}
+		}
+
 		if (parser->job->ring && parser->job->ring != ring)
 			return -EINVAL;
 
@@ -849,7 +1007,7 @@ static int amdgpu_cs_submit(struct amdgpu_cs_parser *p,
 	}
 
 	job->owner = p->filp;
-	job->ctx = entity->fence_context;
+	job->fence_ctx = entity->fence_context;
 	p->fence = fence_get(&job->base.s_fence->finished);
 	cs->out.handle = amdgpu_ctx_add_fence(p->ctx, ring, p->fence);
 	job->uf_sequence = cs->out.handle;
@@ -1015,3 +1173,29 @@ amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser,
 
 	return NULL;
 }
+
+/**
+ * amdgpu_cs_sysvm_access_required - make BOs accessible by the system VM
+ *
+ * @parser: command submission parser context
+ *
+ * Helper for UVD/VCE VM emulation, make sure BOs are accessible by the system VM.
+ */
+int amdgpu_cs_sysvm_access_required(struct amdgpu_cs_parser *parser)
+{
+	unsigned i;
+	int r;
+
+	if (!parser->bo_list)
+		return 0;
+
+	for (i = 0; i < parser->bo_list->num_entries; i++) {
+		struct amdgpu_bo *bo = parser->bo_list->array[i].robj;
+
+		r = amdgpu_ttm_bind(&bo->tbo, &bo->tbo.mem);
+		if (unlikely(r))
+			return r;
+	}
+
+	return 0;
+}

drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c

@@ -60,6 +60,7 @@ static int amdgpu_ctx_init(struct amdgpu_device *adev, struct amdgpu_ctx *ctx)
 			amd_sched_entity_fini(&adev->rings[j]->sched,
 					      &ctx->rings[j].entity);
 		kfree(ctx->fences);
+		ctx->fences = NULL;
 		return r;
 	}
 	return 0;
@@ -77,6 +78,7 @@ static void amdgpu_ctx_fini(struct amdgpu_ctx *ctx)
 		for (j = 0; j < amdgpu_sched_jobs; ++j)
 			fence_put(ctx->rings[i].fences[j]);
 	kfree(ctx->fences);
+	ctx->fences = NULL;
 
 	for (i = 0; i < adev->num_rings; i++)
 		amd_sched_entity_fini(&adev->rings[i]->sched,

drivers/gpu/drm/amd/amdgpu/amdgpu_device.c

@@ -41,16 +41,26 @@
 #include "atom.h"
 #include "amdgpu_atombios.h"
 #include "amd_pcie.h"
+#ifdef CONFIG_DRM_AMDGPU_SI
+#include "si.h"
+#endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 #include "cik.h"
 #endif
 #include "vi.h"
 #include "bif/bif_4_1_d.h"
+#include <linux/pci.h>
+#include <linux/firmware.h>
 
 static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev);
 static void amdgpu_debugfs_regs_cleanup(struct amdgpu_device *adev);
 
 static const char *amdgpu_asic_name[] = {
+	"TAHITI",
+	"PITCAIRN",
+	"VERDE",
+	"OLAND",
+	"HAINAN",
 	"BONAIRE",
 	"KAVERI",
 	"KABINI",
@@ -101,7 +111,7 @@ void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
 		    bool always_indirect)
 {
 	trace_amdgpu_mm_wreg(adev->pdev->device, reg, v);
-	
+
 	if ((reg * 4) < adev->rmmio_size && !always_indirect)
 		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
 	else {
@@ -642,6 +652,46 @@ bool amdgpu_card_posted(struct amdgpu_device *adev)
 
 }
 
+static bool amdgpu_vpost_needed(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev))
+		return false;
+
+	if (amdgpu_passthrough(adev)) {
+		/* for FIJI: In whole GPU pass-through virtualization case
+		 * old smc fw won't clear some registers (e.g. MEM_SIZE, BIOS_SCRATCH)
+		 * so amdgpu_card_posted return false and driver will incorrectly skip vPost.
+		 * but if we force vPost do in pass-through case, the driver reload will hang.
+		 * whether doing vPost depends on amdgpu_card_posted if smc version is above
+		 * 00160e00 for FIJI.
+		 */
+		if (adev->asic_type == CHIP_FIJI) {
+			int err;
+			uint32_t fw_ver;
+			err = request_firmware(&adev->pm.fw, "amdgpu/fiji_smc.bin", adev->dev);
+			/* force vPost if error occured */
+			if (err)
+				return true;
+
+			fw_ver = *((uint32_t *)adev->pm.fw->data + 69);
+			if (fw_ver >= 0x00160e00)
+				return !amdgpu_card_posted(adev);
+		}
+	} else {
+		/* in bare-metal case, amdgpu_card_posted return false
+		 * after system reboot/boot, and return true if driver
+		 * reloaded.
+		 * we shouldn't do vPost after driver reload otherwise GPU
+		 * could hang.
+		 */
+		if (amdgpu_card_posted(adev))
+			return false;
+	}
+
+	/* we assume vPost is neede for all other cases */
+	return true;
+}
+
 /**
  * amdgpu_dummy_page_init - init dummy page used by the driver
  *
@@ -1026,7 +1076,7 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero
 		/* don't suspend or resume card normally */
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
-		amdgpu_resume_kms(dev, true, true);
+		amdgpu_device_resume(dev, true, true);
 
 		dev->pdev->d3_delay = d3_delay;
 
@@ -1036,7 +1086,7 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero
 		printk(KERN_INFO "amdgpu: switched off\n");
 		drm_kms_helper_poll_disable(dev);
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
-		amdgpu_suspend_kms(dev, true, true);
+		amdgpu_device_suspend(dev, true, true);
 		dev->switch_power_state = DRM_SWITCH_POWER_OFF;
 	}
 }
@@ -1181,10 +1231,38 @@ int amdgpu_ip_block_version_cmp(struct amdgpu_device *adev,
 	return 1;
 }
 
+static void amdgpu_whether_enable_virtual_display(struct amdgpu_device *adev)
+{
+	adev->enable_virtual_display = false;
+
+	if (amdgpu_virtual_display) {
+		struct drm_device *ddev = adev->ddev;
+		const char *pci_address_name = pci_name(ddev->pdev);
+		char *pciaddstr, *pciaddstr_tmp, *pciaddname;
+
+		pciaddstr = kstrdup(amdgpu_virtual_display, GFP_KERNEL);
+		pciaddstr_tmp = pciaddstr;
+		while ((pciaddname = strsep(&pciaddstr_tmp, ";"))) {
+			if (!strcmp(pci_address_name, pciaddname)) {
+				adev->enable_virtual_display = true;
+				break;
+			}
+		}
+
+		DRM_INFO("virtual display string:%s, %s:virtual_display:%d\n",
+				 amdgpu_virtual_display, pci_address_name,
+				 adev->enable_virtual_display);
+
+		kfree(pciaddstr);
+	}
+}
+
 static int amdgpu_early_init(struct amdgpu_device *adev)
 {
 	int i, r;
 
+	amdgpu_whether_enable_virtual_display(adev);
+
 	switch (adev->asic_type) {
 	case CHIP_TOPAZ:
 	case CHIP_TONGA:
@@ -1202,6 +1280,18 @@ static int amdgpu_early_init(struct amdgpu_device *adev)
 		if (r)
 			return r;
 		break;
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_VERDE:
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_OLAND:
+	case CHIP_HAINAN:
+		adev->family = AMDGPU_FAMILY_SI;
+		r = si_set_ip_blocks(adev);
+		if (r)
+			return r;
+		break;
+#endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	case CHIP_BONAIRE:
 	case CHIP_HAWAII:
@@ -1318,6 +1408,9 @@ static int amdgpu_late_init(struct amdgpu_device *adev)
 	for (i = 0; i < adev->num_ip_blocks; i++) {
 		if (!adev->ip_block_status[i].valid)
 			continue;
+		if (adev->ip_blocks[i].type == AMD_IP_BLOCK_TYPE_UVD ||
+			adev->ip_blocks[i].type == AMD_IP_BLOCK_TYPE_VCE)
+			continue;
 		/* enable clockgating to save power */
 		r = adev->ip_blocks[i].funcs->set_clockgating_state((void *)adev,
 								    AMD_CG_STATE_GATE);
@@ -1331,6 +1424,7 @@ static int amdgpu_late_init(struct amdgpu_device *adev)
 				DRM_ERROR("late_init of IP block <%s> failed %d\n", adev->ip_blocks[i].funcs->name, r);
 				return r;
 			}
+			adev->ip_block_status[i].late_initialized = true;
 		}
 	}
 
@@ -1376,8 +1470,11 @@ static int amdgpu_fini(struct amdgpu_device *adev)
 	}
 
 	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+		if (!adev->ip_block_status[i].late_initialized)
+			continue;
 		if (adev->ip_blocks[i].funcs->late_fini)
 			adev->ip_blocks[i].funcs->late_fini((void *)adev);
+		adev->ip_block_status[i].late_initialized = false;
 	}
 
 	return 0;
@@ -1433,13 +1530,10 @@ static int amdgpu_resume(struct amdgpu_device *adev)
 	return 0;
 }
 
-static bool amdgpu_device_is_virtual(void)
+static void amdgpu_device_detect_sriov_bios(struct amdgpu_device *adev)
 {
-#ifdef CONFIG_X86
-	return boot_cpu_has(X86_FEATURE_HYPERVISOR);
-#else
-	return false;
-#endif
+	if (amdgpu_atombios_has_gpu_virtualization_table(adev))
+		adev->virtualization.virtual_caps |= AMDGPU_SRIOV_CAPS_SRIOV_VBIOS;
 }
 
 /**
@@ -1461,6 +1555,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 {
 	int r, i;
 	bool runtime = false;
+	u32 max_MBps;
 
 	adev->shutdown = false;
 	adev->dev = &pdev->dev;
@@ -1484,6 +1579,8 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	adev->smc_wreg = &amdgpu_invalid_wreg;
 	adev->pcie_rreg = &amdgpu_invalid_rreg;
 	adev->pcie_wreg = &amdgpu_invalid_wreg;
+	adev->pciep_rreg = &amdgpu_invalid_rreg;
+	adev->pciep_wreg = &amdgpu_invalid_wreg;
 	adev->uvd_ctx_rreg = &amdgpu_invalid_rreg;
 	adev->uvd_ctx_wreg = &amdgpu_invalid_wreg;
 	adev->didt_rreg = &amdgpu_invalid_rreg;
@@ -1520,9 +1617,22 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	spin_lock_init(&adev->didt_idx_lock);
 	spin_lock_init(&adev->gc_cac_idx_lock);
 	spin_lock_init(&adev->audio_endpt_idx_lock);
+	spin_lock_init(&adev->mm_stats.lock);
+
+	INIT_LIST_HEAD(&adev->shadow_list);
+	mutex_init(&adev->shadow_list_lock);
+
+	INIT_LIST_HEAD(&adev->gtt_list);
+	spin_lock_init(&adev->gtt_list_lock);
+
+	if (adev->asic_type >= CHIP_BONAIRE) {
+		adev->rmmio_base = pci_resource_start(adev->pdev, 5);
+		adev->rmmio_size = pci_resource_len(adev->pdev, 5);
+	} else {
+		adev->rmmio_base = pci_resource_start(adev->pdev, 2);
+		adev->rmmio_size = pci_resource_len(adev->pdev, 2);
+	}
 
-	adev->rmmio_base = pci_resource_start(adev->pdev, 5);
-	adev->rmmio_size = pci_resource_len(adev->pdev, 5);
 	adev->rmmio = ioremap(adev->rmmio_base, adev->rmmio_size);
 	if (adev->rmmio == NULL) {
 		return -ENOMEM;
@@ -1530,8 +1640,9 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)adev->rmmio_base);
 	DRM_INFO("register mmio size: %u\n", (unsigned)adev->rmmio_size);
 
-	/* doorbell bar mapping */
-	amdgpu_doorbell_init(adev);
+	if (adev->asic_type >= CHIP_BONAIRE)
+		/* doorbell bar mapping */
+		amdgpu_doorbell_init(adev);
 
 	/* io port mapping */
 	for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) {
@@ -1579,25 +1690,24 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 		goto failed;
 	}
 
-	/* See if the asic supports SR-IOV */
-	adev->virtualization.supports_sr_iov =
-		amdgpu_atombios_has_gpu_virtualization_table(adev);
-
-	/* Check if we are executing in a virtualized environment */
-	adev->virtualization.is_virtual = amdgpu_device_is_virtual();
-	adev->virtualization.caps = amdgpu_asic_get_virtual_caps(adev);
+	/* detect if we are with an SRIOV vbios */
+	amdgpu_device_detect_sriov_bios(adev);
 
 	/* Post card if necessary */
-	if (!amdgpu_card_posted(adev) ||
-	    (adev->virtualization.is_virtual &&
-	     !(adev->virtualization.caps & AMDGPU_VIRT_CAPS_SRIOV_EN))) {
+	if (amdgpu_vpost_needed(adev)) {
 		if (!adev->bios) {
-			dev_err(adev->dev, "Card not posted and no BIOS - ignoring\n");
+			dev_err(adev->dev, "no vBIOS found\n");
 			r = -EINVAL;
 			goto failed;
 		}
-		DRM_INFO("GPU not posted. posting now...\n");
-		amdgpu_atom_asic_init(adev->mode_info.atom_context);
+		DRM_INFO("GPU posting now...\n");
+		r = amdgpu_atom_asic_init(adev->mode_info.atom_context);
+		if (r) {
+			dev_err(adev->dev, "gpu post error!\n");
+			goto failed;
+		}
+	} else {
+		DRM_INFO("GPU post is not needed\n");
 	}
 
 	/* Initialize clocks */
@@ -1628,6 +1738,14 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 
 	adev->accel_working = true;
 
+	/* Initialize the buffer migration limit. */
+	if (amdgpu_moverate >= 0)
+		max_MBps = amdgpu_moverate;
+	else
+		max_MBps = 8; /* Allow 8 MB/s. */
+	/* Get a log2 for easy divisions. */
+	adev->mm_stats.log2_max_MBps = ilog2(max(1u, max_MBps));
+
 	amdgpu_fbdev_init(adev);
 
 	r = amdgpu_ib_pool_init(adev);
@@ -1732,7 +1850,8 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
 	adev->rio_mem = NULL;
 	iounmap(adev->rmmio);
 	adev->rmmio = NULL;
-	amdgpu_doorbell_fini(adev);
+	if (adev->asic_type >= CHIP_BONAIRE)
+		amdgpu_doorbell_fini(adev);
 	amdgpu_debugfs_regs_cleanup(adev);
 	amdgpu_debugfs_remove_files(adev);
 }
@@ -1742,7 +1861,7 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
  * Suspend & resume.
  */
 /**
- * amdgpu_suspend_kms - initiate device suspend
+ * amdgpu_device_suspend - initiate device suspend
  *
  * @pdev: drm dev pointer
  * @state: suspend state
@@ -1751,7 +1870,7 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
  * Returns 0 for success or an error on failure.
  * Called at driver suspend.
  */
-int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
+int amdgpu_device_suspend(struct drm_device *dev, bool suspend, bool fbcon)
 {
 	struct amdgpu_device *adev;
 	struct drm_crtc *crtc;
@@ -1819,6 +1938,10 @@ int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
 		/* Shut down the device */
 		pci_disable_device(dev->pdev);
 		pci_set_power_state(dev->pdev, PCI_D3hot);
+	} else {
+		r = amdgpu_asic_reset(adev);
+		if (r)
+			DRM_ERROR("amdgpu asic reset failed\n");
 	}
 
 	if (fbcon) {
@@ -1830,7 +1953,7 @@ int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
 }
 
 /**
- * amdgpu_resume_kms - initiate device resume
+ * amdgpu_device_resume - initiate device resume
  *
  * @pdev: drm dev pointer
  *
@@ -1838,7 +1961,7 @@ int amdgpu_suspend_kms(struct drm_device *dev, bool suspend, bool fbcon)
  * Returns 0 for success or an error on failure.
  * Called at driver resume.
  */
-int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon)
+int amdgpu_device_resume(struct drm_device *dev, bool resume, bool fbcon)
 {
 	struct drm_connector *connector;
 	struct amdgpu_device *adev = dev->dev_private;
@@ -1848,22 +1971,26 @@ int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon)
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
-	if (fbcon) {
+	if (fbcon)
 		console_lock();
-	}
+
 	if (resume) {
 		pci_set_power_state(dev->pdev, PCI_D0);
 		pci_restore_state(dev->pdev);
-		if (pci_enable_device(dev->pdev)) {
+		r = pci_enable_device(dev->pdev);
+		if (r) {
 			if (fbcon)
 				console_unlock();
-			return -1;
+			return r;
 		}
 	}
 
 	/* post card */
-	if (!amdgpu_card_posted(adev))
-		amdgpu_atom_asic_init(adev->mode_info.atom_context);
+	if (!amdgpu_card_posted(adev) || !resume) {
+		r = amdgpu_atom_asic_init(adev->mode_info.atom_context);
+		if (r)
+			DRM_ERROR("amdgpu asic init failed\n");
+	}
 
 	r = amdgpu_resume(adev);
 	if (r)
@@ -1937,6 +2064,126 @@ int amdgpu_resume_kms(struct drm_device *dev, bool resume, bool fbcon)
 	return 0;
 }
 
+static bool amdgpu_check_soft_reset(struct amdgpu_device *adev)
+{
+	int i;
+	bool asic_hang = false;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_block_status[i].valid)
+			continue;
+		if (adev->ip_blocks[i].funcs->check_soft_reset)
+			adev->ip_blocks[i].funcs->check_soft_reset(adev);
+		if (adev->ip_block_status[i].hang) {
+			DRM_INFO("IP block:%d is hang!\n", i);
+			asic_hang = true;
+		}
+	}
+	return asic_hang;
+}
+
+static int amdgpu_pre_soft_reset(struct amdgpu_device *adev)
+{
+	int i, r = 0;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_block_status[i].valid)
+			continue;
+		if (adev->ip_block_status[i].hang &&
+		    adev->ip_blocks[i].funcs->pre_soft_reset) {
+			r = adev->ip_blocks[i].funcs->pre_soft_reset(adev);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static bool amdgpu_need_full_reset(struct amdgpu_device *adev)
+{
+	if (adev->ip_block_status[AMD_IP_BLOCK_TYPE_GMC].hang ||
+	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_SMC].hang ||
+	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_ACP].hang ||
+	    adev->ip_block_status[AMD_IP_BLOCK_TYPE_DCE].hang) {
+		DRM_INFO("Some block need full reset!\n");
+		return true;
+	}
+	return false;
+}
+
+static int amdgpu_soft_reset(struct amdgpu_device *adev)
+{
+	int i, r = 0;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_block_status[i].valid)
+			continue;
+		if (adev->ip_block_status[i].hang &&
+		    adev->ip_blocks[i].funcs->soft_reset) {
+			r = adev->ip_blocks[i].funcs->soft_reset(adev);
+			if (r)
+				return r;
+		}
+	}
+
+	return 0;
+}
+
+static int amdgpu_post_soft_reset(struct amdgpu_device *adev)
+{
+	int i, r = 0;
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!adev->ip_block_status[i].valid)
+			continue;
+		if (adev->ip_block_status[i].hang &&
+		    adev->ip_blocks[i].funcs->post_soft_reset)
+			r = adev->ip_blocks[i].funcs->post_soft_reset(adev);
+		if (r)
+			return r;
+	}
+
+	return 0;
+}
+
+bool amdgpu_need_backup(struct amdgpu_device *adev)
+{
+	if (adev->flags & AMD_IS_APU)
+		return false;
+
+	return amdgpu_lockup_timeout > 0 ? true : false;
+}
+
+static int amdgpu_recover_vram_from_shadow(struct amdgpu_device *adev,
+					   struct amdgpu_ring *ring,
+					   struct amdgpu_bo *bo,
+					   struct fence **fence)
+{
+	uint32_t domain;
+	int r;
+
+       if (!bo->shadow)
+               return 0;
+
+       r = amdgpu_bo_reserve(bo, false);
+       if (r)
+               return r;
+       domain = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type);
+       /* if bo has been evicted, then no need to recover */
+       if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
+               r = amdgpu_bo_restore_from_shadow(adev, ring, bo,
+						 NULL, fence, true);
+               if (r) {
+                       DRM_ERROR("recover page table failed!\n");
+                       goto err;
+               }
+       }
+err:
+       amdgpu_bo_unreserve(bo);
+       return r;
+}
+
 /**
  * amdgpu_gpu_reset - reset the asic
  *
@@ -1949,6 +2196,12 @@ int amdgpu_gpu_reset(struct amdgpu_device *adev)
 {
 	int i, r;
 	int resched;
+	bool need_full_reset;
+
+	if (!amdgpu_check_soft_reset(adev)) {
+		DRM_INFO("No hardware hang detected. Did some blocks stall?\n");
+		return 0;
+	}
 
 	atomic_inc(&adev->gpu_reset_counter);
 
@@ -1967,40 +2220,93 @@ int amdgpu_gpu_reset(struct amdgpu_device *adev)
 	/* after all hw jobs are reset, hw fence is meaningless, so force_completion */
 	amdgpu_fence_driver_force_completion(adev);
 
-	/* save scratch */
-	amdgpu_atombios_scratch_regs_save(adev);
-	r = amdgpu_suspend(adev);
+	need_full_reset = amdgpu_need_full_reset(adev);
 
-retry:
-	/* Disable fb access */
-	if (adev->mode_info.num_crtc) {
-		struct amdgpu_mode_mc_save save;
-		amdgpu_display_stop_mc_access(adev, &save);
-		amdgpu_wait_for_idle(adev, AMD_IP_BLOCK_TYPE_GMC);
+	if (!need_full_reset) {
+		amdgpu_pre_soft_reset(adev);
+		r = amdgpu_soft_reset(adev);
+		amdgpu_post_soft_reset(adev);
+		if (r || amdgpu_check_soft_reset(adev)) {
+			DRM_INFO("soft reset failed, will fallback to full reset!\n");
+			need_full_reset = true;
+		}
 	}
 
-	r = amdgpu_asic_reset(adev);
-	/* post card */
-	amdgpu_atom_asic_init(adev->mode_info.atom_context);
+	if (need_full_reset) {
+		/* save scratch */
+		amdgpu_atombios_scratch_regs_save(adev);
+		r = amdgpu_suspend(adev);
 
-	if (!r) {
-		dev_info(adev->dev, "GPU reset succeeded, trying to resume\n");
-		r = amdgpu_resume(adev);
+retry:
+		/* Disable fb access */
+		if (adev->mode_info.num_crtc) {
+			struct amdgpu_mode_mc_save save;
+			amdgpu_display_stop_mc_access(adev, &save);
+			amdgpu_wait_for_idle(adev, AMD_IP_BLOCK_TYPE_GMC);
+		}
+
+		r = amdgpu_asic_reset(adev);
+		/* post card */
+		amdgpu_atom_asic_init(adev->mode_info.atom_context);
+
+		if (!r) {
+			dev_info(adev->dev, "GPU reset succeeded, trying to resume\n");
+			r = amdgpu_resume(adev);
+		}
+		/* restore scratch */
+		amdgpu_atombios_scratch_regs_restore(adev);
 	}
-	/* restore scratch */
-	amdgpu_atombios_scratch_regs_restore(adev);
 	if (!r) {
+		amdgpu_irq_gpu_reset_resume_helper(adev);
+		if (need_full_reset && amdgpu_need_backup(adev)) {
+			r = amdgpu_ttm_recover_gart(adev);
+			if (r)
+				DRM_ERROR("gart recovery failed!!!\n");
+		}
 		r = amdgpu_ib_ring_tests(adev);
 		if (r) {
 			dev_err(adev->dev, "ib ring test failed (%d).\n", r);
 			r = amdgpu_suspend(adev);
+			need_full_reset = true;
 			goto retry;
 		}
+		/**
+		 * recovery vm page tables, since we cannot depend on VRAM is
+		 * consistent after gpu full reset.
+		 */
+		if (need_full_reset && amdgpu_need_backup(adev)) {
+			struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
+			struct amdgpu_bo *bo, *tmp;
+			struct fence *fence = NULL, *next = NULL;
+
+			DRM_INFO("recover vram bo from shadow\n");
+			mutex_lock(&adev->shadow_list_lock);
+			list_for_each_entry_safe(bo, tmp, &adev->shadow_list, shadow_list) {
+				amdgpu_recover_vram_from_shadow(adev, ring, bo, &next);
+				if (fence) {
+					r = fence_wait(fence, false);
+					if (r) {
+						WARN(r, "recovery from shadow isn't comleted\n");
+						break;
+					}
+				}
 
+				fence_put(fence);
+				fence = next;
+			}
+			mutex_unlock(&adev->shadow_list_lock);
+			if (fence) {
+				r = fence_wait(fence, false);
+				if (r)
+					WARN(r, "recovery from shadow isn't comleted\n");
+			}
+			fence_put(fence);
+		}
 		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 			struct amdgpu_ring *ring = adev->rings[i];
 			if (!ring)
 				continue;
+
 			amd_sched_job_recovery(&ring->sched);
 			kthread_unpark(ring->sched.thread);
 		}
@@ -2020,7 +2326,6 @@ retry:
 		/* bad news, how to tell it to userspace ? */
 		dev_info(adev->dev, "GPU reset failed\n");
 	}
-	amdgpu_irq_gpu_reset_resume_helper(adev);
 
 	return r;
 }
@@ -2178,22 +2483,26 @@ static ssize_t amdgpu_debugfs_regs_read(struct file *f, char __user *buf,
 	struct amdgpu_device *adev = f->f_inode->i_private;
 	ssize_t result = 0;
 	int r;
-	bool use_bank;
+	bool pm_pg_lock, use_bank;
 	unsigned instance_bank, sh_bank, se_bank;
 
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
+	/* are we reading registers for which a PG lock is necessary? */
+	pm_pg_lock = (*pos >> 23) & 1;
+
 	if (*pos & (1ULL << 62)) {
 		se_bank = (*pos >> 24) & 0x3FF;
 		sh_bank = (*pos >> 34) & 0x3FF;
 		instance_bank = (*pos >> 44) & 0x3FF;
 		use_bank = 1;
-		*pos &= 0xFFFFFF;
 	} else {
 		use_bank = 0;
 	}
 
+	*pos &= 0x3FFFF;
+
 	if (use_bank) {
 		if (sh_bank >= adev->gfx.config.max_sh_per_se ||
 		    se_bank >= adev->gfx.config.max_shader_engines)
@@ -2203,6 +2512,9 @@ static ssize_t amdgpu_debugfs_regs_read(struct file *f, char __user *buf,
 					sh_bank, instance_bank);
 	}
 
+	if (pm_pg_lock)
+		mutex_lock(&adev->pm.mutex);
+
 	while (size) {
 		uint32_t value;
 
@@ -2228,6 +2540,9 @@ end:
 		mutex_unlock(&adev->grbm_idx_mutex);
 	}
 
+	if (pm_pg_lock)
+		mutex_unlock(&adev->pm.mutex);
+
 	return result;
 }
 
@@ -2385,7 +2700,7 @@ static ssize_t amdgpu_debugfs_regs_smc_read(struct file *f, char __user *buf,
 	while (size) {
 		uint32_t value;
 
-		value = RREG32_SMC(*pos >> 2);
+		value = RREG32_SMC(*pos);
 		r = put_user(value, (uint32_t *)buf);
 		if (r)
 			return r;
@@ -2416,7 +2731,7 @@ static ssize_t amdgpu_debugfs_regs_smc_write(struct file *f, const char __user *
 		if (r)
 			return r;
 
-		WREG32_SMC(*pos >> 2, value);
+		WREG32_SMC(*pos, value);
 
 		result += 4;
 		buf += 4;
@@ -2438,12 +2753,12 @@ static ssize_t amdgpu_debugfs_gca_config_read(struct file *f, char __user *buf,
 	if (size & 0x3 || *pos & 0x3)
 		return -EINVAL;
 
-	config = kmalloc(256 * sizeof(*config), GFP_KERNEL);
+	config = kmalloc_array(256, sizeof(*config), GFP_KERNEL);
 	if (!config)
 		return -ENOMEM;
 
 	/* version, increment each time something is added */
-	config[no_regs++] = 0;
+	config[no_regs++] = 2;
 	config[no_regs++] = adev->gfx.config.max_shader_engines;
 	config[no_regs++] = adev->gfx.config.max_tile_pipes;
 	config[no_regs++] = adev->gfx.config.max_cu_per_sh;
@@ -2468,6 +2783,15 @@ static ssize_t amdgpu_debugfs_gca_config_read(struct file *f, char __user *buf,
 	config[no_regs++] = adev->gfx.config.gb_addr_config;
 	config[no_regs++] = adev->gfx.config.num_rbs;
 
+	/* rev==1 */
+	config[no_regs++] = adev->rev_id;
+	config[no_regs++] = adev->pg_flags;
+	config[no_regs++] = adev->cg_flags;
+
+	/* rev==2 */
+	config[no_regs++] = adev->family;
+	config[no_regs++] = adev->external_rev_id;
+
 	while (size && (*pos < no_regs * 4)) {
 		uint32_t value;
 
@@ -2488,6 +2812,29 @@ static ssize_t amdgpu_debugfs_gca_config_read(struct file *f, char __user *buf,
 	return result;
 }
 
+static ssize_t amdgpu_debugfs_sensor_read(struct file *f, char __user *buf,
+					size_t size, loff_t *pos)
+{
+	struct amdgpu_device *adev = f->f_inode->i_private;
+	int idx, r;
+	int32_t value;
+
+	if (size != 4 || *pos & 0x3)
+		return -EINVAL;
+
+	/* convert offset to sensor number */
+	idx = *pos >> 2;
+
+	if (adev->powerplay.pp_funcs && adev->powerplay.pp_funcs->read_sensor)
+		r = adev->powerplay.pp_funcs->read_sensor(adev->powerplay.pp_handle, idx, &value);
+	else
+		return -EINVAL;
+
+	if (!r)
+		r = put_user(value, (int32_t *)buf);
+
+	return !r ? 4 : r;
+}
 
 static const struct file_operations amdgpu_debugfs_regs_fops = {
 	.owner = THIS_MODULE,
@@ -2520,12 +2867,19 @@ static const struct file_operations amdgpu_debugfs_gca_config_fops = {
 	.llseek = default_llseek
 };
 
+static const struct file_operations amdgpu_debugfs_sensors_fops = {
+	.owner = THIS_MODULE,
+	.read = amdgpu_debugfs_sensor_read,
+	.llseek = default_llseek
+};
+
 static const struct file_operations *debugfs_regs[] = {
 	&amdgpu_debugfs_regs_fops,
 	&amdgpu_debugfs_regs_didt_fops,
 	&amdgpu_debugfs_regs_pcie_fops,
 	&amdgpu_debugfs_regs_smc_fops,
 	&amdgpu_debugfs_gca_config_fops,
+	&amdgpu_debugfs_sensors_fops,
 };
 
 static const char *debugfs_regs_names[] = {
@@ -2534,6 +2888,7 @@ static const char *debugfs_regs_names[] = {
 	"amdgpu_regs_pcie",
 	"amdgpu_regs_smc",
 	"amdgpu_gca_config",
+	"amdgpu_sensors",
 };
 
 static int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)

drivers/gpu/drm/amd/amdgpu/amdgpu_display.c

@@ -41,7 +41,7 @@ static void amdgpu_flip_callback(struct fence *f, struct fence_cb *cb)
 		container_of(cb, struct amdgpu_flip_work, cb);
 
 	fence_put(f);
-	schedule_work(&work->flip_work);
+	schedule_work(&work->flip_work.work);
 }
 
 static bool amdgpu_flip_handle_fence(struct amdgpu_flip_work *work,
@@ -63,16 +63,17 @@ static bool amdgpu_flip_handle_fence(struct amdgpu_flip_work *work,
 
 static void amdgpu_flip_work_func(struct work_struct *__work)
 {
+	struct delayed_work *delayed_work =
+		container_of(__work, struct delayed_work, work);
 	struct amdgpu_flip_work *work =
-		container_of(__work, struct amdgpu_flip_work, flip_work);
+		container_of(delayed_work, struct amdgpu_flip_work, flip_work);
 	struct amdgpu_device *adev = work->adev;
 	struct amdgpu_crtc *amdgpuCrtc = adev->mode_info.crtcs[work->crtc_id];
 
 	struct drm_crtc *crtc = &amdgpuCrtc->base;
 	unsigned long flags;
-	unsigned i, repcnt = 4;
-	int vpos, hpos, stat, min_udelay = 0;
-	struct drm_vblank_crtc *vblank = &crtc->dev->vblank[work->crtc_id];
+	unsigned i;
+	int vpos, hpos;
 
 	if (amdgpu_flip_handle_fence(work, &work->excl))
 		return;
@@ -81,55 +82,23 @@ static void amdgpu_flip_work_func(struct work_struct *__work)
 		if (amdgpu_flip_handle_fence(work, &work->shared[i]))
 			return;
 
-	/* We borrow the event spin lock for protecting flip_status */
-	spin_lock_irqsave(&crtc->dev->event_lock, flags);
-
-	/* If this happens to execute within the "virtually extended" vblank
-	 * interval before the start of the real vblank interval then it needs
-	 * to delay programming the mmio flip until the real vblank is entered.
-	 * This prevents completing a flip too early due to the way we fudge
-	 * our vblank counter and vblank timestamps in order to work around the
-	 * problem that the hw fires vblank interrupts before actual start of
-	 * vblank (when line buffer refilling is done for a frame). It
-	 * complements the fudging logic in amdgpu_get_crtc_scanoutpos() for
-	 * timestamping and amdgpu_get_vblank_counter_kms() for vblank counts.
-	 *
-	 * In practice this won't execute very often unless on very fast
-	 * machines because the time window for this to happen is very small.
+	/* Wait until we're out of the vertical blank period before the one
+	 * targeted by the flip
 	 */
-	while (amdgpuCrtc->enabled && --repcnt) {
-		/* GET_DISTANCE_TO_VBLANKSTART returns distance to real vblank
-		 * start in hpos, and to the "fudged earlier" vblank start in
-		 * vpos.
-		 */
-		stat = amdgpu_get_crtc_scanoutpos(adev->ddev, work->crtc_id,
-						  GET_DISTANCE_TO_VBLANKSTART,
-						  &vpos, &hpos, NULL, NULL,
-						  &crtc->hwmode);
-
-		if ((stat & (DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_ACCURATE)) !=
-		    (DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_ACCURATE) ||
-		    !(vpos >= 0 && hpos <= 0))
-			break;
-
-		/* Sleep at least until estimated real start of hw vblank */
-		min_udelay = (-hpos + 1) * max(vblank->linedur_ns / 1000, 5);
-		if (min_udelay > vblank->framedur_ns / 2000) {
-			/* Don't wait ridiculously long - something is wrong */
-			repcnt = 0;
-			break;
-		}
-		spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
-		usleep_range(min_udelay, 2 * min_udelay);
-		spin_lock_irqsave(&crtc->dev->event_lock, flags);
+	if (amdgpuCrtc->enabled &&
+	    (amdgpu_get_crtc_scanoutpos(adev->ddev, work->crtc_id, 0,
+					&vpos, &hpos, NULL, NULL,
+					&crtc->hwmode)
+	     & (DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_IN_VBLANK)) ==
+	    (DRM_SCANOUTPOS_VALID | DRM_SCANOUTPOS_IN_VBLANK) &&
+	    (int)(work->target_vblank -
+		  amdgpu_get_vblank_counter_kms(adev->ddev, amdgpuCrtc->crtc_id)) > 0) {
+		schedule_delayed_work(&work->flip_work, usecs_to_jiffies(1000));
+		return;
 	}
 
-	if (!repcnt)
-		DRM_DEBUG_DRIVER("Delay problem on crtc %d: min_udelay %d, "
-				 "framedur %d, linedur %d, stat %d, vpos %d, "
-				 "hpos %d\n", work->crtc_id, min_udelay,
-				 vblank->framedur_ns / 1000,
-				 vblank->linedur_ns / 1000, stat, vpos, hpos);
+	/* We borrow the event spin lock for protecting flip_status */
+	spin_lock_irqsave(&crtc->dev->event_lock, flags);
 
 	/* Do the flip (mmio) */
 	adev->mode_info.funcs->page_flip(adev, work->crtc_id, work->base, work->async);
@@ -154,25 +123,25 @@ static void amdgpu_unpin_work_func(struct work_struct *__work)
 	int r;
 
 	/* unpin of the old buffer */
-	r = amdgpu_bo_reserve(work->old_rbo, false);
+	r = amdgpu_bo_reserve(work->old_abo, false);
 	if (likely(r == 0)) {
-		r = amdgpu_bo_unpin(work->old_rbo);
+		r = amdgpu_bo_unpin(work->old_abo);
 		if (unlikely(r != 0)) {
 			DRM_ERROR("failed to unpin buffer after flip\n");
 		}
-		amdgpu_bo_unreserve(work->old_rbo);
+		amdgpu_bo_unreserve(work->old_abo);
 	} else
 		DRM_ERROR("failed to reserve buffer after flip\n");
 
-	amdgpu_bo_unref(&work->old_rbo);
+	amdgpu_bo_unref(&work->old_abo);
 	kfree(work->shared);
 	kfree(work);
 }
 
-int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
-			  struct drm_framebuffer *fb,
-			  struct drm_pending_vblank_event *event,
-			  uint32_t page_flip_flags)
+int amdgpu_crtc_page_flip_target(struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_pending_vblank_event *event,
+				 uint32_t page_flip_flags, uint32_t target)
 {
 	struct drm_device *dev = crtc->dev;
 	struct amdgpu_device *adev = dev->dev_private;
@@ -181,7 +150,7 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
 	struct amdgpu_framebuffer *new_amdgpu_fb;
 	struct drm_gem_object *obj;
 	struct amdgpu_flip_work *work;
-	struct amdgpu_bo *new_rbo;
+	struct amdgpu_bo *new_abo;
 	unsigned long flags;
 	u64 tiling_flags;
 	u64 base;
@@ -191,7 +160,7 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
 	if (work == NULL)
 		return -ENOMEM;
 
-	INIT_WORK(&work->flip_work, amdgpu_flip_work_func);
+	INIT_DELAYED_WORK(&work->flip_work, amdgpu_flip_work_func);
 	INIT_WORK(&work->unpin_work, amdgpu_unpin_work_func);
 
 	work->event = event;
@@ -204,28 +173,28 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
 	obj = old_amdgpu_fb->obj;
 
 	/* take a reference to the old object */
-	work->old_rbo = gem_to_amdgpu_bo(obj);
-	amdgpu_bo_ref(work->old_rbo);
+	work->old_abo = gem_to_amdgpu_bo(obj);
+	amdgpu_bo_ref(work->old_abo);
 
 	new_amdgpu_fb = to_amdgpu_framebuffer(fb);
 	obj = new_amdgpu_fb->obj;
-	new_rbo = gem_to_amdgpu_bo(obj);
+	new_abo = gem_to_amdgpu_bo(obj);
 
 	/* pin the new buffer */
-	r = amdgpu_bo_reserve(new_rbo, false);
+	r = amdgpu_bo_reserve(new_abo, false);
 	if (unlikely(r != 0)) {
-		DRM_ERROR("failed to reserve new rbo buffer before flip\n");
+		DRM_ERROR("failed to reserve new abo buffer before flip\n");
 		goto cleanup;
 	}
 
-	r = amdgpu_bo_pin_restricted(new_rbo, AMDGPU_GEM_DOMAIN_VRAM, 0, 0, &base);
+	r = amdgpu_bo_pin_restricted(new_abo, AMDGPU_GEM_DOMAIN_VRAM, 0, 0, &base);
 	if (unlikely(r != 0)) {
 		r = -EINVAL;
-		DRM_ERROR("failed to pin new rbo buffer before flip\n");
+		DRM_ERROR("failed to pin new abo buffer before flip\n");
 		goto unreserve;
 	}
 
-	r = reservation_object_get_fences_rcu(new_rbo->tbo.resv, &work->excl,
+	r = reservation_object_get_fences_rcu(new_abo->tbo.resv, &work->excl,
 					      &work->shared_count,
 					      &work->shared);
 	if (unlikely(r != 0)) {
@@ -233,16 +202,12 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
 		goto unpin;
 	}
 
-	amdgpu_bo_get_tiling_flags(new_rbo, &tiling_flags);
-	amdgpu_bo_unreserve(new_rbo);
+	amdgpu_bo_get_tiling_flags(new_abo, &tiling_flags);
+	amdgpu_bo_unreserve(new_abo);
 
 	work->base = base;
-
-	r = drm_crtc_vblank_get(crtc);
-	if (r) {
-		DRM_ERROR("failed to get vblank before flip\n");
-		goto pflip_cleanup;
-	}
+	work->target_vblank = target - drm_crtc_vblank_count(crtc) +
+		amdgpu_get_vblank_counter_kms(dev, work->crtc_id);
 
 	/* we borrow the event spin lock for protecting flip_wrok */
 	spin_lock_irqsave(&crtc->dev->event_lock, flags);
@@ -250,7 +215,7 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
 		DRM_DEBUG_DRIVER("flip queue: crtc already busy\n");
 		spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
 		r = -EBUSY;
-		goto vblank_cleanup;
+		goto pflip_cleanup;
 	}
 
 	amdgpu_crtc->pflip_status = AMDGPU_FLIP_PENDING;
@@ -262,26 +227,23 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
 	/* update crtc fb */
 	crtc->primary->fb = fb;
 	spin_unlock_irqrestore(&crtc->dev->event_lock, flags);
-	amdgpu_flip_work_func(&work->flip_work);
+	amdgpu_flip_work_func(&work->flip_work.work);
 	return 0;
 
-vblank_cleanup:
-	drm_crtc_vblank_put(crtc);
-
 pflip_cleanup:
-	if (unlikely(amdgpu_bo_reserve(new_rbo, false) != 0)) {
-		DRM_ERROR("failed to reserve new rbo in error path\n");
+	if (unlikely(amdgpu_bo_reserve(new_abo, false) != 0)) {
+		DRM_ERROR("failed to reserve new abo in error path\n");
 		goto cleanup;
 	}
 unpin:
-	if (unlikely(amdgpu_bo_unpin(new_rbo) != 0)) {
-		DRM_ERROR("failed to unpin new rbo in error path\n");
+	if (unlikely(amdgpu_bo_unpin(new_abo) != 0)) {
+		DRM_ERROR("failed to unpin new abo in error path\n");
 	}
 unreserve:
-	amdgpu_bo_unreserve(new_rbo);
+	amdgpu_bo_unreserve(new_abo);
 
 cleanup:
-	amdgpu_bo_unref(&work->old_rbo);
+	amdgpu_bo_unref(&work->old_abo);
 	fence_put(work->excl);
 	for (i = 0; i < work->shared_count; ++i)
 		fence_put(work->shared[i]);
@@ -335,7 +297,7 @@ int amdgpu_crtc_set_config(struct drm_mode_set *set)
 	return ret;
 }
 
-static const char *encoder_names[38] = {
+static const char *encoder_names[41] = {
 	"NONE",
 	"INTERNAL_LVDS",
 	"INTERNAL_TMDS1",
@@ -374,6 +336,9 @@ static const char *encoder_names[38] = {
 	"TRAVIS",
 	"INTERNAL_VCE",
 	"INTERNAL_UNIPHY3",
+	"HDMI_ANX9805",
+	"INTERNAL_AMCLK",
+	"VIRTUAL",
 };
 
 static const char *hpd_names[6] = {

drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c

@@ -53,13 +53,19 @@
  * - 3.2.0 - GFX8: Uses EOP_TC_WB_ACTION_EN, so UMDs don't have to do the same
  *           at the end of IBs.
  * - 3.3.0 - Add VM support for UVD on supported hardware.
+ * - 3.4.0 - Add AMDGPU_INFO_NUM_EVICTIONS.
+ * - 3.5.0 - Add support for new UVD_NO_OP register.
+ * - 3.6.0 - kmd involves use CONTEXT_CONTROL in ring buffer.
+ * - 3.7.0 - Add support for VCE clock list packet
+ * - 3.8.0 - Add support raster config init in the kernel
  */
 #define KMS_DRIVER_MAJOR	3
-#define KMS_DRIVER_MINOR	3
+#define KMS_DRIVER_MINOR	8
 #define KMS_DRIVER_PATCHLEVEL	0
 
 int amdgpu_vram_limit = 0;
 int amdgpu_gart_size = -1; /* auto */
+int amdgpu_moverate = -1; /* auto */
 int amdgpu_benchmarking = 0;
 int amdgpu_testing = 0;
 int amdgpu_audio = -1;
@@ -84,11 +90,14 @@ int amdgpu_sched_jobs = 32;
 int amdgpu_sched_hw_submission = 2;
 int amdgpu_powerplay = -1;
 int amdgpu_powercontainment = 1;
+int amdgpu_sclk_deep_sleep_en = 1;
 unsigned amdgpu_pcie_gen_cap = 0;
 unsigned amdgpu_pcie_lane_cap = 0;
 unsigned amdgpu_cg_mask = 0xffffffff;
 unsigned amdgpu_pg_mask = 0xffffffff;
 char *amdgpu_disable_cu = NULL;
+char *amdgpu_virtual_display = NULL;
+unsigned amdgpu_pp_feature_mask = 0xffffffff;
 
 MODULE_PARM_DESC(vramlimit, "Restrict VRAM for testing, in megabytes");
 module_param_named(vramlimit, amdgpu_vram_limit, int, 0600);
@@ -96,6 +105,9 @@ module_param_named(vramlimit, amdgpu_vram_limit, int, 0600);
 MODULE_PARM_DESC(gartsize, "Size of PCIE/IGP gart to setup in megabytes (32, 64, etc., -1 = auto)");
 module_param_named(gartsize, amdgpu_gart_size, int, 0600);
 
+MODULE_PARM_DESC(moverate, "Maximum buffer migration rate in MB/s. (32, 64, etc., -1=auto, 0=1=disabled)");
+module_param_named(moverate, amdgpu_moverate, int, 0600);
+
 MODULE_PARM_DESC(benchmark, "Run benchmark");
 module_param_named(benchmark, amdgpu_benchmarking, int, 0444);
 
@@ -162,13 +174,17 @@ module_param_named(sched_jobs, amdgpu_sched_jobs, int, 0444);
 MODULE_PARM_DESC(sched_hw_submission, "the max number of HW submissions (default 2)");
 module_param_named(sched_hw_submission, amdgpu_sched_hw_submission, int, 0444);
 
-#ifdef CONFIG_DRM_AMD_POWERPLAY
 MODULE_PARM_DESC(powerplay, "Powerplay component (1 = enable, 0 = disable, -1 = auto (default))");
 module_param_named(powerplay, amdgpu_powerplay, int, 0444);
 
 MODULE_PARM_DESC(powercontainment, "Power Containment (1 = enable (default), 0 = disable)");
 module_param_named(powercontainment, amdgpu_powercontainment, int, 0444);
-#endif
+
+MODULE_PARM_DESC(ppfeaturemask, "all power features enabled (default))");
+module_param_named(ppfeaturemask, amdgpu_pp_feature_mask, int, 0444);
+
+MODULE_PARM_DESC(sclkdeepsleep, "SCLK Deep Sleep (1 = enable (default), 0 = disable)");
+module_param_named(sclkdeepsleep, amdgpu_sclk_deep_sleep_en, int, 0444);
 
 MODULE_PARM_DESC(pcie_gen_cap, "PCIE Gen Caps (0: autodetect (default))");
 module_param_named(pcie_gen_cap, amdgpu_pcie_gen_cap, uint, 0444);
@@ -185,7 +201,84 @@ module_param_named(pg_mask, amdgpu_pg_mask, uint, 0444);
 MODULE_PARM_DESC(disable_cu, "Disable CUs (se.sh.cu,...)");
 module_param_named(disable_cu, amdgpu_disable_cu, charp, 0444);
 
+MODULE_PARM_DESC(virtual_display, "Enable virtual display feature (the virtual_display will be set like xxxx:xx:xx.x;xxxx:xx:xx.x)");
+module_param_named(virtual_display, amdgpu_virtual_display, charp, 0444);
+
 static const struct pci_device_id pciidlist[] = {
+#ifdef  CONFIG_DRM_AMDGPU_SI
+	{0x1002, 0x6780, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x6784, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x6788, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x678A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x6790, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x6791, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x6792, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x6798, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x6799, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x679A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x679B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x679E, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x679F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_TAHITI},
+	{0x1002, 0x6800, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|AMD_IS_MOBILITY},
+	{0x1002, 0x6801, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|AMD_IS_MOBILITY},
+	{0x1002, 0x6802, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN|AMD_IS_MOBILITY},
+	{0x1002, 0x6806, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6808, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6809, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6810, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6811, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6816, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6817, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6818, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6819, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_PITCAIRN},
+	{0x1002, 0x6600, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6601, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6602, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6603, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6604, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6605, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6606, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6607, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6608, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND},
+	{0x1002, 0x6610, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND},
+	{0x1002, 0x6611, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND},
+	{0x1002, 0x6613, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND},
+	{0x1002, 0x6617, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6620, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6621, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6623, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND|AMD_IS_MOBILITY},
+	{0x1002, 0x6631, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_OLAND},
+	{0x1002, 0x6820, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6821, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6822, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6823, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6824, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6825, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6826, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6827, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6828, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x6829, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x682A, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x682B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x682C, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x682D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x682F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6830, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6831, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE|AMD_IS_MOBILITY},
+	{0x1002, 0x6835, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x6837, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x6838, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x6839, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x683B, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x683D, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x683F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_VERDE},
+	{0x1002, 0x6660, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
+	{0x1002, 0x6663, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
+	{0x1002, 0x6664, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
+	{0x1002, 0x6665, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
+	{0x1002, 0x6667, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
+	{0x1002, 0x666F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_HAINAN|AMD_IS_MOBILITY},
+#endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	/* Kaveri */
 	{0x1002, 0x1304, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_KAVERI|AMD_IS_MOBILITY|AMD_IS_APU},
@@ -341,7 +434,7 @@ static int amdgpu_kick_out_firmware_fb(struct pci_dev *pdev)
 #ifdef CONFIG_X86
 	primary = pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
 #endif
-	remove_conflicting_framebuffers(ap, "amdgpudrmfb", primary);
+	drm_fb_helper_remove_conflicting_framebuffers(ap, "amdgpudrmfb", primary);
 	kfree(ap);
 
 	return 0;
@@ -383,32 +476,70 @@ amdgpu_pci_remove(struct pci_dev *pdev)
 	drm_put_dev(dev);
 }
 
+static void
+amdgpu_pci_shutdown(struct pci_dev *pdev)
+{
+	/* if we are running in a VM, make sure the device
+	 * torn down properly on reboot/shutdown.
+	 * unfortunately we can't detect certain
+	 * hypervisors so just do this all the time.
+	 */
+	amdgpu_pci_remove(pdev);
+}
+
 static int amdgpu_pmops_suspend(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
+
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	return amdgpu_suspend_kms(drm_dev, true, true);
+	return amdgpu_device_suspend(drm_dev, true, true);
 }
 
 static int amdgpu_pmops_resume(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	return amdgpu_resume_kms(drm_dev, true, true);
+
+	/* GPU comes up enabled by the bios on resume */
+	if (amdgpu_device_is_px(drm_dev)) {
+		pm_runtime_disable(dev);
+		pm_runtime_set_active(dev);
+		pm_runtime_enable(dev);
+	}
+
+	return amdgpu_device_resume(drm_dev, true, true);
 }
 
 static int amdgpu_pmops_freeze(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
+
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	return amdgpu_suspend_kms(drm_dev, false, true);
+	return amdgpu_device_suspend(drm_dev, false, true);
 }
 
 static int amdgpu_pmops_thaw(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
+
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	return amdgpu_device_resume(drm_dev, false, true);
+}
+
+static int amdgpu_pmops_poweroff(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	return amdgpu_device_suspend(drm_dev, true, true);
+}
+
+static int amdgpu_pmops_restore(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+
 	struct drm_device *drm_dev = pci_get_drvdata(pdev);
-	return amdgpu_resume_kms(drm_dev, false, true);
+	return amdgpu_device_resume(drm_dev, false, true);
 }
 
 static int amdgpu_pmops_runtime_suspend(struct device *dev)
@@ -426,7 +557,7 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 	drm_kms_helper_poll_disable(drm_dev);
 	vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_OFF);
 
-	ret = amdgpu_suspend_kms(drm_dev, false, false);
+	ret = amdgpu_device_suspend(drm_dev, false, false);
 	pci_save_state(pdev);
 	pci_disable_device(pdev);
 	pci_ignore_hotplug(pdev);
@@ -459,7 +590,7 @@ static int amdgpu_pmops_runtime_resume(struct device *dev)
 		return ret;
 	pci_set_master(pdev);
 
-	ret = amdgpu_resume_kms(drm_dev, false, false);
+	ret = amdgpu_device_resume(drm_dev, false, false);
 	drm_kms_helper_poll_enable(drm_dev);
 	vga_switcheroo_set_dynamic_switch(pdev, VGA_SWITCHEROO_ON);
 	drm_dev->switch_power_state = DRM_SWITCH_POWER_ON;
@@ -513,8 +644,8 @@ static const struct dev_pm_ops amdgpu_pm_ops = {
 	.resume = amdgpu_pmops_resume,
 	.freeze = amdgpu_pmops_freeze,
 	.thaw = amdgpu_pmops_thaw,
-	.poweroff = amdgpu_pmops_freeze,
-	.restore = amdgpu_pmops_resume,
+	.poweroff = amdgpu_pmops_poweroff,
+	.restore = amdgpu_pmops_restore,
 	.runtime_suspend = amdgpu_pmops_runtime_suspend,
 	.runtime_resume = amdgpu_pmops_runtime_resume,
 	.runtime_idle = amdgpu_pmops_runtime_idle,
@@ -596,6 +727,7 @@ static struct pci_driver amdgpu_kms_pci_driver = {
 	.id_table = pciidlist,
 	.probe = amdgpu_pci_probe,
 	.remove = amdgpu_pci_remove,
+	.shutdown = amdgpu_pci_shutdown,
 	.driver.pm = &amdgpu_pm_ops,
 };
 

drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c

@@ -25,7 +25,7 @@
  */
 #include <linux/module.h>
 #include <linux/slab.h>
-#include <linux/fb.h>
+#include <linux/pm_runtime.h>
 
 #include <drm/drmP.h>
 #include <drm/drm_crtc.h>
@@ -48,8 +48,35 @@ struct amdgpu_fbdev {
 	struct amdgpu_device *adev;
 };
 
+static int
+amdgpufb_open(struct fb_info *info, int user)
+{
+	struct amdgpu_fbdev *rfbdev = info->par;
+	struct amdgpu_device *adev = rfbdev->adev;
+	int ret = pm_runtime_get_sync(adev->ddev->dev);
+	if (ret < 0 && ret != -EACCES) {
+		pm_runtime_mark_last_busy(adev->ddev->dev);
+		pm_runtime_put_autosuspend(adev->ddev->dev);
+		return ret;
+	}
+	return 0;
+}
+
+static int
+amdgpufb_release(struct fb_info *info, int user)
+{
+	struct amdgpu_fbdev *rfbdev = info->par;
+	struct amdgpu_device *adev = rfbdev->adev;
+
+	pm_runtime_mark_last_busy(adev->ddev->dev);
+	pm_runtime_put_autosuspend(adev->ddev->dev);
+	return 0;
+}
+
 static struct fb_ops amdgpufb_ops = {
 	.owner = THIS_MODULE,
+	.fb_open = amdgpufb_open,
+	.fb_release = amdgpufb_release,
 	.fb_check_var = drm_fb_helper_check_var,
 	.fb_set_par = drm_fb_helper_set_par,
 	.fb_fillrect = drm_fb_helper_cfb_fillrect,
@@ -88,14 +115,14 @@ int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int bpp, bool tile
 
 static void amdgpufb_destroy_pinned_object(struct drm_gem_object *gobj)
 {
-	struct amdgpu_bo *rbo = gem_to_amdgpu_bo(gobj);
+	struct amdgpu_bo *abo = gem_to_amdgpu_bo(gobj);
 	int ret;
 
-	ret = amdgpu_bo_reserve(rbo, false);
+	ret = amdgpu_bo_reserve(abo, false);
 	if (likely(ret == 0)) {
-		amdgpu_bo_kunmap(rbo);
-		amdgpu_bo_unpin(rbo);
-		amdgpu_bo_unreserve(rbo);
+		amdgpu_bo_kunmap(abo);
+		amdgpu_bo_unpin(abo);
+		amdgpu_bo_unreserve(abo);
 	}
 	drm_gem_object_unreference_unlocked(gobj);
 }
@@ -106,7 +133,7 @@ static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev,
 {
 	struct amdgpu_device *adev = rfbdev->adev;
 	struct drm_gem_object *gobj = NULL;
-	struct amdgpu_bo *rbo = NULL;
+	struct amdgpu_bo *abo = NULL;
 	bool fb_tiled = false; /* useful for testing */
 	u32 tiling_flags = 0;
 	int ret;
@@ -132,30 +159,30 @@ static int amdgpufb_create_pinned_object(struct amdgpu_fbdev *rfbdev,
 		       aligned_size);
 		return -ENOMEM;
 	}
-	rbo = gem_to_amdgpu_bo(gobj);
+	abo = gem_to_amdgpu_bo(gobj);
 
 	if (fb_tiled)
 		tiling_flags = AMDGPU_TILING_SET(ARRAY_MODE, GRPH_ARRAY_2D_TILED_THIN1);
 
-	ret = amdgpu_bo_reserve(rbo, false);
+	ret = amdgpu_bo_reserve(abo, false);
 	if (unlikely(ret != 0))
 		goto out_unref;
 
 	if (tiling_flags) {
-		ret = amdgpu_bo_set_tiling_flags(rbo,
+		ret = amdgpu_bo_set_tiling_flags(abo,
 						 tiling_flags);
 		if (ret)
 			dev_err(adev->dev, "FB failed to set tiling flags\n");
 	}
 
 
-	ret = amdgpu_bo_pin_restricted(rbo, AMDGPU_GEM_DOMAIN_VRAM, 0, 0, NULL);
+	ret = amdgpu_bo_pin_restricted(abo, AMDGPU_GEM_DOMAIN_VRAM, 0, 0, NULL);
 	if (ret) {
-		amdgpu_bo_unreserve(rbo);
+		amdgpu_bo_unreserve(abo);
 		goto out_unref;
 	}
-	ret = amdgpu_bo_kmap(rbo, NULL);
-	amdgpu_bo_unreserve(rbo);
+	ret = amdgpu_bo_kmap(abo, NULL);
+	amdgpu_bo_unreserve(abo);
 	if (ret) {
 		goto out_unref;
 	}
@@ -177,7 +204,7 @@ static int amdgpufb_create(struct drm_fb_helper *helper,
 	struct drm_framebuffer *fb = NULL;
 	struct drm_mode_fb_cmd2 mode_cmd;
 	struct drm_gem_object *gobj = NULL;
-	struct amdgpu_bo *rbo = NULL;
+	struct amdgpu_bo *abo = NULL;
 	int ret;
 	unsigned long tmp;
 
@@ -196,7 +223,7 @@ static int amdgpufb_create(struct drm_fb_helper *helper,
 		return ret;
 	}
 
-	rbo = gem_to_amdgpu_bo(gobj);
+	abo = gem_to_amdgpu_bo(gobj);
 
 	/* okay we have an object now allocate the framebuffer */
 	info = drm_fb_helper_alloc_fbi(helper);
@@ -219,7 +246,7 @@ static int amdgpufb_create(struct drm_fb_helper *helper,
 	/* setup helper */
 	rfbdev->helper.fb = fb;
 
-	memset_io(rbo->kptr, 0x0, amdgpu_bo_size(rbo));
+	memset_io(abo->kptr, 0x0, amdgpu_bo_size(abo));
 
 	strcpy(info->fix.id, "amdgpudrmfb");
 
@@ -228,11 +255,11 @@ static int amdgpufb_create(struct drm_fb_helper *helper,
 	info->flags = FBINFO_DEFAULT | FBINFO_CAN_FORCE_OUTPUT;
 	info->fbops = &amdgpufb_ops;
 
-	tmp = amdgpu_bo_gpu_offset(rbo) - adev->mc.vram_start;
+	tmp = amdgpu_bo_gpu_offset(abo) - adev->mc.vram_start;
 	info->fix.smem_start = adev->mc.aper_base + tmp;
-	info->fix.smem_len = amdgpu_bo_size(rbo);
-	info->screen_base = rbo->kptr;
-	info->screen_size = amdgpu_bo_size(rbo);
+	info->fix.smem_len = amdgpu_bo_size(abo);
+	info->screen_base = abo->kptr;
+	info->screen_size = amdgpu_bo_size(abo);
 
 	drm_fb_helper_fill_var(info, &rfbdev->helper, sizes->fb_width, sizes->fb_height);
 
@@ -249,7 +276,7 @@ static int amdgpufb_create(struct drm_fb_helper *helper,
 
 	DRM_INFO("fb mappable at 0x%lX\n",  info->fix.smem_start);
 	DRM_INFO("vram apper at 0x%lX\n",  (unsigned long)adev->mc.aper_base);
-	DRM_INFO("size %lu\n", (unsigned long)amdgpu_bo_size(rbo));
+	DRM_INFO("size %lu\n", (unsigned long)amdgpu_bo_size(abo));
 	DRM_INFO("fb depth is %d\n", fb->depth);
 	DRM_INFO("   pitch is %d\n", fb->pitches[0]);
 
@@ -259,7 +286,7 @@ static int amdgpufb_create(struct drm_fb_helper *helper,
 out_destroy_fbi:
 	drm_fb_helper_release_fbi(helper);
 out_unref:
-	if (rbo) {
+	if (abo) {
 
 	}
 	if (fb && ret) {

drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c

@@ -454,6 +454,7 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev)
 		for (j = 0; j <= ring->fence_drv.num_fences_mask; ++j)
 			fence_put(ring->fence_drv.fences[j]);
 		kfree(ring->fence_drv.fences);
+		ring->fence_drv.fences = NULL;
 		ring->fence_drv.initialized = false;
 	}
 }

drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c

@@ -238,7 +238,7 @@ void amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
 	t = offset / AMDGPU_GPU_PAGE_SIZE;
 	p = t / (PAGE_SIZE / AMDGPU_GPU_PAGE_SIZE);
 	for (i = 0; i < pages; i++, p++) {
-#ifdef CONFIG_AMDGPU_GART_DEBUGFS
+#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
 		adev->gart.pages[p] = NULL;
 #endif
 		page_base = adev->dummy_page.addr;
@@ -286,7 +286,7 @@ int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
 	p = t / (PAGE_SIZE / AMDGPU_GPU_PAGE_SIZE);
 
 	for (i = 0; i < pages; i++, p++) {
-#ifdef CONFIG_AMDGPU_GART_DEBUGFS
+#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
 		adev->gart.pages[p] = pagelist[i];
 #endif
 		if (adev->gart.ptr) {
@@ -331,7 +331,7 @@ int amdgpu_gart_init(struct amdgpu_device *adev)
 	DRM_INFO("GART: num cpu pages %u, num gpu pages %u\n",
 		 adev->gart.num_cpu_pages, adev->gart.num_gpu_pages);
 
-#ifdef CONFIG_AMDGPU_GART_DEBUGFS
+#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
 	/* Allocate pages table */
 	adev->gart.pages = vzalloc(sizeof(void *) * adev->gart.num_cpu_pages);
 	if (adev->gart.pages == NULL) {
@@ -357,7 +357,7 @@ void amdgpu_gart_fini(struct amdgpu_device *adev)
 		amdgpu_gart_unbind(adev, 0, adev->gart.num_cpu_pages);
 	}
 	adev->gart.ready = false;
-#ifdef CONFIG_AMDGPU_GART_DEBUGFS
+#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
 	vfree(adev->gart.pages);
 	adev->gart.pages = NULL;
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_gds.h

@@ -31,14 +31,6 @@
 #define AMDGPU_GWS_SHIFT	PAGE_SHIFT
 #define AMDGPU_OA_SHIFT		PAGE_SHIFT
 
-#define AMDGPU_PL_GDS		TTM_PL_PRIV0
-#define AMDGPU_PL_GWS		TTM_PL_PRIV1
-#define AMDGPU_PL_OA		TTM_PL_PRIV2
-
-#define AMDGPU_PL_FLAG_GDS		TTM_PL_FLAG_PRIV0
-#define AMDGPU_PL_FLAG_GWS		TTM_PL_FLAG_PRIV1
-#define AMDGPU_PL_FLAG_OA		TTM_PL_FLAG_PRIV2
-
 struct amdgpu_ring;
 struct amdgpu_bo;
 

drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c

@@ -118,23 +118,23 @@ void amdgpu_gem_force_release(struct amdgpu_device *adev)
  */
 int amdgpu_gem_object_open(struct drm_gem_object *obj, struct drm_file *file_priv)
 {
-	struct amdgpu_bo *rbo = gem_to_amdgpu_bo(obj);
-	struct amdgpu_device *adev = rbo->adev;
+	struct amdgpu_bo *abo = gem_to_amdgpu_bo(obj);
+	struct amdgpu_device *adev = abo->adev;
 	struct amdgpu_fpriv *fpriv = file_priv->driver_priv;
 	struct amdgpu_vm *vm = &fpriv->vm;
 	struct amdgpu_bo_va *bo_va;
 	int r;
-	r = amdgpu_bo_reserve(rbo, false);
+	r = amdgpu_bo_reserve(abo, false);
 	if (r)
 		return r;
 
-	bo_va = amdgpu_vm_bo_find(vm, rbo);
+	bo_va = amdgpu_vm_bo_find(vm, abo);
 	if (!bo_va) {
-		bo_va = amdgpu_vm_bo_add(adev, vm, rbo);
+		bo_va = amdgpu_vm_bo_add(adev, vm, abo);
 	} else {
 		++bo_va->ref_count;
 	}
-	amdgpu_bo_unreserve(rbo);
+	amdgpu_bo_unreserve(abo);
 	return 0;
 }
 
@@ -528,7 +528,7 @@ static void amdgpu_gem_va_update_vm(struct amdgpu_device *adev,
 		goto error_unreserve;
 
 	if (operation == AMDGPU_VA_OP_MAP)
-		r = amdgpu_vm_bo_update(adev, bo_va, &bo_va->bo->tbo.mem);
+		r = amdgpu_vm_bo_update(adev, bo_va, false);
 
 error_unreserve:
 	ttm_eu_backoff_reservation(&ticket, &list);
@@ -547,7 +547,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	struct drm_gem_object *gobj;
 	struct amdgpu_device *adev = dev->dev_private;
 	struct amdgpu_fpriv *fpriv = filp->driver_priv;
-	struct amdgpu_bo *rbo;
+	struct amdgpu_bo *abo;
 	struct amdgpu_bo_va *bo_va;
 	struct ttm_validate_buffer tv, tv_pd;
 	struct ww_acquire_ctx ticket;
@@ -587,10 +587,10 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	gobj = drm_gem_object_lookup(filp, args->handle);
 	if (gobj == NULL)
 		return -ENOENT;
-	rbo = gem_to_amdgpu_bo(gobj);
+	abo = gem_to_amdgpu_bo(gobj);
 	INIT_LIST_HEAD(&list);
 	INIT_LIST_HEAD(&duplicates);
-	tv.bo = &rbo->tbo;
+	tv.bo = &abo->tbo;
 	tv.shared = true;
 	list_add(&tv.head, &list);
 
@@ -604,7 +604,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 		return r;
 	}
 
-	bo_va = amdgpu_vm_bo_find(&fpriv->vm, rbo);
+	bo_va = amdgpu_vm_bo_find(&fpriv->vm, abo);
 	if (!bo_va) {
 		ttm_eu_backoff_reservation(&ticket, &list);
 		drm_gem_object_unreference_unlocked(gobj);

drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c

@@ -0,0 +1,239 @@
+/*
+ * Copyright 2016 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Christian König
+ */
+
+#include <drm/drmP.h>
+#include "amdgpu.h"
+
+struct amdgpu_gtt_mgr {
+	struct drm_mm mm;
+	spinlock_t lock;
+	uint64_t available;
+};
+
+/**
+ * amdgpu_gtt_mgr_init - init GTT manager and DRM MM
+ *
+ * @man: TTM memory type manager
+ * @p_size: maximum size of GTT
+ *
+ * Allocate and initialize the GTT manager.
+ */
+static int amdgpu_gtt_mgr_init(struct ttm_mem_type_manager *man,
+			       unsigned long p_size)
+{
+	struct amdgpu_gtt_mgr *mgr;
+
+	mgr = kzalloc(sizeof(*mgr), GFP_KERNEL);
+	if (!mgr)
+		return -ENOMEM;
+
+	drm_mm_init(&mgr->mm, 0, p_size);
+	spin_lock_init(&mgr->lock);
+	mgr->available = p_size;
+	man->priv = mgr;
+	return 0;
+}
+
+/**
+ * amdgpu_gtt_mgr_fini - free and destroy GTT manager
+ *
+ * @man: TTM memory type manager
+ *
+ * Destroy and free the GTT manager, returns -EBUSY if ranges are still
+ * allocated inside it.
+ */
+static int amdgpu_gtt_mgr_fini(struct ttm_mem_type_manager *man)
+{
+	struct amdgpu_gtt_mgr *mgr = man->priv;
+
+	spin_lock(&mgr->lock);
+	if (!drm_mm_clean(&mgr->mm)) {
+		spin_unlock(&mgr->lock);
+		return -EBUSY;
+	}
+
+	drm_mm_takedown(&mgr->mm);
+	spin_unlock(&mgr->lock);
+	kfree(mgr);
+	man->priv = NULL;
+	return 0;
+}
+
+/**
+ * amdgpu_gtt_mgr_alloc - allocate new ranges
+ *
+ * @man: TTM memory type manager
+ * @tbo: TTM BO we need this range for
+ * @place: placement flags and restrictions
+ * @mem: the resulting mem object
+ *
+ * Allocate the address space for a node.
+ */
+int amdgpu_gtt_mgr_alloc(struct ttm_mem_type_manager *man,
+			 struct ttm_buffer_object *tbo,
+			 const struct ttm_place *place,
+			 struct ttm_mem_reg *mem)
+{
+	struct amdgpu_gtt_mgr *mgr = man->priv;
+	struct drm_mm_node *node = mem->mm_node;
+	enum drm_mm_search_flags sflags = DRM_MM_SEARCH_BEST;
+	enum drm_mm_allocator_flags aflags = DRM_MM_CREATE_DEFAULT;
+	unsigned long fpfn, lpfn;
+	int r;
+
+	if (node->start != AMDGPU_BO_INVALID_OFFSET)
+		return 0;
+
+	if (place)
+		fpfn = place->fpfn;
+	else
+		fpfn = 0;
+
+	if (place && place->lpfn)
+		lpfn = place->lpfn;
+	else
+		lpfn = man->size;
+
+	if (place && place->flags & TTM_PL_FLAG_TOPDOWN) {
+		sflags = DRM_MM_SEARCH_BELOW;
+		aflags = DRM_MM_CREATE_TOP;
+	}
+
+	spin_lock(&mgr->lock);
+	r = drm_mm_insert_node_in_range_generic(&mgr->mm, node, mem->num_pages,
+						mem->page_alignment, 0,
+						fpfn, lpfn, sflags, aflags);
+	spin_unlock(&mgr->lock);
+
+	if (!r) {
+		mem->start = node->start;
+		if (&tbo->mem == mem)
+			tbo->offset = (tbo->mem.start << PAGE_SHIFT) +
+			    tbo->bdev->man[tbo->mem.mem_type].gpu_offset;
+	}
+
+	return r;
+}
+
+/**
+ * amdgpu_gtt_mgr_new - allocate a new node
+ *
+ * @man: TTM memory type manager
+ * @tbo: TTM BO we need this range for
+ * @place: placement flags and restrictions
+ * @mem: the resulting mem object
+ *
+ * Dummy, allocate the node but no space for it yet.
+ */
+static int amdgpu_gtt_mgr_new(struct ttm_mem_type_manager *man,
+			      struct ttm_buffer_object *tbo,
+			      const struct ttm_place *place,
+			      struct ttm_mem_reg *mem)
+{
+	struct amdgpu_gtt_mgr *mgr = man->priv;
+	struct drm_mm_node *node;
+	int r;
+
+	spin_lock(&mgr->lock);
+	if (mgr->available < mem->num_pages) {
+		spin_unlock(&mgr->lock);
+		return 0;
+	}
+	mgr->available -= mem->num_pages;
+	spin_unlock(&mgr->lock);
+
+	node = kzalloc(sizeof(*node), GFP_KERNEL);
+	if (!node)
+		return -ENOMEM;
+
+	node->start = AMDGPU_BO_INVALID_OFFSET;
+	mem->mm_node = node;
+
+	if (place->fpfn || place->lpfn || place->flags & TTM_PL_FLAG_TOPDOWN) {
+		r = amdgpu_gtt_mgr_alloc(man, tbo, place, mem);
+		if (unlikely(r)) {
+			kfree(node);
+			mem->mm_node = NULL;
+		}
+	} else {
+		mem->start = node->start;
+	}
+
+	return 0;
+}
+
+/**
+ * amdgpu_gtt_mgr_del - free ranges
+ *
+ * @man: TTM memory type manager
+ * @tbo: TTM BO we need this range for
+ * @place: placement flags and restrictions
+ * @mem: TTM memory object
+ *
+ * Free the allocated GTT again.
+ */
+static void amdgpu_gtt_mgr_del(struct ttm_mem_type_manager *man,
+			       struct ttm_mem_reg *mem)
+{
+	struct amdgpu_gtt_mgr *mgr = man->priv;
+	struct drm_mm_node *node = mem->mm_node;
+
+	if (!node)
+		return;
+
+	spin_lock(&mgr->lock);
+	if (node->start != AMDGPU_BO_INVALID_OFFSET)
+		drm_mm_remove_node(node);
+	mgr->available += mem->num_pages;
+	spin_unlock(&mgr->lock);
+
+	kfree(node);
+	mem->mm_node = NULL;
+}
+
+/**
+ * amdgpu_gtt_mgr_debug - dump VRAM table
+ *
+ * @man: TTM memory type manager
+ * @prefix: text prefix
+ *
+ * Dump the table content using printk.
+ */
+static void amdgpu_gtt_mgr_debug(struct ttm_mem_type_manager *man,
+				  const char *prefix)
+{
+	struct amdgpu_gtt_mgr *mgr = man->priv;
+
+	spin_lock(&mgr->lock);
+	drm_mm_debug_table(&mgr->mm, prefix);
+	spin_unlock(&mgr->lock);
+}
+
+const struct ttm_mem_type_manager_func amdgpu_gtt_mgr_func = {
+	amdgpu_gtt_mgr_init,
+	amdgpu_gtt_mgr_fini,
+	amdgpu_gtt_mgr_new,
+	amdgpu_gtt_mgr_del,
+	amdgpu_gtt_mgr_debug
+};

drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.c

@@ -158,8 +158,8 @@ static const struct i2c_algorithm amdgpu_atombios_i2c_algo = {
 };
 
 struct amdgpu_i2c_chan *amdgpu_i2c_create(struct drm_device *dev,
-					    struct amdgpu_i2c_bus_rec *rec,
-					    const char *name)
+					  const struct amdgpu_i2c_bus_rec *rec,
+					  const char *name)
 {
 	struct amdgpu_i2c_chan *i2c;
 	int ret;
@@ -186,10 +186,8 @@ struct amdgpu_i2c_chan *amdgpu_i2c_create(struct drm_device *dev,
 			 "AMDGPU i2c hw bus %s", name);
 		i2c->adapter.algo = &amdgpu_atombios_i2c_algo;
 		ret = i2c_add_adapter(&i2c->adapter);
-		if (ret) {
-			DRM_ERROR("Failed to register hw i2c %s\n", name);
+		if (ret)
 			goto out_free;
-		}
 	} else {
 		/* set the amdgpu bit adapter */
 		snprintf(i2c->adapter.name, sizeof(i2c->adapter.name),
@@ -222,6 +220,7 @@ void amdgpu_i2c_destroy(struct amdgpu_i2c_chan *i2c)
 {
 	if (!i2c)
 		return;
+	WARN_ON(i2c->has_aux);
 	i2c_del_adapter(&i2c->adapter);
 	kfree(i2c);
 }
@@ -251,8 +250,8 @@ void amdgpu_i2c_fini(struct amdgpu_device *adev)
 
 /* Add additional buses */
 void amdgpu_i2c_add(struct amdgpu_device *adev,
-		     struct amdgpu_i2c_bus_rec *rec,
-		     const char *name)
+		    const struct amdgpu_i2c_bus_rec *rec,
+		    const char *name)
 {
 	struct drm_device *dev = adev->ddev;
 	int i;
@@ -268,7 +267,7 @@ void amdgpu_i2c_add(struct amdgpu_device *adev,
 /* looks up bus based on id */
 struct amdgpu_i2c_chan *
 amdgpu_i2c_lookup(struct amdgpu_device *adev,
-		   struct amdgpu_i2c_bus_rec *i2c_bus)
+		  const struct amdgpu_i2c_bus_rec *i2c_bus)
 {
 	int i;
 
@@ -338,7 +337,7 @@ static void amdgpu_i2c_put_byte(struct amdgpu_i2c_chan *i2c_bus,
 
 /* ddc router switching */
 void
-amdgpu_i2c_router_select_ddc_port(struct amdgpu_connector *amdgpu_connector)
+amdgpu_i2c_router_select_ddc_port(const struct amdgpu_connector *amdgpu_connector)
 {
 	u8 val;
 
@@ -367,7 +366,7 @@ amdgpu_i2c_router_select_ddc_port(struct amdgpu_connector *amdgpu_connector)
 
 /* clock/data router switching */
 void
-amdgpu_i2c_router_select_cd_port(struct amdgpu_connector *amdgpu_connector)
+amdgpu_i2c_router_select_cd_port(const struct amdgpu_connector *amdgpu_connector)
 {
 	u8 val;
 

drivers/gpu/drm/amd/amdgpu/amdgpu_i2c.h

@@ -25,20 +25,20 @@
 #define __AMDGPU_I2C_H__
 
 struct amdgpu_i2c_chan *amdgpu_i2c_create(struct drm_device *dev,
-					    struct amdgpu_i2c_bus_rec *rec,
-					    const char *name);
+					  const struct amdgpu_i2c_bus_rec *rec,
+					  const char *name);
 void amdgpu_i2c_destroy(struct amdgpu_i2c_chan *i2c);
 void amdgpu_i2c_init(struct amdgpu_device *adev);
 void amdgpu_i2c_fini(struct amdgpu_device *adev);
 void amdgpu_i2c_add(struct amdgpu_device *adev,
-		     struct amdgpu_i2c_bus_rec *rec,
-		     const char *name);
+		    const struct amdgpu_i2c_bus_rec *rec,
+		    const char *name);
 struct amdgpu_i2c_chan *
 amdgpu_i2c_lookup(struct amdgpu_device *adev,
-		   struct amdgpu_i2c_bus_rec *i2c_bus);
+		  const struct amdgpu_i2c_bus_rec *i2c_bus);
 void
-amdgpu_i2c_router_select_ddc_port(struct amdgpu_connector *amdgpu_connector);
+amdgpu_i2c_router_select_ddc_port(const struct amdgpu_connector *connector);
 void
-amdgpu_i2c_router_select_cd_port(struct amdgpu_connector *amdgpu_connector);
+amdgpu_i2c_router_select_cd_port(const struct amdgpu_connector *connector);
 
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

@@ -124,7 +124,8 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	bool skip_preamble, need_ctx_switch;
 	unsigned patch_offset = ~0;
 	struct amdgpu_vm *vm;
-	uint64_t ctx;
+	uint64_t fence_ctx;
+	uint32_t status = 0, alloc_size;
 
 	unsigned i;
 	int r = 0;
@@ -135,14 +136,14 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	/* ring tests don't use a job */
 	if (job) {
 		vm = job->vm;
-		ctx = job->ctx;
+		fence_ctx = job->fence_ctx;
 	} else {
 		vm = NULL;
-		ctx = 0;
+		fence_ctx = 0;
 	}
 
 	if (!ring->ready) {
-		dev_err(adev->dev, "couldn't schedule ib\n");
+		dev_err(adev->dev, "couldn't schedule ib on ring <%s>\n", ring->name);
 		return -EINVAL;
 	}
 
@@ -151,7 +152,10 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 		return -EINVAL;
 	}
 
-	r = amdgpu_ring_alloc(ring, 256 * num_ibs);
+	alloc_size = amdgpu_ring_get_dma_frame_size(ring) +
+		num_ibs * amdgpu_ring_get_emit_ib_size(ring);
+
+	r = amdgpu_ring_alloc(ring, alloc_size);
 	if (r) {
 		dev_err(adev->dev, "scheduling IB failed (%d).\n", r);
 		return r;
@@ -174,13 +178,22 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	/* always set cond_exec_polling to CONTINUE */
 	*ring->cond_exe_cpu_addr = 1;
 
-	skip_preamble = ring->current_ctx == ctx;
-	need_ctx_switch = ring->current_ctx != ctx;
+	skip_preamble = ring->current_ctx == fence_ctx;
+	need_ctx_switch = ring->current_ctx != fence_ctx;
+	if (job && ring->funcs->emit_cntxcntl) {
+		if (need_ctx_switch)
+			status |= AMDGPU_HAVE_CTX_SWITCH;
+		status |= job->preamble_status;
+		amdgpu_ring_emit_cntxcntl(ring, status);
+	}
+
 	for (i = 0; i < num_ibs; ++i) {
 		ib = &ibs[i];
 
 		/* drop preamble IBs if we don't have a context switch */
-		if ((ib->flags & AMDGPU_IB_FLAG_PREAMBLE) && skip_preamble)
+		if ((ib->flags & AMDGPU_IB_FLAG_PREAMBLE) &&
+			skip_preamble &&
+			!(status & AMDGPU_PREAMBLE_IB_PRESENT_FIRST))
 			continue;
 
 		amdgpu_ring_emit_ib(ring, ib, job ? job->vm_id : 0,
@@ -209,7 +222,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs,
 	if (patch_offset != ~0 && ring->funcs->patch_cond_exec)
 		amdgpu_ring_patch_cond_exec(ring, patch_offset);
 
-	ring->current_ctx = ctx;
+	ring->current_ctx = fence_ctx;
+	if (ring->funcs->emit_switch_buffer)
+		amdgpu_ring_emit_switch_buffer(ring);
 	amdgpu_ring_commit(ring);
 	return 0;
 }

drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c

@@ -40,32 +40,15 @@ static int amdgpu_ih_ring_alloc(struct amdgpu_device *adev)
 
 	/* Allocate ring buffer */
 	if (adev->irq.ih.ring_obj == NULL) {
-		r = amdgpu_bo_create(adev, adev->irq.ih.ring_size,
-				     PAGE_SIZE, true,
-				     AMDGPU_GEM_DOMAIN_GTT, 0,
-				     NULL, NULL, &adev->irq.ih.ring_obj);
+		r = amdgpu_bo_create_kernel(adev, adev->irq.ih.ring_size,
+					    PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
+					    &adev->irq.ih.ring_obj,
+					    &adev->irq.ih.gpu_addr,
+					    (void **)&adev->irq.ih.ring);
 		if (r) {
 			DRM_ERROR("amdgpu: failed to create ih ring buffer (%d).\n", r);
 			return r;
 		}
-		r = amdgpu_bo_reserve(adev->irq.ih.ring_obj, false);
-		if (unlikely(r != 0))
-			return r;
-		r = amdgpu_bo_pin(adev->irq.ih.ring_obj,
-				  AMDGPU_GEM_DOMAIN_GTT,
-				  &adev->irq.ih.gpu_addr);
-		if (r) {
-			amdgpu_bo_unreserve(adev->irq.ih.ring_obj);
-			DRM_ERROR("amdgpu: failed to pin ih ring buffer (%d).\n", r);
-			return r;
-		}
-		r = amdgpu_bo_kmap(adev->irq.ih.ring_obj,
-				   (void **)&adev->irq.ih.ring);
-		amdgpu_bo_unreserve(adev->irq.ih.ring_obj);
-		if (r) {
-			DRM_ERROR("amdgpu: failed to map ih ring buffer (%d).\n", r);
-			return r;
-		}
 	}
 	return 0;
 }
@@ -136,8 +119,6 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, unsigned ring_size,
  */
 void amdgpu_ih_ring_fini(struct amdgpu_device *adev)
 {
-	int r;
-
 	if (adev->irq.ih.use_bus_addr) {
 		if (adev->irq.ih.ring) {
 			/* add 8 bytes for the rptr/wptr shadows and
@@ -149,17 +130,9 @@ void amdgpu_ih_ring_fini(struct amdgpu_device *adev)
 			adev->irq.ih.ring = NULL;
 		}
 	} else {
-		if (adev->irq.ih.ring_obj) {
-			r = amdgpu_bo_reserve(adev->irq.ih.ring_obj, false);
-			if (likely(r == 0)) {
-				amdgpu_bo_kunmap(adev->irq.ih.ring_obj);
-				amdgpu_bo_unpin(adev->irq.ih.ring_obj);
-				amdgpu_bo_unreserve(adev->irq.ih.ring_obj);
-			}
-			amdgpu_bo_unref(&adev->irq.ih.ring_obj);
-			adev->irq.ih.ring = NULL;
-			adev->irq.ih.ring_obj = NULL;
-		}
+		amdgpu_bo_free_kernel(&adev->irq.ih.ring_obj,
+				      &adev->irq.ih.gpu_addr,
+				      (void **)&adev->irq.ih.ring);
 		amdgpu_wb_free(adev, adev->irq.ih.wptr_offs);
 		amdgpu_wb_free(adev, adev->irq.ih.rptr_offs);
 	}

drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h

@@ -70,6 +70,7 @@ struct amdgpu_irq {
 	/* gen irq stuff */
 	struct irq_domain		*domain; /* GPU irq controller domain */
 	unsigned			virq[AMDGPU_MAX_IRQ_SRC_ID];
+	uint32_t                        srbm_soft_reset;
 };
 
 void amdgpu_irq_preinstall(struct drm_device *dev);

drivers/gpu/drm/amd/amdgpu/amdgpu_job.c

@@ -91,7 +91,7 @@ void amdgpu_job_free_resources(struct amdgpu_job *job)
 		amdgpu_ib_free(job->adev, &job->ibs[i], f);
 }
 
-void amdgpu_job_free_cb(struct amd_sched_job *s_job)
+static void amdgpu_job_free_cb(struct amd_sched_job *s_job)
 {
 	struct amdgpu_job *job = container_of(s_job, struct amdgpu_job, base);
 
@@ -124,7 +124,7 @@ int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring,
 		return r;
 
 	job->owner = owner;
-	job->ctx = entity->fence_context;
+	job->fence_ctx = entity->fence_context;
 	*f = fence_get(&job->base.s_fence->finished);
 	amdgpu_job_free_resources(job);
 	amd_sched_entity_push_job(&job->base);

drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c

@@ -292,14 +292,14 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 			type = AMD_IP_BLOCK_TYPE_UVD;
 			ring_mask = adev->uvd.ring.ready ? 1 : 0;
 			ib_start_alignment = AMDGPU_GPU_PAGE_SIZE;
-			ib_size_alignment = 8;
+			ib_size_alignment = 16;
 			break;
 		case AMDGPU_HW_IP_VCE:
 			type = AMD_IP_BLOCK_TYPE_VCE;
-			for (i = 0; i < AMDGPU_MAX_VCE_RINGS; i++)
+			for (i = 0; i < adev->vce.num_rings; i++)
 				ring_mask |= ((adev->vce.ring[i].ready ? 1 : 0) << i);
 			ib_start_alignment = AMDGPU_GPU_PAGE_SIZE;
-			ib_size_alignment = 8;
+			ib_size_alignment = 1;
 			break;
 		default:
 			return -EINVAL;
@@ -373,6 +373,9 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file
 	case AMDGPU_INFO_NUM_BYTES_MOVED:
 		ui64 = atomic64_read(&adev->num_bytes_moved);
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
+	case AMDGPU_INFO_NUM_EVICTIONS:
+		ui64 = atomic64_read(&adev->num_evictions);
+		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
 	case AMDGPU_INFO_VRAM_USAGE:
 		ui64 = atomic64_read(&adev->vram_usage);
 		return copy_to_user(out, &ui64, min(size, 8u)) ? -EFAULT : 0;
@@ -539,12 +542,16 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 		return r;
 
 	fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
-	if (unlikely(!fpriv))
-		return -ENOMEM;
+	if (unlikely(!fpriv)) {
+		r = -ENOMEM;
+		goto out_suspend;
+	}
 
 	r = amdgpu_vm_init(adev, &fpriv->vm);
-	if (r)
-		goto error_free;
+	if (r) {
+		kfree(fpriv);
+		goto out_suspend;
+	}
 
 	mutex_init(&fpriv->bo_list_lock);
 	idr_init(&fpriv->bo_list_handles);
@@ -553,12 +560,9 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 
 	file_priv->driver_priv = fpriv;
 
+out_suspend:
 	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
-	return 0;
-
-error_free:
-	kfree(fpriv);
 
 	return r;
 }
@@ -597,6 +601,9 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
 
 	kfree(fpriv);
 	file_priv->driver_priv = NULL;
+
+	pm_runtime_mark_last_busy(dev->dev);
+	pm_runtime_put_autosuspend(dev->dev);
 }
 
 /**
@@ -611,6 +618,7 @@ void amdgpu_driver_postclose_kms(struct drm_device *dev,
 void amdgpu_driver_preclose_kms(struct drm_device *dev,
 				struct drm_file *file_priv)
 {
+	pm_runtime_get_sync(dev->dev);
 }
 
 /*

drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h

@@ -39,6 +39,8 @@
 #include <drm/drm_plane_helper.h>
 #include <linux/i2c.h>
 #include <linux/i2c-algo-bit.h>
+#include <linux/hrtimer.h>
+#include "amdgpu_irq.h"
 
 struct amdgpu_bo;
 struct amdgpu_device;
@@ -339,6 +341,8 @@ struct amdgpu_mode_info {
 	int			num_dig; /* number of dig blocks */
 	int			disp_priority;
 	const struct amdgpu_display_funcs *funcs;
+	struct hrtimer vblank_timer;
+	enum amdgpu_interrupt_state vsync_timer_enabled;
 };
 
 #define AMDGPU_MAX_BL_LEVEL 0xFF
@@ -587,10 +591,10 @@ int amdgpu_align_pitch(struct amdgpu_device *adev, int width, int bpp, bool tile
 void amdgpu_print_display_setup(struct drm_device *dev);
 int amdgpu_modeset_create_props(struct amdgpu_device *adev);
 int amdgpu_crtc_set_config(struct drm_mode_set *set);
-int amdgpu_crtc_page_flip(struct drm_crtc *crtc,
-			  struct drm_framebuffer *fb,
-			  struct drm_pending_vblank_event *event,
-			  uint32_t page_flip_flags);
+int amdgpu_crtc_page_flip_target(struct drm_crtc *crtc,
+				 struct drm_framebuffer *fb,
+				 struct drm_pending_vblank_event *event,
+				 uint32_t page_flip_flags, uint32_t target);
 extern const struct drm_mode_config_funcs amdgpu_mode_funcs;
 
 #endif

drivers/gpu/drm/amd/amdgpu/amdgpu_object.c

@@ -38,20 +38,17 @@
 #include "amdgpu_trace.h"
 
 
-int amdgpu_ttm_init(struct amdgpu_device *adev);
-void amdgpu_ttm_fini(struct amdgpu_device *adev);
 
 static u64 amdgpu_get_vis_part_size(struct amdgpu_device *adev,
 						struct ttm_mem_reg *mem)
 {
-	u64 ret = 0;
-	if (mem->start << PAGE_SHIFT < adev->mc.visible_vram_size) {
-		ret = (u64)((mem->start << PAGE_SHIFT) + mem->size) >
-			   adev->mc.visible_vram_size ?
-			   adev->mc.visible_vram_size - (mem->start << PAGE_SHIFT) :
-			   mem->size;
-	}
-	return ret;
+	if (mem->start << PAGE_SHIFT >= adev->mc.visible_vram_size)
+		return 0;
+
+	return ((mem->start << PAGE_SHIFT) + mem->size) >
+		adev->mc.visible_vram_size ?
+		adev->mc.visible_vram_size - (mem->start << PAGE_SHIFT) :
+		mem->size;
 }
 
 static void amdgpu_update_memory_usage(struct amdgpu_device *adev,
@@ -99,6 +96,11 @@ static void amdgpu_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 
 	drm_gem_object_release(&bo->gem_base);
 	amdgpu_bo_unref(&bo->parent);
+	if (!list_empty(&bo->shadow_list)) {
+		mutex_lock(&bo->adev->shadow_list_lock);
+		list_del_init(&bo->shadow_list);
+		mutex_unlock(&bo->adev->shadow_list_lock);
+	}
 	kfree(bo->metadata);
 	kfree(bo);
 }
@@ -112,90 +114,99 @@ bool amdgpu_ttm_bo_is_amdgpu_bo(struct ttm_buffer_object *bo)
 
 static void amdgpu_ttm_placement_init(struct amdgpu_device *adev,
 				      struct ttm_placement *placement,
-				      struct ttm_place *placements,
+				      struct ttm_place *places,
 				      u32 domain, u64 flags)
 {
-	u32 c = 0, i;
-
-	placement->placement = placements;
-	placement->busy_placement = placements;
+	u32 c = 0;
 
 	if (domain & AMDGPU_GEM_DOMAIN_VRAM) {
+		unsigned visible_pfn = adev->mc.visible_vram_size >> PAGE_SHIFT;
+
 		if (flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS &&
-			adev->mc.visible_vram_size < adev->mc.real_vram_size) {
-			placements[c].fpfn =
-				adev->mc.visible_vram_size >> PAGE_SHIFT;
-			placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
-				TTM_PL_FLAG_VRAM | TTM_PL_FLAG_TOPDOWN;
+		    !(flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) &&
+		    adev->mc.visible_vram_size < adev->mc.real_vram_size) {
+			places[c].fpfn = visible_pfn;
+			places[c].lpfn = 0;
+			places[c].flags = TTM_PL_FLAG_WC |
+				TTM_PL_FLAG_UNCACHED | TTM_PL_FLAG_VRAM |
+				TTM_PL_FLAG_TOPDOWN;
+			c++;
 		}
-		placements[c].fpfn = 0;
-		placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
+
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_UNCACHED |
 			TTM_PL_FLAG_VRAM;
-		if (!(flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED))
-			placements[c - 1].flags |= TTM_PL_FLAG_TOPDOWN;
+		if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED)
+			places[c].lpfn = visible_pfn;
+		else
+			places[c].flags |= TTM_PL_FLAG_TOPDOWN;
+		c++;
 	}
 
 	if (domain & AMDGPU_GEM_DOMAIN_GTT) {
-		if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC) {
-			placements[c].fpfn = 0;
-			placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_TT |
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].flags = TTM_PL_FLAG_TT;
+		if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
+			places[c].flags |= TTM_PL_FLAG_WC |
 				TTM_PL_FLAG_UNCACHED;
-		} else {
-			placements[c].fpfn = 0;
-			placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_TT;
-		}
+		else
+			places[c].flags |= TTM_PL_FLAG_CACHED;
+		c++;
 	}
 
 	if (domain & AMDGPU_GEM_DOMAIN_CPU) {
-		if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC) {
-			placements[c].fpfn = 0;
-			placements[c++].flags = TTM_PL_FLAG_WC | TTM_PL_FLAG_SYSTEM |
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].flags = TTM_PL_FLAG_SYSTEM;
+		if (flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC)
+			places[c].flags |= TTM_PL_FLAG_WC |
 				TTM_PL_FLAG_UNCACHED;
-		} else {
-			placements[c].fpfn = 0;
-			placements[c++].flags = TTM_PL_FLAG_CACHED | TTM_PL_FLAG_SYSTEM;
-		}
+		else
+			places[c].flags |= TTM_PL_FLAG_CACHED;
+		c++;
 	}
 
 	if (domain & AMDGPU_GEM_DOMAIN_GDS) {
-		placements[c].fpfn = 0;
-		placements[c++].flags = TTM_PL_FLAG_UNCACHED |
-			AMDGPU_PL_FLAG_GDS;
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_GDS;
+		c++;
 	}
+
 	if (domain & AMDGPU_GEM_DOMAIN_GWS) {
-		placements[c].fpfn = 0;
-		placements[c++].flags = TTM_PL_FLAG_UNCACHED |
-			AMDGPU_PL_FLAG_GWS;
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_GWS;
+		c++;
 	}
+
 	if (domain & AMDGPU_GEM_DOMAIN_OA) {
-		placements[c].fpfn = 0;
-		placements[c++].flags = TTM_PL_FLAG_UNCACHED |
-			AMDGPU_PL_FLAG_OA;
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].flags = TTM_PL_FLAG_UNCACHED | AMDGPU_PL_FLAG_OA;
+		c++;
 	}
 
 	if (!c) {
-		placements[c].fpfn = 0;
-		placements[c++].flags = TTM_PL_MASK_CACHING |
-			TTM_PL_FLAG_SYSTEM;
+		places[c].fpfn = 0;
+		places[c].lpfn = 0;
+		places[c].flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM;
+		c++;
 	}
+
 	placement->num_placement = c;
-	placement->num_busy_placement = c;
+	placement->placement = places;
 
-	for (i = 0; i < c; i++) {
-		if ((flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) &&
-			(placements[i].flags & TTM_PL_FLAG_VRAM) &&
-			!placements[i].fpfn)
-			placements[i].lpfn =
-				adev->mc.visible_vram_size >> PAGE_SHIFT;
-		else
-			placements[i].lpfn = 0;
-	}
+	placement->num_busy_placement = c;
+	placement->busy_placement = places;
 }
 
-void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *rbo, u32 domain)
+void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *abo, u32 domain)
 {
-	amdgpu_ttm_placement_init(rbo->adev, &rbo->placement,
-				  rbo->placements, domain, rbo->flags);
+	amdgpu_ttm_placement_init(abo->adev, &abo->placement,
+				  abo->placements, domain, abo->flags);
 }
 
 static void amdgpu_fill_placement_to_bo(struct amdgpu_bo *bo,
@@ -211,6 +222,98 @@ static void amdgpu_fill_placement_to_bo(struct amdgpu_bo *bo,
 	bo->placement.busy_placement = bo->placements;
 }
 
+/**
+ * amdgpu_bo_create_kernel - create BO for kernel use
+ *
+ * @adev: amdgpu device object
+ * @size: size for the new BO
+ * @align: alignment for the new BO
+ * @domain: where to place it
+ * @bo_ptr: resulting BO
+ * @gpu_addr: GPU addr of the pinned BO
+ * @cpu_addr: optional CPU address mapping
+ *
+ * Allocates and pins a BO for kernel internal use.
+ *
+ * Returns 0 on success, negative error code otherwise.
+ */
+int amdgpu_bo_create_kernel(struct amdgpu_device *adev,
+			    unsigned long size, int align,
+			    u32 domain, struct amdgpu_bo **bo_ptr,
+			    u64 *gpu_addr, void **cpu_addr)
+{
+	int r;
+
+	r = amdgpu_bo_create(adev, size, align, true, domain,
+			     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
+			     NULL, NULL, bo_ptr);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to allocate kernel bo\n", r);
+		return r;
+	}
+
+	r = amdgpu_bo_reserve(*bo_ptr, false);
+	if (r) {
+		dev_err(adev->dev, "(%d) failed to reserve kernel bo\n", r);
+		goto error_free;
+	}
+
+	r = amdgpu_bo_pin(*bo_ptr, domain, gpu_addr);
+	if (r) {
+		dev_err(adev->dev, "(%d) kernel bo pin failed\n", r);
+		goto error_unreserve;
+	}
+
+	if (cpu_addr) {
+		r = amdgpu_bo_kmap(*bo_ptr, cpu_addr);
+		if (r) {
+			dev_err(adev->dev, "(%d) kernel bo map failed\n", r);
+			goto error_unreserve;
+		}
+	}
+
+	amdgpu_bo_unreserve(*bo_ptr);
+
+	return 0;
+
+error_unreserve:
+	amdgpu_bo_unreserve(*bo_ptr);
+
+error_free:
+	amdgpu_bo_unref(bo_ptr);
+
+	return r;
+}
+
+/**
+ * amdgpu_bo_free_kernel - free BO for kernel use
+ *
+ * @bo: amdgpu BO to free
+ *
+ * unmaps and unpin a BO for kernel internal use.
+ */
+void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
+			   void **cpu_addr)
+{
+	if (*bo == NULL)
+		return;
+
+	if (likely(amdgpu_bo_reserve(*bo, false) == 0)) {
+		if (cpu_addr)
+			amdgpu_bo_kunmap(*bo);
+
+		amdgpu_bo_unpin(*bo);
+		amdgpu_bo_unreserve(*bo);
+	}
+	amdgpu_bo_unref(bo);
+
+	if (gpu_addr)
+		*gpu_addr = 0;
+
+	if (cpu_addr)
+		*cpu_addr = NULL;
+}
+
 int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
 				unsigned long size, int byte_align,
 				bool kernel, u32 domain, u64 flags,
@@ -249,7 +352,7 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
 		return r;
 	}
 	bo->adev = adev;
-	INIT_LIST_HEAD(&bo->list);
+	INIT_LIST_HEAD(&bo->shadow_list);
 	INIT_LIST_HEAD(&bo->va);
 	bo->prefered_domains = domain & (AMDGPU_GEM_DOMAIN_VRAM |
 					 AMDGPU_GEM_DOMAIN_GTT |
@@ -277,11 +380,79 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
 	if (unlikely(r != 0)) {
 		return r;
 	}
+
+	if (flags & AMDGPU_GEM_CREATE_VRAM_CLEARED &&
+	    bo->tbo.mem.placement & TTM_PL_FLAG_VRAM) {
+		struct fence *fence;
+
+		if (adev->mman.buffer_funcs_ring == NULL ||
+		   !adev->mman.buffer_funcs_ring->ready) {
+			r = -EBUSY;
+			goto fail_free;
+		}
+
+		r = amdgpu_bo_reserve(bo, false);
+		if (unlikely(r != 0))
+			goto fail_free;
+
+		amdgpu_ttm_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_VRAM);
+		r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
+		if (unlikely(r != 0))
+			goto fail_unreserve;
+
+		amdgpu_fill_buffer(bo, 0, bo->tbo.resv, &fence);
+		amdgpu_bo_fence(bo, fence, false);
+		amdgpu_bo_unreserve(bo);
+		fence_put(bo->tbo.moving);
+		bo->tbo.moving = fence_get(fence);
+		fence_put(fence);
+	}
 	*bo_ptr = bo;
 
 	trace_amdgpu_bo_create(bo);
 
 	return 0;
+
+fail_unreserve:
+	amdgpu_bo_unreserve(bo);
+fail_free:
+	amdgpu_bo_unref(&bo);
+	return r;
+}
+
+static int amdgpu_bo_create_shadow(struct amdgpu_device *adev,
+				   unsigned long size, int byte_align,
+				   struct amdgpu_bo *bo)
+{
+	struct ttm_placement placement = {0};
+	struct ttm_place placements[AMDGPU_GEM_DOMAIN_MAX + 1];
+	int r;
+
+	if (bo->shadow)
+		return 0;
+
+	bo->flags |= AMDGPU_GEM_CREATE_SHADOW;
+	memset(&placements, 0,
+	       (AMDGPU_GEM_DOMAIN_MAX + 1) * sizeof(struct ttm_place));
+
+	amdgpu_ttm_placement_init(adev, &placement,
+				  placements, AMDGPU_GEM_DOMAIN_GTT,
+				  AMDGPU_GEM_CREATE_CPU_GTT_USWC);
+
+	r = amdgpu_bo_create_restricted(adev, size, byte_align, true,
+					AMDGPU_GEM_DOMAIN_GTT,
+					AMDGPU_GEM_CREATE_CPU_GTT_USWC,
+					NULL, &placement,
+					bo->tbo.resv,
+					&bo->shadow);
+	if (!r) {
+		bo->shadow->parent = amdgpu_bo_ref(bo);
+		mutex_lock(&adev->shadow_list_lock);
+		list_add_tail(&bo->shadow_list, &adev->shadow_list);
+		mutex_unlock(&adev->shadow_list_lock);
+	}
+
+	return r;
 }
 
 int amdgpu_bo_create(struct amdgpu_device *adev,
@@ -293,6 +464,7 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 {
 	struct ttm_placement placement = {0};
 	struct ttm_place placements[AMDGPU_GEM_DOMAIN_MAX + 1];
+	int r;
 
 	memset(&placements, 0,
 	       (AMDGPU_GEM_DOMAIN_MAX + 1) * sizeof(struct ttm_place));
@@ -300,9 +472,83 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 	amdgpu_ttm_placement_init(adev, &placement,
 				  placements, domain, flags);
 
-	return amdgpu_bo_create_restricted(adev, size, byte_align, kernel,
-					   domain, flags, sg, &placement,
-					   resv, bo_ptr);
+	r = amdgpu_bo_create_restricted(adev, size, byte_align, kernel,
+					domain, flags, sg, &placement,
+					resv, bo_ptr);
+	if (r)
+		return r;
+
+	if (amdgpu_need_backup(adev) && (flags & AMDGPU_GEM_CREATE_SHADOW)) {
+		r = amdgpu_bo_create_shadow(adev, size, byte_align, (*bo_ptr));
+		if (r)
+			amdgpu_bo_unref(bo_ptr);
+	}
+
+	return r;
+}
+
+int amdgpu_bo_backup_to_shadow(struct amdgpu_device *adev,
+			       struct amdgpu_ring *ring,
+			       struct amdgpu_bo *bo,
+			       struct reservation_object *resv,
+			       struct fence **fence,
+			       bool direct)
+
+{
+	struct amdgpu_bo *shadow = bo->shadow;
+	uint64_t bo_addr, shadow_addr;
+	int r;
+
+	if (!shadow)
+		return -EINVAL;
+
+	bo_addr = amdgpu_bo_gpu_offset(bo);
+	shadow_addr = amdgpu_bo_gpu_offset(bo->shadow);
+
+	r = reservation_object_reserve_shared(bo->tbo.resv);
+	if (r)
+		goto err;
+
+	r = amdgpu_copy_buffer(ring, bo_addr, shadow_addr,
+			       amdgpu_bo_size(bo), resv, fence,
+			       direct);
+	if (!r)
+		amdgpu_bo_fence(bo, *fence, true);
+
+err:
+	return r;
+}
+
+int amdgpu_bo_restore_from_shadow(struct amdgpu_device *adev,
+				  struct amdgpu_ring *ring,
+				  struct amdgpu_bo *bo,
+				  struct reservation_object *resv,
+				  struct fence **fence,
+				  bool direct)
+
+{
+	struct amdgpu_bo *shadow = bo->shadow;
+	uint64_t bo_addr, shadow_addr;
+	int r;
+
+	if (!shadow)
+		return -EINVAL;
+
+	bo_addr = amdgpu_bo_gpu_offset(bo);
+	shadow_addr = amdgpu_bo_gpu_offset(bo->shadow);
+
+	r = reservation_object_reserve_shared(bo->tbo.resv);
+	if (r)
+		goto err;
+
+	r = amdgpu_copy_buffer(ring, shadow_addr, bo_addr,
+			       amdgpu_bo_size(bo), resv, fence,
+			       direct);
+	if (!r)
+		amdgpu_bo_fence(bo, *fence, true);
+
+err:
+	return r;
 }
 
 int amdgpu_bo_kmap(struct amdgpu_bo *bo, void **ptr)
@@ -380,16 +626,17 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 		return -EINVAL;
 
 	if (bo->pin_count) {
+		uint32_t mem_type = bo->tbo.mem.mem_type;
+
+		if (domain != amdgpu_mem_type_to_domain(mem_type))
+			return -EINVAL;
+
 		bo->pin_count++;
 		if (gpu_addr)
 			*gpu_addr = amdgpu_bo_gpu_offset(bo);
 
 		if (max_offset != 0) {
-			u64 domain_start;
-			if (domain == AMDGPU_GEM_DOMAIN_VRAM)
-				domain_start = bo->adev->mc.vram_start;
-			else
-				domain_start = bo->adev->mc.gtt_start;
+			u64 domain_start = bo->tbo.bdev->man[mem_type].gpu_offset;
 			WARN_ON_ONCE(max_offset <
 				     (amdgpu_bo_gpu_offset(bo) - domain_start));
 		}
@@ -401,7 +648,8 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 		/* force to pin into visible video ram */
 		if ((bo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
 		    !(bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS) &&
-		    (!max_offset || max_offset > bo->adev->mc.visible_vram_size)) {
+		    (!max_offset || max_offset >
+		     bo->adev->mc.visible_vram_size)) {
 			if (WARN_ON_ONCE(min_offset >
 					 bo->adev->mc.visible_vram_size))
 				return -EINVAL;
@@ -420,19 +668,28 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain,
 	}
 
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
-	if (likely(r == 0)) {
-		bo->pin_count = 1;
-		if (gpu_addr != NULL)
-			*gpu_addr = amdgpu_bo_gpu_offset(bo);
-		if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
-			bo->adev->vram_pin_size += amdgpu_bo_size(bo);
-			if (bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS)
-				bo->adev->invisible_pin_size += amdgpu_bo_size(bo);
-		} else
-			bo->adev->gart_pin_size += amdgpu_bo_size(bo);
-	} else {
+	if (unlikely(r)) {
 		dev_err(bo->adev->dev, "%p pin failed\n", bo);
+		goto error;
+	}
+	r = amdgpu_ttm_bind(&bo->tbo, &bo->tbo.mem);
+	if (unlikely(r)) {
+		dev_err(bo->adev->dev, "%p bind failed\n", bo);
+		goto error;
 	}
+
+	bo->pin_count = 1;
+	if (gpu_addr != NULL)
+		*gpu_addr = amdgpu_bo_gpu_offset(bo);
+	if (domain == AMDGPU_GEM_DOMAIN_VRAM) {
+		bo->adev->vram_pin_size += amdgpu_bo_size(bo);
+		if (bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS)
+			bo->adev->invisible_pin_size += amdgpu_bo_size(bo);
+	} else if (domain == AMDGPU_GEM_DOMAIN_GTT) {
+		bo->adev->gart_pin_size += amdgpu_bo_size(bo);
+	}
+
+error:
 	return r;
 }
 
@@ -457,16 +714,20 @@ int amdgpu_bo_unpin(struct amdgpu_bo *bo)
 		bo->placements[i].flags &= ~TTM_PL_FLAG_NO_EVICT;
 	}
 	r = ttm_bo_validate(&bo->tbo, &bo->placement, false, false);
-	if (likely(r == 0)) {
-		if (bo->tbo.mem.mem_type == TTM_PL_VRAM) {
-			bo->adev->vram_pin_size -= amdgpu_bo_size(bo);
-			if (bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS)
-				bo->adev->invisible_pin_size -= amdgpu_bo_size(bo);
-		} else
-			bo->adev->gart_pin_size -= amdgpu_bo_size(bo);
-	} else {
+	if (unlikely(r)) {
 		dev_err(bo->adev->dev, "%p validate failed for unpin\n", bo);
+		goto error;
 	}
+
+	if (bo->tbo.mem.mem_type == TTM_PL_VRAM) {
+		bo->adev->vram_pin_size -= amdgpu_bo_size(bo);
+		if (bo->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS)
+			bo->adev->invisible_pin_size -= amdgpu_bo_size(bo);
+	} else if (bo->tbo.mem.mem_type == TTM_PL_TT) {
+		bo->adev->gart_pin_size -= amdgpu_bo_size(bo);
+	}
+
+error:
 	return r;
 }
 
@@ -588,23 +849,23 @@ int amdgpu_bo_get_metadata(struct amdgpu_bo *bo, void *buffer,
 void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 			   struct ttm_mem_reg *new_mem)
 {
-	struct amdgpu_bo *rbo;
+	struct amdgpu_bo *abo;
 	struct ttm_mem_reg *old_mem = &bo->mem;
 
 	if (!amdgpu_ttm_bo_is_amdgpu_bo(bo))
 		return;
 
-	rbo = container_of(bo, struct amdgpu_bo, tbo);
-	amdgpu_vm_bo_invalidate(rbo->adev, rbo);
+	abo = container_of(bo, struct amdgpu_bo, tbo);
+	amdgpu_vm_bo_invalidate(abo->adev, abo);
 
 	/* update statistics */
 	if (!new_mem)
 		return;
 
 	/* move_notify is called before move happens */
-	amdgpu_update_memory_usage(rbo->adev, &bo->mem, new_mem);
+	amdgpu_update_memory_usage(abo->adev, &bo->mem, new_mem);
 
-	trace_amdgpu_ttm_bo_move(rbo, new_mem->mem_type, old_mem->mem_type);
+	trace_amdgpu_ttm_bo_move(abo, new_mem->mem_type, old_mem->mem_type);
 }
 
 int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
@@ -637,7 +898,8 @@ int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 	for (i = 0; i < abo->placement.num_placement; i++) {
 		/* Force into visible VRAM */
 		if ((abo->placements[i].flags & TTM_PL_FLAG_VRAM) &&
-		    (!abo->placements[i].lpfn || abo->placements[i].lpfn > lpfn))
+		    (!abo->placements[i].lpfn ||
+		     abo->placements[i].lpfn > lpfn))
 			abo->placements[i].lpfn = lpfn;
 	}
 	r = ttm_bo_validate(bo, &abo->placement, false, false);
@@ -674,3 +936,24 @@ void amdgpu_bo_fence(struct amdgpu_bo *bo, struct fence *fence,
 	else
 		reservation_object_add_excl_fence(resv, fence);
 }
+
+/**
+ * amdgpu_bo_gpu_offset - return GPU offset of bo
+ * @bo:	amdgpu object for which we query the offset
+ *
+ * Returns current GPU offset of the object.
+ *
+ * Note: object should either be pinned or reserved when calling this
+ * function, it might be useful to add check for this for debugging.
+ */
+u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo)
+{
+	WARN_ON_ONCE(bo->tbo.mem.mem_type == TTM_PL_SYSTEM);
+	WARN_ON_ONCE(bo->tbo.mem.mem_type == TTM_PL_TT &&
+		     !amdgpu_ttm_is_bound(bo->tbo.ttm));
+	WARN_ON_ONCE(!ww_mutex_is_locked(&bo->tbo.resv->lock) &&
+		     !bo->pin_count);
+	WARN_ON_ONCE(bo->tbo.mem.start == AMDGPU_BO_INVALID_OFFSET);
+
+	return bo->tbo.offset;
+}

drivers/gpu/drm/amd/amdgpu/amdgpu_object.h

@@ -31,6 +31,8 @@
 #include <drm/amdgpu_drm.h>
 #include "amdgpu.h"
 
+#define AMDGPU_BO_INVALID_OFFSET	LONG_MAX
+
 /**
  * amdgpu_mem_type_to_domain - return domain corresponding to mem_type
  * @mem_type:	ttm memory type
@@ -85,21 +87,6 @@ static inline void amdgpu_bo_unreserve(struct amdgpu_bo *bo)
 	ttm_bo_unreserve(&bo->tbo);
 }
 
-/**
- * amdgpu_bo_gpu_offset - return GPU offset of bo
- * @bo:	amdgpu object for which we query the offset
- *
- * Returns current GPU offset of the object.
- *
- * Note: object should either be pinned or reserved when calling this
- * function, it might be useful to add check for this for debugging.
- */
-static inline u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo)
-{
-	WARN_ON_ONCE(bo->tbo.mem.mem_type == TTM_PL_SYSTEM);
-	return bo->tbo.offset;
-}
-
 static inline unsigned long amdgpu_bo_size(struct amdgpu_bo *bo)
 {
 	return bo->tbo.num_pages << PAGE_SHIFT;
@@ -139,6 +126,12 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev,
 				struct ttm_placement *placement,
 			        struct reservation_object *resv,
 				struct amdgpu_bo **bo_ptr);
+int amdgpu_bo_create_kernel(struct amdgpu_device *adev,
+			    unsigned long size, int align,
+			    u32 domain, struct amdgpu_bo **bo_ptr,
+			    u64 *gpu_addr, void **cpu_addr);
+void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr,
+			   void **cpu_addr);
 int amdgpu_bo_kmap(struct amdgpu_bo *bo, void **ptr);
 void amdgpu_bo_kunmap(struct amdgpu_bo *bo);
 struct amdgpu_bo *amdgpu_bo_ref(struct amdgpu_bo *bo);
@@ -165,6 +158,19 @@ void amdgpu_bo_move_notify(struct ttm_buffer_object *bo,
 int amdgpu_bo_fault_reserve_notify(struct ttm_buffer_object *bo);
 void amdgpu_bo_fence(struct amdgpu_bo *bo, struct fence *fence,
 		     bool shared);
+u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo);
+int amdgpu_bo_backup_to_shadow(struct amdgpu_device *adev,
+			       struct amdgpu_ring *ring,
+			       struct amdgpu_bo *bo,
+			       struct reservation_object *resv,
+			       struct fence **fence, bool direct);
+int amdgpu_bo_restore_from_shadow(struct amdgpu_device *adev,
+				  struct amdgpu_ring *ring,
+				  struct amdgpu_bo *bo,
+				  struct reservation_object *resv,
+				  struct fence **fence,
+				  bool direct);
+
 
 /*
  * sub allocation

drivers/gpu/drm/amd/amdgpu/amdgpu_pll.c

@@ -25,6 +25,7 @@
 #include "amdgpu.h"
 #include "atom.h"
 #include "atombios_encoders.h"
+#include "amdgpu_pll.h"
 #include <asm/div64.h>
 #include <linux/gcd.h>
 

drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c

@@ -1103,54 +1103,46 @@ force:
 
 void amdgpu_dpm_enable_uvd(struct amdgpu_device *adev, bool enable)
 {
-	if (adev->pp_enabled)
+	if (adev->pp_enabled || adev->pm.funcs->powergate_uvd) {
+		/* enable/disable UVD */
+		mutex_lock(&adev->pm.mutex);
 		amdgpu_dpm_powergate_uvd(adev, !enable);
-	else {
-		if (adev->pm.funcs->powergate_uvd) {
+		mutex_unlock(&adev->pm.mutex);
+	} else {
+		if (enable) {
 			mutex_lock(&adev->pm.mutex);
-			/* enable/disable UVD */
-			amdgpu_dpm_powergate_uvd(adev, !enable);
+			adev->pm.dpm.uvd_active = true;
+			adev->pm.dpm.state = POWER_STATE_TYPE_INTERNAL_UVD;
 			mutex_unlock(&adev->pm.mutex);
 		} else {
-			if (enable) {
-				mutex_lock(&adev->pm.mutex);
-				adev->pm.dpm.uvd_active = true;
-				adev->pm.dpm.state = POWER_STATE_TYPE_INTERNAL_UVD;
-				mutex_unlock(&adev->pm.mutex);
-			} else {
-				mutex_lock(&adev->pm.mutex);
-				adev->pm.dpm.uvd_active = false;
-				mutex_unlock(&adev->pm.mutex);
-			}
-			amdgpu_pm_compute_clocks(adev);
+			mutex_lock(&adev->pm.mutex);
+			adev->pm.dpm.uvd_active = false;
+			mutex_unlock(&adev->pm.mutex);
 		}
-
+		amdgpu_pm_compute_clocks(adev);
 	}
 }
 
 void amdgpu_dpm_enable_vce(struct amdgpu_device *adev, bool enable)
 {
-	if (adev->pp_enabled)
+	if (adev->pp_enabled || adev->pm.funcs->powergate_vce) {
+		/* enable/disable VCE */
+		mutex_lock(&adev->pm.mutex);
 		amdgpu_dpm_powergate_vce(adev, !enable);
-	else {
-		if (adev->pm.funcs->powergate_vce) {
+		mutex_unlock(&adev->pm.mutex);
+	} else {
+		if (enable) {
 			mutex_lock(&adev->pm.mutex);
-			amdgpu_dpm_powergate_vce(adev, !enable);
+			adev->pm.dpm.vce_active = true;
+			/* XXX select vce level based on ring/task */
+			adev->pm.dpm.vce_level = AMDGPU_VCE_LEVEL_AC_ALL;
 			mutex_unlock(&adev->pm.mutex);
 		} else {
-			if (enable) {
-				mutex_lock(&adev->pm.mutex);
-				adev->pm.dpm.vce_active = true;
-				/* XXX select vce level based on ring/task */
-				adev->pm.dpm.vce_level = AMDGPU_VCE_LEVEL_AC_ALL;
-				mutex_unlock(&adev->pm.mutex);
-			} else {
-				mutex_lock(&adev->pm.mutex);
-				adev->pm.dpm.vce_active = false;
-				mutex_unlock(&adev->pm.mutex);
-			}
-			amdgpu_pm_compute_clocks(adev);
+			mutex_lock(&adev->pm.mutex);
+			adev->pm.dpm.vce_active = false;
+			mutex_unlock(&adev->pm.mutex);
 		}
+		amdgpu_pm_compute_clocks(adev);
 	}
 }
 
@@ -1330,6 +1322,64 @@ void amdgpu_pm_compute_clocks(struct amdgpu_device *adev)
  */
 #if defined(CONFIG_DEBUG_FS)
 
+static int amdgpu_debugfs_pm_info_pp(struct seq_file *m, struct amdgpu_device *adev)
+{
+	int32_t value;
+
+	/* sanity check PP is enabled */
+	if (!(adev->powerplay.pp_funcs &&
+	      adev->powerplay.pp_funcs->read_sensor))
+	      return -EINVAL;
+
+	/* GPU Clocks */
+	seq_printf(m, "GFX Clocks and Power:\n");
+	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GFX_MCLK, &value))
+		seq_printf(m, "\t%u MHz (MCLK)\n", value/100);
+	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GFX_SCLK, &value))
+		seq_printf(m, "\t%u MHz (SCLK)\n", value/100);
+	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_VDDGFX, &value))
+		seq_printf(m, "\t%u mV (VDDGFX)\n", value);
+	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_VDDNB, &value))
+		seq_printf(m, "\t%u mV (VDDNB)\n", value);
+	seq_printf(m, "\n");
+
+	/* GPU Temp */
+	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GPU_TEMP, &value))
+		seq_printf(m, "GPU Temperature: %u C\n", value/1000);
+
+	/* GPU Load */
+	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_GPU_LOAD, &value))
+		seq_printf(m, "GPU Load: %u %%\n", value);
+	seq_printf(m, "\n");
+
+	/* UVD clocks */
+	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_UVD_POWER, &value)) {
+		if (!value) {
+			seq_printf(m, "UVD: Disabled\n");
+		} else {
+			seq_printf(m, "UVD: Enabled\n");
+			if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_UVD_DCLK, &value))
+				seq_printf(m, "\t%u MHz (DCLK)\n", value/100);
+			if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_UVD_VCLK, &value))
+				seq_printf(m, "\t%u MHz (VCLK)\n", value/100);
+		}
+	}
+	seq_printf(m, "\n");
+
+	/* VCE clocks */
+	if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_VCE_POWER, &value)) {
+		if (!value) {
+			seq_printf(m, "VCE: Disabled\n");
+		} else {
+			seq_printf(m, "VCE: Enabled\n");
+			if (!amdgpu_dpm_read_sensor(adev, AMDGPU_PP_SENSOR_VCE_ECCLK, &value))
+				seq_printf(m, "\t%u MHz (ECCLK)\n", value/100);
+		}
+	}
+
+	return 0;
+}
+
 static int amdgpu_debugfs_pm_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -1345,11 +1395,11 @@ static int amdgpu_debugfs_pm_info(struct seq_file *m, void *data)
 	     (ddev->switch_power_state != DRM_SWITCH_POWER_ON)) {
 		seq_printf(m, "PX asic powered off\n");
 	} else if (adev->pp_enabled) {
-		amdgpu_dpm_debugfs_print_current_performance_level(adev, m);
+		return amdgpu_debugfs_pm_info_pp(m, adev);
 	} else {
 		mutex_lock(&adev->pm.mutex);
 		if (adev->pm.funcs->debugfs_print_current_performance_level)
-			amdgpu_dpm_debugfs_print_current_performance_level(adev, m);
+			adev->pm.funcs->debugfs_print_current_performance_level(adev, m);
 		else
 			seq_printf(m, "Debugfs support not implemented for this asic\n");
 		mutex_unlock(&adev->pm.mutex);

drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c

@@ -30,6 +30,7 @@
 #include "amdgpu_pm.h"
 #include <drm/amdgpu_drm.h>
 #include "amdgpu_powerplay.h"
+#include "si_dpm.h"
 #include "cik_dpm.h"
 #include "vi_dpm.h"
 
@@ -41,7 +42,6 @@ static int amdgpu_powerplay_init(struct amdgpu_device *adev)
 	amd_pp = &(adev->powerplay);
 
 	if (adev->pp_enabled) {
-#ifdef CONFIG_DRM_AMD_POWERPLAY
 		struct amd_pp_init *pp_init;
 
 		pp_init = kzalloc(sizeof(struct amd_pp_init), GFP_KERNEL);
@@ -52,15 +52,21 @@ static int amdgpu_powerplay_init(struct amdgpu_device *adev)
 		pp_init->chip_family = adev->family;
 		pp_init->chip_id = adev->asic_type;
 		pp_init->device = amdgpu_cgs_create_device(adev);
-		pp_init->powercontainment_enabled = amdgpu_powercontainment;
-
 		ret = amd_powerplay_init(pp_init, amd_pp);
 		kfree(pp_init);
-#endif
 	} else {
 		amd_pp->pp_handle = (void *)adev;
 
 		switch (adev->asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_SI
+		case CHIP_TAHITI:
+		case CHIP_PITCAIRN:
+		case CHIP_VERDE:
+		case CHIP_OLAND:
+		case CHIP_HAINAN:
+			amd_pp->ip_funcs = &si_dpm_ip_funcs;
+		break;
+#endif
 #ifdef CONFIG_DRM_AMDGPU_CIK
 		case CHIP_BONAIRE:
 		case CHIP_HAWAII:
@@ -72,15 +78,6 @@ static int amdgpu_powerplay_init(struct amdgpu_device *adev)
 			amd_pp->ip_funcs = &kv_dpm_ip_funcs;
 			break;
 #endif
-		case CHIP_TOPAZ:
-			amd_pp->ip_funcs = &iceland_dpm_ip_funcs;
-			break;
-		case CHIP_TONGA:
-			amd_pp->ip_funcs = &tonga_dpm_ip_funcs;
-			break;
-		case CHIP_FIJI:
-			amd_pp->ip_funcs = &fiji_dpm_ip_funcs;
-			break;
 		case CHIP_CARRIZO:
 		case CHIP_STONEY:
 			amd_pp->ip_funcs = &cz_dpm_ip_funcs;
@@ -98,19 +95,17 @@ static int amdgpu_pp_early_init(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int ret = 0;
 
-#ifdef CONFIG_DRM_AMD_POWERPLAY
 	switch (adev->asic_type) {
 	case CHIP_POLARIS11:
 	case CHIP_POLARIS10:
-		adev->pp_enabled = true;
-		break;
 	case CHIP_TONGA:
 	case CHIP_FIJI:
-		adev->pp_enabled = (amdgpu_powerplay == 0) ? false : true;
+	case CHIP_TOPAZ:
+		adev->pp_enabled = true;
 		break;
 	case CHIP_CARRIZO:
 	case CHIP_STONEY:
-		adev->pp_enabled = (amdgpu_powerplay > 0) ? true : false;
+		adev->pp_enabled = (amdgpu_powerplay == 0) ? false : true;
 		break;
 	/* These chips don't have powerplay implemenations */
 	case CHIP_BONAIRE:
@@ -118,14 +113,10 @@ static int amdgpu_pp_early_init(void *handle)
 	case CHIP_KABINI:
 	case CHIP_MULLINS:
 	case CHIP_KAVERI:
-	case CHIP_TOPAZ:
 	default:
 		adev->pp_enabled = false;
 		break;
 	}
-#else
-	adev->pp_enabled = false;
-#endif
 
 	ret = amdgpu_powerplay_init(adev);
 	if (ret)
@@ -147,12 +138,11 @@ static int amdgpu_pp_late_init(void *handle)
 		ret = adev->powerplay.ip_funcs->late_init(
 					adev->powerplay.pp_handle);
 
-#ifdef CONFIG_DRM_AMD_POWERPLAY
 	if (adev->pp_enabled && adev->pm.dpm_enabled) {
 		amdgpu_pm_sysfs_init(adev);
 		amdgpu_dpm_dispatch_task(adev, AMD_PP_EVENT_COMPLETE_INIT, NULL, NULL);
 	}
-#endif
+
 	return ret;
 }
 
@@ -165,10 +155,8 @@ static int amdgpu_pp_sw_init(void *handle)
 		ret = adev->powerplay.ip_funcs->sw_init(
 					adev->powerplay.pp_handle);
 
-#ifdef CONFIG_DRM_AMD_POWERPLAY
 	if (adev->pp_enabled)
 		adev->pm.dpm_enabled = true;
-#endif
 
 	return ret;
 }
@@ -219,7 +207,6 @@ static int amdgpu_pp_hw_fini(void *handle)
 
 static void amdgpu_pp_late_fini(void *handle)
 {
-#ifdef CONFIG_DRM_AMD_POWERPLAY
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
 	if (adev->pp_enabled) {
@@ -230,7 +217,6 @@ static void amdgpu_pp_late_fini(void *handle)
 	if (adev->powerplay.ip_funcs->late_fini)
 		adev->powerplay.ip_funcs->late_fini(
 			  adev->powerplay.pp_handle);
-#endif
 }
 
 static int amdgpu_pp_suspend(void *handle)

drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c

@@ -222,33 +222,16 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 
 	/* Allocate ring buffer */
 	if (ring->ring_obj == NULL) {
-		r = amdgpu_bo_create(adev, ring->ring_size, PAGE_SIZE, true,
-				     AMDGPU_GEM_DOMAIN_GTT, 0,
-				     NULL, NULL, &ring->ring_obj);
+		r = amdgpu_bo_create_kernel(adev, ring->ring_size, PAGE_SIZE,
+					    AMDGPU_GEM_DOMAIN_GTT,
+					    &ring->ring_obj,
+					    &ring->gpu_addr,
+					    (void **)&ring->ring);
 		if (r) {
 			dev_err(adev->dev, "(%d) ring create failed\n", r);
 			return r;
 		}
-		r = amdgpu_bo_reserve(ring->ring_obj, false);
-		if (unlikely(r != 0))
-			return r;
-		r = amdgpu_bo_pin(ring->ring_obj, AMDGPU_GEM_DOMAIN_GTT,
-					&ring->gpu_addr);
-		if (r) {
-			amdgpu_bo_unreserve(ring->ring_obj);
-			dev_err(adev->dev, "(%d) ring pin failed\n", r);
-			return r;
-		}
-		r = amdgpu_bo_kmap(ring->ring_obj,
-				       (void **)&ring->ring);
-
 		memset((void *)ring->ring, 0, ring->ring_size);
-
-		amdgpu_bo_unreserve(ring->ring_obj);
-		if (r) {
-			dev_err(adev->dev, "(%d) ring map failed\n", r);
-			return r;
-		}
 	}
 	ring->ptr_mask = (ring->ring_size / 4) - 1;
 	ring->max_dw = max_dw;
@@ -269,29 +252,20 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
  */
 void amdgpu_ring_fini(struct amdgpu_ring *ring)
 {
-	int r;
-	struct amdgpu_bo *ring_obj;
-
-	ring_obj = ring->ring_obj;
 	ring->ready = false;
-	ring->ring = NULL;
-	ring->ring_obj = NULL;
 
 	amdgpu_wb_free(ring->adev, ring->cond_exe_offs);
 	amdgpu_wb_free(ring->adev, ring->fence_offs);
 	amdgpu_wb_free(ring->adev, ring->rptr_offs);
 	amdgpu_wb_free(ring->adev, ring->wptr_offs);
 
-	if (ring_obj) {
-		r = amdgpu_bo_reserve(ring_obj, false);
-		if (likely(r == 0)) {
-			amdgpu_bo_kunmap(ring_obj);
-			amdgpu_bo_unpin(ring_obj);
-			amdgpu_bo_unreserve(ring_obj);
-		}
-		amdgpu_bo_unref(&ring_obj);
-	}
+	amdgpu_bo_free_kernel(&ring->ring_obj,
+			      &ring->gpu_addr,
+			      (void **)&ring->ring);
+
 	amdgpu_debugfs_ring_fini(ring);
+
+	ring->adev->rings[ring->idx] = NULL;
 }
 
 /*

drivers/gpu/drm/amd/amdgpu/amdgpu_test.c

@@ -111,7 +111,7 @@ static void amdgpu_do_test_moves(struct amdgpu_device *adev)
 		amdgpu_bo_kunmap(gtt_obj[i]);
 
 		r = amdgpu_copy_buffer(ring, gtt_addr, vram_addr,
-				       size, NULL, &fence);
+				       size, NULL, &fence, false);
 
 		if (r) {
 			DRM_ERROR("Failed GTT->VRAM copy %d\n", i);
@@ -156,7 +156,7 @@ static void amdgpu_do_test_moves(struct amdgpu_device *adev)
 		amdgpu_bo_kunmap(vram_obj);
 
 		r = amdgpu_copy_buffer(ring, vram_addr, gtt_addr,
-				       size, NULL, &fence);
+				       size, NULL, &fence, false);
 
 		if (r) {
 			DRM_ERROR("Failed VRAM->GTT copy %d\n", i);

drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h

@@ -247,7 +247,7 @@ DEFINE_EVENT(amdgpu_vm_mapping, amdgpu_vm_bo_mapping,
 	    TP_ARGS(mapping)
 );
 
-TRACE_EVENT(amdgpu_vm_set_page,
+TRACE_EVENT(amdgpu_vm_set_ptes,
 	    TP_PROTO(uint64_t pe, uint64_t addr, unsigned count,
 		     uint32_t incr, uint32_t flags),
 	    TP_ARGS(pe, addr, count, incr, flags),
@@ -271,6 +271,24 @@ TRACE_EVENT(amdgpu_vm_set_page,
 		      __entry->flags, __entry->count)
 );
 
+TRACE_EVENT(amdgpu_vm_copy_ptes,
+	    TP_PROTO(uint64_t pe, uint64_t src, unsigned count),
+	    TP_ARGS(pe, src, count),
+	    TP_STRUCT__entry(
+			     __field(u64, pe)
+			     __field(u64, src)
+			     __field(u32, count)
+			     ),
+
+	    TP_fast_assign(
+			   __entry->pe = pe;
+			   __entry->src = src;
+			   __entry->count = count;
+			   ),
+	    TP_printk("pe=%010Lx, src=%010Lx, count=%u",
+		      __entry->pe, __entry->src, __entry->count)
+);
+
 TRACE_EVENT(amdgpu_vm_flush,
 	    TP_PROTO(uint64_t pd_addr, unsigned ring, unsigned id),
 	    TP_ARGS(pd_addr, ring, id),

drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c

@@ -34,6 +34,7 @@
 #include <ttm/ttm_placement.h>
 #include <ttm/ttm_module.h>
 #include <ttm/ttm_page_alloc.h>
+#include <ttm/ttm_memory.h>
 #include <drm/drmP.h>
 #include <drm/amdgpu_drm.h>
 #include <linux/seq_file.h>
@@ -74,7 +75,7 @@ static void amdgpu_ttm_mem_global_release(struct drm_global_reference *ref)
 	ttm_mem_global_release(ref->object);
 }
 
-static int amdgpu_ttm_global_init(struct amdgpu_device *adev)
+int amdgpu_ttm_global_init(struct amdgpu_device *adev)
 {
 	struct drm_global_reference *global_ref;
 	struct amdgpu_ring *ring;
@@ -88,10 +89,10 @@ static int amdgpu_ttm_global_init(struct amdgpu_device *adev)
 	global_ref->init = &amdgpu_ttm_mem_global_init;
 	global_ref->release = &amdgpu_ttm_mem_global_release;
 	r = drm_global_item_ref(global_ref);
-	if (r != 0) {
+	if (r) {
 		DRM_ERROR("Failed setting up TTM memory accounting "
 			  "subsystem.\n");
-		return r;
+		goto error_mem;
 	}
 
 	adev->mman.bo_global_ref.mem_glob =
@@ -102,26 +103,30 @@ static int amdgpu_ttm_global_init(struct amdgpu_device *adev)
 	global_ref->init = &ttm_bo_global_init;
 	global_ref->release = &ttm_bo_global_release;
 	r = drm_global_item_ref(global_ref);
-	if (r != 0) {
+	if (r) {
 		DRM_ERROR("Failed setting up TTM BO subsystem.\n");
-		drm_global_item_unref(&adev->mman.mem_global_ref);
-		return r;
+		goto error_bo;
 	}
 
 	ring = adev->mman.buffer_funcs_ring;
 	rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_KERNEL];
 	r = amd_sched_entity_init(&ring->sched, &adev->mman.entity,
 				  rq, amdgpu_sched_jobs);
-	if (r != 0) {
+	if (r) {
 		DRM_ERROR("Failed setting up TTM BO move run queue.\n");
-		drm_global_item_unref(&adev->mman.mem_global_ref);
-		drm_global_item_unref(&adev->mman.bo_global_ref.ref);
-		return r;
+		goto error_entity;
 	}
 
 	adev->mman.mem_global_referenced = true;
 
 	return 0;
+
+error_entity:
+	drm_global_item_unref(&adev->mman.bo_global_ref.ref);
+error_bo:
+	drm_global_item_unref(&adev->mman.mem_global_ref);
+error_mem:
+	return r;
 }
 
 static void amdgpu_ttm_global_fini(struct amdgpu_device *adev)
@@ -155,7 +160,7 @@ static int amdgpu_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 		man->default_caching = TTM_PL_FLAG_CACHED;
 		break;
 	case TTM_PL_TT:
-		man->func = &ttm_bo_manager_func;
+		man->func = &amdgpu_gtt_mgr_func;
 		man->gpu_offset = adev->mc.gtt_start;
 		man->available_caching = TTM_PL_MASK_CACHING;
 		man->default_caching = TTM_PL_FLAG_CACHED;
@@ -190,12 +195,13 @@ static int amdgpu_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
 				struct ttm_placement *placement)
 {
-	struct amdgpu_bo *rbo;
+	struct amdgpu_bo *abo;
 	static struct ttm_place placements = {
 		.fpfn = 0,
 		.lpfn = 0,
 		.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_SYSTEM
 	};
+	unsigned i;
 
 	if (!amdgpu_ttm_bo_is_amdgpu_bo(bo)) {
 		placement->placement = &placements;
@@ -204,28 +210,44 @@ static void amdgpu_evict_flags(struct ttm_buffer_object *bo,
 		placement->num_busy_placement = 1;
 		return;
 	}
-	rbo = container_of(bo, struct amdgpu_bo, tbo);
+	abo = container_of(bo, struct amdgpu_bo, tbo);
 	switch (bo->mem.mem_type) {
 	case TTM_PL_VRAM:
-		if (rbo->adev->mman.buffer_funcs_ring->ready == false)
-			amdgpu_ttm_placement_from_domain(rbo, AMDGPU_GEM_DOMAIN_CPU);
-		else
-			amdgpu_ttm_placement_from_domain(rbo, AMDGPU_GEM_DOMAIN_GTT);
+		if (abo->adev->mman.buffer_funcs_ring->ready == false) {
+			amdgpu_ttm_placement_from_domain(abo, AMDGPU_GEM_DOMAIN_CPU);
+		} else {
+			amdgpu_ttm_placement_from_domain(abo, AMDGPU_GEM_DOMAIN_GTT);
+			for (i = 0; i < abo->placement.num_placement; ++i) {
+				if (!(abo->placements[i].flags &
+				      TTM_PL_FLAG_TT))
+					continue;
+
+				if (abo->placements[i].lpfn)
+					continue;
+
+				/* set an upper limit to force directly
+				 * allocating address space for the BO.
+				 */
+				abo->placements[i].lpfn =
+					abo->adev->mc.gtt_size >> PAGE_SHIFT;
+			}
+		}
 		break;
 	case TTM_PL_TT:
 	default:
-		amdgpu_ttm_placement_from_domain(rbo, AMDGPU_GEM_DOMAIN_CPU);
+		amdgpu_ttm_placement_from_domain(abo, AMDGPU_GEM_DOMAIN_CPU);
 	}
-	*placement = rbo->placement;
+	*placement = abo->placement;
 }
 
 static int amdgpu_verify_access(struct ttm_buffer_object *bo, struct file *filp)
 {
-	struct amdgpu_bo *rbo = container_of(bo, struct amdgpu_bo, tbo);
+	struct amdgpu_bo *abo = container_of(bo, struct amdgpu_bo, tbo);
 
 	if (amdgpu_ttm_tt_get_usermm(bo->ttm))
 		return -EPERM;
-	return drm_vma_node_verify_access(&rbo->gem_base.vma_node, filp);
+	return drm_vma_node_verify_access(&abo->gem_base.vma_node,
+					  filp->private_data);
 }
 
 static void amdgpu_move_null(struct ttm_buffer_object *bo,
@@ -251,26 +273,30 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo,
 
 	adev = amdgpu_get_adev(bo->bdev);
 	ring = adev->mman.buffer_funcs_ring;
-	old_start = (u64)old_mem->start << PAGE_SHIFT;
-	new_start = (u64)new_mem->start << PAGE_SHIFT;
 
 	switch (old_mem->mem_type) {
-	case TTM_PL_VRAM:
-		old_start += adev->mc.vram_start;
-		break;
 	case TTM_PL_TT:
-		old_start += adev->mc.gtt_start;
+		r = amdgpu_ttm_bind(bo, old_mem);
+		if (r)
+			return r;
+
+	case TTM_PL_VRAM:
+		old_start = (u64)old_mem->start << PAGE_SHIFT;
+		old_start += bo->bdev->man[old_mem->mem_type].gpu_offset;
 		break;
 	default:
 		DRM_ERROR("Unknown placement %d\n", old_mem->mem_type);
 		return -EINVAL;
 	}
 	switch (new_mem->mem_type) {
-	case TTM_PL_VRAM:
-		new_start += adev->mc.vram_start;
-		break;
 	case TTM_PL_TT:
-		new_start += adev->mc.gtt_start;
+		r = amdgpu_ttm_bind(bo, new_mem);
+		if (r)
+			return r;
+
+	case TTM_PL_VRAM:
+		new_start = (u64)new_mem->start << PAGE_SHIFT;
+		new_start += bo->bdev->man[new_mem->mem_type].gpu_offset;
 		break;
 	default:
 		DRM_ERROR("Unknown placement %d\n", old_mem->mem_type);
@@ -285,7 +311,7 @@ static int amdgpu_move_blit(struct ttm_buffer_object *bo,
 
 	r = amdgpu_copy_buffer(ring, old_start, new_start,
 			       new_mem->num_pages * PAGE_SIZE, /* bytes */
-			       bo->resv, &fence);
+			       bo->resv, &fence, false);
 	if (r)
 		return r;
 
@@ -314,7 +340,7 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo,
 	placement.num_busy_placement = 1;
 	placement.busy_placement = &placements;
 	placements.fpfn = 0;
-	placements.lpfn = 0;
+	placements.lpfn = adev->mc.gtt_size >> PAGE_SHIFT;
 	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
 			     interruptible, no_wait_gpu);
@@ -335,7 +361,7 @@ static int amdgpu_move_vram_ram(struct ttm_buffer_object *bo,
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
-	r = ttm_bo_move_ttm(bo, true, interruptible, no_wait_gpu, new_mem);
+	r = ttm_bo_move_ttm(bo, interruptible, no_wait_gpu, new_mem);
 out_cleanup:
 	ttm_bo_mem_put(bo, &tmp_mem);
 	return r;
@@ -361,14 +387,14 @@ static int amdgpu_move_ram_vram(struct ttm_buffer_object *bo,
 	placement.num_busy_placement = 1;
 	placement.busy_placement = &placements;
 	placements.fpfn = 0;
-	placements.lpfn = 0;
+	placements.lpfn = adev->mc.gtt_size >> PAGE_SHIFT;
 	placements.flags = TTM_PL_MASK_CACHING | TTM_PL_FLAG_TT;
 	r = ttm_bo_mem_space(bo, &placement, &tmp_mem,
 			     interruptible, no_wait_gpu);
 	if (unlikely(r)) {
 		return r;
 	}
-	r = ttm_bo_move_ttm(bo, true, interruptible, no_wait_gpu, &tmp_mem);
+	r = ttm_bo_move_ttm(bo, interruptible, no_wait_gpu, &tmp_mem);
 	if (unlikely(r)) {
 		goto out_cleanup;
 	}
@@ -435,8 +461,7 @@ static int amdgpu_bo_move(struct ttm_buffer_object *bo,
 
 	if (r) {
 memcpy:
-		r = ttm_bo_move_memcpy(bo, evict, interruptible,
-				       no_wait_gpu, new_mem);
+		r = ttm_bo_move_memcpy(bo, interruptible, no_wait_gpu, new_mem);
 		if (r) {
 			return r;
 		}
@@ -524,6 +549,7 @@ struct amdgpu_ttm_tt {
 	spinlock_t              guptasklock;
 	struct list_head        guptasks;
 	atomic_t		mmu_invalidations;
+	struct list_head        list;
 };
 
 int amdgpu_ttm_tt_get_user_pages(struct ttm_tt *ttm, struct page **pages)
@@ -641,7 +667,6 @@ static int amdgpu_ttm_backend_bind(struct ttm_tt *ttm,
 				   struct ttm_mem_reg *bo_mem)
 {
 	struct amdgpu_ttm_tt *gtt = (void*)ttm;
-	uint32_t flags = amdgpu_ttm_tt_pte_flags(gtt->adev, ttm, bo_mem);
 	int r;
 
 	if (gtt->userptr) {
@@ -651,7 +676,6 @@ static int amdgpu_ttm_backend_bind(struct ttm_tt *ttm,
 			return r;
 		}
 	}
-	gtt->offset = (unsigned long)(bo_mem->start << PAGE_SHIFT);
 	if (!ttm->num_pages) {
 		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
 		     ttm->num_pages, bo_mem, ttm);
@@ -662,14 +686,71 @@ static int amdgpu_ttm_backend_bind(struct ttm_tt *ttm,
 	    bo_mem->mem_type == AMDGPU_PL_OA)
 		return -EINVAL;
 
+	return 0;
+}
+
+bool amdgpu_ttm_is_bound(struct ttm_tt *ttm)
+{
+	struct amdgpu_ttm_tt *gtt = (void *)ttm;
+
+	return gtt && !list_empty(&gtt->list);
+}
+
+int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem)
+{
+	struct ttm_tt *ttm = bo->ttm;
+	struct amdgpu_ttm_tt *gtt = (void *)bo->ttm;
+	uint32_t flags;
+	int r;
+
+	if (!ttm || amdgpu_ttm_is_bound(ttm))
+		return 0;
+
+	r = amdgpu_gtt_mgr_alloc(&bo->bdev->man[TTM_PL_TT], bo,
+				 NULL, bo_mem);
+	if (r) {
+		DRM_ERROR("Failed to allocate GTT address space (%d)\n", r);
+		return r;
+	}
+
+	flags = amdgpu_ttm_tt_pte_flags(gtt->adev, ttm, bo_mem);
+	gtt->offset = (u64)bo_mem->start << PAGE_SHIFT;
 	r = amdgpu_gart_bind(gtt->adev, gtt->offset, ttm->num_pages,
 		ttm->pages, gtt->ttm.dma_address, flags);
 
 	if (r) {
-		DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
-			  ttm->num_pages, (unsigned)gtt->offset);
+		DRM_ERROR("failed to bind %lu pages at 0x%08llX\n",
+			  ttm->num_pages, gtt->offset);
 		return r;
 	}
+	spin_lock(&gtt->adev->gtt_list_lock);
+	list_add_tail(&gtt->list, &gtt->adev->gtt_list);
+	spin_unlock(&gtt->adev->gtt_list_lock);
+	return 0;
+}
+
+int amdgpu_ttm_recover_gart(struct amdgpu_device *adev)
+{
+	struct amdgpu_ttm_tt *gtt, *tmp;
+	struct ttm_mem_reg bo_mem;
+	uint32_t flags;
+	int r;
+
+	bo_mem.mem_type = TTM_PL_TT;
+	spin_lock(&adev->gtt_list_lock);
+	list_for_each_entry_safe(gtt, tmp, &adev->gtt_list, list) {
+		flags = amdgpu_ttm_tt_pte_flags(gtt->adev, &gtt->ttm.ttm, &bo_mem);
+		r = amdgpu_gart_bind(adev, gtt->offset, gtt->ttm.ttm.num_pages,
+				     gtt->ttm.ttm.pages, gtt->ttm.dma_address,
+				     flags);
+		if (r) {
+			spin_unlock(&adev->gtt_list_lock);
+			DRM_ERROR("failed to bind %lu pages at 0x%08llX\n",
+				  gtt->ttm.ttm.num_pages, gtt->offset);
+			return r;
+		}
+	}
+	spin_unlock(&adev->gtt_list_lock);
 	return 0;
 }
 
@@ -677,12 +758,19 @@ static int amdgpu_ttm_backend_unbind(struct ttm_tt *ttm)
 {
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 
+	if (gtt->userptr)
+		amdgpu_ttm_tt_unpin_userptr(ttm);
+
+	if (!amdgpu_ttm_is_bound(ttm))
+		return 0;
+
 	/* unbind shouldn't be done for GDS/GWS/OA in ttm_bo_clean_mm */
 	if (gtt->adev->gart.ready)
 		amdgpu_gart_unbind(gtt->adev, gtt->offset, ttm->num_pages);
 
-	if (gtt->userptr)
-		amdgpu_ttm_tt_unpin_userptr(ttm);
+	spin_lock(&gtt->adev->gtt_list_lock);
+	list_del_init(&gtt->list);
+	spin_unlock(&gtt->adev->gtt_list_lock);
 
 	return 0;
 }
@@ -720,6 +808,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_bo_device *bdev,
 		kfree(gtt);
 		return NULL;
 	}
+	INIT_LIST_HEAD(&gtt->list);
 	return &gtt->ttm.ttm;
 }
 
@@ -991,10 +1080,6 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	unsigned i, j;
 	int r;
 
-	r = amdgpu_ttm_global_init(adev);
-	if (r) {
-		return r;
-	}
 	/* No others user of address space so set it to 0 */
 	r = ttm_bo_device_init(&adev->mman.bdev,
 			       adev->mman.bo_global_ref.ref.object,
@@ -1159,7 +1244,7 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring,
 		       uint64_t dst_offset,
 		       uint32_t byte_count,
 		       struct reservation_object *resv,
-		       struct fence **fence)
+		       struct fence **fence, bool direct_submit)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_job *job;
@@ -1201,10 +1286,81 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring,
 		byte_count -= cur_size_in_bytes;
 	}
 
+	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
+	WARN_ON(job->ibs[0].length_dw > num_dw);
+	if (direct_submit) {
+		r = amdgpu_ib_schedule(ring, job->num_ibs, job->ibs,
+				       NULL, NULL, fence);
+		job->fence = fence_get(*fence);
+		if (r)
+			DRM_ERROR("Error scheduling IBs (%d)\n", r);
+		amdgpu_job_free(job);
+	} else {
+		r = amdgpu_job_submit(job, ring, &adev->mman.entity,
+				      AMDGPU_FENCE_OWNER_UNDEFINED, fence);
+		if (r)
+			goto error_free;
+	}
+
+	return r;
+
+error_free:
+	amdgpu_job_free(job);
+	return r;
+}
+
+int amdgpu_fill_buffer(struct amdgpu_bo *bo,
+		uint32_t src_data,
+		struct reservation_object *resv,
+		struct fence **fence)
+{
+	struct amdgpu_device *adev = bo->adev;
+	struct amdgpu_job *job;
+	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
+
+	uint32_t max_bytes, byte_count;
+	uint64_t dst_offset;
+	unsigned int num_loops, num_dw;
+	unsigned int i;
+	int r;
+
+	byte_count = bo->tbo.num_pages << PAGE_SHIFT;
+	max_bytes = adev->mman.buffer_funcs->fill_max_bytes;
+	num_loops = DIV_ROUND_UP(byte_count, max_bytes);
+	num_dw = num_loops * adev->mman.buffer_funcs->fill_num_dw;
+
+	/* for IB padding */
+	while (num_dw & 0x7)
+		num_dw++;
+
+	r = amdgpu_job_alloc_with_ib(adev, num_dw * 4, &job);
+	if (r)
+		return r;
+
+	if (resv) {
+		r = amdgpu_sync_resv(adev, &job->sync, resv,
+				AMDGPU_FENCE_OWNER_UNDEFINED);
+		if (r) {
+			DRM_ERROR("sync failed (%d).\n", r);
+			goto error_free;
+		}
+	}
+
+	dst_offset = bo->tbo.mem.start << PAGE_SHIFT;
+	for (i = 0; i < num_loops; i++) {
+		uint32_t cur_size_in_bytes = min(byte_count, max_bytes);
+
+		amdgpu_emit_fill_buffer(adev, &job->ibs[0], src_data,
+				dst_offset, cur_size_in_bytes);
+
+		dst_offset += cur_size_in_bytes;
+		byte_count -= cur_size_in_bytes;
+	}
+
 	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
 	WARN_ON(job->ibs[0].length_dw > num_dw);
 	r = amdgpu_job_submit(job, ring, &adev->mman.entity,
-			      AMDGPU_FENCE_OWNER_UNDEFINED, fence);
+			AMDGPU_FENCE_OWNER_UNDEFINED, fence);
 	if (r)
 		goto error_free;
 
@@ -1395,3 +1551,8 @@ static void amdgpu_ttm_debugfs_fini(struct amdgpu_device *adev)
 
 #endif
 }
+
+u64 amdgpu_ttm_get_gtt_mem_size(struct amdgpu_device *adev)
+{
+	return ttm_get_kernel_zone_memory_size(adev->mman.mem_global_ref.object);
+}

drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h

@@ -0,0 +1,90 @@
+/*
+ * Copyright 2016 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_TTM_H__
+#define __AMDGPU_TTM_H__
+
+#include "gpu_scheduler.h"
+
+#define AMDGPU_PL_GDS		(TTM_PL_PRIV + 0)
+#define AMDGPU_PL_GWS		(TTM_PL_PRIV + 1)
+#define AMDGPU_PL_OA		(TTM_PL_PRIV + 2)
+
+#define AMDGPU_PL_FLAG_GDS		(TTM_PL_FLAG_PRIV << 0)
+#define AMDGPU_PL_FLAG_GWS		(TTM_PL_FLAG_PRIV << 1)
+#define AMDGPU_PL_FLAG_OA		(TTM_PL_FLAG_PRIV << 2)
+
+#define AMDGPU_TTM_LRU_SIZE	20
+
+struct amdgpu_mman_lru {
+	struct list_head		*lru[TTM_NUM_MEM_TYPES];
+	struct list_head		*swap_lru;
+};
+
+struct amdgpu_mman {
+	struct ttm_bo_global_ref        bo_global_ref;
+	struct drm_global_reference	mem_global_ref;
+	struct ttm_bo_device		bdev;
+	bool				mem_global_referenced;
+	bool				initialized;
+
+#if defined(CONFIG_DEBUG_FS)
+	struct dentry			*vram;
+	struct dentry			*gtt;
+#endif
+
+	/* buffer handling */
+	const struct amdgpu_buffer_funcs	*buffer_funcs;
+	struct amdgpu_ring			*buffer_funcs_ring;
+	/* Scheduler entity for buffer moves */
+	struct amd_sched_entity			entity;
+
+	/* custom LRU management */
+	struct amdgpu_mman_lru			log2_size[AMDGPU_TTM_LRU_SIZE];
+	/* guard for log2_size array, don't add anything in between */
+	struct amdgpu_mman_lru			guard;
+};
+
+extern const struct ttm_mem_type_manager_func amdgpu_gtt_mgr_func;
+
+int amdgpu_gtt_mgr_alloc(struct ttm_mem_type_manager *man,
+			 struct ttm_buffer_object *tbo,
+			 const struct ttm_place *place,
+			 struct ttm_mem_reg *mem);
+
+int amdgpu_copy_buffer(struct amdgpu_ring *ring,
+		       uint64_t src_offset,
+		       uint64_t dst_offset,
+		       uint32_t byte_count,
+		       struct reservation_object *resv,
+		       struct fence **fence, bool direct_submit);
+int amdgpu_fill_buffer(struct amdgpu_bo *bo,
+			uint32_t src_data,
+			struct reservation_object *resv,
+			struct fence **fence);
+
+int amdgpu_mmap(struct file *filp, struct vm_area_struct *vma);
+bool amdgpu_ttm_is_bound(struct ttm_tt *ttm);
+int amdgpu_ttm_bind(struct ttm_buffer_object *bo, struct ttm_mem_reg *bo_mem);
+
+#endif

drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c

@@ -247,40 +247,32 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 	const struct common_firmware_header *header = NULL;
 
 	err = amdgpu_bo_create(adev, adev->firmware.fw_size, PAGE_SIZE, true,
-			AMDGPU_GEM_DOMAIN_GTT, 0, NULL, NULL, bo);
+			       AMDGPU_GEM_DOMAIN_GTT, 0, NULL, NULL, bo);
 	if (err) {
 		dev_err(adev->dev, "(%d) Firmware buffer allocate failed\n", err);
-		err = -ENOMEM;
 		goto failed;
 	}
 
 	err = amdgpu_bo_reserve(*bo, false);
 	if (err) {
-		amdgpu_bo_unref(bo);
 		dev_err(adev->dev, "(%d) Firmware buffer reserve failed\n", err);
-		goto failed;
+		goto failed_reserve;
 	}
 
 	err = amdgpu_bo_pin(*bo, AMDGPU_GEM_DOMAIN_GTT, &fw_mc_addr);
 	if (err) {
-		amdgpu_bo_unreserve(*bo);
-		amdgpu_bo_unref(bo);
 		dev_err(adev->dev, "(%d) Firmware buffer pin failed\n", err);
-		goto failed;
+		goto failed_pin;
 	}
 
 	err = amdgpu_bo_kmap(*bo, &fw_buf_ptr);
 	if (err) {
 		dev_err(adev->dev, "(%d) Firmware buffer kmap failed\n", err);
-		amdgpu_bo_unpin(*bo);
-		amdgpu_bo_unreserve(*bo);
-		amdgpu_bo_unref(bo);
-		goto failed;
+		goto failed_kmap;
 	}
 
 	amdgpu_bo_unreserve(*bo);
 
-	fw_offset = 0;
 	for (i = 0; i < AMDGPU_UCODE_ID_MAXIMUM; i++) {
 		ucode = &adev->firmware.ucode[i];
 		if (ucode->fw) {
@@ -290,10 +282,16 @@ int amdgpu_ucode_init_bo(struct amdgpu_device *adev)
 			fw_offset += ALIGN(le32_to_cpu(header->ucode_size_bytes), PAGE_SIZE);
 		}
 	}
+	return 0;
 
+failed_kmap:
+	amdgpu_bo_unpin(*bo);
+failed_pin:
+	amdgpu_bo_unreserve(*bo);
+failed_reserve:
+	amdgpu_bo_unref(bo);
 failed:
-	if (err)
-		adev->firmware.smu_load = false;
+	adev->firmware.smu_load = false;
 
 	return err;
 }

drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c

@@ -201,39 +201,14 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
 	bo_size = AMDGPU_GPU_PAGE_ALIGN(le32_to_cpu(hdr->ucode_size_bytes) + 8)
 		  +  AMDGPU_UVD_STACK_SIZE + AMDGPU_UVD_HEAP_SIZE
 		  +  AMDGPU_UVD_SESSION_SIZE * adev->uvd.max_handles;
-	r = amdgpu_bo_create(adev, bo_size, PAGE_SIZE, true,
-			     AMDGPU_GEM_DOMAIN_VRAM,
-			     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
-			     NULL, NULL, &adev->uvd.vcpu_bo);
+	r = amdgpu_bo_create_kernel(adev, bo_size, PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_VRAM, &adev->uvd.vcpu_bo,
+				    &adev->uvd.gpu_addr, &adev->uvd.cpu_addr);
 	if (r) {
 		dev_err(adev->dev, "(%d) failed to allocate UVD bo\n", r);
 		return r;
 	}
 
-	r = amdgpu_bo_reserve(adev->uvd.vcpu_bo, false);
-	if (r) {
-		amdgpu_bo_unref(&adev->uvd.vcpu_bo);
-		dev_err(adev->dev, "(%d) failed to reserve UVD bo\n", r);
-		return r;
-	}
-
-	r = amdgpu_bo_pin(adev->uvd.vcpu_bo, AMDGPU_GEM_DOMAIN_VRAM,
-			  &adev->uvd.gpu_addr);
-	if (r) {
-		amdgpu_bo_unreserve(adev->uvd.vcpu_bo);
-		amdgpu_bo_unref(&adev->uvd.vcpu_bo);
-		dev_err(adev->dev, "(%d) UVD bo pin failed\n", r);
-		return r;
-	}
-
-	r = amdgpu_bo_kmap(adev->uvd.vcpu_bo, &adev->uvd.cpu_addr);
-	if (r) {
-		dev_err(adev->dev, "(%d) UVD map failed\n", r);
-		return r;
-	}
-
-	amdgpu_bo_unreserve(adev->uvd.vcpu_bo);
-
 	ring = &adev->uvd.ring;
 	rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL];
 	r = amd_sched_entity_init(&ring->sched, &adev->uvd.entity,
@@ -274,22 +249,13 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev)
 
 int amdgpu_uvd_sw_fini(struct amdgpu_device *adev)
 {
-	int r;
-
 	kfree(adev->uvd.saved_bo);
 
 	amd_sched_entity_fini(&adev->uvd.ring.sched, &adev->uvd.entity);
 
-	if (adev->uvd.vcpu_bo) {
-		r = amdgpu_bo_reserve(adev->uvd.vcpu_bo, false);
-		if (!r) {
-			amdgpu_bo_kunmap(adev->uvd.vcpu_bo);
-			amdgpu_bo_unpin(adev->uvd.vcpu_bo);
-			amdgpu_bo_unreserve(adev->uvd.vcpu_bo);
-		}
-
-		amdgpu_bo_unref(&adev->uvd.vcpu_bo);
-	}
+	amdgpu_bo_free_kernel(&adev->uvd.vcpu_bo,
+			      &adev->uvd.gpu_addr,
+			      (void **)&adev->uvd.cpu_addr);
 
 	amdgpu_ring_fini(&adev->uvd.ring);
 
@@ -323,7 +289,7 @@ int amdgpu_uvd_suspend(struct amdgpu_device *adev)
 	if (!adev->uvd.saved_bo)
 		return -ENOMEM;
 
-	memcpy(adev->uvd.saved_bo, ptr, size);
+	memcpy_fromio(adev->uvd.saved_bo, ptr, size);
 
 	return 0;
 }
@@ -340,7 +306,7 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev)
 	ptr = adev->uvd.cpu_addr;
 
 	if (adev->uvd.saved_bo != NULL) {
-		memcpy(ptr, adev->uvd.saved_bo, size);
+		memcpy_toio(ptr, adev->uvd.saved_bo, size);
 		kfree(adev->uvd.saved_bo);
 		adev->uvd.saved_bo = NULL;
 	} else {
@@ -349,11 +315,11 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev)
 
 		hdr = (const struct common_firmware_header *)adev->uvd.fw->data;
 		offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
-		memcpy(adev->uvd.cpu_addr, (adev->uvd.fw->data) + offset,
-			(adev->uvd.fw->size) - offset);
+		memcpy_toio(adev->uvd.cpu_addr, adev->uvd.fw->data + offset,
+			    le32_to_cpu(hdr->ucode_size_bytes));
 		size -= le32_to_cpu(hdr->ucode_size_bytes);
 		ptr += le32_to_cpu(hdr->ucode_size_bytes);
-		memset(ptr, 0, size);
+		memset_io(ptr, 0, size);
 	}
 
 	return 0;
@@ -385,12 +351,12 @@ void amdgpu_uvd_free_handles(struct amdgpu_device *adev, struct drm_file *filp)
 	}
 }
 
-static void amdgpu_uvd_force_into_uvd_segment(struct amdgpu_bo *rbo)
+static void amdgpu_uvd_force_into_uvd_segment(struct amdgpu_bo *abo)
 {
 	int i;
-	for (i = 0; i < rbo->placement.num_placement; ++i) {
-		rbo->placements[i].fpfn = 0 >> PAGE_SHIFT;
-		rbo->placements[i].lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
+	for (i = 0; i < abo->placement.num_placement; ++i) {
+		abo->placements[i].fpfn = 0 >> PAGE_SHIFT;
+		abo->placements[i].lpfn = (256 * 1024 * 1024) >> PAGE_SHIFT;
 	}
 }
 
@@ -843,6 +809,7 @@ static int amdgpu_uvd_cs_reg(struct amdgpu_uvd_cs_ctx *ctx,
 				return r;
 			break;
 		case mmUVD_ENGINE_CNTL:
+		case mmUVD_NO_OP:
 			break;
 		default:
 			DRM_ERROR("Invalid reg 0x%X!\n", reg);
@@ -915,6 +882,10 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx)
 		return -EINVAL;
 	}
 
+	r = amdgpu_cs_sysvm_access_required(parser);
+	if (r)
+		return r;
+
 	ctx.parser = parser;
 	ctx.buf_sizes = buf_sizes;
 	ctx.ib_idx = ib_idx;
@@ -981,8 +952,10 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo,
 	ib->ptr[3] = addr >> 32;
 	ib->ptr[4] = PACKET0(mmUVD_GPCOM_VCPU_CMD, 0);
 	ib->ptr[5] = 0;
-	for (i = 6; i < 16; ++i)
-		ib->ptr[i] = PACKET2(0);
+	for (i = 6; i < 16; i += 2) {
+		ib->ptr[i] = PACKET0(mmUVD_NO_OP, 0);
+		ib->ptr[i+1] = 0;
+	}
 	ib->length_dw = 16;
 
 	if (direct) {
@@ -1114,15 +1087,9 @@ static void amdgpu_uvd_idle_work_handler(struct work_struct *work)
 {
 	struct amdgpu_device *adev =
 		container_of(work, struct amdgpu_device, uvd.idle_work.work);
-	unsigned i, fences, handles = 0;
-
-	fences = amdgpu_fence_count_emitted(&adev->uvd.ring);
-
-	for (i = 0; i < adev->uvd.max_handles; ++i)
-		if (atomic_read(&adev->uvd.handles[i]))
-			++handles;
+	unsigned fences = amdgpu_fence_count_emitted(&adev->uvd.ring);
 
-	if (fences == 0 && handles == 0) {
+	if (fences == 0) {
 		if (adev->pm.dpm_enabled) {
 			amdgpu_dpm_enable_uvd(adev, false);
 		} else {

drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c

@@ -210,6 +210,8 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size)
  */
 int amdgpu_vce_sw_fini(struct amdgpu_device *adev)
 {
+	unsigned i;
+
 	if (adev->vce.vcpu_bo == NULL)
 		return 0;
 
@@ -217,8 +219,8 @@ int amdgpu_vce_sw_fini(struct amdgpu_device *adev)
 
 	amdgpu_bo_unref(&adev->vce.vcpu_bo);
 
-	amdgpu_ring_fini(&adev->vce.ring[0]);
-	amdgpu_ring_fini(&adev->vce.ring[1]);
+	for (i = 0; i < adev->vce.num_rings; i++)
+		amdgpu_ring_fini(&adev->vce.ring[i]);
 
 	release_firmware(adev->vce.fw);
 	mutex_destroy(&adev->vce.idle_mutex);
@@ -282,8 +284,8 @@ int amdgpu_vce_resume(struct amdgpu_device *adev)
 
 	hdr = (const struct common_firmware_header *)adev->vce.fw->data;
 	offset = le32_to_cpu(hdr->ucode_array_offset_bytes);
-	memcpy(cpu_addr, (adev->vce.fw->data) + offset,
-		(adev->vce.fw->size) - offset);
+	memcpy_toio(cpu_addr, adev->vce.fw->data + offset,
+		    adev->vce.fw->size - offset);
 
 	amdgpu_bo_kunmap(adev->vce.vcpu_bo);
 
@@ -303,9 +305,12 @@ static void amdgpu_vce_idle_work_handler(struct work_struct *work)
 {
 	struct amdgpu_device *adev =
 		container_of(work, struct amdgpu_device, vce.idle_work.work);
+	unsigned i, count = 0;
+
+	for (i = 0; i < adev->vce.num_rings; i++)
+		count += amdgpu_fence_count_emitted(&adev->vce.ring[i]);
 
-	if ((amdgpu_fence_count_emitted(&adev->vce.ring[0]) == 0) &&
-	    (amdgpu_fence_count_emitted(&adev->vce.ring[1]) == 0)) {
+	if (count == 0) {
 		if (adev->pm.dpm_enabled) {
 			amdgpu_dpm_enable_vce(adev, false);
 		} else {
@@ -634,7 +639,11 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 	uint32_t allocated = 0;
 	uint32_t tmp, handle = 0;
 	uint32_t *size = &tmp;
-	int i, r = 0, idx = 0;
+	int i, r, idx = 0;
+
+	r = amdgpu_cs_sysvm_access_required(p);
+	if (r)
+		return r;
 
 	while (idx < ib->length_dw) {
 		uint32_t len = amdgpu_get_ib_value(p, ib_idx, idx);
@@ -687,6 +696,21 @@ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx)
 		case 0x04000008: /* rdo */
 		case 0x04000009: /* vui */
 		case 0x05000002: /* auxiliary buffer */
+		case 0x05000009: /* clock table */
+			break;
+
+		case 0x0500000c: /* hw config */
+			switch (p->adev->asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_CIK
+			case CHIP_KAVERI:
+			case CHIP_MULLINS:
+#endif
+			case CHIP_CARRIZO:
+				break;
+			default:
+				r = -EINVAL;
+				goto out;
+			}
 			break;
 
 		case 0x03000001: /* encode */
@@ -799,6 +823,18 @@ void amdgpu_vce_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq,
 	amdgpu_ring_write(ring, VCE_CMD_END);
 }
 
+unsigned amdgpu_vce_ring_get_emit_ib_size(struct amdgpu_ring *ring)
+{
+	return
+		4; /* amdgpu_vce_ring_emit_ib */
+}
+
+unsigned amdgpu_vce_ring_get_dma_frame_size(struct amdgpu_ring *ring)
+{
+	return
+		6; /* amdgpu_vce_ring_emit_fence  x1 no user fence */
+}
+
 /**
  * amdgpu_vce_ring_test_ring - test if VCE ring is working
  *
@@ -850,8 +886,8 @@ int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring, long timeout)
 	struct fence *fence = NULL;
 	long r;
 
-	/* skip vce ring1 ib test for now, since it's not reliable */
-	if (ring == &ring->adev->vce.ring[1])
+	/* skip vce ring1/2 ib test for now, since it's not reliable */
+	if (ring != &ring->adev->vce.ring[0])
 		return 0;
 
 	r = amdgpu_vce_get_create_msg(ring, 1, NULL);

drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h

@@ -42,5 +42,7 @@ int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring);
 int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring, long timeout);
 void amdgpu_vce_ring_begin_use(struct amdgpu_ring *ring);
 void amdgpu_vce_ring_end_use(struct amdgpu_ring *ring);
+unsigned amdgpu_vce_ring_get_emit_ib_size(struct amdgpu_ring *ring);
+unsigned amdgpu_vce_ring_get_dma_frame_size(struct amdgpu_ring *ring);
 
 #endif

drivers/gpu/drm/amd/powerplay/hwmgr/polaris10_clockpowergating.h → drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h

@@ -19,22 +19,39 @@
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
  * OTHER DEALINGS IN THE SOFTWARE.
  *
+ * Author: Monk.liu@amd.com
  */
+#ifndef AMDGPU_VIRT_H
+#define AMDGPU_VIRT_H
 
-#ifndef _POLARIS10_CLOCK_POWER_GATING_H_
-#define _POLARIS10_CLOCK_POWER_GATING_H_
+#define AMDGPU_SRIOV_CAPS_SRIOV_VBIOS  (1 << 0) /* vBIOS is sr-iov ready */
+#define AMDGPU_SRIOV_CAPS_ENABLE_IOV   (1 << 1) /* sr-iov is enabled on this GPU */
+#define AMDGPU_SRIOV_CAPS_IS_VF        (1 << 2) /* this GPU is a virtual function */
+#define AMDGPU_PASSTHROUGH_MODE        (1 << 3) /* thw whole GPU is pass through for VM */
+/* GPU virtualization */
+struct amdgpu_virtualization {
+	uint32_t virtual_caps;
+};
 
-#include "polaris10_hwmgr.h"
-#include "pp_asicblocks.h"
+#define amdgpu_sriov_enabled(adev) \
+((adev)->virtualization.virtual_caps & AMDGPU_SRIOV_CAPS_ENABLE_IOV)
 
-int polaris10_phm_powergate_vce(struct pp_hwmgr *hwmgr, bool bgate);
-int polaris10_phm_powergate_uvd(struct pp_hwmgr *hwmgr, bool bgate);
-int polaris10_phm_powerdown_uvd(struct pp_hwmgr *hwmgr);
-int polaris10_phm_powergate_samu(struct pp_hwmgr *hwmgr, bool bgate);
-int polaris10_phm_powergate_acp(struct pp_hwmgr *hwmgr, bool bgate);
-int polaris10_phm_disable_clock_power_gating(struct pp_hwmgr *hwmgr);
-int polaris10_phm_update_clock_gatings(struct pp_hwmgr *hwmgr,
-					const uint32_t *msg_id);
-int polaris10_phm_enable_per_cu_power_gating(struct pp_hwmgr *hwmgr, bool enable);
+#define amdgpu_sriov_vf(adev) \
+((adev)->virtualization.virtual_caps & AMDGPU_SRIOV_CAPS_IS_VF)
 
-#endif /* _POLARIS10_CLOCK_POWER_GATING_H_ */
+#define amdgpu_sriov_bios(adev) \
+((adev)->virtualization.virtual_caps & AMDGPU_SRIOV_CAPS_SRIOV_VBIOS)
+
+#define amdgpu_passthrough(adev) \
+((adev)->virtualization.virtual_caps & AMDGPU_PASSTHROUGH_MODE)
+
+static inline bool is_virtual_machine(void)
+{
+#ifdef CONFIG_X86
+	return boot_cpu_has(X86_FEATURE_HYPERVISOR);
+#else
+	return false;
+#endif
+}
+
+#endif

drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c

@@ -51,19 +51,22 @@
  * SI supports 16.
  */
 
-/* Special value that no flush is necessary */
-#define AMDGPU_VM_NO_FLUSH (~0ll)
-
 /* Local structure. Encapsulate some VM table update parameters to reduce
  * the number of function parameters
  */
-struct amdgpu_vm_update_params {
+struct amdgpu_pte_update_params {
+	/* amdgpu device we do this update for */
+	struct amdgpu_device *adev;
 	/* address where to copy page table entries from */
 	uint64_t src;
-	/* DMA addresses to use for mapping */
-	dma_addr_t *pages_addr;
 	/* indirect buffer to fill with commands */
 	struct amdgpu_ib *ib;
+	/* Function which actually does the update */
+	void (*func)(struct amdgpu_pte_update_params *params, uint64_t pe,
+		     uint64_t addr, unsigned count, uint32_t incr,
+		     uint32_t flags);
+	/* indicate update pt or its shadow */
+	bool shadow;
 };
 
 /**
@@ -467,10 +470,9 @@ struct amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm,
 }
 
 /**
- * amdgpu_vm_update_pages - helper to call the right asic function
+ * amdgpu_vm_do_set_ptes - helper to call the right asic function
  *
- * @adev: amdgpu_device pointer
- * @vm_update_params: see amdgpu_vm_update_params definition
+ * @params: see amdgpu_pte_update_params definition
  * @pe: addr of the page entry
  * @addr: dst addr to write into pe
  * @count: number of page entries to update
@@ -480,32 +482,46 @@ struct amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm,
  * Traces the parameters and calls the right asic functions
  * to setup the page table using the DMA.
  */
-static void amdgpu_vm_update_pages(struct amdgpu_device *adev,
-				   struct amdgpu_vm_update_params
-					*vm_update_params,
+static void amdgpu_vm_do_set_ptes(struct amdgpu_pte_update_params *params,
+				  uint64_t pe, uint64_t addr,
+				  unsigned count, uint32_t incr,
+				  uint32_t flags)
+{
+	trace_amdgpu_vm_set_ptes(pe, addr, count, incr, flags);
+
+	if (count < 3) {
+		amdgpu_vm_write_pte(params->adev, params->ib, pe,
+				    addr | flags, count, incr);
+
+	} else {
+		amdgpu_vm_set_pte_pde(params->adev, params->ib, pe, addr,
+				      count, incr, flags);
+	}
+}
+
+/**
+ * amdgpu_vm_do_copy_ptes - copy the PTEs from the GART
+ *
+ * @params: see amdgpu_pte_update_params definition
+ * @pe: addr of the page entry
+ * @addr: dst addr to write into pe
+ * @count: number of page entries to update
+ * @incr: increase next addr by incr bytes
+ * @flags: hw access flags
+ *
+ * Traces the parameters and calls the DMA function to copy the PTEs.
+ */
+static void amdgpu_vm_do_copy_ptes(struct amdgpu_pte_update_params *params,
 				   uint64_t pe, uint64_t addr,
 				   unsigned count, uint32_t incr,
 				   uint32_t flags)
 {
-	trace_amdgpu_vm_set_page(pe, addr, count, incr, flags);
-
-	if (vm_update_params->src) {
-		amdgpu_vm_copy_pte(adev, vm_update_params->ib,
-			pe, (vm_update_params->src + (addr >> 12) * 8), count);
+	uint64_t src = (params->src + (addr >> 12) * 8);
 
-	} else if (vm_update_params->pages_addr) {
-		amdgpu_vm_write_pte(adev, vm_update_params->ib,
-			vm_update_params->pages_addr,
-			pe, addr, count, incr, flags);
 
-	} else if (count < 3) {
-		amdgpu_vm_write_pte(adev, vm_update_params->ib, NULL, pe, addr,
-				    count, incr, flags);
+	trace_amdgpu_vm_copy_ptes(pe, src, count);
 
-	} else {
-		amdgpu_vm_set_pte_pde(adev, vm_update_params->ib, pe, addr,
-				      count, incr, flags);
-	}
+	amdgpu_vm_copy_pte(params->adev, params->ib, pe, src, count);
 }
 
 /**
@@ -523,12 +539,11 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 	struct amdgpu_ring *ring;
 	struct fence *fence = NULL;
 	struct amdgpu_job *job;
-	struct amdgpu_vm_update_params vm_update_params;
+	struct amdgpu_pte_update_params params;
 	unsigned entries;
 	uint64_t addr;
 	int r;
 
-	memset(&vm_update_params, 0, sizeof(vm_update_params));
 	ring = container_of(vm->entity.sched, struct amdgpu_ring, sched);
 
 	r = reservation_object_reserve_shared(bo->tbo.resv);
@@ -539,6 +554,10 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 	if (r)
 		goto error;
 
+	r = amdgpu_ttm_bind(&bo->tbo, &bo->tbo.mem);
+	if (r)
+		goto error;
+
 	addr = amdgpu_bo_gpu_offset(bo);
 	entries = amdgpu_bo_size(bo) / 8;
 
@@ -546,9 +565,10 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev,
 	if (r)
 		goto error;
 
-	vm_update_params.ib = &job->ibs[0];
-	amdgpu_vm_update_pages(adev, &vm_update_params, addr, 0, entries,
-			       0, 0);
+	memset(&params, 0, sizeof(params));
+	params.adev = adev;
+	params.ib = &job->ibs[0];
+	amdgpu_vm_do_set_ptes(&params, addr, 0, entries, 0, 0);
 	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
 
 	WARN_ON(job->ibs[0].length_dw > 64);
@@ -577,55 +597,46 @@ error:
  * Look up the physical address of the page that the pte resolves
  * to and return the pointer for the page table entry.
  */
-uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr)
+static uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr)
 {
 	uint64_t result;
 
-	if (pages_addr) {
-		/* page table offset */
-		result = pages_addr[addr >> PAGE_SHIFT];
-
-		/* in case cpu page size != gpu page size*/
-		result |= addr & (~PAGE_MASK);
+	/* page table offset */
+	result = pages_addr[addr >> PAGE_SHIFT];
 
-	} else {
-		/* No mapping required */
-		result = addr;
-	}
+	/* in case cpu page size != gpu page size*/
+	result |= addr & (~PAGE_MASK);
 
 	result &= 0xFFFFFFFFFFFFF000ULL;
 
 	return result;
 }
 
-/**
- * amdgpu_vm_update_pdes - make sure that page directory is valid
- *
- * @adev: amdgpu_device pointer
- * @vm: requested vm
- * @start: start of GPU address range
- * @end: end of GPU address range
- *
- * Allocates new page tables if necessary
- * and updates the page directory.
- * Returns 0 for success, error for failure.
- */
-int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
-				    struct amdgpu_vm *vm)
+static int amdgpu_vm_update_pd_or_shadow(struct amdgpu_device *adev,
+					 struct amdgpu_vm *vm,
+					 bool shadow)
 {
 	struct amdgpu_ring *ring;
-	struct amdgpu_bo *pd = vm->page_directory;
-	uint64_t pd_addr = amdgpu_bo_gpu_offset(pd);
+	struct amdgpu_bo *pd = shadow ? vm->page_directory->shadow :
+		vm->page_directory;
+	uint64_t pd_addr;
 	uint32_t incr = AMDGPU_VM_PTE_COUNT * 8;
 	uint64_t last_pde = ~0, last_pt = ~0;
 	unsigned count = 0, pt_idx, ndw;
 	struct amdgpu_job *job;
-	struct amdgpu_vm_update_params vm_update_params;
+	struct amdgpu_pte_update_params params;
 	struct fence *fence = NULL;
 
 	int r;
 
-	memset(&vm_update_params, 0, sizeof(vm_update_params));
+	if (!pd)
+		return 0;
+
+	r = amdgpu_ttm_bind(&pd->tbo, &pd->tbo.mem);
+	if (r)
+		return r;
+
+	pd_addr = amdgpu_bo_gpu_offset(pd);
 	ring = container_of(vm->entity.sched, struct amdgpu_ring, sched);
 
 	/* padding, etc. */
@@ -638,7 +649,9 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
 	if (r)
 		return r;
 
-	vm_update_params.ib = &job->ibs[0];
+	memset(&params, 0, sizeof(params));
+	params.adev = adev;
+	params.ib = &job->ibs[0];
 
 	/* walk over the address space and update the page directory */
 	for (pt_idx = 0; pt_idx <= vm->max_pde_used; ++pt_idx) {
@@ -648,20 +661,34 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
 		if (bo == NULL)
 			continue;
 
+		if (bo->shadow) {
+			struct amdgpu_bo *shadow = bo->shadow;
+
+			r = amdgpu_ttm_bind(&shadow->tbo, &shadow->tbo.mem);
+			if (r)
+				return r;
+		}
+
 		pt = amdgpu_bo_gpu_offset(bo);
-		if (vm->page_tables[pt_idx].addr == pt)
-			continue;
-		vm->page_tables[pt_idx].addr = pt;
+		if (!shadow) {
+			if (vm->page_tables[pt_idx].addr == pt)
+				continue;
+			vm->page_tables[pt_idx].addr = pt;
+		} else {
+			if (vm->page_tables[pt_idx].shadow_addr == pt)
+				continue;
+			vm->page_tables[pt_idx].shadow_addr = pt;
+		}
 
 		pde = pd_addr + pt_idx * 8;
 		if (((last_pde + 8 * count) != pde) ||
-		    ((last_pt + incr * count) != pt)) {
+		    ((last_pt + incr * count) != pt) ||
+		    (count == AMDGPU_VM_MAX_UPDATE_SIZE)) {
 
 			if (count) {
-				amdgpu_vm_update_pages(adev, &vm_update_params,
-						       last_pde, last_pt,
-						       count, incr,
-						       AMDGPU_PTE_VALID);
+				amdgpu_vm_do_set_ptes(&params, last_pde,
+						      last_pt, count, incr,
+						      AMDGPU_PTE_VALID);
 			}
 
 			count = 1;
@@ -673,15 +700,14 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
 	}
 
 	if (count)
-		amdgpu_vm_update_pages(adev, &vm_update_params,
-					last_pde, last_pt,
-					count, incr, AMDGPU_PTE_VALID);
+		amdgpu_vm_do_set_ptes(&params, last_pde, last_pt,
+				      count, incr, AMDGPU_PTE_VALID);
 
-	if (vm_update_params.ib->length_dw != 0) {
-		amdgpu_ring_pad_ib(ring, vm_update_params.ib);
+	if (params.ib->length_dw != 0) {
+		amdgpu_ring_pad_ib(ring, params.ib);
 		amdgpu_sync_resv(adev, &job->sync, pd->tbo.resv,
 				 AMDGPU_FENCE_OWNER_VM);
-		WARN_ON(vm_update_params.ib->length_dw > ndw);
+		WARN_ON(params.ib->length_dw > ndw);
 		r = amdgpu_job_submit(job, ring, &vm->entity,
 				      AMDGPU_FENCE_OWNER_VM, &fence);
 		if (r)
@@ -703,92 +729,33 @@ error_free:
 	return r;
 }
 
-/**
- * amdgpu_vm_frag_ptes - add fragment information to PTEs
+/*
+ * amdgpu_vm_update_pdes - make sure that page directory is valid
  *
  * @adev: amdgpu_device pointer
- * @vm_update_params: see amdgpu_vm_update_params definition
- * @pe_start: first PTE to handle
- * @pe_end: last PTE to handle
- * @addr: addr those PTEs should point to
- * @flags: hw mapping flags
+ * @vm: requested vm
+ * @start: start of GPU address range
+ * @end: end of GPU address range
+ *
+ * Allocates new page tables if necessary
+ * and updates the page directory.
+ * Returns 0 for success, error for failure.
  */
-static void amdgpu_vm_frag_ptes(struct amdgpu_device *adev,
-				struct amdgpu_vm_update_params
-					*vm_update_params,
-				uint64_t pe_start, uint64_t pe_end,
-				uint64_t addr, uint32_t flags)
+int amdgpu_vm_update_page_directory(struct amdgpu_device *adev,
+                                   struct amdgpu_vm *vm)
 {
-	/**
-	 * The MC L1 TLB supports variable sized pages, based on a fragment
-	 * field in the PTE. When this field is set to a non-zero value, page
-	 * granularity is increased from 4KB to (1 << (12 + frag)). The PTE
-	 * flags are considered valid for all PTEs within the fragment range
-	 * and corresponding mappings are assumed to be physically contiguous.
-	 *
-	 * The L1 TLB can store a single PTE for the whole fragment,
-	 * significantly increasing the space available for translation
-	 * caching. This leads to large improvements in throughput when the
-	 * TLB is under pressure.
-	 *
-	 * The L2 TLB distributes small and large fragments into two
-	 * asymmetric partitions. The large fragment cache is significantly
-	 * larger. Thus, we try to use large fragments wherever possible.
-	 * Userspace can support this by aligning virtual base address and
-	 * allocation size to the fragment size.
-	 */
-
-	/* SI and newer are optimized for 64KB */
-	uint64_t frag_flags = AMDGPU_PTE_FRAG_64KB;
-	uint64_t frag_align = 0x80;
-
-	uint64_t frag_start = ALIGN(pe_start, frag_align);
-	uint64_t frag_end = pe_end & ~(frag_align - 1);
-
-	unsigned count;
-
-	/* Abort early if there isn't anything to do */
-	if (pe_start == pe_end)
-		return;
-
-	/* system pages are non continuously */
-	if (vm_update_params->src || vm_update_params->pages_addr ||
-		!(flags & AMDGPU_PTE_VALID) || (frag_start >= frag_end)) {
-
-		count = (pe_end - pe_start) / 8;
-		amdgpu_vm_update_pages(adev, vm_update_params, pe_start,
-				       addr, count, AMDGPU_GPU_PAGE_SIZE,
-				       flags);
-		return;
-	}
-
-	/* handle the 4K area at the beginning */
-	if (pe_start != frag_start) {
-		count = (frag_start - pe_start) / 8;
-		amdgpu_vm_update_pages(adev, vm_update_params, pe_start, addr,
-				       count, AMDGPU_GPU_PAGE_SIZE, flags);
-		addr += AMDGPU_GPU_PAGE_SIZE * count;
-	}
-
-	/* handle the area in the middle */
-	count = (frag_end - frag_start) / 8;
-	amdgpu_vm_update_pages(adev, vm_update_params, frag_start, addr, count,
-			       AMDGPU_GPU_PAGE_SIZE, flags | frag_flags);
+	int r;
 
-	/* handle the 4K area at the end */
-	if (frag_end != pe_end) {
-		addr += AMDGPU_GPU_PAGE_SIZE * count;
-		count = (pe_end - frag_end) / 8;
-		amdgpu_vm_update_pages(adev, vm_update_params, frag_end, addr,
-				       count, AMDGPU_GPU_PAGE_SIZE, flags);
-	}
+	r = amdgpu_vm_update_pd_or_shadow(adev, vm, true);
+	if (r)
+		return r;
+	return amdgpu_vm_update_pd_or_shadow(adev, vm, false);
 }
 
 /**
  * amdgpu_vm_update_ptes - make sure that page tables are valid
  *
- * @adev: amdgpu_device pointer
- * @vm_update_params: see amdgpu_vm_update_params definition
+ * @params: see amdgpu_pte_update_params definition
  * @vm: requested vm
  * @start: start of GPU address range
  * @end: end of GPU address range
@@ -797,16 +764,14 @@ static void amdgpu_vm_frag_ptes(struct amdgpu_device *adev,
  *
  * Update the page tables in the range @start - @end.
  */
-static void amdgpu_vm_update_ptes(struct amdgpu_device *adev,
-				  struct amdgpu_vm_update_params
-					*vm_update_params,
+static void amdgpu_vm_update_ptes(struct amdgpu_pte_update_params *params,
 				  struct amdgpu_vm *vm,
 				  uint64_t start, uint64_t end,
 				  uint64_t dst, uint32_t flags)
 {
 	const uint64_t mask = AMDGPU_VM_PTE_COUNT - 1;
 
-	uint64_t cur_pe_start, cur_pe_end, cur_dst;
+	uint64_t cur_pe_start, cur_nptes, cur_dst;
 	uint64_t addr; /* next GPU address to be updated */
 	uint64_t pt_idx;
 	struct amdgpu_bo *pt;
@@ -817,7 +782,11 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev,
 	addr = start;
 	pt_idx = addr >> amdgpu_vm_block_size;
 	pt = vm->page_tables[pt_idx].entry.robj;
-
+	if (params->shadow) {
+		if (!pt->shadow)
+			return;
+		pt = vm->page_tables[pt_idx].entry.robj->shadow;
+	}
 	if ((addr & ~mask) == (end & ~mask))
 		nptes = end - addr;
 	else
@@ -825,7 +794,7 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev,
 
 	cur_pe_start = amdgpu_bo_gpu_offset(pt);
 	cur_pe_start += (addr & mask) * 8;
-	cur_pe_end = cur_pe_start + 8 * nptes;
+	cur_nptes = nptes;
 	cur_dst = dst;
 
 	/* for next ptb*/
@@ -836,6 +805,11 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev,
 	while (addr < end) {
 		pt_idx = addr >> amdgpu_vm_block_size;
 		pt = vm->page_tables[pt_idx].entry.robj;
+		if (params->shadow) {
+			if (!pt->shadow)
+				return;
+			pt = vm->page_tables[pt_idx].entry.robj->shadow;
+		}
 
 		if ((addr & ~mask) == (end & ~mask))
 			nptes = end - addr;
@@ -845,19 +819,19 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev,
 		next_pe_start = amdgpu_bo_gpu_offset(pt);
 		next_pe_start += (addr & mask) * 8;
 
-		if (cur_pe_end == next_pe_start) {
+		if ((cur_pe_start + 8 * cur_nptes) == next_pe_start &&
+		    ((cur_nptes + nptes) <= AMDGPU_VM_MAX_UPDATE_SIZE)) {
 			/* The next ptb is consecutive to current ptb.
-			 * Don't call amdgpu_vm_frag_ptes now.
+			 * Don't call the update function now.
 			 * Will update two ptbs together in future.
 			*/
-			cur_pe_end += 8 * nptes;
+			cur_nptes += nptes;
 		} else {
-			amdgpu_vm_frag_ptes(adev, vm_update_params,
-					    cur_pe_start, cur_pe_end,
-					    cur_dst, flags);
+			params->func(params, cur_pe_start, cur_dst, cur_nptes,
+				     AMDGPU_GPU_PAGE_SIZE, flags);
 
 			cur_pe_start = next_pe_start;
-			cur_pe_end = next_pe_start + 8 * nptes;
+			cur_nptes = nptes;
 			cur_dst = dst;
 		}
 
@@ -866,8 +840,75 @@ static void amdgpu_vm_update_ptes(struct amdgpu_device *adev,
 		dst += nptes * AMDGPU_GPU_PAGE_SIZE;
 	}
 
-	amdgpu_vm_frag_ptes(adev, vm_update_params, cur_pe_start,
-			    cur_pe_end, cur_dst, flags);
+	params->func(params, cur_pe_start, cur_dst, cur_nptes,
+		     AMDGPU_GPU_PAGE_SIZE, flags);
+}
+
+/*
+ * amdgpu_vm_frag_ptes - add fragment information to PTEs
+ *
+ * @params: see amdgpu_pte_update_params definition
+ * @vm: requested vm
+ * @start: first PTE to handle
+ * @end: last PTE to handle
+ * @dst: addr those PTEs should point to
+ * @flags: hw mapping flags
+ */
+static void amdgpu_vm_frag_ptes(struct amdgpu_pte_update_params	*params,
+				struct amdgpu_vm *vm,
+				uint64_t start, uint64_t end,
+				uint64_t dst, uint32_t flags)
+{
+	/**
+	 * The MC L1 TLB supports variable sized pages, based on a fragment
+	 * field in the PTE. When this field is set to a non-zero value, page
+	 * granularity is increased from 4KB to (1 << (12 + frag)). The PTE
+	 * flags are considered valid for all PTEs within the fragment range
+	 * and corresponding mappings are assumed to be physically contiguous.
+	 *
+	 * The L1 TLB can store a single PTE for the whole fragment,
+	 * significantly increasing the space available for translation
+	 * caching. This leads to large improvements in throughput when the
+	 * TLB is under pressure.
+	 *
+	 * The L2 TLB distributes small and large fragments into two
+	 * asymmetric partitions. The large fragment cache is significantly
+	 * larger. Thus, we try to use large fragments wherever possible.
+	 * Userspace can support this by aligning virtual base address and
+	 * allocation size to the fragment size.
+	 */
+
+	/* SI and newer are optimized for 64KB */
+	uint64_t frag_flags = AMDGPU_PTE_FRAG(AMDGPU_LOG2_PAGES_PER_FRAG);
+	uint64_t frag_align = 1 << AMDGPU_LOG2_PAGES_PER_FRAG;
+
+	uint64_t frag_start = ALIGN(start, frag_align);
+	uint64_t frag_end = end & ~(frag_align - 1);
+
+	/* system pages are non continuously */
+	if (params->src || !(flags & AMDGPU_PTE_VALID) ||
+	    (frag_start >= frag_end)) {
+
+		amdgpu_vm_update_ptes(params, vm, start, end, dst, flags);
+		return;
+	}
+
+	/* handle the 4K area at the beginning */
+	if (start != frag_start) {
+		amdgpu_vm_update_ptes(params, vm, start, frag_start,
+				      dst, flags);
+		dst += (frag_start - start) * AMDGPU_GPU_PAGE_SIZE;
+	}
+
+	/* handle the area in the middle */
+	amdgpu_vm_update_ptes(params, vm, frag_start, frag_end, dst,
+			      flags | frag_flags);
+
+	/* handle the 4K area at the end */
+	if (frag_end != end) {
+		dst += (frag_end - frag_start) * AMDGPU_GPU_PAGE_SIZE;
+		amdgpu_vm_update_ptes(params, vm, frag_end, end, dst, flags);
+	}
 }
 
 /**
@@ -900,14 +941,19 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 	void *owner = AMDGPU_FENCE_OWNER_VM;
 	unsigned nptes, ncmds, ndw;
 	struct amdgpu_job *job;
-	struct amdgpu_vm_update_params vm_update_params;
+	struct amdgpu_pte_update_params params;
 	struct fence *f = NULL;
 	int r;
 
+	memset(&params, 0, sizeof(params));
+	params.adev = adev;
+	params.src = src;
+
 	ring = container_of(vm->entity.sched, struct amdgpu_ring, sched);
-	memset(&vm_update_params, 0, sizeof(vm_update_params));
-	vm_update_params.src = src;
-	vm_update_params.pages_addr = pages_addr;
+
+	memset(&params, 0, sizeof(params));
+	params.adev = adev;
+	params.src = src;
 
 	/* sync to everything on unmapping */
 	if (!(flags & AMDGPU_PTE_VALID))
@@ -924,30 +970,53 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 	/* padding, etc. */
 	ndw = 64;
 
-	if (vm_update_params.src) {
+	if (src) {
 		/* only copy commands needed */
 		ndw += ncmds * 7;
 
-	} else if (vm_update_params.pages_addr) {
-		/* header for write data commands */
-		ndw += ncmds * 4;
+		params.func = amdgpu_vm_do_copy_ptes;
+
+	} else if (pages_addr) {
+		/* copy commands needed */
+		ndw += ncmds * 7;
 
-		/* body of write data command */
+		/* and also PTEs */
 		ndw += nptes * 2;
 
+		params.func = amdgpu_vm_do_copy_ptes;
+
 	} else {
 		/* set page commands needed */
 		ndw += ncmds * 10;
 
 		/* two extra commands for begin/end of fragment */
 		ndw += 2 * 10;
+
+		params.func = amdgpu_vm_do_set_ptes;
 	}
 
 	r = amdgpu_job_alloc_with_ib(adev, ndw * 4, &job);
 	if (r)
 		return r;
 
-	vm_update_params.ib = &job->ibs[0];
+	params.ib = &job->ibs[0];
+
+	if (!src && pages_addr) {
+		uint64_t *pte;
+		unsigned i;
+
+		/* Put the PTEs at the end of the IB. */
+		i = ndw - nptes * 2;
+		pte= (uint64_t *)&(job->ibs->ptr[i]);
+		params.src = job->ibs->gpu_addr + i * 4;
+
+		for (i = 0; i < nptes; ++i) {
+			pte[i] = amdgpu_vm_map_gart(pages_addr, addr + i *
+						    AMDGPU_GPU_PAGE_SIZE);
+			pte[i] |= flags;
+		}
+		addr = 0;
+	}
 
 	r = amdgpu_sync_fence(adev, &job->sync, exclusive);
 	if (r)
@@ -962,11 +1031,13 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev,
 	if (r)
 		goto error_free;
 
-	amdgpu_vm_update_ptes(adev, &vm_update_params, vm, start,
-			      last + 1, addr, flags);
+	params.shadow = true;
+	amdgpu_vm_frag_ptes(&params, vm, start, last + 1, addr, flags);
+	params.shadow = false;
+	amdgpu_vm_frag_ptes(&params, vm, start, last + 1, addr, flags);
 
-	amdgpu_ring_pad_ib(ring, vm_update_params.ib);
-	WARN_ON(vm_update_params.ib->length_dw > ndw);
+	amdgpu_ring_pad_ib(ring, params.ib);
+	WARN_ON(params.ib->length_dw > ndw);
 	r = amdgpu_job_submit(job, ring, &vm->entity,
 			      AMDGPU_FENCE_OWNER_VM, &f);
 	if (r)
@@ -1062,28 +1133,32 @@ static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev,
  *
  * @adev: amdgpu_device pointer
  * @bo_va: requested BO and VM object
- * @mem: ttm mem
+ * @clear: if true clear the entries
  *
  * Fill in the page table entries for @bo_va.
  * Returns 0 for success, -EINVAL for failure.
- *
- * Object have to be reserved and mutex must be locked!
  */
 int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 			struct amdgpu_bo_va *bo_va,
-			struct ttm_mem_reg *mem)
+			bool clear)
 {
 	struct amdgpu_vm *vm = bo_va->vm;
 	struct amdgpu_bo_va_mapping *mapping;
 	dma_addr_t *pages_addr = NULL;
 	uint32_t gtt_flags, flags;
+	struct ttm_mem_reg *mem;
 	struct fence *exclusive;
 	uint64_t addr;
 	int r;
 
-	if (mem) {
+	if (clear) {
+		mem = NULL;
+		addr = 0;
+		exclusive = NULL;
+	} else {
 		struct ttm_dma_tt *ttm;
 
+		mem = &bo_va->bo->tbo.mem;
 		addr = (u64)mem->start << PAGE_SHIFT;
 		switch (mem->mem_type) {
 		case TTM_PL_TT:
@@ -1101,13 +1176,11 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 		}
 
 		exclusive = reservation_object_get_excl(bo_va->bo->tbo.resv);
-	} else {
-		addr = 0;
-		exclusive = NULL;
 	}
 
 	flags = amdgpu_ttm_tt_pte_flags(adev, bo_va->bo->tbo.ttm, mem);
-	gtt_flags = (adev == bo_va->bo->adev) ? flags : 0;
+	gtt_flags = (amdgpu_ttm_is_bound(bo_va->bo->tbo.ttm) &&
+		adev == bo_va->bo->adev) ? flags : 0;
 
 	spin_lock(&vm->status_lock);
 	if (!list_empty(&bo_va->vm_status))
@@ -1134,7 +1207,7 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev,
 	spin_lock(&vm->status_lock);
 	list_splice_init(&bo_va->invalids, &bo_va->valids);
 	list_del_init(&bo_va->vm_status);
-	if (!mem)
+	if (clear)
 		list_add(&bo_va->vm_status, &vm->cleared);
 	spin_unlock(&vm->status_lock);
 
@@ -1197,7 +1270,7 @@ int amdgpu_vm_clear_invalids(struct amdgpu_device *adev,
 			struct amdgpu_bo_va, vm_status);
 		spin_unlock(&vm->status_lock);
 
-		r = amdgpu_vm_bo_update(adev, bo_va, NULL);
+		r = amdgpu_vm_bo_update(adev, bo_va, true);
 		if (r)
 			return r;
 
@@ -1342,7 +1415,8 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev,
 		r = amdgpu_bo_create(adev, AMDGPU_VM_PTE_COUNT * 8,
 				     AMDGPU_GPU_PAGE_SIZE, true,
 				     AMDGPU_GEM_DOMAIN_VRAM,
-				     AMDGPU_GEM_CREATE_NO_CPU_ACCESS,
+				     AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
+				     AMDGPU_GEM_CREATE_SHADOW,
 				     NULL, resv, &pt);
 		if (r)
 			goto error_free;
@@ -1354,10 +1428,20 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev,
 
 		r = amdgpu_vm_clear_bo(adev, vm, pt);
 		if (r) {
+			amdgpu_bo_unref(&pt->shadow);
 			amdgpu_bo_unref(&pt);
 			goto error_free;
 		}
 
+		if (pt->shadow) {
+			r = amdgpu_vm_clear_bo(adev, vm, pt->shadow);
+			if (r) {
+				amdgpu_bo_unref(&pt->shadow);
+				amdgpu_bo_unref(&pt);
+				goto error_free;
+			}
+		}
+
 		entry->robj = pt;
 		entry->priority = 0;
 		entry->tv.bo = &entry->robj->tbo;
@@ -1541,7 +1625,8 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 
 	r = amdgpu_bo_create(adev, pd_size, align, true,
 			     AMDGPU_GEM_DOMAIN_VRAM,
-			     AMDGPU_GEM_CREATE_NO_CPU_ACCESS,
+			     AMDGPU_GEM_CREATE_NO_CPU_ACCESS |
+			     AMDGPU_GEM_CREATE_SHADOW,
 			     NULL, NULL, &vm->page_directory);
 	if (r)
 		goto error_free_sched_entity;
@@ -1551,14 +1636,25 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 		goto error_free_page_directory;
 
 	r = amdgpu_vm_clear_bo(adev, vm, vm->page_directory);
-	amdgpu_bo_unreserve(vm->page_directory);
 	if (r)
-		goto error_free_page_directory;
+		goto error_unreserve;
+
+	if (vm->page_directory->shadow) {
+		r = amdgpu_vm_clear_bo(adev, vm, vm->page_directory->shadow);
+		if (r)
+			goto error_unreserve;
+	}
+
 	vm->last_eviction_counter = atomic64_read(&adev->num_evictions);
+	amdgpu_bo_unreserve(vm->page_directory);
 
 	return 0;
 
+error_unreserve:
+	amdgpu_bo_unreserve(vm->page_directory);
+
 error_free_page_directory:
+	amdgpu_bo_unref(&vm->page_directory->shadow);
 	amdgpu_bo_unref(&vm->page_directory);
 	vm->page_directory = NULL;
 
@@ -1600,10 +1696,18 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 		kfree(mapping);
 	}
 
-	for (i = 0; i < amdgpu_vm_num_pdes(adev); i++)
-		amdgpu_bo_unref(&vm->page_tables[i].entry.robj);
+	for (i = 0; i < amdgpu_vm_num_pdes(adev); i++) {
+		struct amdgpu_bo *pt = vm->page_tables[i].entry.robj;
+
+		if (!pt)
+			continue;
+
+		amdgpu_bo_unref(&pt->shadow);
+		amdgpu_bo_unref(&pt);
+	}
 	drm_free_large(vm->page_tables);
 
+	amdgpu_bo_unref(&vm->page_directory->shadow);
 	amdgpu_bo_unref(&vm->page_directory);
 	fence_put(vm->page_directory_fence);
 }

drivers/gpu/drm/amd/amdgpu/atombios_crtc.c

@@ -497,7 +497,13 @@ void amdgpu_atombios_crtc_set_disp_eng_pll(struct amdgpu_device *adev,
 			 * SetPixelClock provides the dividers
 			 */
 			args.v6.ulDispEngClkFreq = cpu_to_le32(dispclk);
-			args.v6.ucPpll = ATOM_EXT_PLL1;
+			if (adev->asic_type == CHIP_TAHITI ||
+			    adev->asic_type == CHIP_PITCAIRN ||
+			    adev->asic_type == CHIP_VERDE ||
+			    adev->asic_type == CHIP_OLAND)
+				args.v6.ucPpll = ATOM_PPLL0;
+			else
+				args.v6.ucPpll = ATOM_EXT_PLL1;
 			break;
 		default:
 			DRM_ERROR("Unknown table version %d %d\n", frev, crev);

drivers/gpu/drm/amd/amdgpu/atombios_dp.c

@@ -88,7 +88,6 @@ static int amdgpu_atombios_dp_process_aux_ch(struct amdgpu_i2c_chan *chan,
 
 	/* timeout */
 	if (args.v2.ucReplyStatus == 1) {
-		DRM_DEBUG_KMS("dp_aux_ch timeout\n");
 		r = -ETIMEDOUT;
 		goto done;
 	}
@@ -339,22 +338,21 @@ int amdgpu_atombios_dp_get_dpcd(struct amdgpu_connector *amdgpu_connector)
 {
 	struct amdgpu_connector_atom_dig *dig_connector = amdgpu_connector->con_priv;
 	u8 msg[DP_DPCD_SIZE];
-	int ret, i;
+	int ret;
 
-	for (i = 0; i < 7; i++) {
-		ret = drm_dp_dpcd_read(&amdgpu_connector->ddc_bus->aux, DP_DPCD_REV, msg,
-				       DP_DPCD_SIZE);
-		if (ret == DP_DPCD_SIZE) {
-			memcpy(dig_connector->dpcd, msg, DP_DPCD_SIZE);
+	ret = drm_dp_dpcd_read(&amdgpu_connector->ddc_bus->aux, DP_DPCD_REV,
+			       msg, DP_DPCD_SIZE);
+	if (ret == DP_DPCD_SIZE) {
+		memcpy(dig_connector->dpcd, msg, DP_DPCD_SIZE);
 
-			DRM_DEBUG_KMS("DPCD: %*ph\n", (int)sizeof(dig_connector->dpcd),
-				      dig_connector->dpcd);
+		DRM_DEBUG_KMS("DPCD: %*ph\n", (int)sizeof(dig_connector->dpcd),
+			      dig_connector->dpcd);
 
-			amdgpu_atombios_dp_probe_oui(amdgpu_connector);
+		amdgpu_atombios_dp_probe_oui(amdgpu_connector);
 
-			return 0;
-		}
+		return 0;
 	}
+
 	dig_connector->dpcd[0] = 0;
 	return -EINVAL;
 }

drivers/gpu/drm/amd/amdgpu/atombios_i2c.c

@@ -27,6 +27,7 @@
 #include "amdgpu.h"
 #include "atom.h"
 #include "amdgpu_atombios.h"
+#include "atombios_i2c.h"
 
 #define TARGET_HW_I2C_CLOCK 50
 

drivers/gpu/drm/amd/amdgpu/ci_dpm.c

@@ -5396,7 +5396,7 @@ static void ci_dpm_disable(struct amdgpu_device *adev)
 	amdgpu_irq_put(adev, &adev->pm.dpm.thermal.irq,
 		       AMDGPU_THERMAL_IRQ_HIGH_TO_LOW);
 
-	ci_dpm_powergate_uvd(adev, false);
+	ci_dpm_powergate_uvd(adev, true);
 
 	if (!amdgpu_ci_is_smc_running(adev))
 		return;
@@ -5874,7 +5874,10 @@ static int ci_dpm_init(struct amdgpu_device *adev)
 	pi->pcie_dpm_key_disabled = 0;
 	pi->thermal_sclk_dpm_enabled = 0;
 
-	pi->caps_sclk_ds = true;
+	if (amdgpu_sclk_deep_sleep_en)
+		pi->caps_sclk_ds = true;
+	else
+		pi->caps_sclk_ds = false;
 
 	pi->mclk_strobe_mode_threshold = 40000;
 	pi->mclk_stutter_mode_threshold = 40000;
@@ -6033,7 +6036,7 @@ static int ci_dpm_init(struct amdgpu_device *adev)
 
 	pi->caps_dynamic_ac_timing = true;
 
-	pi->uvd_power_gated = false;
+	pi->uvd_power_gated = true;
 
 	/* make sure dc limits are valid */
 	if ((adev->pm.dpm.dyn_state.max_clock_voltage_on_dc.sclk == 0) ||
@@ -6176,8 +6179,6 @@ static int ci_dpm_late_init(void *handle)
 	if (ret)
 		return ret;
 
-	ci_dpm_powergate_uvd(adev, true);
-
 	return 0;
 }
 

drivers/gpu/drm/amd/amdgpu/cik.c

@@ -67,6 +67,7 @@
 
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_powerplay.h"
+#include "dce_virtual.h"
 
 /*
  * Indirect registers accessor
@@ -962,12 +963,6 @@ static bool cik_read_bios_from_rom(struct amdgpu_device *adev,
 	return true;
 }
 
-static u32 cik_get_virtual_caps(struct amdgpu_device *adev)
-{
-	/* CIK does not support SR-IOV */
-	return 0;
-}
-
 static const struct amdgpu_allowed_register_entry cik_allowed_read_registers[] = {
 	{mmGRBM_STATUS, false},
 	{mmGB_ADDR_CONFIG, false},
@@ -1640,6 +1635,12 @@ static uint32_t cik_get_rev_id(struct amdgpu_device *adev)
 		>> CC_DRM_ID_STRAPS__ATI_REV_ID__SHIFT;
 }
 
+static void cik_detect_hw_virtualization(struct amdgpu_device *adev)
+{
+	if (is_virtual_machine()) /* passthrough mode */
+		adev->virtualization.virtual_caps |= AMDGPU_PASSTHROUGH_MODE;
+}
+
 static const struct amdgpu_ip_block_version bonaire_ip_blocks[] =
 {
 	/* ORDER MATTERS! */
@@ -1708,6 +1709,74 @@ static const struct amdgpu_ip_block_version bonaire_ip_blocks[] =
 	},
 };
 
+static const struct amdgpu_ip_block_version bonaire_ip_blocks_vd[] =
+{
+	/* ORDER MATTERS! */
+	{
+		.type = AMD_IP_BLOCK_TYPE_COMMON,
+		.major = 1,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_common_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &gmc_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_IH,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_ih_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &amdgpu_pp_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_DCE,
+		.major = 8,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &dce_virtual_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GFX,
+		.major = 7,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &gfx_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SDMA,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_sdma_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_UVD,
+		.major = 4,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &uvd_v4_2_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_VCE,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &vce_v2_0_ip_funcs,
+	},
+};
+
 static const struct amdgpu_ip_block_version hawaii_ip_blocks[] =
 {
 	/* ORDER MATTERS! */
@@ -1776,6 +1845,74 @@ static const struct amdgpu_ip_block_version hawaii_ip_blocks[] =
 	},
 };
 
+static const struct amdgpu_ip_block_version hawaii_ip_blocks_vd[] =
+{
+	/* ORDER MATTERS! */
+	{
+		.type = AMD_IP_BLOCK_TYPE_COMMON,
+		.major = 1,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_common_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &gmc_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_IH,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_ih_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &amdgpu_pp_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_DCE,
+		.major = 8,
+		.minor = 5,
+		.rev = 0,
+		.funcs = &dce_virtual_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GFX,
+		.major = 7,
+		.minor = 3,
+		.rev = 0,
+		.funcs = &gfx_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SDMA,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_sdma_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_UVD,
+		.major = 4,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &uvd_v4_2_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_VCE,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &vce_v2_0_ip_funcs,
+	},
+};
+
 static const struct amdgpu_ip_block_version kabini_ip_blocks[] =
 {
 	/* ORDER MATTERS! */
@@ -1844,6 +1981,74 @@ static const struct amdgpu_ip_block_version kabini_ip_blocks[] =
 	},
 };
 
+static const struct amdgpu_ip_block_version kabini_ip_blocks_vd[] =
+{
+	/* ORDER MATTERS! */
+	{
+		.type = AMD_IP_BLOCK_TYPE_COMMON,
+		.major = 1,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_common_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &gmc_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_IH,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_ih_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &amdgpu_pp_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_DCE,
+		.major = 8,
+		.minor = 3,
+		.rev = 0,
+		.funcs = &dce_virtual_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GFX,
+		.major = 7,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &gfx_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SDMA,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_sdma_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_UVD,
+		.major = 4,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &uvd_v4_2_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_VCE,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &vce_v2_0_ip_funcs,
+	},
+};
+
 static const struct amdgpu_ip_block_version mullins_ip_blocks[] =
 {
 	/* ORDER MATTERS! */
@@ -1912,6 +2117,74 @@ static const struct amdgpu_ip_block_version mullins_ip_blocks[] =
 	},
 };
 
+static const struct amdgpu_ip_block_version mullins_ip_blocks_vd[] =
+{
+	/* ORDER MATTERS! */
+	{
+		.type = AMD_IP_BLOCK_TYPE_COMMON,
+		.major = 1,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_common_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &gmc_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_IH,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_ih_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &amdgpu_pp_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_DCE,
+		.major = 8,
+		.minor = 3,
+		.rev = 0,
+		.funcs = &dce_virtual_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GFX,
+		.major = 7,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &gfx_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SDMA,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_sdma_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_UVD,
+		.major = 4,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &uvd_v4_2_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_VCE,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &vce_v2_0_ip_funcs,
+	},
+};
+
 static const struct amdgpu_ip_block_version kaveri_ip_blocks[] =
 {
 	/* ORDER MATTERS! */
@@ -1980,32 +2253,128 @@ static const struct amdgpu_ip_block_version kaveri_ip_blocks[] =
 	},
 };
 
+static const struct amdgpu_ip_block_version kaveri_ip_blocks_vd[] =
+{
+	/* ORDER MATTERS! */
+	{
+		.type = AMD_IP_BLOCK_TYPE_COMMON,
+		.major = 1,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_common_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &gmc_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_IH,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_ih_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SMC,
+		.major = 7,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &amdgpu_pp_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_DCE,
+		.major = 8,
+		.minor = 1,
+		.rev = 0,
+		.funcs = &dce_virtual_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_GFX,
+		.major = 7,
+		.minor = 1,
+		.rev = 0,
+		.funcs = &gfx_v7_0_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_SDMA,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &cik_sdma_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_UVD,
+		.major = 4,
+		.minor = 2,
+		.rev = 0,
+		.funcs = &uvd_v4_2_ip_funcs,
+	},
+	{
+		.type = AMD_IP_BLOCK_TYPE_VCE,
+		.major = 2,
+		.minor = 0,
+		.rev = 0,
+		.funcs = &vce_v2_0_ip_funcs,
+	},
+};
+
 int cik_set_ip_blocks(struct amdgpu_device *adev)
 {
-	switch (adev->asic_type) {
-	case CHIP_BONAIRE:
-		adev->ip_blocks = bonaire_ip_blocks;
-		adev->num_ip_blocks = ARRAY_SIZE(bonaire_ip_blocks);
-		break;
-	case CHIP_HAWAII:
-		adev->ip_blocks = hawaii_ip_blocks;
-		adev->num_ip_blocks = ARRAY_SIZE(hawaii_ip_blocks);
-		break;
-	case CHIP_KAVERI:
-		adev->ip_blocks = kaveri_ip_blocks;
-		adev->num_ip_blocks = ARRAY_SIZE(kaveri_ip_blocks);
-		break;
-	case CHIP_KABINI:
-		adev->ip_blocks = kabini_ip_blocks;
-		adev->num_ip_blocks = ARRAY_SIZE(kabini_ip_blocks);
-		break;
-	case CHIP_MULLINS:
-		adev->ip_blocks = mullins_ip_blocks;
-		adev->num_ip_blocks = ARRAY_SIZE(mullins_ip_blocks);
-		break;
-	default:
-		/* FIXME: not supported yet */
-		return -EINVAL;
+	if (adev->enable_virtual_display) {
+		switch (adev->asic_type) {
+		case CHIP_BONAIRE:
+			adev->ip_blocks = bonaire_ip_blocks_vd;
+			adev->num_ip_blocks = ARRAY_SIZE(bonaire_ip_blocks_vd);
+			break;
+		case CHIP_HAWAII:
+			adev->ip_blocks = hawaii_ip_blocks_vd;
+			adev->num_ip_blocks = ARRAY_SIZE(hawaii_ip_blocks_vd);
+			break;
+		case CHIP_KAVERI:
+			adev->ip_blocks = kaveri_ip_blocks_vd;
+			adev->num_ip_blocks = ARRAY_SIZE(kaveri_ip_blocks_vd);
+			break;
+		case CHIP_KABINI:
+			adev->ip_blocks = kabini_ip_blocks_vd;
+			adev->num_ip_blocks = ARRAY_SIZE(kabini_ip_blocks_vd);
+			break;
+		case CHIP_MULLINS:
+			adev->ip_blocks = mullins_ip_blocks_vd;
+			adev->num_ip_blocks = ARRAY_SIZE(mullins_ip_blocks_vd);
+			break;
+		default:
+			/* FIXME: not supported yet */
+			return -EINVAL;
+		}
+	} else {
+		switch (adev->asic_type) {
+		case CHIP_BONAIRE:
+			adev->ip_blocks = bonaire_ip_blocks;
+			adev->num_ip_blocks = ARRAY_SIZE(bonaire_ip_blocks);
+			break;
+		case CHIP_HAWAII:
+			adev->ip_blocks = hawaii_ip_blocks;
+			adev->num_ip_blocks = ARRAY_SIZE(hawaii_ip_blocks);
+			break;
+		case CHIP_KAVERI:
+			adev->ip_blocks = kaveri_ip_blocks;
+			adev->num_ip_blocks = ARRAY_SIZE(kaveri_ip_blocks);
+			break;
+		case CHIP_KABINI:
+			adev->ip_blocks = kabini_ip_blocks;
+			adev->num_ip_blocks = ARRAY_SIZE(kabini_ip_blocks);
+			break;
+		case CHIP_MULLINS:
+			adev->ip_blocks = mullins_ip_blocks;
+			adev->num_ip_blocks = ARRAY_SIZE(mullins_ip_blocks);
+			break;
+		default:
+			/* FIXME: not supported yet */
+			return -EINVAL;
+		}
 	}
 
 	return 0;
@@ -2015,13 +2384,13 @@ static const struct amdgpu_asic_funcs cik_asic_funcs =
 {
 	.read_disabled_bios = &cik_read_disabled_bios,
 	.read_bios_from_rom = &cik_read_bios_from_rom,
+	.detect_hw_virtualization = cik_detect_hw_virtualization,
 	.read_register = &cik_read_register,
 	.reset = &cik_asic_reset,
 	.set_vga_state = &cik_vga_set_state,
 	.get_xclk = &cik_get_xclk,
 	.set_uvd_clocks = &cik_set_uvd_clocks,
 	.set_vce_clocks = &cik_set_vce_clocks,
-	.get_virtual_caps = &cik_get_virtual_caps,
 };
 
 static int cik_common_early_init(void *handle)

drivers/gpu/drm/amd/amdgpu/cik_sdma.c

@@ -695,24 +695,16 @@ static void cik_sdma_vm_copy_pte(struct amdgpu_ib *ib,
 				 uint64_t pe, uint64_t src,
 				 unsigned count)
 {
-	while (count) {
-		unsigned bytes = count * 8;
-		if (bytes > 0x1FFFF8)
-			bytes = 0x1FFFF8;
-
-		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_COPY,
-			SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
-		ib->ptr[ib->length_dw++] = bytes;
-		ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
-		ib->ptr[ib->length_dw++] = lower_32_bits(src);
-		ib->ptr[ib->length_dw++] = upper_32_bits(src);
-		ib->ptr[ib->length_dw++] = lower_32_bits(pe);
-		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-
-		pe += bytes;
-		src += bytes;
-		count -= bytes / 8;
-	}
+	unsigned bytes = count * 8;
+
+	ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_COPY,
+		SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+	ib->ptr[ib->length_dw++] = bytes;
+	ib->ptr[ib->length_dw++] = 0; /* src/dst endian swap */
+	ib->ptr[ib->length_dw++] = lower_32_bits(src);
+	ib->ptr[ib->length_dw++] = upper_32_bits(src);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
 }
 
 /**
@@ -720,39 +712,27 @@ static void cik_sdma_vm_copy_pte(struct amdgpu_ib *ib,
  *
  * @ib: indirect buffer to fill with commands
  * @pe: addr of the page entry
- * @addr: dst addr to write into pe
+ * @value: dst addr to write into pe
  * @count: number of page entries to update
  * @incr: increase next addr by incr bytes
- * @flags: access flags
  *
  * Update PTEs by writing them manually using sDMA (CIK).
  */
-static void cik_sdma_vm_write_pte(struct amdgpu_ib *ib,
-				  const dma_addr_t *pages_addr, uint64_t pe,
-				  uint64_t addr, unsigned count,
-				  uint32_t incr, uint32_t flags)
+static void cik_sdma_vm_write_pte(struct amdgpu_ib *ib, uint64_t pe,
+				  uint64_t value, unsigned count,
+				  uint32_t incr)
 {
-	uint64_t value;
-	unsigned ndw;
-
-	while (count) {
-		ndw = count * 2;
-		if (ndw > 0xFFFFE)
-			ndw = 0xFFFFE;
-
-		/* for non-physically contiguous pages (system) */
-		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE,
-			SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
-		ib->ptr[ib->length_dw++] = pe;
-		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-		ib->ptr[ib->length_dw++] = ndw;
-		for (; ndw > 0; ndw -= 2, --count, pe += 8) {
-			value = amdgpu_vm_map_gart(pages_addr, addr);
-			addr += incr;
-			value |= flags;
-			ib->ptr[ib->length_dw++] = value;
-			ib->ptr[ib->length_dw++] = upper_32_bits(value);
-		}
+	unsigned ndw = count * 2;
+
+	ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_WRITE,
+		SDMA_WRITE_SUB_OPCODE_LINEAR, 0);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe);
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = ndw;
+	for (; ndw > 0; ndw -= 2) {
+		ib->ptr[ib->length_dw++] = lower_32_bits(value);
+		ib->ptr[ib->length_dw++] = upper_32_bits(value);
+		value += incr;
 	}
 }
 
@@ -768,40 +748,21 @@ static void cik_sdma_vm_write_pte(struct amdgpu_ib *ib,
  *
  * Update the page tables using sDMA (CIK).
  */
-static void cik_sdma_vm_set_pte_pde(struct amdgpu_ib *ib,
-				    uint64_t pe,
+static void cik_sdma_vm_set_pte_pde(struct amdgpu_ib *ib, uint64_t pe,
 				    uint64_t addr, unsigned count,
 				    uint32_t incr, uint32_t flags)
 {
-	uint64_t value;
-	unsigned ndw;
-
-	while (count) {
-		ndw = count;
-		if (ndw > 0x7FFFF)
-			ndw = 0x7FFFF;
-
-		if (flags & AMDGPU_PTE_VALID)
-			value = addr;
-		else
-			value = 0;
-
-		/* for physically contiguous pages (vram) */
-		ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0);
-		ib->ptr[ib->length_dw++] = pe; /* dst addr */
-		ib->ptr[ib->length_dw++] = upper_32_bits(pe);
-		ib->ptr[ib->length_dw++] = flags; /* mask */
-		ib->ptr[ib->length_dw++] = 0;
-		ib->ptr[ib->length_dw++] = value; /* value */
-		ib->ptr[ib->length_dw++] = upper_32_bits(value);
-		ib->ptr[ib->length_dw++] = incr; /* increment size */
-		ib->ptr[ib->length_dw++] = 0;
-		ib->ptr[ib->length_dw++] = ndw; /* number of entries */
-
-		pe += ndw * 8;
-		addr += ndw * incr;
-		count -= ndw;
-	}
+	/* for physically contiguous pages (vram) */
+	ib->ptr[ib->length_dw++] = SDMA_PACKET(SDMA_OPCODE_GENERATE_PTE_PDE, 0, 0);
+	ib->ptr[ib->length_dw++] = lower_32_bits(pe); /* dst addr */
+	ib->ptr[ib->length_dw++] = upper_32_bits(pe);
+	ib->ptr[ib->length_dw++] = flags; /* mask */
+	ib->ptr[ib->length_dw++] = 0;
+	ib->ptr[ib->length_dw++] = lower_32_bits(addr); /* value */
+	ib->ptr[ib->length_dw++] = upper_32_bits(addr);
+	ib->ptr[ib->length_dw++] = incr; /* increment size */
+	ib->ptr[ib->length_dw++] = 0;
+	ib->ptr[ib->length_dw++] = count; /* number of entries */
 }
 
 /**
@@ -887,6 +848,22 @@ static void cik_sdma_ring_emit_vm_flush(struct amdgpu_ring *ring,
 	amdgpu_ring_write(ring, (0xfff << 16) | 10); /* retry count, poll interval */
 }
 
+static unsigned cik_sdma_ring_get_emit_ib_size(struct amdgpu_ring *ring)
+{
+	return
+		7 + 4; /* cik_sdma_ring_emit_ib */
+}
+
+static unsigned cik_sdma_ring_get_dma_frame_size(struct amdgpu_ring *ring)
+{
+	return
+		6 + /* cik_sdma_ring_emit_hdp_flush */
+		3 + /* cik_sdma_ring_emit_hdp_invalidate */
+		6 + /* cik_sdma_ring_emit_pipeline_sync */
+		12 + /* cik_sdma_ring_emit_vm_flush */
+		9 + 9 + 9; /* cik_sdma_ring_emit_fence x3 for user fence, vm fence */
+}
+
 static void cik_enable_sdma_mgcg(struct amdgpu_device *adev,
 				 bool enable)
 {
@@ -1262,6 +1239,8 @@ static const struct amdgpu_ring_funcs cik_sdma_ring_funcs = {
 	.test_ib = cik_sdma_ring_test_ib,
 	.insert_nop = cik_sdma_ring_insert_nop,
 	.pad_ib = cik_sdma_ring_pad_ib,
+	.get_emit_ib_size = cik_sdma_ring_get_emit_ib_size,
+	.get_dma_frame_size = cik_sdma_ring_get_dma_frame_size,
 };
 
 static void cik_sdma_set_ring_funcs(struct amdgpu_device *adev)

drivers/gpu/drm/amd/amdgpu/cikd.h

@@ -562,4 +562,40 @@ enum {
 	MTYPE_NONCACHED = 3
 };
 
+/* mmPA_SC_RASTER_CONFIG mask */
+#define RB_MAP_PKR0(x)				((x) << 0)
+#define RB_MAP_PKR0_MASK			(0x3 << 0)
+#define RB_MAP_PKR1(x)				((x) << 2)
+#define RB_MAP_PKR1_MASK			(0x3 << 2)
+#define RB_XSEL2(x)				((x) << 4)
+#define RB_XSEL2_MASK				(0x3 << 4)
+#define RB_XSEL					(1 << 6)
+#define RB_YSEL					(1 << 7)
+#define PKR_MAP(x)				((x) << 8)
+#define PKR_MAP_MASK				(0x3 << 8)
+#define PKR_XSEL(x)				((x) << 10)
+#define PKR_XSEL_MASK				(0x3 << 10)
+#define PKR_YSEL(x)				((x) << 12)
+#define PKR_YSEL_MASK				(0x3 << 12)
+#define SC_MAP(x)				((x) << 16)
+#define SC_MAP_MASK				(0x3 << 16)
+#define SC_XSEL(x)				((x) << 18)
+#define SC_XSEL_MASK				(0x3 << 18)
+#define SC_YSEL(x)				((x) << 20)
+#define SC_YSEL_MASK				(0x3 << 20)
+#define SE_MAP(x)				((x) << 24)
+#define SE_MAP_MASK				(0x3 << 24)
+#define SE_XSEL(x)				((x) << 26)
+#define SE_XSEL_MASK				(0x3 << 26)
+#define SE_YSEL(x)				((x) << 28)
+#define SE_YSEL_MASK				(0x3 << 28)
+
+/* mmPA_SC_RASTER_CONFIG_1 mask */
+#define SE_PAIR_MAP(x)				((x) << 0)
+#define SE_PAIR_MAP_MASK			(0x3 << 0)
+#define SE_PAIR_XSEL(x)				((x) << 2)
+#define SE_PAIR_XSEL_MASK			(0x3 << 2)
+#define SE_PAIR_YSEL(x)				((x) << 4)
+#define SE_PAIR_YSEL_MASK			(0x3 << 4)
+
 #endif

drivers/gpu/drm/amd/amdgpu/cz_dpm.c

@@ -44,6 +44,7 @@
 
 static void cz_dpm_powergate_uvd(struct amdgpu_device *adev, bool gate);
 static void cz_dpm_powergate_vce(struct amdgpu_device *adev, bool gate);
+static void cz_dpm_fini(struct amdgpu_device *adev);
 
 static struct cz_ps *cz_get_ps(struct amdgpu_ps *rps)
 {
@@ -350,6 +351,8 @@ static int cz_parse_power_table(struct amdgpu_device *adev)
 
 		ps = kzalloc(sizeof(struct cz_ps), GFP_KERNEL);
 		if (ps == NULL) {
+			for (j = 0; j < i; j++)
+				kfree(adev->pm.dpm.ps[j].ps_priv);
 			kfree(adev->pm.dpm.ps);
 			return -ENOMEM;
 		}
@@ -409,11 +412,11 @@ static int cz_dpm_init(struct amdgpu_device *adev)
 
 	ret = amdgpu_get_platform_caps(adev);
 	if (ret)
-		return ret;
+		goto err;
 
 	ret = amdgpu_parse_extended_power_table(adev);
 	if (ret)
-		return ret;
+		goto err;
 
 	pi->sram_end = SMC_RAM_END;
 
@@ -435,7 +438,11 @@ static int cz_dpm_init(struct amdgpu_device *adev)
 		pi->caps_td_ramping = true;
 		pi->caps_tcp_ramping = true;
 	}
-	pi->caps_sclk_ds = true;
+	if (amdgpu_sclk_deep_sleep_en)
+		pi->caps_sclk_ds = true;
+	else
+		pi->caps_sclk_ds = false;
+
 	pi->voting_clients = 0x00c00033;
 	pi->auto_thermal_throttling_enabled = true;
 	pi->bapm_enabled = false;
@@ -463,23 +470,26 @@ static int cz_dpm_init(struct amdgpu_device *adev)
 
 	ret = cz_parse_sys_info_table(adev);
 	if (ret)
-		return ret;
+		goto err;
 
 	cz_patch_voltage_values(adev);
 	cz_construct_boot_state(adev);
 
 	ret = cz_parse_power_table(adev);
 	if (ret)
-		return ret;
+		goto err;
 
 	ret = cz_process_firmware_header(adev);
 	if (ret)
-		return ret;
+		goto err;
 
 	pi->dpm_enabled = true;
 	pi->uvd_dynamic_pg = false;
 
 	return 0;
+err:
+	cz_dpm_fini(adev);
+	return ret;
 }
 
 static void cz_dpm_fini(struct amdgpu_device *adev)
@@ -668,17 +678,12 @@ static void cz_reset_ap_mask(struct amdgpu_device *adev)
 	struct cz_power_info *pi = cz_get_pi(adev);
 
 	pi->active_process_mask = 0;
-
 }
 
 static int cz_dpm_download_pptable_from_smu(struct amdgpu_device *adev,
 							void **table)
 {
-	int ret = 0;
-
-	ret = cz_smu_download_pptable(adev, table);
-
-	return ret;
+	return cz_smu_download_pptable(adev, table);
 }
 
 static int cz_dpm_upload_pptable_to_smu(struct amdgpu_device *adev)
@@ -818,9 +823,9 @@ static void cz_init_sclk_limit(struct amdgpu_device *adev)
 	pi->sclk_dpm.hard_min_clk = 0;
 	cz_send_msg_to_smc(adev, PPSMC_MSG_GetMaxSclkLevel);
 	level = cz_get_argument(adev);
-	if (level < table->count)
+	if (level < table->count) {
 		clock = table->entries[level].clk;
-	else {
+	} else {
 		DRM_ERROR("Invalid SLCK Voltage Dependency table entry.\n");
 		clock = table->entries[table->count - 1].clk;
 	}
@@ -846,9 +851,9 @@ static void cz_init_uvd_limit(struct amdgpu_device *adev)
 	pi->uvd_dpm.hard_min_clk = 0;
 	cz_send_msg_to_smc(adev, PPSMC_MSG_GetMaxUvdLevel);
 	level = cz_get_argument(adev);
-	if (level < table->count)
+	if (level < table->count) {
 		clock = table->entries[level].vclk;
-	else {
+	} else {
 		DRM_ERROR("Invalid UVD Voltage Dependency table entry.\n");
 		clock = table->entries[table->count - 1].vclk;
 	}
@@ -874,9 +879,9 @@ static void cz_init_vce_limit(struct amdgpu_device *adev)
 	pi->vce_dpm.hard_min_clk = table->entries[0].ecclk;
 	cz_send_msg_to_smc(adev, PPSMC_MSG_GetMaxEclkLevel);
 	level = cz_get_argument(adev);
-	if (level < table->count)
+	if (level < table->count) {
 		clock = table->entries[level].ecclk;
-	else {
+	} else {
 		/* future BIOS would fix this error */
 		DRM_ERROR("Invalid VCE Voltage Dependency table entry.\n");
 		clock = table->entries[table->count - 1].ecclk;
@@ -903,9 +908,9 @@ static void cz_init_acp_limit(struct amdgpu_device *adev)
 	pi->acp_dpm.hard_min_clk = 0;
 	cz_send_msg_to_smc(adev, PPSMC_MSG_GetMaxAclkLevel);
 	level = cz_get_argument(adev);
-	if (level < table->count)
+	if (level < table->count) {
 		clock = table->entries[level].clk;
-	else {
+	} else {
 		DRM_ERROR("Invalid ACP Voltage Dependency table entry.\n");
 		clock = table->entries[table->count - 1].clk;
 	}
@@ -930,7 +935,6 @@ static void cz_init_sclk_threshold(struct amdgpu_device *adev)
 	struct cz_power_info *pi = cz_get_pi(adev);
 
 	pi->low_sclk_interrupt_threshold = 0;
-
 }
 
 static void cz_dpm_setup_asic(struct amdgpu_device *adev)
@@ -1203,7 +1207,7 @@ static int cz_enable_didt(struct amdgpu_device *adev, bool enable)
 	int ret;
 
 	if (pi->caps_sq_ramping || pi->caps_db_ramping ||
-			pi->caps_td_ramping || pi->caps_tcp_ramping) {
+	    pi->caps_td_ramping || pi->caps_tcp_ramping) {
 		if (adev->gfx.gfx_current_status != AMDGPU_GFX_SAFE_MODE) {
 			ret = cz_disable_cgpg(adev);
 			if (ret) {
@@ -1277,7 +1281,7 @@ static void cz_apply_state_adjust_rules(struct amdgpu_device *adev,
 	ps->force_high = false;
 	ps->need_dfs_bypass = true;
 	pi->video_start = new_rps->dclk || new_rps->vclk ||
-				new_rps->evclk || new_rps->ecclk;
+			  new_rps->evclk || new_rps->ecclk;
 
 	if ((new_rps->class & ATOM_PPLIB_CLASSIFICATION_UI_MASK) ==
 			ATOM_PPLIB_CLASSIFICATION_UI_BATTERY)
@@ -1335,7 +1339,6 @@ static int cz_dpm_enable(struct amdgpu_device *adev)
 	}
 
 	cz_reset_acp_boot_level(adev);
-
 	cz_update_current_ps(adev, adev->pm.dpm.boot_ps);
 
 	return 0;
@@ -1665,7 +1668,6 @@ static void cz_dpm_post_set_power_state(struct amdgpu_device *adev)
 	struct amdgpu_ps *ps = &pi->requested_rps;
 
 	cz_update_current_ps(adev, ps);
-
 }
 
 static int cz_dpm_force_highest(struct amdgpu_device *adev)
@@ -2108,29 +2110,58 @@ static void cz_dpm_powergate_uvd(struct amdgpu_device *adev, bool gate)
 			/* disable clockgating so we can properly shut down the block */
 			ret = amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
 							    AMD_CG_STATE_UNGATE);
+			if (ret) {
+				DRM_ERROR("UVD DPM Power Gating failed to set clockgating state\n");
+				return;
+			}
+
 			/* shutdown the UVD block */
 			ret = amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
 							    AMD_PG_STATE_GATE);
-			/* XXX: check for errors */
+
+			if (ret) {
+				DRM_ERROR("UVD DPM Power Gating failed to set powergating state\n");
+				return;
+			}
 		}
 		cz_update_uvd_dpm(adev, gate);
-		if (pi->caps_uvd_pg)
+		if (pi->caps_uvd_pg) {
 			/* power off the UVD block */
-			cz_send_msg_to_smc(adev, PPSMC_MSG_UVDPowerOFF);
+			ret = cz_send_msg_to_smc(adev, PPSMC_MSG_UVDPowerOFF);
+			if (ret) {
+				DRM_ERROR("UVD DPM Power Gating failed to send SMU PowerOFF message\n");
+				return;
+			}
+		}
 	} else {
 		if (pi->caps_uvd_pg) {
 			/* power on the UVD block */
 			if (pi->uvd_dynamic_pg)
-				cz_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_UVDPowerON, 1);
+				ret = cz_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_UVDPowerON, 1);
 			else
-				cz_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_UVDPowerON, 0);
+				ret = cz_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_UVDPowerON, 0);
+
+			if (ret) {
+				DRM_ERROR("UVD DPM Power Gating Failed to send SMU PowerON message\n");
+				return;
+			}
+
 			/* re-init the UVD block */
 			ret = amdgpu_set_powergating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
 							    AMD_PG_STATE_UNGATE);
+
+			if (ret) {
+				DRM_ERROR("UVD DPM Power Gating Failed to set powergating state\n");
+				return;
+			}
+
 			/* enable clockgating. hw will dynamically gate/ungate clocks on the fly */
 			ret = amdgpu_set_clockgating_state(adev, AMD_IP_BLOCK_TYPE_UVD,
 							    AMD_CG_STATE_GATE);
-			/* XXX: check for errors */
+			if (ret) {
+				DRM_ERROR("UVD DPM Power Gating Failed to set clockgating state\n");
+				return;
+			}
 		}
 		cz_update_uvd_dpm(adev, gate);
 	}
@@ -2168,7 +2199,6 @@ static int cz_update_vce_dpm(struct amdgpu_device *adev)
 	/* Stable Pstate is enabled and we need to set the VCE DPM to highest level */
 	if (pi->caps_stable_power_state) {
 		pi->vce_dpm.hard_min_clk = table->entries[table->count-1].ecclk;
-
 	} else { /* non-stable p-state cases. without vce.Arbiter.EcclkHardMin */
 		/* leave it as set by user */
 		/*pi->vce_dpm.hard_min_clk = table->entries[0].ecclk;*/

drivers/gpu/drm/amd/amdgpu/cz_smc.c

@@ -29,6 +29,8 @@
 #include "cz_smumgr.h"
 #include "smu_ucode_xfer_cz.h"
 #include "amdgpu_ucode.h"
+#include "cz_dpm.h"
+#include "vi_dpm.h"
 
 #include "smu/smu_8_0_d.h"
 #include "smu/smu_8_0_sh_mask.h"
@@ -48,7 +50,7 @@ static struct cz_smu_private_data *cz_smu_get_priv(struct amdgpu_device *adev)
 	return priv;
 }
 
-int cz_send_msg_to_smc_async(struct amdgpu_device *adev, u16 msg)
+static int cz_send_msg_to_smc_async(struct amdgpu_device *adev, u16 msg)
 {
 	int i;
 	u32 content = 0, tmp;
@@ -99,13 +101,6 @@ int cz_send_msg_to_smc(struct amdgpu_device *adev, u16 msg)
 	return 0;
 }
 
-int cz_send_msg_to_smc_with_parameter_async(struct amdgpu_device *adev,
-						u16 msg, u32 parameter)
-{
-	WREG32(mmSMU_MP1_SRBM2P_ARG_0, parameter);
-	return cz_send_msg_to_smc_async(adev, msg);
-}
-
 int cz_send_msg_to_smc_with_parameter(struct amdgpu_device *adev,
 						u16 msg, u32 parameter)
 {
@@ -140,7 +135,7 @@ int cz_read_smc_sram_dword(struct amdgpu_device *adev, u32 smc_address,
 	return 0;
 }
 
-int cz_write_smc_sram_dword(struct amdgpu_device *adev, u32 smc_address,
+static int cz_write_smc_sram_dword(struct amdgpu_device *adev, u32 smc_address,
 						u32 value, u32 limit)
 {
 	int ret;

drivers/gpu/drm/amd/amdgpu/dce_v10_0.c

@@ -221,7 +221,7 @@ static bool dce_v10_0_is_counter_moving(struct amdgpu_device *adev, int crtc)
  */
 static void dce_v10_0_vblank_wait(struct amdgpu_device *adev, int crtc)
 {
-	unsigned i = 0;
+	unsigned i = 100;
 
 	if (crtc >= adev->mode_info.num_crtc)
 		return;
@@ -233,14 +233,16 @@ static void dce_v10_0_vblank_wait(struct amdgpu_device *adev, int crtc)
 	 * wait for another frame.
 	 */
 	while (dce_v10_0_is_in_vblank(adev, crtc)) {
-		if (i++ % 100 == 0) {
+		if (i++ == 100) {
+			i = 0;
 			if (!dce_v10_0_is_counter_moving(adev, crtc))
 				break;
 		}
 	}
 
 	while (!dce_v10_0_is_in_vblank(adev, crtc)) {
-		if (i++ % 100 == 0) {
+		if (i++ == 100) {
+			i = 0;
 			if (!dce_v10_0_is_counter_moving(adev, crtc))
 				break;
 		}
@@ -425,16 +427,6 @@ static void dce_v10_0_hpd_init(struct amdgpu_device *adev)
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
 
-		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
-		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
-			/* don't try to enable hpd on eDP or LVDS avoid breaking the
-			 * aux dp channel on imac and help (but not completely fix)
-			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
-			 * also avoid interrupt storms during dpms.
-			 */
-			continue;
-		}
-
 		switch (amdgpu_connector->hpd.hpd) {
 		case AMDGPU_HPD_1:
 			idx = 0;
@@ -458,6 +450,19 @@ static void dce_v10_0_hpd_init(struct amdgpu_device *adev)
 			continue;
 		}
 
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
+		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
+			/* don't try to enable hpd on eDP or LVDS avoid breaking the
+			 * aux dp channel on imac and help (but not completely fix)
+			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
+			 * also avoid interrupt storms during dpms.
+			 */
+			tmp = RREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[idx]);
+			tmp = REG_SET_FIELD(tmp, DC_HPD_INT_CONTROL, DC_HPD_INT_EN, 0);
+			WREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[idx], tmp);
+			continue;
+		}
+
 		tmp = RREG32(mmDC_HPD_CONTROL + hpd_offsets[idx]);
 		tmp = REG_SET_FIELD(tmp, DC_HPD_CONTROL, DC_HPD_EN, 1);
 		WREG32(mmDC_HPD_CONTROL + hpd_offsets[idx], tmp);
@@ -646,8 +651,8 @@ static void dce_v10_0_resume_mc_access(struct amdgpu_device *adev,
 
 		if (save->crtc_enabled[i]) {
 			tmp = RREG32(mmMASTER_UPDATE_MODE + crtc_offsets[i]);
-			if (REG_GET_FIELD(tmp, MASTER_UPDATE_MODE, MASTER_UPDATE_MODE) != 3) {
-				tmp = REG_SET_FIELD(tmp, MASTER_UPDATE_MODE, MASTER_UPDATE_MODE, 3);
+			if (REG_GET_FIELD(tmp, MASTER_UPDATE_MODE, MASTER_UPDATE_MODE) != 0) {
+				tmp = REG_SET_FIELD(tmp, MASTER_UPDATE_MODE, MASTER_UPDATE_MODE, 0);
 				WREG32(mmMASTER_UPDATE_MODE + crtc_offsets[i], tmp);
 			}
 			tmp = RREG32(mmGRPH_UPDATE + crtc_offsets[i]);
@@ -712,6 +717,45 @@ static void dce_v10_0_set_vga_render_state(struct amdgpu_device *adev,
 	WREG32(mmVGA_RENDER_CONTROL, tmp);
 }
 
+static int dce_v10_0_get_num_crtc(struct amdgpu_device *adev)
+{
+	int num_crtc = 0;
+
+	switch (adev->asic_type) {
+	case CHIP_FIJI:
+	case CHIP_TONGA:
+		num_crtc = 6;
+		break;
+	default:
+		num_crtc = 0;
+	}
+	return num_crtc;
+}
+
+void dce_v10_0_disable_dce(struct amdgpu_device *adev)
+{
+	/*Disable VGA render and enabled crtc, if has DCE engine*/
+	if (amdgpu_atombios_has_dce_engine_info(adev)) {
+		u32 tmp;
+		int crtc_enabled, i;
+
+		dce_v10_0_set_vga_render_state(adev, false);
+
+		/*Disable crtc*/
+		for (i = 0; i < dce_v10_0_get_num_crtc(adev); i++) {
+			crtc_enabled = REG_GET_FIELD(RREG32(mmCRTC_CONTROL + crtc_offsets[i]),
+									 CRTC_CONTROL, CRTC_MASTER_EN);
+			if (crtc_enabled) {
+				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+				tmp = RREG32(mmCRTC_CONTROL + crtc_offsets[i]);
+				tmp = REG_SET_FIELD(tmp, CRTC_CONTROL, CRTC_MASTER_EN, 0);
+				WREG32(mmCRTC_CONTROL + crtc_offsets[i], tmp);
+				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 0);
+			}
+		}
+	}
+}
+
 static void dce_v10_0_program_fmt(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
@@ -2063,7 +2107,7 @@ static int dce_v10_0_crtc_do_set_base(struct drm_crtc *crtc,
 	struct amdgpu_framebuffer *amdgpu_fb;
 	struct drm_framebuffer *target_fb;
 	struct drm_gem_object *obj;
-	struct amdgpu_bo *rbo;
+	struct amdgpu_bo *abo;
 	uint64_t fb_location, tiling_flags;
 	uint32_t fb_format, fb_pitch_pixels;
 	u32 fb_swap = REG_SET_FIELD(0, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP, ENDIAN_NONE);
@@ -2071,6 +2115,7 @@ static int dce_v10_0_crtc_do_set_base(struct drm_crtc *crtc,
 	u32 tmp, viewport_w, viewport_h;
 	int r;
 	bool bypass_lut = false;
+	char *format_name;
 
 	/* no fb bound */
 	if (!atomic && !crtc->primary->fb) {
@@ -2090,23 +2135,23 @@ static int dce_v10_0_crtc_do_set_base(struct drm_crtc *crtc,
 	 * just update base pointers
 	 */
 	obj = amdgpu_fb->obj;
-	rbo = gem_to_amdgpu_bo(obj);
-	r = amdgpu_bo_reserve(rbo, false);
+	abo = gem_to_amdgpu_bo(obj);
+	r = amdgpu_bo_reserve(abo, false);
 	if (unlikely(r != 0))
 		return r;
 
 	if (atomic) {
-		fb_location = amdgpu_bo_gpu_offset(rbo);
+		fb_location = amdgpu_bo_gpu_offset(abo);
 	} else {
-		r = amdgpu_bo_pin(rbo, AMDGPU_GEM_DOMAIN_VRAM, &fb_location);
+		r = amdgpu_bo_pin(abo, AMDGPU_GEM_DOMAIN_VRAM, &fb_location);
 		if (unlikely(r != 0)) {
-			amdgpu_bo_unreserve(rbo);
+			amdgpu_bo_unreserve(abo);
 			return -EINVAL;
 		}
 	}
 
-	amdgpu_bo_get_tiling_flags(rbo, &tiling_flags);
-	amdgpu_bo_unreserve(rbo);
+	amdgpu_bo_get_tiling_flags(abo, &tiling_flags);
+	amdgpu_bo_unreserve(abo);
 
 	pipe_config = AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
 
@@ -2182,8 +2227,9 @@ static int dce_v10_0_crtc_do_set_base(struct drm_crtc *crtc,
 		bypass_lut = true;
 		break;
 	default:
-		DRM_ERROR("Unsupported screen format %s\n",
-			drm_get_format_name(target_fb->pixel_format));
+		format_name = drm_get_format_name(target_fb->pixel_format);
+		DRM_ERROR("Unsupported screen format %s\n", format_name);
+		kfree(format_name);
 		return -EINVAL;
 	}
 
@@ -2275,17 +2321,17 @@ static int dce_v10_0_crtc_do_set_base(struct drm_crtc *crtc,
 	WREG32(mmVIEWPORT_SIZE + amdgpu_crtc->crtc_offset,
 	       (viewport_w << 16) | viewport_h);
 
-	/* set pageflip to happen only at start of vblank interval (front porch) */
-	WREG32(mmMASTER_UPDATE_MODE + amdgpu_crtc->crtc_offset, 3);
+	/* set pageflip to happen anywhere in vblank interval */
+	WREG32(mmMASTER_UPDATE_MODE + amdgpu_crtc->crtc_offset, 0);
 
 	if (!atomic && fb && fb != crtc->primary->fb) {
 		amdgpu_fb = to_amdgpu_framebuffer(fb);
-		rbo = gem_to_amdgpu_bo(amdgpu_fb->obj);
-		r = amdgpu_bo_reserve(rbo, false);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
 		if (unlikely(r != 0))
 			return r;
-		amdgpu_bo_unpin(rbo);
-		amdgpu_bo_unreserve(rbo);
+		amdgpu_bo_unpin(abo);
+		amdgpu_bo_unreserve(abo);
 	}
 
 	/* Bytes per pixel may have changed */
@@ -2698,7 +2744,7 @@ static const struct drm_crtc_funcs dce_v10_0_crtc_funcs = {
 	.gamma_set = dce_v10_0_crtc_gamma_set,
 	.set_config = amdgpu_crtc_set_config,
 	.destroy = dce_v10_0_crtc_destroy,
-	.page_flip = amdgpu_crtc_page_flip,
+	.page_flip_target = amdgpu_crtc_page_flip_target,
 };
 
 static void dce_v10_0_crtc_dpms(struct drm_crtc *crtc, int mode)
@@ -2765,16 +2811,16 @@ static void dce_v10_0_crtc_disable(struct drm_crtc *crtc)
 	if (crtc->primary->fb) {
 		int r;
 		struct amdgpu_framebuffer *amdgpu_fb;
-		struct amdgpu_bo *rbo;
+		struct amdgpu_bo *abo;
 
 		amdgpu_fb = to_amdgpu_framebuffer(crtc->primary->fb);
-		rbo = gem_to_amdgpu_bo(amdgpu_fb->obj);
-		r = amdgpu_bo_reserve(rbo, false);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
 		if (unlikely(r))
-			DRM_ERROR("failed to reserve rbo before unpin\n");
+			DRM_ERROR("failed to reserve abo before unpin\n");
 		else {
-			amdgpu_bo_unpin(rbo);
-			amdgpu_bo_unreserve(rbo);
+			amdgpu_bo_unpin(abo);
+			amdgpu_bo_unreserve(abo);
 		}
 	}
 	/* disable the GRPH */
@@ -2962,10 +3008,11 @@ static int dce_v10_0_early_init(void *handle)
 	dce_v10_0_set_display_funcs(adev);
 	dce_v10_0_set_irq_funcs(adev);
 
+	adev->mode_info.num_crtc = dce_v10_0_get_num_crtc(adev);
+
 	switch (adev->asic_type) {
 	case CHIP_FIJI:
 	case CHIP_TONGA:
-		adev->mode_info.num_crtc = 6; /* XXX 7??? */
 		adev->mode_info.num_hpd = 6;
 		adev->mode_info.num_dig = 7;
 		break;
@@ -3141,11 +3188,26 @@ static int dce_v10_0_wait_for_idle(void *handle)
 	return 0;
 }
 
+static int dce_v10_0_check_soft_reset(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	if (dce_v10_0_is_display_hung(adev))
+		adev->ip_block_status[AMD_IP_BLOCK_TYPE_DCE].hang = true;
+	else
+		adev->ip_block_status[AMD_IP_BLOCK_TYPE_DCE].hang = false;
+
+	return 0;
+}
+
 static int dce_v10_0_soft_reset(void *handle)
 {
 	u32 srbm_soft_reset = 0, tmp;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	if (!adev->ip_block_status[AMD_IP_BLOCK_TYPE_DCE].hang)
+		return 0;
+
 	if (dce_v10_0_is_display_hung(adev))
 		srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_DC_MASK;
 
@@ -3512,6 +3574,7 @@ const struct amd_ip_funcs dce_v10_0_ip_funcs = {
 	.resume = dce_v10_0_resume,
 	.is_idle = dce_v10_0_is_idle,
 	.wait_for_idle = dce_v10_0_wait_for_idle,
+	.check_soft_reset = dce_v10_0_check_soft_reset,
 	.soft_reset = dce_v10_0_soft_reset,
 	.set_clockgating_state = dce_v10_0_set_clockgating_state,
 	.set_powergating_state = dce_v10_0_set_powergating_state,

drivers/gpu/drm/amd/amdgpu/dce_v10_0.h

@@ -26,4 +26,6 @@
 
 extern const struct amd_ip_funcs dce_v10_0_ip_funcs;
 
+void dce_v10_0_disable_dce(struct amdgpu_device *adev);
+
 #endif

drivers/gpu/drm/amd/amdgpu/dce_v11_0.c

@@ -443,16 +443,6 @@ static void dce_v11_0_hpd_init(struct amdgpu_device *adev)
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
 
-		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
-		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
-			/* don't try to enable hpd on eDP or LVDS avoid breaking the
-			 * aux dp channel on imac and help (but not completely fix)
-			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
-			 * also avoid interrupt storms during dpms.
-			 */
-			continue;
-		}
-
 		switch (amdgpu_connector->hpd.hpd) {
 		case AMDGPU_HPD_1:
 			idx = 0;
@@ -476,6 +466,19 @@ static void dce_v11_0_hpd_init(struct amdgpu_device *adev)
 			continue;
 		}
 
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
+		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
+			/* don't try to enable hpd on eDP or LVDS avoid breaking the
+			 * aux dp channel on imac and help (but not completely fix)
+			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
+			 * also avoid interrupt storms during dpms.
+			 */
+			tmp = RREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[idx]);
+			tmp = REG_SET_FIELD(tmp, DC_HPD_INT_CONTROL, DC_HPD_INT_EN, 0);
+			WREG32(mmDC_HPD_INT_CONTROL + hpd_offsets[idx], tmp);
+			continue;
+		}
+
 		tmp = RREG32(mmDC_HPD_CONTROL + hpd_offsets[idx]);
 		tmp = REG_SET_FIELD(tmp, DC_HPD_CONTROL, DC_HPD_EN, 1);
 		WREG32(mmDC_HPD_CONTROL + hpd_offsets[idx], tmp);
@@ -673,6 +676,53 @@ static void dce_v11_0_set_vga_render_state(struct amdgpu_device *adev,
 	WREG32(mmVGA_RENDER_CONTROL, tmp);
 }
 
+static int dce_v11_0_get_num_crtc (struct amdgpu_device *adev)
+{
+	int num_crtc = 0;
+
+	switch (adev->asic_type) {
+	case CHIP_CARRIZO:
+		num_crtc = 3;
+		break;
+	case CHIP_STONEY:
+		num_crtc = 2;
+		break;
+	case CHIP_POLARIS10:
+		num_crtc = 6;
+		break;
+	case CHIP_POLARIS11:
+		num_crtc = 5;
+		break;
+	default:
+		num_crtc = 0;
+	}
+	return num_crtc;
+}
+
+void dce_v11_0_disable_dce(struct amdgpu_device *adev)
+{
+	/*Disable VGA render and enabled crtc, if has DCE engine*/
+	if (amdgpu_atombios_has_dce_engine_info(adev)) {
+		u32 tmp;
+		int crtc_enabled, i;
+
+		dce_v11_0_set_vga_render_state(adev, false);
+
+		/*Disable crtc*/
+		for (i = 0; i < dce_v11_0_get_num_crtc(adev); i++) {
+			crtc_enabled = REG_GET_FIELD(RREG32(mmCRTC_CONTROL + crtc_offsets[i]),
+									 CRTC_CONTROL, CRTC_MASTER_EN);
+			if (crtc_enabled) {
+				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+				tmp = RREG32(mmCRTC_CONTROL + crtc_offsets[i]);
+				tmp = REG_SET_FIELD(tmp, CRTC_CONTROL, CRTC_MASTER_EN, 0);
+				WREG32(mmCRTC_CONTROL + crtc_offsets[i], tmp);
+				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 0);
+			}
+		}
+	}
+}
+
 static void dce_v11_0_program_fmt(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
@@ -2038,7 +2088,7 @@ static int dce_v11_0_crtc_do_set_base(struct drm_crtc *crtc,
 	struct amdgpu_framebuffer *amdgpu_fb;
 	struct drm_framebuffer *target_fb;
 	struct drm_gem_object *obj;
-	struct amdgpu_bo *rbo;
+	struct amdgpu_bo *abo;
 	uint64_t fb_location, tiling_flags;
 	uint32_t fb_format, fb_pitch_pixels;
 	u32 fb_swap = REG_SET_FIELD(0, GRPH_SWAP_CNTL, GRPH_ENDIAN_SWAP, ENDIAN_NONE);
@@ -2046,6 +2096,7 @@ static int dce_v11_0_crtc_do_set_base(struct drm_crtc *crtc,
 	u32 tmp, viewport_w, viewport_h;
 	int r;
 	bool bypass_lut = false;
+	char *format_name;
 
 	/* no fb bound */
 	if (!atomic && !crtc->primary->fb) {
@@ -2065,23 +2116,23 @@ static int dce_v11_0_crtc_do_set_base(struct drm_crtc *crtc,
 	 * just update base pointers
 	 */
 	obj = amdgpu_fb->obj;
-	rbo = gem_to_amdgpu_bo(obj);
-	r = amdgpu_bo_reserve(rbo, false);
+	abo = gem_to_amdgpu_bo(obj);
+	r = amdgpu_bo_reserve(abo, false);
 	if (unlikely(r != 0))
 		return r;
 
 	if (atomic) {
-		fb_location = amdgpu_bo_gpu_offset(rbo);
+		fb_location = amdgpu_bo_gpu_offset(abo);
 	} else {
-		r = amdgpu_bo_pin(rbo, AMDGPU_GEM_DOMAIN_VRAM, &fb_location);
+		r = amdgpu_bo_pin(abo, AMDGPU_GEM_DOMAIN_VRAM, &fb_location);
 		if (unlikely(r != 0)) {
-			amdgpu_bo_unreserve(rbo);
+			amdgpu_bo_unreserve(abo);
 			return -EINVAL;
 		}
 	}
 
-	amdgpu_bo_get_tiling_flags(rbo, &tiling_flags);
-	amdgpu_bo_unreserve(rbo);
+	amdgpu_bo_get_tiling_flags(abo, &tiling_flags);
+	amdgpu_bo_unreserve(abo);
 
 	pipe_config = AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
 
@@ -2157,8 +2208,9 @@ static int dce_v11_0_crtc_do_set_base(struct drm_crtc *crtc,
 		bypass_lut = true;
 		break;
 	default:
-		DRM_ERROR("Unsupported screen format %s\n",
-			drm_get_format_name(target_fb->pixel_format));
+		format_name = drm_get_format_name(target_fb->pixel_format);
+		DRM_ERROR("Unsupported screen format %s\n", format_name);
+		kfree(format_name);
 		return -EINVAL;
 	}
 
@@ -2250,17 +2302,17 @@ static int dce_v11_0_crtc_do_set_base(struct drm_crtc *crtc,
 	WREG32(mmVIEWPORT_SIZE + amdgpu_crtc->crtc_offset,
 	       (viewport_w << 16) | viewport_h);
 
-	/* set pageflip to happen only at start of vblank interval (front porch) */
-	WREG32(mmCRTC_MASTER_UPDATE_MODE + amdgpu_crtc->crtc_offset, 3);
+	/* set pageflip to happen anywhere in vblank interval */
+	WREG32(mmCRTC_MASTER_UPDATE_MODE + amdgpu_crtc->crtc_offset, 0);
 
 	if (!atomic && fb && fb != crtc->primary->fb) {
 		amdgpu_fb = to_amdgpu_framebuffer(fb);
-		rbo = gem_to_amdgpu_bo(amdgpu_fb->obj);
-		r = amdgpu_bo_reserve(rbo, false);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
 		if (unlikely(r != 0))
 			return r;
-		amdgpu_bo_unpin(rbo);
-		amdgpu_bo_unreserve(rbo);
+		amdgpu_bo_unpin(abo);
+		amdgpu_bo_unreserve(abo);
 	}
 
 	/* Bytes per pixel may have changed */
@@ -2708,7 +2760,7 @@ static const struct drm_crtc_funcs dce_v11_0_crtc_funcs = {
 	.gamma_set = dce_v11_0_crtc_gamma_set,
 	.set_config = amdgpu_crtc_set_config,
 	.destroy = dce_v11_0_crtc_destroy,
-	.page_flip = amdgpu_crtc_page_flip,
+	.page_flip_target = amdgpu_crtc_page_flip_target,
 };
 
 static void dce_v11_0_crtc_dpms(struct drm_crtc *crtc, int mode)
@@ -2775,16 +2827,16 @@ static void dce_v11_0_crtc_disable(struct drm_crtc *crtc)
 	if (crtc->primary->fb) {
 		int r;
 		struct amdgpu_framebuffer *amdgpu_fb;
-		struct amdgpu_bo *rbo;
+		struct amdgpu_bo *abo;
 
 		amdgpu_fb = to_amdgpu_framebuffer(crtc->primary->fb);
-		rbo = gem_to_amdgpu_bo(amdgpu_fb->obj);
-		r = amdgpu_bo_reserve(rbo, false);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
 		if (unlikely(r))
-			DRM_ERROR("failed to reserve rbo before unpin\n");
+			DRM_ERROR("failed to reserve abo before unpin\n");
 		else {
-			amdgpu_bo_unpin(rbo);
-			amdgpu_bo_unreserve(rbo);
+			amdgpu_bo_unpin(abo);
+			amdgpu_bo_unreserve(abo);
 		}
 	}
 	/* disable the GRPH */
@@ -2999,24 +3051,22 @@ static int dce_v11_0_early_init(void *handle)
 	dce_v11_0_set_display_funcs(adev);
 	dce_v11_0_set_irq_funcs(adev);
 
+	adev->mode_info.num_crtc = dce_v11_0_get_num_crtc(adev);
+
 	switch (adev->asic_type) {
 	case CHIP_CARRIZO:
-		adev->mode_info.num_crtc = 3;
 		adev->mode_info.num_hpd = 6;
 		adev->mode_info.num_dig = 9;
 		break;
 	case CHIP_STONEY:
-		adev->mode_info.num_crtc = 2;
 		adev->mode_info.num_hpd = 6;
 		adev->mode_info.num_dig = 9;
 		break;
 	case CHIP_POLARIS10:
-		adev->mode_info.num_crtc = 6;
 		adev->mode_info.num_hpd = 6;
 		adev->mode_info.num_dig = 6;
 		break;
 	case CHIP_POLARIS11:
-		adev->mode_info.num_crtc = 5;
 		adev->mode_info.num_hpd = 5;
 		adev->mode_info.num_dig = 5;
 		break;
@@ -3109,6 +3159,7 @@ static int dce_v11_0_sw_fini(void *handle)
 
 	dce_v11_0_afmt_fini(adev);
 
+	drm_mode_config_cleanup(adev->ddev);
 	adev->mode_info.mode_config_initialized = false;
 
 	return 0;

drivers/gpu/drm/amd/amdgpu/dce_v11_0.h

@@ -26,4 +26,6 @@
 
 extern const struct amd_ip_funcs dce_v11_0_ip_funcs;
 
+void dce_v11_0_disable_dce(struct amdgpu_device *adev);
+
 #endif

drivers/gpu/drm/amd/amdgpu/dce_v6_0.c

@@ -0,0 +1,3176 @@
+/*
+ * Copyright 2015 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "drmP.h"
+#include "amdgpu.h"
+#include "amdgpu_pm.h"
+#include "amdgpu_i2c.h"
+#include "atom.h"
+#include "amdgpu_atombios.h"
+#include "atombios_crtc.h"
+#include "atombios_encoders.h"
+#include "amdgpu_pll.h"
+#include "amdgpu_connectors.h"
+#include "si/si_reg.h"
+#include "si/sid.h"
+
+static void dce_v6_0_set_display_funcs(struct amdgpu_device *adev);
+static void dce_v6_0_set_irq_funcs(struct amdgpu_device *adev);
+
+static const u32 crtc_offsets[6] =
+{
+	SI_CRTC0_REGISTER_OFFSET,
+	SI_CRTC1_REGISTER_OFFSET,
+	SI_CRTC2_REGISTER_OFFSET,
+	SI_CRTC3_REGISTER_OFFSET,
+	SI_CRTC4_REGISTER_OFFSET,
+	SI_CRTC5_REGISTER_OFFSET
+};
+
+static const uint32_t dig_offsets[] = {
+	SI_CRTC0_REGISTER_OFFSET,
+	SI_CRTC1_REGISTER_OFFSET,
+	SI_CRTC2_REGISTER_OFFSET,
+	SI_CRTC3_REGISTER_OFFSET,
+	SI_CRTC4_REGISTER_OFFSET,
+	SI_CRTC5_REGISTER_OFFSET,
+	(0x13830 - 0x7030) >> 2,
+};
+
+static const struct {
+	uint32_t	reg;
+	uint32_t	vblank;
+	uint32_t	vline;
+	uint32_t	hpd;
+
+} interrupt_status_offsets[6] = { {
+	.reg = DISP_INTERRUPT_STATUS,
+	.vblank = DISP_INTERRUPT_STATUS__LB_D1_VBLANK_INTERRUPT_MASK,
+	.vline = DISP_INTERRUPT_STATUS__LB_D1_VLINE_INTERRUPT_MASK,
+	.hpd = DISP_INTERRUPT_STATUS__DC_HPD1_INTERRUPT_MASK
+}, {
+	.reg = DISP_INTERRUPT_STATUS_CONTINUE,
+	.vblank = DISP_INTERRUPT_STATUS_CONTINUE__LB_D2_VBLANK_INTERRUPT_MASK,
+	.vline = DISP_INTERRUPT_STATUS_CONTINUE__LB_D2_VLINE_INTERRUPT_MASK,
+	.hpd = DISP_INTERRUPT_STATUS_CONTINUE__DC_HPD2_INTERRUPT_MASK
+}, {
+	.reg = DISP_INTERRUPT_STATUS_CONTINUE2,
+	.vblank = DISP_INTERRUPT_STATUS_CONTINUE2__LB_D3_VBLANK_INTERRUPT_MASK,
+	.vline = DISP_INTERRUPT_STATUS_CONTINUE2__LB_D3_VLINE_INTERRUPT_MASK,
+	.hpd = DISP_INTERRUPT_STATUS_CONTINUE2__DC_HPD3_INTERRUPT_MASK
+}, {
+	.reg = DISP_INTERRUPT_STATUS_CONTINUE3,
+	.vblank = DISP_INTERRUPT_STATUS_CONTINUE3__LB_D4_VBLANK_INTERRUPT_MASK,
+	.vline = DISP_INTERRUPT_STATUS_CONTINUE3__LB_D4_VLINE_INTERRUPT_MASK,
+	.hpd = DISP_INTERRUPT_STATUS_CONTINUE3__DC_HPD4_INTERRUPT_MASK
+}, {
+	.reg = DISP_INTERRUPT_STATUS_CONTINUE4,
+	.vblank = DISP_INTERRUPT_STATUS_CONTINUE4__LB_D5_VBLANK_INTERRUPT_MASK,
+	.vline = DISP_INTERRUPT_STATUS_CONTINUE4__LB_D5_VLINE_INTERRUPT_MASK,
+	.hpd = DISP_INTERRUPT_STATUS_CONTINUE4__DC_HPD5_INTERRUPT_MASK
+}, {
+	.reg = DISP_INTERRUPT_STATUS_CONTINUE5,
+	.vblank = DISP_INTERRUPT_STATUS_CONTINUE5__LB_D6_VBLANK_INTERRUPT_MASK,
+	.vline = DISP_INTERRUPT_STATUS_CONTINUE5__LB_D6_VLINE_INTERRUPT_MASK,
+	.hpd = DISP_INTERRUPT_STATUS_CONTINUE5__DC_HPD6_INTERRUPT_MASK
+} };
+
+static const uint32_t hpd_int_control_offsets[6] = {
+	DC_HPD1_INT_CONTROL,
+	DC_HPD2_INT_CONTROL,
+	DC_HPD3_INT_CONTROL,
+	DC_HPD4_INT_CONTROL,
+	DC_HPD5_INT_CONTROL,
+	DC_HPD6_INT_CONTROL,
+};
+
+static u32 dce_v6_0_audio_endpt_rreg(struct amdgpu_device *adev,
+				     u32 block_offset, u32 reg)
+{
+	DRM_INFO("xxxx: dce_v6_0_audio_endpt_rreg ----no impl!!!!\n");
+	return 0;
+}
+
+static void dce_v6_0_audio_endpt_wreg(struct amdgpu_device *adev,
+				      u32 block_offset, u32 reg, u32 v)
+{
+	DRM_INFO("xxxx: dce_v6_0_audio_endpt_wreg ----no impl!!!!\n");
+}
+
+static bool dce_v6_0_is_in_vblank(struct amdgpu_device *adev, int crtc)
+{
+	if (RREG32(EVERGREEN_CRTC_STATUS + crtc_offsets[crtc]) & EVERGREEN_CRTC_V_BLANK)
+		return true;
+	else
+		return false;
+}
+
+static bool dce_v6_0_is_counter_moving(struct amdgpu_device *adev, int crtc)
+{
+	u32 pos1, pos2;
+
+	pos1 = RREG32(EVERGREEN_CRTC_STATUS_POSITION + crtc_offsets[crtc]);
+	pos2 = RREG32(EVERGREEN_CRTC_STATUS_POSITION + crtc_offsets[crtc]);
+
+	if (pos1 != pos2)
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce_v6_0_wait_for_vblank - vblank wait asic callback.
+ *
+ * @crtc: crtc to wait for vblank on
+ *
+ * Wait for vblank on the requested crtc (evergreen+).
+ */
+static void dce_v6_0_vblank_wait(struct amdgpu_device *adev, int crtc)
+{
+	unsigned i = 100;
+
+	if (crtc >= adev->mode_info.num_crtc)
+		return;
+
+	if (!(RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[crtc]) & EVERGREEN_CRTC_MASTER_EN))
+		return;
+
+	/* depending on when we hit vblank, we may be close to active; if so,
+	 * wait for another frame.
+	 */
+	while (dce_v6_0_is_in_vblank(adev, crtc)) {
+		if (i++ == 100) {
+			i = 0;
+			if (!dce_v6_0_is_counter_moving(adev, crtc))
+				break;
+		}
+	}
+
+	while (!dce_v6_0_is_in_vblank(adev, crtc)) {
+		if (i++ == 100) {
+			i = 0;
+			if (!dce_v6_0_is_counter_moving(adev, crtc))
+				break;
+		}
+	}
+}
+
+static u32 dce_v6_0_vblank_get_counter(struct amdgpu_device *adev, int crtc)
+{
+	if (crtc >= adev->mode_info.num_crtc)
+		return 0;
+	else
+		return RREG32(CRTC_STATUS_FRAME_COUNT + crtc_offsets[crtc]);
+}
+
+static void dce_v6_0_pageflip_interrupt_init(struct amdgpu_device *adev)
+{
+	unsigned i;
+
+	/* Enable pflip interrupts */
+	for (i = 0; i < adev->mode_info.num_crtc; i++)
+		amdgpu_irq_get(adev, &adev->pageflip_irq, i);
+}
+
+static void dce_v6_0_pageflip_interrupt_fini(struct amdgpu_device *adev)
+{
+	unsigned i;
+
+	/* Disable pflip interrupts */
+	for (i = 0; i < adev->mode_info.num_crtc; i++)
+		amdgpu_irq_put(adev, &adev->pageflip_irq, i);
+}
+
+/**
+ * dce_v6_0_page_flip - pageflip callback.
+ *
+ * @adev: amdgpu_device pointer
+ * @crtc_id: crtc to cleanup pageflip on
+ * @crtc_base: new address of the crtc (GPU MC address)
+ *
+ * Does the actual pageflip (evergreen+).
+ * During vblank we take the crtc lock and wait for the update_pending
+ * bit to go high, when it does, we release the lock, and allow the
+ * double buffered update to take place.
+ * Returns the current update pending status.
+ */
+static void dce_v6_0_page_flip(struct amdgpu_device *adev,
+			       int crtc_id, u64 crtc_base, bool async)
+{
+	struct amdgpu_crtc *amdgpu_crtc = adev->mode_info.crtcs[crtc_id];
+
+	/* flip at hsync for async, default is vsync */
+	WREG32(EVERGREEN_GRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset, async ?
+	       EVERGREEN_GRPH_SURFACE_UPDATE_H_RETRACE_EN : 0);
+	/* update the scanout addresses */
+	WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
+	       upper_32_bits(crtc_base));
+	WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
+	       (u32)crtc_base);
+
+	/* post the write */
+	RREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset);
+}
+
+static int dce_v6_0_crtc_get_scanoutpos(struct amdgpu_device *adev, int crtc,
+					u32 *vbl, u32 *position)
+{
+	if ((crtc < 0) || (crtc >= adev->mode_info.num_crtc))
+		return -EINVAL;
+	*vbl = RREG32(EVERGREEN_CRTC_V_BLANK_START_END + crtc_offsets[crtc]);
+	*position = RREG32(EVERGREEN_CRTC_STATUS_POSITION + crtc_offsets[crtc]);
+
+	return 0;
+
+}
+
+/**
+ * dce_v6_0_hpd_sense - hpd sense callback.
+ *
+ * @adev: amdgpu_device pointer
+ * @hpd: hpd (hotplug detect) pin
+ *
+ * Checks if a digital monitor is connected (evergreen+).
+ * Returns true if connected, false if not connected.
+ */
+static bool dce_v6_0_hpd_sense(struct amdgpu_device *adev,
+			       enum amdgpu_hpd_id hpd)
+{
+	bool connected = false;
+
+	switch (hpd) {
+	case AMDGPU_HPD_1:
+		if (RREG32(DC_HPD1_INT_STATUS) & DC_HPDx_SENSE)
+			connected = true;
+		break;
+	case AMDGPU_HPD_2:
+		if (RREG32(DC_HPD2_INT_STATUS) & DC_HPDx_SENSE)
+			connected = true;
+		break;
+	case AMDGPU_HPD_3:
+		if (RREG32(DC_HPD3_INT_STATUS) & DC_HPDx_SENSE)
+			connected = true;
+		break;
+	case AMDGPU_HPD_4:
+		if (RREG32(DC_HPD4_INT_STATUS) & DC_HPDx_SENSE)
+			connected = true;
+		break;
+	case AMDGPU_HPD_5:
+		if (RREG32(DC_HPD5_INT_STATUS) & DC_HPDx_SENSE)
+			connected = true;
+		break;
+	case AMDGPU_HPD_6:
+		if (RREG32(DC_HPD6_INT_STATUS) & DC_HPDx_SENSE)
+			connected = true;
+		break;
+	default:
+		break;
+	}
+
+	return connected;
+}
+
+/**
+ * dce_v6_0_hpd_set_polarity - hpd set polarity callback.
+ *
+ * @adev: amdgpu_device pointer
+ * @hpd: hpd (hotplug detect) pin
+ *
+ * Set the polarity of the hpd pin (evergreen+).
+ */
+static void dce_v6_0_hpd_set_polarity(struct amdgpu_device *adev,
+				      enum amdgpu_hpd_id hpd)
+{
+	u32 tmp;
+	bool connected = dce_v6_0_hpd_sense(adev, hpd);
+
+	switch (hpd) {
+	case AMDGPU_HPD_1:
+		tmp = RREG32(DC_HPD1_INT_CONTROL);
+		if (connected)
+			tmp &= ~DC_HPDx_INT_POLARITY;
+		else
+			tmp |= DC_HPDx_INT_POLARITY;
+		WREG32(DC_HPD1_INT_CONTROL, tmp);
+		break;
+	case AMDGPU_HPD_2:
+		tmp = RREG32(DC_HPD2_INT_CONTROL);
+		if (connected)
+			tmp &= ~DC_HPDx_INT_POLARITY;
+		else
+			tmp |= DC_HPDx_INT_POLARITY;
+		WREG32(DC_HPD2_INT_CONTROL, tmp);
+		break;
+	case AMDGPU_HPD_3:
+		tmp = RREG32(DC_HPD3_INT_CONTROL);
+		if (connected)
+			tmp &= ~DC_HPDx_INT_POLARITY;
+		else
+			tmp |= DC_HPDx_INT_POLARITY;
+		WREG32(DC_HPD3_INT_CONTROL, tmp);
+		break;
+	case AMDGPU_HPD_4:
+		tmp = RREG32(DC_HPD4_INT_CONTROL);
+		if (connected)
+			tmp &= ~DC_HPDx_INT_POLARITY;
+		else
+			tmp |= DC_HPDx_INT_POLARITY;
+		WREG32(DC_HPD4_INT_CONTROL, tmp);
+		break;
+	case AMDGPU_HPD_5:
+		tmp = RREG32(DC_HPD5_INT_CONTROL);
+		if (connected)
+			tmp &= ~DC_HPDx_INT_POLARITY;
+		else
+			tmp |= DC_HPDx_INT_POLARITY;
+		WREG32(DC_HPD5_INT_CONTROL, tmp);
+			break;
+	case AMDGPU_HPD_6:
+		tmp = RREG32(DC_HPD6_INT_CONTROL);
+		if (connected)
+			tmp &= ~DC_HPDx_INT_POLARITY;
+		else
+			tmp |= DC_HPDx_INT_POLARITY;
+		WREG32(DC_HPD6_INT_CONTROL, tmp);
+		break;
+	default:
+		break;
+	}
+}
+
+/**
+ * dce_v6_0_hpd_init - hpd setup callback.
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Setup the hpd pins used by the card (evergreen+).
+ * Enable the pin, set the polarity, and enable the hpd interrupts.
+ */
+static void dce_v6_0_hpd_init(struct amdgpu_device *adev)
+{
+	struct drm_device *dev = adev->ddev;
+	struct drm_connector *connector;
+	u32 tmp = DC_HPDx_CONNECTION_TIMER(0x9c4) |
+		DC_HPDx_RX_INT_TIMER(0xfa) | DC_HPDx_EN;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
+
+		switch (amdgpu_connector->hpd.hpd) {
+		case AMDGPU_HPD_1:
+			WREG32(DC_HPD1_CONTROL, tmp);
+			break;
+		case AMDGPU_HPD_2:
+			WREG32(DC_HPD2_CONTROL, tmp);
+			break;
+		case AMDGPU_HPD_3:
+			WREG32(DC_HPD3_CONTROL, tmp);
+			break;
+		case AMDGPU_HPD_4:
+			WREG32(DC_HPD4_CONTROL, tmp);
+			break;
+		case AMDGPU_HPD_5:
+			WREG32(DC_HPD5_CONTROL, tmp);
+			break;
+		case AMDGPU_HPD_6:
+			WREG32(DC_HPD6_CONTROL, tmp);
+			break;
+		default:
+			break;
+		}
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
+		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
+			/* don't try to enable hpd on eDP or LVDS avoid breaking the
+			 * aux dp channel on imac and help (but not completely fix)
+			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
+			 * also avoid interrupt storms during dpms.
+			 */
+			u32 dc_hpd_int_cntl_reg, dc_hpd_int_cntl;
+
+			switch (amdgpu_connector->hpd.hpd) {
+			case AMDGPU_HPD_1:
+				dc_hpd_int_cntl_reg = DC_HPD1_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_2:
+				dc_hpd_int_cntl_reg = DC_HPD2_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_3:
+				dc_hpd_int_cntl_reg = DC_HPD3_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_4:
+				dc_hpd_int_cntl_reg = DC_HPD4_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_5:
+				dc_hpd_int_cntl_reg = DC_HPD5_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_6:
+				dc_hpd_int_cntl_reg = DC_HPD6_INT_CONTROL;
+				break;
+			default:
+				continue;
+			}
+
+			dc_hpd_int_cntl = RREG32(dc_hpd_int_cntl_reg);
+			dc_hpd_int_cntl &= ~DC_HPDx_INT_EN;
+			WREG32(dc_hpd_int_cntl_reg, dc_hpd_int_cntl);
+			continue;
+		}
+
+		dce_v6_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
+		amdgpu_irq_get(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
+	}
+
+}
+
+/**
+ * dce_v6_0_hpd_fini - hpd tear down callback.
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Tear down the hpd pins used by the card (evergreen+).
+ * Disable the hpd interrupts.
+ */
+static void dce_v6_0_hpd_fini(struct amdgpu_device *adev)
+{
+	struct drm_device *dev = adev->ddev;
+	struct drm_connector *connector;
+
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
+
+		switch (amdgpu_connector->hpd.hpd) {
+		case AMDGPU_HPD_1:
+			WREG32(DC_HPD1_CONTROL, 0);
+			break;
+		case AMDGPU_HPD_2:
+			WREG32(DC_HPD2_CONTROL, 0);
+			break;
+		case AMDGPU_HPD_3:
+			WREG32(DC_HPD3_CONTROL, 0);
+			break;
+		case AMDGPU_HPD_4:
+			WREG32(DC_HPD4_CONTROL, 0);
+			break;
+		case AMDGPU_HPD_5:
+			WREG32(DC_HPD5_CONTROL, 0);
+			break;
+		case AMDGPU_HPD_6:
+			WREG32(DC_HPD6_CONTROL, 0);
+			break;
+		default:
+			break;
+		}
+		amdgpu_irq_put(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
+	}
+}
+
+static u32 dce_v6_0_hpd_get_gpio_reg(struct amdgpu_device *adev)
+{
+	return SI_DC_GPIO_HPD_A;
+}
+
+static bool dce_v6_0_is_display_hung(struct amdgpu_device *adev)
+{
+	DRM_INFO("xxxx: dce_v6_0_is_display_hung ----no imp!!!!!\n");
+
+	return true;
+}
+
+static u32 evergreen_get_vblank_counter(struct amdgpu_device* adev, int crtc)
+{
+	if (crtc >= adev->mode_info.num_crtc)
+		return 0;
+	else
+		return RREG32(CRTC_STATUS_FRAME_COUNT + crtc_offsets[crtc]);
+}
+
+static void dce_v6_0_stop_mc_access(struct amdgpu_device *adev,
+				    struct amdgpu_mode_mc_save *save)
+{
+	u32 crtc_enabled, tmp, frame_count;
+	int i, j;
+
+	save->vga_render_control = RREG32(VGA_RENDER_CONTROL);
+	save->vga_hdp_control = RREG32(VGA_HDP_CONTROL);
+
+	/* disable VGA render */
+	WREG32(VGA_RENDER_CONTROL, 0);
+
+	/* blank the display controllers */
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		crtc_enabled = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]) & EVERGREEN_CRTC_MASTER_EN;
+		if (crtc_enabled) {
+			save->crtc_enabled[i] = true;
+			tmp = RREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i]);
+
+			if (!(tmp & EVERGREEN_CRTC_BLANK_DATA_EN)) {
+				dce_v6_0_vblank_wait(adev, i);
+				WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+				tmp |= EVERGREEN_CRTC_BLANK_DATA_EN;
+				WREG32(EVERGREEN_CRTC_BLANK_CONTROL + crtc_offsets[i], tmp);
+				WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
+			}
+			/* wait for the next frame */
+			frame_count = evergreen_get_vblank_counter(adev, i);
+			for (j = 0; j < adev->usec_timeout; j++) {
+				if (evergreen_get_vblank_counter(adev, i) != frame_count)
+					break;
+				udelay(1);
+			}
+
+			/* XXX this is a hack to avoid strange behavior with EFI on certain systems */
+			WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+			tmp = RREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i]);
+			tmp &= ~EVERGREEN_CRTC_MASTER_EN;
+			WREG32(EVERGREEN_CRTC_CONTROL + crtc_offsets[i], tmp);
+			WREG32(EVERGREEN_CRTC_UPDATE_LOCK + crtc_offsets[i], 0);
+			save->crtc_enabled[i] = false;
+			/* ***** */
+		} else {
+			save->crtc_enabled[i] = false;
+		}
+	}
+}
+
+static void dce_v6_0_resume_mc_access(struct amdgpu_device *adev,
+				      struct amdgpu_mode_mc_save *save)
+{
+	u32 tmp;
+	int i, j;
+
+	/* update crtc base addresses */
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + crtc_offsets[i],
+		       upper_32_bits(adev->mc.vram_start));
+		WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + crtc_offsets[i],
+		       upper_32_bits(adev->mc.vram_start));
+		WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + crtc_offsets[i],
+		       (u32)adev->mc.vram_start);
+		WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + crtc_offsets[i],
+		       (u32)adev->mc.vram_start);
+	}
+
+	WREG32(EVERGREEN_VGA_MEMORY_BASE_ADDRESS_HIGH, upper_32_bits(adev->mc.vram_start));
+	WREG32(EVERGREEN_VGA_MEMORY_BASE_ADDRESS, (u32)adev->mc.vram_start);
+
+	/* unlock regs and wait for update */
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		if (save->crtc_enabled[i]) {
+			tmp = RREG32(EVERGREEN_MASTER_UPDATE_MODE + crtc_offsets[i]);
+			if ((tmp & 0x7) != 3) {
+				tmp &= ~0x7;
+				tmp |= 0x3;
+				WREG32(EVERGREEN_MASTER_UPDATE_MODE + crtc_offsets[i], tmp);
+			}
+			tmp = RREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i]);
+			if (tmp & EVERGREEN_GRPH_UPDATE_LOCK) {
+				tmp &= ~EVERGREEN_GRPH_UPDATE_LOCK;
+				WREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i], tmp);
+			}
+			tmp = RREG32(EVERGREEN_MASTER_UPDATE_LOCK + crtc_offsets[i]);
+			if (tmp & 1) {
+				tmp &= ~1;
+				WREG32(EVERGREEN_MASTER_UPDATE_LOCK + crtc_offsets[i], tmp);
+			}
+			for (j = 0; j < adev->usec_timeout; j++) {
+				tmp = RREG32(EVERGREEN_GRPH_UPDATE + crtc_offsets[i]);
+				if ((tmp & EVERGREEN_GRPH_SURFACE_UPDATE_PENDING) == 0)
+					break;
+				udelay(1);
+			}
+		}
+	}
+
+	/* Unlock vga access */
+	WREG32(VGA_HDP_CONTROL, save->vga_hdp_control);
+	mdelay(1);
+	WREG32(VGA_RENDER_CONTROL, save->vga_render_control);
+
+}
+
+static void dce_v6_0_set_vga_render_state(struct amdgpu_device *adev,
+					  bool render)
+{
+	if (!render) 
+		WREG32(R_000300_VGA_RENDER_CONTROL,
+			RREG32(R_000300_VGA_RENDER_CONTROL) & C_000300_VGA_VSTATUS_CNTL);
+
+}
+
+static void dce_v6_0_program_fmt(struct drm_encoder *encoder)
+{
+
+	struct drm_device *dev = encoder->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+	struct drm_connector *connector = amdgpu_get_connector_for_encoder(encoder);
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(encoder->crtc);
+	int bpc = 0;
+	u32 tmp = 0;
+	enum amdgpu_connector_dither dither = AMDGPU_FMT_DITHER_DISABLE;
+
+	if (connector) {
+		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
+		bpc = amdgpu_connector_get_monitor_bpc(connector);
+		dither = amdgpu_connector->dither;
+	}
+
+	/* LVDS FMT is set up by atom */
+	if (amdgpu_encoder->devices & ATOM_DEVICE_LCD_SUPPORT)
+		return;
+
+	if (bpc == 0)
+		return;
+
+
+	switch (bpc) {
+	case 6:
+		if (dither == AMDGPU_FMT_DITHER_ENABLE)
+			/* XXX sort out optimal dither settings */
+			tmp |= (FMT_FRAME_RANDOM_ENABLE | FMT_HIGHPASS_RANDOM_ENABLE |
+				FMT_SPATIAL_DITHER_EN);
+		else
+			tmp |= FMT_TRUNCATE_EN;
+		break;
+	case 8:
+		if (dither == AMDGPU_FMT_DITHER_ENABLE)
+			/* XXX sort out optimal dither settings */
+			tmp |= (FMT_FRAME_RANDOM_ENABLE | FMT_HIGHPASS_RANDOM_ENABLE |
+				FMT_RGB_RANDOM_ENABLE |
+				FMT_SPATIAL_DITHER_EN | FMT_SPATIAL_DITHER_DEPTH);
+		else
+			tmp |= (FMT_TRUNCATE_EN | FMT_TRUNCATE_DEPTH);
+		break;
+	case 10:
+	default:
+		/* not needed */
+		break;
+	}
+
+	WREG32(FMT_BIT_DEPTH_CONTROL + amdgpu_crtc->crtc_offset, tmp);
+}
+
+/**
+ * cik_get_number_of_dram_channels - get the number of dram channels
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Look up the number of video ram channels (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the number of dram channels
+ */
+static u32 si_get_number_of_dram_channels(struct amdgpu_device *adev)
+{
+	u32 tmp = RREG32(MC_SHARED_CHMAP);
+
+	switch ((tmp & MC_SHARED_CHMAP__NOOFCHAN_MASK) >> MC_SHARED_CHMAP__NOOFCHAN__SHIFT) {
+	case 0:
+	default:
+		return 1;
+	case 1:
+		return 2;
+	case 2:
+		return 4;
+	case 3:
+		return 8;
+	case 4:
+		return 3;
+	case 5:
+		return 6;
+	case 6:
+		return 10;
+	case 7:
+		return 12;
+	case 8:
+		return 16;
+	}
+}
+
+struct dce6_wm_params {
+	u32 dram_channels; /* number of dram channels */
+	u32 yclk;          /* bandwidth per dram data pin in kHz */
+	u32 sclk;          /* engine clock in kHz */
+	u32 disp_clk;      /* display clock in kHz */
+	u32 src_width;     /* viewport width */
+	u32 active_time;   /* active display time in ns */
+	u32 blank_time;    /* blank time in ns */
+	bool interlaced;    /* mode is interlaced */
+	fixed20_12 vsc;    /* vertical scale ratio */
+	u32 num_heads;     /* number of active crtcs */
+	u32 bytes_per_pixel; /* bytes per pixel display + overlay */
+	u32 lb_size;       /* line buffer allocated to pipe */
+	u32 vtaps;         /* vertical scaler taps */
+};
+
+/**
+ * dce_v6_0_dram_bandwidth - get the dram bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the raw dram bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dram bandwidth in MBytes/s
+ */
+static u32 dce_v6_0_dram_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate raw DRAM Bandwidth */
+	fixed20_12 dram_efficiency; /* 0.7 */
+	fixed20_12 yclk, dram_channels, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	yclk.full = dfixed_const(wm->yclk);
+	yclk.full = dfixed_div(yclk, a);
+	dram_channels.full = dfixed_const(wm->dram_channels * 4);
+	a.full = dfixed_const(10);
+	dram_efficiency.full = dfixed_const(7);
+	dram_efficiency.full = dfixed_div(dram_efficiency, a);
+	bandwidth.full = dfixed_mul(dram_channels, yclk);
+	bandwidth.full = dfixed_mul(bandwidth, dram_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce_v6_0_dram_bandwidth_for_display - get the dram bandwidth for display
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the dram bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dram bandwidth for display in MBytes/s
+ */
+static u32 dce_v6_0_dram_bandwidth_for_display(struct dce6_wm_params *wm)
+{
+	/* Calculate DRAM Bandwidth and the part allocated to display. */
+	fixed20_12 disp_dram_allocation; /* 0.3 to 0.7 */
+	fixed20_12 yclk, dram_channels, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	yclk.full = dfixed_const(wm->yclk);
+	yclk.full = dfixed_div(yclk, a);
+	dram_channels.full = dfixed_const(wm->dram_channels * 4);
+	a.full = dfixed_const(10);
+	disp_dram_allocation.full = dfixed_const(3); /* XXX worse case value 0.3 */
+	disp_dram_allocation.full = dfixed_div(disp_dram_allocation, a);
+	bandwidth.full = dfixed_mul(dram_channels, yclk);
+	bandwidth.full = dfixed_mul(bandwidth, disp_dram_allocation);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce_v6_0_data_return_bandwidth - get the data return bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the data return bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the data return bandwidth in MBytes/s
+ */
+static u32 dce_v6_0_data_return_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate the display Data return Bandwidth */
+	fixed20_12 return_efficiency; /* 0.8 */
+	fixed20_12 sclk, bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	sclk.full = dfixed_const(wm->sclk);
+	sclk.full = dfixed_div(sclk, a);
+	a.full = dfixed_const(10);
+	return_efficiency.full = dfixed_const(8);
+	return_efficiency.full = dfixed_div(return_efficiency, a);
+	a.full = dfixed_const(32);
+	bandwidth.full = dfixed_mul(a, sclk);
+	bandwidth.full = dfixed_mul(bandwidth, return_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce_v6_0_dmif_request_bandwidth - get the dmif bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the dmif bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the dmif bandwidth in MBytes/s
+ */
+static u32 dce_v6_0_dmif_request_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate the DMIF Request Bandwidth */
+	fixed20_12 disp_clk_request_efficiency; /* 0.8 */
+	fixed20_12 disp_clk, bandwidth;
+	fixed20_12 a, b;
+
+	a.full = dfixed_const(1000);
+	disp_clk.full = dfixed_const(wm->disp_clk);
+	disp_clk.full = dfixed_div(disp_clk, a);
+	a.full = dfixed_const(32);
+	b.full = dfixed_mul(a, disp_clk);
+
+	a.full = dfixed_const(10);
+	disp_clk_request_efficiency.full = dfixed_const(8);
+	disp_clk_request_efficiency.full = dfixed_div(disp_clk_request_efficiency, a);
+
+	bandwidth.full = dfixed_mul(b, disp_clk_request_efficiency);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce_v6_0_available_bandwidth - get the min available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the min available bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the min available bandwidth in MBytes/s
+ */
+static u32 dce_v6_0_available_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate the Available bandwidth. Display can use this temporarily but not in average. */
+	u32 dram_bandwidth = dce_v6_0_dram_bandwidth(wm);
+	u32 data_return_bandwidth = dce_v6_0_data_return_bandwidth(wm);
+	u32 dmif_req_bandwidth = dce_v6_0_dmif_request_bandwidth(wm);
+
+	return min(dram_bandwidth, min(data_return_bandwidth, dmif_req_bandwidth));
+}
+
+/**
+ * dce_v6_0_average_bandwidth - get the average available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the average available bandwidth used for display (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the average available bandwidth in MBytes/s
+ */
+static u32 dce_v6_0_average_bandwidth(struct dce6_wm_params *wm)
+{
+	/* Calculate the display mode Average Bandwidth
+	 * DisplayMode should contain the source and destination dimensions,
+	 * timing, etc.
+	 */
+	fixed20_12 bpp;
+	fixed20_12 line_time;
+	fixed20_12 src_width;
+	fixed20_12 bandwidth;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1000);
+	line_time.full = dfixed_const(wm->active_time + wm->blank_time);
+	line_time.full = dfixed_div(line_time, a);
+	bpp.full = dfixed_const(wm->bytes_per_pixel);
+	src_width.full = dfixed_const(wm->src_width);
+	bandwidth.full = dfixed_mul(src_width, bpp);
+	bandwidth.full = dfixed_mul(bandwidth, wm->vsc);
+	bandwidth.full = dfixed_div(bandwidth, line_time);
+
+	return dfixed_trunc(bandwidth);
+}
+
+/**
+ * dce_v6_0_latency_watermark - get the latency watermark
+ *
+ * @wm: watermark calculation data
+ *
+ * Calculate the latency watermark (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns the latency watermark in ns
+ */
+static u32 dce_v6_0_latency_watermark(struct dce6_wm_params *wm)
+{
+	/* First calculate the latency in ns */
+	u32 mc_latency = 2000; /* 2000 ns. */
+	u32 available_bandwidth = dce_v6_0_available_bandwidth(wm);
+	u32 worst_chunk_return_time = (512 * 8 * 1000) / available_bandwidth;
+	u32 cursor_line_pair_return_time = (128 * 4 * 1000) / available_bandwidth;
+	u32 dc_latency = 40000000 / wm->disp_clk; /* dc pipe latency */
+	u32 other_heads_data_return_time = ((wm->num_heads + 1) * worst_chunk_return_time) +
+		(wm->num_heads * cursor_line_pair_return_time);
+	u32 latency = mc_latency + other_heads_data_return_time + dc_latency;
+	u32 max_src_lines_per_dst_line, lb_fill_bw, line_fill_time;
+	u32 tmp, dmif_size = 12288;
+	fixed20_12 a, b, c;
+
+	if (wm->num_heads == 0)
+		return 0;
+
+	a.full = dfixed_const(2);
+	b.full = dfixed_const(1);
+	if ((wm->vsc.full > a.full) ||
+	    ((wm->vsc.full > b.full) && (wm->vtaps >= 3)) ||
+	    (wm->vtaps >= 5) ||
+	    ((wm->vsc.full >= a.full) && wm->interlaced))
+		max_src_lines_per_dst_line = 4;
+	else
+		max_src_lines_per_dst_line = 2;
+
+	a.full = dfixed_const(available_bandwidth);
+	b.full = dfixed_const(wm->num_heads);
+	a.full = dfixed_div(a, b);
+
+	b.full = dfixed_const(mc_latency + 512);
+	c.full = dfixed_const(wm->disp_clk);
+	b.full = dfixed_div(b, c);
+
+	c.full = dfixed_const(dmif_size);
+	b.full = dfixed_div(c, b);
+
+	tmp = min(dfixed_trunc(a), dfixed_trunc(b));
+
+	b.full = dfixed_const(1000);
+	c.full = dfixed_const(wm->disp_clk);
+	b.full = dfixed_div(c, b);
+	c.full = dfixed_const(wm->bytes_per_pixel);
+	b.full = dfixed_mul(b, c);
+
+	lb_fill_bw = min(tmp, dfixed_trunc(b));
+
+	a.full = dfixed_const(max_src_lines_per_dst_line * wm->src_width * wm->bytes_per_pixel);
+	b.full = dfixed_const(1000);
+	c.full = dfixed_const(lb_fill_bw);
+	b.full = dfixed_div(c, b);
+	a.full = dfixed_div(a, b);
+	line_fill_time = dfixed_trunc(a);
+
+	if (line_fill_time < wm->active_time)
+		return latency;
+	else
+		return latency + (line_fill_time - wm->active_time);
+
+}
+
+/**
+ * dce_v6_0_average_bandwidth_vs_dram_bandwidth_for_display - check
+ * average and available dram bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Check if the display average bandwidth fits in the display
+ * dram bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce_v6_0_average_bandwidth_vs_dram_bandwidth_for_display(struct dce6_wm_params *wm)
+{
+	if (dce_v6_0_average_bandwidth(wm) <=
+	    (dce_v6_0_dram_bandwidth_for_display(wm) / wm->num_heads))
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce_v6_0_average_bandwidth_vs_available_bandwidth - check
+ * average and available bandwidth
+ *
+ * @wm: watermark calculation data
+ *
+ * Check if the display average bandwidth fits in the display
+ * available bandwidth (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce_v6_0_average_bandwidth_vs_available_bandwidth(struct dce6_wm_params *wm)
+{
+	if (dce_v6_0_average_bandwidth(wm) <=
+	    (dce_v6_0_available_bandwidth(wm) / wm->num_heads))
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce_v6_0_check_latency_hiding - check latency hiding
+ *
+ * @wm: watermark calculation data
+ *
+ * Check latency hiding (CIK).
+ * Used for display watermark bandwidth calculations
+ * Returns true if the display fits, false if not.
+ */
+static bool dce_v6_0_check_latency_hiding(struct dce6_wm_params *wm)
+{
+	u32 lb_partitions = wm->lb_size / wm->src_width;
+	u32 line_time = wm->active_time + wm->blank_time;
+	u32 latency_tolerant_lines;
+	u32 latency_hiding;
+	fixed20_12 a;
+
+	a.full = dfixed_const(1);
+	if (wm->vsc.full > a.full)
+		latency_tolerant_lines = 1;
+	else {
+		if (lb_partitions <= (wm->vtaps + 1))
+			latency_tolerant_lines = 1;
+		else
+			latency_tolerant_lines = 2;
+	}
+
+	latency_hiding = (latency_tolerant_lines * line_time + wm->blank_time);
+
+	if (dce_v6_0_latency_watermark(wm) <= latency_hiding)
+		return true;
+	else
+		return false;
+}
+
+/**
+ * dce_v6_0_program_watermarks - program display watermarks
+ *
+ * @adev: amdgpu_device pointer
+ * @amdgpu_crtc: the selected display controller
+ * @lb_size: line buffer size
+ * @num_heads: number of display controllers in use
+ *
+ * Calculate and program the display watermarks for the
+ * selected display controller (CIK).
+ */
+static void dce_v6_0_program_watermarks(struct amdgpu_device *adev,
+					struct amdgpu_crtc *amdgpu_crtc,
+					u32 lb_size, u32 num_heads)
+{
+	struct drm_display_mode *mode = &amdgpu_crtc->base.mode;
+	struct dce6_wm_params wm_low, wm_high;
+	u32 dram_channels;
+	u32 pixel_period;
+	u32 line_time = 0;
+	u32 latency_watermark_a = 0, latency_watermark_b = 0;
+	u32 priority_a_mark = 0, priority_b_mark = 0;
+	u32 priority_a_cnt = PRIORITY_OFF;
+	u32 priority_b_cnt = PRIORITY_OFF;
+	u32 tmp, arb_control3;
+	fixed20_12 a, b, c;
+
+	if (amdgpu_crtc->base.enabled && num_heads && mode) {
+		pixel_period = 1000000 / (u32)mode->clock;
+		line_time = min((u32)mode->crtc_htotal * pixel_period, (u32)65535);
+		priority_a_cnt = 0;
+		priority_b_cnt = 0;
+
+		dram_channels = si_get_number_of_dram_channels(adev);
+
+		/* watermark for high clocks */
+		if (adev->pm.dpm_enabled) {
+			wm_high.yclk =
+				amdgpu_dpm_get_mclk(adev, false) * 10;
+			wm_high.sclk =
+				amdgpu_dpm_get_sclk(adev, false) * 10;
+		} else {
+			wm_high.yclk = adev->pm.current_mclk * 10;
+			wm_high.sclk = adev->pm.current_sclk * 10;
+		}
+
+		wm_high.disp_clk = mode->clock;
+		wm_high.src_width = mode->crtc_hdisplay;
+		wm_high.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_high.blank_time = line_time - wm_high.active_time;
+		wm_high.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm_high.interlaced = true;
+		wm_high.vsc = amdgpu_crtc->vsc;
+		wm_high.vtaps = 1;
+		if (amdgpu_crtc->rmx_type != RMX_OFF)
+			wm_high.vtaps = 2;
+		wm_high.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_high.lb_size = lb_size;
+		wm_high.dram_channels = dram_channels;
+		wm_high.num_heads = num_heads;
+
+		if (adev->pm.dpm_enabled) {
+		/* watermark for low clocks */
+			wm_low.yclk =
+				amdgpu_dpm_get_mclk(adev, true) * 10;
+			wm_low.sclk =
+				amdgpu_dpm_get_sclk(adev, true) * 10;
+		} else {
+			wm_low.yclk = adev->pm.current_mclk * 10;
+			wm_low.sclk = adev->pm.current_sclk * 10;
+		}
+
+		wm_low.disp_clk = mode->clock;
+		wm_low.src_width = mode->crtc_hdisplay;
+		wm_low.active_time = mode->crtc_hdisplay * pixel_period;
+		wm_low.blank_time = line_time - wm_low.active_time;
+		wm_low.interlaced = false;
+		if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+			wm_low.interlaced = true;
+		wm_low.vsc = amdgpu_crtc->vsc;
+		wm_low.vtaps = 1;
+		if (amdgpu_crtc->rmx_type != RMX_OFF)
+			wm_low.vtaps = 2;
+		wm_low.bytes_per_pixel = 4; /* XXX: get this from fb config */
+		wm_low.lb_size = lb_size;
+		wm_low.dram_channels = dram_channels;
+		wm_low.num_heads = num_heads;
+
+		/* set for high clocks */
+		latency_watermark_a = min(dce_v6_0_latency_watermark(&wm_high), (u32)65535);
+		/* set for low clocks */
+		latency_watermark_b = min(dce_v6_0_latency_watermark(&wm_low), (u32)65535);
+
+		/* possibly force display priority to high */
+		/* should really do this at mode validation time... */
+		if (!dce_v6_0_average_bandwidth_vs_dram_bandwidth_for_display(&wm_high) ||
+		    !dce_v6_0_average_bandwidth_vs_available_bandwidth(&wm_high) ||
+		    !dce_v6_0_check_latency_hiding(&wm_high) ||
+		    (adev->mode_info.disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority to high\n");
+			priority_a_cnt |= PRIORITY_ALWAYS_ON;
+			priority_b_cnt |= PRIORITY_ALWAYS_ON;
+		}
+		if (!dce_v6_0_average_bandwidth_vs_dram_bandwidth_for_display(&wm_low) ||
+		    !dce_v6_0_average_bandwidth_vs_available_bandwidth(&wm_low) ||
+		    !dce_v6_0_check_latency_hiding(&wm_low) ||
+		    (adev->mode_info.disp_priority == 2)) {
+			DRM_DEBUG_KMS("force priority to high\n");
+			priority_a_cnt |= PRIORITY_ALWAYS_ON;
+			priority_b_cnt |= PRIORITY_ALWAYS_ON;
+		}
+
+		a.full = dfixed_const(1000);
+		b.full = dfixed_const(mode->clock);
+		b.full = dfixed_div(b, a);
+		c.full = dfixed_const(latency_watermark_a);
+		c.full = dfixed_mul(c, b);
+		c.full = dfixed_mul(c, amdgpu_crtc->hsc);
+		c.full = dfixed_div(c, a);
+		a.full = dfixed_const(16);
+		c.full = dfixed_div(c, a);
+		priority_a_mark = dfixed_trunc(c);
+		priority_a_cnt |= priority_a_mark & PRIORITY_MARK_MASK;
+
+		a.full = dfixed_const(1000);
+		b.full = dfixed_const(mode->clock);
+		b.full = dfixed_div(b, a);
+		c.full = dfixed_const(latency_watermark_b);
+		c.full = dfixed_mul(c, b);
+		c.full = dfixed_mul(c, amdgpu_crtc->hsc);
+		c.full = dfixed_div(c, a);
+		a.full = dfixed_const(16);
+		c.full = dfixed_div(c, a);
+		priority_b_mark = dfixed_trunc(c);
+		priority_b_cnt |= priority_b_mark & PRIORITY_MARK_MASK;
+	}
+
+	/* select wm A */
+	arb_control3 = RREG32(DPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset);
+	tmp = arb_control3;
+	tmp &= ~LATENCY_WATERMARK_MASK(3);
+	tmp |= LATENCY_WATERMARK_MASK(1);
+	WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset, tmp);
+	WREG32(DPG_PIPE_LATENCY_CONTROL + amdgpu_crtc->crtc_offset,
+	       (LATENCY_LOW_WATERMARK(latency_watermark_a) |
+		LATENCY_HIGH_WATERMARK(line_time)));
+	/* select wm B */
+	tmp = RREG32(DPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset);
+	tmp &= ~LATENCY_WATERMARK_MASK(3);
+	tmp |= LATENCY_WATERMARK_MASK(2);
+	WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset, tmp);
+	WREG32(DPG_PIPE_LATENCY_CONTROL + amdgpu_crtc->crtc_offset,
+	       (LATENCY_LOW_WATERMARK(latency_watermark_b) |
+		LATENCY_HIGH_WATERMARK(line_time)));
+	/* restore original selection */
+	WREG32(DPG_PIPE_ARBITRATION_CONTROL3 + amdgpu_crtc->crtc_offset, arb_control3);
+
+	/* write the priority marks */
+	WREG32(PRIORITY_A_CNT + amdgpu_crtc->crtc_offset, priority_a_cnt);
+	WREG32(PRIORITY_B_CNT + amdgpu_crtc->crtc_offset, priority_b_cnt);
+
+	/* save values for DPM */
+	amdgpu_crtc->line_time = line_time;
+	amdgpu_crtc->wm_high = latency_watermark_a;
+}
+
+/* watermark setup */
+static u32 dce_v6_0_line_buffer_adjust(struct amdgpu_device *adev,
+				   struct amdgpu_crtc *amdgpu_crtc,
+				   struct drm_display_mode *mode,
+				   struct drm_display_mode *other_mode)
+{
+	u32 tmp, buffer_alloc, i;
+	u32 pipe_offset = amdgpu_crtc->crtc_id * 0x8;
+	/*
+	 * Line Buffer Setup
+	 * There are 3 line buffers, each one shared by 2 display controllers.
+	 * DC_LB_MEMORY_SPLIT controls how that line buffer is shared between
+	 * the display controllers.  The paritioning is done via one of four
+	 * preset allocations specified in bits 21:20:
+	 *  0 - half lb
+	 *  2 - whole lb, other crtc must be disabled
+	 */
+	/* this can get tricky if we have two large displays on a paired group
+	 * of crtcs.  Ideally for multiple large displays we'd assign them to
+	 * non-linked crtcs for maximum line buffer allocation.
+	 */
+	if (amdgpu_crtc->base.enabled && mode) {
+		if (other_mode) {
+			tmp = 0; /* 1/2 */
+			buffer_alloc = 1;
+		} else {
+			tmp = 2; /* whole */
+			buffer_alloc = 2;
+		}
+	} else {
+		tmp = 0;
+		buffer_alloc = 0;
+	}
+
+	WREG32(DC_LB_MEMORY_SPLIT + amdgpu_crtc->crtc_offset,
+	       DC_LB_MEMORY_CONFIG(tmp));
+
+	WREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset,
+	       DMIF_BUFFERS_ALLOCATED(buffer_alloc));
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (RREG32(PIPE0_DMIF_BUFFER_CONTROL + pipe_offset) &
+		    DMIF_BUFFERS_ALLOCATED_COMPLETED)
+			break;
+		udelay(1);
+	}
+
+	if (amdgpu_crtc->base.enabled && mode) {
+		switch (tmp) {
+		case 0:
+		default:
+			return 4096 * 2;
+		case 2:
+			return 8192 * 2;
+		}
+	}
+
+	/* controller not enabled, so no lb used */
+	return 0;
+}
+
+
+/**
+ *
+ * dce_v6_0_bandwidth_update - program display watermarks
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Calculate and program the display watermarks and line
+ * buffer allocation (CIK).
+ */
+static void dce_v6_0_bandwidth_update(struct amdgpu_device *adev)
+{
+	struct drm_display_mode *mode0 = NULL;
+	struct drm_display_mode *mode1 = NULL;
+	u32 num_heads = 0, lb_size;
+	int i;
+
+	if (!adev->mode_info.mode_config_initialized)
+		return;
+
+	amdgpu_update_display_priority(adev);
+
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		if (adev->mode_info.crtcs[i]->base.enabled)
+			num_heads++;
+	}
+	for (i = 0; i < adev->mode_info.num_crtc; i += 2) {
+		mode0 = &adev->mode_info.crtcs[i]->base.mode;
+		mode1 = &adev->mode_info.crtcs[i+1]->base.mode;
+		lb_size = dce_v6_0_line_buffer_adjust(adev, adev->mode_info.crtcs[i], mode0, mode1);
+		dce_v6_0_program_watermarks(adev, adev->mode_info.crtcs[i], lb_size, num_heads);
+		lb_size = dce_v6_0_line_buffer_adjust(adev, adev->mode_info.crtcs[i+1], mode1, mode0);
+		dce_v6_0_program_watermarks(adev, adev->mode_info.crtcs[i+1], lb_size, num_heads);
+	}
+}
+/*
+static void dce_v6_0_audio_get_connected_pins(struct amdgpu_device *adev)
+{
+	int i;
+	u32 offset, tmp;
+
+	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
+		offset = adev->mode_info.audio.pin[i].offset;
+		tmp = RREG32_AUDIO_ENDPT(offset,
+				      AZ_F0_CODEC_PIN_CONTROL_RESPONSE_CONFIGURATION_DEFAULT);
+		if (((tmp & PORT_CONNECTIVITY_MASK) >> PORT_CONNECTIVITY_SHIFT) == 1)
+			adev->mode_info.audio.pin[i].connected = false;
+		else
+			adev->mode_info.audio.pin[i].connected = true;
+	}
+
+}
+
+static struct amdgpu_audio_pin *dce_v6_0_audio_get_pin(struct amdgpu_device *adev)
+{
+	int i;
+
+	dce_v6_0_audio_get_connected_pins(adev);
+
+	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
+		if (adev->mode_info.audio.pin[i].connected)
+			return &adev->mode_info.audio.pin[i];
+	}
+	DRM_ERROR("No connected audio pins found!\n");
+	return NULL;
+}
+
+static void dce_v6_0_afmt_audio_select_pin(struct drm_encoder *encoder)
+{
+	struct amdgpu_device *adev = encoder->dev->dev_private;
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
+	u32 offset;
+
+	if (!dig || !dig->afmt || !dig->afmt->pin)
+		return;
+
+	offset = dig->afmt->offset;
+
+	WREG32(AFMT_AUDIO_SRC_CONTROL + offset,
+	       AFMT_AUDIO_SRC_SELECT(dig->afmt->pin->id));
+
+}
+
+static void dce_v6_0_audio_write_latency_fields(struct drm_encoder *encoder,
+						struct drm_display_mode *mode)
+{
+	DRM_INFO("xxxx: dce_v6_0_audio_write_latency_fields---no imp!!!!!\n");
+}
+
+static void dce_v6_0_audio_write_speaker_allocation(struct drm_encoder *encoder)
+{
+	DRM_INFO("xxxx: dce_v6_0_audio_write_speaker_allocation---no imp!!!!!\n");
+}
+
+static void dce_v6_0_audio_write_sad_regs(struct drm_encoder *encoder)
+{
+	DRM_INFO("xxxx: dce_v6_0_audio_write_sad_regs---no imp!!!!!\n");
+
+}
+*/
+static void dce_v6_0_audio_enable(struct amdgpu_device *adev,
+				  struct amdgpu_audio_pin *pin,
+				  bool enable)
+{
+	DRM_INFO("xxxx: dce_v6_0_audio_enable---no imp!!!!!\n");
+}
+
+static const u32 pin_offsets[7] =
+{
+	(0x1780 - 0x1780),
+	(0x1786 - 0x1780),
+	(0x178c - 0x1780),
+	(0x1792 - 0x1780),
+	(0x1798 - 0x1780),
+	(0x179d - 0x1780),
+	(0x17a4 - 0x1780),
+};
+
+static int dce_v6_0_audio_init(struct amdgpu_device *adev)
+{
+	return 0;
+}
+
+static void dce_v6_0_audio_fini(struct amdgpu_device *adev)
+{
+
+}
+
+/*
+static void dce_v6_0_afmt_update_ACR(struct drm_encoder *encoder, uint32_t clock)
+{
+	DRM_INFO("xxxx: dce_v6_0_afmt_update_ACR---no imp!!!!!\n");
+}
+*/
+/*
+ * build a HDMI Video Info Frame
+ */
+/*
+static void dce_v6_0_afmt_update_avi_infoframe(struct drm_encoder *encoder,
+					       void *buffer, size_t size)
+{
+	DRM_INFO("xxxx: dce_v6_0_afmt_update_avi_infoframe---no imp!!!!!\n");
+}
+
+static void dce_v6_0_audio_set_dto(struct drm_encoder *encoder, u32 clock)
+{
+	DRM_INFO("xxxx: dce_v6_0_audio_set_dto---no imp!!!!!\n");
+}
+*/
+/*
+ * update the info frames with the data from the current display mode
+ */
+static void dce_v6_0_afmt_setmode(struct drm_encoder *encoder,
+				  struct drm_display_mode *mode)
+{
+	DRM_INFO("xxxx: dce_v6_0_afmt_setmode ----no impl !!!!!!!!\n");
+}
+
+static void dce_v6_0_afmt_enable(struct drm_encoder *encoder, bool enable)
+{
+	struct drm_device *dev = encoder->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
+
+	if (!dig || !dig->afmt)
+		return;
+
+	/* Silent, r600_hdmi_enable will raise WARN for us */
+	if (enable && dig->afmt->enabled)
+		return;
+	if (!enable && !dig->afmt->enabled)
+		return;
+
+	if (!enable && dig->afmt->pin) {
+		dce_v6_0_audio_enable(adev, dig->afmt->pin, false);
+		dig->afmt->pin = NULL;
+	}
+
+	dig->afmt->enabled = enable;
+
+	DRM_DEBUG("%sabling AFMT interface @ 0x%04X for encoder 0x%x\n",
+		  enable ? "En" : "Dis", dig->afmt->offset, amdgpu_encoder->encoder_id);
+}
+
+static int dce_v6_0_afmt_init(struct amdgpu_device *adev)
+{
+	int i, j;
+
+	for (i = 0; i < adev->mode_info.num_dig; i++)
+		adev->mode_info.afmt[i] = NULL;
+
+	/* DCE6 has audio blocks tied to DIG encoders */
+	for (i = 0; i < adev->mode_info.num_dig; i++) {
+		adev->mode_info.afmt[i] = kzalloc(sizeof(struct amdgpu_afmt), GFP_KERNEL);
+		if (adev->mode_info.afmt[i]) {
+			adev->mode_info.afmt[i]->offset = dig_offsets[i];
+			adev->mode_info.afmt[i]->id = i;
+		} else {
+			for (j = 0; j < i; j++) {
+				kfree(adev->mode_info.afmt[j]);
+				adev->mode_info.afmt[j] = NULL;
+			}
+			DRM_ERROR("Out of memory allocating afmt table\n");
+			return -ENOMEM;
+		}
+	}
+	return 0;
+}
+
+static void dce_v6_0_afmt_fini(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->mode_info.num_dig; i++) {
+		kfree(adev->mode_info.afmt[i]);
+		adev->mode_info.afmt[i] = NULL;
+	}
+}
+
+static const u32 vga_control_regs[6] =
+{
+	AVIVO_D1VGA_CONTROL,
+	AVIVO_D2VGA_CONTROL,
+	EVERGREEN_D3VGA_CONTROL,
+	EVERGREEN_D4VGA_CONTROL,
+	EVERGREEN_D5VGA_CONTROL,
+	EVERGREEN_D6VGA_CONTROL,
+};
+
+static void dce_v6_0_vga_enable(struct drm_crtc *crtc, bool enable)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	u32 vga_control;
+
+	vga_control = RREG32(vga_control_regs[amdgpu_crtc->crtc_id]) & ~1;
+	WREG32(vga_control_regs[amdgpu_crtc->crtc_id], vga_control | (enable ? 1 : 0));
+}
+
+static void dce_v6_0_grph_enable(struct drm_crtc *crtc, bool enable)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+
+	WREG32(EVERGREEN_GRPH_ENABLE + amdgpu_crtc->crtc_offset, enable ? 1 : 0);
+}
+
+static int dce_v6_0_crtc_do_set_base(struct drm_crtc *crtc,
+				     struct drm_framebuffer *fb,
+				     int x, int y, int atomic)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_framebuffer *amdgpu_fb;
+	struct drm_framebuffer *target_fb;
+	struct drm_gem_object *obj;
+	struct amdgpu_bo *abo;
+	uint64_t fb_location, tiling_flags;
+	uint32_t fb_format, fb_pitch_pixels, pipe_config;
+	u32 fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_NONE);
+	u32 viewport_w, viewport_h;
+	int r;
+	bool bypass_lut = false;
+
+	/* no fb bound */
+	if (!atomic && !crtc->primary->fb) {
+		DRM_DEBUG_KMS("No FB bound\n");
+		return 0;
+	}
+
+	if (atomic) {
+		amdgpu_fb = to_amdgpu_framebuffer(fb);
+		target_fb = fb;
+	} else {
+		amdgpu_fb = to_amdgpu_framebuffer(crtc->primary->fb);
+		target_fb = crtc->primary->fb;
+	}
+
+	/* If atomic, assume fb object is pinned & idle & fenced and
+	 * just update base pointers
+	 */
+	obj = amdgpu_fb->obj;
+	abo = gem_to_amdgpu_bo(obj);
+	r = amdgpu_bo_reserve(abo, false);
+	if (unlikely(r != 0))
+		return r;
+
+	if (atomic) {
+		fb_location = amdgpu_bo_gpu_offset(abo);
+	} else {
+		r = amdgpu_bo_pin(abo, AMDGPU_GEM_DOMAIN_VRAM, &fb_location);
+		if (unlikely(r != 0)) {
+			amdgpu_bo_unreserve(abo);
+			return -EINVAL;
+		}
+	}
+
+	amdgpu_bo_get_tiling_flags(abo, &tiling_flags);
+	amdgpu_bo_unreserve(abo);
+
+	switch (target_fb->pixel_format) {
+	case DRM_FORMAT_C8:
+		fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_8BPP) |
+			     EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_INDEXED));
+		break;
+	case DRM_FORMAT_XRGB4444:
+	case DRM_FORMAT_ARGB4444:
+		fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) |
+			     EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB4444));
+#ifdef __BIG_ENDIAN
+		fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN16);
+#endif
+		break;
+	case DRM_FORMAT_XRGB1555:
+	case DRM_FORMAT_ARGB1555:
+		fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) |
+			     EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB1555));
+#ifdef __BIG_ENDIAN
+		fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN16);
+#endif
+		break;
+	case DRM_FORMAT_BGRX5551:
+	case DRM_FORMAT_BGRA5551:
+		fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) |
+			     EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_BGRA5551));
+#ifdef __BIG_ENDIAN
+		fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN16);
+#endif
+		break;
+	case DRM_FORMAT_RGB565:
+		fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_16BPP) |
+			     EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB565));
+#ifdef __BIG_ENDIAN
+		fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN16);
+#endif
+		break;
+	case DRM_FORMAT_XRGB8888:
+	case DRM_FORMAT_ARGB8888:
+		fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_32BPP) |
+			     EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB8888));
+#ifdef __BIG_ENDIAN
+		fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN32);
+#endif
+		break;
+	case DRM_FORMAT_XRGB2101010:
+	case DRM_FORMAT_ARGB2101010:
+		fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_32BPP) |
+			     EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_ARGB2101010));
+#ifdef __BIG_ENDIAN
+		fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN32);
+#endif
+		/* Greater 8 bpc fb needs to bypass hw-lut to retain precision */
+		bypass_lut = true;
+		break;
+	case DRM_FORMAT_BGRX1010102:
+	case DRM_FORMAT_BGRA1010102:
+		fb_format = (EVERGREEN_GRPH_DEPTH(EVERGREEN_GRPH_DEPTH_32BPP) |
+			     EVERGREEN_GRPH_FORMAT(EVERGREEN_GRPH_FORMAT_BGRA1010102));
+#ifdef __BIG_ENDIAN
+		fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_8IN32);
+#endif
+		/* Greater 8 bpc fb needs to bypass hw-lut to retain precision */
+		bypass_lut = true;
+		break;
+	default:
+		DRM_ERROR("Unsupported screen format %s\n",
+			  drm_get_format_name(target_fb->pixel_format));
+		return -EINVAL;
+	}
+
+	if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == ARRAY_2D_TILED_THIN1) {
+		unsigned bankw, bankh, mtaspect, tile_split, num_banks;
+
+		bankw = AMDGPU_TILING_GET(tiling_flags, BANK_WIDTH);
+		bankh = AMDGPU_TILING_GET(tiling_flags, BANK_HEIGHT);
+		mtaspect = AMDGPU_TILING_GET(tiling_flags, MACRO_TILE_ASPECT);
+		tile_split = AMDGPU_TILING_GET(tiling_flags, TILE_SPLIT);
+		num_banks = AMDGPU_TILING_GET(tiling_flags, NUM_BANKS);
+
+		fb_format |= EVERGREEN_GRPH_NUM_BANKS(num_banks);
+		fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_2D_TILED_THIN1);
+		fb_format |= EVERGREEN_GRPH_TILE_SPLIT(tile_split);
+		fb_format |= EVERGREEN_GRPH_BANK_WIDTH(bankw);
+		fb_format |= EVERGREEN_GRPH_BANK_HEIGHT(bankh);
+		fb_format |= EVERGREEN_GRPH_MACRO_TILE_ASPECT(mtaspect);
+	} else if (AMDGPU_TILING_GET(tiling_flags, ARRAY_MODE) == ARRAY_1D_TILED_THIN1) {
+		fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1);
+	}
+
+	pipe_config = AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
+	fb_format |= SI_GRPH_PIPE_CONFIG(pipe_config);
+
+	dce_v6_0_vga_enable(crtc, false);
+
+	/* Make sure surface address is updated at vertical blank rather than
+	 * horizontal blank
+	 */
+	WREG32(EVERGREEN_GRPH_FLIP_CONTROL + amdgpu_crtc->crtc_offset, 0);
+
+	WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
+	       upper_32_bits(fb_location));
+	WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
+	       upper_32_bits(fb_location));
+	WREG32(EVERGREEN_GRPH_PRIMARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
+	       (u32)fb_location & EVERGREEN_GRPH_SURFACE_ADDRESS_MASK);
+	WREG32(EVERGREEN_GRPH_SECONDARY_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
+	       (u32) fb_location & EVERGREEN_GRPH_SURFACE_ADDRESS_MASK);
+	WREG32(EVERGREEN_GRPH_CONTROL + amdgpu_crtc->crtc_offset, fb_format);
+	WREG32(EVERGREEN_GRPH_SWAP_CONTROL + amdgpu_crtc->crtc_offset, fb_swap);
+
+	/*
+	 * The LUT only has 256 slots for indexing by a 8 bpc fb. Bypass the LUT
+	 * for > 8 bpc scanout to avoid truncation of fb indices to 8 msb's, to
+	 * retain the full precision throughout the pipeline.
+	 */
+	WREG32_P(EVERGREEN_GRPH_LUT_10BIT_BYPASS_CONTROL + amdgpu_crtc->crtc_offset,
+		 (bypass_lut ? EVERGREEN_LUT_10BIT_BYPASS_EN : 0),
+		 ~EVERGREEN_LUT_10BIT_BYPASS_EN);
+
+	if (bypass_lut)
+		DRM_DEBUG_KMS("Bypassing hardware LUT due to 10 bit fb scanout.\n");
+
+	WREG32(EVERGREEN_GRPH_SURFACE_OFFSET_X + amdgpu_crtc->crtc_offset, 0);
+	WREG32(EVERGREEN_GRPH_SURFACE_OFFSET_Y + amdgpu_crtc->crtc_offset, 0);
+	WREG32(EVERGREEN_GRPH_X_START + amdgpu_crtc->crtc_offset, 0);
+	WREG32(EVERGREEN_GRPH_Y_START + amdgpu_crtc->crtc_offset, 0);
+	WREG32(EVERGREEN_GRPH_X_END + amdgpu_crtc->crtc_offset, target_fb->width);
+	WREG32(EVERGREEN_GRPH_Y_END + amdgpu_crtc->crtc_offset, target_fb->height);
+
+	fb_pitch_pixels = target_fb->pitches[0] / (target_fb->bits_per_pixel / 8);
+	WREG32(EVERGREEN_GRPH_PITCH + amdgpu_crtc->crtc_offset, fb_pitch_pixels);
+
+	dce_v6_0_grph_enable(crtc, true);
+
+	WREG32(EVERGREEN_DESKTOP_HEIGHT + amdgpu_crtc->crtc_offset,
+		       target_fb->height);
+	x &= ~3;
+	y &= ~1;
+	WREG32(EVERGREEN_VIEWPORT_START + amdgpu_crtc->crtc_offset,
+	       (x << 16) | y);
+	viewport_w = crtc->mode.hdisplay;
+	viewport_h = (crtc->mode.vdisplay + 1) & ~1;
+
+	WREG32(EVERGREEN_VIEWPORT_SIZE + amdgpu_crtc->crtc_offset,
+	       (viewport_w << 16) | viewport_h);
+
+	/* set pageflip to happen anywhere in vblank interval */
+	WREG32(EVERGREEN_MASTER_UPDATE_MODE + amdgpu_crtc->crtc_offset, 0);
+
+	if (!atomic && fb && fb != crtc->primary->fb) {
+		amdgpu_fb = to_amdgpu_framebuffer(fb);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
+		if (unlikely(r != 0))
+			return r;
+		amdgpu_bo_unpin(abo);
+		amdgpu_bo_unreserve(abo);
+	}
+
+	/* Bytes per pixel may have changed */
+	dce_v6_0_bandwidth_update(adev);
+
+	return 0;
+
+}
+
+static void dce_v6_0_set_interleave(struct drm_crtc *crtc,
+				    struct drm_display_mode *mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+
+	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+		WREG32(EVERGREEN_DATA_FORMAT + amdgpu_crtc->crtc_offset,
+		       EVERGREEN_INTERLEAVE_EN);
+	else
+		WREG32(EVERGREEN_DATA_FORMAT + amdgpu_crtc->crtc_offset, 0);
+}
+
+static void dce_v6_0_crtc_load_lut(struct drm_crtc *crtc)
+{
+
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	int i;
+
+	DRM_DEBUG_KMS("%d\n", amdgpu_crtc->crtc_id);
+
+	WREG32(NI_INPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset,
+	       (NI_INPUT_CSC_GRPH_MODE(NI_INPUT_CSC_BYPASS) |
+		NI_INPUT_CSC_OVL_MODE(NI_INPUT_CSC_BYPASS)));
+	WREG32(NI_PRESCALE_GRPH_CONTROL + amdgpu_crtc->crtc_offset,
+	       NI_GRPH_PRESCALE_BYPASS);
+	WREG32(NI_PRESCALE_OVL_CONTROL + amdgpu_crtc->crtc_offset,
+	       NI_OVL_PRESCALE_BYPASS);
+	WREG32(NI_INPUT_GAMMA_CONTROL + amdgpu_crtc->crtc_offset,
+	       (NI_GRPH_INPUT_GAMMA_MODE(NI_INPUT_GAMMA_USE_LUT) |
+		NI_OVL_INPUT_GAMMA_MODE(NI_INPUT_GAMMA_USE_LUT)));
+
+
+
+	WREG32(EVERGREEN_DC_LUT_CONTROL + amdgpu_crtc->crtc_offset, 0);
+
+	WREG32(EVERGREEN_DC_LUT_BLACK_OFFSET_BLUE + amdgpu_crtc->crtc_offset, 0);
+	WREG32(EVERGREEN_DC_LUT_BLACK_OFFSET_GREEN + amdgpu_crtc->crtc_offset, 0);
+	WREG32(EVERGREEN_DC_LUT_BLACK_OFFSET_RED + amdgpu_crtc->crtc_offset, 0);
+
+	WREG32(EVERGREEN_DC_LUT_WHITE_OFFSET_BLUE + amdgpu_crtc->crtc_offset, 0xffff);
+	WREG32(EVERGREEN_DC_LUT_WHITE_OFFSET_GREEN + amdgpu_crtc->crtc_offset, 0xffff);
+	WREG32(EVERGREEN_DC_LUT_WHITE_OFFSET_RED + amdgpu_crtc->crtc_offset, 0xffff);
+
+	WREG32(EVERGREEN_DC_LUT_RW_MODE + amdgpu_crtc->crtc_offset, 0);
+	WREG32(EVERGREEN_DC_LUT_WRITE_EN_MASK + amdgpu_crtc->crtc_offset, 0x00000007);
+
+	WREG32(EVERGREEN_DC_LUT_RW_INDEX + amdgpu_crtc->crtc_offset, 0);
+	for (i = 0; i < 256; i++) {
+		WREG32(EVERGREEN_DC_LUT_30_COLOR + amdgpu_crtc->crtc_offset,
+		       (amdgpu_crtc->lut_r[i] << 20) |
+		       (amdgpu_crtc->lut_g[i] << 10) |
+		       (amdgpu_crtc->lut_b[i] << 0));
+	}
+
+	WREG32(NI_DEGAMMA_CONTROL + amdgpu_crtc->crtc_offset,
+	       (NI_GRPH_DEGAMMA_MODE(NI_DEGAMMA_BYPASS) |
+		NI_OVL_DEGAMMA_MODE(NI_DEGAMMA_BYPASS) |
+		NI_ICON_DEGAMMA_MODE(NI_DEGAMMA_BYPASS) |
+		NI_CURSOR_DEGAMMA_MODE(NI_DEGAMMA_BYPASS)));
+	WREG32(NI_GAMUT_REMAP_CONTROL + amdgpu_crtc->crtc_offset,
+	       (NI_GRPH_GAMUT_REMAP_MODE(NI_GAMUT_REMAP_BYPASS) |
+		NI_OVL_GAMUT_REMAP_MODE(NI_GAMUT_REMAP_BYPASS)));
+	WREG32(NI_REGAMMA_CONTROL + amdgpu_crtc->crtc_offset,
+	       (NI_GRPH_REGAMMA_MODE(NI_REGAMMA_BYPASS) |
+		NI_OVL_REGAMMA_MODE(NI_REGAMMA_BYPASS)));
+	WREG32(NI_OUTPUT_CSC_CONTROL + amdgpu_crtc->crtc_offset,
+	       (NI_OUTPUT_CSC_GRPH_MODE(0) |
+		NI_OUTPUT_CSC_OVL_MODE(NI_OUTPUT_CSC_BYPASS)));
+	/* XXX match this to the depth of the crtc fmt block, move to modeset? */
+	WREG32(0x1a50 + amdgpu_crtc->crtc_offset, 0);
+
+
+}
+
+static int dce_v6_0_pick_dig_encoder(struct drm_encoder *encoder)
+{
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
+
+	switch (amdgpu_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+		return dig->linkb ? 1 : 0;
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+		return dig->linkb ? 3 : 2;
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+		return dig->linkb ? 5 : 4;
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
+		return 6;
+	default:
+		DRM_ERROR("invalid encoder_id: 0x%x\n", amdgpu_encoder->encoder_id);
+		return 0;
+	}
+}
+
+/**
+ * dce_v6_0_pick_pll - Allocate a PPLL for use by the crtc.
+ *
+ * @crtc: drm crtc
+ *
+ * Returns the PPLL (Pixel PLL) to be used by the crtc.  For DP monitors
+ * a single PPLL can be used for all DP crtcs/encoders.  For non-DP
+ * monitors a dedicated PPLL must be used.  If a particular board has
+ * an external DP PLL, return ATOM_PPLL_INVALID to skip PLL programming
+ * as there is no need to program the PLL itself.  If we are not able to
+ * allocate a PLL, return ATOM_PPLL_INVALID to skip PLL programming to
+ * avoid messing up an existing monitor.
+ *
+ *
+ */
+static u32 dce_v6_0_pick_pll(struct drm_crtc *crtc)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	u32 pll_in_use;
+	int pll;
+
+	if (ENCODER_MODE_IS_DP(amdgpu_atombios_encoder_get_encoder_mode(amdgpu_crtc->encoder))) {
+		if (adev->clock.dp_extclk)
+			/* skip PPLL programming if using ext clock */
+			return ATOM_PPLL_INVALID;
+		else
+			return ATOM_PPLL0;
+	} else {
+		/* use the same PPLL for all monitors with the same clock */
+		pll = amdgpu_pll_get_shared_nondp_ppll(crtc);
+		if (pll != ATOM_PPLL_INVALID)
+			return pll;
+	}
+
+	/*  PPLL1, and PPLL2 */
+	pll_in_use = amdgpu_pll_get_use_mask(crtc);
+	if (!(pll_in_use & (1 << ATOM_PPLL2)))
+		return ATOM_PPLL2;
+	if (!(pll_in_use & (1 << ATOM_PPLL1)))
+		return ATOM_PPLL1;
+	DRM_ERROR("unable to allocate a PPLL\n");
+	return ATOM_PPLL_INVALID;
+}
+
+static void dce_v6_0_lock_cursor(struct drm_crtc *crtc, bool lock)
+{
+	struct amdgpu_device *adev = crtc->dev->dev_private;
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	uint32_t cur_lock;
+
+	cur_lock = RREG32(EVERGREEN_CUR_UPDATE + amdgpu_crtc->crtc_offset);
+	if (lock)
+		cur_lock |= EVERGREEN_CURSOR_UPDATE_LOCK;
+	else
+		cur_lock &= ~EVERGREEN_CURSOR_UPDATE_LOCK;
+	WREG32(EVERGREEN_CUR_UPDATE + amdgpu_crtc->crtc_offset, cur_lock);
+}
+
+static void dce_v6_0_hide_cursor(struct drm_crtc *crtc)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct amdgpu_device *adev = crtc->dev->dev_private;
+
+	WREG32_IDX(EVERGREEN_CUR_CONTROL + amdgpu_crtc->crtc_offset,
+		   EVERGREEN_CURSOR_MODE(EVERGREEN_CURSOR_24_8_PRE_MULT) |
+		   EVERGREEN_CURSOR_URGENT_CONTROL(EVERGREEN_CURSOR_URGENT_1_2));
+
+
+}
+
+static void dce_v6_0_show_cursor(struct drm_crtc *crtc)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct amdgpu_device *adev = crtc->dev->dev_private;
+
+	WREG32(EVERGREEN_CUR_SURFACE_ADDRESS_HIGH + amdgpu_crtc->crtc_offset,
+	       upper_32_bits(amdgpu_crtc->cursor_addr));
+	WREG32(EVERGREEN_CUR_SURFACE_ADDRESS + amdgpu_crtc->crtc_offset,
+	       lower_32_bits(amdgpu_crtc->cursor_addr));
+
+	WREG32_IDX(EVERGREEN_CUR_CONTROL + amdgpu_crtc->crtc_offset,
+		   EVERGREEN_CURSOR_EN |
+		   EVERGREEN_CURSOR_MODE(EVERGREEN_CURSOR_24_8_PRE_MULT) |
+		   EVERGREEN_CURSOR_URGENT_CONTROL(EVERGREEN_CURSOR_URGENT_1_2));
+
+}
+
+static int dce_v6_0_cursor_move_locked(struct drm_crtc *crtc,
+				       int x, int y)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct amdgpu_device *adev = crtc->dev->dev_private;
+	int xorigin = 0, yorigin = 0;
+
+	int w = amdgpu_crtc->cursor_width;
+
+	/* avivo cursor are offset into the total surface */
+	x += crtc->x;
+	y += crtc->y;
+	DRM_DEBUG("x %d y %d c->x %d c->y %d\n", x, y, crtc->x, crtc->y);
+
+	if (x < 0) {
+		xorigin = min(-x, amdgpu_crtc->max_cursor_width - 1);
+		x = 0;
+	}
+	if (y < 0) {
+		yorigin = min(-y, amdgpu_crtc->max_cursor_height - 1);
+		y = 0;
+	}
+
+	WREG32(EVERGREEN_CUR_POSITION + amdgpu_crtc->crtc_offset, (x << 16) | y);
+	WREG32(EVERGREEN_CUR_HOT_SPOT + amdgpu_crtc->crtc_offset, (xorigin << 16) | yorigin);
+	WREG32(EVERGREEN_CUR_SIZE + amdgpu_crtc->crtc_offset,
+	       ((w - 1) << 16) | (amdgpu_crtc->cursor_height - 1));
+
+	amdgpu_crtc->cursor_x = x;
+	amdgpu_crtc->cursor_y = y;
+	return 0;
+}
+
+static int dce_v6_0_crtc_cursor_move(struct drm_crtc *crtc,
+				     int x, int y)
+{
+	int ret;
+
+	dce_v6_0_lock_cursor(crtc, true);
+	ret = dce_v6_0_cursor_move_locked(crtc, x, y);
+	dce_v6_0_lock_cursor(crtc, false);
+
+	return ret;
+}
+
+static int dce_v6_0_crtc_cursor_set2(struct drm_crtc *crtc,
+				     struct drm_file *file_priv,
+				     uint32_t handle,
+				     uint32_t width,
+				     uint32_t height,
+				     int32_t hot_x,
+				     int32_t hot_y)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_gem_object *obj;
+	struct amdgpu_bo *aobj;
+	int ret;
+
+	if (!handle) {
+		/* turn off cursor */
+		dce_v6_0_hide_cursor(crtc);
+		obj = NULL;
+		goto unpin;
+	}
+
+	if ((width > amdgpu_crtc->max_cursor_width) ||
+	    (height > amdgpu_crtc->max_cursor_height)) {
+		DRM_ERROR("bad cursor width or height %d x %d\n", width, height);
+		return -EINVAL;
+	}
+
+	obj = drm_gem_object_lookup(file_priv, handle);
+	if (!obj) {
+		DRM_ERROR("Cannot find cursor object %x for crtc %d\n", handle, amdgpu_crtc->crtc_id);
+		return -ENOENT;
+	}
+
+	aobj = gem_to_amdgpu_bo(obj);
+	ret = amdgpu_bo_reserve(aobj, false);
+	if (ret != 0) {
+		drm_gem_object_unreference_unlocked(obj);
+		return ret;
+	}
+
+	ret = amdgpu_bo_pin(aobj, AMDGPU_GEM_DOMAIN_VRAM, &amdgpu_crtc->cursor_addr);
+	amdgpu_bo_unreserve(aobj);
+	if (ret) {
+		DRM_ERROR("Failed to pin new cursor BO (%d)\n", ret);
+		drm_gem_object_unreference_unlocked(obj);
+		return ret;
+	}
+
+	amdgpu_crtc->cursor_width = width;
+	amdgpu_crtc->cursor_height = height;
+
+	dce_v6_0_lock_cursor(crtc, true);
+
+	if (hot_x != amdgpu_crtc->cursor_hot_x ||
+	    hot_y != amdgpu_crtc->cursor_hot_y) {
+		int x, y;
+
+		x = amdgpu_crtc->cursor_x + amdgpu_crtc->cursor_hot_x - hot_x;
+		y = amdgpu_crtc->cursor_y + amdgpu_crtc->cursor_hot_y - hot_y;
+
+		dce_v6_0_cursor_move_locked(crtc, x, y);
+
+		amdgpu_crtc->cursor_hot_x = hot_x;
+		amdgpu_crtc->cursor_hot_y = hot_y;
+	}
+
+	dce_v6_0_show_cursor(crtc);
+	dce_v6_0_lock_cursor(crtc, false);
+
+unpin:
+	if (amdgpu_crtc->cursor_bo) {
+		struct amdgpu_bo *aobj = gem_to_amdgpu_bo(amdgpu_crtc->cursor_bo);
+		ret = amdgpu_bo_reserve(aobj, false);
+		if (likely(ret == 0)) {
+			amdgpu_bo_unpin(aobj);
+			amdgpu_bo_unreserve(aobj);
+		}
+		drm_gem_object_unreference_unlocked(amdgpu_crtc->cursor_bo);
+	}
+
+	amdgpu_crtc->cursor_bo = obj;
+	return 0;
+}
+
+static void dce_v6_0_cursor_reset(struct drm_crtc *crtc)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+
+	if (amdgpu_crtc->cursor_bo) {
+		dce_v6_0_lock_cursor(crtc, true);
+
+		dce_v6_0_cursor_move_locked(crtc, amdgpu_crtc->cursor_x,
+					    amdgpu_crtc->cursor_y);
+
+		dce_v6_0_show_cursor(crtc);
+		dce_v6_0_lock_cursor(crtc, false);
+	}
+}
+
+static int dce_v6_0_crtc_gamma_set(struct drm_crtc *crtc, u16 *red, u16 *green,
+				   u16 *blue, uint32_t size)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	int i;
+
+	/* userspace palettes are always correct as is */
+	for (i = 0; i < size; i++) {
+		amdgpu_crtc->lut_r[i] = red[i] >> 6;
+		amdgpu_crtc->lut_g[i] = green[i] >> 6;
+		amdgpu_crtc->lut_b[i] = blue[i] >> 6;
+	}
+	dce_v6_0_crtc_load_lut(crtc);
+
+	return 0;
+}
+
+static void dce_v6_0_crtc_destroy(struct drm_crtc *crtc)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+
+	drm_crtc_cleanup(crtc);
+	kfree(amdgpu_crtc);
+}
+
+static const struct drm_crtc_funcs dce_v6_0_crtc_funcs = {
+	.cursor_set2 = dce_v6_0_crtc_cursor_set2,
+	.cursor_move = dce_v6_0_crtc_cursor_move,
+	.gamma_set = dce_v6_0_crtc_gamma_set,
+	.set_config = amdgpu_crtc_set_config,
+	.destroy = dce_v6_0_crtc_destroy,
+	.page_flip_target = amdgpu_crtc_page_flip_target,
+};
+
+static void dce_v6_0_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	unsigned type;
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		amdgpu_crtc->enabled = true;
+		amdgpu_atombios_crtc_enable(crtc, ATOM_ENABLE);
+		amdgpu_atombios_crtc_blank(crtc, ATOM_DISABLE);
+		/* Make sure VBLANK and PFLIP interrupts are still enabled */
+		type = amdgpu_crtc_idx_to_irq_type(adev, amdgpu_crtc->crtc_id);
+		amdgpu_irq_update(adev, &adev->crtc_irq, type);
+		amdgpu_irq_update(adev, &adev->pageflip_irq, type);
+		drm_vblank_post_modeset(dev, amdgpu_crtc->crtc_id);
+		dce_v6_0_crtc_load_lut(crtc);
+		break;
+	case DRM_MODE_DPMS_STANDBY:
+	case DRM_MODE_DPMS_SUSPEND:
+	case DRM_MODE_DPMS_OFF:
+		drm_vblank_pre_modeset(dev, amdgpu_crtc->crtc_id);
+		if (amdgpu_crtc->enabled)
+			amdgpu_atombios_crtc_blank(crtc, ATOM_ENABLE);
+		amdgpu_atombios_crtc_enable(crtc, ATOM_DISABLE);
+		amdgpu_crtc->enabled = false;
+		break;
+	}
+	/* adjust pm to dpms */
+	amdgpu_pm_compute_clocks(adev);
+}
+
+static void dce_v6_0_crtc_prepare(struct drm_crtc *crtc)
+{
+	/* disable crtc pair power gating before programming */
+	amdgpu_atombios_crtc_powergate(crtc, ATOM_DISABLE);
+	amdgpu_atombios_crtc_lock(crtc, ATOM_ENABLE);
+	dce_v6_0_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+}
+
+static void dce_v6_0_crtc_commit(struct drm_crtc *crtc)
+{
+	dce_v6_0_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
+	amdgpu_atombios_crtc_lock(crtc, ATOM_DISABLE);
+}
+
+static void dce_v6_0_crtc_disable(struct drm_crtc *crtc)
+{
+
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_atom_ss ss;
+	int i;
+
+	dce_v6_0_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+	if (crtc->primary->fb) {
+		int r;
+		struct amdgpu_framebuffer *amdgpu_fb;
+		struct amdgpu_bo *abo;
+
+		amdgpu_fb = to_amdgpu_framebuffer(crtc->primary->fb);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
+		if (unlikely(r))
+			DRM_ERROR("failed to reserve abo before unpin\n");
+		else {
+			amdgpu_bo_unpin(abo);
+			amdgpu_bo_unreserve(abo);
+		}
+	}
+	/* disable the GRPH */
+	dce_v6_0_grph_enable(crtc, false);
+
+	amdgpu_atombios_crtc_powergate(crtc, ATOM_ENABLE);
+
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		if (adev->mode_info.crtcs[i] &&
+		    adev->mode_info.crtcs[i]->enabled &&
+		    i != amdgpu_crtc->crtc_id &&
+		    amdgpu_crtc->pll_id == adev->mode_info.crtcs[i]->pll_id) {
+			/* one other crtc is using this pll don't turn
+			 * off the pll
+			 */
+			goto done;
+		}
+	}
+
+	switch (amdgpu_crtc->pll_id) {
+	case ATOM_PPLL1:
+	case ATOM_PPLL2:
+		/* disable the ppll */
+		amdgpu_atombios_crtc_program_pll(crtc, amdgpu_crtc->crtc_id, amdgpu_crtc->pll_id,
+						 0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
+		break;
+	default:
+		break;
+	}
+done:
+	amdgpu_crtc->pll_id = ATOM_PPLL_INVALID;
+	amdgpu_crtc->adjusted_clock = 0;
+	amdgpu_crtc->encoder = NULL;
+	amdgpu_crtc->connector = NULL;
+}
+
+static int dce_v6_0_crtc_mode_set(struct drm_crtc *crtc,
+				  struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode,
+				  int x, int y, struct drm_framebuffer *old_fb)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+
+	if (!amdgpu_crtc->adjusted_clock)
+		return -EINVAL;
+
+	amdgpu_atombios_crtc_set_pll(crtc, adjusted_mode);
+	amdgpu_atombios_crtc_set_dtd_timing(crtc, adjusted_mode);
+	dce_v6_0_crtc_do_set_base(crtc, old_fb, x, y, 0);
+	amdgpu_atombios_crtc_overscan_setup(crtc, mode, adjusted_mode);
+	amdgpu_atombios_crtc_scaler_setup(crtc);
+	dce_v6_0_cursor_reset(crtc);
+	/* update the hw version fpr dpm */
+	amdgpu_crtc->hw_mode = *adjusted_mode;
+
+	return 0;
+}
+
+static bool dce_v6_0_crtc_mode_fixup(struct drm_crtc *crtc,
+				     const struct drm_display_mode *mode,
+				     struct drm_display_mode *adjusted_mode)
+{
+
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct drm_encoder *encoder;
+
+	/* assign the encoder to the amdgpu crtc to avoid repeated lookups later */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		if (encoder->crtc == crtc) {
+			amdgpu_crtc->encoder = encoder;
+			amdgpu_crtc->connector = amdgpu_get_connector_for_encoder(encoder);
+			break;
+		}
+	}
+	if ((amdgpu_crtc->encoder == NULL) || (amdgpu_crtc->connector == NULL)) {
+		amdgpu_crtc->encoder = NULL;
+		amdgpu_crtc->connector = NULL;
+		return false;
+	}
+	if (!amdgpu_crtc_scaling_mode_fixup(crtc, mode, adjusted_mode))
+		return false;
+	if (amdgpu_atombios_crtc_prepare_pll(crtc, adjusted_mode))
+		return false;
+	/* pick pll */
+	amdgpu_crtc->pll_id = dce_v6_0_pick_pll(crtc);
+	/* if we can't get a PPLL for a non-DP encoder, fail */
+	if ((amdgpu_crtc->pll_id == ATOM_PPLL_INVALID) &&
+	    !ENCODER_MODE_IS_DP(amdgpu_atombios_encoder_get_encoder_mode(amdgpu_crtc->encoder)))
+		return false;
+
+	return true;
+}
+
+static int dce_v6_0_crtc_set_base(struct drm_crtc *crtc, int x, int y,
+				  struct drm_framebuffer *old_fb)
+{
+	return dce_v6_0_crtc_do_set_base(crtc, old_fb, x, y, 0);
+}
+
+static int dce_v6_0_crtc_set_base_atomic(struct drm_crtc *crtc,
+					 struct drm_framebuffer *fb,
+					 int x, int y, enum mode_set_atomic state)
+{
+       return dce_v6_0_crtc_do_set_base(crtc, fb, x, y, 1);
+}
+
+static const struct drm_crtc_helper_funcs dce_v6_0_crtc_helper_funcs = {
+	.dpms = dce_v6_0_crtc_dpms,
+	.mode_fixup = dce_v6_0_crtc_mode_fixup,
+	.mode_set = dce_v6_0_crtc_mode_set,
+	.mode_set_base = dce_v6_0_crtc_set_base,
+	.mode_set_base_atomic = dce_v6_0_crtc_set_base_atomic,
+	.prepare = dce_v6_0_crtc_prepare,
+	.commit = dce_v6_0_crtc_commit,
+	.load_lut = dce_v6_0_crtc_load_lut,
+	.disable = dce_v6_0_crtc_disable,
+};
+
+static int dce_v6_0_crtc_init(struct amdgpu_device *adev, int index)
+{
+	struct amdgpu_crtc *amdgpu_crtc;
+	int i;
+
+	amdgpu_crtc = kzalloc(sizeof(struct amdgpu_crtc) +
+			      (AMDGPUFB_CONN_LIMIT * sizeof(struct drm_connector *)), GFP_KERNEL);
+	if (amdgpu_crtc == NULL)
+		return -ENOMEM;
+
+	drm_crtc_init(adev->ddev, &amdgpu_crtc->base, &dce_v6_0_crtc_funcs);
+
+	drm_mode_crtc_set_gamma_size(&amdgpu_crtc->base, 256);
+	amdgpu_crtc->crtc_id = index;
+	adev->mode_info.crtcs[index] = amdgpu_crtc;
+
+	amdgpu_crtc->max_cursor_width = CURSOR_WIDTH;
+	amdgpu_crtc->max_cursor_height = CURSOR_HEIGHT;
+	adev->ddev->mode_config.cursor_width = amdgpu_crtc->max_cursor_width;
+	adev->ddev->mode_config.cursor_height = amdgpu_crtc->max_cursor_height;
+
+	for (i = 0; i < 256; i++) {
+		amdgpu_crtc->lut_r[i] = i << 2;
+		amdgpu_crtc->lut_g[i] = i << 2;
+		amdgpu_crtc->lut_b[i] = i << 2;
+	}
+
+	amdgpu_crtc->crtc_offset = crtc_offsets[amdgpu_crtc->crtc_id];
+
+	amdgpu_crtc->pll_id = ATOM_PPLL_INVALID;
+	amdgpu_crtc->adjusted_clock = 0;
+	amdgpu_crtc->encoder = NULL;
+	amdgpu_crtc->connector = NULL;
+	drm_crtc_helper_add(&amdgpu_crtc->base, &dce_v6_0_crtc_helper_funcs);
+
+	return 0;
+}
+
+static int dce_v6_0_early_init(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	adev->audio_endpt_rreg = &dce_v6_0_audio_endpt_rreg;
+	adev->audio_endpt_wreg = &dce_v6_0_audio_endpt_wreg;
+
+	dce_v6_0_set_display_funcs(adev);
+	dce_v6_0_set_irq_funcs(adev);
+
+	switch (adev->asic_type) {
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+		adev->mode_info.num_crtc = 6;
+		adev->mode_info.num_hpd = 6;
+		adev->mode_info.num_dig = 6;
+		break;
+	case CHIP_OLAND:
+		adev->mode_info.num_crtc = 2;
+		adev->mode_info.num_hpd = 2;
+		adev->mode_info.num_dig = 2;
+		break;
+	default:
+		/* FIXME: not supported yet */
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int dce_v6_0_sw_init(void *handle)
+{
+	int r, i;
+	bool ret;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		r = amdgpu_irq_add_id(adev, i + 1, &adev->crtc_irq);
+		if (r)
+			return r;
+	}
+
+	for (i = 8; i < 20; i += 2) {
+		r = amdgpu_irq_add_id(adev, i, &adev->pageflip_irq);
+		if (r)
+			return r;
+	}
+
+	/* HPD hotplug */
+	r = amdgpu_irq_add_id(adev, 42, &adev->hpd_irq);
+	if (r)
+		return r;
+
+	adev->mode_info.mode_config_initialized = true;
+
+	adev->ddev->mode_config.funcs = &amdgpu_mode_funcs;
+	adev->ddev->mode_config.async_page_flip = true;
+	adev->ddev->mode_config.max_width = 16384;
+	adev->ddev->mode_config.max_height = 16384;
+	adev->ddev->mode_config.preferred_depth = 24;
+	adev->ddev->mode_config.prefer_shadow = 1;
+	adev->ddev->mode_config.fb_base = adev->mc.aper_base;
+
+	r = amdgpu_modeset_create_props(adev);
+	if (r)
+		return r;
+
+	adev->ddev->mode_config.max_width = 16384;
+	adev->ddev->mode_config.max_height = 16384;
+
+	/* allocate crtcs */
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		r = dce_v6_0_crtc_init(adev, i);
+		if (r)
+			return r;
+	}
+
+	ret = amdgpu_atombios_get_connector_info_from_object_table(adev);
+	if (ret)
+		amdgpu_print_display_setup(adev->ddev);
+	else
+		return -EINVAL;
+
+	/* setup afmt */
+	r = dce_v6_0_afmt_init(adev);
+	if (r)
+		return r;
+
+	r = dce_v6_0_audio_init(adev);
+	if (r)
+		return r;
+
+	drm_kms_helper_poll_init(adev->ddev);
+
+	return r;
+}
+
+static int dce_v6_0_sw_fini(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	kfree(adev->mode_info.bios_hardcoded_edid);
+
+	drm_kms_helper_poll_fini(adev->ddev);
+
+	dce_v6_0_audio_fini(adev);
+	dce_v6_0_afmt_fini(adev);
+
+	drm_mode_config_cleanup(adev->ddev);
+	adev->mode_info.mode_config_initialized = false;
+
+	return 0;
+}
+
+static int dce_v6_0_hw_init(void *handle)
+{
+	int i;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	/* init dig PHYs, disp eng pll */
+	amdgpu_atombios_encoder_init_dig(adev);
+	amdgpu_atombios_crtc_set_disp_eng_pll(adev, adev->clock.default_dispclk);
+
+	/* initialize hpd */
+	dce_v6_0_hpd_init(adev);
+
+	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
+		dce_v6_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
+	}
+
+	dce_v6_0_pageflip_interrupt_init(adev);
+
+	return 0;
+}
+
+static int dce_v6_0_hw_fini(void *handle)
+{
+	int i;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	dce_v6_0_hpd_fini(adev);
+
+	for (i = 0; i < adev->mode_info.audio.num_pins; i++) {
+		dce_v6_0_audio_enable(adev, &adev->mode_info.audio.pin[i], false);
+	}
+
+	dce_v6_0_pageflip_interrupt_fini(adev);
+
+	return 0;
+}
+
+static int dce_v6_0_suspend(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	amdgpu_atombios_scratch_regs_save(adev);
+
+	return dce_v6_0_hw_fini(handle);
+}
+
+static int dce_v6_0_resume(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int ret;
+
+	ret = dce_v6_0_hw_init(handle);
+
+	amdgpu_atombios_scratch_regs_restore(adev);
+
+	/* turn on the BL */
+	if (adev->mode_info.bl_encoder) {
+		u8 bl_level = amdgpu_display_backlight_get_level(adev,
+								  adev->mode_info.bl_encoder);
+		amdgpu_display_backlight_set_level(adev, adev->mode_info.bl_encoder,
+						    bl_level);
+	}
+
+	return ret;
+}
+
+static bool dce_v6_0_is_idle(void *handle)
+{
+	return true;
+}
+
+static int dce_v6_0_wait_for_idle(void *handle)
+{
+	return 0;
+}
+
+static int dce_v6_0_soft_reset(void *handle)
+{
+	DRM_INFO("xxxx: dce_v6_0_soft_reset --- no impl!!\n");
+	return 0;
+}
+
+static void dce_v6_0_set_crtc_vblank_interrupt_state(struct amdgpu_device *adev,
+						     int crtc,
+						     enum amdgpu_interrupt_state state)
+{
+	u32 reg_block, interrupt_mask;
+
+	if (crtc >= adev->mode_info.num_crtc) {
+		DRM_DEBUG("invalid crtc %d\n", crtc);
+		return;
+	}
+
+	switch (crtc) {
+	case 0:
+		reg_block = SI_CRTC0_REGISTER_OFFSET;
+		break;
+	case 1:
+		reg_block = SI_CRTC1_REGISTER_OFFSET;
+		break;
+	case 2:
+		reg_block = SI_CRTC2_REGISTER_OFFSET;
+		break;
+	case 3:
+		reg_block = SI_CRTC3_REGISTER_OFFSET;
+		break;
+	case 4:
+		reg_block = SI_CRTC4_REGISTER_OFFSET;
+		break;
+	case 5:
+		reg_block = SI_CRTC5_REGISTER_OFFSET;
+		break;
+	default:
+		DRM_DEBUG("invalid crtc %d\n", crtc);
+		return;
+	}
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		interrupt_mask = RREG32(INT_MASK + reg_block);
+		interrupt_mask &= ~VBLANK_INT_MASK;
+		WREG32(INT_MASK + reg_block, interrupt_mask);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		interrupt_mask = RREG32(INT_MASK + reg_block);
+		interrupt_mask |= VBLANK_INT_MASK;
+		WREG32(INT_MASK + reg_block, interrupt_mask);
+		break;
+	default:
+		break;
+	}
+}
+
+static void dce_v6_0_set_crtc_vline_interrupt_state(struct amdgpu_device *adev,
+						    int crtc,
+						    enum amdgpu_interrupt_state state)
+{
+
+}
+
+static int dce_v6_0_set_hpd_interrupt_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *src,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	u32 dc_hpd_int_cntl_reg, dc_hpd_int_cntl;
+
+	switch (type) {
+	case AMDGPU_HPD_1:
+		dc_hpd_int_cntl_reg = DC_HPD1_INT_CONTROL;
+		break;
+	case AMDGPU_HPD_2:
+		dc_hpd_int_cntl_reg = DC_HPD2_INT_CONTROL;
+		break;
+	case AMDGPU_HPD_3:
+		dc_hpd_int_cntl_reg = DC_HPD3_INT_CONTROL;
+		break;
+	case AMDGPU_HPD_4:
+		dc_hpd_int_cntl_reg = DC_HPD4_INT_CONTROL;
+		break;
+	case AMDGPU_HPD_5:
+		dc_hpd_int_cntl_reg = DC_HPD5_INT_CONTROL;
+		break;
+	case AMDGPU_HPD_6:
+		dc_hpd_int_cntl_reg = DC_HPD6_INT_CONTROL;
+		break;
+	default:
+		DRM_DEBUG("invalid hdp %d\n", type);
+		return 0;
+	}
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		dc_hpd_int_cntl = RREG32(dc_hpd_int_cntl_reg);
+		dc_hpd_int_cntl &= ~(DC_HPDx_INT_EN | DC_HPDx_RX_INT_EN);
+		WREG32(dc_hpd_int_cntl_reg, dc_hpd_int_cntl);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		dc_hpd_int_cntl = RREG32(dc_hpd_int_cntl_reg);
+		dc_hpd_int_cntl |= (DC_HPDx_INT_EN | DC_HPDx_RX_INT_EN);
+		WREG32(dc_hpd_int_cntl_reg, dc_hpd_int_cntl);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int dce_v6_0_set_crtc_interrupt_state(struct amdgpu_device *adev,
+					     struct amdgpu_irq_src *src,
+					     unsigned type,
+					     enum amdgpu_interrupt_state state)
+{
+	switch (type) {
+	case AMDGPU_CRTC_IRQ_VBLANK1:
+		dce_v6_0_set_crtc_vblank_interrupt_state(adev, 0, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VBLANK2:
+		dce_v6_0_set_crtc_vblank_interrupt_state(adev, 1, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VBLANK3:
+		dce_v6_0_set_crtc_vblank_interrupt_state(adev, 2, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VBLANK4:
+		dce_v6_0_set_crtc_vblank_interrupt_state(adev, 3, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VBLANK5:
+		dce_v6_0_set_crtc_vblank_interrupt_state(adev, 4, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VBLANK6:
+		dce_v6_0_set_crtc_vblank_interrupt_state(adev, 5, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VLINE1:
+		dce_v6_0_set_crtc_vline_interrupt_state(adev, 0, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VLINE2:
+		dce_v6_0_set_crtc_vline_interrupt_state(adev, 1, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VLINE3:
+		dce_v6_0_set_crtc_vline_interrupt_state(adev, 2, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VLINE4:
+		dce_v6_0_set_crtc_vline_interrupt_state(adev, 3, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VLINE5:
+		dce_v6_0_set_crtc_vline_interrupt_state(adev, 4, state);
+		break;
+	case AMDGPU_CRTC_IRQ_VLINE6:
+		dce_v6_0_set_crtc_vline_interrupt_state(adev, 5, state);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int dce_v6_0_crtc_irq(struct amdgpu_device *adev,
+			     struct amdgpu_irq_src *source,
+			     struct amdgpu_iv_entry *entry)
+{
+	unsigned crtc = entry->src_id - 1;
+	uint32_t disp_int = RREG32(interrupt_status_offsets[crtc].reg);
+	unsigned irq_type = amdgpu_crtc_idx_to_irq_type(adev, crtc);
+
+	switch (entry->src_data) {
+	case 0: /* vblank */
+		if (disp_int & interrupt_status_offsets[crtc].vblank)
+			WREG32(VBLANK_STATUS + crtc_offsets[crtc], VBLANK_ACK);
+		else
+			DRM_DEBUG("IH: IH event w/o asserted irq bit?\n");
+
+		if (amdgpu_irq_enabled(adev, source, irq_type)) {
+			drm_handle_vblank(adev->ddev, crtc);
+		}
+		DRM_DEBUG("IH: D%d vblank\n", crtc + 1);
+		break;
+	case 1: /* vline */
+		if (disp_int & interrupt_status_offsets[crtc].vline)
+			WREG32(VLINE_STATUS + crtc_offsets[crtc], VLINE_ACK);
+		else
+			DRM_DEBUG("IH: IH event w/o asserted irq bit?\n");
+
+		DRM_DEBUG("IH: D%d vline\n", crtc + 1);
+		break;
+	default:
+		DRM_DEBUG("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data);
+		break;
+	}
+
+	return 0;
+}
+
+static int dce_v6_0_set_pageflip_interrupt_state(struct amdgpu_device *adev,
+						 struct amdgpu_irq_src *src,
+						 unsigned type,
+						 enum amdgpu_interrupt_state state)
+{
+	u32 reg;
+
+	if (type >= adev->mode_info.num_crtc) {
+		DRM_ERROR("invalid pageflip crtc %d\n", type);
+		return -EINVAL;
+	}
+
+	reg = RREG32(GRPH_INT_CONTROL + crtc_offsets[type]);
+	if (state == AMDGPU_IRQ_STATE_DISABLE)
+		WREG32(GRPH_INT_CONTROL + crtc_offsets[type],
+		       reg & ~GRPH_INTERRUPT_CONTROL__GRPH_PFLIP_INT_MASK_MASK);
+	else
+		WREG32(GRPH_INT_CONTROL + crtc_offsets[type],
+		       reg | GRPH_INTERRUPT_CONTROL__GRPH_PFLIP_INT_MASK_MASK);
+
+	return 0;
+}
+
+static int dce_v6_0_pageflip_irq(struct amdgpu_device *adev,
+				 struct amdgpu_irq_src *source,
+				 struct amdgpu_iv_entry *entry)
+{
+		unsigned long flags;
+	unsigned crtc_id;
+	struct amdgpu_crtc *amdgpu_crtc;
+	struct amdgpu_flip_work *works;
+
+	crtc_id = (entry->src_id - 8) >> 1;
+	amdgpu_crtc = adev->mode_info.crtcs[crtc_id];
+
+	if (crtc_id >= adev->mode_info.num_crtc) {
+		DRM_ERROR("invalid pageflip crtc %d\n", crtc_id);
+		return -EINVAL;
+	}
+
+	if (RREG32(GRPH_INT_STATUS + crtc_offsets[crtc_id]) &
+	    GRPH_INTERRUPT_STATUS__GRPH_PFLIP_INT_OCCURRED_MASK)
+		WREG32(GRPH_INT_STATUS + crtc_offsets[crtc_id],
+		       GRPH_INTERRUPT_STATUS__GRPH_PFLIP_INT_CLEAR_MASK);
+
+	/* IRQ could occur when in initial stage */
+	if (amdgpu_crtc == NULL)
+		return 0;
+
+	spin_lock_irqsave(&adev->ddev->event_lock, flags);
+	works = amdgpu_crtc->pflip_works;
+	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED){
+		DRM_DEBUG_DRIVER("amdgpu_crtc->pflip_status = %d != "
+						"AMDGPU_FLIP_SUBMITTED(%d)\n",
+						amdgpu_crtc->pflip_status,
+						AMDGPU_FLIP_SUBMITTED);
+		spin_unlock_irqrestore(&adev->ddev->event_lock, flags);
+		return 0;
+	}
+
+	/* page flip completed. clean up */
+	amdgpu_crtc->pflip_status = AMDGPU_FLIP_NONE;
+	amdgpu_crtc->pflip_works = NULL;
+
+	/* wakeup usersapce */
+	if (works->event)
+		drm_crtc_send_vblank_event(&amdgpu_crtc->base, works->event);
+
+	spin_unlock_irqrestore(&adev->ddev->event_lock, flags);
+
+	drm_crtc_vblank_put(&amdgpu_crtc->base);
+	schedule_work(&works->unpin_work);
+
+	return 0;
+}
+
+static int dce_v6_0_hpd_irq(struct amdgpu_device *adev,
+			    struct amdgpu_irq_src *source,
+			    struct amdgpu_iv_entry *entry)
+{
+	uint32_t disp_int, mask, int_control, tmp;
+	unsigned hpd;
+
+	if (entry->src_data >= adev->mode_info.num_hpd) {
+		DRM_DEBUG("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data);
+		return 0;
+	}
+
+	hpd = entry->src_data;
+	disp_int = RREG32(interrupt_status_offsets[hpd].reg);
+	mask = interrupt_status_offsets[hpd].hpd;
+	int_control = hpd_int_control_offsets[hpd];
+
+	if (disp_int & mask) {
+		tmp = RREG32(int_control);
+		tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
+		WREG32(int_control, tmp);
+		schedule_work(&adev->hotplug_work);
+		DRM_INFO("IH: HPD%d\n", hpd + 1);
+	}
+
+	return 0;
+
+}
+
+static int dce_v6_0_set_clockgating_state(void *handle,
+					  enum amd_clockgating_state state)
+{
+	return 0;
+}
+
+static int dce_v6_0_set_powergating_state(void *handle,
+					  enum amd_powergating_state state)
+{
+	return 0;
+}
+
+const struct amd_ip_funcs dce_v6_0_ip_funcs = {
+	.name = "dce_v6_0",
+	.early_init = dce_v6_0_early_init,
+	.late_init = NULL,
+	.sw_init = dce_v6_0_sw_init,
+	.sw_fini = dce_v6_0_sw_fini,
+	.hw_init = dce_v6_0_hw_init,
+	.hw_fini = dce_v6_0_hw_fini,
+	.suspend = dce_v6_0_suspend,
+	.resume = dce_v6_0_resume,
+	.is_idle = dce_v6_0_is_idle,
+	.wait_for_idle = dce_v6_0_wait_for_idle,
+	.soft_reset = dce_v6_0_soft_reset,
+	.set_clockgating_state = dce_v6_0_set_clockgating_state,
+	.set_powergating_state = dce_v6_0_set_powergating_state,
+};
+
+static void
+dce_v6_0_encoder_mode_set(struct drm_encoder *encoder,
+			  struct drm_display_mode *mode,
+			  struct drm_display_mode *adjusted_mode)
+{
+
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+
+	amdgpu_encoder->pixel_clock = adjusted_mode->clock;
+
+	/* need to call this here rather than in prepare() since we need some crtc info */
+	amdgpu_atombios_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+
+	/* set scaler clears this on some chips */
+	dce_v6_0_set_interleave(encoder->crtc, mode);
+
+	if (amdgpu_atombios_encoder_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI) {
+		dce_v6_0_afmt_enable(encoder, true);
+		dce_v6_0_afmt_setmode(encoder, adjusted_mode);
+	}
+}
+
+static void dce_v6_0_encoder_prepare(struct drm_encoder *encoder)
+{
+
+	struct amdgpu_device *adev = encoder->dev->dev_private;
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+	struct drm_connector *connector = amdgpu_get_connector_for_encoder(encoder);
+
+	if ((amdgpu_encoder->active_device &
+	     (ATOM_DEVICE_DFP_SUPPORT | ATOM_DEVICE_LCD_SUPPORT)) ||
+	    (amdgpu_encoder_get_dp_bridge_encoder_id(encoder) !=
+	     ENCODER_OBJECT_ID_NONE)) {
+		struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
+		if (dig) {
+			dig->dig_encoder = dce_v6_0_pick_dig_encoder(encoder);
+			if (amdgpu_encoder->active_device & ATOM_DEVICE_DFP_SUPPORT)
+				dig->afmt = adev->mode_info.afmt[dig->dig_encoder];
+		}
+	}
+
+	amdgpu_atombios_scratch_regs_lock(adev, true);
+
+	if (connector) {
+		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
+
+		/* select the clock/data port if it uses a router */
+		if (amdgpu_connector->router.cd_valid)
+			amdgpu_i2c_router_select_cd_port(amdgpu_connector);
+
+		/* turn eDP panel on for mode set */
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP)
+			amdgpu_atombios_encoder_set_edp_panel_power(connector,
+							     ATOM_TRANSMITTER_ACTION_POWER_ON);
+	}
+
+	/* this is needed for the pll/ss setup to work correctly in some cases */
+	amdgpu_atombios_encoder_set_crtc_source(encoder);
+	/* set up the FMT blocks */
+	dce_v6_0_program_fmt(encoder);
+}
+
+static void dce_v6_0_encoder_commit(struct drm_encoder *encoder)
+{
+
+	struct drm_device *dev = encoder->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+
+	/* need to call this here as we need the crtc set up */
+	amdgpu_atombios_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
+	amdgpu_atombios_scratch_regs_lock(adev, false);
+}
+
+static void dce_v6_0_encoder_disable(struct drm_encoder *encoder)
+{
+
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+	struct amdgpu_encoder_atom_dig *dig;
+
+	amdgpu_atombios_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+
+	if (amdgpu_atombios_encoder_is_digital(encoder)) {
+		if (amdgpu_atombios_encoder_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI)
+			dce_v6_0_afmt_enable(encoder, false);
+		dig = amdgpu_encoder->enc_priv;
+		dig->dig_encoder = -1;
+	}
+	amdgpu_encoder->active_device = 0;
+}
+
+/* these are handled by the primary encoders */
+static void dce_v6_0_ext_prepare(struct drm_encoder *encoder)
+{
+
+}
+
+static void dce_v6_0_ext_commit(struct drm_encoder *encoder)
+{
+
+}
+
+static void
+dce_v6_0_ext_mode_set(struct drm_encoder *encoder,
+		      struct drm_display_mode *mode,
+		      struct drm_display_mode *adjusted_mode)
+{
+
+}
+
+static void dce_v6_0_ext_disable(struct drm_encoder *encoder)
+{
+
+}
+
+static void
+dce_v6_0_ext_dpms(struct drm_encoder *encoder, int mode)
+{
+
+}
+
+static bool dce_v6_0_ext_mode_fixup(struct drm_encoder *encoder,
+				    const struct drm_display_mode *mode,
+				    struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static const struct drm_encoder_helper_funcs dce_v6_0_ext_helper_funcs = {
+	.dpms = dce_v6_0_ext_dpms,
+	.mode_fixup = dce_v6_0_ext_mode_fixup,
+	.prepare = dce_v6_0_ext_prepare,
+	.mode_set = dce_v6_0_ext_mode_set,
+	.commit = dce_v6_0_ext_commit,
+	.disable = dce_v6_0_ext_disable,
+	/* no detect for TMDS/LVDS yet */
+};
+
+static const struct drm_encoder_helper_funcs dce_v6_0_dig_helper_funcs = {
+	.dpms = amdgpu_atombios_encoder_dpms,
+	.mode_fixup = amdgpu_atombios_encoder_mode_fixup,
+	.prepare = dce_v6_0_encoder_prepare,
+	.mode_set = dce_v6_0_encoder_mode_set,
+	.commit = dce_v6_0_encoder_commit,
+	.disable = dce_v6_0_encoder_disable,
+	.detect = amdgpu_atombios_encoder_dig_detect,
+};
+
+static const struct drm_encoder_helper_funcs dce_v6_0_dac_helper_funcs = {
+	.dpms = amdgpu_atombios_encoder_dpms,
+	.mode_fixup = amdgpu_atombios_encoder_mode_fixup,
+	.prepare = dce_v6_0_encoder_prepare,
+	.mode_set = dce_v6_0_encoder_mode_set,
+	.commit = dce_v6_0_encoder_commit,
+	.detect = amdgpu_atombios_encoder_dac_detect,
+};
+
+static void dce_v6_0_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+	if (amdgpu_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
+		amdgpu_atombios_encoder_fini_backlight(amdgpu_encoder);
+	kfree(amdgpu_encoder->enc_priv);
+	drm_encoder_cleanup(encoder);
+	kfree(amdgpu_encoder);
+}
+
+static const struct drm_encoder_funcs dce_v6_0_encoder_funcs = {
+	.destroy = dce_v6_0_encoder_destroy,
+};
+
+static void dce_v6_0_encoder_add(struct amdgpu_device *adev,
+				 uint32_t encoder_enum,
+				 uint32_t supported_device,
+				 u16 caps)
+{
+	struct drm_device *dev = adev->ddev;
+	struct drm_encoder *encoder;
+	struct amdgpu_encoder *amdgpu_encoder;
+
+	/* see if we already added it */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		amdgpu_encoder = to_amdgpu_encoder(encoder);
+		if (amdgpu_encoder->encoder_enum == encoder_enum) {
+			amdgpu_encoder->devices |= supported_device;
+			return;
+		}
+
+	}
+
+	/* add a new one */
+	amdgpu_encoder = kzalloc(sizeof(struct amdgpu_encoder), GFP_KERNEL);
+	if (!amdgpu_encoder)
+		return;
+
+	encoder = &amdgpu_encoder->base;
+	switch (adev->mode_info.num_crtc) {
+	case 1:
+		encoder->possible_crtcs = 0x1;
+		break;
+	case 2:
+	default:
+		encoder->possible_crtcs = 0x3;
+		break;
+	case 4:
+		encoder->possible_crtcs = 0xf;
+		break;
+	case 6:
+		encoder->possible_crtcs = 0x3f;
+		break;
+	}
+
+	amdgpu_encoder->enc_priv = NULL;
+	amdgpu_encoder->encoder_enum = encoder_enum;
+	amdgpu_encoder->encoder_id = (encoder_enum & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
+	amdgpu_encoder->devices = supported_device;
+	amdgpu_encoder->rmx_type = RMX_OFF;
+	amdgpu_encoder->underscan_type = UNDERSCAN_OFF;
+	amdgpu_encoder->is_ext_encoder = false;
+	amdgpu_encoder->caps = caps;
+
+	switch (amdgpu_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
+		drm_encoder_init(dev, encoder, &dce_v6_0_encoder_funcs,
+				 DRM_MODE_ENCODER_DAC, NULL);
+		drm_encoder_helper_add(encoder, &dce_v6_0_dac_helper_funcs);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY3:
+		if (amdgpu_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
+			amdgpu_encoder->rmx_type = RMX_FULL;
+			drm_encoder_init(dev, encoder, &dce_v6_0_encoder_funcs,
+					 DRM_MODE_ENCODER_LVDS, NULL);
+			amdgpu_encoder->enc_priv = amdgpu_atombios_encoder_get_lcd_info(amdgpu_encoder);
+		} else if (amdgpu_encoder->devices & (ATOM_DEVICE_CRT_SUPPORT)) {
+			drm_encoder_init(dev, encoder, &dce_v6_0_encoder_funcs,
+					 DRM_MODE_ENCODER_DAC, NULL);
+			amdgpu_encoder->enc_priv = amdgpu_atombios_encoder_get_dig_info(amdgpu_encoder);
+		} else {
+			drm_encoder_init(dev, encoder, &dce_v6_0_encoder_funcs,
+					 DRM_MODE_ENCODER_TMDS, NULL);
+			amdgpu_encoder->enc_priv = amdgpu_atombios_encoder_get_dig_info(amdgpu_encoder);
+		}
+		drm_encoder_helper_add(encoder, &dce_v6_0_dig_helper_funcs);
+		break;
+	case ENCODER_OBJECT_ID_SI170B:
+	case ENCODER_OBJECT_ID_CH7303:
+	case ENCODER_OBJECT_ID_EXTERNAL_SDVOA:
+	case ENCODER_OBJECT_ID_EXTERNAL_SDVOB:
+	case ENCODER_OBJECT_ID_TITFP513:
+	case ENCODER_OBJECT_ID_VT1623:
+	case ENCODER_OBJECT_ID_HDMI_SI1930:
+	case ENCODER_OBJECT_ID_TRAVIS:
+	case ENCODER_OBJECT_ID_NUTMEG:
+		/* these are handled by the primary encoders */
+		amdgpu_encoder->is_ext_encoder = true;
+		if (amdgpu_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
+			drm_encoder_init(dev, encoder, &dce_v6_0_encoder_funcs,
+					 DRM_MODE_ENCODER_LVDS, NULL);
+		else if (amdgpu_encoder->devices & (ATOM_DEVICE_CRT_SUPPORT))
+			drm_encoder_init(dev, encoder, &dce_v6_0_encoder_funcs,
+					 DRM_MODE_ENCODER_DAC, NULL);
+		else
+			drm_encoder_init(dev, encoder, &dce_v6_0_encoder_funcs,
+					 DRM_MODE_ENCODER_TMDS, NULL);
+		drm_encoder_helper_add(encoder, &dce_v6_0_ext_helper_funcs);
+		break;
+	}
+}
+
+static const struct amdgpu_display_funcs dce_v6_0_display_funcs = {
+	.set_vga_render_state = &dce_v6_0_set_vga_render_state,
+	.bandwidth_update = &dce_v6_0_bandwidth_update,
+	.vblank_get_counter = &dce_v6_0_vblank_get_counter,
+	.vblank_wait = &dce_v6_0_vblank_wait,
+	.is_display_hung = &dce_v6_0_is_display_hung,
+	.backlight_set_level = &amdgpu_atombios_encoder_set_backlight_level,
+	.backlight_get_level = &amdgpu_atombios_encoder_get_backlight_level,
+	.hpd_sense = &dce_v6_0_hpd_sense,
+	.hpd_set_polarity = &dce_v6_0_hpd_set_polarity,
+	.hpd_get_gpio_reg = &dce_v6_0_hpd_get_gpio_reg,
+	.page_flip = &dce_v6_0_page_flip,
+	.page_flip_get_scanoutpos = &dce_v6_0_crtc_get_scanoutpos,
+	.add_encoder = &dce_v6_0_encoder_add,
+	.add_connector = &amdgpu_connector_add,
+	.stop_mc_access = &dce_v6_0_stop_mc_access,
+	.resume_mc_access = &dce_v6_0_resume_mc_access,
+};
+
+static void dce_v6_0_set_display_funcs(struct amdgpu_device *adev)
+{
+	if (adev->mode_info.funcs == NULL)
+		adev->mode_info.funcs = &dce_v6_0_display_funcs;
+}
+
+static const struct amdgpu_irq_src_funcs dce_v6_0_crtc_irq_funcs = {
+	.set = dce_v6_0_set_crtc_interrupt_state,
+	.process = dce_v6_0_crtc_irq,
+};
+
+static const struct amdgpu_irq_src_funcs dce_v6_0_pageflip_irq_funcs = {
+	.set = dce_v6_0_set_pageflip_interrupt_state,
+	.process = dce_v6_0_pageflip_irq,
+};
+
+static const struct amdgpu_irq_src_funcs dce_v6_0_hpd_irq_funcs = {
+	.set = dce_v6_0_set_hpd_interrupt_state,
+	.process = dce_v6_0_hpd_irq,
+};
+
+static void dce_v6_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_LAST;
+	adev->crtc_irq.funcs = &dce_v6_0_crtc_irq_funcs;
+
+	adev->pageflip_irq.num_types = AMDGPU_PAGEFLIP_IRQ_LAST;
+	adev->pageflip_irq.funcs = &dce_v6_0_pageflip_irq_funcs;
+
+	adev->hpd_irq.num_types = AMDGPU_HPD_LAST;
+	adev->hpd_irq.funcs = &dce_v6_0_hpd_irq_funcs;
+}

drivers/gpu/drm/amd/amdgpu/tonga_smum.h → drivers/gpu/drm/amd/amdgpu/dce_v6_0.h

@@ -1,5 +1,5 @@
 /*
- * Copyright 2014 Advanced Micro Devices, Inc.
+ * Copyright 2015 Advanced Micro Devices, Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -21,22 +21,9 @@
  *
  */
 
-#ifndef TONGA_SMUMGR_H
-#define TONGA_SMUMGR_H
+#ifndef __DCE_V6_0_H__
+#define __DCE_V6_0_H__
 
-#include "tonga_ppsmc.h"
-
-int tonga_smu_init(struct amdgpu_device *adev);
-int tonga_smu_fini(struct amdgpu_device *adev);
-int tonga_smu_start(struct amdgpu_device *adev);
-
-struct tonga_smu_private_data
-{
-	uint8_t *header;
-	uint32_t smu_buffer_addr_high;
-	uint32_t smu_buffer_addr_low;
-	uint32_t header_addr_high;
-	uint32_t header_addr_low;
-};
+extern const struct amd_ip_funcs dce_v6_0_ip_funcs;
 
 #endif

drivers/gpu/drm/amd/amdgpu/dce_v8_0.c

@@ -170,7 +170,7 @@ static bool dce_v8_0_is_counter_moving(struct amdgpu_device *adev, int crtc)
  */
 static void dce_v8_0_vblank_wait(struct amdgpu_device *adev, int crtc)
 {
-	unsigned i = 0;
+	unsigned i = 100;
 
 	if (crtc >= adev->mode_info.num_crtc)
 		return;
@@ -182,14 +182,16 @@ static void dce_v8_0_vblank_wait(struct amdgpu_device *adev, int crtc)
 	 * wait for another frame.
 	 */
 	while (dce_v8_0_is_in_vblank(adev, crtc)) {
-		if (i++ % 100 == 0) {
+		if (i++ == 100) {
+			i = 0;
 			if (!dce_v8_0_is_counter_moving(adev, crtc))
 				break;
 		}
 	}
 
 	while (!dce_v8_0_is_in_vblank(adev, crtc)) {
-		if (i++ % 100 == 0) {
+		if (i++ == 100) {
+			i = 0;
 			if (!dce_v8_0_is_counter_moving(adev, crtc))
 				break;
 		}
@@ -395,15 +397,6 @@ static void dce_v8_0_hpd_init(struct amdgpu_device *adev)
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		struct amdgpu_connector *amdgpu_connector = to_amdgpu_connector(connector);
 
-		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
-		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
-			/* don't try to enable hpd on eDP or LVDS avoid breaking the
-			 * aux dp channel on imac and help (but not completely fix)
-			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
-			 * also avoid interrupt storms during dpms.
-			 */
-			continue;
-		}
 		switch (amdgpu_connector->hpd.hpd) {
 		case AMDGPU_HPD_1:
 			WREG32(mmDC_HPD1_CONTROL, tmp);
@@ -426,6 +419,45 @@ static void dce_v8_0_hpd_init(struct amdgpu_device *adev)
 		default:
 			break;
 		}
+
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP ||
+		    connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
+			/* don't try to enable hpd on eDP or LVDS avoid breaking the
+			 * aux dp channel on imac and help (but not completely fix)
+			 * https://bugzilla.redhat.com/show_bug.cgi?id=726143
+			 * also avoid interrupt storms during dpms.
+			 */
+			u32 dc_hpd_int_cntl_reg, dc_hpd_int_cntl;
+
+			switch (amdgpu_connector->hpd.hpd) {
+			case AMDGPU_HPD_1:
+				dc_hpd_int_cntl_reg = mmDC_HPD1_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_2:
+				dc_hpd_int_cntl_reg = mmDC_HPD2_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_3:
+				dc_hpd_int_cntl_reg = mmDC_HPD3_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_4:
+				dc_hpd_int_cntl_reg = mmDC_HPD4_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_5:
+				dc_hpd_int_cntl_reg = mmDC_HPD5_INT_CONTROL;
+				break;
+			case AMDGPU_HPD_6:
+				dc_hpd_int_cntl_reg = mmDC_HPD6_INT_CONTROL;
+				break;
+			default:
+				continue;
+			}
+
+			dc_hpd_int_cntl = RREG32(dc_hpd_int_cntl_reg);
+			dc_hpd_int_cntl &= ~DC_HPD1_INT_CONTROL__DC_HPD1_INT_EN_MASK;
+			WREG32(dc_hpd_int_cntl_reg, dc_hpd_int_cntl);
+			continue;
+		}
+
 		dce_v8_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
 		amdgpu_irq_get(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
 	}
@@ -604,6 +636,52 @@ static void dce_v8_0_set_vga_render_state(struct amdgpu_device *adev,
 	WREG32(mmVGA_RENDER_CONTROL, tmp);
 }
 
+static int dce_v8_0_get_num_crtc(struct amdgpu_device *adev)
+{
+	int num_crtc = 0;
+
+	switch (adev->asic_type) {
+	case CHIP_BONAIRE:
+	case CHIP_HAWAII:
+		num_crtc = 6;
+		break;
+	case CHIP_KAVERI:
+		num_crtc = 4;
+		break;
+	case CHIP_KABINI:
+	case CHIP_MULLINS:
+		num_crtc = 2;
+		break;
+	default:
+		num_crtc = 0;
+	}
+	return num_crtc;
+}
+
+void dce_v8_0_disable_dce(struct amdgpu_device *adev)
+{
+	/*Disable VGA render and enabled crtc, if has DCE engine*/
+	if (amdgpu_atombios_has_dce_engine_info(adev)) {
+		u32 tmp;
+		int crtc_enabled, i;
+
+		dce_v8_0_set_vga_render_state(adev, false);
+
+		/*Disable crtc*/
+		for (i = 0; i < dce_v8_0_get_num_crtc(adev); i++) {
+			crtc_enabled = REG_GET_FIELD(RREG32(mmCRTC_CONTROL + crtc_offsets[i]),
+									 CRTC_CONTROL, CRTC_MASTER_EN);
+			if (crtc_enabled) {
+				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 1);
+				tmp = RREG32(mmCRTC_CONTROL + crtc_offsets[i]);
+				tmp = REG_SET_FIELD(tmp, CRTC_CONTROL, CRTC_MASTER_EN, 0);
+				WREG32(mmCRTC_CONTROL + crtc_offsets[i], tmp);
+				WREG32(mmCRTC_UPDATE_LOCK + crtc_offsets[i], 0);
+			}
+		}
+	}
+}
+
 static void dce_v8_0_program_fmt(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
@@ -1501,13 +1579,13 @@ static void dce_v8_0_audio_write_sad_regs(struct drm_encoder *encoder)
 
 			if (sad->format == eld_reg_to_type[i][1]) {
 				if (sad->channels > max_channels) {
-				value = (sad->channels <<
-				 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__MAX_CHANNELS__SHIFT) |
-				(sad->byte2 <<
-				 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__DESCRIPTOR_BYTE_2__SHIFT) |
-				(sad->freq <<
-				 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__SUPPORTED_FREQUENCIES__SHIFT);
-				max_channels = sad->channels;
+					value = (sad->channels <<
+						 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__MAX_CHANNELS__SHIFT) |
+					        (sad->byte2 <<
+						 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__DESCRIPTOR_BYTE_2__SHIFT) |
+					        (sad->freq <<
+						 AZALIA_F0_CODEC_PIN_CONTROL_AUDIO_DESCRIPTOR0__SUPPORTED_FREQUENCIES__SHIFT);
+					max_channels = sad->channels;
 				}
 
 				if (sad->format == HDMI_AUDIO_CODING_TYPE_PCM)
@@ -1613,7 +1691,7 @@ static void dce_v8_0_afmt_update_ACR(struct drm_encoder *encoder, uint32_t clock
 	struct amdgpu_encoder_atom_dig *dig = amdgpu_encoder->enc_priv;
 	uint32_t offset = dig->afmt->offset;
 
-	WREG32(mmHDMI_ACR_32_0 + offset, (acr.cts_32khz << HDMI_ACR_44_0__HDMI_ACR_CTS_44__SHIFT));
+	WREG32(mmHDMI_ACR_32_0 + offset, (acr.cts_32khz << HDMI_ACR_32_0__HDMI_ACR_CTS_32__SHIFT));
 	WREG32(mmHDMI_ACR_32_1 + offset, acr.n_32khz);
 
 	WREG32(mmHDMI_ACR_44_0 + offset, (acr.cts_44_1khz << HDMI_ACR_44_0__HDMI_ACR_CTS_44__SHIFT));
@@ -1693,6 +1771,7 @@ static void dce_v8_0_afmt_setmode(struct drm_encoder *encoder,
 	/* Silent, r600_hdmi_enable will raise WARN for us */
 	if (!dig->afmt->enabled)
 		return;
+
 	offset = dig->afmt->offset;
 
 	/* hdmi deep color mode general control packets setup, if bpc > 8 */
@@ -1817,7 +1896,7 @@ static void dce_v8_0_afmt_setmode(struct drm_encoder *encoder,
 
 	WREG32_OR(mmHDMI_INFOFRAME_CONTROL0 + offset,
 		  HDMI_INFOFRAME_CONTROL0__HDMI_AVI_INFO_SEND_MASK | /* enable AVI info frames */
-		  HDMI_INFOFRAME_CONTROL0__HDMI_AVI_INFO_SEND_MASK); /* required for audio info values to be updated */
+		  HDMI_INFOFRAME_CONTROL0__HDMI_AVI_INFO_CONT_MASK); /* required for audio info values to be updated */
 
 	WREG32_P(mmHDMI_INFOFRAME_CONTROL1 + offset,
 		 (2 << HDMI_INFOFRAME_CONTROL1__HDMI_AVI_INFO_LINE__SHIFT), /* anything other than 0 */
@@ -1826,13 +1905,12 @@ static void dce_v8_0_afmt_setmode(struct drm_encoder *encoder,
 	WREG32_OR(mmAFMT_AUDIO_PACKET_CONTROL + offset,
 		  AFMT_AUDIO_PACKET_CONTROL__AFMT_AUDIO_SAMPLE_SEND_MASK); /* send audio packets */
 
-	/* it's unknown what these bits do excatly, but it's indeed quite useful for debugging */
 	WREG32(mmAFMT_RAMP_CONTROL0 + offset, 0x00FFFFFF);
 	WREG32(mmAFMT_RAMP_CONTROL1 + offset, 0x007FFFFF);
 	WREG32(mmAFMT_RAMP_CONTROL2 + offset, 0x00000001);
 	WREG32(mmAFMT_RAMP_CONTROL3 + offset, 0x00000001);
 
-	/* enable audio after to setting up hw */
+	/* enable audio after setting up hw */
 	dce_v8_0_audio_enable(adev, dig->afmt->pin, true);
 }
 
@@ -1944,7 +2022,7 @@ static int dce_v8_0_crtc_do_set_base(struct drm_crtc *crtc,
 	struct amdgpu_framebuffer *amdgpu_fb;
 	struct drm_framebuffer *target_fb;
 	struct drm_gem_object *obj;
-	struct amdgpu_bo *rbo;
+	struct amdgpu_bo *abo;
 	uint64_t fb_location, tiling_flags;
 	uint32_t fb_format, fb_pitch_pixels;
 	u32 fb_swap = (GRPH_ENDIAN_NONE << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
@@ -1952,6 +2030,7 @@ static int dce_v8_0_crtc_do_set_base(struct drm_crtc *crtc,
 	u32 viewport_w, viewport_h;
 	int r;
 	bool bypass_lut = false;
+	char *format_name;
 
 	/* no fb bound */
 	if (!atomic && !crtc->primary->fb) {
@@ -1971,23 +2050,23 @@ static int dce_v8_0_crtc_do_set_base(struct drm_crtc *crtc,
 	 * just update base pointers
 	 */
 	obj = amdgpu_fb->obj;
-	rbo = gem_to_amdgpu_bo(obj);
-	r = amdgpu_bo_reserve(rbo, false);
+	abo = gem_to_amdgpu_bo(obj);
+	r = amdgpu_bo_reserve(abo, false);
 	if (unlikely(r != 0))
 		return r;
 
 	if (atomic) {
-		fb_location = amdgpu_bo_gpu_offset(rbo);
+		fb_location = amdgpu_bo_gpu_offset(abo);
 	} else {
-		r = amdgpu_bo_pin(rbo, AMDGPU_GEM_DOMAIN_VRAM, &fb_location);
+		r = amdgpu_bo_pin(abo, AMDGPU_GEM_DOMAIN_VRAM, &fb_location);
 		if (unlikely(r != 0)) {
-			amdgpu_bo_unreserve(rbo);
+			amdgpu_bo_unreserve(abo);
 			return -EINVAL;
 		}
 	}
 
-	amdgpu_bo_get_tiling_flags(rbo, &tiling_flags);
-	amdgpu_bo_unreserve(rbo);
+	amdgpu_bo_get_tiling_flags(abo, &tiling_flags);
+	amdgpu_bo_unreserve(abo);
 
 	pipe_config = AMDGPU_TILING_GET(tiling_flags, PIPE_CONFIG);
 
@@ -1999,7 +2078,7 @@ static int dce_v8_0_crtc_do_set_base(struct drm_crtc *crtc,
 	case DRM_FORMAT_XRGB4444:
 	case DRM_FORMAT_ARGB4444:
 		fb_format = ((GRPH_DEPTH_16BPP << GRPH_CONTROL__GRPH_DEPTH__SHIFT) |
-			     (GRPH_FORMAT_ARGB1555 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
+			     (GRPH_FORMAT_ARGB4444 << GRPH_CONTROL__GRPH_FORMAT__SHIFT));
 #ifdef __BIG_ENDIAN
 		fb_swap = (GRPH_ENDIAN_8IN16 << GRPH_SWAP_CNTL__GRPH_ENDIAN_SWAP__SHIFT);
 #endif
@@ -2056,8 +2135,9 @@ static int dce_v8_0_crtc_do_set_base(struct drm_crtc *crtc,
 		bypass_lut = true;
 		break;
 	default:
-		DRM_ERROR("Unsupported screen format %s\n",
-			  drm_get_format_name(target_fb->pixel_format));
+		format_name = drm_get_format_name(target_fb->pixel_format);
+		DRM_ERROR("Unsupported screen format %s\n", format_name);
+		kfree(format_name);
 		return -EINVAL;
 	}
 
@@ -2137,17 +2217,17 @@ static int dce_v8_0_crtc_do_set_base(struct drm_crtc *crtc,
 	WREG32(mmVIEWPORT_SIZE + amdgpu_crtc->crtc_offset,
 	       (viewport_w << 16) | viewport_h);
 
-	/* set pageflip to happen only at start of vblank interval (front porch) */
-	WREG32(mmMASTER_UPDATE_MODE + amdgpu_crtc->crtc_offset, 3);
+	/* set pageflip to happen anywhere in vblank interval */
+	WREG32(mmMASTER_UPDATE_MODE + amdgpu_crtc->crtc_offset, 0);
 
 	if (!atomic && fb && fb != crtc->primary->fb) {
 		amdgpu_fb = to_amdgpu_framebuffer(fb);
-		rbo = gem_to_amdgpu_bo(amdgpu_fb->obj);
-		r = amdgpu_bo_reserve(rbo, false);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
 		if (unlikely(r != 0))
 			return r;
-		amdgpu_bo_unpin(rbo);
-		amdgpu_bo_unreserve(rbo);
+		amdgpu_bo_unpin(abo);
+		amdgpu_bo_unreserve(abo);
 	}
 
 	/* Bytes per pixel may have changed */
@@ -2552,7 +2632,7 @@ static const struct drm_crtc_funcs dce_v8_0_crtc_funcs = {
 	.gamma_set = dce_v8_0_crtc_gamma_set,
 	.set_config = amdgpu_crtc_set_config,
 	.destroy = dce_v8_0_crtc_destroy,
-	.page_flip = amdgpu_crtc_page_flip,
+	.page_flip_target = amdgpu_crtc_page_flip_target,
 };
 
 static void dce_v8_0_crtc_dpms(struct drm_crtc *crtc, int mode)
@@ -2619,16 +2699,16 @@ static void dce_v8_0_crtc_disable(struct drm_crtc *crtc)
 	if (crtc->primary->fb) {
 		int r;
 		struct amdgpu_framebuffer *amdgpu_fb;
-		struct amdgpu_bo *rbo;
+		struct amdgpu_bo *abo;
 
 		amdgpu_fb = to_amdgpu_framebuffer(crtc->primary->fb);
-		rbo = gem_to_amdgpu_bo(amdgpu_fb->obj);
-		r = amdgpu_bo_reserve(rbo, false);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
 		if (unlikely(r))
-			DRM_ERROR("failed to reserve rbo before unpin\n");
+			DRM_ERROR("failed to reserve abo before unpin\n");
 		else {
-			amdgpu_bo_unpin(rbo);
-			amdgpu_bo_unreserve(rbo);
+			amdgpu_bo_unpin(abo);
+			amdgpu_bo_unreserve(abo);
 		}
 	}
 	/* disable the GRPH */
@@ -2653,7 +2733,7 @@ static void dce_v8_0_crtc_disable(struct drm_crtc *crtc)
 	case ATOM_PPLL2:
 		/* disable the ppll */
 		amdgpu_atombios_crtc_program_pll(crtc, amdgpu_crtc->crtc_id, amdgpu_crtc->pll_id,
-					  0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
+                                                 0, 0, ATOM_DISABLE, 0, 0, 0, 0, 0, false, &ss);
 		break;
 	case ATOM_PPLL0:
 		/* disable the ppll */
@@ -2803,21 +2883,20 @@ static int dce_v8_0_early_init(void *handle)
 	dce_v8_0_set_display_funcs(adev);
 	dce_v8_0_set_irq_funcs(adev);
 
+	adev->mode_info.num_crtc = dce_v8_0_get_num_crtc(adev);
+
 	switch (adev->asic_type) {
 	case CHIP_BONAIRE:
 	case CHIP_HAWAII:
-		adev->mode_info.num_crtc = 6;
 		adev->mode_info.num_hpd = 6;
 		adev->mode_info.num_dig = 6;
 		break;
 	case CHIP_KAVERI:
-		adev->mode_info.num_crtc = 4;
 		adev->mode_info.num_hpd = 6;
 		adev->mode_info.num_dig = 7;
 		break;
 	case CHIP_KABINI:
 	case CHIP_MULLINS:
-		adev->mode_info.num_crtc = 2;
 		adev->mode_info.num_hpd = 6;
 		adev->mode_info.num_dig = 6; /* ? */
 		break;
@@ -3236,7 +3315,6 @@ static int dce_v8_0_crtc_irq(struct amdgpu_device *adev,
 			drm_handle_vblank(adev->ddev, crtc);
 		}
 		DRM_DEBUG("IH: D%d vblank\n", crtc + 1);
-
 		break;
 	case 1: /* vline */
 		if (disp_int & interrupt_status_offsets[crtc].vline)
@@ -3245,7 +3323,6 @@ static int dce_v8_0_crtc_irq(struct amdgpu_device *adev,
 			DRM_DEBUG("IH: IH event w/o asserted irq bit?\n");
 
 		DRM_DEBUG("IH: D%d vline\n", crtc + 1);
-
 		break;
 	default:
 		DRM_DEBUG("Unhandled interrupt: %d %d\n", entry->src_id, entry->src_data);

drivers/gpu/drm/amd/amdgpu/dce_v8_0.h

@@ -26,4 +26,6 @@
 
 extern const struct amd_ip_funcs dce_v8_0_ip_funcs;
 
+void dce_v8_0_disable_dce(struct amdgpu_device *adev);
+
 #endif

drivers/gpu/drm/amd/amdgpu/dce_virtual.c

@@ -0,0 +1,802 @@
+/*
+ * Copyright 2014 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include "drmP.h"
+#include "amdgpu.h"
+#include "amdgpu_pm.h"
+#include "amdgpu_i2c.h"
+#include "atom.h"
+#include "amdgpu_pll.h"
+#include "amdgpu_connectors.h"
+#ifdef CONFIG_DRM_AMDGPU_CIK
+#include "dce_v8_0.h"
+#endif
+#include "dce_v10_0.h"
+#include "dce_v11_0.h"
+#include "dce_virtual.h"
+
+static void dce_virtual_set_display_funcs(struct amdgpu_device *adev);
+static void dce_virtual_set_irq_funcs(struct amdgpu_device *adev);
+static int dce_virtual_pageflip_irq(struct amdgpu_device *adev,
+				  struct amdgpu_irq_src *source,
+				  struct amdgpu_iv_entry *entry);
+
+/**
+ * dce_virtual_vblank_wait - vblank wait asic callback.
+ *
+ * @adev: amdgpu_device pointer
+ * @crtc: crtc to wait for vblank on
+ *
+ * Wait for vblank on the requested crtc (evergreen+).
+ */
+static void dce_virtual_vblank_wait(struct amdgpu_device *adev, int crtc)
+{
+	return;
+}
+
+static u32 dce_virtual_vblank_get_counter(struct amdgpu_device *adev, int crtc)
+{
+	return 0;
+}
+
+static void dce_virtual_page_flip(struct amdgpu_device *adev,
+			      int crtc_id, u64 crtc_base, bool async)
+{
+	return;
+}
+
+static int dce_virtual_crtc_get_scanoutpos(struct amdgpu_device *adev, int crtc,
+					u32 *vbl, u32 *position)
+{
+	*vbl = 0;
+	*position = 0;
+
+	return -EINVAL;
+}
+
+static bool dce_virtual_hpd_sense(struct amdgpu_device *adev,
+			       enum amdgpu_hpd_id hpd)
+{
+	return true;
+}
+
+static void dce_virtual_hpd_set_polarity(struct amdgpu_device *adev,
+				      enum amdgpu_hpd_id hpd)
+{
+	return;
+}
+
+static u32 dce_virtual_hpd_get_gpio_reg(struct amdgpu_device *adev)
+{
+	return 0;
+}
+
+static bool dce_virtual_is_display_hung(struct amdgpu_device *adev)
+{
+	return false;
+}
+
+static void dce_virtual_stop_mc_access(struct amdgpu_device *adev,
+			      struct amdgpu_mode_mc_save *save)
+{
+	switch (adev->asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_CIK
+	case CHIP_BONAIRE:
+	case CHIP_HAWAII:
+	case CHIP_KAVERI:
+	case CHIP_KABINI:
+	case CHIP_MULLINS:
+		dce_v8_0_disable_dce(adev);
+		break;
+#endif
+	case CHIP_FIJI:
+	case CHIP_TONGA:
+		dce_v10_0_disable_dce(adev);
+		break;
+	case CHIP_CARRIZO:
+	case CHIP_STONEY:
+	case CHIP_POLARIS11:
+	case CHIP_POLARIS10:
+		dce_v11_0_disable_dce(adev);
+		break;
+	case CHIP_TOPAZ:
+		/* no DCE */
+		return;
+	default:
+		DRM_ERROR("Virtual display unsupported ASIC type: 0x%X\n", adev->asic_type);
+	}
+
+	return;
+}
+static void dce_virtual_resume_mc_access(struct amdgpu_device *adev,
+				struct amdgpu_mode_mc_save *save)
+{
+	return;
+}
+
+static void dce_virtual_set_vga_render_state(struct amdgpu_device *adev,
+				    bool render)
+{
+	return;
+}
+
+/**
+ * dce_virtual_bandwidth_update - program display watermarks
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Calculate and program the display watermarks and line
+ * buffer allocation (CIK).
+ */
+static void dce_virtual_bandwidth_update(struct amdgpu_device *adev)
+{
+	return;
+}
+
+static int dce_virtual_crtc_gamma_set(struct drm_crtc *crtc, u16 *red,
+				      u16 *green, u16 *blue, uint32_t size)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	int i;
+
+	/* userspace palettes are always correct as is */
+	for (i = 0; i < size; i++) {
+		amdgpu_crtc->lut_r[i] = red[i] >> 6;
+		amdgpu_crtc->lut_g[i] = green[i] >> 6;
+		amdgpu_crtc->lut_b[i] = blue[i] >> 6;
+	}
+
+	return 0;
+}
+
+static void dce_virtual_crtc_destroy(struct drm_crtc *crtc)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+
+	drm_crtc_cleanup(crtc);
+	kfree(amdgpu_crtc);
+}
+
+static const struct drm_crtc_funcs dce_virtual_crtc_funcs = {
+	.cursor_set2 = NULL,
+	.cursor_move = NULL,
+	.gamma_set = dce_virtual_crtc_gamma_set,
+	.set_config = amdgpu_crtc_set_config,
+	.destroy = dce_virtual_crtc_destroy,
+	.page_flip_target = amdgpu_crtc_page_flip_target,
+};
+
+static void dce_virtual_crtc_dpms(struct drm_crtc *crtc, int mode)
+{
+	struct drm_device *dev = crtc->dev;
+	struct amdgpu_device *adev = dev->dev_private;
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	unsigned type;
+
+	switch (mode) {
+	case DRM_MODE_DPMS_ON:
+		amdgpu_crtc->enabled = true;
+		/* Make sure VBLANK and PFLIP interrupts are still enabled */
+		type = amdgpu_crtc_idx_to_irq_type(adev, amdgpu_crtc->crtc_id);
+		amdgpu_irq_update(adev, &adev->crtc_irq, type);
+		amdgpu_irq_update(adev, &adev->pageflip_irq, type);
+		drm_vblank_on(dev, amdgpu_crtc->crtc_id);
+		break;
+	case DRM_MODE_DPMS_STANDBY:
+	case DRM_MODE_DPMS_SUSPEND:
+	case DRM_MODE_DPMS_OFF:
+		drm_vblank_off(dev, amdgpu_crtc->crtc_id);
+		amdgpu_crtc->enabled = false;
+		break;
+	}
+}
+
+
+static void dce_virtual_crtc_prepare(struct drm_crtc *crtc)
+{
+	dce_virtual_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+}
+
+static void dce_virtual_crtc_commit(struct drm_crtc *crtc)
+{
+	dce_virtual_crtc_dpms(crtc, DRM_MODE_DPMS_ON);
+}
+
+static void dce_virtual_crtc_disable(struct drm_crtc *crtc)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+
+	dce_virtual_crtc_dpms(crtc, DRM_MODE_DPMS_OFF);
+	if (crtc->primary->fb) {
+		int r;
+		struct amdgpu_framebuffer *amdgpu_fb;
+		struct amdgpu_bo *abo;
+
+		amdgpu_fb = to_amdgpu_framebuffer(crtc->primary->fb);
+		abo = gem_to_amdgpu_bo(amdgpu_fb->obj);
+		r = amdgpu_bo_reserve(abo, false);
+		if (unlikely(r))
+			DRM_ERROR("failed to reserve abo before unpin\n");
+		else {
+			amdgpu_bo_unpin(abo);
+			amdgpu_bo_unreserve(abo);
+		}
+	}
+
+	amdgpu_crtc->pll_id = ATOM_PPLL_INVALID;
+	amdgpu_crtc->encoder = NULL;
+	amdgpu_crtc->connector = NULL;
+}
+
+static int dce_virtual_crtc_mode_set(struct drm_crtc *crtc,
+				  struct drm_display_mode *mode,
+				  struct drm_display_mode *adjusted_mode,
+				  int x, int y, struct drm_framebuffer *old_fb)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+
+	/* update the hw version fpr dpm */
+	amdgpu_crtc->hw_mode = *adjusted_mode;
+
+	return 0;
+}
+
+static bool dce_virtual_crtc_mode_fixup(struct drm_crtc *crtc,
+				     const struct drm_display_mode *mode,
+				     struct drm_display_mode *adjusted_mode)
+{
+	struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct drm_encoder *encoder;
+
+	/* assign the encoder to the amdgpu crtc to avoid repeated lookups later */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		if (encoder->crtc == crtc) {
+			amdgpu_crtc->encoder = encoder;
+			amdgpu_crtc->connector = amdgpu_get_connector_for_encoder(encoder);
+			break;
+		}
+	}
+	if ((amdgpu_crtc->encoder == NULL) || (amdgpu_crtc->connector == NULL)) {
+		amdgpu_crtc->encoder = NULL;
+		amdgpu_crtc->connector = NULL;
+		return false;
+	}
+
+	return true;
+}
+
+
+static int dce_virtual_crtc_set_base(struct drm_crtc *crtc, int x, int y,
+				  struct drm_framebuffer *old_fb)
+{
+	return 0;
+}
+
+static void dce_virtual_crtc_load_lut(struct drm_crtc *crtc)
+{
+	return;
+}
+
+static int dce_virtual_crtc_set_base_atomic(struct drm_crtc *crtc,
+					 struct drm_framebuffer *fb,
+					 int x, int y, enum mode_set_atomic state)
+{
+	return 0;
+}
+
+static const struct drm_crtc_helper_funcs dce_virtual_crtc_helper_funcs = {
+	.dpms = dce_virtual_crtc_dpms,
+	.mode_fixup = dce_virtual_crtc_mode_fixup,
+	.mode_set = dce_virtual_crtc_mode_set,
+	.mode_set_base = dce_virtual_crtc_set_base,
+	.mode_set_base_atomic = dce_virtual_crtc_set_base_atomic,
+	.prepare = dce_virtual_crtc_prepare,
+	.commit = dce_virtual_crtc_commit,
+	.load_lut = dce_virtual_crtc_load_lut,
+	.disable = dce_virtual_crtc_disable,
+};
+
+static int dce_virtual_crtc_init(struct amdgpu_device *adev, int index)
+{
+	struct amdgpu_crtc *amdgpu_crtc;
+	int i;
+
+	amdgpu_crtc = kzalloc(sizeof(struct amdgpu_crtc) +
+			      (AMDGPUFB_CONN_LIMIT * sizeof(struct drm_connector *)), GFP_KERNEL);
+	if (amdgpu_crtc == NULL)
+		return -ENOMEM;
+
+	drm_crtc_init(adev->ddev, &amdgpu_crtc->base, &dce_virtual_crtc_funcs);
+
+	drm_mode_crtc_set_gamma_size(&amdgpu_crtc->base, 256);
+	amdgpu_crtc->crtc_id = index;
+	adev->mode_info.crtcs[index] = amdgpu_crtc;
+
+	for (i = 0; i < 256; i++) {
+		amdgpu_crtc->lut_r[i] = i << 2;
+		amdgpu_crtc->lut_g[i] = i << 2;
+		amdgpu_crtc->lut_b[i] = i << 2;
+	}
+
+	amdgpu_crtc->pll_id = ATOM_PPLL_INVALID;
+	amdgpu_crtc->encoder = NULL;
+	amdgpu_crtc->connector = NULL;
+	drm_crtc_helper_add(&amdgpu_crtc->base, &dce_virtual_crtc_helper_funcs);
+
+	return 0;
+}
+
+static int dce_virtual_early_init(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	adev->mode_info.vsync_timer_enabled = AMDGPU_IRQ_STATE_DISABLE;
+	dce_virtual_set_display_funcs(adev);
+	dce_virtual_set_irq_funcs(adev);
+
+	adev->mode_info.num_crtc = 1;
+	adev->mode_info.num_hpd = 1;
+	adev->mode_info.num_dig = 1;
+	return 0;
+}
+
+static bool dce_virtual_get_connector_info(struct amdgpu_device *adev)
+{
+	struct amdgpu_i2c_bus_rec ddc_bus;
+	struct amdgpu_router router;
+	struct amdgpu_hpd hpd;
+
+	/* look up gpio for ddc, hpd */
+	ddc_bus.valid = false;
+	hpd.hpd = AMDGPU_HPD_NONE;
+	/* needed for aux chan transactions */
+	ddc_bus.hpd = hpd.hpd;
+
+	memset(&router, 0, sizeof(router));
+	router.ddc_valid = false;
+	router.cd_valid = false;
+	amdgpu_display_add_connector(adev,
+				      0,
+				      ATOM_DEVICE_CRT1_SUPPORT,
+				      DRM_MODE_CONNECTOR_VIRTUAL, &ddc_bus,
+				      CONNECTOR_OBJECT_ID_VIRTUAL,
+				      &hpd,
+				      &router);
+
+	amdgpu_display_add_encoder(adev, ENCODER_VIRTUAL_ENUM_VIRTUAL,
+							ATOM_DEVICE_CRT1_SUPPORT,
+							0);
+
+	amdgpu_link_encoder_connector(adev->ddev);
+
+	return true;
+}
+
+static int dce_virtual_sw_init(void *handle)
+{
+	int r, i;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	r = amdgpu_irq_add_id(adev, 229, &adev->crtc_irq);
+	if (r)
+		return r;
+
+	adev->ddev->max_vblank_count = 0;
+
+	adev->ddev->mode_config.funcs = &amdgpu_mode_funcs;
+
+	adev->ddev->mode_config.max_width = 16384;
+	adev->ddev->mode_config.max_height = 16384;
+
+	adev->ddev->mode_config.preferred_depth = 24;
+	adev->ddev->mode_config.prefer_shadow = 1;
+
+	adev->ddev->mode_config.fb_base = adev->mc.aper_base;
+
+	r = amdgpu_modeset_create_props(adev);
+	if (r)
+		return r;
+
+	adev->ddev->mode_config.max_width = 16384;
+	adev->ddev->mode_config.max_height = 16384;
+
+	/* allocate crtcs */
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		r = dce_virtual_crtc_init(adev, i);
+		if (r)
+			return r;
+	}
+
+	dce_virtual_get_connector_info(adev);
+	amdgpu_print_display_setup(adev->ddev);
+
+	drm_kms_helper_poll_init(adev->ddev);
+
+	adev->mode_info.mode_config_initialized = true;
+	return 0;
+}
+
+static int dce_virtual_sw_fini(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	kfree(adev->mode_info.bios_hardcoded_edid);
+
+	drm_kms_helper_poll_fini(adev->ddev);
+
+	drm_mode_config_cleanup(adev->ddev);
+	adev->mode_info.mode_config_initialized = false;
+	return 0;
+}
+
+static int dce_virtual_hw_init(void *handle)
+{
+	return 0;
+}
+
+static int dce_virtual_hw_fini(void *handle)
+{
+	return 0;
+}
+
+static int dce_virtual_suspend(void *handle)
+{
+	return dce_virtual_hw_fini(handle);
+}
+
+static int dce_virtual_resume(void *handle)
+{
+	return dce_virtual_hw_init(handle);
+}
+
+static bool dce_virtual_is_idle(void *handle)
+{
+	return true;
+}
+
+static int dce_virtual_wait_for_idle(void *handle)
+{
+	return 0;
+}
+
+static int dce_virtual_soft_reset(void *handle)
+{
+	return 0;
+}
+
+static int dce_virtual_set_clockgating_state(void *handle,
+					  enum amd_clockgating_state state)
+{
+	return 0;
+}
+
+static int dce_virtual_set_powergating_state(void *handle,
+					  enum amd_powergating_state state)
+{
+	return 0;
+}
+
+const struct amd_ip_funcs dce_virtual_ip_funcs = {
+	.name = "dce_virtual",
+	.early_init = dce_virtual_early_init,
+	.late_init = NULL,
+	.sw_init = dce_virtual_sw_init,
+	.sw_fini = dce_virtual_sw_fini,
+	.hw_init = dce_virtual_hw_init,
+	.hw_fini = dce_virtual_hw_fini,
+	.suspend = dce_virtual_suspend,
+	.resume = dce_virtual_resume,
+	.is_idle = dce_virtual_is_idle,
+	.wait_for_idle = dce_virtual_wait_for_idle,
+	.soft_reset = dce_virtual_soft_reset,
+	.set_clockgating_state = dce_virtual_set_clockgating_state,
+	.set_powergating_state = dce_virtual_set_powergating_state,
+};
+
+/* these are handled by the primary encoders */
+static void dce_virtual_encoder_prepare(struct drm_encoder *encoder)
+{
+	return;
+}
+
+static void dce_virtual_encoder_commit(struct drm_encoder *encoder)
+{
+	return;
+}
+
+static void
+dce_virtual_encoder_mode_set(struct drm_encoder *encoder,
+		      struct drm_display_mode *mode,
+		      struct drm_display_mode *adjusted_mode)
+{
+	return;
+}
+
+static void dce_virtual_encoder_disable(struct drm_encoder *encoder)
+{
+	return;
+}
+
+static void
+dce_virtual_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	return;
+}
+
+static bool dce_virtual_encoder_mode_fixup(struct drm_encoder *encoder,
+				    const struct drm_display_mode *mode,
+				    struct drm_display_mode *adjusted_mode)
+{
+
+	/* set the active encoder to connector routing */
+	amdgpu_encoder_set_active_device(encoder);
+
+	return true;
+}
+
+static const struct drm_encoder_helper_funcs dce_virtual_encoder_helper_funcs = {
+	.dpms = dce_virtual_encoder_dpms,
+	.mode_fixup = dce_virtual_encoder_mode_fixup,
+	.prepare = dce_virtual_encoder_prepare,
+	.mode_set = dce_virtual_encoder_mode_set,
+	.commit = dce_virtual_encoder_commit,
+	.disable = dce_virtual_encoder_disable,
+};
+
+static void dce_virtual_encoder_destroy(struct drm_encoder *encoder)
+{
+	struct amdgpu_encoder *amdgpu_encoder = to_amdgpu_encoder(encoder);
+
+	kfree(amdgpu_encoder->enc_priv);
+	drm_encoder_cleanup(encoder);
+	kfree(amdgpu_encoder);
+}
+
+static const struct drm_encoder_funcs dce_virtual_encoder_funcs = {
+	.destroy = dce_virtual_encoder_destroy,
+};
+
+static void dce_virtual_encoder_add(struct amdgpu_device *adev,
+				 uint32_t encoder_enum,
+				 uint32_t supported_device,
+				 u16 caps)
+{
+	struct drm_device *dev = adev->ddev;
+	struct drm_encoder *encoder;
+	struct amdgpu_encoder *amdgpu_encoder;
+
+	/* see if we already added it */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		amdgpu_encoder = to_amdgpu_encoder(encoder);
+		if (amdgpu_encoder->encoder_enum == encoder_enum) {
+			amdgpu_encoder->devices |= supported_device;
+			return;
+		}
+
+	}
+
+	/* add a new one */
+	amdgpu_encoder = kzalloc(sizeof(struct amdgpu_encoder), GFP_KERNEL);
+	if (!amdgpu_encoder)
+		return;
+
+	encoder = &amdgpu_encoder->base;
+	encoder->possible_crtcs = 0x1;
+	amdgpu_encoder->enc_priv = NULL;
+	amdgpu_encoder->encoder_enum = encoder_enum;
+	amdgpu_encoder->encoder_id = (encoder_enum & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
+	amdgpu_encoder->devices = supported_device;
+	amdgpu_encoder->rmx_type = RMX_OFF;
+	amdgpu_encoder->underscan_type = UNDERSCAN_OFF;
+	amdgpu_encoder->is_ext_encoder = false;
+	amdgpu_encoder->caps = caps;
+
+	drm_encoder_init(dev, encoder, &dce_virtual_encoder_funcs,
+					 DRM_MODE_ENCODER_VIRTUAL, NULL);
+	drm_encoder_helper_add(encoder, &dce_virtual_encoder_helper_funcs);
+	DRM_INFO("[FM]encoder: %d is VIRTUAL\n", amdgpu_encoder->encoder_id);
+}
+
+static const struct amdgpu_display_funcs dce_virtual_display_funcs = {
+	.set_vga_render_state = &dce_virtual_set_vga_render_state,
+	.bandwidth_update = &dce_virtual_bandwidth_update,
+	.vblank_get_counter = &dce_virtual_vblank_get_counter,
+	.vblank_wait = &dce_virtual_vblank_wait,
+	.is_display_hung = &dce_virtual_is_display_hung,
+	.backlight_set_level = NULL,
+	.backlight_get_level = NULL,
+	.hpd_sense = &dce_virtual_hpd_sense,
+	.hpd_set_polarity = &dce_virtual_hpd_set_polarity,
+	.hpd_get_gpio_reg = &dce_virtual_hpd_get_gpio_reg,
+	.page_flip = &dce_virtual_page_flip,
+	.page_flip_get_scanoutpos = &dce_virtual_crtc_get_scanoutpos,
+	.add_encoder = &dce_virtual_encoder_add,
+	.add_connector = &amdgpu_connector_add,
+	.stop_mc_access = &dce_virtual_stop_mc_access,
+	.resume_mc_access = &dce_virtual_resume_mc_access,
+};
+
+static void dce_virtual_set_display_funcs(struct amdgpu_device *adev)
+{
+	if (adev->mode_info.funcs == NULL)
+		adev->mode_info.funcs = &dce_virtual_display_funcs;
+}
+
+static enum hrtimer_restart dce_virtual_vblank_timer_handle(struct hrtimer *vblank_timer)
+{
+	struct amdgpu_mode_info *mode_info = container_of(vblank_timer, struct amdgpu_mode_info ,vblank_timer);
+	struct amdgpu_device *adev = container_of(mode_info, struct amdgpu_device ,mode_info);
+	unsigned crtc = 0;
+	drm_handle_vblank(adev->ddev, crtc);
+	dce_virtual_pageflip_irq(adev, NULL, NULL);
+	hrtimer_start(vblank_timer, ktime_set(0, DCE_VIRTUAL_VBLANK_PERIOD), HRTIMER_MODE_REL);
+	return HRTIMER_NORESTART;
+}
+
+static void dce_virtual_set_crtc_vblank_interrupt_state(struct amdgpu_device *adev,
+						     int crtc,
+						     enum amdgpu_interrupt_state state)
+{
+	if (crtc >= adev->mode_info.num_crtc) {
+		DRM_DEBUG("invalid crtc %d\n", crtc);
+		return;
+	}
+
+	if (state && !adev->mode_info.vsync_timer_enabled) {
+		DRM_DEBUG("Enable software vsync timer\n");
+		hrtimer_init(&adev->mode_info.vblank_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+		hrtimer_set_expires(&adev->mode_info.vblank_timer, ktime_set(0, DCE_VIRTUAL_VBLANK_PERIOD));
+		adev->mode_info.vblank_timer.function = dce_virtual_vblank_timer_handle;
+		hrtimer_start(&adev->mode_info.vblank_timer, ktime_set(0, DCE_VIRTUAL_VBLANK_PERIOD), HRTIMER_MODE_REL);
+	} else if (!state && adev->mode_info.vsync_timer_enabled) {
+		DRM_DEBUG("Disable software vsync timer\n");
+		hrtimer_cancel(&adev->mode_info.vblank_timer);
+	}
+
+	adev->mode_info.vsync_timer_enabled = state;
+	DRM_DEBUG("[FM]set crtc %d vblank interrupt state %d\n", crtc, state);
+}
+
+
+static int dce_virtual_set_crtc_irq_state(struct amdgpu_device *adev,
+                                       struct amdgpu_irq_src *source,
+                                       unsigned type,
+                                       enum amdgpu_interrupt_state state)
+{
+	switch (type) {
+	case AMDGPU_CRTC_IRQ_VBLANK1:
+		dce_virtual_set_crtc_vblank_interrupt_state(adev, 0, state);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static void dce_virtual_crtc_vblank_int_ack(struct amdgpu_device *adev,
+					  int crtc)
+{
+	if (crtc >= adev->mode_info.num_crtc) {
+		DRM_DEBUG("invalid crtc %d\n", crtc);
+		return;
+	}
+}
+
+static int dce_virtual_crtc_irq(struct amdgpu_device *adev,
+			      struct amdgpu_irq_src *source,
+			      struct amdgpu_iv_entry *entry)
+{
+	unsigned crtc = 0;
+	unsigned irq_type = AMDGPU_CRTC_IRQ_VBLANK1;
+
+	dce_virtual_crtc_vblank_int_ack(adev, crtc);
+
+	if (amdgpu_irq_enabled(adev, source, irq_type)) {
+		drm_handle_vblank(adev->ddev, crtc);
+	}
+	dce_virtual_pageflip_irq(adev, NULL, NULL);
+	DRM_DEBUG("IH: D%d vblank\n", crtc + 1);
+	return 0;
+}
+
+static int dce_virtual_set_pageflip_irq_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *src,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	if (type >= adev->mode_info.num_crtc) {
+		DRM_ERROR("invalid pageflip crtc %d\n", type);
+		return -EINVAL;
+	}
+	DRM_DEBUG("[FM]set pageflip irq type %d state %d\n", type, state);
+
+	return 0;
+}
+
+static int dce_virtual_pageflip_irq(struct amdgpu_device *adev,
+				  struct amdgpu_irq_src *source,
+				  struct amdgpu_iv_entry *entry)
+{
+	unsigned long flags;
+	unsigned crtc_id = 0;
+	struct amdgpu_crtc *amdgpu_crtc;
+	struct amdgpu_flip_work *works;
+
+	crtc_id = 0;
+	amdgpu_crtc = adev->mode_info.crtcs[crtc_id];
+
+	if (crtc_id >= adev->mode_info.num_crtc) {
+		DRM_ERROR("invalid pageflip crtc %d\n", crtc_id);
+		return -EINVAL;
+	}
+
+	/* IRQ could occur when in initial stage */
+	if (amdgpu_crtc == NULL)
+		return 0;
+
+	spin_lock_irqsave(&adev->ddev->event_lock, flags);
+	works = amdgpu_crtc->pflip_works;
+	if (amdgpu_crtc->pflip_status != AMDGPU_FLIP_SUBMITTED) {
+		DRM_DEBUG_DRIVER("amdgpu_crtc->pflip_status = %d != "
+			"AMDGPU_FLIP_SUBMITTED(%d)\n",
+			amdgpu_crtc->pflip_status,
+			AMDGPU_FLIP_SUBMITTED);
+		spin_unlock_irqrestore(&adev->ddev->event_lock, flags);
+		return 0;
+	}
+
+	/* page flip completed. clean up */
+	amdgpu_crtc->pflip_status = AMDGPU_FLIP_NONE;
+	amdgpu_crtc->pflip_works = NULL;
+
+	/* wakeup usersapce */
+	if (works->event)
+		drm_crtc_send_vblank_event(&amdgpu_crtc->base, works->event);
+
+	spin_unlock_irqrestore(&adev->ddev->event_lock, flags);
+
+	drm_crtc_vblank_put(&amdgpu_crtc->base);
+	schedule_work(&works->unpin_work);
+
+	return 0;
+}
+
+static const struct amdgpu_irq_src_funcs dce_virtual_crtc_irq_funcs = {
+	.set = dce_virtual_set_crtc_irq_state,
+	.process = dce_virtual_crtc_irq,
+};
+
+static const struct amdgpu_irq_src_funcs dce_virtual_pageflip_irq_funcs = {
+	.set = dce_virtual_set_pageflip_irq_state,
+	.process = dce_virtual_pageflip_irq,
+};
+
+static void dce_virtual_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->crtc_irq.num_types = AMDGPU_CRTC_IRQ_LAST;
+	adev->crtc_irq.funcs = &dce_virtual_crtc_irq_funcs;
+
+	adev->pageflip_irq.num_types = AMDGPU_PAGEFLIP_IRQ_LAST;
+	adev->pageflip_irq.funcs = &dce_virtual_pageflip_irq_funcs;
+}
+

drivers/gpu/drm/amd/amdgpu/iceland_smum.h → drivers/gpu/drm/amd/amdgpu/dce_virtual.h

@@ -21,21 +21,11 @@
  *
  */
 
-#ifndef ICELAND_SMUM_H
-#define ICELAND_SMUM_H
+#ifndef __DCE_VIRTUAL_H__
+#define __DCE_VIRTUAL_H__
 
-#include "ppsmc.h"
-
-extern int iceland_smu_init(struct amdgpu_device *adev);
-extern int iceland_smu_fini(struct amdgpu_device *adev);
-extern int iceland_smu_start(struct amdgpu_device *adev);
-
-struct iceland_smu_private_data
-{
-	uint8_t *header;
-	uint8_t *mec_image;
-	uint32_t header_addr_high;
-	uint32_t header_addr_low;
-};
+extern const struct amd_ip_funcs dce_virtual_ip_funcs;
+#define DCE_VIRTUAL_VBLANK_PERIOD 16666666
 
 #endif
+

drivers/gpu/drm/amd/amdgpu/fiji_dpm.c

@@ -1,186 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#include <linux/firmware.h>
-#include "drmP.h"
-#include "amdgpu.h"
-#include "fiji_smum.h"
-
-MODULE_FIRMWARE("amdgpu/fiji_smc.bin");
-
-static void fiji_dpm_set_funcs(struct amdgpu_device *adev);
-
-static int fiji_dpm_early_init(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	fiji_dpm_set_funcs(adev);
-
-	return 0;
-}
-
-static int fiji_dpm_init_microcode(struct amdgpu_device *adev)
-{
-	char fw_name[30] = "amdgpu/fiji_smc.bin";
-	int err;
-
-	err = request_firmware(&adev->pm.fw, fw_name, adev->dev);
-	if (err)
-		goto out;
-	err = amdgpu_ucode_validate(adev->pm.fw);
-
-out:
-	if (err) {
-		DRM_ERROR("Failed to load firmware \"%s\"", fw_name);
-		release_firmware(adev->pm.fw);
-		adev->pm.fw = NULL;
-	}
-	return err;
-}
-
-static int fiji_dpm_sw_init(void *handle)
-{
-	int ret;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	ret = fiji_dpm_init_microcode(adev);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-static int fiji_dpm_sw_fini(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	release_firmware(adev->pm.fw);
-	adev->pm.fw = NULL;
-
-	return 0;
-}
-
-static int fiji_dpm_hw_init(void *handle)
-{
-	int ret;
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	mutex_lock(&adev->pm.mutex);
-
-	ret = fiji_smu_init(adev);
-	if (ret) {
-		DRM_ERROR("SMU initialization failed\n");
-		goto fail;
-	}
-
-	ret = fiji_smu_start(adev);
-	if (ret) {
-		DRM_ERROR("SMU start failed\n");
-		goto fail;
-	}
-
-	mutex_unlock(&adev->pm.mutex);
-	return 0;
-
-fail:
-	adev->firmware.smu_load = false;
-	mutex_unlock(&adev->pm.mutex);
-	return -EINVAL;
-}
-
-static int fiji_dpm_hw_fini(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	mutex_lock(&adev->pm.mutex);
-	fiji_smu_fini(adev);
-	mutex_unlock(&adev->pm.mutex);
-	return 0;
-}
-
-static int fiji_dpm_suspend(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	fiji_dpm_hw_fini(adev);
-
-	return 0;
-}
-
-static int fiji_dpm_resume(void *handle)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	fiji_dpm_hw_init(adev);
-
-	return 0;
-}
-
-static int fiji_dpm_set_clockgating_state(void *handle,
-			enum amd_clockgating_state state)
-{
-	return 0;
-}
-
-static int fiji_dpm_set_powergating_state(void *handle,
-			enum amd_powergating_state state)
-{
-	return 0;
-}
-
-const struct amd_ip_funcs fiji_dpm_ip_funcs = {
-	.name = "fiji_dpm",
-	.early_init = fiji_dpm_early_init,
-	.late_init = NULL,
-	.sw_init = fiji_dpm_sw_init,
-	.sw_fini = fiji_dpm_sw_fini,
-	.hw_init = fiji_dpm_hw_init,
-	.hw_fini = fiji_dpm_hw_fini,
-	.suspend = fiji_dpm_suspend,
-	.resume = fiji_dpm_resume,
-	.is_idle = NULL,
-	.wait_for_idle = NULL,
-	.soft_reset = NULL,
-	.set_clockgating_state = fiji_dpm_set_clockgating_state,
-	.set_powergating_state = fiji_dpm_set_powergating_state,
-};
-
-static const struct amdgpu_dpm_funcs fiji_dpm_funcs = {
-	.get_temperature = NULL,
-	.pre_set_power_state = NULL,
-	.set_power_state = NULL,
-	.post_set_power_state = NULL,
-	.display_configuration_changed = NULL,
-	.get_sclk = NULL,
-	.get_mclk = NULL,
-	.print_power_state = NULL,
-	.debugfs_print_current_performance_level = NULL,
-	.force_performance_level = NULL,
-	.vblank_too_short = NULL,
-	.powergate_uvd = NULL,
-};
-
-static void fiji_dpm_set_funcs(struct amdgpu_device *adev)
-{
-	if (NULL == adev->pm.funcs)
-		adev->pm.funcs = &fiji_dpm_funcs;
-}

drivers/gpu/drm/amd/amdgpu/fiji_smc.c

@@ -1,863 +0,0 @@
-/*
- * Copyright 2014 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- */
-
-#include <linux/firmware.h>
-#include "drmP.h"
-#include "amdgpu.h"
-#include "fiji_ppsmc.h"
-#include "fiji_smum.h"
-#include "smu_ucode_xfer_vi.h"
-#include "amdgpu_ucode.h"
-
-#include "smu/smu_7_1_3_d.h"
-#include "smu/smu_7_1_3_sh_mask.h"
-
-#define FIJI_SMC_SIZE 0x20000
-
-static int fiji_set_smc_sram_address(struct amdgpu_device *adev, uint32_t smc_address, uint32_t limit)
-{
-	uint32_t val;
-
-	if (smc_address & 3)
-		return -EINVAL;
-
-	if ((smc_address + 3) > limit)
-		return -EINVAL;
-
-	WREG32(mmSMC_IND_INDEX_0, smc_address);
-
-	val = RREG32(mmSMC_IND_ACCESS_CNTL);
-	val = REG_SET_FIELD(val, SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, 0);
-	WREG32(mmSMC_IND_ACCESS_CNTL, val);
-
-	return 0;
-}
-
-static int fiji_copy_bytes_to_smc(struct amdgpu_device *adev, uint32_t smc_start_address, const uint8_t *src, uint32_t byte_count, uint32_t limit)
-{
-	uint32_t addr;
-	uint32_t data, orig_data;
-	int result = 0;
-	uint32_t extra_shift;
-	unsigned long flags;
-
-	if (smc_start_address & 3)
-		return -EINVAL;
-
-	if ((smc_start_address + byte_count) > limit)
-		return -EINVAL;
-
-	addr = smc_start_address;
-
-	spin_lock_irqsave(&adev->smc_idx_lock, flags);
-	while (byte_count >= 4) {
-		/* Bytes are written into the SMC addres space with the MSB first */
-		data = (src[0] << 24) + (src[1] << 16) + (src[2] << 8) + src[3];
-
-		result = fiji_set_smc_sram_address(adev, addr, limit);
-
-		if (result)
-			goto out;
-
-		WREG32(mmSMC_IND_DATA_0, data);
-
-		src += 4;
-		byte_count -= 4;
-		addr += 4;
-	}
-
-	if (0 != byte_count) {
-		/* Now write odd bytes left, do a read modify write cycle */
-		data = 0;
-
-		result = fiji_set_smc_sram_address(adev, addr, limit);
-		if (result)
-			goto out;
-
-		orig_data = RREG32(mmSMC_IND_DATA_0);
-		extra_shift = 8 * (4 - byte_count);
-
-		while (byte_count > 0) {
-			data = (data << 8) + *src++;
-			byte_count--;
-		}
-
-		data <<= extra_shift;
-		data |= (orig_data & ~((~0UL) << extra_shift));
-
-		result = fiji_set_smc_sram_address(adev, addr, limit);
-		if (result)
-			goto out;
-
-		WREG32(mmSMC_IND_DATA_0, data);
-	}
-
-out:
-	spin_unlock_irqrestore(&adev->smc_idx_lock, flags);
-	return result;
-}
-
-static int fiji_program_jump_on_start(struct amdgpu_device *adev)
-{
-	static unsigned char data[] = {0xE0, 0x00, 0x80, 0x40};
-	fiji_copy_bytes_to_smc(adev, 0x0, data, 4, sizeof(data)+1);
-
-	return 0;
-}
-
-static bool fiji_is_smc_ram_running(struct amdgpu_device *adev)
-{
-	uint32_t val = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0);
-	val = REG_GET_FIELD(val, SMC_SYSCON_CLOCK_CNTL_0, ck_disable);
-
-	return ((0 == val) && (0x20100 <= RREG32_SMC(ixSMC_PC_C)));
-}
-
-static int wait_smu_response(struct amdgpu_device *adev)
-{
-	int i;
-	uint32_t val;
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		val = RREG32(mmSMC_RESP_0);
-		if (REG_GET_FIELD(val, SMC_RESP_0, SMC_RESP))
-			break;
-		udelay(1);
-	}
-
-	if (i == adev->usec_timeout)
-		return -EINVAL;
-
-	return 0;
-}
-
-static int fiji_send_msg_to_smc_offset(struct amdgpu_device *adev)
-{
-	if (wait_smu_response(adev)) {
-		DRM_ERROR("Failed to send previous message\n");
-		return -EINVAL;
-	}
-
-	WREG32(mmSMC_MSG_ARG_0, 0x20000);
-	WREG32(mmSMC_MESSAGE_0, PPSMC_MSG_Test);
-
-	if (wait_smu_response(adev)) {
-		DRM_ERROR("Failed to send message\n");
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int fiji_send_msg_to_smc(struct amdgpu_device *adev, PPSMC_Msg msg)
-{
-	if (!fiji_is_smc_ram_running(adev))
-	{
-		return -EINVAL;
-	}
-
-	if (wait_smu_response(adev)) {
-		DRM_ERROR("Failed to send previous message\n");
-		return -EINVAL;
-	}
-
-	WREG32(mmSMC_MESSAGE_0, msg);
-
-	if (wait_smu_response(adev)) {
-		DRM_ERROR("Failed to send message\n");
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int fiji_send_msg_to_smc_without_waiting(struct amdgpu_device *adev,
-						PPSMC_Msg msg)
-{
-	if (wait_smu_response(adev)) {
-		DRM_ERROR("Failed to send previous message\n");
-		return -EINVAL;
-	}
-
-	WREG32(mmSMC_MESSAGE_0, msg);
-
-	return 0;
-}
-
-static int fiji_send_msg_to_smc_with_parameter(struct amdgpu_device *adev,
-						PPSMC_Msg msg,
-						uint32_t parameter)
-{
-	if (!fiji_is_smc_ram_running(adev))
-		return -EINVAL;
-
-	if (wait_smu_response(adev)) {
-		DRM_ERROR("Failed to send previous message\n");
-		return -EINVAL;
-	}
-
-	WREG32(mmSMC_MSG_ARG_0, parameter);
-
-	return fiji_send_msg_to_smc(adev, msg);
-}
-
-static int fiji_send_msg_to_smc_with_parameter_without_waiting(
-					struct amdgpu_device *adev,
-					PPSMC_Msg msg, uint32_t parameter)
-{
-	if (wait_smu_response(adev)) {
-		DRM_ERROR("Failed to send previous message\n");
-		return -EINVAL;
-	}
-
-	WREG32(mmSMC_MSG_ARG_0, parameter);
-
-	return fiji_send_msg_to_smc_without_waiting(adev, msg);
-}
-
-#if 0 /* not used yet */
-static int fiji_wait_for_smc_inactive(struct amdgpu_device *adev)
-{
-	int i;
-	uint32_t val;
-
-	if (!fiji_is_smc_ram_running(adev))
-		return -EINVAL;
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		val = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0);
-		if (REG_GET_FIELD(val, SMC_SYSCON_CLOCK_CNTL_0, cken) == 0)
-			break;
-		udelay(1);
-	}
-
-	if (i == adev->usec_timeout)
-		return -EINVAL;
-
-	return 0;
-}
-#endif
-
-static int fiji_smu_upload_firmware_image(struct amdgpu_device *adev)
-{
-	const struct smc_firmware_header_v1_0 *hdr;
-	uint32_t ucode_size;
-	uint32_t ucode_start_address;
-	const uint8_t *src;
-	uint32_t val;
-	uint32_t byte_count;
-	uint32_t *data;
-	unsigned long flags;
-
-	if (!adev->pm.fw)
-		return -EINVAL;
-
-	/* Skip SMC ucode loading on SR-IOV capable boards.
-	 * vbios does this for us in asic_init in that case.
-	 */
-	if (adev->virtualization.supports_sr_iov)
-		return 0;
-
-	hdr = (const struct smc_firmware_header_v1_0 *)adev->pm.fw->data;
-	amdgpu_ucode_print_smc_hdr(&hdr->header);
-
-	adev->pm.fw_version = le32_to_cpu(hdr->header.ucode_version);
-	ucode_size = le32_to_cpu(hdr->header.ucode_size_bytes);
-	ucode_start_address = le32_to_cpu(hdr->ucode_start_addr);
-	src = (const uint8_t *)
-		(adev->pm.fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
-
-	if (ucode_size & 3) {
-		DRM_ERROR("SMC ucode is not 4 bytes aligned\n");
-		return -EINVAL;
-	}
-
-	if (ucode_size > FIJI_SMC_SIZE) {
-		DRM_ERROR("SMC address is beyond the SMC RAM area\n");
-		return -EINVAL;
-	}
-
-	spin_lock_irqsave(&adev->smc_idx_lock, flags);
-	WREG32(mmSMC_IND_INDEX_0, ucode_start_address);
-
-	val = RREG32(mmSMC_IND_ACCESS_CNTL);
-	val = REG_SET_FIELD(val, SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, 1);
-	WREG32(mmSMC_IND_ACCESS_CNTL, val);
-
-	byte_count = ucode_size;
-	data = (uint32_t *)src;
-	for (; byte_count >= 4; data++, byte_count -= 4)
-		WREG32(mmSMC_IND_DATA_0, data[0]);
-
-	val = RREG32(mmSMC_IND_ACCESS_CNTL);
-	val = REG_SET_FIELD(val, SMC_IND_ACCESS_CNTL, AUTO_INCREMENT_IND_0, 0);
-	WREG32(mmSMC_IND_ACCESS_CNTL, val);
-	spin_unlock_irqrestore(&adev->smc_idx_lock, flags);
-
-	return 0;
-}
-
-#if 0 /* not used yet */
-static int fiji_read_smc_sram_dword(struct amdgpu_device *adev,
-				uint32_t smc_address,
-				uint32_t *value,
-				uint32_t limit)
-{
-	int result;
-	unsigned long flags;
-
-	spin_lock_irqsave(&adev->smc_idx_lock, flags);
-	result = fiji_set_smc_sram_address(adev, smc_address, limit);
-	if (result == 0)
-		*value = RREG32(mmSMC_IND_DATA_0);
-	spin_unlock_irqrestore(&adev->smc_idx_lock, flags);
-	return result;
-}
-
-static int fiji_write_smc_sram_dword(struct amdgpu_device *adev,
-				uint32_t smc_address,
-				uint32_t value,
-				uint32_t limit)
-{
-	int result;
-	unsigned long flags;
-
-	spin_lock_irqsave(&adev->smc_idx_lock, flags);
-	result = fiji_set_smc_sram_address(adev, smc_address, limit);
-	if (result == 0)
-		WREG32(mmSMC_IND_DATA_0, value);
-	spin_unlock_irqrestore(&adev->smc_idx_lock, flags);
-	return result;
-}
-
-static int fiji_smu_stop_smc(struct amdgpu_device *adev)
-{
-	uint32_t val = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL);
-	val = REG_SET_FIELD(val, SMC_SYSCON_RESET_CNTL, rst_reg, 1);
-	WREG32_SMC(ixSMC_SYSCON_RESET_CNTL, val);
-
-	val = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0);
-	val = REG_SET_FIELD(val, SMC_SYSCON_CLOCK_CNTL_0, ck_disable, 1);
-	WREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0, val);
-
-	return 0;
-}
-#endif
-
-static enum AMDGPU_UCODE_ID fiji_convert_fw_type(uint32_t fw_type)
-{
-	switch (fw_type) {
-		case UCODE_ID_SDMA0:
-			return AMDGPU_UCODE_ID_SDMA0;
-		case UCODE_ID_SDMA1:
-			return AMDGPU_UCODE_ID_SDMA1;
-		case UCODE_ID_CP_CE:
-			return AMDGPU_UCODE_ID_CP_CE;
-		case UCODE_ID_CP_PFP:
-			return AMDGPU_UCODE_ID_CP_PFP;
-		case UCODE_ID_CP_ME:
-			return AMDGPU_UCODE_ID_CP_ME;
-		case UCODE_ID_CP_MEC:
-		case UCODE_ID_CP_MEC_JT1:
-		case UCODE_ID_CP_MEC_JT2:
-			return AMDGPU_UCODE_ID_CP_MEC1;
-		case UCODE_ID_RLC_G:
-			return AMDGPU_UCODE_ID_RLC_G;
-		default:
-			DRM_ERROR("ucode type is out of range!\n");
-			return AMDGPU_UCODE_ID_MAXIMUM;
-	}
-}
-
-static int fiji_smu_populate_single_firmware_entry(struct amdgpu_device *adev,
-						uint32_t fw_type,
-						struct SMU_Entry *entry)
-{
-	enum AMDGPU_UCODE_ID id = fiji_convert_fw_type(fw_type);
-	struct amdgpu_firmware_info *ucode = &adev->firmware.ucode[id];
-	const struct gfx_firmware_header_v1_0 *header = NULL;
-	uint64_t gpu_addr;
-	uint32_t data_size;
-
-	if (ucode->fw == NULL)
-		return -EINVAL;
-	gpu_addr  = ucode->mc_addr;
-	header = (const struct gfx_firmware_header_v1_0 *)ucode->fw->data;
-	data_size = le32_to_cpu(header->header.ucode_size_bytes);
-
-	if ((fw_type == UCODE_ID_CP_MEC_JT1) ||
-		(fw_type == UCODE_ID_CP_MEC_JT2)) {
-		gpu_addr += le32_to_cpu(header->jt_offset) << 2;
-		data_size = le32_to_cpu(header->jt_size) << 2;
-	}
-
-	entry->version = (uint16_t)le32_to_cpu(header->header.ucode_version);
-	entry->id = (uint16_t)fw_type;
-	entry->image_addr_high = upper_32_bits(gpu_addr);
-	entry->image_addr_low = lower_32_bits(gpu_addr);
-	entry->meta_data_addr_high = 0;
-	entry->meta_data_addr_low = 0;
-	entry->data_size_byte = data_size;
-	entry->num_register_entries = 0;
-
-	if (fw_type == UCODE_ID_RLC_G)
-		entry->flags = 1;
-	else
-		entry->flags = 0;
-
-	return 0;
-}
-
-static int fiji_smu_request_load_fw(struct amdgpu_device *adev)
-{
-	struct fiji_smu_private_data *private = (struct fiji_smu_private_data *)adev->smu.priv;
-	struct SMU_DRAMData_TOC *toc;
-	uint32_t fw_to_load;
-
-	WREG32_SMC(ixSOFT_REGISTERS_TABLE_28, 0);
-
-	fiji_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_SMU_DRAM_ADDR_HI, private->smu_buffer_addr_high);
-	fiji_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_SMU_DRAM_ADDR_LO, private->smu_buffer_addr_low);
-
-	toc = (struct SMU_DRAMData_TOC *)private->header;
-	toc->num_entries = 0;
-	toc->structure_version = 1;
-
-	if (!adev->firmware.smu_load)
-		return 0;
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_RLC_G,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for RLC\n");
-		return -EINVAL;
-	}
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_CP_CE,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for CE\n");
-		return -EINVAL;
-	}
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_CP_PFP,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for PFP\n");
-		return -EINVAL;
-	}
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_CP_ME,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for ME\n");
-		return -EINVAL;
-	}
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_CP_MEC,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for MEC\n");
-		return -EINVAL;
-	}
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_CP_MEC_JT1,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for MEC_JT1\n");
-		return -EINVAL;
-	}
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_CP_MEC_JT2,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for MEC_JT2\n");
-		return -EINVAL;
-	}
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_SDMA0,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for SDMA0\n");
-		return -EINVAL;
-	}
-
-	if (fiji_smu_populate_single_firmware_entry(adev, UCODE_ID_SDMA1,
-			&toc->entry[toc->num_entries++])) {
-		DRM_ERROR("Failed to get firmware entry for SDMA1\n");
-		return -EINVAL;
-	}
-
-	fiji_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_DRV_DRAM_ADDR_HI, private->header_addr_high);
-	fiji_send_msg_to_smc_with_parameter(adev, PPSMC_MSG_DRV_DRAM_ADDR_LO, private->header_addr_low);
-
-	fw_to_load = UCODE_ID_RLC_G_MASK |
-			UCODE_ID_SDMA0_MASK |
-			UCODE_ID_SDMA1_MASK |
-			UCODE_ID_CP_CE_MASK |
-			UCODE_ID_CP_ME_MASK |
-			UCODE_ID_CP_PFP_MASK |
-			UCODE_ID_CP_MEC_MASK;
-
-	if (fiji_send_msg_to_smc_with_parameter_without_waiting(adev, PPSMC_MSG_LoadUcodes, fw_to_load)) {
-		DRM_ERROR("Fail to request SMU load ucode\n");
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static uint32_t fiji_smu_get_mask_for_fw_type(uint32_t fw_type)
-{
-	switch (fw_type) {
-		case AMDGPU_UCODE_ID_SDMA0:
-			return UCODE_ID_SDMA0_MASK;
-		case AMDGPU_UCODE_ID_SDMA1:
-			return UCODE_ID_SDMA1_MASK;
-		case AMDGPU_UCODE_ID_CP_CE:
-			return UCODE_ID_CP_CE_MASK;
-		case AMDGPU_UCODE_ID_CP_PFP:
-			return UCODE_ID_CP_PFP_MASK;
-		case AMDGPU_UCODE_ID_CP_ME:
-			return UCODE_ID_CP_ME_MASK;
-		case AMDGPU_UCODE_ID_CP_MEC1:
-			return UCODE_ID_CP_MEC_MASK;
-		case AMDGPU_UCODE_ID_CP_MEC2:
-			return UCODE_ID_CP_MEC_MASK;
-		case AMDGPU_UCODE_ID_RLC_G:
-			return UCODE_ID_RLC_G_MASK;
-		default:
-			DRM_ERROR("ucode type is out of range!\n");
-			return 0;
-	}
-}
-
-static int fiji_smu_check_fw_load_finish(struct amdgpu_device *adev,
-					uint32_t fw_type)
-{
-	uint32_t fw_mask = fiji_smu_get_mask_for_fw_type(fw_type);
-	int i;
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		if (fw_mask == (RREG32_SMC(ixSOFT_REGISTERS_TABLE_28) & fw_mask))
-			break;
-		udelay(1);
-	}
-
-	if (i == adev->usec_timeout) {
-		DRM_ERROR("check firmware loading failed\n");
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int fiji_smu_start_in_protection_mode(struct amdgpu_device *adev)
-{
-	int result;
-	uint32_t val;
-	int i;
-
-	/* Assert reset */
-	val = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL);
-	val = REG_SET_FIELD(val, SMC_SYSCON_RESET_CNTL, rst_reg, 1);
-	WREG32_SMC(ixSMC_SYSCON_RESET_CNTL, val);
-
-	result = fiji_smu_upload_firmware_image(adev);
-	if (result)
-		return result;
-
-	/* Clear status */
-	WREG32_SMC(ixSMU_STATUS, 0);
-
-	/* Enable clock */
-	val = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0);
-	val = REG_SET_FIELD(val, SMC_SYSCON_CLOCK_CNTL_0, ck_disable, 0);
-	WREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0, val);
-
-	/* De-assert reset */
-	val = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL);
-	val = REG_SET_FIELD(val, SMC_SYSCON_RESET_CNTL, rst_reg, 0);
-	WREG32_SMC(ixSMC_SYSCON_RESET_CNTL, val);
-
-	/* Set SMU Auto Start */
-	val = RREG32_SMC(ixSMU_INPUT_DATA);
-	val = REG_SET_FIELD(val, SMU_INPUT_DATA, AUTO_START, 1);
-	WREG32_SMC(ixSMU_INPUT_DATA, val);
-
-	/* Clear firmware interrupt enable flag */
-	WREG32_SMC(ixFIRMWARE_FLAGS, 0);
-
-	for (i = 0; i < adev->usec_timeout; i++) {
-		val = RREG32_SMC(ixRCU_UC_EVENTS);
-		if (REG_GET_FIELD(val, RCU_UC_EVENTS, INTERRUPTS_ENABLED))
-			break;
-		udelay(1);
-	}
-
-	if (i == adev->usec_timeout) {
-		DRM_ERROR("Interrupt is not enabled by firmware\n");
-		return -EINVAL;
-	}
-
-	/* Call Test SMU message with 0x20000 offset
-	 * to trigger SMU start
-	 */
-	fiji_send_msg_to_smc_offset(adev);
-	DRM_INFO("[FM]try triger smu start\n");
-	/* Wait for done bit to be set */
-	for (i = 0; i < adev->usec_timeout; i++) {
-		val = RREG32_SMC(ixSMU_STATUS);
-		if (REG_GET_FIELD(val, SMU_STATUS, SMU_DONE))
-			break;
-		udelay(1);
-	}
-
-	if (i == adev->usec_timeout) {
-		DRM_ERROR("Timeout for SMU start\n");
-		return -EINVAL;
-	}
-
-	/* Check pass/failed indicator */
-	val = RREG32_SMC(ixSMU_STATUS);
-	if (!REG_GET_FIELD(val, SMU_STATUS, SMU_PASS)) {
-		DRM_ERROR("SMU Firmware start failed\n");
-		return -EINVAL;
-	}
-	DRM_INFO("[FM]smu started\n");
-	/* Wait for firmware to initialize */
-	for (i = 0; i < adev->usec_timeout; i++) {
-		val = RREG32_SMC(ixFIRMWARE_FLAGS);
-		if(REG_GET_FIELD(val, FIRMWARE_FLAGS, INTERRUPTS_ENABLED))
-			break;
-		udelay(1);
-	}
-
-	if (i == adev->usec_timeout) {
-		DRM_ERROR("SMU firmware initialization failed\n");
-		return -EINVAL;
-	}
-	DRM_INFO("[FM]smu initialized\n");
-
-	return 0;
-}
-
-static int fiji_smu_start_in_non_protection_mode(struct amdgpu_device *adev)
-{
-	int i, result;
-	uint32_t val;
-
-	/* wait for smc boot up */
-	for (i = 0; i < adev->usec_timeout; i++) {
-		val = RREG32_SMC(ixRCU_UC_EVENTS);
-		val = REG_GET_FIELD(val, RCU_UC_EVENTS, boot_seq_done);
-		if (val)
-			break;
-		udelay(1);
-	}
-
-	if (i == adev->usec_timeout) {
-		DRM_ERROR("SMC boot sequence is not completed\n");
-		return -EINVAL;
-	}
-
-	/* Clear firmware interrupt enable flag */
-	WREG32_SMC(ixFIRMWARE_FLAGS, 0);
-
-	/* Assert reset */
-	val = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL);
-	val = REG_SET_FIELD(val, SMC_SYSCON_RESET_CNTL, rst_reg, 1);
-	WREG32_SMC(ixSMC_SYSCON_RESET_CNTL, val);
-
-	result = fiji_smu_upload_firmware_image(adev);
-	if (result)
-		return result;
-
-	/* Set smc instruct start point at 0x0 */
-	fiji_program_jump_on_start(adev);
-
-	/* Enable clock */
-	val = RREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0);
-	val = REG_SET_FIELD(val, SMC_SYSCON_CLOCK_CNTL_0, ck_disable, 0);
-	WREG32_SMC(ixSMC_SYSCON_CLOCK_CNTL_0, val);
-
-	/* De-assert reset */
-	val = RREG32_SMC(ixSMC_SYSCON_RESET_CNTL);
-	val = REG_SET_FIELD(val, SMC_SYSCON_RESET_CNTL, rst_reg, 0);
-	WREG32_SMC(ixSMC_SYSCON_RESET_CNTL, val);
-
-	/* Wait for firmware to initialize */
-	for (i = 0; i < adev->usec_timeout; i++) {
-		val = RREG32_SMC(ixFIRMWARE_FLAGS);
-		if (REG_GET_FIELD(val, FIRMWARE_FLAGS, INTERRUPTS_ENABLED))
-			break;
-		udelay(1);
-	}
-
-	if (i == adev->usec_timeout) {
-		DRM_ERROR("Timeout for SMC firmware initialization\n");
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-int fiji_smu_start(struct amdgpu_device *adev)
-{
-	int result;
-	uint32_t val;
-
-	if (!fiji_is_smc_ram_running(adev)) {
-		val = RREG32_SMC(ixSMU_FIRMWARE);
-		if (!REG_GET_FIELD(val, SMU_FIRMWARE, SMU_MODE)) {
-			DRM_INFO("[FM]start smu in nonprotection mode\n");
-			result = fiji_smu_start_in_non_protection_mode(adev);
-			if (result)
-				return result;
-		} else {
-			DRM_INFO("[FM]start smu in protection mode\n");
-			result = fiji_smu_start_in_protection_mode(adev);
-			if (result)
-				return result;
-		}
-	}
-
-	return fiji_smu_request_load_fw(adev);
-}
-
-static const struct amdgpu_smumgr_funcs fiji_smumgr_funcs = {
-	.check_fw_load_finish = fiji_smu_check_fw_load_finish,
-	.request_smu_load_fw = NULL,
-	.request_smu_specific_fw = NULL,
-};
-
-int fiji_smu_init(struct amdgpu_device *adev)
-{
-	struct fiji_smu_private_data *private;
-	uint32_t image_size = ((sizeof(struct SMU_DRAMData_TOC) / 4096) + 1) * 4096;
-	uint32_t smu_internal_buffer_size = 200*4096;
-	struct amdgpu_bo **toc_buf = &adev->smu.toc_buf;
-	struct amdgpu_bo **smu_buf = &adev->smu.smu_buf;
-	uint64_t mc_addr;
-	void *toc_buf_ptr;
-	void *smu_buf_ptr;
-	int ret;
-
-	private = kzalloc(sizeof(struct fiji_smu_private_data), GFP_KERNEL);
-	if (NULL == private)
-		return -ENOMEM;
-
-	/* allocate firmware buffers */
-	if (adev->firmware.smu_load)
-		amdgpu_ucode_init_bo(adev);
-
-	adev->smu.priv = private;
-	adev->smu.fw_flags = 0;
-
-	/* Allocate FW image data structure and header buffer */
-	ret = amdgpu_bo_create(adev, image_size, PAGE_SIZE,
-			       true, AMDGPU_GEM_DOMAIN_VRAM,
-			       AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
-			       NULL, NULL, toc_buf);
-	if (ret) {
-		DRM_ERROR("Failed to allocate memory for TOC buffer\n");
-		return -ENOMEM;
-	}
-
-	/* Allocate buffer for SMU internal buffer */
-	ret = amdgpu_bo_create(adev, smu_internal_buffer_size, PAGE_SIZE,
-			       true, AMDGPU_GEM_DOMAIN_VRAM,
-			       AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
-			       NULL, NULL, smu_buf);
-	if (ret) {
-		DRM_ERROR("Failed to allocate memory for SMU internal buffer\n");
-		return -ENOMEM;
-	}
-
-	/* Retrieve GPU address for header buffer and internal buffer */
-	ret = amdgpu_bo_reserve(adev->smu.toc_buf, false);
-	if (ret) {
-		amdgpu_bo_unref(&adev->smu.toc_buf);
-		DRM_ERROR("Failed to reserve the TOC buffer\n");
-		return -EINVAL;
-	}
-
-	ret = amdgpu_bo_pin(adev->smu.toc_buf, AMDGPU_GEM_DOMAIN_VRAM, &mc_addr);
-	if (ret) {
-		amdgpu_bo_unreserve(adev->smu.toc_buf);
-		amdgpu_bo_unref(&adev->smu.toc_buf);
-		DRM_ERROR("Failed to pin the TOC buffer\n");
-		return -EINVAL;
-	}
-
-	ret = amdgpu_bo_kmap(*toc_buf, &toc_buf_ptr);
-	if (ret) {
-		amdgpu_bo_unreserve(adev->smu.toc_buf);
-		amdgpu_bo_unref(&adev->smu.toc_buf);
-		DRM_ERROR("Failed to map the TOC buffer\n");
-		return -EINVAL;
-	}
-
-	amdgpu_bo_unreserve(adev->smu.toc_buf);
-	private->header_addr_low = lower_32_bits(mc_addr);
-	private->header_addr_high = upper_32_bits(mc_addr);
-	private->header = toc_buf_ptr;
-
-	ret = amdgpu_bo_reserve(adev->smu.smu_buf, false);
-	if (ret) {
-		amdgpu_bo_unref(&adev->smu.smu_buf);
-		amdgpu_bo_unref(&adev->smu.toc_buf);
-		DRM_ERROR("Failed to reserve the SMU internal buffer\n");
-		return -EINVAL;
-	}
-
-	ret = amdgpu_bo_pin(adev->smu.smu_buf, AMDGPU_GEM_DOMAIN_VRAM, &mc_addr);
-	if (ret) {
-		amdgpu_bo_unreserve(adev->smu.smu_buf);
-		amdgpu_bo_unref(&adev->smu.smu_buf);
-		amdgpu_bo_unref(&adev->smu.toc_buf);
-		DRM_ERROR("Failed to pin the SMU internal buffer\n");
-		return -EINVAL;
-	}
-
-	ret = amdgpu_bo_kmap(*smu_buf, &smu_buf_ptr);
-	if (ret) {
-		amdgpu_bo_unreserve(adev->smu.smu_buf);
-		amdgpu_bo_unref(&adev->smu.smu_buf);
-		amdgpu_bo_unref(&adev->smu.toc_buf);
-		DRM_ERROR("Failed to map the SMU internal buffer\n");
-		return -EINVAL;
-	}
-
-	amdgpu_bo_unreserve(adev->smu.smu_buf);
-	private->smu_buffer_addr_low = lower_32_bits(mc_addr);
-	private->smu_buffer_addr_high = upper_32_bits(mc_addr);
-
-	adev->smu.smumgr_funcs = &fiji_smumgr_funcs;
-
-	return 0;
-}
-
-int fiji_smu_fini(struct amdgpu_device *adev)
-{
-	amdgpu_bo_unref(&adev->smu.toc_buf);
-	amdgpu_bo_unref(&adev->smu.smu_buf);
-	kfree(adev->smu.priv);
-	adev->smu.priv = NULL;
-	if (adev->firmware.fw_buf)
-		amdgpu_ucode_fini_bo(adev);
-
-	return 0;
-}

drivers/gpu/drm/amd/amdgpu/gfx_v6_0.c

@@ -0,0 +1,3362 @@
+/*
+ * Copyright 2015 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+#include <linux/firmware.h>
+#include "amdgpu.h"
+#include "amdgpu_ih.h"
+#include "amdgpu_gfx.h"
+#include "amdgpu_ucode.h"
+#include "si/clearstate_si.h"
+#include "si/sid.h"
+
+#define GFX6_NUM_GFX_RINGS     1
+#define GFX6_NUM_COMPUTE_RINGS 2
+#define STATIC_PER_CU_PG_ENABLE                    (1 << 3)
+#define DYN_PER_CU_PG_ENABLE                       (1 << 2)
+#define RLC_SAVE_AND_RESTORE_STARTING_OFFSET 0x90
+#define RLC_CLEAR_STATE_DESCRIPTOR_OFFSET    0x3D
+
+
+static void gfx_v6_0_set_ring_funcs(struct amdgpu_device *adev);
+static void gfx_v6_0_set_irq_funcs(struct amdgpu_device *adev);
+static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev);
+
+MODULE_FIRMWARE("radeon/tahiti_pfp.bin");
+MODULE_FIRMWARE("radeon/tahiti_me.bin");
+MODULE_FIRMWARE("radeon/tahiti_ce.bin");
+MODULE_FIRMWARE("radeon/tahiti_rlc.bin");
+
+MODULE_FIRMWARE("radeon/pitcairn_pfp.bin");
+MODULE_FIRMWARE("radeon/pitcairn_me.bin");
+MODULE_FIRMWARE("radeon/pitcairn_ce.bin");
+MODULE_FIRMWARE("radeon/pitcairn_rlc.bin");
+
+MODULE_FIRMWARE("radeon/verde_pfp.bin");
+MODULE_FIRMWARE("radeon/verde_me.bin");
+MODULE_FIRMWARE("radeon/verde_ce.bin");
+MODULE_FIRMWARE("radeon/verde_rlc.bin");
+
+MODULE_FIRMWARE("radeon/oland_pfp.bin");
+MODULE_FIRMWARE("radeon/oland_me.bin");
+MODULE_FIRMWARE("radeon/oland_ce.bin");
+MODULE_FIRMWARE("radeon/oland_rlc.bin");
+
+MODULE_FIRMWARE("radeon/hainan_pfp.bin");
+MODULE_FIRMWARE("radeon/hainan_me.bin");
+MODULE_FIRMWARE("radeon/hainan_ce.bin");
+MODULE_FIRMWARE("radeon/hainan_rlc.bin");
+
+static u32 gfx_v6_0_get_csb_size(struct amdgpu_device *adev);
+static void gfx_v6_0_get_csb_buffer(struct amdgpu_device *adev, volatile u32 *buffer);
+//static void gfx_v6_0_init_cp_pg_table(struct amdgpu_device *adev);
+static void gfx_v6_0_init_pg(struct amdgpu_device *adev);
+
+
+static const u32 verde_rlc_save_restore_register_list[] =
+{
+	(0x8000 << 16) | (0x98f4 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x98f4 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0xe80 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0xe80 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x89bc >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x89bc >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x8c1c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x8c1c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x98f0 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xe7c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9148 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9148 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9150 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x897c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8d8c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac54 >> 2),
+	0X00000000,
+	0x3,
+	(0x9c00 << 16) | (0x98f8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9910 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9914 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9918 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x991c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9920 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9924 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9928 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x992c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9930 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9934 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9938 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x993c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9940 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9944 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9948 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x994c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9950 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9954 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9958 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x995c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9960 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9964 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9968 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x996c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9970 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9974 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9978 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x997c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9980 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9984 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9988 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x998c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c00 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c04 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c08 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9b7c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9b7c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0xe84 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0xe84 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x89c0 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x89c0 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x914c >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x914c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x8c20 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x8c20 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9354 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9354 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9060 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9364 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9100 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x913c >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e0 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e4 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x90e8 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e0 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e4 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x90e8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8bcc >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8b24 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88c4 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e50 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8c0c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e58 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8e5c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9508 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x950c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9494 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac0c >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac10 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xae00 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0xac08 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88d4 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88c8 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x88cc >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x89b0 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8b10 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x8a14 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9830 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9834 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9838 >> 2),
+	0x00000000,
+	(0x9c00 << 16) | (0x9a10 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8000 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8001 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8001 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8040 << 16) | (0x9874 >> 2),
+	0x00000000,
+	(0x8041 << 16) | (0x9870 >> 2),
+	0x00000000,
+	(0x8041 << 16) | (0x9874 >> 2),
+	0x00000000,
+	0x00000000
+};
+
+static int gfx_v6_0_init_microcode(struct amdgpu_device *adev)
+{
+	const char *chip_name;
+	char fw_name[30];
+	int err;
+	const struct gfx_firmware_header_v1_0 *cp_hdr;
+	const struct rlc_firmware_header_v1_0 *rlc_hdr;
+
+	DRM_DEBUG("\n");
+
+	switch (adev->asic_type) {
+	case CHIP_TAHITI:
+		chip_name = "tahiti";
+		break;
+	case CHIP_PITCAIRN:
+		chip_name = "pitcairn";
+		break;
+	case CHIP_VERDE:
+		chip_name = "verde";
+		break;
+	case CHIP_OLAND:
+		chip_name = "oland";
+		break;
+	case CHIP_HAINAN:
+		chip_name = "hainan";
+		break;
+	default: BUG();
+	}
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_pfp.bin", chip_name);
+	err = request_firmware(&adev->gfx.pfp_fw, fw_name, adev->dev);
+	if (err)
+		goto out;
+	err = amdgpu_ucode_validate(adev->gfx.pfp_fw);
+	if (err)
+		goto out;
+	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data;
+	adev->gfx.pfp_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
+	adev->gfx.pfp_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_me.bin", chip_name);
+	err = request_firmware(&adev->gfx.me_fw, fw_name, adev->dev);
+	if (err)
+		goto out;
+	err = amdgpu_ucode_validate(adev->gfx.me_fw);
+	if (err)
+		goto out;
+	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data;
+	adev->gfx.me_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
+	adev->gfx.me_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_ce.bin", chip_name);
+	err = request_firmware(&adev->gfx.ce_fw, fw_name, adev->dev);
+	if (err)
+		goto out;
+	err = amdgpu_ucode_validate(adev->gfx.ce_fw);
+	if (err)
+		goto out;
+	cp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data;
+	adev->gfx.ce_fw_version = le32_to_cpu(cp_hdr->header.ucode_version);
+	adev->gfx.ce_feature_version = le32_to_cpu(cp_hdr->ucode_feature_version);
+
+	snprintf(fw_name, sizeof(fw_name), "radeon/%s_rlc.bin", chip_name);
+	err = request_firmware(&adev->gfx.rlc_fw, fw_name, adev->dev);
+	if (err)
+		goto out;
+	err = amdgpu_ucode_validate(adev->gfx.rlc_fw);
+	rlc_hdr = (const struct rlc_firmware_header_v1_0 *)adev->gfx.rlc_fw->data;
+	adev->gfx.rlc_fw_version = le32_to_cpu(rlc_hdr->header.ucode_version);
+	adev->gfx.rlc_feature_version = le32_to_cpu(rlc_hdr->ucode_feature_version);
+
+out:
+	if (err) {
+		printk(KERN_ERR
+		       "gfx6: Failed to load firmware \"%s\"\n",
+		       fw_name);
+		release_firmware(adev->gfx.pfp_fw);
+		adev->gfx.pfp_fw = NULL;
+		release_firmware(adev->gfx.me_fw);
+		adev->gfx.me_fw = NULL;
+		release_firmware(adev->gfx.ce_fw);
+		adev->gfx.ce_fw = NULL;
+		release_firmware(adev->gfx.rlc_fw);
+		adev->gfx.rlc_fw = NULL;
+	}
+	return err;
+}
+
+static void gfx_v6_0_tiling_mode_table_init(struct amdgpu_device *adev)
+{
+	const u32 num_tile_mode_states = 32;
+	u32 reg_offset, gb_tile_moden, split_equal_to_row_size;
+
+	switch (adev->gfx.config.mem_row_size_in_kb) {
+	case 1:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_1KB;
+		break;
+	case 2:
+	default:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_2KB;
+		break;
+	case 4:
+		split_equal_to_row_size = ADDR_SURF_TILE_SPLIT_4KB;
+		break;
+	}
+
+	if (adev->asic_type == CHIP_VERDE ||
+		adev->asic_type == CHIP_OLAND ||
+		adev->asic_type == CHIP_HAINAN) {
+		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 1: 
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 2:
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 3:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 4:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 5:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 6:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 7:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 8: 
+				gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 9:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 10:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 11:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 12:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 13:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 14:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 15:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 16:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 17:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P4_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 21:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 22:  
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 23: 
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 24: 
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 25: 
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) |
+						 NUM_BANKS(ADDR_SURF_8_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			adev->gfx.config.tile_mode_array[reg_offset] = gb_tile_moden;
+			WREG32(GB_TILE_MODE0 + reg_offset, gb_tile_moden);
+		}
+	} else if ((adev->asic_type == CHIP_TAHITI) || (adev->asic_type == CHIP_PITCAIRN)) {
+		for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) {
+			switch (reg_offset) {
+			case 0:  /* non-AA compressed depth or any compressed stencil */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 1:  /* 2xAA/4xAA compressed depth only */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 2:  /* 8xAA compressed depth only */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 3:  /* 2xAA/4xAA compressed depth with stencil (for depth buffer) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 4:  /* Maps w/ a dimension less than the 2D macro-tile dimensions (for mipmapped depth textures) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 5:  /* Uncompressed 16bpp depth - and stencil buffer allocated with it */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 6:  /* Uncompressed 32bpp depth - and stencil buffer allocated with it */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 7:  /* Uncompressed 8bpp stencil without depth (drivers typically do not use) */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DEPTH_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 8:  /* 1D and 1D Array Surfaces */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 9:  /* Displayable maps. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 10:  /* Display 8bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 11:  /* Display 16bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 12:  /* Display 32bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_DISPLAY_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 13:  /* Thin. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 14:  /* Thin 8 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 15:  /* Thin 16 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 16:  /* Thin 32 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 17:  /* Thin 64 bpp. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(split_equal_to_row_size) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			case 21:  /* 8 bpp PRT. */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 22:  /* 16 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4));
+				break;
+			case 23:  /* 32 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 24:  /* 64 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) |
+						 NUM_BANKS(ADDR_SURF_16_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2));
+				break;
+			case 25:  /* 128 bpp PRT */
+				gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) |
+						 MICRO_TILE_MODE(ADDR_SURF_THIN_MICRO_TILING) |
+						 PIPE_CONFIG(ADDR_SURF_P8_32x32_8x16) |
+						 TILE_SPLIT(ADDR_SURF_TILE_SPLIT_1KB) |
+						 NUM_BANKS(ADDR_SURF_8_BANK) |
+						 BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) |
+						 BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) |
+						 MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1));
+				break;
+			default:
+				gb_tile_moden = 0;
+				break;
+			}
+			adev->gfx.config.tile_mode_array[reg_offset] = gb_tile_moden;
+			WREG32(GB_TILE_MODE0 + reg_offset, gb_tile_moden);
+		}
+	} else{
+
+		DRM_ERROR("unknown asic: 0x%x\n", adev->asic_type);
+	}
+
+}
+
+static void gfx_v6_0_select_se_sh(struct amdgpu_device *adev, u32 se_num,
+				  u32 sh_num, u32 instance)
+{
+	u32 data;
+
+	if (instance == 0xffffffff)
+		data = INSTANCE_BROADCAST_WRITES;
+	else
+		data = INSTANCE_INDEX(instance);
+
+	if ((se_num == 0xffffffff) && (sh_num == 0xffffffff))
+		data |= SH_BROADCAST_WRITES | SE_BROADCAST_WRITES;
+	else if (se_num == 0xffffffff)
+		data |= SE_BROADCAST_WRITES | SH_INDEX(sh_num);
+	else if (sh_num == 0xffffffff)
+		data |= SH_BROADCAST_WRITES | SE_INDEX(se_num);
+	else
+		data |= SH_INDEX(sh_num) | SE_INDEX(se_num);
+	WREG32(GRBM_GFX_INDEX, data);
+}
+
+static u32 gfx_v6_0_create_bitmask(u32 bit_width)
+{
+	return (u32)(((u64)1 << bit_width) - 1);
+}
+
+static u32 gfx_v6_0_get_rb_disabled(struct amdgpu_device *adev,
+				    u32 max_rb_num_per_se,
+				    u32 sh_per_se)
+{
+	u32 data, mask;
+
+	data = RREG32(CC_RB_BACKEND_DISABLE);
+	data &= BACKEND_DISABLE_MASK;
+	data |= RREG32(GC_USER_RB_BACKEND_DISABLE);
+
+	data >>= BACKEND_DISABLE_SHIFT;
+
+	mask = gfx_v6_0_create_bitmask(max_rb_num_per_se / sh_per_se);
+
+	return data & mask;
+}
+
+static void gfx_v6_0_raster_config(struct amdgpu_device *adev, u32 *rconf)
+{
+	switch (adev->asic_type) {
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+		*rconf |= RB_XSEL2(2) | RB_XSEL | PKR_MAP(2) | PKR_YSEL(1) |
+			  SE_MAP(2) | SE_XSEL(2) | SE_YSEL(2);
+		break;
+	case CHIP_VERDE:
+		*rconf |= RB_XSEL | PKR_MAP(2) | PKR_YSEL(1);
+		break;
+	case CHIP_OLAND:
+		*rconf |= RB_YSEL;
+		break;
+	case CHIP_HAINAN:
+		*rconf |= 0x0;
+		break;
+	default:
+		DRM_ERROR("unknown asic: 0x%x\n", adev->asic_type);
+		break;
+	}
+}
+
+static void gfx_v6_0_write_harvested_raster_configs(struct amdgpu_device *adev,
+						    u32 raster_config, unsigned rb_mask,
+						    unsigned num_rb)
+{
+	unsigned sh_per_se = max_t(unsigned, adev->gfx.config.max_sh_per_se, 1);
+	unsigned num_se = max_t(unsigned, adev->gfx.config.max_shader_engines, 1);
+	unsigned rb_per_pkr = min_t(unsigned, num_rb / num_se / sh_per_se, 2);
+	unsigned rb_per_se = num_rb / num_se;
+	unsigned se_mask[4];
+	unsigned se;
+
+	se_mask[0] = ((1 << rb_per_se) - 1) & rb_mask;
+	se_mask[1] = (se_mask[0] << rb_per_se) & rb_mask;
+	se_mask[2] = (se_mask[1] << rb_per_se) & rb_mask;
+	se_mask[3] = (se_mask[2] << rb_per_se) & rb_mask;
+
+	WARN_ON(!(num_se == 1 || num_se == 2 || num_se == 4));
+	WARN_ON(!(sh_per_se == 1 || sh_per_se == 2));
+	WARN_ON(!(rb_per_pkr == 1 || rb_per_pkr == 2));
+
+	for (se = 0; se < num_se; se++) {
+		unsigned raster_config_se = raster_config;
+		unsigned pkr0_mask = ((1 << rb_per_pkr) - 1) << (se * rb_per_se);
+		unsigned pkr1_mask = pkr0_mask << rb_per_pkr;
+		int idx = (se / 2) * 2;
+
+		if ((num_se > 1) && (!se_mask[idx] || !se_mask[idx + 1])) {
+			raster_config_se &= ~SE_MAP_MASK;
+
+			if (!se_mask[idx]) {
+				raster_config_se |= SE_MAP(RASTER_CONFIG_SE_MAP_3);
+			} else {
+				raster_config_se |= SE_MAP(RASTER_CONFIG_SE_MAP_0);
+			}
+		}
+
+		pkr0_mask &= rb_mask;
+		pkr1_mask &= rb_mask;
+		if (rb_per_se > 2 && (!pkr0_mask || !pkr1_mask)) {
+			raster_config_se &= ~PKR_MAP_MASK;
+
+			if (!pkr0_mask) {
+				raster_config_se |= PKR_MAP(RASTER_CONFIG_PKR_MAP_3);
+			} else {
+				raster_config_se |= PKR_MAP(RASTER_CONFIG_PKR_MAP_0);
+			}
+		}
+
+		if (rb_per_se >= 2) {
+			unsigned rb0_mask = 1 << (se * rb_per_se);
+			unsigned rb1_mask = rb0_mask << 1;
+
+			rb0_mask &= rb_mask;
+			rb1_mask &= rb_mask;
+			if (!rb0_mask || !rb1_mask) {
+				raster_config_se &= ~RB_MAP_PKR0_MASK;
+
+				if (!rb0_mask) {
+					raster_config_se |=
+						RB_MAP_PKR0(RASTER_CONFIG_RB_MAP_3);
+				} else {
+					raster_config_se |=
+						RB_MAP_PKR0(RASTER_CONFIG_RB_MAP_0);
+				}
+			}
+
+			if (rb_per_se > 2) {
+				rb0_mask = 1 << (se * rb_per_se + rb_per_pkr);
+				rb1_mask = rb0_mask << 1;
+				rb0_mask &= rb_mask;
+				rb1_mask &= rb_mask;
+				if (!rb0_mask || !rb1_mask) {
+					raster_config_se &= ~RB_MAP_PKR1_MASK;
+
+					if (!rb0_mask) {
+						raster_config_se |=
+							RB_MAP_PKR1(RASTER_CONFIG_RB_MAP_3);
+					} else {
+						raster_config_se |=
+							RB_MAP_PKR1(RASTER_CONFIG_RB_MAP_0);
+					}
+				}
+			}
+		}
+
+		/* GRBM_GFX_INDEX has a different offset on SI */
+		gfx_v6_0_select_se_sh(adev, se, 0xffffffff, 0xffffffff);
+		WREG32(PA_SC_RASTER_CONFIG, raster_config_se);
+	}
+
+	/* GRBM_GFX_INDEX has a different offset on SI */
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+}
+
+static void gfx_v6_0_setup_rb(struct amdgpu_device *adev,
+			      u32 se_num, u32 sh_per_se,
+			      u32 max_rb_num_per_se)
+{
+	int i, j;
+	u32 data, mask;
+	u32 disabled_rbs = 0;
+	u32 enabled_rbs = 0;
+	unsigned num_rb_pipes;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	for (i = 0; i < se_num; i++) {
+		for (j = 0; j < sh_per_se; j++) {
+			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
+			data = gfx_v6_0_get_rb_disabled(adev, max_rb_num_per_se, sh_per_se);
+			disabled_rbs |= data << ((i * sh_per_se + j) * TAHITI_RB_BITMAP_WIDTH_PER_SH);
+		}
+	}
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	mask = 1;
+	for (i = 0; i < max_rb_num_per_se * se_num; i++) {
+		if (!(disabled_rbs & mask))
+			enabled_rbs |= mask;
+		mask <<= 1;
+	}
+
+	adev->gfx.config.backend_enable_mask = enabled_rbs;
+	adev->gfx.config.num_rbs = hweight32(enabled_rbs);
+
+	num_rb_pipes = min_t(unsigned, adev->gfx.config.max_backends_per_se *
+			     adev->gfx.config.max_shader_engines, 16);
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	for (i = 0; i < se_num; i++) {
+		gfx_v6_0_select_se_sh(adev, i, 0xffffffff, 0xffffffff);
+		data = 0;
+		for (j = 0; j < sh_per_se; j++) {
+			switch (enabled_rbs & 3) {
+			case 1:
+				data |= (RASTER_CONFIG_RB_MAP_0 << (i * sh_per_se + j) * 2);
+				break;
+			case 2:
+				data |= (RASTER_CONFIG_RB_MAP_3 << (i * sh_per_se + j) * 2);
+				break;
+			case 3:
+			default:
+				data |= (RASTER_CONFIG_RB_MAP_2 << (i * sh_per_se + j) * 2);
+				break;
+			}
+			enabled_rbs >>= 2;
+		}
+		gfx_v6_0_raster_config(adev, &data);
+
+		if (!adev->gfx.config.backend_enable_mask ||
+				adev->gfx.config.num_rbs >= num_rb_pipes)
+			WREG32(PA_SC_RASTER_CONFIG, data);
+		else
+			gfx_v6_0_write_harvested_raster_configs(adev, data,
+								adev->gfx.config.backend_enable_mask,
+								num_rb_pipes);
+	}
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	mutex_unlock(&adev->grbm_idx_mutex);
+}
+/*
+static void gmc_v6_0_init_compute_vmid(struct amdgpu_device *adev)
+{
+}
+*/
+
+static u32 gfx_v6_0_get_cu_enabled(struct amdgpu_device *adev, u32 cu_per_sh)
+{
+	u32 data, mask;
+
+	data = RREG32(CC_GC_SHADER_ARRAY_CONFIG);
+	data &= INACTIVE_CUS_MASK;
+	data |= RREG32(GC_USER_SHADER_ARRAY_CONFIG);
+
+	data >>= INACTIVE_CUS_SHIFT;
+
+	mask = gfx_v6_0_create_bitmask(cu_per_sh);
+
+	return ~data & mask;
+}
+
+
+static void gfx_v6_0_setup_spi(struct amdgpu_device *adev,
+			 u32 se_num, u32 sh_per_se,
+			 u32 cu_per_sh)
+{
+	int i, j, k;
+	u32 data, mask;
+	u32 active_cu = 0;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	for (i = 0; i < se_num; i++) {
+		for (j = 0; j < sh_per_se; j++) {
+			gfx_v6_0_select_se_sh(adev, i, j, 0xffffffff);
+			data = RREG32(SPI_STATIC_THREAD_MGMT_3);
+			active_cu = gfx_v6_0_get_cu_enabled(adev, cu_per_sh);
+
+			mask = 1;
+			for (k = 0; k < 16; k++) {
+				mask <<= k;
+				if (active_cu & mask) {
+					data &= ~mask;
+					WREG32(SPI_STATIC_THREAD_MGMT_3, data);
+					break;
+				}
+			}
+		}
+	}
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	mutex_unlock(&adev->grbm_idx_mutex);
+}
+
+static void gfx_v6_0_gpu_init(struct amdgpu_device *adev)
+{
+	u32 gb_addr_config = 0;
+	u32 mc_shared_chmap, mc_arb_ramcfg;
+	u32 sx_debug_1;
+	u32 hdp_host_path_cntl;
+	u32 tmp;
+
+	switch (adev->asic_type) {
+	case CHIP_TAHITI:
+		adev->gfx.config.max_shader_engines = 2;
+		adev->gfx.config.max_tile_pipes = 12;
+		adev->gfx.config.max_cu_per_sh = 8;
+		adev->gfx.config.max_sh_per_se = 2;
+		adev->gfx.config.max_backends_per_se = 4;
+		adev->gfx.config.max_texture_channel_caches = 12;
+		adev->gfx.config.max_gprs = 256;
+		adev->gfx.config.max_gs_threads = 32;
+		adev->gfx.config.max_hw_contexts = 8;
+
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0x30;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = TAHITI_GB_ADDR_CONFIG_GOLDEN;
+		break;
+	case CHIP_PITCAIRN:
+		adev->gfx.config.max_shader_engines = 2;
+		adev->gfx.config.max_tile_pipes = 8;
+		adev->gfx.config.max_cu_per_sh = 5;
+		adev->gfx.config.max_sh_per_se = 2;
+		adev->gfx.config.max_backends_per_se = 4;
+		adev->gfx.config.max_texture_channel_caches = 8;
+		adev->gfx.config.max_gprs = 256;
+		adev->gfx.config.max_gs_threads = 32;
+		adev->gfx.config.max_hw_contexts = 8;
+
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x100;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0x30;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = TAHITI_GB_ADDR_CONFIG_GOLDEN;
+		break;
+
+	case CHIP_VERDE:
+		adev->gfx.config.max_shader_engines = 1;
+		adev->gfx.config.max_tile_pipes = 4;
+		adev->gfx.config.max_cu_per_sh = 5;
+		adev->gfx.config.max_sh_per_se = 2;
+		adev->gfx.config.max_backends_per_se = 4;
+		adev->gfx.config.max_texture_channel_caches = 4;
+		adev->gfx.config.max_gprs = 256;
+		adev->gfx.config.max_gs_threads = 32;
+		adev->gfx.config.max_hw_contexts = 8;
+
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x40;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0x30;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = VERDE_GB_ADDR_CONFIG_GOLDEN;
+		break;
+	case CHIP_OLAND:
+		adev->gfx.config.max_shader_engines = 1;
+		adev->gfx.config.max_tile_pipes = 4;
+		adev->gfx.config.max_cu_per_sh = 6;
+		adev->gfx.config.max_sh_per_se = 1;
+		adev->gfx.config.max_backends_per_se = 2;
+		adev->gfx.config.max_texture_channel_caches = 4;
+		adev->gfx.config.max_gprs = 256;
+		adev->gfx.config.max_gs_threads = 16;
+		adev->gfx.config.max_hw_contexts = 8;
+
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x40;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0x30;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = VERDE_GB_ADDR_CONFIG_GOLDEN;
+		break;
+	case CHIP_HAINAN:
+		adev->gfx.config.max_shader_engines = 1;
+		adev->gfx.config.max_tile_pipes = 4;
+		adev->gfx.config.max_cu_per_sh = 5;
+		adev->gfx.config.max_sh_per_se = 1;
+		adev->gfx.config.max_backends_per_se = 1;
+		adev->gfx.config.max_texture_channel_caches = 2;
+		adev->gfx.config.max_gprs = 256;
+		adev->gfx.config.max_gs_threads = 16;
+		adev->gfx.config.max_hw_contexts = 8;
+
+		adev->gfx.config.sc_prim_fifo_size_frontend = 0x20;
+		adev->gfx.config.sc_prim_fifo_size_backend = 0x40;
+		adev->gfx.config.sc_hiz_tile_fifo_size = 0x30;
+		adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130;
+		gb_addr_config = HAINAN_GB_ADDR_CONFIG_GOLDEN;
+		break;
+	default:
+		BUG();
+		break;
+	}
+
+	WREG32(GRBM_CNTL, GRBM_READ_TIMEOUT(0xff));
+	WREG32(SRBM_INT_CNTL, 1);
+	WREG32(SRBM_INT_ACK, 1);
+
+	WREG32(BIF_FB_EN, FB_READ_EN | FB_WRITE_EN);
+
+	mc_shared_chmap = RREG32(MC_SHARED_CHMAP);
+	mc_arb_ramcfg = RREG32(MC_ARB_RAMCFG);
+
+	adev->gfx.config.num_tile_pipes = adev->gfx.config.max_tile_pipes;
+	adev->gfx.config.mem_max_burst_length_bytes = 256;
+	tmp = (mc_arb_ramcfg & NOOFCOLS_MASK) >> NOOFCOLS_SHIFT;
+	adev->gfx.config.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024;
+	if (adev->gfx.config.mem_row_size_in_kb > 4)
+		adev->gfx.config.mem_row_size_in_kb = 4;
+	adev->gfx.config.shader_engine_tile_size = 32;
+	adev->gfx.config.num_gpus = 1;
+	adev->gfx.config.multi_gpu_tile_size = 64;
+
+	gb_addr_config &= ~ROW_SIZE_MASK;
+	switch (adev->gfx.config.mem_row_size_in_kb) {
+	case 1:
+	default:
+		gb_addr_config |= ROW_SIZE(0);
+		break;
+	case 2:
+		gb_addr_config |= ROW_SIZE(1);
+		break;
+	case 4:
+		gb_addr_config |= ROW_SIZE(2);
+		break;
+	}
+	adev->gfx.config.gb_addr_config = gb_addr_config;
+
+	WREG32(GB_ADDR_CONFIG, gb_addr_config);
+	WREG32(DMIF_ADDR_CONFIG, gb_addr_config);
+	WREG32(DMIF_ADDR_CALC, gb_addr_config);
+	WREG32(HDP_ADDR_CONFIG, gb_addr_config);
+	WREG32(DMA_TILING_CONFIG + DMA0_REGISTER_OFFSET, gb_addr_config);
+	WREG32(DMA_TILING_CONFIG + DMA1_REGISTER_OFFSET, gb_addr_config);
+#if 0
+	if (adev->has_uvd) {
+		WREG32(UVD_UDEC_ADDR_CONFIG, gb_addr_config);
+		WREG32(UVD_UDEC_DB_ADDR_CONFIG, gb_addr_config);
+		WREG32(UVD_UDEC_DBW_ADDR_CONFIG, gb_addr_config);
+	}
+#endif
+	gfx_v6_0_tiling_mode_table_init(adev);
+
+	gfx_v6_0_setup_rb(adev, adev->gfx.config.max_shader_engines,
+		    adev->gfx.config.max_sh_per_se,
+		    adev->gfx.config.max_backends_per_se);
+
+	gfx_v6_0_setup_spi(adev, adev->gfx.config.max_shader_engines,
+		     adev->gfx.config.max_sh_per_se,
+		     adev->gfx.config.max_cu_per_sh);
+
+	gfx_v6_0_get_cu_info(adev);
+
+	WREG32(CP_QUEUE_THRESHOLDS, (ROQ_IB1_START(0x16) |
+				     ROQ_IB2_START(0x2b)));
+	WREG32(CP_MEQ_THRESHOLDS, MEQ1_START(0x30) | MEQ2_START(0x60));
+
+	sx_debug_1 = RREG32(SX_DEBUG_1);
+	WREG32(SX_DEBUG_1, sx_debug_1);
+
+	WREG32(SPI_CONFIG_CNTL_1, VTX_DONE_DELAY(4));
+
+	WREG32(PA_SC_FIFO_SIZE, (SC_FRONTEND_PRIM_FIFO_SIZE(adev->gfx.config.sc_prim_fifo_size_frontend) |
+				 SC_BACKEND_PRIM_FIFO_SIZE(adev->gfx.config.sc_prim_fifo_size_backend) |
+				 SC_HIZ_TILE_FIFO_SIZE(adev->gfx.config.sc_hiz_tile_fifo_size) |
+				 SC_EARLYZ_TILE_FIFO_SIZE(adev->gfx.config.sc_earlyz_tile_fifo_size)));
+
+	WREG32(VGT_NUM_INSTANCES, 1);
+	WREG32(CP_PERFMON_CNTL, 0);
+	WREG32(SQ_CONFIG, 0);
+	WREG32(PA_SC_FORCE_EOV_MAX_CNTS, (FORCE_EOV_MAX_CLK_CNT(4095) |
+					  FORCE_EOV_MAX_REZ_CNT(255)));
+
+	WREG32(VGT_CACHE_INVALIDATION, CACHE_INVALIDATION(VC_AND_TC) |
+	       AUTO_INVLD_EN(ES_AND_GS_AUTO));
+
+	WREG32(VGT_GS_VERTEX_REUSE, 16);
+	WREG32(PA_SC_LINE_STIPPLE_STATE, 0);
+
+	WREG32(CB_PERFCOUNTER0_SELECT0, 0);
+	WREG32(CB_PERFCOUNTER0_SELECT1, 0);
+	WREG32(CB_PERFCOUNTER1_SELECT0, 0);
+	WREG32(CB_PERFCOUNTER1_SELECT1, 0);
+	WREG32(CB_PERFCOUNTER2_SELECT0, 0);
+	WREG32(CB_PERFCOUNTER2_SELECT1, 0);
+	WREG32(CB_PERFCOUNTER3_SELECT0, 0);
+	WREG32(CB_PERFCOUNTER3_SELECT1, 0);
+
+	hdp_host_path_cntl = RREG32(HDP_HOST_PATH_CNTL);
+	WREG32(HDP_HOST_PATH_CNTL, hdp_host_path_cntl);
+
+	WREG32(PA_CL_ENHANCE, CLIP_VTX_REORDER_ENA | NUM_CLIP_SEQ(3));
+
+	udelay(50);
+}
+
+
+static void gfx_v6_0_scratch_init(struct amdgpu_device *adev)
+{
+	int i;
+
+	adev->gfx.scratch.num_reg = 7;
+	adev->gfx.scratch.reg_base = SCRATCH_REG0;
+	for (i = 0; i < adev->gfx.scratch.num_reg; i++) {
+		adev->gfx.scratch.free[i] = true;
+		adev->gfx.scratch.reg[i] = adev->gfx.scratch.reg_base + i;
+	}
+}
+
+static int gfx_v6_0_ring_test_ring(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	uint32_t scratch;
+	uint32_t tmp = 0;
+	unsigned i;
+	int r;
+
+	r = amdgpu_gfx_scratch_get(adev, &scratch);
+	if (r) {
+		DRM_ERROR("amdgpu: cp failed to get scratch reg (%d).\n", r);
+		return r;
+	}
+	WREG32(scratch, 0xCAFEDEAD);
+
+	r = amdgpu_ring_alloc(ring, 3);
+	if (r) {
+		DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n", ring->idx, r);
+		amdgpu_gfx_scratch_free(adev, scratch);
+		return r;
+	}
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	amdgpu_ring_write(ring, (scratch - PACKET3_SET_CONFIG_REG_START));
+	amdgpu_ring_write(ring, 0xDEADBEEF);
+	amdgpu_ring_commit(ring);
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		tmp = RREG32(scratch);
+		if (tmp == 0xDEADBEEF)
+			break;
+		DRM_UDELAY(1);
+	}
+	if (i < adev->usec_timeout) {
+		DRM_INFO("ring test on %d succeeded in %d usecs\n", ring->idx, i);
+	} else {
+		DRM_ERROR("amdgpu: ring %d test failed (scratch(0x%04X)=0x%08X)\n",
+			  ring->idx, scratch, tmp);
+		r = -EINVAL;
+	}
+	amdgpu_gfx_scratch_free(adev, scratch);
+	return r;
+}
+
+static void gfx_v6_0_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+{
+	/* flush hdp cache */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
+				 WRITE_DATA_DST_SEL(0)));
+	amdgpu_ring_write(ring, HDP_MEM_COHERENCY_FLUSH_CNTL);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, 0x1);
+}
+
+/**
+ * gfx_v6_0_ring_emit_hdp_invalidate - emit an hdp invalidate on the cp
+ *
+ * @adev: amdgpu_device pointer
+ * @ridx: amdgpu ring index
+ *
+ * Emits an hdp invalidate on the cp.
+ */
+static void gfx_v6_0_ring_emit_hdp_invalidate(struct amdgpu_ring *ring)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
+				 WRITE_DATA_DST_SEL(0)));
+	amdgpu_ring_write(ring, HDP_DEBUG0);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, 0x1);
+}
+
+static void gfx_v6_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr,
+				     u64 seq, unsigned flags)
+{
+	bool write64bit = flags & AMDGPU_FENCE_FLAG_64BIT;
+	bool int_sel = flags & AMDGPU_FENCE_FLAG_INT;
+	/* flush read cache over gart */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	amdgpu_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START));
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
+	amdgpu_ring_write(ring, PACKET3_TCL1_ACTION_ENA |
+			  PACKET3_TC_ACTION_ENA |
+			  PACKET3_SH_KCACHE_ACTION_ENA |
+			  PACKET3_SH_ICACHE_ACTION_ENA);
+	amdgpu_ring_write(ring, 0xFFFFFFFF);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, 10); /* poll interval */
+	/* EVENT_WRITE_EOP - flush caches, send int */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
+	amdgpu_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_TS_EVENT) | EVENT_INDEX(5));
+	amdgpu_ring_write(ring, addr & 0xfffffffc);
+	amdgpu_ring_write(ring, (upper_32_bits(addr) & 0xffff) |
+				DATA_SEL(write64bit ? 2 : 1) | INT_SEL(int_sel ? 2 : 0));
+	amdgpu_ring_write(ring, lower_32_bits(seq));
+	amdgpu_ring_write(ring, upper_32_bits(seq));
+}
+
+static void gfx_v6_0_ring_emit_ib(struct amdgpu_ring *ring,
+				  struct amdgpu_ib *ib,
+				  unsigned vm_id, bool ctx_switch)
+{
+	u32 header, control = 0;
+
+	/* insert SWITCH_BUFFER packet before first IB in the ring frame */
+	if (ctx_switch) {
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
+		amdgpu_ring_write(ring, 0);
+	}
+
+	if (ib->flags & AMDGPU_IB_FLAG_CE)
+		header = PACKET3(PACKET3_INDIRECT_BUFFER_CONST, 2);
+	else
+		header = PACKET3(PACKET3_INDIRECT_BUFFER, 2);
+
+	control |= ib->length_dw | (vm_id << 24);
+
+	amdgpu_ring_write(ring, header);
+	amdgpu_ring_write(ring,
+#ifdef __BIG_ENDIAN
+			  (2 << 0) |
+#endif
+			  (ib->gpu_addr & 0xFFFFFFFC));
+	amdgpu_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFFFF);
+	amdgpu_ring_write(ring, control);
+}
+
+/**
+ * gfx_v6_0_ring_test_ib - basic ring IB test
+ *
+ * @ring: amdgpu_ring structure holding ring information
+ *
+ * Allocate an IB and execute it on the gfx ring (SI).
+ * Provides a basic gfx ring test to verify that IBs are working.
+ * Returns 0 on success, error on failure.
+ */
+static int gfx_v6_0_ring_test_ib(struct amdgpu_ring *ring, long timeout)
+{
+	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_ib ib;
+	struct fence *f = NULL;
+	uint32_t scratch;
+	uint32_t tmp = 0;
+	long r;
+
+	r = amdgpu_gfx_scratch_get(adev, &scratch);
+	if (r) {
+		DRM_ERROR("amdgpu: failed to get scratch reg (%ld).\n", r);
+		return r;
+	}
+	WREG32(scratch, 0xCAFEDEAD);
+	memset(&ib, 0, sizeof(ib));
+	r = amdgpu_ib_get(adev, NULL, 256, &ib);
+	if (r) {
+		DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r);
+		goto err1;
+	}
+	ib.ptr[0] = PACKET3(PACKET3_SET_CONFIG_REG, 1);
+	ib.ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_START));
+	ib.ptr[2] = 0xDEADBEEF;
+	ib.length_dw = 3;
+
+	r = amdgpu_ib_schedule(ring, 1, &ib, NULL, NULL, &f);
+	if (r)
+		goto err2;
+
+	r = fence_wait_timeout(f, false, timeout);
+	if (r == 0) {
+		DRM_ERROR("amdgpu: IB test timed out\n");
+		r = -ETIMEDOUT;
+		goto err2;
+	} else if (r < 0) {
+		DRM_ERROR("amdgpu: fence wait failed (%ld).\n", r);
+		goto err2;
+	}
+	tmp = RREG32(scratch);
+	if (tmp == 0xDEADBEEF) {
+		DRM_INFO("ib test on ring %d succeeded\n", ring->idx);
+		r = 0;
+	} else {
+		DRM_ERROR("amdgpu: ib test failed (scratch(0x%04X)=0x%08X)\n",
+			  scratch, tmp);
+		r = -EINVAL;
+	}
+
+err2:
+	amdgpu_ib_free(adev, &ib, NULL);
+	fence_put(f);
+err1:
+	amdgpu_gfx_scratch_free(adev, scratch);
+	return r;
+}
+
+static void gfx_v6_0_cp_gfx_enable(struct amdgpu_device *adev, bool enable)
+{
+	int i;
+	if (enable)
+		WREG32(CP_ME_CNTL, 0);
+	else {
+		WREG32(CP_ME_CNTL, (CP_ME_HALT | CP_PFP_HALT | CP_CE_HALT));
+		WREG32(SCRATCH_UMSK, 0);
+		for (i = 0; i < adev->gfx.num_gfx_rings; i++)
+			adev->gfx.gfx_ring[i].ready = false;
+		for (i = 0; i < adev->gfx.num_compute_rings; i++)
+			adev->gfx.compute_ring[i].ready = false;
+	}
+	udelay(50);
+}
+
+static int gfx_v6_0_cp_gfx_load_microcode(struct amdgpu_device *adev)
+{
+	unsigned i;
+	const struct gfx_firmware_header_v1_0 *pfp_hdr;
+	const struct gfx_firmware_header_v1_0 *ce_hdr;
+	const struct gfx_firmware_header_v1_0 *me_hdr;
+	const __le32 *fw_data;
+	u32 fw_size;
+
+	if (!adev->gfx.me_fw || !adev->gfx.pfp_fw || !adev->gfx.ce_fw)
+		return -EINVAL;
+
+	gfx_v6_0_cp_gfx_enable(adev, false);
+	pfp_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data;
+	ce_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data;
+	me_hdr = (const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data;
+
+	amdgpu_ucode_print_gfx_hdr(&pfp_hdr->header);
+	amdgpu_ucode_print_gfx_hdr(&ce_hdr->header);
+	amdgpu_ucode_print_gfx_hdr(&me_hdr->header);
+
+	/* PFP */
+	fw_data = (const __le32 *)
+		(adev->gfx.pfp_fw->data + le32_to_cpu(pfp_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(pfp_hdr->header.ucode_size_bytes) / 4;
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+	for (i = 0; i < fw_size; i++)
+		WREG32(CP_PFP_UCODE_DATA, le32_to_cpup(fw_data++));
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+
+	/* CE */
+	fw_data = (const __le32 *)
+		(adev->gfx.ce_fw->data + le32_to_cpu(ce_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(ce_hdr->header.ucode_size_bytes) / 4;
+	WREG32(CP_CE_UCODE_ADDR, 0);
+	for (i = 0; i < fw_size; i++)
+		WREG32(CP_CE_UCODE_DATA, le32_to_cpup(fw_data++));
+	WREG32(CP_CE_UCODE_ADDR, 0);
+
+	/* ME */
+	fw_data = (const __be32 *)
+		(adev->gfx.me_fw->data + le32_to_cpu(me_hdr->header.ucode_array_offset_bytes));
+	fw_size = le32_to_cpu(me_hdr->header.ucode_size_bytes) / 4;
+	WREG32(CP_ME_RAM_WADDR, 0);
+	for (i = 0; i < fw_size; i++)
+		WREG32(CP_ME_RAM_DATA, le32_to_cpup(fw_data++));
+	WREG32(CP_ME_RAM_WADDR, 0);
+
+
+	WREG32(CP_PFP_UCODE_ADDR, 0);
+	WREG32(CP_CE_UCODE_ADDR, 0);
+	WREG32(CP_ME_RAM_WADDR, 0);
+	WREG32(CP_ME_RAM_RADDR, 0);
+	return 0;
+}
+
+static int gfx_v6_0_cp_gfx_start(struct amdgpu_device *adev)
+{
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+	struct amdgpu_ring *ring = &adev->gfx.gfx_ring[0];
+	int r, i;
+
+	r = amdgpu_ring_alloc(ring, 7 + 4);
+	if (r) {
+		DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r);
+		return r;
+	}
+	amdgpu_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5));
+	amdgpu_ring_write(ring, 0x1);
+	amdgpu_ring_write(ring, 0x0);
+	amdgpu_ring_write(ring, adev->gfx.config.max_hw_contexts - 1);
+	amdgpu_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, 0);
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_BASE, 2));
+	amdgpu_ring_write(ring, PACKET3_BASE_INDEX(CE_PARTITION_BASE));
+	amdgpu_ring_write(ring, 0xc000);
+	amdgpu_ring_write(ring, 0xe000);
+	amdgpu_ring_commit(ring);
+
+	gfx_v6_0_cp_gfx_enable(adev, true);
+
+	r = amdgpu_ring_alloc(ring, gfx_v6_0_get_csb_size(adev) + 10);
+	if (r) {
+		DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r);
+		return r;
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	amdgpu_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+
+	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
+		for (ext = sect->section; ext->extent != NULL; ++ext) {
+			if (sect->id == SECT_CONTEXT) {
+				amdgpu_ring_write(ring,
+						  PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
+				amdgpu_ring_write(ring, ext->reg_index - PACKET3_SET_CONTEXT_REG_START);
+				for (i = 0; i < ext->reg_count; i++)
+					amdgpu_ring_write(ring, ext->extent[i]);
+			}
+		}
+	}
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	amdgpu_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE);
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
+	amdgpu_ring_write(ring, 0);
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	amdgpu_ring_write(ring, 0x00000316);
+	amdgpu_ring_write(ring, 0x0000000e);
+	amdgpu_ring_write(ring, 0x00000010);
+
+	amdgpu_ring_commit(ring);
+
+	return 0;
+}
+
+static int gfx_v6_0_cp_gfx_resume(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	u32 tmp;
+	u32 rb_bufsz;
+	int r;
+	u64 rptr_addr;
+
+	WREG32(CP_SEM_WAIT_TIMER, 0x0);
+	WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0);
+
+	/* Set the write pointer delay */
+	WREG32(CP_RB_WPTR_DELAY, 0);
+
+	WREG32(CP_DEBUG, 0);
+	WREG32(SCRATCH_ADDR, 0);
+
+	/* ring 0 - compute and gfx */
+	/* Set ring buffer size */
+	ring = &adev->gfx.gfx_ring[0];
+	rb_bufsz = order_base_2(ring->ring_size / 8);
+	tmp = (order_base_2(AMDGPU_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+
+#ifdef __BIG_ENDIAN
+	tmp |= BUF_SWAP_32BIT;
+#endif
+	WREG32(CP_RB0_CNTL, tmp);
+
+	/* Initialize the ring buffer's read and write pointers */
+	WREG32(CP_RB0_CNTL, tmp | RB_RPTR_WR_ENA);
+	ring->wptr = 0;
+	WREG32(CP_RB0_WPTR, ring->wptr);
+
+	/* set the wb address whether it's enabled or not */
+	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	WREG32(CP_RB0_RPTR_ADDR, lower_32_bits(rptr_addr));
+	WREG32(CP_RB0_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
+
+	WREG32(SCRATCH_UMSK, 0);
+
+	mdelay(1);
+	WREG32(CP_RB0_CNTL, tmp);
+
+	WREG32(CP_RB0_BASE, ring->gpu_addr >> 8);
+
+	/* start the rings */
+	gfx_v6_0_cp_gfx_start(adev);
+	ring->ready = true;
+	r = amdgpu_ring_test_ring(ring);
+	if (r) {
+		ring->ready = false;
+		return r;
+	}
+
+	return 0;
+}
+
+static u32 gfx_v6_0_ring_get_rptr(struct amdgpu_ring *ring)
+{
+	return ring->adev->wb.wb[ring->rptr_offs];
+}
+
+static u32 gfx_v6_0_ring_get_wptr(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring == &adev->gfx.gfx_ring[0])
+		return RREG32(CP_RB0_WPTR);
+	else if (ring == &adev->gfx.compute_ring[0])
+		return RREG32(CP_RB1_WPTR);
+	else if (ring == &adev->gfx.compute_ring[1])
+		return RREG32(CP_RB2_WPTR);
+	else
+		BUG();
+}
+
+static void gfx_v6_0_ring_set_wptr_gfx(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	WREG32(CP_RB0_WPTR, ring->wptr);
+	(void)RREG32(CP_RB0_WPTR);
+}
+
+static void gfx_v6_0_ring_set_wptr_compute(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+
+	if (ring == &adev->gfx.compute_ring[0]) {
+		WREG32(CP_RB1_WPTR, ring->wptr);
+		(void)RREG32(CP_RB1_WPTR);
+	} else if (ring == &adev->gfx.compute_ring[1]) {
+		WREG32(CP_RB2_WPTR, ring->wptr);
+		(void)RREG32(CP_RB2_WPTR);
+	} else {
+		BUG();
+	}
+
+}
+
+static int gfx_v6_0_cp_compute_resume(struct amdgpu_device *adev)
+{
+	struct amdgpu_ring *ring;
+	u32 tmp;
+	u32 rb_bufsz;
+	int r;
+	u64 rptr_addr;
+
+	/* ring1  - compute only */
+	/* Set ring buffer size */
+
+	ring = &adev->gfx.compute_ring[0];
+	rb_bufsz = order_base_2(ring->ring_size / 8);
+	tmp = (order_base_2(AMDGPU_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+#ifdef __BIG_ENDIAN
+	tmp |= BUF_SWAP_32BIT;
+#endif
+	WREG32(CP_RB1_CNTL, tmp);
+
+	WREG32(CP_RB1_CNTL, tmp | RB_RPTR_WR_ENA);
+	ring->wptr = 0;
+	WREG32(CP_RB1_WPTR, ring->wptr);
+
+	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	WREG32(CP_RB1_RPTR_ADDR, lower_32_bits(rptr_addr));
+	WREG32(CP_RB1_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
+
+	mdelay(1);
+	WREG32(CP_RB1_CNTL, tmp);
+	WREG32(CP_RB1_BASE, ring->gpu_addr >> 8);
+
+	ring = &adev->gfx.compute_ring[1];
+	rb_bufsz = order_base_2(ring->ring_size / 8);
+	tmp = (order_base_2(AMDGPU_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
+#ifdef __BIG_ENDIAN
+	tmp |= BUF_SWAP_32BIT;
+#endif
+	WREG32(CP_RB2_CNTL, tmp);
+
+	WREG32(CP_RB2_CNTL, tmp | RB_RPTR_WR_ENA);
+	ring->wptr = 0;
+	WREG32(CP_RB2_WPTR, ring->wptr);
+	rptr_addr = adev->wb.gpu_addr + (ring->rptr_offs * 4);
+	WREG32(CP_RB2_RPTR_ADDR, lower_32_bits(rptr_addr));
+	WREG32(CP_RB2_RPTR_ADDR_HI, upper_32_bits(rptr_addr) & 0xFF);
+
+	mdelay(1);
+	WREG32(CP_RB2_CNTL, tmp);
+	WREG32(CP_RB2_BASE, ring->gpu_addr >> 8);
+
+	adev->gfx.compute_ring[0].ready = true;
+	adev->gfx.compute_ring[1].ready = true;
+
+	r = amdgpu_ring_test_ring(&adev->gfx.compute_ring[0]);
+	if (r) {
+		adev->gfx.compute_ring[0].ready = false;
+		return r;
+	}
+
+	r = amdgpu_ring_test_ring(&adev->gfx.compute_ring[1]);
+	if (r) {
+		adev->gfx.compute_ring[1].ready = false;
+		return r;
+	}
+
+	return 0;
+}
+
+static void gfx_v6_0_cp_enable(struct amdgpu_device *adev, bool enable)
+{
+	gfx_v6_0_cp_gfx_enable(adev, enable);
+}
+
+static int gfx_v6_0_cp_load_microcode(struct amdgpu_device *adev)
+{
+	return gfx_v6_0_cp_gfx_load_microcode(adev);
+}
+
+static void gfx_v6_0_enable_gui_idle_interrupt(struct amdgpu_device *adev,
+					       bool enable)
+{	
+	u32 tmp = RREG32(CP_INT_CNTL_RING0);
+	u32 mask;
+	int i;
+
+	if (enable)
+		tmp |= (CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	else
+		tmp &= ~(CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL_RING0, tmp);
+
+	if (!enable) {
+		/* read a gfx register */
+		tmp = RREG32(DB_DEPTH_INFO);
+
+		mask = RLC_BUSY_STATUS | GFX_POWER_STATUS | GFX_CLOCK_STATUS | GFX_LS_STATUS;
+		for (i = 0; i < adev->usec_timeout; i++) {
+			if ((RREG32(RLC_STAT) & mask) == (GFX_CLOCK_STATUS | GFX_POWER_STATUS))
+				break;
+			udelay(1);
+		}
+	}
+}
+
+static int gfx_v6_0_cp_resume(struct amdgpu_device *adev)
+{
+	int r;
+
+	gfx_v6_0_enable_gui_idle_interrupt(adev, false);
+
+	r = gfx_v6_0_cp_load_microcode(adev);
+	if (r)
+		return r;
+
+	r = gfx_v6_0_cp_gfx_resume(adev);
+	if (r)
+		return r;
+	r = gfx_v6_0_cp_compute_resume(adev);
+	if (r)
+		return r;
+
+	gfx_v6_0_enable_gui_idle_interrupt(adev, true);
+
+	return 0;
+}
+
+static void gfx_v6_0_ring_emit_pipeline_sync(struct amdgpu_ring *ring)
+{
+	int usepfp = (ring->type == AMDGPU_RING_TYPE_GFX);
+	uint32_t seq = ring->fence_drv.sync_seq;
+	uint64_t addr = ring->fence_drv.gpu_addr;
+
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+	amdgpu_ring_write(ring, (WAIT_REG_MEM_MEM_SPACE(1) | /* memory */
+				 WAIT_REG_MEM_FUNCTION(3) | /* equal */
+				 WAIT_REG_MEM_ENGINE(usepfp)));   /* pfp or me */
+	amdgpu_ring_write(ring, addr & 0xfffffffc);
+	amdgpu_ring_write(ring, upper_32_bits(addr) & 0xffffffff);
+	amdgpu_ring_write(ring, seq);
+	amdgpu_ring_write(ring, 0xffffffff);
+	amdgpu_ring_write(ring, 4); /* poll interval */
+
+	if (usepfp) {
+		/* synce CE with ME to prevent CE fetch CEIB before context switch done */
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
+		amdgpu_ring_write(ring, 0);
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
+		amdgpu_ring_write(ring, 0);
+	}
+}
+
+static void gfx_v6_0_ring_emit_vm_flush(struct amdgpu_ring *ring,
+					unsigned vm_id, uint64_t pd_addr)
+{
+	int usepfp = (ring->type == AMDGPU_RING_TYPE_GFX);
+
+	/* write new base address */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
+				 WRITE_DATA_DST_SEL(0)));
+	if (vm_id < 8) {
+		amdgpu_ring_write(ring, (VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + vm_id ));
+	} else {
+		amdgpu_ring_write(ring, (VM_CONTEXT8_PAGE_TABLE_BASE_ADDR + (vm_id - 8)));
+	}
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, pd_addr >> 12);
+
+	/* bits 0-15 are the VM contexts0-15 */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WRITE_DATA, 3));
+	amdgpu_ring_write(ring, (WRITE_DATA_ENGINE_SEL(1) |
+				 WRITE_DATA_DST_SEL(0)));
+	amdgpu_ring_write(ring, VM_INVALIDATE_REQUEST);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, 1 << vm_id);
+
+	/* wait for the invalidate to complete */
+	amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5));
+	amdgpu_ring_write(ring, (WAIT_REG_MEM_FUNCTION(0) |  /* always */
+				 WAIT_REG_MEM_ENGINE(0))); /* me */
+	amdgpu_ring_write(ring, VM_INVALIDATE_REQUEST);
+	amdgpu_ring_write(ring, 0);
+	amdgpu_ring_write(ring, 0); /* ref */
+	amdgpu_ring_write(ring, 0); /* mask */
+	amdgpu_ring_write(ring, 0x20); /* poll interval */
+
+	if (usepfp) {
+		/* sync PFP to ME, otherwise we might get invalid PFP reads */
+		amdgpu_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0));
+		amdgpu_ring_write(ring, 0x0);
+
+		/* synce CE with ME to prevent CE fetch CEIB before context switch done */
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
+		amdgpu_ring_write(ring, 0);
+		amdgpu_ring_write(ring, PACKET3(PACKET3_SWITCH_BUFFER, 0));
+		amdgpu_ring_write(ring, 0);
+	}
+}
+
+
+static void gfx_v6_0_rlc_fini(struct amdgpu_device *adev)
+{
+	int r;
+
+	if (adev->gfx.rlc.save_restore_obj) {
+		r = amdgpu_bo_reserve(adev->gfx.rlc.save_restore_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(adev->dev, "(%d) reserve RLC sr bo failed\n", r);
+		amdgpu_bo_unpin(adev->gfx.rlc.save_restore_obj);
+		amdgpu_bo_unreserve(adev->gfx.rlc.save_restore_obj);
+
+		amdgpu_bo_unref(&adev->gfx.rlc.save_restore_obj);
+		adev->gfx.rlc.save_restore_obj = NULL;
+	}
+
+	if (adev->gfx.rlc.clear_state_obj) {
+		r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(adev->dev, "(%d) reserve RLC c bo failed\n", r);
+		amdgpu_bo_unpin(adev->gfx.rlc.clear_state_obj);
+		amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
+
+		amdgpu_bo_unref(&adev->gfx.rlc.clear_state_obj);
+		adev->gfx.rlc.clear_state_obj = NULL;
+	}
+
+	if (adev->gfx.rlc.cp_table_obj) {
+		r = amdgpu_bo_reserve(adev->gfx.rlc.cp_table_obj, false);
+		if (unlikely(r != 0))
+			dev_warn(adev->dev, "(%d) reserve RLC cp table bo failed\n", r);
+		amdgpu_bo_unpin(adev->gfx.rlc.cp_table_obj);
+		amdgpu_bo_unreserve(adev->gfx.rlc.cp_table_obj);
+
+		amdgpu_bo_unref(&adev->gfx.rlc.cp_table_obj);
+		adev->gfx.rlc.cp_table_obj = NULL;
+	}
+}
+
+static int gfx_v6_0_rlc_init(struct amdgpu_device *adev)
+{
+	const u32 *src_ptr;
+	volatile u32 *dst_ptr;
+	u32 dws, i;
+	u64 reg_list_mc_addr;
+	const struct cs_section_def *cs_data;
+	int r;
+
+	adev->gfx.rlc.reg_list = verde_rlc_save_restore_register_list;
+	adev->gfx.rlc.reg_list_size =
+			(u32)ARRAY_SIZE(verde_rlc_save_restore_register_list);
+
+	adev->gfx.rlc.cs_data = si_cs_data;
+	src_ptr = adev->gfx.rlc.reg_list;
+	dws = adev->gfx.rlc.reg_list_size;
+	cs_data = adev->gfx.rlc.cs_data;
+
+	if (src_ptr) {
+		/* save restore block */
+		if (adev->gfx.rlc.save_restore_obj == NULL) {
+
+			r = amdgpu_bo_create(adev, dws * 4, PAGE_SIZE, true,
+					     AMDGPU_GEM_DOMAIN_VRAM,
+					     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
+					     NULL, NULL,
+					     &adev->gfx.rlc.save_restore_obj);
+
+			if (r) {
+				dev_warn(adev->dev, "(%d) create RLC sr bo failed\n", r);
+				return r;
+			}
+		}
+
+		r = amdgpu_bo_reserve(adev->gfx.rlc.save_restore_obj, false);
+		if (unlikely(r != 0)) {
+			gfx_v6_0_rlc_fini(adev);
+			return r;
+		}
+		r = amdgpu_bo_pin(adev->gfx.rlc.save_restore_obj, AMDGPU_GEM_DOMAIN_VRAM,
+				  &adev->gfx.rlc.save_restore_gpu_addr);
+		if (r) {
+			amdgpu_bo_unreserve(adev->gfx.rlc.save_restore_obj);
+			dev_warn(adev->dev, "(%d) pin RLC sr bo failed\n", r);
+			gfx_v6_0_rlc_fini(adev);
+			return r;
+		}
+
+		r = amdgpu_bo_kmap(adev->gfx.rlc.save_restore_obj, (void **)&adev->gfx.rlc.sr_ptr);
+		if (r) {
+			dev_warn(adev->dev, "(%d) map RLC sr bo failed\n", r);
+			gfx_v6_0_rlc_fini(adev);
+			return r;
+		}
+		/* write the sr buffer */
+		dst_ptr = adev->gfx.rlc.sr_ptr;
+		for (i = 0; i < adev->gfx.rlc.reg_list_size; i++)
+			dst_ptr[i] = cpu_to_le32(src_ptr[i]);
+		amdgpu_bo_kunmap(adev->gfx.rlc.save_restore_obj);
+		amdgpu_bo_unreserve(adev->gfx.rlc.save_restore_obj);
+	}
+
+	if (cs_data) {
+		/* clear state block */
+		adev->gfx.rlc.clear_state_size = gfx_v6_0_get_csb_size(adev);
+		dws = adev->gfx.rlc.clear_state_size + (256 / 4);
+
+		if (adev->gfx.rlc.clear_state_obj == NULL) {
+			r = amdgpu_bo_create(adev, dws * 4, PAGE_SIZE, true,
+					     AMDGPU_GEM_DOMAIN_VRAM,
+					     AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED,
+					     NULL, NULL,
+					     &adev->gfx.rlc.clear_state_obj);
+
+			if (r) {
+				dev_warn(adev->dev, "(%d) create RLC c bo failed\n", r);
+				gfx_v6_0_rlc_fini(adev);
+				return r;
+			}
+		}
+		r = amdgpu_bo_reserve(adev->gfx.rlc.clear_state_obj, false);
+		if (unlikely(r != 0)) {
+			gfx_v6_0_rlc_fini(adev);
+			return r;
+		}
+		r = amdgpu_bo_pin(adev->gfx.rlc.clear_state_obj, AMDGPU_GEM_DOMAIN_VRAM,
+				  &adev->gfx.rlc.clear_state_gpu_addr);
+		if (r) {
+			amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
+			dev_warn(adev->dev, "(%d) pin RLC c bo failed\n", r);
+			gfx_v6_0_rlc_fini(adev);
+			return r;
+		}
+
+		r = amdgpu_bo_kmap(adev->gfx.rlc.clear_state_obj, (void **)&adev->gfx.rlc.cs_ptr);
+		if (r) {
+			dev_warn(adev->dev, "(%d) map RLC c bo failed\n", r);
+			gfx_v6_0_rlc_fini(adev);
+			return r;
+		}
+		/* set up the cs buffer */
+		dst_ptr = adev->gfx.rlc.cs_ptr;
+		reg_list_mc_addr = adev->gfx.rlc.clear_state_gpu_addr + 256;
+		dst_ptr[0] = cpu_to_le32(upper_32_bits(reg_list_mc_addr));
+		dst_ptr[1] = cpu_to_le32(lower_32_bits(reg_list_mc_addr));
+		dst_ptr[2] = cpu_to_le32(adev->gfx.rlc.clear_state_size);
+		gfx_v6_0_get_csb_buffer(adev, &dst_ptr[(256/4)]);
+		amdgpu_bo_kunmap(adev->gfx.rlc.clear_state_obj);
+		amdgpu_bo_unreserve(adev->gfx.rlc.clear_state_obj);
+	}
+
+	return 0;
+}
+
+static void gfx_v6_0_enable_lbpw(struct amdgpu_device *adev, bool enable)
+{
+	u32 tmp;
+
+	tmp = RREG32(RLC_LB_CNTL);
+	if (enable)
+		tmp |= LOAD_BALANCE_ENABLE;
+	else
+		tmp &= ~LOAD_BALANCE_ENABLE;
+	WREG32(RLC_LB_CNTL, tmp);
+
+	if (!enable) {
+		gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+		WREG32(SPI_LB_CU_MASK, 0x00ff);
+	}
+
+}
+
+static void gfx_v6_0_wait_for_rlc_serdes(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (RREG32(RLC_SERDES_MASTER_BUSY_0) == 0)
+			break;
+		udelay(1);
+	}
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (RREG32(RLC_SERDES_MASTER_BUSY_1) == 0)
+			break;
+		udelay(1);
+	}
+}
+
+static void gfx_v6_0_update_rlc(struct amdgpu_device *adev, u32 rlc)
+{
+	u32 tmp;
+
+	tmp = RREG32(RLC_CNTL);
+	if (tmp != rlc)
+		WREG32(RLC_CNTL, rlc);
+}
+
+static u32 gfx_v6_0_halt_rlc(struct amdgpu_device *adev)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(RLC_CNTL);
+
+	if (data & RLC_ENABLE) {
+		data &= ~RLC_ENABLE;
+		WREG32(RLC_CNTL, data);
+
+		gfx_v6_0_wait_for_rlc_serdes(adev);
+	}
+
+	return orig;
+}
+
+static void gfx_v6_0_rlc_stop(struct amdgpu_device *adev)
+{
+	WREG32(RLC_CNTL, 0);
+
+	gfx_v6_0_enable_gui_idle_interrupt(adev, false);
+	gfx_v6_0_wait_for_rlc_serdes(adev);
+}
+
+static void gfx_v6_0_rlc_start(struct amdgpu_device *adev)
+{
+	WREG32(RLC_CNTL, RLC_ENABLE);
+
+	gfx_v6_0_enable_gui_idle_interrupt(adev, true);
+
+	udelay(50);
+}
+
+static void gfx_v6_0_rlc_reset(struct amdgpu_device *adev)
+{
+	u32 tmp = RREG32(GRBM_SOFT_RESET);
+
+	tmp |= SOFT_RESET_RLC;
+	WREG32(GRBM_SOFT_RESET, tmp);
+	udelay(50);
+	tmp &= ~SOFT_RESET_RLC;
+	WREG32(GRBM_SOFT_RESET, tmp);
+	udelay(50);
+}
+
+static bool gfx_v6_0_lbpw_supported(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	/* Enable LBPW only for DDR3 */
+	tmp = RREG32(MC_SEQ_MISC0);
+	if ((tmp & 0xF0000000) == 0xB0000000)
+		return true;
+	return false;
+}
+static void gfx_v6_0_init_cg(struct amdgpu_device *adev)
+{
+}
+
+static int gfx_v6_0_rlc_resume(struct amdgpu_device *adev)
+{
+	u32 i;
+	const struct rlc_firmware_header_v1_0 *hdr;
+	const __le32 *fw_data;
+	u32 fw_size;
+
+
+	if (!adev->gfx.rlc_fw)
+		return -EINVAL;
+
+	gfx_v6_0_rlc_stop(adev);
+	gfx_v6_0_rlc_reset(adev);
+	gfx_v6_0_init_pg(adev);
+	gfx_v6_0_init_cg(adev);
+
+	WREG32(RLC_RL_BASE, 0);
+	WREG32(RLC_RL_SIZE, 0);
+	WREG32(RLC_LB_CNTL, 0);
+	WREG32(RLC_LB_CNTR_MAX, 0xffffffff);
+	WREG32(RLC_LB_CNTR_INIT, 0);
+	WREG32(RLC_LB_INIT_CU_MASK, 0xffffffff);
+
+	WREG32(RLC_MC_CNTL, 0);
+	WREG32(RLC_UCODE_CNTL, 0);
+
+	hdr = (const struct rlc_firmware_header_v1_0 *)adev->gfx.rlc_fw->data;
+	fw_size = le32_to_cpu(hdr->header.ucode_size_bytes) / 4;
+	fw_data = (const __le32 *)
+		(adev->gfx.rlc_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+
+	amdgpu_ucode_print_rlc_hdr(&hdr->header);
+
+	for (i = 0; i < fw_size; i++) {
+		WREG32(RLC_UCODE_ADDR, i);
+		WREG32(RLC_UCODE_DATA, le32_to_cpup(fw_data++));
+	}
+	WREG32(RLC_UCODE_ADDR, 0);
+
+	gfx_v6_0_enable_lbpw(adev, gfx_v6_0_lbpw_supported(adev));
+	gfx_v6_0_rlc_start(adev);
+
+	return 0;
+}
+
+static void gfx_v6_0_enable_cgcg(struct amdgpu_device *adev, bool enable)
+{
+	u32 data, orig, tmp;
+
+	orig = data = RREG32(RLC_CGCG_CGLS_CTRL);
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_CGCG)) {
+		gfx_v6_0_enable_gui_idle_interrupt(adev, true);
+
+		WREG32(RLC_GCPM_GENERAL_3, 0x00000080);
+
+		tmp = gfx_v6_0_halt_rlc(adev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00b000ff);
+
+		gfx_v6_0_wait_for_rlc_serdes(adev);
+		gfx_v6_0_update_rlc(adev, tmp);
+
+		WREG32(RLC_SERDES_WR_CTRL, 0x007000ff);
+
+		data |= CGCG_EN | CGLS_EN;
+	} else {
+		gfx_v6_0_enable_gui_idle_interrupt(adev, false);
+
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+		RREG32(CB_CGTT_SCLK_CTRL);
+
+		data &= ~(CGCG_EN | CGLS_EN);
+	}
+
+	if (orig != data)
+		WREG32(RLC_CGCG_CGLS_CTRL, data);
+
+}
+
+static void gfx_v6_0_enable_mgcg(struct amdgpu_device *adev, bool enable)
+{
+
+	u32 data, orig, tmp = 0;
+
+	if (enable && (adev->cg_flags & AMD_CG_SUPPORT_GFX_MGCG)) {
+		orig = data = RREG32(CGTS_SM_CTRL_REG);
+		data = 0x96940200;
+		if (orig != data)
+			WREG32(CGTS_SM_CTRL_REG, data);
+
+		if (adev->cg_flags & AMD_CG_SUPPORT_GFX_CP_LS) {
+			orig = data = RREG32(CP_MEM_SLP_CNTL);
+			data |= CP_MEM_LS_EN;
+			if (orig != data)
+				WREG32(CP_MEM_SLP_CNTL, data);
+		}
+
+		orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
+		data &= 0xffffffc0;
+		if (orig != data)
+			WREG32(RLC_CGTT_MGCG_OVERRIDE, data);
+
+		tmp = gfx_v6_0_halt_rlc(adev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00d000ff);
+
+		gfx_v6_0_update_rlc(adev, tmp);
+	} else {
+		orig = data = RREG32(RLC_CGTT_MGCG_OVERRIDE);
+		data |= 0x00000003;
+		if (orig != data)
+			WREG32(RLC_CGTT_MGCG_OVERRIDE, data);
+
+		data = RREG32(CP_MEM_SLP_CNTL);
+		if (data & CP_MEM_LS_EN) {
+			data &= ~CP_MEM_LS_EN;
+			WREG32(CP_MEM_SLP_CNTL, data);
+		}
+		orig = data = RREG32(CGTS_SM_CTRL_REG);
+		data |= LS_OVERRIDE | OVERRIDE;
+		if (orig != data)
+			WREG32(CGTS_SM_CTRL_REG, data);
+
+		tmp = gfx_v6_0_halt_rlc(adev);
+
+		WREG32(RLC_SERDES_WR_MASTER_MASK_0, 0xffffffff);
+		WREG32(RLC_SERDES_WR_MASTER_MASK_1, 0xffffffff);
+		WREG32(RLC_SERDES_WR_CTRL, 0x00e000ff);
+
+		gfx_v6_0_update_rlc(adev, tmp);
+	}
+}
+/*
+static void gfx_v6_0_update_cg(struct amdgpu_device *adev,
+			       bool enable)
+{
+	gfx_v6_0_enable_gui_idle_interrupt(adev, false);
+	if (enable) {
+		gfx_v6_0_enable_mgcg(adev, true);
+		gfx_v6_0_enable_cgcg(adev, true);
+	} else {
+		gfx_v6_0_enable_cgcg(adev, false);
+		gfx_v6_0_enable_mgcg(adev, false);
+	}
+	gfx_v6_0_enable_gui_idle_interrupt(adev, true);
+}
+*/
+static void gfx_v6_0_enable_sclk_slowdown_on_pu(struct amdgpu_device *adev,
+						bool enable)
+{
+}
+
+static void gfx_v6_0_enable_sclk_slowdown_on_pd(struct amdgpu_device *adev,
+						bool enable)
+{
+}
+
+static void gfx_v6_0_enable_cp_pg(struct amdgpu_device *adev, bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(RLC_PG_CNTL);
+	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_CP))
+		data &= ~0x8000;
+	else
+		data |= 0x8000;
+	if (orig != data)
+		WREG32(RLC_PG_CNTL, data);
+}
+
+static void gfx_v6_0_enable_gds_pg(struct amdgpu_device *adev, bool enable)
+{
+}
+/*
+static void gfx_v6_0_init_cp_pg_table(struct amdgpu_device *adev)
+{
+	const __le32 *fw_data;
+	volatile u32 *dst_ptr;
+	int me, i, max_me = 4;
+	u32 bo_offset = 0;
+	u32 table_offset, table_size;
+
+	if (adev->asic_type == CHIP_KAVERI)
+		max_me = 5;
+
+	if (adev->gfx.rlc.cp_table_ptr == NULL)
+		return;
+
+	dst_ptr = adev->gfx.rlc.cp_table_ptr;
+	for (me = 0; me < max_me; me++) {
+		if (me == 0) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.ce_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.ce_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		} else if (me == 1) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.pfp_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.pfp_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		} else if (me == 2) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.me_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.me_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		} else if (me == 3) {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.mec_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.mec_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		} else {
+			const struct gfx_firmware_header_v1_0 *hdr =
+				(const struct gfx_firmware_header_v1_0 *)adev->gfx.mec2_fw->data;
+			fw_data = (const __le32 *)
+				(adev->gfx.mec2_fw->data +
+				 le32_to_cpu(hdr->header.ucode_array_offset_bytes));
+			table_offset = le32_to_cpu(hdr->jt_offset);
+			table_size = le32_to_cpu(hdr->jt_size);
+		}
+
+		for (i = 0; i < table_size; i ++) {
+			dst_ptr[bo_offset + i] =
+				cpu_to_le32(le32_to_cpu(fw_data[table_offset + i]));
+		}
+
+		bo_offset += table_size;
+	}
+}
+*/
+static void gfx_v6_0_enable_gfx_cgpg(struct amdgpu_device *adev,
+				     bool enable)
+{
+
+	u32 tmp;
+
+	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG)) {
+		tmp = RLC_PUD(0x10) | RLC_PDD(0x10) | RLC_TTPD(0x10) | RLC_MSD(0x10);
+		WREG32(RLC_TTOP_D, tmp);
+
+		tmp = RREG32(RLC_PG_CNTL);
+		tmp |= GFX_PG_ENABLE;
+		WREG32(RLC_PG_CNTL, tmp);
+
+		tmp = RREG32(RLC_AUTO_PG_CTRL);
+		tmp |= AUTO_PG_EN;
+		WREG32(RLC_AUTO_PG_CTRL, tmp);
+	} else {
+		tmp = RREG32(RLC_AUTO_PG_CTRL);
+		tmp &= ~AUTO_PG_EN;
+		WREG32(RLC_AUTO_PG_CTRL, tmp);
+
+		tmp = RREG32(DB_RENDER_CONTROL);
+	}
+}
+
+static u32 gfx_v6_0_get_cu_active_bitmap(struct amdgpu_device *adev,
+					 u32 se, u32 sh)
+{
+
+	u32 mask = 0, tmp, tmp1;
+	int i;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+	gfx_v6_0_select_se_sh(adev, se, sh, 0xffffffff);
+	tmp = RREG32(CC_GC_SHADER_ARRAY_CONFIG);
+	tmp1 = RREG32(GC_USER_SHADER_ARRAY_CONFIG);
+	gfx_v6_0_select_se_sh(adev, 0xffffffff, 0xffffffff, 0xffffffff);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	tmp &= 0xffff0000;
+
+	tmp |= tmp1;
+	tmp >>= 16;
+
+	for (i = 0; i < adev->gfx.config.max_cu_per_sh; i ++) {
+		mask <<= 1;
+		mask |= 1;
+	}
+
+	return (~tmp) & mask;
+}
+
+static void gfx_v6_0_init_ao_cu_mask(struct amdgpu_device *adev)
+{
+	u32 i, j, k, active_cu_number = 0;
+
+	u32 mask, counter, cu_bitmap;
+	u32 tmp = 0;
+
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
+			mask = 1;
+			cu_bitmap = 0;
+			counter  = 0;
+			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k++) {
+				if (gfx_v6_0_get_cu_active_bitmap(adev, i, j) & mask) {
+					if (counter < 2)
+						cu_bitmap |= mask;
+					counter++;
+				}
+				mask <<= 1;
+			}
+
+			active_cu_number += counter;
+			tmp |= (cu_bitmap << (i * 16 + j * 8));
+		}
+	}
+
+	WREG32(RLC_PG_AO_CU_MASK, tmp);
+
+	tmp = RREG32(RLC_MAX_PG_CU);
+	tmp &= ~MAX_PU_CU_MASK;
+	tmp |= MAX_PU_CU(active_cu_number);
+	WREG32(RLC_MAX_PG_CU, tmp);
+}
+
+static void gfx_v6_0_enable_gfx_static_mgpg(struct amdgpu_device *adev,
+					    bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(RLC_PG_CNTL);
+	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_SMG))
+		data |= STATIC_PER_CU_PG_ENABLE;
+	else
+		data &= ~STATIC_PER_CU_PG_ENABLE;
+	if (orig != data)
+		WREG32(RLC_PG_CNTL, data);
+}
+
+static void gfx_v6_0_enable_gfx_dynamic_mgpg(struct amdgpu_device *adev,
+					     bool enable)
+{
+	u32 data, orig;
+
+	orig = data = RREG32(RLC_PG_CNTL);
+	if (enable && (adev->pg_flags & AMD_PG_SUPPORT_GFX_DMG))
+		data |= DYN_PER_CU_PG_ENABLE;
+	else
+		data &= ~DYN_PER_CU_PG_ENABLE;
+	if (orig != data)
+		WREG32(RLC_PG_CNTL, data);
+}
+
+static void gfx_v6_0_init_gfx_cgpg(struct amdgpu_device *adev)
+{
+	u32 tmp;
+
+	WREG32(RLC_SAVE_AND_RESTORE_BASE, adev->gfx.rlc.save_restore_gpu_addr >> 8);
+
+	tmp = RREG32(RLC_PG_CNTL);
+	tmp |= GFX_PG_SRC;
+	WREG32(RLC_PG_CNTL, tmp);
+
+	WREG32(RLC_CLEAR_STATE_RESTORE_BASE, adev->gfx.rlc.clear_state_gpu_addr >> 8);
+
+	tmp = RREG32(RLC_AUTO_PG_CTRL);
+
+	tmp &= ~GRBM_REG_SGIT_MASK;
+	tmp |= GRBM_REG_SGIT(0x700);
+	tmp &= ~PG_AFTER_GRBM_REG_ST_MASK;
+	WREG32(RLC_AUTO_PG_CTRL, tmp);
+}
+
+static void gfx_v6_0_update_gfx_pg(struct amdgpu_device *adev, bool enable)
+{
+	gfx_v6_0_enable_gfx_cgpg(adev, enable);
+	gfx_v6_0_enable_gfx_static_mgpg(adev, enable);
+	gfx_v6_0_enable_gfx_dynamic_mgpg(adev, enable);
+}
+
+static u32 gfx_v6_0_get_csb_size(struct amdgpu_device *adev)
+{
+	u32 count = 0;
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+
+	if (adev->gfx.rlc.cs_data == NULL)
+		return 0;
+
+	/* begin clear state */
+	count += 2;
+	/* context control state */
+	count += 3;
+
+	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
+		for (ext = sect->section; ext->extent != NULL; ++ext) {
+			if (sect->id == SECT_CONTEXT)
+				count += 2 + ext->reg_count;
+			else
+				return 0;
+		}
+	}
+	/* pa_sc_raster_config */
+	count += 3;
+	/* end clear state */
+	count += 2;
+	/* clear state */
+	count += 2;
+
+	return count;
+}
+
+static void gfx_v6_0_get_csb_buffer(struct amdgpu_device *adev,
+				    volatile u32 *buffer)
+{
+	u32 count = 0, i;
+	const struct cs_section_def *sect = NULL;
+	const struct cs_extent_def *ext = NULL;
+
+	if (adev->gfx.rlc.cs_data == NULL)
+		return;
+	if (buffer == NULL)
+		return;
+
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	buffer[count++] = cpu_to_le32(0x80000000);
+	buffer[count++] = cpu_to_le32(0x80000000);
+
+	for (sect = adev->gfx.rlc.cs_data; sect->section != NULL; ++sect) {
+		for (ext = sect->section; ext->extent != NULL; ++ext) {
+			if (sect->id == SECT_CONTEXT) {
+				buffer[count++] =
+					cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, ext->reg_count));
+				buffer[count++] = cpu_to_le32(ext->reg_index - 0xa000);
+				for (i = 0; i < ext->reg_count; i++)
+					buffer[count++] = cpu_to_le32(ext->extent[i]);
+			} else {
+				return;
+			}
+		}
+	}
+
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	buffer[count++] = cpu_to_le32(PA_SC_RASTER_CONFIG - PACKET3_SET_CONTEXT_REG_START);
+
+	switch (adev->asic_type) {
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+		buffer[count++] = cpu_to_le32(0x2a00126a);
+		break;
+	case CHIP_VERDE:
+		buffer[count++] = cpu_to_le32(0x0000124a);
+		break;
+	case CHIP_OLAND:
+		buffer[count++] = cpu_to_le32(0x00000082);
+		break;
+	case CHIP_HAINAN:
+		buffer[count++] = cpu_to_le32(0x00000000);
+		break;
+	default:
+		buffer[count++] = cpu_to_le32(0x00000000);
+		break;
+	}
+
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	buffer[count++] = cpu_to_le32(PACKET3_PREAMBLE_END_CLEAR_STATE);
+
+	buffer[count++] = cpu_to_le32(PACKET3(PACKET3_CLEAR_STATE, 0));
+	buffer[count++] = cpu_to_le32(0);
+}
+
+static void gfx_v6_0_init_pg(struct amdgpu_device *adev)
+{
+	if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG |
+			      AMD_PG_SUPPORT_GFX_SMG |
+			      AMD_PG_SUPPORT_GFX_DMG |
+			      AMD_PG_SUPPORT_CP |
+			      AMD_PG_SUPPORT_GDS |
+			      AMD_PG_SUPPORT_RLC_SMU_HS)) {
+		gfx_v6_0_enable_sclk_slowdown_on_pu(adev, true);
+		gfx_v6_0_enable_sclk_slowdown_on_pd(adev, true);
+		if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) {
+			gfx_v6_0_init_gfx_cgpg(adev);
+			gfx_v6_0_enable_cp_pg(adev, true);
+			gfx_v6_0_enable_gds_pg(adev, true);
+		} else {
+			WREG32(RLC_SAVE_AND_RESTORE_BASE, adev->gfx.rlc.save_restore_gpu_addr >> 8);
+			WREG32(RLC_CLEAR_STATE_RESTORE_BASE, adev->gfx.rlc.clear_state_gpu_addr >> 8);
+
+		}
+		gfx_v6_0_init_ao_cu_mask(adev);
+		gfx_v6_0_update_gfx_pg(adev, true);
+	} else {
+
+		WREG32(RLC_SAVE_AND_RESTORE_BASE, adev->gfx.rlc.save_restore_gpu_addr >> 8);
+		WREG32(RLC_CLEAR_STATE_RESTORE_BASE, adev->gfx.rlc.clear_state_gpu_addr >> 8);
+	}
+}
+
+static void gfx_v6_0_fini_pg(struct amdgpu_device *adev)
+{
+	if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG |
+			      AMD_PG_SUPPORT_GFX_SMG |
+			      AMD_PG_SUPPORT_GFX_DMG |
+			      AMD_PG_SUPPORT_CP |
+			      AMD_PG_SUPPORT_GDS |
+			      AMD_PG_SUPPORT_RLC_SMU_HS)) {
+		gfx_v6_0_update_gfx_pg(adev, false);
+		if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) {
+			gfx_v6_0_enable_cp_pg(adev, false);
+			gfx_v6_0_enable_gds_pg(adev, false);
+		}
+	}
+}
+
+static uint64_t gfx_v6_0_get_gpu_clock_counter(struct amdgpu_device *adev)
+{
+	uint64_t clock;
+
+	mutex_lock(&adev->gfx.gpu_clock_mutex);
+	WREG32(RLC_CAPTURE_GPU_CLOCK_COUNT, 1);
+	clock = (uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_LSB) |
+	        ((uint64_t)RREG32(RLC_GPU_CLOCK_COUNT_MSB) << 32ULL);
+	mutex_unlock(&adev->gfx.gpu_clock_mutex);
+	return clock;
+}
+
+static void gfx_v6_ring_emit_cntxcntl(struct amdgpu_ring *ring, uint32_t flags)
+{
+	amdgpu_ring_write(ring, PACKET3(PACKET3_CONTEXT_CONTROL, 1));
+	amdgpu_ring_write(ring, 0x80000000);
+	amdgpu_ring_write(ring, 0);
+}
+
+static unsigned gfx_v6_0_ring_get_emit_ib_size(struct amdgpu_ring *ring)
+{
+	return
+		6; /* gfx_v6_0_ring_emit_ib */
+}
+
+static unsigned gfx_v6_0_ring_get_dma_frame_size_gfx(struct amdgpu_ring *ring)
+{
+	return
+		5 + /* gfx_v6_0_ring_emit_hdp_flush */
+		5 + /* gfx_v6_0_ring_emit_hdp_invalidate */
+		14 + 14 + 14 + /* gfx_v6_0_ring_emit_fence x3 for user fence, vm fence */
+		7 + 4 + /* gfx_v6_0_ring_emit_pipeline_sync */
+		17 + 6 + /* gfx_v6_0_ring_emit_vm_flush */
+		3; /* gfx_v6_ring_emit_cntxcntl */
+}
+
+static unsigned gfx_v6_0_ring_get_dma_frame_size_compute(struct amdgpu_ring *ring)
+{
+	return
+		5 + /* gfx_v6_0_ring_emit_hdp_flush */
+		5 + /* gfx_v6_0_ring_emit_hdp_invalidate */
+		7 + /* gfx_v6_0_ring_emit_pipeline_sync */
+		17 + /* gfx_v6_0_ring_emit_vm_flush */
+		14 + 14 + 14; /* gfx_v6_0_ring_emit_fence x3 for user fence, vm fence */
+}
+
+static const struct amdgpu_gfx_funcs gfx_v6_0_gfx_funcs = {
+	.get_gpu_clock_counter = &gfx_v6_0_get_gpu_clock_counter,
+	.select_se_sh = &gfx_v6_0_select_se_sh,
+};
+
+static int gfx_v6_0_early_init(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	adev->gfx.num_gfx_rings = GFX6_NUM_GFX_RINGS;
+	adev->gfx.num_compute_rings = GFX6_NUM_COMPUTE_RINGS;
+	adev->gfx.funcs = &gfx_v6_0_gfx_funcs;
+	gfx_v6_0_set_ring_funcs(adev);
+	gfx_v6_0_set_irq_funcs(adev);
+
+	return 0;
+}
+
+static int gfx_v6_0_sw_init(void *handle)
+{
+	struct amdgpu_ring *ring;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+	int i, r;
+
+	r = amdgpu_irq_add_id(adev, 181, &adev->gfx.eop_irq);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_add_id(adev, 184, &adev->gfx.priv_reg_irq);
+	if (r)
+		return r;
+
+	r = amdgpu_irq_add_id(adev, 185, &adev->gfx.priv_inst_irq);
+	if (r)
+		return r;
+
+	gfx_v6_0_scratch_init(adev);
+
+	r = gfx_v6_0_init_microcode(adev);
+	if (r) {
+		DRM_ERROR("Failed to load gfx firmware!\n");
+		return r;
+	}
+
+	r = gfx_v6_0_rlc_init(adev);
+	if (r) {
+		DRM_ERROR("Failed to init rlc BOs!\n");
+		return r;
+	}
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++) {
+		ring = &adev->gfx.gfx_ring[i];
+		ring->ring_obj = NULL;
+		sprintf(ring->name, "gfx");
+		r = amdgpu_ring_init(adev, ring, 1024,
+				     0x80000000, 0xf,
+				     &adev->gfx.eop_irq, AMDGPU_CP_IRQ_GFX_EOP,
+				     AMDGPU_RING_TYPE_GFX);
+		if (r)
+			return r;
+	}
+
+	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
+		unsigned irq_type;
+
+		if ((i >= 32) || (i >= AMDGPU_MAX_COMPUTE_RINGS)) {
+			DRM_ERROR("Too many (%d) compute rings!\n", i);
+			break;
+		}
+		ring = &adev->gfx.compute_ring[i];
+		ring->ring_obj = NULL;
+		ring->use_doorbell = false;
+		ring->doorbell_index = 0;
+		ring->me = 1;
+		ring->pipe = i;
+		ring->queue = i;
+		sprintf(ring->name, "comp %d.%d.%d", ring->me, ring->pipe, ring->queue);
+		irq_type = AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP + ring->pipe;
+		r = amdgpu_ring_init(adev, ring, 1024,
+				     0x80000000, 0xf,
+				     &adev->gfx.eop_irq, irq_type,
+				     AMDGPU_RING_TYPE_COMPUTE);
+		if (r)
+			return r;
+	}
+
+	return r;
+}
+
+static int gfx_v6_0_sw_fini(void *handle)
+{
+	int i;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	amdgpu_bo_unref(&adev->gds.oa_gfx_bo);
+	amdgpu_bo_unref(&adev->gds.gws_gfx_bo);
+	amdgpu_bo_unref(&adev->gds.gds_gfx_bo);
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
+		amdgpu_ring_fini(&adev->gfx.gfx_ring[i]);
+	for (i = 0; i < adev->gfx.num_compute_rings; i++)
+		amdgpu_ring_fini(&adev->gfx.compute_ring[i]);
+
+	gfx_v6_0_rlc_fini(adev);
+
+	return 0;
+}
+
+static int gfx_v6_0_hw_init(void *handle)
+{
+	int r;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	gfx_v6_0_gpu_init(adev);
+
+	r = gfx_v6_0_rlc_resume(adev);
+	if (r)
+		return r;
+
+	r = gfx_v6_0_cp_resume(adev);
+	if (r)
+		return r;
+
+	adev->gfx.ce_ram_size = 0x8000;
+
+	return r;
+}
+
+static int gfx_v6_0_hw_fini(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	gfx_v6_0_cp_enable(adev, false);
+	gfx_v6_0_rlc_stop(adev);
+	gfx_v6_0_fini_pg(adev);
+
+	return 0;
+}
+
+static int gfx_v6_0_suspend(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	return gfx_v6_0_hw_fini(adev);
+}
+
+static int gfx_v6_0_resume(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	return gfx_v6_0_hw_init(adev);
+}
+
+static bool gfx_v6_0_is_idle(void *handle)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	if (RREG32(GRBM_STATUS) & GRBM_STATUS__GUI_ACTIVE_MASK)
+		return false;
+	else
+		return true;
+}
+
+static int gfx_v6_0_wait_for_idle(void *handle)
+{
+	unsigned i;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	for (i = 0; i < adev->usec_timeout; i++) {
+		if (gfx_v6_0_is_idle(handle))
+			return 0;
+		udelay(1);
+	}
+	return -ETIMEDOUT;
+}
+
+static int gfx_v6_0_soft_reset(void *handle)
+{
+	return 0;
+}
+
+static void gfx_v6_0_set_gfx_eop_interrupt_state(struct amdgpu_device *adev,
+						 enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		cp_int_cntl = RREG32(CP_INT_CNTL_RING0);
+		cp_int_cntl &= ~CP_INT_CNTL_RING__TIME_STAMP_INT_ENABLE_MASK;
+		WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		cp_int_cntl = RREG32(CP_INT_CNTL_RING0);
+		cp_int_cntl |= CP_INT_CNTL_RING__TIME_STAMP_INT_ENABLE_MASK;
+		WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+		break;
+	default:
+		break;
+	}
+}
+
+static void gfx_v6_0_set_compute_eop_interrupt_state(struct amdgpu_device *adev,
+						     int ring,
+						     enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl;
+	switch (state){
+	case AMDGPU_IRQ_STATE_DISABLE:
+		if (ring == 0) {
+			cp_int_cntl = RREG32(CP_INT_CNTL_RING1);
+			cp_int_cntl &= ~CP_INT_CNTL_RING__TIME_STAMP_INT_ENABLE_MASK;
+			WREG32(CP_INT_CNTL_RING1, cp_int_cntl);
+			break;
+		} else {
+			cp_int_cntl = RREG32(CP_INT_CNTL_RING2);
+			cp_int_cntl &= ~CP_INT_CNTL_RING__TIME_STAMP_INT_ENABLE_MASK;
+			WREG32(CP_INT_CNTL_RING2, cp_int_cntl);
+			break;
+
+		}
+	case AMDGPU_IRQ_STATE_ENABLE:
+		if (ring == 0) {
+			cp_int_cntl = RREG32(CP_INT_CNTL_RING1);
+			cp_int_cntl |= CP_INT_CNTL_RING__TIME_STAMP_INT_ENABLE_MASK;
+			WREG32(CP_INT_CNTL_RING1, cp_int_cntl);
+			break;
+		} else {
+			cp_int_cntl = RREG32(CP_INT_CNTL_RING2);
+			cp_int_cntl |= CP_INT_CNTL_RING__TIME_STAMP_INT_ENABLE_MASK;
+			WREG32(CP_INT_CNTL_RING2, cp_int_cntl);
+			break;
+
+		}
+
+	default:
+		BUG();
+		break;
+
+	}
+}
+
+static int gfx_v6_0_set_priv_reg_fault_state(struct amdgpu_device *adev,
+					     struct amdgpu_irq_src *src,
+					     unsigned type,
+					     enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		cp_int_cntl = RREG32(CP_INT_CNTL_RING0);
+		cp_int_cntl &= ~CP_INT_CNTL_RING0__PRIV_REG_INT_ENABLE_MASK;
+		WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		cp_int_cntl = RREG32(CP_INT_CNTL_RING0);
+		cp_int_cntl |= CP_INT_CNTL_RING0__PRIV_REG_INT_ENABLE_MASK;
+		WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v6_0_set_priv_inst_fault_state(struct amdgpu_device *adev,
+					      struct amdgpu_irq_src *src,
+					      unsigned type,
+					      enum amdgpu_interrupt_state state)
+{
+	u32 cp_int_cntl;
+
+	switch (state) {
+	case AMDGPU_IRQ_STATE_DISABLE:
+		cp_int_cntl = RREG32(CP_INT_CNTL_RING0);
+		cp_int_cntl &= ~CP_INT_CNTL_RING0__PRIV_INSTR_INT_ENABLE_MASK;
+		WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+		break;
+	case AMDGPU_IRQ_STATE_ENABLE:
+		cp_int_cntl = RREG32(CP_INT_CNTL_RING0);
+		cp_int_cntl |= CP_INT_CNTL_RING0__PRIV_INSTR_INT_ENABLE_MASK;
+		WREG32(CP_INT_CNTL_RING0, cp_int_cntl);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int gfx_v6_0_set_eop_interrupt_state(struct amdgpu_device *adev,
+					    struct amdgpu_irq_src *src,
+					    unsigned type,
+					    enum amdgpu_interrupt_state state)
+{
+	switch (type) {
+	case AMDGPU_CP_IRQ_GFX_EOP:
+		gfx_v6_0_set_gfx_eop_interrupt_state(adev, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE0_EOP:
+		gfx_v6_0_set_compute_eop_interrupt_state(adev, 0, state);
+		break;
+	case AMDGPU_CP_IRQ_COMPUTE_MEC1_PIPE1_EOP:
+		gfx_v6_0_set_compute_eop_interrupt_state(adev, 1, state);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int gfx_v6_0_eop_irq(struct amdgpu_device *adev,
+			    struct amdgpu_irq_src *source,
+			    struct amdgpu_iv_entry *entry)
+{
+	switch (entry->ring_id) {
+	case 0:
+		amdgpu_fence_process(&adev->gfx.gfx_ring[0]);
+		break;
+	case 1:
+	case 2:
+		amdgpu_fence_process(&adev->gfx.compute_ring[entry->ring_id -1]);
+		break;
+	default:
+		break;
+	}
+	return 0;
+}
+
+static int gfx_v6_0_priv_reg_irq(struct amdgpu_device *adev,
+				 struct amdgpu_irq_src *source,
+				 struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal register access in command stream\n");
+	schedule_work(&adev->reset_work);
+	return 0;
+}
+
+static int gfx_v6_0_priv_inst_irq(struct amdgpu_device *adev,
+				  struct amdgpu_irq_src *source,
+				  struct amdgpu_iv_entry *entry)
+{
+	DRM_ERROR("Illegal instruction in command stream\n");
+	schedule_work(&adev->reset_work);
+	return 0;
+}
+
+static int gfx_v6_0_set_clockgating_state(void *handle,
+					  enum amd_clockgating_state state)
+{
+	bool gate = false;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	if (state == AMD_CG_STATE_GATE)
+		gate = true;
+
+	gfx_v6_0_enable_gui_idle_interrupt(adev, false);
+	if (gate) {
+		gfx_v6_0_enable_mgcg(adev, true);
+		gfx_v6_0_enable_cgcg(adev, true);
+	} else {
+		gfx_v6_0_enable_cgcg(adev, false);
+		gfx_v6_0_enable_mgcg(adev, false);
+	}
+	gfx_v6_0_enable_gui_idle_interrupt(adev, true);
+
+	return 0;
+}
+
+static int gfx_v6_0_set_powergating_state(void *handle,
+					  enum amd_powergating_state state)
+{
+	bool gate = false;
+	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
+
+	if (state == AMD_PG_STATE_GATE)
+		gate = true;
+
+	if (adev->pg_flags & (AMD_PG_SUPPORT_GFX_PG |
+			      AMD_PG_SUPPORT_GFX_SMG |
+			      AMD_PG_SUPPORT_GFX_DMG |
+			      AMD_PG_SUPPORT_CP |
+			      AMD_PG_SUPPORT_GDS |
+			      AMD_PG_SUPPORT_RLC_SMU_HS)) {
+		gfx_v6_0_update_gfx_pg(adev, gate);
+		if (adev->pg_flags & AMD_PG_SUPPORT_GFX_PG) {
+			gfx_v6_0_enable_cp_pg(adev, gate);
+			gfx_v6_0_enable_gds_pg(adev, gate);
+		}
+	}
+
+	return 0;
+}
+
+const struct amd_ip_funcs gfx_v6_0_ip_funcs = {
+	.name = "gfx_v6_0",
+	.early_init = gfx_v6_0_early_init,
+	.late_init = NULL,
+	.sw_init = gfx_v6_0_sw_init,
+	.sw_fini = gfx_v6_0_sw_fini,
+	.hw_init = gfx_v6_0_hw_init,
+	.hw_fini = gfx_v6_0_hw_fini,
+	.suspend = gfx_v6_0_suspend,
+	.resume = gfx_v6_0_resume,
+	.is_idle = gfx_v6_0_is_idle,
+	.wait_for_idle = gfx_v6_0_wait_for_idle,
+	.soft_reset = gfx_v6_0_soft_reset,
+	.set_clockgating_state = gfx_v6_0_set_clockgating_state,
+	.set_powergating_state = gfx_v6_0_set_powergating_state,
+};
+
+static const struct amdgpu_ring_funcs gfx_v6_0_ring_funcs_gfx = {
+	.get_rptr = gfx_v6_0_ring_get_rptr,
+	.get_wptr = gfx_v6_0_ring_get_wptr,
+	.set_wptr = gfx_v6_0_ring_set_wptr_gfx,
+	.parse_cs = NULL,
+	.emit_ib = gfx_v6_0_ring_emit_ib,
+	.emit_fence = gfx_v6_0_ring_emit_fence,
+	.emit_pipeline_sync = gfx_v6_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = gfx_v6_0_ring_emit_vm_flush,
+	.emit_hdp_flush = gfx_v6_0_ring_emit_hdp_flush,
+	.emit_hdp_invalidate = gfx_v6_0_ring_emit_hdp_invalidate,
+	.test_ring = gfx_v6_0_ring_test_ring,
+	.test_ib = gfx_v6_0_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.emit_cntxcntl = gfx_v6_ring_emit_cntxcntl,
+	.get_emit_ib_size = gfx_v6_0_ring_get_emit_ib_size,
+	.get_dma_frame_size = gfx_v6_0_ring_get_dma_frame_size_gfx,
+};
+
+static const struct amdgpu_ring_funcs gfx_v6_0_ring_funcs_compute = {
+	.get_rptr = gfx_v6_0_ring_get_rptr,
+	.get_wptr = gfx_v6_0_ring_get_wptr,
+	.set_wptr = gfx_v6_0_ring_set_wptr_compute,
+	.parse_cs = NULL,
+	.emit_ib = gfx_v6_0_ring_emit_ib,
+	.emit_fence = gfx_v6_0_ring_emit_fence,
+	.emit_pipeline_sync = gfx_v6_0_ring_emit_pipeline_sync,
+	.emit_vm_flush = gfx_v6_0_ring_emit_vm_flush,
+	.emit_hdp_flush = gfx_v6_0_ring_emit_hdp_flush,
+	.emit_hdp_invalidate = gfx_v6_0_ring_emit_hdp_invalidate,
+	.test_ring = gfx_v6_0_ring_test_ring,
+	.test_ib = gfx_v6_0_ring_test_ib,
+	.insert_nop = amdgpu_ring_insert_nop,
+	.get_emit_ib_size = gfx_v6_0_ring_get_emit_ib_size,
+	.get_dma_frame_size = gfx_v6_0_ring_get_dma_frame_size_compute,
+};
+
+static void gfx_v6_0_set_ring_funcs(struct amdgpu_device *adev)
+{
+	int i;
+
+	for (i = 0; i < adev->gfx.num_gfx_rings; i++)
+		adev->gfx.gfx_ring[i].funcs = &gfx_v6_0_ring_funcs_gfx;
+	for (i = 0; i < adev->gfx.num_compute_rings; i++)
+		adev->gfx.compute_ring[i].funcs = &gfx_v6_0_ring_funcs_compute;
+}
+
+static const struct amdgpu_irq_src_funcs gfx_v6_0_eop_irq_funcs = {
+	.set = gfx_v6_0_set_eop_interrupt_state,
+	.process = gfx_v6_0_eop_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v6_0_priv_reg_irq_funcs = {
+	.set = gfx_v6_0_set_priv_reg_fault_state,
+	.process = gfx_v6_0_priv_reg_irq,
+};
+
+static const struct amdgpu_irq_src_funcs gfx_v6_0_priv_inst_irq_funcs = {
+	.set = gfx_v6_0_set_priv_inst_fault_state,
+	.process = gfx_v6_0_priv_inst_irq,
+};
+
+static void gfx_v6_0_set_irq_funcs(struct amdgpu_device *adev)
+{
+	adev->gfx.eop_irq.num_types = AMDGPU_CP_IRQ_LAST;
+	adev->gfx.eop_irq.funcs = &gfx_v6_0_eop_irq_funcs;
+
+	adev->gfx.priv_reg_irq.num_types = 1;
+	adev->gfx.priv_reg_irq.funcs = &gfx_v6_0_priv_reg_irq_funcs;
+
+	adev->gfx.priv_inst_irq.num_types = 1;
+	adev->gfx.priv_inst_irq.funcs = &gfx_v6_0_priv_inst_irq_funcs;
+}
+
+static void gfx_v6_0_get_cu_info(struct amdgpu_device *adev)
+{
+	int i, j, k, counter, active_cu_number = 0;
+	u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0;
+	struct amdgpu_cu_info *cu_info = &adev->gfx.cu_info;
+
+	memset(cu_info, 0, sizeof(*cu_info));
+
+	for (i = 0; i < adev->gfx.config.max_shader_engines; i++) {
+		for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) {
+			mask = 1;
+			ao_bitmap = 0;
+			counter = 0;
+			bitmap = gfx_v6_0_get_cu_active_bitmap(adev, i, j);
+			cu_info->bitmap[i][j] = bitmap;
+
+			for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) {
+				if (bitmap & mask) {
+					if (counter < 2)
+						ao_bitmap |= mask;
+					counter ++;
+				}
+				mask <<= 1;
+			}
+			active_cu_number += counter;
+			ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8));
+		}
+	}
+
+	cu_info->number = active_cu_number;
+	cu_info->ao_cu_mask = ao_cu_mask;
+}