drm for 5.13-rc1

- printk fourcc modifier support added %p4cc
 
 core:
 - drm_crtc_commit_wait
 - atomic plane state helpers reworked for full state
 - dma-buf heaps API rework
 - edid: rework and improvements for displayid
 
 dp-mst:
  - better topology logging
 
 bridge:
 - Chipone ICN6211
 - Lontium LT8912B
 - anx7625 regulator support
 
 panel:
 - fix lt9611 4k panels handling
 
 simple-kms:
 - add plane state helpers
 
 ttm:
 - debugfs support
 - removal of unused sysfs
 - ignore signaled moved fences
 - ioremap buffer according to mem caching
 
 i915:
 - Alderlake S enablement
 - Conversion to dma_resv_locking
 - Bring back watchdog timeout support
 - legacy ioctl cleanups
 - add GEM TDDO and RFC process
 - DG1 LMEM preparation work
 - intel_display.c refactoring
 - Gen9/TGL PCH combination support
 - eDP MSO Support
 - multiple PSR instance support
 - Link training debug updates
 - Disable PSR2 support on JSL/EHL
 - DDR5/LPDDR5 support for bw calcs
 - LSPCON limited to gen9/10 platforms
 - HSW/BDW async flip/VTd corruption workaround
 = SAGV watermakr fixes
 - SNB hard hang on ring resume fix
 - Limit imported dma-buf size
 - move to use new tasklet API
 - refactor KBL/TGL/ADL-S display/gt steppings
 - refactoring legacy DP/HDMI, FB plane code out
 
 amdgpu:
 - uapi: add ioctl to query video capabilities
 - Iniital AMD Freesync HDMI support
 - Initial Adebaran support
 - 10bpc dithering improvements
 - DCN secure display support
 - Drop legacy IO BAR requirements
 - PCIE/S0ix/RAS/Prime/Reset fixes
 - Display ASSR support
 - SMU gfx busy queues for RV/PCO
 - Initial LTTPR display work
 
 amdkfd:
 - MMU notifier fixes
 - APU fixes
 
 radeon:
 - debugfs cleanps
 - fw error handling ifix
 - Flexible array cleanups
 
 msm:
 - big DSI phy/pll cleanup
 - sc7280 initial support
 - commong bandwidth scaling path
 - shrinker locking contention fixes
 - unpin/swap support for GEM objcets
 
 ast:
 - cursor plane handling reworked
 
 tegra:
 - don't register DP AUX channels before connectors
 
 zynqmp:
 - fix OOB struct padding memset
 
 gma500:
 - drop ttm and medfield support
 
 exynos:
 - request_irq cleanup function
 
 mediatek:
 - fine tune line time for EOTp
 - MT8192 dpi support
 - atomic crtc config updates
 - don't support HDMI connector creation
 
 mxsdb:
 - imx8mm support
 
 panfrost:
 -= MMU IRQ handling rework
 
 qxl:
 - locking fixes
 - resource deallocation changes
 
 sun4i:
 - add alpha properties to UI/VI layers
 
 vc4:
 - RPi4 CEC support
 
 vmwgfx:
 - doc cleanups
 
 arc:
 - moved to drm/tiny
 -----BEGIN PGP SIGNATURE-----
 
 iQIcBAABAgAGBQJgiNSVAAoJEAx081l5xIa+fvYP/1206BfOYOx5opt5K3By06ZY
 zrOsbeaqFdHzfUR7xVwO4vqQNhkX4Pt8H/U7uYZx8PRdrXzGENwWLIaIskyUrKOd
 BtwNqUr0ZXJGDlGg26StnUHKeAXuYXlpBKLta5y4LUTkI+bm6V/oVaDMq4dnah70
 2CXS4C2mnaFRLBzuRlraxoGFN4eZkz6Waeyo6PJxn/l2GE2gw+jho0Yrh8e8F2w5
 EjQeNF22/uHwznov03XFJlyugecuBDbE8A6Ma/znnkVdBXcT94eUMugbKOKi4Nn6
 PuJOEdJxmj/9s3oi6kBERc8dvpOj0O+8Vp+xOzn2U3BVXebvu7VoJsq6FcAvL5lN
 ltj4iErxUlEud2GRIVUMx8OTFiKj4ThRFJ2/8Uf22r3P7RHO5E9BLnZBzqIAhDVr
 s2cDBMItcxcVHRCmE04h12XAO4libZBb2TVjbqG94Acq7beR76pMszFrmxPmHBEm
 NGe1s7+ajxMzsq/NIsk4XAhqSmJo6+ujKyyVnrgvKUVeEaWW1U4YvjhJaetnP4fB
 47gF24wOSNFwiCUZlqaIpp/MR4Z8YmaJ7tayWQq4Oj/neWe/yc8xQgQIuE8GL20j
 P9eNQNvlBnoxkz275M9x4kVhJ5FRjr7OYnd3sFVnALuj6fnL3Z1RXLqI1lNtIz1d
 YM89veZuNxMaiDz8roPH
 =bLWZ
 -----END PGP SIGNATURE-----

Merge tag 'drm-next-2021-04-28' of git://anongit.freedesktop.org/drm/drm

Pull drm updates from Dave Airlie:
 "The usual lots of work all over the place.

  i915 has gotten some Alderlake work and prelim DG1 code, along with a
  major locking rework over the GEM code, and brings back the property
  of timing out long running jobs using a watchdog. amdgpu has some
  Alderbran support (new GPU), freesync HDMI support along with a lot
  other fixes.

  Outside of the drm, there is a new printf specifier added which should
  have all the correct acks/sobs:

   - printk fourcc modifier support added %p4cc

  Summary:

  core:
   - drm_crtc_commit_wait
   - atomic plane state helpers reworked for full state
   - dma-buf heaps API rework
   - edid: rework and improvements for displayid

  dp-mst:
   - better topology logging

  bridge:
   - Chipone ICN6211
   - Lontium LT8912B
   - anx7625 regulator support

  panel:
   - fix lt9611 4k panels handling

  simple-kms:
   - add plane state helpers

  ttm:
   - debugfs support
   - removal of unused sysfs
   - ignore signaled moved fences
   - ioremap buffer according to mem caching

  i915:
   - Alderlake S enablement
   - Conversion to dma_resv_locking
   - Bring back watchdog timeout support
   - legacy ioctl cleanups
   - add GEM TDDO and RFC process
   - DG1 LMEM preparation work
   - intel_display.c refactoring
   - Gen9/TGL PCH combination support
   - eDP MSO Support
   - multiple PSR instance support
   - Link training debug updates
   - Disable PSR2 support on JSL/EHL
   - DDR5/LPDDR5 support for bw calcs
   - LSPCON limited to gen9/10 platforms
   - HSW/BDW async flip/VTd corruption workaround
   - SAGV watermark fixes
   - SNB hard hang on ring resume fix
   - Limit imported dma-buf size
   - move to use new tasklet API
   - refactor KBL/TGL/ADL-S display/gt steppings
   - refactoring legacy DP/HDMI, FB plane code out

  amdgpu:
   - uapi: add ioctl to query video capabilities
   - Iniital AMD Freesync HDMI support
   - Initial Adebaran support
   - 10bpc dithering improvements
   - DCN secure display support
   - Drop legacy IO BAR requirements
   - PCIE/S0ix/RAS/Prime/Reset fixes
   - Display ASSR support
   - SMU gfx busy queues for RV/PCO
   - Initial LTTPR display work

  amdkfd:
   - MMU notifier fixes
   - APU fixes

  radeon:
   - debugfs cleanps
   - fw error handling ifix
   - Flexible array cleanups

  msm:
   - big DSI phy/pll cleanup
   - sc7280 initial support
   - commong bandwidth scaling path
   - shrinker locking contention fixes
   - unpin/swap support for GEM objcets

  ast:
   - cursor plane handling reworked

  tegra:
   - don't register DP AUX channels before connectors

  zynqmp:
   - fix OOB struct padding memset

  gma500:
   - drop ttm and medfield support

  exynos:
   - request_irq cleanup function

  mediatek:
   - fine tune line time for EOTp
   - MT8192 dpi support
   - atomic crtc config updates
   - don't support HDMI connector creation

  mxsdb:
   - imx8mm support

  panfrost:
   - MMU IRQ handling rework

  qxl:
   - locking fixes
   - resource deallocation changes

  sun4i:
   - add alpha properties to UI/VI layers

  vc4:
   - RPi4 CEC support

  vmwgfx:
   - doc cleanups

  arc:
   - moved to drm/tiny"

* tag 'drm-next-2021-04-28' of git://anongit.freedesktop.org/drm/drm: (1390 commits)
  drm/ttm: Don't count pages in SG BOs against pages_limit
  drm/ttm: fix return value check
  drm/bridge: lt8912b: fix incorrect handling of of_* return values
  drm: bridge: fix LONTIUM use of mipi_dsi_() functions
  drm: bridge: fix ANX7625 use of mipi_dsi_() functions
  drm/amdgpu: page retire over debugfs mechanism
  drm/radeon: Fix a missing check bug in radeon_dp_mst_detect()
  drm/amd/display: Fix the Wunused-function warning
  drm/radeon/r600: Fix variables that are not used after assignment
  drm/amdgpu/smu7: fix CAC setting on TOPAZ
  drm/amd/display: Update DCN302 SR Exit Latency
  drm/amdgpu: enable ras eeprom on aldebaran
  drm/amdgpu: RAS harvest on driver load
  drm/amdgpu: add ras aldebaran ras eeprom driver
  drm/amd/pm: increase time out value when sending msg to SMU
  drm/amdgpu: add DMUB outbox event IRQ source define/complete/debug flag
  drm/amd/pm: add the callback to get vbios bootup values for vangogh
  drm/radeon: Fix size overflow
  drm/amdgpu: Fix size overflow
  drm/amdgpu: move mmhub ras_func init to ip specific file
  ...

Documentation/core-api/printk-formats.rst

@@ -591,6 +591,24 @@ For printing netdev_features_t.
 
 Passed by reference.
 
+V4L2 and DRM FourCC code (pixel format)
+---------------------------------------
+
+::
+
+	%p4cc
+
+Print a FourCC code used by V4L2 or DRM, including format endianness and
+its numerical value as hexadecimal.
+
+Passed by reference.
+
+Examples::
+
+	%p4cc	BG12 little-endian (0x32314742)
+	%p4cc	Y10  little-endian (0x20303159)
+	%p4cc	NV12 big-endian (0xb231564e)
+
 Thanks
 ======
 

Documentation/devicetree/bindings/display/allwinner,sun8i-a83t-dw-hdmi.yaml

@@ -12,8 +12,8 @@ description: |
   and CEC.
 
   These DT bindings follow the Synopsys DWC HDMI TX bindings defined
-  in Documentation/devicetree/bindings/display/bridge/dw_hdmi.txt with
+  in bridge/synopsys,dw-hdmi.yaml with the following device-specific
-  the following device-specific properties.
+  properties.
 
 maintainers:
   - Chen-Yu Tsai <wens@csie.org>

Documentation/devicetree/bindings/display/brcm,bcm2711-hdmi.yaml

@@ -109,7 +109,7 @@ required:
   - resets
   - ddc
 
-additionalProperties: false
+unevaluatedProperties: false
 
 examples:
   - |

Documentation/devicetree/bindings/display/bridge/analogix,anx7625.yaml

@@ -34,6 +34,15 @@ properties:
     description: used for reset chip control, RESET_N pin B7.
     maxItems: 1
 
+  vdd10-supply:
+    description: Regulator that provides the supply 1.0V power.
+
+  vdd18-supply:
+    description: Regulator that provides the supply 1.8V power.
+
+  vdd33-supply:
+    description: Regulator that provides the supply 3.3V power.
+
   ports:
     $ref: /schemas/graph.yaml#/properties/ports
 
@@ -55,6 +64,9 @@ properties:
 required:
   - compatible
   - reg
+  - vdd10-supply
+  - vdd18-supply
+  - vdd33-supply
   - ports
 
 additionalProperties: false
@@ -72,6 +84,9 @@ examples:
             reg = <0x58>;
             enable-gpios = <&pio 45 GPIO_ACTIVE_HIGH>;
             reset-gpios = <&pio 73 GPIO_ACTIVE_HIGH>;
+            vdd10-supply = <&pp1000_mipibrdg>;
+            vdd18-supply = <&pp1800_mipibrdg>;
+            vdd33-supply = <&pp3300_mipibrdg>;
 
             ports {
                 #address-cells = <1>;

Documentation/devicetree/bindings/display/bridge/chipone,icn6211.yaml

@@ -0,0 +1,99 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/bridge/chipone,icn6211.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Chipone ICN6211 MIPI-DSI to RGB Converter bridge
+
+maintainers:
+  - Jagan Teki <jagan@amarulasolutions.com>
+
+description: |
+  ICN6211 is MIPI-DSI to RGB Converter bridge from chipone.
+
+  It has a flexible configuration of MIPI DSI signal input and
+  produce RGB565, RGB666, RGB888 output format.
+
+properties:
+  compatible:
+    enum:
+      - chipone,icn6211
+
+  reg:
+    maxItems: 1
+    description: virtual channel number of a DSI peripheral
+
+  enable-gpios:
+    description: Bridge EN pin, chip is reset when EN is low.
+
+  vdd1-supply:
+    description: A 1.8V/2.5V/3.3V supply that power the MIPI RX.
+
+  vdd2-supply:
+    description: A 1.8V/2.5V/3.3V supply that power the PLL.
+
+  vdd3-supply:
+    description: A 1.8V/2.5V/3.3V supply that power the RGB output.
+
+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description:
+          Video port for MIPI DSI input
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description:
+          Video port for MIPI DPI output (panel or connector).
+
+    required:
+      - port@0
+      - port@1
+
+required:
+  - compatible
+  - reg
+  - enable-gpios
+  - ports
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+
+    dsi {
+      #address-cells = <1>;
+      #size-cells = <0>;
+
+      bridge@0 {
+        compatible = "chipone,icn6211";
+        reg = <0>;
+        enable-gpios = <&r_pio 0 5 GPIO_ACTIVE_HIGH>; /* LCD-RST: PL5 */
+
+        ports {
+          #address-cells = <1>;
+          #size-cells = <0>;
+
+          port@0 {
+            reg = <0>;
+
+            bridge_in_dsi: endpoint {
+              remote-endpoint = <&dsi_out_bridge>;
+            };
+          };
+
+          port@1 {
+            reg = <1>;
+
+            bridge_out_panel: endpoint {
+              remote-endpoint = <&panel_out_bridge>;
+            };
+          };
+        };
+      };
+    };

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

@@ -1,33 +0,0 @@
-Synopsys DesignWare HDMI TX Encoder
-===================================
-
-This document defines device tree properties for the Synopsys DesignWare HDMI
-TX Encoder (DWC HDMI TX). It doesn't constitue a device tree binding
-specification by itself but is meant to be referenced by platform-specific
-device tree bindings.
-
-When referenced from platform device tree bindings the properties defined in
-this document are defined as follows. The platform device tree bindings are
-responsible for defining whether each property is required or optional.
-
-- reg: Memory mapped base address and length of the DWC HDMI TX registers.
-
-- reg-io-width: Width of the registers specified by the reg property. The
-  value is expressed in bytes and must be equal to 1 or 4 if specified. The
-  register width defaults to 1 if the property is not present.
-
-- interrupts: Reference to the DWC HDMI TX interrupt.
-
-- clocks: References to all the clocks specified in the clock-names property
-  as specified in Documentation/devicetree/bindings/clock/clock-bindings.txt.
-
-- clock-names: The DWC HDMI TX uses the following clocks.
-
-  - "iahb" is the bus clock for either AHB and APB (mandatory).
-  - "isfr" is the internal register configuration clock (mandatory).
-  - "cec" is the HDMI CEC controller main clock (optional).
-
-- ports: The connectivity of the DWC HDMI TX with the rest of the system is
-  expressed in using ports as specified in the device graph bindings defined
-  in Documentation/devicetree/bindings/graph.txt. The numbering of the ports
-  is platform-specific.

Documentation/devicetree/bindings/display/bridge/lontium,lt8912b.yaml

@@ -0,0 +1,102 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/bridge/lontium,lt8912b.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Lontium LT8912B MIPI to HDMI Bridge
+
+maintainers:
+  - Adrien Grassein <adrien.grassein@gmail.com>
+
+description: |
+  The LT8912B is a bridge device which convert DSI to HDMI
+
+properties:
+  compatible:
+    enum:
+      - lontium,lt8912b
+
+  reg:
+    maxItems: 1
+
+  reset-gpios:
+    maxItems: 1
+    description: GPIO connected to active high RESET pin.
+
+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description:
+          Primary MIPI port for MIPI input
+
+        properties:
+          endpoint:
+            $ref: /schemas/media/video-interfaces.yaml#
+            unevaluatedProperties: false
+
+            properties:
+              data-lanes: true
+
+            required:
+              - data-lanes
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: |
+          HDMI port, should be connected to a node compatible with the
+          hdmi-connector binding.
+
+    required:
+      - port@0
+      - port@1
+
+required:
+  - compatible
+  - reg
+  - reset-gpios
+  - ports
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+
+    i2c4 {
+      #address-cells = <1>;
+      #size-cells = <0>;
+
+      hdmi-bridge@48 {
+        compatible = "lontium,lt8912b";
+        reg = <0x48>;
+        reset-gpios = <&max7323 0 GPIO_ACTIVE_LOW>;
+
+        ports {
+          #address-cells = <1>;
+          #size-cells = <0>;
+
+          port@0 {
+            reg = <0>;
+
+            hdmi_out_in: endpoint {
+              data-lanes = <0 1 2 3>;
+              remote-endpoint = <&mipi_dsi_out>;
+            };
+          };
+
+          port@1 {
+              reg = <1>;
+
+              endpoint {
+                  remote-endpoint = <&hdmi_in>;
+              };
+          };
+        };
+      };
+    };
+
+...

Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.txt

@@ -1,88 +0,0 @@
-Renesas Gen3 DWC HDMI TX Encoder
-================================
-
-The HDMI transmitter is a Synopsys DesignWare HDMI 1.4 TX controller IP
-with a companion PHY IP.
-
-These DT bindings follow the Synopsys DWC HDMI TX bindings defined in
-Documentation/devicetree/bindings/display/bridge/dw_hdmi.txt with the
-following device-specific properties.
-
-
-Required properties:
-
-- compatible : Shall contain one or more of
-  - "renesas,r8a774a1-hdmi" for R8A774A1 (RZ/G2M) compatible HDMI TX
-  - "renesas,r8a774b1-hdmi" for R8A774B1 (RZ/G2N) compatible HDMI TX
-  - "renesas,r8a774e1-hdmi" for R8A774E1 (RZ/G2H) compatible HDMI TX
-  - "renesas,r8a7795-hdmi" for R8A7795 (R-Car H3) compatible HDMI TX
-  - "renesas,r8a7796-hdmi" for R8A7796 (R-Car M3-W) compatible HDMI TX
-  - "renesas,r8a77961-hdmi" for R8A77961 (R-Car M3-W+) compatible HDMI TX
-  - "renesas,r8a77965-hdmi" for R8A77965 (R-Car M3-N) compatible HDMI TX
-  - "renesas,rcar-gen3-hdmi" for the generic R-Car Gen3 and RZ/G2 compatible
-			     HDMI TX
-
-    When compatible with generic versions, nodes must list the SoC-specific
-    version corresponding to the platform first, followed by the
-    family-specific version.
-
-- reg: See dw_hdmi.txt.
-- interrupts: HDMI interrupt number
-- clocks: See dw_hdmi.txt.
-- clock-names: Shall contain "iahb" and "isfr" as defined in dw_hdmi.txt.
-- ports: See dw_hdmi.txt. The DWC HDMI shall have one port numbered 0
-  corresponding to the video input of the controller and one port numbered 1
-  corresponding to its HDMI output, and one port numbered 2 corresponding to
-  sound input of the controller. Each port shall have a single endpoint.
-
-Optional properties:
-
-- power-domains: Shall reference the power domain that contains the DWC HDMI,
-  if any.
-
-
-Example:
-
-	hdmi0: hdmi@fead0000 {
-		compatible = "renesas,r8a7795-hdmi", "renesas,rcar-gen3-hdmi";
-		reg = <0 0xfead0000 0 0x10000>;
-		interrupts = <0 389 IRQ_TYPE_LEVEL_HIGH>;
-		clocks = <&cpg CPG_CORE R8A7795_CLK_S0D4>, <&cpg CPG_MOD 729>;
-		clock-names = "iahb", "isfr";
-		power-domains = <&sysc R8A7795_PD_ALWAYS_ON>;
-
-		ports {
-			#address-cells = <1>;
-			#size-cells = <0>;
-			port@0 {
-				reg = <0>;
-				dw_hdmi0_in: endpoint {
-					remote-endpoint = <&du_out_hdmi0>;
-				};
-			};
-			port@1 {
-				reg = <1>;
-				rcar_dw_hdmi0_out: endpoint {
-					remote-endpoint = <&hdmi0_con>;
-				};
-			};
-			port@2 {
-				reg = <2>;
-				rcar_dw_hdmi0_sound_in: endpoint {
-					remote-endpoint = <&hdmi_sound_out>;
-				};
-			};
-		};
-	};
-
-	hdmi0-out {
-		compatible = "hdmi-connector";
-		label = "HDMI0 OUT";
-		type = "a";
-
-		port {
-			hdmi0_con: endpoint {
-				remote-endpoint = <&rcar_dw_hdmi0_out>;
-			};
-		};
-	};

Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.yaml

@@ -0,0 +1,125 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/bridge/renesas,dw-hdmi.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Renesas R-Car DWC HDMI TX Encoder
+
+maintainers:
+  - Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
+
+description: |
+  The HDMI transmitter is a Synopsys DesignWare HDMI 1.4 TX controller IP
+  with a companion PHY IP.
+
+allOf:
+  - $ref: synopsys,dw-hdmi.yaml#
+
+properties:
+  compatible:
+    items:
+      - enum:
+          - renesas,r8a774a1-hdmi # for RZ/G2M compatible HDMI TX
+          - renesas,r8a774b1-hdmi # for RZ/G2N compatible HDMI TX
+          - renesas,r8a774e1-hdmi # for RZ/G2H compatible HDMI TX
+          - renesas,r8a7795-hdmi # for R-Car H3 compatible HDMI TX
+          - renesas,r8a7796-hdmi # for R-Car M3-W compatible HDMI TX
+          - renesas,r8a77961-hdmi # for R-Car M3-W+ compatible HDMI TX
+          - renesas,r8a77965-hdmi # for R-Car M3-N compatible HDMI TX
+      - const: renesas,rcar-gen3-hdmi
+
+  reg-io-width:
+    const: 1
+
+  clocks:
+    maxItems: 2
+
+  clock-names:
+    maxItems: 2
+
+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: Parallel RGB input port
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: HDMI output port
+
+      port@2:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: Sound input port
+
+    required:
+      - port@0
+      - port@1
+      - port@2
+
+  power-domains:
+    maxItems: 1
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - clock-names
+  - interrupts
+  - ports
+
+unevaluatedProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/clock/r8a7795-cpg-mssr.h>
+    #include <dt-bindings/interrupt-controller/irq.h>
+    #include <dt-bindings/power/r8a7795-sysc.h>
+
+    hdmi@fead0000 {
+        compatible = "renesas,r8a7795-hdmi", "renesas,rcar-gen3-hdmi";
+        reg = <0xfead0000 0x10000>;
+        interrupts = <0 389 IRQ_TYPE_LEVEL_HIGH>;
+        clocks = <&cpg CPG_CORE R8A7795_CLK_S0D4>, <&cpg CPG_MOD 729>;
+        clock-names = "iahb", "isfr";
+        power-domains = <&sysc R8A7795_PD_ALWAYS_ON>;
+
+        ports {
+            #address-cells = <1>;
+            #size-cells = <0>;
+            port@0 {
+                reg = <0>;
+                dw_hdmi0_in: endpoint {
+                    remote-endpoint = <&du_out_hdmi0>;
+                };
+            };
+            port@1 {
+                reg = <1>;
+                rcar_dw_hdmi0_out: endpoint {
+                    remote-endpoint = <&hdmi0_con>;
+                };
+            };
+            port@2 {
+                reg = <2>;
+                rcar_dw_hdmi0_sound_in: endpoint {
+                    remote-endpoint = <&hdmi_sound_out>;
+                };
+            };
+        };
+    };
+
+    hdmi0-out {
+        compatible = "hdmi-connector";
+        label = "HDMI0 OUT";
+        type = "a";
+
+        port {
+            hdmi0_con: endpoint {
+                remote-endpoint = <&rcar_dw_hdmi0_out>;
+            };
+        };
+    };
+
+...

Documentation/devicetree/bindings/display/bridge/synopsys,dw-hdmi.yaml

@@ -0,0 +1,55 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/bridge/synopsys,dw-hdmi.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Common Properties for Synopsys DesignWare HDMI TX Controller
+
+maintainers:
+  - Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
+
+description: |
+  This document defines device tree properties for the Synopsys DesignWare HDMI
+  TX controller (DWC HDMI TX) IP core. It doesn't constitute a full device tree
+  binding specification by itself but is meant to be referenced by device tree
+  bindings for the platform-specific integrations of the DWC HDMI TX.
+
+  When referenced from platform device tree bindings the properties defined in
+  this document are defined as follows. The platform device tree bindings are
+  responsible for defining whether each property is required or optional.
+
+properties:
+  reg:
+    maxItems: 1
+
+  reg-io-width:
+    description:
+      Width (in bytes) of the registers specified by the reg property.
+    allOf:
+      - $ref: /schemas/types.yaml#/definitions/uint32
+      - enum: [1, 4]
+    default: 1
+
+  clocks:
+    minItems: 2
+    maxItems: 5
+    items:
+      - description: The bus clock for either AHB and APB
+      - description: The internal register configuration clock
+    additionalItems: true
+
+  clock-names:
+    minItems: 2
+    maxItems: 5
+    items:
+      - const: iahb
+      - const: isfr
+    additionalItems: true
+
+  interrupts:
+    maxItems: 1
+
+additionalProperties: true
+
+...

Documentation/devicetree/bindings/display/fsl,lcdif.yaml

@@ -0,0 +1,110 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/fsl,lcdif.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Freescale/NXP i.MX LCD Interface (LCDIF)
+
+maintainers:
+  - Marek Vasut <marex@denx.de>
+  - Stefan Agner <stefan@agner.ch>
+
+description: |
+  (e)LCDIF display controller found in the Freescale/NXP i.MX SoCs.
+
+properties:
+  compatible:
+    oneOf:
+      - enum:
+          - fsl,imx23-lcdif
+          - fsl,imx28-lcdif
+          - fsl,imx6sx-lcdif
+      - items:
+          - enum:
+              - fsl,imx6sl-lcdif
+              - fsl,imx6sll-lcdif
+              - fsl,imx6ul-lcdif
+              - fsl,imx7d-lcdif
+              - fsl,imx8mm-lcdif
+              - fsl,imx8mq-lcdif
+          - const: fsl,imx6sx-lcdif
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    items:
+      - description: Pixel clock
+      - description: Bus clock
+      - description: Display AXI clock
+    minItems: 1
+
+  clock-names:
+    items:
+      - const: pix
+      - const: axi
+      - const: disp_axi
+    minItems: 1
+
+  interrupts:
+    maxItems: 1
+
+  port:
+    $ref: /schemas/graph.yaml#/properties/port
+    description: The LCDIF output port
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - interrupts
+  - port
+
+additionalProperties: false
+
+allOf:
+  - if:
+      properties:
+        compatible:
+          contains:
+            const: fsl,imx6sx-lcdif
+    then:
+      properties:
+        clocks:
+          minItems: 2
+          maxItems: 3
+        clock-names:
+          minItems: 2
+          maxItems: 3
+      required:
+        - clock-names
+    else:
+      properties:
+        clocks:
+          maxItems: 1
+        clock-names:
+          maxItems: 1
+
+examples:
+  - |
+    #include <dt-bindings/clock/imx6sx-clock.h>
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+
+    display-controller@2220000 {
+        compatible = "fsl,imx6sx-lcdif";
+        reg = <0x02220000 0x4000>;
+        interrupts = <GIC_SPI 5 IRQ_TYPE_LEVEL_HIGH>;
+        clocks = <&clks IMX6SX_CLK_LCDIF1_PIX>,
+                 <&clks IMX6SX_CLK_LCDIF_APB>,
+                 <&clks IMX6SX_CLK_DISPLAY_AXI>;
+        clock-names = "pix", "axi", "disp_axi";
+
+        port {
+            endpoint {
+                remote-endpoint = <&panel_in>;
+            };
+        };
+    };
+
+...

Documentation/devicetree/bindings/display/imx/fsl,imx6-hdmi.yaml

@@ -0,0 +1,126 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/imx/fsl,imx6-hdmi.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Freescale i.MX6 DWC HDMI TX Encoder
+
+maintainers:
+  - Philipp Zabel <p.zabel@pengutronix.de>
+
+description: |
+  The HDMI transmitter is a Synopsys DesignWare HDMI 1.4 TX controller IP
+  with a companion PHY IP.
+
+allOf:
+  - $ref: ../bridge/synopsys,dw-hdmi.yaml#
+
+properties:
+  compatible:
+    enum:
+      - fsl,imx6dl-hdmi
+      - fsl,imx6q-hdmi
+
+  reg-io-width:
+    const: 1
+
+  clocks:
+    maxItems: 2
+
+  clock-names:
+    maxItems: 2
+
+  ddc-i2c-bus:
+    $ref: /schemas/types.yaml#/definitions/phandle
+    description:
+      The HDMI DDC bus can be connected to either a system I2C master or the
+      functionally-reduced I2C master contained in the DWC HDMI. When connected
+      to a system I2C master this property contains a phandle to that I2C
+      master controller.
+
+  gpr:
+    $ref: /schemas/types.yaml#/definitions/phandle
+    description:
+      phandle to the iomuxc-gpr region containing the HDMI multiplexer control
+      register.
+
+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+    description: |
+      This device has four video ports, corresponding to the four inputs of the
+      HDMI multiplexer. Each port shall have a single endpoint.
+
+    properties:
+      port@0:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: First input of the HDMI multiplexer
+
+      port@1:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: Second input of the HDMI multiplexer
+
+      port@2:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: Third input of the HDMI multiplexer
+
+      port@3:
+        $ref: /schemas/graph.yaml#/properties/port
+        description: Fourth input of the HDMI multiplexer
+
+    anyOf:
+      - required:
+          - port@0
+      - required:
+          - port@1
+      - required:
+          - port@2
+      - required:
+          - port@3
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - clock-names
+  - gpr
+  - interrupts
+  - ports
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/clock/imx6qdl-clock.h>
+
+    hdmi: hdmi@120000 {
+        reg = <0x00120000 0x9000>;
+        interrupts = <0 115 0x04>;
+        gpr = <&gpr>;
+        clocks = <&clks IMX6QDL_CLK_HDMI_IAHB>,
+                 <&clks IMX6QDL_CLK_HDMI_ISFR>;
+        clock-names = "iahb", "isfr";
+
+        ports {
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            port@0 {
+                reg = <0>;
+
+                hdmi_mux_0: endpoint {
+                    remote-endpoint = <&ipu1_di0_hdmi>;
+                };
+            };
+
+            port@1 {
+                reg = <1>;
+
+                hdmi_mux_1: endpoint {
+                    remote-endpoint = <&ipu1_di1_hdmi>;
+                };
+            };
+        };
+    };
+
+...

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

@@ -1,65 +0,0 @@
-Freescale i.MX6 DWC HDMI TX Encoder
-===================================
-
-The HDMI transmitter is a Synopsys DesignWare HDMI 1.4 TX controller IP
-with a companion PHY IP.
-
-These DT bindings follow the Synopsys DWC HDMI TX bindings defined in
-Documentation/devicetree/bindings/display/bridge/dw_hdmi.txt with the
-following device-specific properties.
-
-
-Required properties:
-
-- compatible : Shall be one of "fsl,imx6q-hdmi" or "fsl,imx6dl-hdmi".
-- reg: See dw_hdmi.txt.
-- interrupts: HDMI interrupt number
-- clocks: See dw_hdmi.txt.
-- clock-names: Shall contain "iahb" and "isfr" as defined in dw_hdmi.txt.
-- ports: See dw_hdmi.txt. The DWC HDMI shall have between one and four ports,
-  numbered 0 to 3, corresponding to the four inputs of the HDMI multiplexer.
-  Each port shall have a single endpoint.
-- gpr : Shall contain a phandle to the iomuxc-gpr region containing the HDMI
-  multiplexer control register.
-
-Optional properties
-
-- ddc-i2c-bus: The HDMI DDC bus can be connected to either a system I2C master
-  or the functionally-reduced I2C master contained in the DWC HDMI. When
-  connected to a system I2C master this property contains a phandle to that
-  I2C master controller.
-
-
-Example:
-
-	gpr: iomuxc-gpr@20e0000 {
-		/* ... */
-	};
-
-        hdmi: hdmi@120000 {
-                #address-cells = <1>;
-                #size-cells = <0>;
-                compatible = "fsl,imx6q-hdmi";
-                reg = <0x00120000 0x9000>;
-                interrupts = <0 115 0x04>;
-                gpr = <&gpr>;
-                clocks = <&clks 123>, <&clks 124>;
-                clock-names = "iahb", "isfr";
-                ddc-i2c-bus = <&i2c2>;
-
-                port@0 {
-                        reg = <0>;
-
-                        hdmi_mux_0: endpoint {
-                                remote-endpoint = <&ipu1_di0_hdmi>;
-                        };
-                };
-
-                port@1 {
-                        reg = <1>;
-
-                        hdmi_mux_1: endpoint {
-                                remote-endpoint = <&ipu1_di1_hdmi>;
-                        };
-                };
-        };

Documentation/devicetree/bindings/display/mediatek/mediatek,dpi.yaml

@@ -22,6 +22,7 @@ properties:
       - mediatek,mt7623-dpi
       - mediatek,mt8173-dpi
       - mediatek,mt8183-dpi
+      - mediatek,mt8192-dpi
 
   reg:
     maxItems: 1
@@ -50,15 +51,10 @@ properties:
       - const: sleep
 
   port:
-    type: object
+    $ref: /schemas/graph.yaml#/properties/port
     description:
-      Output port node with endpoint definitions as described in
+      Output port node. This port should be connected to the input port of an
-      Documentation/devicetree/bindings/graph.txt. This port should be connected
+      attached HDMI or LVDS encoder chip.
-      to the input port of an attached HDMI or LVDS encoder chip.
-
-    properties:
-      endpoint:
-        type: object
 
 required:
   - compatible

Documentation/devicetree/bindings/display/mxsfb.txt

@@ -1,87 +0,0 @@
-* Freescale MXS LCD Interface (LCDIF)
-
-New bindings:
-=============
-Required properties:
-- compatible:	Should be "fsl,imx23-lcdif" for i.MX23.
-		Should be "fsl,imx28-lcdif" for i.MX28.
-		Should be "fsl,imx6sx-lcdif" for i.MX6SX.
-		Should be "fsl,imx8mq-lcdif" for i.MX8MQ.
-- reg:		Address and length of the register set for LCDIF
-- interrupts:	Should contain LCDIF interrupt
-- clocks:	A list of phandle + clock-specifier pairs, one for each
-		entry in 'clock-names'.
-- clock-names:	A list of clock names. For MXSFB it should contain:
-    - "pix" for the LCDIF block clock
-    - (MX6SX-only) "axi", "disp_axi" for the bus interface clock
-
-Required sub-nodes:
-  - port: The connection to an encoder chip.
-
-Example:
-
-	lcdif1: display-controller@2220000 {
-		compatible = "fsl,imx6sx-lcdif", "fsl,imx28-lcdif";
-		reg = <0x02220000 0x4000>;
-		interrupts = <GIC_SPI 5 IRQ_TYPE_LEVEL_HIGH>;
-		clocks = <&clks IMX6SX_CLK_LCDIF1_PIX>,
-			 <&clks IMX6SX_CLK_LCDIF_APB>,
-			 <&clks IMX6SX_CLK_DISPLAY_AXI>;
-		clock-names = "pix", "axi", "disp_axi";
-
-		port {
-			parallel_out: endpoint {
-				remote-endpoint = <&panel_in_parallel>;
-			};
-		};
-	};
-
-Deprecated bindings:
-====================
-Required properties:
-- compatible:	Should be "fsl,imx23-lcdif" for i.MX23.
-		Should be "fsl,imx28-lcdif" for i.MX28.
-- reg:		Address and length of the register set for LCDIF
-- interrupts:	Should contain LCDIF interrupts
-- display:	phandle to display node (see below for details)
-
-* display node
-
-Required properties:
-- bits-per-pixel:	<16> for RGB565, <32> for RGB888/666.
-- bus-width:		number of data lines.  Could be <8>, <16>, <18> or <24>.
-
-Required sub-node:
-- display-timings:	Refer to binding doc display-timing.txt for details.
-
-Examples:
-
-lcdif@80030000 {
-	compatible = "fsl,imx28-lcdif";
-	reg = <0x80030000 2000>;
-	interrupts = <38 86>;
-
-	display: display {
-		bits-per-pixel = <32>;
-		bus-width = <24>;
-
-		display-timings {
-			native-mode = <&timing0>;
-			timing0: timing0 {
-				clock-frequency = <33500000>;
-				hactive = <800>;
-				vactive = <480>;
-				hfront-porch = <164>;
-				hback-porch = <89>;
-				hsync-len = <10>;
-				vback-porch = <23>;
-				vfront-porch = <10>;
-				vsync-len = <10>;
-				hsync-active = <0>;
-				vsync-active = <0>;
-				de-active = <1>;
-				pixelclk-active = <0>;
-			};
-		};
-	};
-};

Documentation/devicetree/bindings/display/panel/panel-simple.yaml

@@ -161,6 +161,8 @@ properties:
         # Innolux Corporation 12.1" G121X1-L03 XGA (1024x768) TFT LCD panel
       - innolux,g121x1-l03
         # Innolux Corporation 11.6" WXGA (1366x768) TFT LCD panel
+      - innolux,n116bca-ea1
+        # Innolux Corporation 11.6" WXGA (1366x768) TFT LCD panel
       - innolux,n116bge
         # InnoLux 13.3" FHD (1920x1080) eDP TFT LCD panel
       - innolux,n125hce-gn1

Documentation/devicetree/bindings/display/renesas,du.txt

@@ -1,145 +0,0 @@
-* Renesas R-Car Display Unit (DU)
-
-Required Properties:
-
-  - compatible: must be one of the following.
-    - "renesas,du-r8a7742" for R8A7742 (RZ/G1H) compatible DU
-    - "renesas,du-r8a7743" for R8A7743 (RZ/G1M) compatible DU
-    - "renesas,du-r8a7744" for R8A7744 (RZ/G1N) compatible DU
-    - "renesas,du-r8a7745" for R8A7745 (RZ/G1E) compatible DU
-    - "renesas,du-r8a77470" for R8A77470 (RZ/G1C) compatible DU
-    - "renesas,du-r8a774a1" for R8A774A1 (RZ/G2M) compatible DU
-    - "renesas,du-r8a774b1" for R8A774B1 (RZ/G2N) compatible DU
-    - "renesas,du-r8a774c0" for R8A774C0 (RZ/G2E) compatible DU
-    - "renesas,du-r8a774e1" for R8A774E1 (RZ/G2H) compatible DU
-    - "renesas,du-r8a7779" for R8A7779 (R-Car H1) compatible DU
-    - "renesas,du-r8a7790" for R8A7790 (R-Car H2) compatible DU
-    - "renesas,du-r8a7791" for R8A7791 (R-Car M2-W) compatible DU
-    - "renesas,du-r8a7792" for R8A7792 (R-Car V2H) compatible DU
-    - "renesas,du-r8a7793" for R8A7793 (R-Car M2-N) compatible DU
-    - "renesas,du-r8a7794" for R8A7794 (R-Car E2) compatible DU
-    - "renesas,du-r8a7795" for R8A7795 (R-Car H3) compatible DU
-    - "renesas,du-r8a7796" for R8A7796 (R-Car M3-W) compatible DU
-    - "renesas,du-r8a77961" for R8A77961 (R-Car M3-W+) compatible DU
-    - "renesas,du-r8a77965" for R8A77965 (R-Car M3-N) compatible DU
-    - "renesas,du-r8a77970" for R8A77970 (R-Car V3M) compatible DU
-    - "renesas,du-r8a77980" for R8A77980 (R-Car V3H) compatible DU
-    - "renesas,du-r8a77990" for R8A77990 (R-Car E3) compatible DU
-    - "renesas,du-r8a77995" for R8A77995 (R-Car D3) compatible DU
-
-  - reg: the memory-mapped I/O registers base address and length
-
-  - interrupts: Interrupt specifiers for the DU interrupts.
-
-  - clocks: A list of phandles + clock-specifier pairs, one for each entry in
-    the clock-names property.
-  - clock-names: Name of the clocks. This property is model-dependent.
-    - R8A7779 uses a single functional clock. The clock doesn't need to be
-      named.
-    - All other DU instances use one functional clock per channel The
-      functional clocks must be named "du.x" with "x" being the channel
-      numerical index.
-    - In addition to the functional clocks, all DU versions also support
-      externally supplied pixel clocks. Those clocks are optional. When
-      supplied they must be named "dclkin.x" with "x" being the input clock
-      numerical index.
-
-  - renesas,cmms: A list of phandles to the CMM instances present in the SoC,
-    one for each available DU channel. The property shall not be specified for
-    SoCs that do not provide any CMM (such as V3M and V3H).
-
-  - renesas,vsps: A list of phandle and channel index tuples to the VSPs that
-    handle the memory interfaces for the DU channels. The phandle identifies the
-    VSP instance that serves the DU channel, and the channel index identifies
-    the LIF instance in that VSP.
-
-Optional properties:
-  - resets: A list of phandle + reset-specifier pairs, one for each entry in
-    the reset-names property.
-  - reset-names: Names of the resets. This property is model-dependent.
-    - All but R8A7779 use one reset for a group of one or more successive
-      channels. The resets must be named "du.x" with "x" being the numerical
-      index of the lowest channel in the group.
-
-Required nodes:
-
-The connections to the DU output video ports are modeled using the OF graph
-bindings specified in Documentation/devicetree/bindings/graph.txt.
-
-The following table lists for each supported model the port number
-corresponding to each DU output.
-
-                        Port0          Port1          Port2          Port3
------------------------------------------------------------------------------
- R8A7742 (RZ/G1H)       DPAD 0         LVDS 0         LVDS 1         -
- R8A7743 (RZ/G1M)       DPAD 0         LVDS 0         -              -
- R8A7744 (RZ/G1N)       DPAD 0         LVDS 0         -              -
- R8A7745 (RZ/G1E)       DPAD 0         DPAD 1         -              -
- R8A77470 (RZ/G1C)      DPAD 0         DPAD 1         LVDS 0         -
- R8A774A1 (RZ/G2M)      DPAD 0         HDMI 0         LVDS 0         -
- R8A774B1 (RZ/G2N)      DPAD 0         HDMI 0         LVDS 0         -
- R8A774C0 (RZ/G2E)      DPAD 0         LVDS 0         LVDS 1         -
- R8A774E1 (RZ/G2H)      DPAD 0         HDMI 0         LVDS 0         -
- R8A7779 (R-Car H1)     DPAD 0         DPAD 1         -              -
- R8A7790 (R-Car H2)     DPAD 0         LVDS 0         LVDS 1         -
- R8A7791 (R-Car M2-W)   DPAD 0         LVDS 0         -              -
- R8A7792 (R-Car V2H)    DPAD 0         DPAD 1         -              -
- R8A7793 (R-Car M2-N)   DPAD 0         LVDS 0         -              -
- R8A7794 (R-Car E2)     DPAD 0         DPAD 1         -              -
- R8A7795 (R-Car H3)     DPAD 0         HDMI 0         HDMI 1         LVDS 0
- R8A7796 (R-Car M3-W)   DPAD 0         HDMI 0         LVDS 0         -
- R8A77961 (R-Car M3-W+) DPAD 0         HDMI 0         LVDS 0         -
- R8A77965 (R-Car M3-N)  DPAD 0         HDMI 0         LVDS 0         -
- R8A77970 (R-Car V3M)   DPAD 0         LVDS 0         -              -
- R8A77980 (R-Car V3H)   DPAD 0         LVDS 0         -              -
- R8A77990 (R-Car E3)    DPAD 0         LVDS 0         LVDS 1         -
- R8A77995 (R-Car D3)    DPAD 0         LVDS 0         LVDS 1         -
-
-
-Example: R8A7795 (R-Car H3) ES2.0 DU
-
-	du: display@feb00000 {
-		compatible = "renesas,du-r8a7795";
-		reg = <0 0xfeb00000 0 0x80000>;
-		interrupts = <GIC_SPI 256 IRQ_TYPE_LEVEL_HIGH>,
-			     <GIC_SPI 268 IRQ_TYPE_LEVEL_HIGH>,
-			     <GIC_SPI 269 IRQ_TYPE_LEVEL_HIGH>,
-			     <GIC_SPI 270 IRQ_TYPE_LEVEL_HIGH>;
-		clocks = <&cpg CPG_MOD 724>,
-			 <&cpg CPG_MOD 723>,
-			 <&cpg CPG_MOD 722>,
-			 <&cpg CPG_MOD 721>;
-		clock-names = "du.0", "du.1", "du.2", "du.3";
-		resets = <&cpg 724>, <&cpg 722>;
-		reset-names = "du.0", "du.2";
-		renesas,cmms = <&cmm0>, <&cmm1>, <&cmm2>, <&cmm3>;
-		renesas,vsps = <&vspd0 0>, <&vspd1 0>, <&vspd2 0>, <&vspd0 1>;
-
-		ports {
-			#address-cells = <1>;
-			#size-cells = <0>;
-
-			port@0 {
-				reg = <0>;
-				du_out_rgb: endpoint {
-				};
-			};
-			port@1 {
-				reg = <1>;
-				du_out_hdmi0: endpoint {
-					remote-endpoint = <&dw_hdmi0_in>;
-				};
-			};
-			port@2 {
-				reg = <2>;
-				du_out_hdmi1: endpoint {
-					remote-endpoint = <&dw_hdmi1_in>;
-				};
-			};
-			port@3 {
-				reg = <3>;
-				du_out_lvds0: endpoint {
-				};
-			};
-		};
-	};

Documentation/devicetree/bindings/display/renesas,du.yaml

@@ -0,0 +1,831 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/renesas,du.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Renesas R-Car Display Unit (DU)
+
+maintainers:
+  - Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
+
+description: |
+  These DT bindings describe the Display Unit embedded in the Renesas R-Car
+  Gen1, R-Car Gen2, R-Car Gen3, RZ/G1 and RZ/G2 SoCs.
+
+properties:
+  compatible:
+    enum:
+      - renesas,du-r8a7742 # for RZ/G1H compatible DU
+      - renesas,du-r8a7743 # for RZ/G1M compatible DU
+      - renesas,du-r8a7744 # for RZ/G1N compatible DU
+      - renesas,du-r8a7745 # for RZ/G1E compatible DU
+      - renesas,du-r8a77470 # for RZ/G1C compatible DU
+      - renesas,du-r8a774a1 # for RZ/G2M compatible DU
+      - renesas,du-r8a774b1 # for RZ/G2N compatible DU
+      - renesas,du-r8a774c0 # for RZ/G2E compatible DU
+      - renesas,du-r8a774e1 # for RZ/G2H compatible DU
+      - renesas,du-r8a7779 # for R-Car H1 compatible DU
+      - renesas,du-r8a7790 # for R-Car H2 compatible DU
+      - renesas,du-r8a7791 # for R-Car M2-W compatible DU
+      - renesas,du-r8a7792 # for R-Car V2H compatible DU
+      - renesas,du-r8a7793 # for R-Car M2-N compatible DU
+      - renesas,du-r8a7794 # for R-Car E2 compatible DU
+      - renesas,du-r8a7795 # for R-Car H3 compatible DU
+      - renesas,du-r8a7796 # for R-Car M3-W compatible DU
+      - renesas,du-r8a77961 # for R-Car M3-W+ compatible DU
+      - renesas,du-r8a77965 # for R-Car M3-N compatible DU
+      - renesas,du-r8a77970 # for R-Car V3M compatible DU
+      - renesas,du-r8a77980 # for R-Car V3H compatible DU
+      - renesas,du-r8a77990 # for R-Car E3 compatible DU
+      - renesas,du-r8a77995 # for R-Car D3 compatible DU
+
+  reg:
+    maxItems: 1
+
+  # See compatible-specific constraints below.
+  clocks: true
+  clock-names: true
+  interrupts:
+    description: Interrupt specifiers, one per DU channel
+  resets: true
+  reset-names: true
+
+  ports:
+    $ref: /schemas/graph.yaml#/properties/port
+    description: |
+      The connections to the DU output video ports are modeled using the OF
+      graph bindings specified in Documentation/devicetree/bindings/graph.txt.
+      The number of ports and their assignment are model-dependent. Each port
+      shall have a single endpoint.
+
+    patternProperties:
+      "^port@[0-3]$":
+        $ref: /schemas/graph.yaml#/properties/port
+        unevaluatedProperties: false
+
+    required:
+      - port@0
+      - port@1
+
+    unevaluatedProperties: false
+
+  renesas,cmms:
+    $ref: "/schemas/types.yaml#/definitions/phandle-array"
+    description:
+      A list of phandles to the CMM instances present in the SoC, one for each
+      available DU channel.
+
+  renesas,vsps:
+    $ref: "/schemas/types.yaml#/definitions/phandle-array"
+    description:
+      A list of phandle and channel index tuples to the VSPs that handle the
+      memory interfaces for the DU channels. The phandle identifies the VSP
+      instance that serves the DU channel, and the channel index identifies
+      the LIF instance in that VSP.
+
+required:
+  - compatible
+  - reg
+  - clocks
+  - interrupts
+  - resets
+  - ports
+
+allOf:
+  - if:
+      properties:
+        compatible:
+          contains:
+            const: renesas,du-r8a7779
+    then:
+      properties:
+        clocks:
+          minItems: 1
+          maxItems: 3
+          items:
+            - description: Functional clock
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+
+        clock-names:
+          minItems: 1
+          maxItems: 3
+          items:
+            - const: du.0
+            - pattern: '^dclkin\.[01]$'
+            - pattern: '^dclkin\.[01]$'
+
+        interrupts:
+          maxItems: 1
+
+        resets:
+          maxItems: 1
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: DPAD 1
+            # port@2 is TCON, not supported yet
+            port@2: false
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+
+      required:
+        - interrupts
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a7743
+              - renesas,du-r8a7744
+              - renesas,du-r8a7791
+              - renesas,du-r8a7793
+    then:
+      properties:
+        clocks:
+          minItems: 2
+          maxItems: 4
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+
+        clock-names:
+          minItems: 2
+          maxItems: 4
+          items:
+            - const: du.0
+            - const: du.1
+            - pattern: '^dclkin\.[01]$'
+            - pattern: '^dclkin\.[01]$'
+
+        interrupts:
+          maxItems: 2
+
+        resets:
+          maxItems: 1
+
+        reset-names:
+          items:
+            - const: du.0
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: LVDS 0
+            # port@2 is TCON, not supported yet
+            port@2: false
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a7745
+              - renesas,du-r8a7792
+    then:
+      properties:
+        clocks:
+          minItems: 2
+          maxItems: 4
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+
+        clock-names:
+          minItems: 2
+          maxItems: 4
+          items:
+            - const: du.0
+            - const: du.1
+            - pattern: '^dclkin\.[01]$'
+            - pattern: '^dclkin\.[01]$'
+
+        interrupts:
+          maxItems: 2
+
+        resets:
+          maxItems: 1
+
+        reset-names:
+          items:
+            - const: du.0
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: DPAD 1
+            port@2: false
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a7794
+    then:
+      properties:
+        clocks:
+          minItems: 2
+          maxItems: 4
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+
+        clock-names:
+          minItems: 2
+          maxItems: 4
+          items:
+            - const: du.0
+            - const: du.1
+            - pattern: '^dclkin\.[01]$'
+            - pattern: '^dclkin\.[01]$'
+
+        interrupts:
+          maxItems: 2
+
+        resets:
+          maxItems: 1
+
+        reset-names:
+          items:
+            - const: du.0
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: DPAD 1
+            # port@2 is TCON, not supported yet
+            port@2: false
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a77470
+    then:
+      properties:
+        clocks:
+          minItems: 2
+          maxItems: 4
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+
+        clock-names:
+          minItems: 2
+          maxItems: 4
+          items:
+            - const: du.0
+            - const: du.1
+            - pattern: '^dclkin\.[01]$'
+            - pattern: '^dclkin\.[01]$'
+
+        interrupts:
+          maxItems: 2
+
+        resets:
+          maxItems: 1
+
+        reset-names:
+          items:
+            - const: du.0
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: DPAD 1
+            port@2:
+              description: LVDS 0
+            # port@3 is DVENC, not supported yet
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+            - port@2
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a7742
+              - renesas,du-r8a7790
+    then:
+      properties:
+        clocks:
+          minItems: 3
+          maxItems: 6
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: Functional clock for DU2
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+            - description: DU_DOTCLKIN2 input clock
+
+        clock-names:
+          minItems: 3
+          maxItems: 6
+          items:
+            - const: du.0
+            - const: du.1
+            - const: du.2
+            - pattern: '^dclkin\.[012]$'
+            - pattern: '^dclkin\.[012]$'
+            - pattern: '^dclkin\.[012]$'
+
+        interrupts:
+          maxItems: 3
+
+        resets:
+          maxItems: 1
+
+        reset-names:
+          items:
+            - const: du.0
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: LVDS 0
+            port@2:
+              description: LVDS 1
+            # port@3 is TCON, not supported yet
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+            - port@2
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a7795
+    then:
+      properties:
+        clocks:
+          minItems: 4
+          maxItems: 8
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: Functional clock for DU2
+            - description: Functional clock for DU4
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+            - description: DU_DOTCLKIN2 input clock
+            - description: DU_DOTCLKIN3 input clock
+
+        clock-names:
+          minItems: 4
+          maxItems: 8
+          items:
+            - const: du.0
+            - const: du.1
+            - const: du.2
+            - const: du.3
+            - pattern: '^dclkin\.[0123]$'
+            - pattern: '^dclkin\.[0123]$'
+            - pattern: '^dclkin\.[0123]$'
+            - pattern: '^dclkin\.[0123]$'
+
+        interrupts:
+          maxItems: 4
+
+        resets:
+          maxItems: 2
+
+        reset-names:
+          items:
+            - const: du.0
+            - const: du.2
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: HDMI 0
+            port@2:
+              description: HDMI 1
+            port@3:
+              description: LVDS 0
+
+          required:
+            - port@0
+            - port@1
+            - port@2
+            - port@3
+
+        renesas,cmms:
+          minItems: 4
+
+        renesas,vsps:
+          minItems: 4
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+        - renesas,vsps
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a774a1
+              - renesas,du-r8a7796
+              - renesas,du-r8a77961
+    then:
+      properties:
+        clocks:
+          minItems: 3
+          maxItems: 6
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: Functional clock for DU2
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+            - description: DU_DOTCLKIN2 input clock
+
+        clock-names:
+          minItems: 3
+          maxItems: 6
+          items:
+            - const: du.0
+            - const: du.1
+            - const: du.2
+            - pattern: '^dclkin\.[012]$'
+            - pattern: '^dclkin\.[012]$'
+            - pattern: '^dclkin\.[012]$'
+
+        interrupts:
+          maxItems: 3
+
+        resets:
+          maxItems: 2
+
+        reset-names:
+          items:
+            - const: du.0
+            - const: du.2
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: HDMI 0
+            port@2:
+              description: LVDS 0
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+            - port@2
+
+        renesas,cmms:
+          minItems: 3
+
+        renesas,vsps:
+          minItems: 3
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+        - renesas,vsps
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a774b1
+              - renesas,du-r8a774e1
+              - renesas,du-r8a77965
+    then:
+      properties:
+        clocks:
+          minItems: 3
+          maxItems: 6
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: Functional clock for DU3
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+            - description: DU_DOTCLKIN3 input clock
+
+        clock-names:
+          minItems: 3
+          maxItems: 6
+          items:
+            - const: du.0
+            - const: du.1
+            - const: du.3
+            - pattern: '^dclkin\.[013]$'
+            - pattern: '^dclkin\.[013]$'
+            - pattern: '^dclkin\.[013]$'
+
+        interrupts:
+          maxItems: 3
+
+        resets:
+          maxItems: 2
+
+        reset-names:
+          items:
+            - const: du.0
+            - const: du.3
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: HDMI 0
+            port@2:
+              description: LVDS 0
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+            - port@2
+
+        renesas,cmms:
+          minItems: 3
+
+        renesas,vsps:
+          minItems: 3
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+        - renesas,vsps
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a77970
+              - renesas,du-r8a77980
+    then:
+      properties:
+        clocks:
+          minItems: 1
+          maxItems: 2
+          items:
+            - description: Functional clock for DU0
+            - description: DU_DOTCLKIN0 input clock
+
+        clock-names:
+          minItems: 1
+          maxItems: 2
+          items:
+            - const: du.0
+            - const: dclkin.0
+
+        interrupts:
+          maxItems: 1
+
+        resets:
+          maxItems: 1
+
+        reset-names:
+          items:
+            - const: du.0
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: LVDS 0
+            port@2: false
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+
+        renesas,vsps:
+          minItems: 1
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+        - renesas,vsps
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,du-r8a774c0
+              - renesas,du-r8a77990
+              - renesas,du-r8a77995
+    then:
+      properties:
+        clocks:
+          minItems: 2
+          maxItems: 4
+          items:
+            - description: Functional clock for DU0
+            - description: Functional clock for DU1
+            - description: DU_DOTCLKIN0 input clock
+            - description: DU_DOTCLKIN1 input clock
+
+        clock-names:
+          minItems: 2
+          maxItems: 4
+          items:
+            - const: du.0
+            - const: du.1
+            - pattern: '^dclkin\.[01]$'
+            - pattern: '^dclkin\.[01]$'
+
+        interrupts:
+          maxItems: 2
+
+        resets:
+          maxItems: 1
+
+        reset-names:
+          items:
+            - const: du.0
+
+        ports:
+          properties:
+            port@0:
+              description: DPAD 0
+            port@1:
+              description: LVDS 0
+            port@2:
+              description: LVDS 1
+            # port@3 is TCON, not supported yet
+            port@3: false
+
+          required:
+            - port@0
+            - port@1
+            - port@2
+
+        renesas,cmms:
+          minItems: 2
+
+        renesas,vsps:
+          minItems: 2
+
+      required:
+        - clock-names
+        - interrupts
+        - resets
+        - reset-names
+        - renesas,vsps
+
+additionalProperties: false
+
+examples:
+  # R-Car H3 ES2.0 DU
+  - |
+    #include <dt-bindings/clock/renesas-cpg-mssr.h>
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+
+    display@feb00000 {
+        compatible = "renesas,du-r8a7795";
+        reg = <0xfeb00000 0x80000>;
+        interrupts = <GIC_SPI 256 IRQ_TYPE_LEVEL_HIGH>,
+                     <GIC_SPI 268 IRQ_TYPE_LEVEL_HIGH>,
+                     <GIC_SPI 269 IRQ_TYPE_LEVEL_HIGH>,
+                     <GIC_SPI 270 IRQ_TYPE_LEVEL_HIGH>;
+        clocks = <&cpg CPG_MOD 724>,
+                 <&cpg CPG_MOD 723>,
+                 <&cpg CPG_MOD 722>,
+                 <&cpg CPG_MOD 721>;
+        clock-names = "du.0", "du.1", "du.2", "du.3";
+        resets = <&cpg 724>, <&cpg 722>;
+        reset-names = "du.0", "du.2";
+
+        renesas,cmms = <&cmm0>, <&cmm1>, <&cmm2>, <&cmm3>;
+        renesas,vsps = <&vspd0 0>, <&vspd1 0>, <&vspd2 0>, <&vspd0 1>;
+
+        ports {
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            port@0 {
+                reg = <0>;
+                endpoint {
+                    remote-endpoint = <&adv7123_in>;
+                };
+            };
+            port@1 {
+                reg = <1>;
+                endpoint {
+                    remote-endpoint = <&dw_hdmi0_in>;
+                };
+            };
+            port@2 {
+                reg = <2>;
+                endpoint {
+                    remote-endpoint = <&dw_hdmi1_in>;
+                };
+            };
+            port@3 {
+                reg = <3>;
+                endpoint {
+                    remote-endpoint = <&lvds0_in>;
+                };
+            };
+        };
+    };
+
+...

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

@@ -1,74 +0,0 @@
-Rockchip DWC HDMI TX Encoder
-============================
-
-The HDMI transmitter is a Synopsys DesignWare HDMI 1.4 TX controller IP
-with a companion PHY IP.
-
-These DT bindings follow the Synopsys DWC HDMI TX bindings defined in
-Documentation/devicetree/bindings/display/bridge/dw_hdmi.txt with the
-following device-specific properties.
-
-
-Required properties:
-
-- compatible: should be one of the following:
-		"rockchip,rk3228-dw-hdmi"
-		"rockchip,rk3288-dw-hdmi"
-		"rockchip,rk3328-dw-hdmi"
-		"rockchip,rk3399-dw-hdmi"
-- reg: See dw_hdmi.txt.
-- reg-io-width: See dw_hdmi.txt. Shall be 4.
-- interrupts: HDMI interrupt number
-- clocks: See dw_hdmi.txt.
-- clock-names: Shall contain "iahb" and "isfr" as defined in dw_hdmi.txt.
-- ports: See dw_hdmi.txt. The DWC HDMI shall have a single port numbered 0
-  corresponding to the video input of the controller. The port shall have two
-  endpoints, numbered 0 and 1, connected respectively to the vopb and vopl.
-- rockchip,grf: Shall reference the GRF to mux vopl/vopb.
-
-Optional properties
-
-- ddc-i2c-bus: The HDMI DDC bus can be connected to either a system I2C master
-  or the functionally-reduced I2C master contained in the DWC HDMI. When
-  connected to a system I2C master this property contains a phandle to that
-  I2C master controller.
-- clock-names: See dw_hdmi.txt. The "cec" clock is optional.
-- clock-names: May contain "cec" as defined in dw_hdmi.txt.
-- clock-names: May contain "grf", power for grf io.
-- clock-names: May contain "vpll", external clock for some hdmi phy.
-- phys: from general PHY binding: the phandle for the PHY device.
-- phy-names: Should be "hdmi" if phys references an external phy.
-
-Optional pinctrl entry:
-- If you have both a "unwedge" and "default" pinctrl entry, dw_hdmi
-  will switch to the unwedge pinctrl state for 10ms if it ever gets an
-  i2c timeout.  It's intended that this unwedge pinctrl entry will
-  cause the SDA line to be driven low to work around a hardware
-  errata.
-
-Example:
-
-hdmi: hdmi@ff980000 {
-	compatible = "rockchip,rk3288-dw-hdmi";
-	reg = <0xff980000 0x20000>;
-	reg-io-width = <4>;
-	ddc-i2c-bus = <&i2c5>;
-	rockchip,grf = <&grf>;
-	interrupts = <GIC_SPI 103 IRQ_TYPE_LEVEL_HIGH>;
-	clocks = <&cru  PCLK_HDMI_CTRL>, <&cru SCLK_HDMI_HDCP>;
-	clock-names = "iahb", "isfr";
-	ports {
-		hdmi_in: port {
-			#address-cells = <1>;
-			#size-cells = <0>;
-			hdmi_in_vopb: endpoint@0 {
-				reg = <0>;
-				remote-endpoint = <&vopb_out_hdmi>;
-			};
-			hdmi_in_vopl: endpoint@1 {
-				reg = <1>;
-				remote-endpoint = <&vopl_out_hdmi>;
-			};
-		};
-	};
-};

Documentation/devicetree/bindings/display/rockchip/rockchip,dw-hdmi.yaml

@@ -0,0 +1,156 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/display/rockchip/rockchip,dw-hdmi.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Rockchip DWC HDMI TX Encoder
+
+maintainers:
+  - Mark Yao <markyao0591@gmail.com>
+
+description: |
+  The HDMI transmitter is a Synopsys DesignWare HDMI 1.4 TX controller IP
+  with a companion PHY IP.
+
+allOf:
+  - $ref: ../bridge/synopsys,dw-hdmi.yaml#
+
+properties:
+  compatible:
+    enum:
+      - rockchip,rk3228-dw-hdmi
+      - rockchip,rk3288-dw-hdmi
+      - rockchip,rk3328-dw-hdmi
+      - rockchip,rk3399-dw-hdmi
+
+  reg-io-width:
+    const: 4
+
+  clocks:
+    minItems: 2
+    maxItems: 5
+    items:
+      - {}
+      - {}
+      # The next three clocks are all optional, but shall be specified in this
+      # order when present.
+      - description: The HDMI CEC controller main clock
+      - description: Power for GRF IO
+      - description: External clock for some HDMI PHY
+
+  clock-names:
+    minItems: 2
+    maxItems: 5
+    items:
+      - {}
+      - {}
+      - enum:
+          - cec
+          - grf
+          - vpll
+      - enum:
+          - grf
+          - vpll
+      - const: vpll
+
+  ddc-i2c-bus:
+    $ref: /schemas/types.yaml#/definitions/phandle
+    description:
+      The HDMI DDC bus can be connected to either a system I2C master or the
+      functionally-reduced I2C master contained in the DWC HDMI. When connected
+      to a system I2C master this property contains a phandle to that I2C
+      master controller.
+
+  phys:
+    maxItems: 1
+    description: The HDMI PHY
+
+  phy-names:
+    const: hdmi
+
+  pinctrl-names:
+    description:
+      The unwedge pinctrl entry shall drive the DDC SDA line low. This is
+      intended to work around a hardware errata that can cause the DDC I2C
+      bus to be wedged.
+    items:
+      - const: default
+      - const: unwedge
+
+  ports:
+    $ref: /schemas/graph.yaml#/properties/ports
+
+    properties:
+      port:
+        $ref: /schemas/graph.yaml#/$defs/port-base
+        unevaluatedProperties: false
+        description: Input of the DWC HDMI TX
+
+        properties:
+          endpoint@0:
+            $ref: /schemas/graph.yaml#/properties/endpoint
+            description: Connection to the VOPB
+
+          endpoint@1:
+            $ref: /schemas/graph.yaml#/properties/endpoint
+            description: Connection to the VOPL
+
+        required:
+          - endpoint@0
+          - endpoint@1
+
+    required:
+      - port
+
+  rockchip,grf:
+    $ref: /schemas/types.yaml#/definitions/phandle
+    description:
+      phandle to the GRF to mux vopl/vopb.
+
+required:
+  - compatible
+  - reg
+  - reg-io-width
+  - clocks
+  - clock-names
+  - interrupts
+  - ports
+  - rockchip,grf
+
+unevaluatedProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/clock/rk3288-cru.h>
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+    #include <dt-bindings/interrupt-controller/irq.h>
+
+    hdmi: hdmi@ff980000 {
+        compatible = "rockchip,rk3288-dw-hdmi";
+        reg = <0xff980000 0x20000>;
+        reg-io-width = <4>;
+        ddc-i2c-bus = <&i2c5>;
+        rockchip,grf = <&grf>;
+        interrupts = <GIC_SPI 103 IRQ_TYPE_LEVEL_HIGH>;
+        clocks = <&cru  PCLK_HDMI_CTRL>, <&cru SCLK_HDMI_HDCP>;
+        clock-names = "iahb", "isfr";
+
+        ports {
+            port {
+                #address-cells = <1>;
+                #size-cells = <0>;
+
+                hdmi_in_vopb: endpoint@0 {
+                    reg = <0>;
+                    remote-endpoint = <&vopb_out_hdmi>;
+                };
+                hdmi_in_vopl: endpoint@1 {
+                    reg = <1>;
+                    remote-endpoint = <&vopl_out_hdmi>;
+                };
+            };
+        };
+    };
+
+...

Documentation/driver-api/dma-buf.rst

@@ -257,3 +257,79 @@ fences in the kernel. This means:
   userspace is allowed to use userspace fencing or long running compute
   workloads. This also means no implicit fencing for shared buffers in these
   cases.
+
+Recoverable Hardware Page Faults Implications
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Modern hardware supports recoverable page faults, which has a lot of
+implications for DMA fences.
+
+First, a pending page fault obviously holds up the work that's running on the
+accelerator and a memory allocation is usually required to resolve the fault.
+But memory allocations are not allowed to gate completion of DMA fences, which
+means any workload using recoverable page faults cannot use DMA fences for
+synchronization. Synchronization fences controlled by userspace must be used
+instead.
+
+On GPUs this poses a problem, because current desktop compositor protocols on
+Linux rely on DMA fences, which means without an entirely new userspace stack
+built on top of userspace fences, they cannot benefit from recoverable page
+faults. Specifically this means implicit synchronization will not be possible.
+The exception is when page faults are only used as migration hints and never to
+on-demand fill a memory request. For now this means recoverable page
+faults on GPUs are limited to pure compute workloads.
+
+Furthermore GPUs usually have shared resources between the 3D rendering and
+compute side, like compute units or command submission engines. If both a 3D
+job with a DMA fence and a compute workload using recoverable page faults are
+pending they could deadlock:
+
+- The 3D workload might need to wait for the compute job to finish and release
+  hardware resources first.
+
+- The compute workload might be stuck in a page fault, because the memory
+  allocation is waiting for the DMA fence of the 3D workload to complete.
+
+There are a few options to prevent this problem, one of which drivers need to
+ensure:
+
+- Compute workloads can always be preempted, even when a page fault is pending
+  and not yet repaired. Not all hardware supports this.
+
+- DMA fence workloads and workloads which need page fault handling have
+  independent hardware resources to guarantee forward progress. This could be
+  achieved through e.g. through dedicated engines and minimal compute unit
+  reservations for DMA fence workloads.
+
+- The reservation approach could be further refined by only reserving the
+  hardware resources for DMA fence workloads when they are in-flight. This must
+  cover the time from when the DMA fence is visible to other threads up to
+  moment when fence is completed through dma_fence_signal().
+
+- As a last resort, if the hardware provides no useful reservation mechanics,
+  all workloads must be flushed from the GPU when switching between jobs
+  requiring DMA fences or jobs requiring page fault handling: This means all DMA
+  fences must complete before a compute job with page fault handling can be
+  inserted into the scheduler queue. And vice versa, before a DMA fence can be
+  made visible anywhere in the system, all compute workloads must be preempted
+  to guarantee all pending GPU page faults are flushed.
+
+- Only a fairly theoretical option would be to untangle these dependencies when
+  allocating memory to repair hardware page faults, either through separate
+  memory blocks or runtime tracking of the full dependency graph of all DMA
+  fences. This results very wide impact on the kernel, since resolving the page
+  on the CPU side can itself involve a page fault. It is much more feasible and
+  robust to limit the impact of handling hardware page faults to the specific
+  driver.
+
+Note that workloads that run on independent hardware like copy engines or other
+GPUs do not have any impact. This allows us to keep using DMA fences internally
+in the kernel even for resolving hardware page faults, e.g. by using copy
+engines to clear or copy memory needed to resolve the page fault.
+
+In some ways this page fault problem is a special case of the `Infinite DMA
+Fences` discussions: Infinite fences from compute workloads are allowed to
+depend on DMA fences, but not the other way around. And not even the page fault
+problem is new, because some other CPU thread in userspace might
+hit a page fault which holds up a userspace fence - supporting page faults on
+GPUs doesn't anything fundamentally new.

Documentation/gpu/drm-kms-helpers.rst

@@ -80,6 +80,18 @@ Atomic State Helper Reference
 .. kernel-doc:: drivers/gpu/drm/drm_atomic_state_helper.c
    :export:
 
+GEM Atomic Helper Reference
+---------------------------
+
+.. kernel-doc:: drivers/gpu/drm/drm_gem_atomic_helper.c
+   :doc: overview
+
+.. kernel-doc:: include/drm/drm_gem_atomic_helper.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_gem_atomic_helper.c
+   :export:
+
 Simple KMS Helper Reference
 ===========================
 

Documentation/gpu/index.rst

@@ -16,6 +16,7 @@ Linux GPU Driver Developer's Guide
    vga-switcheroo
    vgaarbiter
    todo
+   rfc/index
 
 .. only::  subproject and html
 

Documentation/gpu/rfc/index.rst

@@ -0,0 +1,17 @@
+===============
+GPU RFC Section
+===============
+
+For complex work, especially new uapi, it is often good to nail the high level
+design issues before getting lost in the code details. This section is meant to
+host such documentation:
+
+* Each RFC should be a section in this file, explaining the goal and main design
+  considerations. Especially for uapi make sure you Cc: all relevant project
+  mailing lists and involved people outside of dri-devel.
+
+* For uapi structures add a file to this directory with and then pull the
+  kerneldoc in like with real uapi headers.
+
+* Once the code has landed move all the documentation to the right places in
+  the main core, helper or driver sections.

Documentation/gpu/todo.rst

@@ -459,52 +459,6 @@ Contact: Emil Velikov, respective driver maintainers
 
 Level: Intermediate
 
-Plumb drm_atomic_state all over
--------------------------------
-
-Currently various atomic functions take just a single or a handful of
-object states (eg. plane state). While that single object state can
-suffice for some simple cases, we often have to dig out additional
-object states for dealing with various dependencies between the individual
-objects or the hardware they represent. The process of digging out the
-additional states is rather non-intuitive and error prone.
-
-To fix that most functions should rather take the overall
-drm_atomic_state as one of their parameters. The other parameters
-would generally be the object(s) we mainly want to interact with.
-
-For example, instead of
-
-.. code-block:: c
-
-   int (*atomic_check)(struct drm_plane *plane, struct drm_plane_state *state);
-
-we would have something like
-
-.. code-block:: c
-
-   int (*atomic_check)(struct drm_plane *plane, struct drm_atomic_state *state);
-
-The implementation can then trivially gain access to any required object
-state(s) via drm_atomic_get_plane_state(), drm_atomic_get_new_plane_state(),
-drm_atomic_get_old_plane_state(), and their equivalents for
-other object types.
-
-Additionally many drivers currently access the object->state pointer
-directly in their commit functions. That is not going to work if we
-eg. want to allow deeper commit pipelines as those pointers could
-then point to the states corresponding to a future commit instead of
-the current commit we're trying to process. Also non-blocking commits
-execute locklessly so there are serious concerns with dereferencing
-the object->state pointers without holding the locks that protect them.
-Use of drm_atomic_get_new_plane_state(), drm_atomic_get_old_plane_state(),
-etc. avoids these problems as well since they relate to a specific
-commit via the passed in drm_atomic_state.
-
-Contact: Ville Syrjälä, Daniel Vetter
-
-Level: Intermediate
-
 Use struct dma_buf_map throughout codebase
 ------------------------------------------
 
@@ -596,20 +550,24 @@ Contact: Daniel Vetter
 
 Level: Intermediate
 
-KMS cleanups
+Object lifetime fixes
-------------
+---------------------
 
-Some of these date from the very introduction of KMS in 2008 ...
+There's two related issues here
 
-- Make ->funcs and ->helper_private vtables optional. There's a bunch of empty
+- Cleanup up the various ->destroy callbacks, which often are all the same
-  function tables in drivers, but before we can remove them we need to make sure
+  simple code.
-  that all the users in helpers and drivers do correctly check for a NULL
-  vtable.
 
-- Cleanup up the various ->destroy callbacks. A lot of them just wrapt the
+- Lots of drivers erroneously allocate DRM modeset objects using devm_kzalloc,
-  drm_*_cleanup implementations and can be removed. Some tack a kfree() at the
+  which results in use-after free issues on driver unload. This can be serious
-  end, for which we could add drm_*_cleanup_kfree(). And then there's the (for
+  trouble even for drivers for hardware integrated on the SoC due to
-  historical reasons) misnamed drm_primary_helper_destroy() function.
+  EPROBE_DEFERRED backoff.
+
+Both these problems can be solved by switching over to drmm_kzalloc(), and the
+various convenience wrappers provided, e.g. drmm_crtc_alloc_with_planes(),
+drmm_universal_plane_alloc(), ... and so on.
+
+Contact: Daniel Vetter
 
 Level: Intermediate
 
@@ -666,8 +624,6 @@ See the documentation of :ref:`VKMS <vkms>` for more details. This is an ideal
 internship task, since it only requires a virtual machine and can be sized to
 fit the available time.
 
-Contact: Daniel Vetter
-
 Level: See details
 
 Backlight Refactoring
@@ -721,7 +677,7 @@ Outside DRM
 Convert fbdev drivers to DRM
 ----------------------------
 
-There are plenty of fbdev drivers for older hardware. Some hwardware has
+There are plenty of fbdev drivers for older hardware. Some hardware has
 become obsolete, but some still provides good(-enough) framebuffers. The
 drivers that are still useful should be converted to DRM and afterwards
 removed from fbdev.

MAINTAINERS

@@ -1326,7 +1326,7 @@ ARC PGU DRM DRIVER
 M:	Alexey Brodkin <abrodkin@synopsys.com>
 S:	Supported
 F:	Documentation/devicetree/bindings/display/snps,arcpgu.txt
-F:	drivers/gpu/drm/arc/
+F:	drivers/gpu/drm/tiny/arcpgu.c
 
 ARCNET NETWORK LAYER
 M:	Michael Grzeschik <m.grzeschik@pengutronix.de>
@@ -5641,6 +5641,12 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/display/panel/boe,himax8279d.yaml
 F:	drivers/gpu/drm/panel/panel-boe-himax8279d.c
 
+DRM DRIVER FOR CHIPONE ICN6211 MIPI-DSI to RGB CONVERTER BRIDGE
+M:	Jagan Teki <jagan@amarulasolutions.com>
+S:	Maintained
+F:	Documentation/devicetree/bindings/display/bridge/chipone,icn6211.yaml
+F:	drivers/gpu/drm/bridge/chipone-icn6211.c
+
 DRM DRIVER FOR FARADAY TVE200 TV ENCODER
 M:	Linus Walleij <linus.walleij@linaro.org>
 S:	Maintained
@@ -5659,6 +5665,14 @@ S:	Maintained
 F:	Documentation/devicetree/bindings/display/panel/feiyang,fy07024di26a30d.yaml
 F:	drivers/gpu/drm/panel/panel-feiyang-fy07024di26a30d.c
 
+DRM DRIVER FOR GENERIC USB DISPLAY
+M:	Noralf Trønnes <noralf@tronnes.org>
+S:	Maintained
+W:	https://github.com/notro/gud/wiki
+T:	git git://anongit.freedesktop.org/drm/drm-misc
+F:	drivers/gpu/drm/gud/
+F:	include/drm/gud.h
+
 DRM DRIVER FOR GRAIN MEDIA GM12U320 PROJECTORS
 M:	Hans de Goede <hdegoede@redhat.com>
 S:	Maintained
@@ -5967,6 +5981,7 @@ F:	drivers/gpu/drm/atmel-hlcdc/
 DRM DRIVERS FOR BRIDGE CHIPS
 M:	Andrzej Hajda <a.hajda@samsung.com>
 M:	Neil Armstrong <narmstrong@baylibre.com>
+M:	Robert Foss <robert.foss@linaro.org>
 R:	Laurent Pinchart <Laurent.pinchart@ideasonboard.com>
 R:	Jonas Karlman <jonas@kwiboo.se>
 R:	Jernej Skrabec <jernej.skrabec@siol.net>
@@ -6036,6 +6051,7 @@ DRM DRIVERS FOR MEDIATEK
 M:	Chun-Kuang Hu <chunkuang.hu@kernel.org>
 M:	Philipp Zabel <p.zabel@pengutronix.de>
 L:	dri-devel@lists.freedesktop.org
+L:	linux-mediatek@lists.infradead.org (moderated for non-subscribers)
 S:	Supported
 F:	Documentation/devicetree/bindings/display/mediatek/
 F:	drivers/gpu/drm/mediatek/
@@ -6061,9 +6077,9 @@ L:	dri-devel@lists.freedesktop.org
 L:	linux-renesas-soc@vger.kernel.org
 S:	Supported
 T:	git git://linuxtv.org/pinchartl/media drm/du/next
-F:	Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.txt
+F:	Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.yaml
 F:	Documentation/devicetree/bindings/display/bridge/renesas,lvds.yaml
-F:	Documentation/devicetree/bindings/display/renesas,du.txt
+F:	Documentation/devicetree/bindings/display/renesas,du.yaml
 F:	drivers/gpu/drm/rcar-du/
 F:	drivers/gpu/drm/shmobile/
 F:	include/linux/platform_data/shmob_drm.h
@@ -10572,6 +10588,12 @@ S:	Maintained
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid.git
 F:	drivers/hid/hid-lg-g15.c
 
+LONTIUM LT8912B MIPI TO HDMI BRIDGE
+M:	Adrien Grassein <adrien.grassein@gmail.com>
+S:	Maintained
+F:	Documentation/devicetree/bindings/display/bridge/lontium,lt8912b.yaml
+F:	drivers/gpu/drm/bridge/lontium-lt8912b.c
+
 LSILOGIC MPT FUSION DRIVERS (FC/SAS/SPI)
 M:	Sathya Prakash <sathya.prakash@broadcom.com>
 M:	Sreekanth Reddy <sreekanth.reddy@broadcom.com>
@@ -12394,7 +12416,7 @@ M:	Stefan Agner <stefan@agner.ch>
 L:	dri-devel@lists.freedesktop.org
 S:	Supported
 T:	git git://anongit.freedesktop.org/drm/drm-misc
-F:	Documentation/devicetree/bindings/display/mxsfb.txt
+F:	Documentation/devicetree/bindings/display/fsl,lcdif.yaml
 F:	drivers/gpu/drm/mxsfb/
 
 MYLEX DAC960 PCI RAID Controller

arch/x86/kernel/early-quirks.c

@@ -551,6 +551,7 @@ static const struct pci_device_id intel_early_ids[] __initconst = {
 	INTEL_EHL_IDS(&gen11_early_ops),
 	INTEL_TGL_12_IDS(&gen11_early_ops),
 	INTEL_RKL_IDS(&gen11_early_ops),
+	INTEL_ADLS_IDS(&gen11_early_ops),
 };
 
 struct resource intel_graphics_stolen_res __ro_after_init = DEFINE_RES_MEM(0, 0);

drivers/clk/clk-mux.c

@@ -8,6 +8,7 @@
  */
 
 #include <linux/clk-provider.h>
+#include <linux/device.h>
 #include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/io.h>
@@ -206,6 +207,40 @@ struct clk_hw *__clk_hw_register_mux(struct device *dev, struct device_node *np,
 }
 EXPORT_SYMBOL_GPL(__clk_hw_register_mux);
 
+static void devm_clk_hw_release_mux(struct device *dev, void *res)
+{
+	clk_hw_unregister_mux(*(struct clk_hw **)res);
+}
+
+struct clk_hw *__devm_clk_hw_register_mux(struct device *dev, struct device_node *np,
+		const char *name, u8 num_parents,
+		const char * const *parent_names,
+		const struct clk_hw **parent_hws,
+		const struct clk_parent_data *parent_data,
+		unsigned long flags, void __iomem *reg, u8 shift, u32 mask,
+		u8 clk_mux_flags, u32 *table, spinlock_t *lock)
+{
+	struct clk_hw **ptr, *hw;
+
+	ptr = devres_alloc(devm_clk_hw_release_mux, sizeof(*ptr), GFP_KERNEL);
+	if (!ptr)
+		return ERR_PTR(-ENOMEM);
+
+	hw = __clk_hw_register_mux(dev, np, name, num_parents, parent_names, parent_hws,
+				       parent_data, flags, reg, shift, mask,
+				       clk_mux_flags, table, lock);
+
+	if (!IS_ERR(hw)) {
+		*ptr = hw;
+		devres_add(dev, ptr);
+	} else {
+		devres_free(ptr);
+	}
+
+	return hw;
+}
+EXPORT_SYMBOL_GPL(__devm_clk_hw_register_mux);
+
 struct clk *clk_register_mux_table(struct device *dev, const char *name,
 		const char * const *parent_names, u8 num_parents,
 		unsigned long flags, void __iomem *reg, u8 shift, u32 mask,

drivers/dma-buf/dma-fence.c

@@ -123,7 +123,9 @@ static const struct dma_fence_ops dma_fence_stub_ops = {
 /**
  * dma_fence_get_stub - return a signaled fence
  *
- * Return a stub fence which is already signaled.
+ * Return a stub fence which is already signaled. The fence's
+ * timestamp corresponds to the first time after boot this
+ * function is called.
  */
 struct dma_fence *dma_fence_get_stub(void)
 {
@@ -141,6 +143,29 @@ struct dma_fence *dma_fence_get_stub(void)
 }
 EXPORT_SYMBOL(dma_fence_get_stub);
 
+/**
+ * dma_fence_allocate_private_stub - return a private, signaled fence
+ *
+ * Return a newly allocated and signaled stub fence.
+ */
+struct dma_fence *dma_fence_allocate_private_stub(void)
+{
+	struct dma_fence *fence;
+
+	fence = kzalloc(sizeof(*fence), GFP_KERNEL);
+	if (fence == NULL)
+		return ERR_PTR(-ENOMEM);
+
+	dma_fence_init(fence,
+		       &dma_fence_stub_ops,
+		       &dma_fence_stub_lock,
+		       0, 0);
+	dma_fence_signal(fence);
+
+	return fence;
+}
+EXPORT_SYMBOL(dma_fence_allocate_private_stub);
+
 /**
  * dma_fence_context_alloc - allocate an array of fence contexts
  * @num: amount of contexts to allocate

drivers/dma-buf/dma-heap.c

@@ -202,6 +202,18 @@ void *dma_heap_get_drvdata(struct dma_heap *heap)
 	return heap->priv;
 }
 
+/**
+ * dma_heap_get_name() - get heap name
+ * @heap: DMA-Heap to retrieve private data for
+ *
+ * Returns:
+ * The char* for the heap name.
+ */
+const char *dma_heap_get_name(struct dma_heap *heap)
+{
+	return heap->name;
+}
+
 struct dma_heap *dma_heap_add(const struct dma_heap_export_info *exp_info)
 {
 	struct dma_heap *heap, *h, *err_ret;

drivers/dma-buf/heaps/cma_heap.c

@@ -339,6 +339,7 @@ static struct dma_buf *cma_heap_allocate(struct dma_heap *heap,
 	buffer->pagecount = pagecount;
 
 	/* create the dmabuf */
+	exp_info.exp_name = dma_heap_get_name(heap);
 	exp_info.ops = &cma_heap_buf_ops;
 	exp_info.size = buffer->len;
 	exp_info.flags = fd_flags;

drivers/dma-buf/heaps/system_heap.c

@@ -390,6 +390,7 @@ static struct dma_buf *system_heap_allocate(struct dma_heap *heap,
 	}
 
 	/* create the dmabuf */
+	exp_info.exp_name = dma_heap_get_name(heap);
 	exp_info.ops = &system_heap_buf_ops;
 	exp_info.size = buffer->len;
 	exp_info.flags = fd_flags;

drivers/gpu/drm/Kconfig

@@ -352,8 +352,6 @@ source "drivers/gpu/drm/vc4/Kconfig"
 
 source "drivers/gpu/drm/etnaviv/Kconfig"
 
-source "drivers/gpu/drm/arc/Kconfig"
-
 source "drivers/gpu/drm/hisilicon/Kconfig"
 
 source "drivers/gpu/drm/mediatek/Kconfig"
@@ -386,6 +384,8 @@ source "drivers/gpu/drm/tidss/Kconfig"
 
 source "drivers/gpu/drm/xlnx/Kconfig"
 
+source "drivers/gpu/drm/gud/Kconfig"
+
 # Keep legacy drivers last
 
 menuconfig DRM_LEGACY

drivers/gpu/drm/Makefile

@@ -7,7 +7,7 @@ drm-y       :=	drm_auth.o drm_cache.o \
 		drm_file.o drm_gem.o drm_ioctl.o drm_irq.o \
 		drm_drv.o \
 		drm_sysfs.o drm_hashtab.o drm_mm.o \
-		drm_crtc.o drm_fourcc.o drm_modes.o drm_edid.o \
+		drm_crtc.o drm_fourcc.o drm_modes.o drm_edid.o drm_displayid.o \
 		drm_encoder_slave.o \
 		drm_trace_points.o drm_prime.o \
 		drm_rect.o drm_vma_manager.o drm_flip_work.o \
@@ -44,7 +44,8 @@ drm_kms_helper-y := drm_bridge_connector.o drm_crtc_helper.o drm_dp_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_modeset_helper.o \
-		drm_scdc_helper.o drm_gem_framebuffer_helper.o \
+		drm_scdc_helper.o drm_gem_atomic_helper.o \
+		drm_gem_framebuffer_helper.o \
 		drm_atomic_state_helper.o drm_damage_helper.o \
 		drm_format_helper.o drm_self_refresh_helper.o
 
@@ -110,7 +111,6 @@ obj-y			+= panel/
 obj-y			+= bridge/
 obj-$(CONFIG_DRM_FSL_DCU) += fsl-dcu/
 obj-$(CONFIG_DRM_ETNAVIV) += etnaviv/
-obj-$(CONFIG_DRM_ARCPGU)+= arc/
 obj-y			+= hisilicon/
 obj-$(CONFIG_DRM_ZTE)	+= zte/
 obj-$(CONFIG_DRM_MXSFB)	+= mxsfb/
@@ -125,3 +125,4 @@ obj-$(CONFIG_DRM_ASPEED_GFX) += aspeed/
 obj-$(CONFIG_DRM_MCDE) += mcde/
 obj-$(CONFIG_DRM_TIDSS) += tidss/
 obj-y			+= xlnx/
+obj-y			+= gud/

drivers/gpu/drm/amd/amdgpu/Kconfig

@@ -34,15 +34,6 @@ config DRM_AMDGPU_USERPTR
 	  This option selects CONFIG_HMM and CONFIG_HMM_MIRROR if it
 	  isn't already selected to enabled full userptr support.
 
-config DRM_AMDGPU_GART_DEBUGFS
-	bool "Allow GART access through debugfs"
-	depends on DRM_AMDGPU
-	depends on DEBUG_FS
-	default n
-	help
-	  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"
 source "drivers/gpu/drm/amd/display/Kconfig"
 source "drivers/gpu/drm/amd/amdkfd/Kconfig"

drivers/gpu/drm/amd/amdgpu/Makefile

@@ -71,7 +71,7 @@ amdgpu-y += \
 	vi.o mxgpu_vi.o nbio_v6_1.o soc15.o emu_soc.o mxgpu_ai.o nbio_v7_0.o vega10_reg_init.o \
 	vega20_reg_init.o nbio_v7_4.o nbio_v2_3.o nv.o navi10_reg_init.o navi14_reg_init.o \
 	arct_reg_init.o navi12_reg_init.o mxgpu_nv.o sienna_cichlid_reg_init.o vangogh_reg_init.o \
-	nbio_v7_2.o dimgrey_cavefish_reg_init.o hdp_v4_0.o hdp_v5_0.o
+	nbio_v7_2.o dimgrey_cavefish_reg_init.o hdp_v4_0.o hdp_v5_0.o aldebaran_reg_init.o aldebaran.o
 
 # add DF block
 amdgpu-y += \
@@ -83,11 +83,12 @@ amdgpu-y += \
 	gmc_v7_0.o \
 	gmc_v8_0.o \
 	gfxhub_v1_0.o mmhub_v1_0.o gmc_v9_0.o gfxhub_v1_1.o mmhub_v9_4.o \
-	gfxhub_v2_0.o mmhub_v2_0.o gmc_v10_0.o gfxhub_v2_1.o mmhub_v2_3.o
+	gfxhub_v2_0.o mmhub_v2_0.o gmc_v10_0.o gfxhub_v2_1.o mmhub_v2_3.o \
+	mmhub_v1_7.o
 
 # add UMC block
 amdgpu-y += \
-	umc_v6_1.o umc_v6_0.o umc_v8_7.o
+	umc_v6_0.o umc_v6_1.o umc_v6_7.o umc_v8_7.o
 
 # add IH block
 amdgpu-y += \
@@ -106,7 +107,8 @@ amdgpu-y += \
 	psp_v3_1.o \
 	psp_v10_0.o \
 	psp_v11_0.o \
-	psp_v12_0.o
+	psp_v12_0.o \
+	psp_v13_0.o
 
 # add DCE block
 amdgpu-y += \
@@ -121,6 +123,7 @@ amdgpu-y += \
 	gfx_v8_0.o \
 	gfx_v9_0.o \
 	gfx_v9_4.o \
+	gfx_v9_4_2.o \
 	gfx_v10_0.o
 
 # add async DMA block
@@ -129,6 +132,7 @@ amdgpu-y += \
 	sdma_v2_4.o \
 	sdma_v3_0.o \
 	sdma_v4_0.o \
+	sdma_v4_4.o \
 	sdma_v5_0.o \
 	sdma_v5_2.o
 
@@ -172,11 +176,17 @@ amdgpu-y += \
 amdgpu-y += \
 	smuio_v9_0.o \
 	smuio_v11_0.o \
-	smuio_v11_0_6.o
+	smuio_v11_0_6.o \
+	smuio_v13_0.o
+
+# add reset block
+amdgpu-y += \
+	amdgpu_reset.o
 
 # add amdkfd interfaces
 amdgpu-y += amdgpu_amdkfd.o
 
+
 ifneq ($(CONFIG_HSA_AMD),)
 AMDKFD_PATH := ../amdkfd
 include $(FULL_AMD_PATH)/amdkfd/Makefile
@@ -187,6 +197,7 @@ amdgpu-y += \
 	amdgpu_amdkfd_gfx_v8.o \
 	amdgpu_amdkfd_gfx_v9.o \
 	amdgpu_amdkfd_arcturus.o \
+	amdgpu_amdkfd_aldebaran.o \
 	amdgpu_amdkfd_gfx_v10.o \
 	amdgpu_amdkfd_gfx_v10_3.o
 

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

@@ -0,0 +1,407 @@
+/*
+ * Copyright 2021 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 "aldebaran.h"
+#include "amdgpu_reset.h"
+#include "amdgpu_amdkfd.h"
+#include "amdgpu_dpm.h"
+#include "amdgpu_job.h"
+#include "amdgpu_ring.h"
+#include "amdgpu_ras.h"
+#include "amdgpu_psp.h"
+#include "amdgpu_xgmi.h"
+
+static struct amdgpu_reset_handler *
+aldebaran_get_reset_handler(struct amdgpu_reset_control *reset_ctl,
+			    struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_reset_handler *handler;
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	if (reset_context->method != AMD_RESET_METHOD_NONE) {
+		dev_dbg(adev->dev, "Getting reset handler for method %d\n",
+			reset_context->method);
+		list_for_each_entry(handler, &reset_ctl->reset_handlers,
+				     handler_list) {
+			if (handler->reset_method == reset_context->method)
+				return handler;
+		}
+	}
+
+	if (adev->gmc.xgmi.connected_to_cpu) {
+		list_for_each_entry(handler, &reset_ctl->reset_handlers,
+				     handler_list) {
+			if (handler->reset_method == AMD_RESET_METHOD_MODE2) {
+				reset_context->method = AMD_RESET_METHOD_MODE2;
+				return handler;
+			}
+		}
+	}
+
+	dev_dbg(adev->dev, "Reset handler not found!\n");
+
+	return NULL;
+}
+
+static int aldebaran_mode2_suspend_ip(struct amdgpu_device *adev)
+{
+	int r, i;
+
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
+
+	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA))
+			continue;
+
+		r = adev->ip_blocks[i].version->funcs->suspend(adev);
+
+		if (r) {
+			dev_err(adev->dev,
+				"suspend of IP block <%s> failed %d\n",
+				adev->ip_blocks[i].version->funcs->name, r);
+			return r;
+		}
+
+		adev->ip_blocks[i].status.hw = false;
+	}
+
+	return r;
+}
+
+static int
+aldebaran_mode2_prepare_hwcontext(struct amdgpu_reset_control *reset_ctl,
+				  struct amdgpu_reset_context *reset_context)
+{
+	int r = 0;
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	dev_dbg(adev->dev, "Aldebaran prepare hw context\n");
+	/* Don't suspend on bare metal if we are not going to HW reset the ASIC */
+	if (!amdgpu_sriov_vf(adev))
+		r = aldebaran_mode2_suspend_ip(adev);
+
+	return r;
+}
+
+static void aldebaran_async_reset(struct work_struct *work)
+{
+	struct amdgpu_reset_handler *handler;
+	struct amdgpu_reset_control *reset_ctl =
+		container_of(work, struct amdgpu_reset_control, reset_work);
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+
+	list_for_each_entry(handler, &reset_ctl->reset_handlers,
+			     handler_list) {
+		if (handler->reset_method == reset_ctl->active_reset) {
+			dev_dbg(adev->dev, "Resetting device\n");
+			handler->do_reset(adev);
+			break;
+		}
+	}
+}
+
+static int aldebaran_mode2_reset(struct amdgpu_device *adev)
+{
+	/* disable BM */
+	pci_clear_master(adev->pdev);
+	adev->asic_reset_res = amdgpu_dpm_mode2_reset(adev);
+	return adev->asic_reset_res;
+}
+
+static int
+aldebaran_mode2_perform_reset(struct amdgpu_reset_control *reset_ctl,
+			      struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_device *tmp_adev = NULL;
+	struct amdgpu_device *adev = (struct amdgpu_device *)reset_ctl->handle;
+	int r = 0;
+
+	dev_dbg(adev->dev, "aldebaran perform hw reset\n");
+	if (reset_context->hive == NULL) {
+		/* Wrong context, return error */
+		return -EINVAL;
+	}
+
+	list_for_each_entry(tmp_adev, &reset_context->hive->device_list,
+			     gmc.xgmi.head) {
+		mutex_lock(&tmp_adev->reset_cntl->reset_lock);
+		tmp_adev->reset_cntl->active_reset = AMD_RESET_METHOD_MODE2;
+	}
+	/*
+	 * Mode2 reset doesn't need any sync between nodes in XGMI hive, instead launch
+	 * them together so that they can be completed asynchronously on multiple nodes
+	 */
+	list_for_each_entry(tmp_adev, &reset_context->hive->device_list,
+			     gmc.xgmi.head) {
+		/* For XGMI run all resets in parallel to speed up the process */
+		if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) {
+			if (!queue_work(system_unbound_wq,
+					&tmp_adev->reset_cntl->reset_work))
+				r = -EALREADY;
+		} else
+			r = aldebaran_mode2_reset(tmp_adev);
+		if (r) {
+			dev_err(tmp_adev->dev,
+				"ASIC reset failed with error, %d for drm dev, %s",
+				r, adev_to_drm(tmp_adev)->unique);
+			break;
+		}
+	}
+
+	/* For XGMI wait for all resets to complete before proceed */
+	if (!r) {
+		list_for_each_entry(tmp_adev,
+				     &reset_context->hive->device_list,
+				     gmc.xgmi.head) {
+			if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) {
+				flush_work(&tmp_adev->reset_cntl->reset_work);
+				r = tmp_adev->asic_reset_res;
+				if (r)
+					break;
+			}
+		}
+	}
+
+	list_for_each_entry(tmp_adev, &reset_context->hive->device_list,
+			     gmc.xgmi.head) {
+		mutex_unlock(&tmp_adev->reset_cntl->reset_lock);
+		tmp_adev->reset_cntl->active_reset = AMD_RESET_METHOD_NONE;
+	}
+
+	return r;
+}
+
+static int aldebaran_mode2_restore_ip(struct amdgpu_device *adev)
+{
+	struct amdgpu_firmware_info *ucode_list[AMDGPU_UCODE_ID_MAXIMUM];
+	struct amdgpu_firmware_info *ucode;
+	struct amdgpu_ip_block *cmn_block;
+	int ucode_count = 0;
+	int i, r;
+
+	dev_dbg(adev->dev, "Reloading ucodes after reset\n");
+	for (i = 0; i < adev->firmware.max_ucodes; i++) {
+		ucode = &adev->firmware.ucode[i];
+		if (!ucode->fw)
+			continue;
+		switch (ucode->ucode_id) {
+		case AMDGPU_UCODE_ID_SDMA0:
+		case AMDGPU_UCODE_ID_SDMA1:
+		case AMDGPU_UCODE_ID_SDMA2:
+		case AMDGPU_UCODE_ID_SDMA3:
+		case AMDGPU_UCODE_ID_SDMA4:
+		case AMDGPU_UCODE_ID_SDMA5:
+		case AMDGPU_UCODE_ID_SDMA6:
+		case AMDGPU_UCODE_ID_SDMA7:
+		case AMDGPU_UCODE_ID_CP_MEC1:
+		case AMDGPU_UCODE_ID_CP_MEC1_JT:
+		case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL:
+		case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM:
+		case AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM:
+		case AMDGPU_UCODE_ID_RLC_G:
+			ucode_list[ucode_count++] = ucode;
+			break;
+		default:
+			break;
+		};
+	}
+
+	/* Reinit NBIF block */
+	cmn_block =
+		amdgpu_device_ip_get_ip_block(adev, AMD_IP_BLOCK_TYPE_COMMON);
+	if (unlikely(!cmn_block)) {
+		dev_err(adev->dev, "Failed to get BIF handle\n");
+		return -EINVAL;
+	}
+	r = cmn_block->version->funcs->resume(adev);
+	if (r)
+		return r;
+
+	/* Reinit GFXHUB */
+	adev->gfxhub.funcs->init(adev);
+	r = adev->gfxhub.funcs->gart_enable(adev);
+	if (r) {
+		dev_err(adev->dev, "GFXHUB gart reenable failed after reset\n");
+		return r;
+	}
+
+	/* Reload GFX firmware */
+	r = psp_load_fw_list(&adev->psp, ucode_list, ucode_count);
+	if (r) {
+		dev_err(adev->dev, "GFX ucode load failed after reset\n");
+		return r;
+	}
+
+	/* Resume RLC, FW needs RLC alive to complete reset process */
+	adev->gfx.rlc.funcs->resume(adev);
+
+	/* Wait for FW reset event complete */
+	r = smu_wait_for_event(adev, SMU_EVENT_RESET_COMPLETE, 0);
+	if (r) {
+		dev_err(adev->dev,
+			"Failed to get response from firmware after reset\n");
+		return r;
+	}
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA))
+			continue;
+		r = adev->ip_blocks[i].version->funcs->resume(adev);
+		if (r) {
+			dev_err(adev->dev,
+				"resume of IP block <%s> failed %d\n",
+				adev->ip_blocks[i].version->funcs->name, r);
+			return r;
+		}
+
+		adev->ip_blocks[i].status.hw = true;
+	}
+
+	for (i = 0; i < adev->num_ip_blocks; i++) {
+		if (!(adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_GFX ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_SDMA ||
+		      adev->ip_blocks[i].version->type ==
+			      AMD_IP_BLOCK_TYPE_COMMON))
+			continue;
+
+		if (adev->ip_blocks[i].version->funcs->late_init) {
+			r = adev->ip_blocks[i].version->funcs->late_init(
+				(void *)adev);
+			if (r) {
+				dev_err(adev->dev,
+					"late_init of IP block <%s> failed %d after reset\n",
+					adev->ip_blocks[i].version->funcs->name,
+					r);
+				return r;
+			}
+		}
+		adev->ip_blocks[i].status.late_initialized = true;
+	}
+
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
+
+	return r;
+}
+
+static int
+aldebaran_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
+				  struct amdgpu_reset_context *reset_context)
+{
+	int r;
+	struct amdgpu_device *tmp_adev = NULL;
+
+	if (reset_context->hive == NULL) {
+		/* Wrong context, return error */
+		return -EINVAL;
+	}
+
+	list_for_each_entry(tmp_adev, &reset_context->hive->device_list,
+			     gmc.xgmi.head) {
+		dev_info(tmp_adev->dev,
+			 "GPU reset succeeded, trying to resume\n");
+		r = aldebaran_mode2_restore_ip(tmp_adev);
+		if (r)
+			goto end;
+
+		/*
+		 * Add this ASIC as tracked as reset was already
+		 * complete successfully.
+		 */
+		amdgpu_register_gpu_instance(tmp_adev);
+
+		/* Resume RAS */
+		amdgpu_ras_resume(tmp_adev);
+
+		/* Update PSP FW topology after reset */
+		if (reset_context->hive &&
+		    tmp_adev->gmc.xgmi.num_physical_nodes > 1)
+			r = amdgpu_xgmi_update_topology(reset_context->hive,
+							tmp_adev);
+
+		if (!r) {
+			amdgpu_irq_gpu_reset_resume_helper(tmp_adev);
+
+			r = amdgpu_ib_ring_tests(tmp_adev);
+			if (r) {
+				dev_err(tmp_adev->dev,
+					"ib ring test failed (%d).\n", r);
+				r = -EAGAIN;
+				tmp_adev->asic_reset_res = r;
+				goto end;
+			}
+		}
+	}
+
+end:
+	return r;
+}
+
+static struct amdgpu_reset_handler aldebaran_mode2_handler = {
+	.reset_method		= AMD_RESET_METHOD_MODE2,
+	.prepare_env		= NULL,
+	.prepare_hwcontext	= aldebaran_mode2_prepare_hwcontext,
+	.perform_reset		= aldebaran_mode2_perform_reset,
+	.restore_hwcontext	= aldebaran_mode2_restore_hwcontext,
+	.restore_env		= NULL,
+	.do_reset		= aldebaran_mode2_reset,
+};
+
+int aldebaran_reset_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_reset_control *reset_ctl;
+
+	reset_ctl = kzalloc(sizeof(*reset_ctl), GFP_KERNEL);
+	if (!reset_ctl)
+		return -ENOMEM;
+
+	reset_ctl->handle = adev;
+	reset_ctl->async_reset = aldebaran_async_reset;
+	reset_ctl->active_reset = AMD_RESET_METHOD_NONE;
+	reset_ctl->get_reset_handler = aldebaran_get_reset_handler;
+
+	INIT_LIST_HEAD(&reset_ctl->reset_handlers);
+	INIT_WORK(&reset_ctl->reset_work, reset_ctl->async_reset);
+	/* Only mode2 is handled through reset control now */
+	amdgpu_reset_add_handler(reset_ctl, &aldebaran_mode2_handler);
+
+	adev->reset_cntl = reset_ctl;
+
+	return 0;
+}
+
+int aldebaran_reset_fini(struct amdgpu_device *adev)
+{
+	kfree(adev->reset_cntl);
+	adev->reset_cntl = NULL;
+	return 0;
+}

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

@@ -0,0 +1,32 @@
+/*
+ * Copyright 2021 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 __ALDEBARAN_H__
+#define __ALDEBARAN_H__
+
+#include "amdgpu.h"
+
+int aldebaran_reset_init(struct amdgpu_device *adev);
+int aldebaran_reset_fini(struct amdgpu_device *adev);
+
+#endif

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

@@ -0,0 +1,54 @@
+/*
+ * Copyright 2020 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 "amdgpu.h"
+#include "soc15.h"
+
+#include "soc15_common.h"
+#include "aldebaran_ip_offset.h"
+
+int aldebaran_reg_base_init(struct amdgpu_device *adev)
+{
+	/* HW has more IP blocks,  only initialized the block needed by our driver  */
+	uint32_t i;
+	for (i = 0 ; i < MAX_INSTANCE ; ++i) {
+		adev->reg_offset[GC_HWIP][i] = (uint32_t *)(&(GC_BASE.instance[i]));
+		adev->reg_offset[HDP_HWIP][i] = (uint32_t *)(&(HDP_BASE.instance[i]));
+		adev->reg_offset[MMHUB_HWIP][i] = (uint32_t *)(&(MMHUB_BASE.instance[i]));
+		adev->reg_offset[ATHUB_HWIP][i] = (uint32_t *)(&(ATHUB_BASE.instance[i]));
+		adev->reg_offset[NBIO_HWIP][i] = (uint32_t *)(&(NBIO_BASE.instance[i]));
+		adev->reg_offset[MP0_HWIP][i] = (uint32_t *)(&(MP0_BASE.instance[i]));
+		adev->reg_offset[MP1_HWIP][i] = (uint32_t *)(&(MP1_BASE.instance[i]));
+		adev->reg_offset[DF_HWIP][i] = (uint32_t *)(&(DF_BASE.instance[i]));
+		adev->reg_offset[OSSSYS_HWIP][i] = (uint32_t *)(&(OSSSYS_BASE.instance[i]));
+		adev->reg_offset[SDMA0_HWIP][i] = (uint32_t *)(&(SDMA0_BASE.instance[i]));
+		adev->reg_offset[SDMA1_HWIP][i] = (uint32_t *)(&(SDMA1_BASE.instance[i]));
+		adev->reg_offset[SDMA2_HWIP][i] = (uint32_t *)(&(SDMA2_BASE.instance[i]));
+		adev->reg_offset[SDMA3_HWIP][i] = (uint32_t *)(&(SDMA3_BASE.instance[i]));
+		adev->reg_offset[SDMA4_HWIP][i] = (uint32_t *)(&(SDMA4_BASE.instance[i]));
+		adev->reg_offset[SMUIO_HWIP][i] = (uint32_t *)(&(SMUIO_BASE.instance[i]));
+		adev->reg_offset[THM_HWIP][i] = (uint32_t *)(&(THM_BASE.instance[i]));
+		adev->reg_offset[UMC_HWIP][i] = (uint32_t *)(&(UMC_BASE.instance[i]));
+		adev->reg_offset[VCN_HWIP][i] = (uint32_t *)(&(VCN_BASE.instance[i]));
+	}
+	return 0;
+}

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

@@ -107,7 +107,6 @@
 #include "amdgpu_gfxhub.h"
 #include "amdgpu_df.h"
 #include "amdgpu_smuio.h"
-#include "amdgpu_hdp.h"
 
 #define MAX_GPU_INSTANCE		16
 
@@ -124,6 +123,16 @@ struct amdgpu_mgpu_info
 	uint32_t			num_gpu;
 	uint32_t			num_dgpu;
 	uint32_t			num_apu;
+
+	/* delayed reset_func for XGMI configuration if necessary */
+	struct delayed_work		delayed_reset_work;
+	bool				pending_reset;
+};
+
+struct amdgpu_watchdog_timer
+{
+	bool timeout_fatal_disable;
+	uint32_t period; /* maxCycles = (1 << period), the number of cycles before a timeout */
 };
 
 #define AMDGPU_MAX_TIMEOUT_PARAM_LENGTH	256
@@ -177,7 +186,9 @@ extern int amdgpu_compute_multipipe;
 extern int amdgpu_gpu_recovery;
 extern int amdgpu_emu_mode;
 extern uint amdgpu_smu_memory_pool_size;
+extern int amdgpu_smu_pptable_id;
 extern uint amdgpu_dc_feature_mask;
+extern uint amdgpu_freesync_vid_mode;
 extern uint amdgpu_dc_debug_mask;
 extern uint amdgpu_dm_abm_level;
 extern int amdgpu_backlight;
@@ -185,6 +196,7 @@ extern struct amdgpu_mgpu_info mgpu_info;
 extern int amdgpu_ras_enable;
 extern uint amdgpu_ras_mask;
 extern int amdgpu_bad_page_threshold;
+extern struct amdgpu_watchdog_timer amdgpu_watchdog_timer;
 extern int amdgpu_async_gfx_ring;
 extern int amdgpu_mcbp;
 extern int amdgpu_discovery;
@@ -258,6 +270,8 @@ struct amdgpu_bo_va_mapping;
 struct amdgpu_atif;
 struct kfd_vm_fault_info;
 struct amdgpu_hive_info;
+struct amdgpu_reset_context;
+struct amdgpu_reset_control;
 
 enum amdgpu_cp_irq {
 	AMDGPU_CP_IRQ_GFX_ME0_PIPE0_EOP = 0,
@@ -576,6 +590,7 @@ struct amdgpu_allowed_register_entry {
 };
 
 enum amd_reset_method {
+	AMD_RESET_METHOD_NONE = -1,
 	AMD_RESET_METHOD_LEGACY = 0,
 	AMD_RESET_METHOD_MODE0,
 	AMD_RESET_METHOD_MODE1,
@@ -584,6 +599,19 @@ enum amd_reset_method {
 	AMD_RESET_METHOD_PCI,
 };
 
+struct amdgpu_video_codec_info {
+	u32 codec_type;
+	u32 max_width;
+	u32 max_height;
+	u32 max_pixels_per_frame;
+	u32 max_level;
+};
+
+struct amdgpu_video_codecs {
+	const u32 codec_count;
+	const struct amdgpu_video_codec_info *codec_array;
+};
+
 /*
  * ASIC specific functions.
  */
@@ -628,6 +656,9 @@ struct amdgpu_asic_funcs {
 	void (*pre_asic_init)(struct amdgpu_device *adev);
 	/* enter/exit umd stable pstate */
 	int (*update_umd_stable_pstate)(struct amdgpu_device *adev, bool enter);
+	/* query video codecs */
+	int (*query_video_codecs)(struct amdgpu_device *adev, bool encode,
+				  const struct amdgpu_video_codecs **codecs);
 };
 
 /*
@@ -792,12 +823,7 @@ struct amdgpu_device {
 	bool				accel_working;
 	struct notifier_block		acpi_nb;
 	struct amdgpu_i2c_chan		*i2c_bus[AMDGPU_MAX_I2C_BUS];
-	struct amdgpu_debugfs		debugfs[AMDGPU_DEBUGFS_MAX_COMPONENTS];
+	struct debugfs_blob_wrapper     debugfs_vbios_blob;
-	unsigned			debugfs_count;
-#if defined(CONFIG_DEBUG_FS)
-	struct dentry                   *debugfs_preempt;
-	struct dentry			*debugfs_regs[AMDGPU_DEBUGFS_MAX_COMPONENTS];
-#endif
 	struct amdgpu_atif		*atif;
 	struct amdgpu_atcs		atcs;
 	struct mutex			srbm_mutex;
@@ -853,8 +879,6 @@ struct amdgpu_device {
 	spinlock_t audio_endpt_idx_lock;
 	amdgpu_block_rreg_t		audio_endpt_rreg;
 	amdgpu_block_wreg_t		audio_endpt_wreg;
-	void __iomem                    *rio_mem;
-	resource_size_t			rio_mem_size;
 	struct amdgpu_doorbell		doorbell;
 
 	/* clock/pll info */
@@ -897,6 +921,8 @@ struct amdgpu_device {
 	struct amdgpu_irq_src		vupdate_irq;
 	struct amdgpu_irq_src		pageflip_irq;
 	struct amdgpu_irq_src		hpd_irq;
+	struct amdgpu_irq_src		dmub_trace_irq;
+	struct amdgpu_irq_src		dmub_outbox_irq;
 
 	/* rings */
 	u64				fence_context;
@@ -1020,6 +1046,7 @@ struct amdgpu_device {
 
 	int asic_reset_res;
 	struct work_struct		xgmi_reset_work;
+	struct list_head		reset_list;
 
 	long				gfx_timeout;
 	long				sdma_timeout;
@@ -1050,6 +1077,8 @@ struct amdgpu_device {
 
 	bool                            in_pci_err_recovery;
 	struct pci_saved_state          *pci_state;
+
+	struct amdgpu_reset_control     *reset_cntl;
 };
 
 static inline struct amdgpu_device *drm_to_adev(struct drm_device *ddev)
@@ -1062,7 +1091,7 @@ static inline struct drm_device *adev_to_drm(struct amdgpu_device *adev)
 	return &adev->ddev;
 }
 
-static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_bo_device *bdev)
+static inline struct amdgpu_device *amdgpu_ttm_adev(struct ttm_device *bdev)
 {
 	return container_of(bdev, struct amdgpu_device, mman.bdev);
 }
@@ -1084,9 +1113,6 @@ void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev,
 void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value);
 uint8_t amdgpu_mm_rreg8(struct amdgpu_device *adev, uint32_t offset);
 
-u32 amdgpu_io_rreg(struct amdgpu_device *adev, u32 reg);
-void amdgpu_io_wreg(struct amdgpu_device *adev, u32 reg, u32 v);
-
 u32 amdgpu_device_indirect_rreg(struct amdgpu_device *adev,
 				u32 pcie_index, u32 pcie_data,
 				u32 reg_addr);
@@ -1103,6 +1129,12 @@ void amdgpu_device_indirect_wreg64(struct amdgpu_device *adev,
 bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type);
 bool amdgpu_device_has_dc_support(struct amdgpu_device *adev);
 
+int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
+				 struct amdgpu_reset_context *reset_context);
+
+int amdgpu_do_asic_reset(struct list_head *device_list_handle,
+			 struct amdgpu_reset_context *reset_context);
+
 int emu_soc_asic_init(struct amdgpu_device *adev);
 
 /*
@@ -1168,8 +1200,6 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 	} while (0)
 
 #define DREG32_SYS(sqf, adev, reg) seq_printf((sqf), #reg " : 0x%08X\n", amdgpu_device_rreg((adev), (reg), false))
-#define RREG32_IO(reg) amdgpu_io_rreg(adev, (reg))
-#define WREG32_IO(reg, v) amdgpu_io_wreg(adev, (reg), (v))
 
 #define REG_FIELD_SHIFT(reg, field) reg##__##field##__SHIFT
 #define REG_FIELD_MASK(reg, field) reg##__##field##_MASK
@@ -1223,6 +1253,7 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define amdgpu_asic_pre_asic_init(adev) (adev)->asic_funcs->pre_asic_init((adev))
 #define amdgpu_asic_update_umd_stable_pstate(adev, enter) \
 	((adev)->asic_funcs->update_umd_stable_pstate ? (adev)->asic_funcs->update_umd_stable_pstate((adev), (enter)) : 0)
+#define amdgpu_asic_query_video_codecs(adev, e, c) (adev)->asic_funcs->query_video_codecs((adev), (e), (c))
 
 #define amdgpu_inc_vram_lost(adev) atomic_inc(&((adev)->vram_lost_counter));
 
@@ -1242,7 +1273,9 @@ void amdgpu_device_program_register_sequence(struct amdgpu_device *adev,
 					     const u32 *registers,
 					     const u32 array_size);
 
+int amdgpu_device_mode1_reset(struct amdgpu_device *adev);
 bool amdgpu_device_supports_atpx(struct drm_device *dev);
+bool amdgpu_device_supports_px(struct drm_device *dev);
 bool amdgpu_device_supports_boco(struct drm_device *dev);
 bool amdgpu_device_supports_baco(struct drm_device *dev);
 bool amdgpu_device_is_peer_accessible(struct amdgpu_device *adev,
@@ -1356,6 +1389,13 @@ void amdgpu_pci_resume(struct pci_dev *pdev);
 bool amdgpu_device_cache_pci_state(struct pci_dev *pdev);
 bool amdgpu_device_load_pci_state(struct pci_dev *pdev);
 
+bool amdgpu_device_skip_hw_access(struct amdgpu_device *adev);
+
+int amdgpu_device_set_cg_state(struct amdgpu_device *adev,
+			       enum amd_clockgating_state state);
+int amdgpu_device_set_pg_state(struct amdgpu_device *adev,
+			       enum amd_powergating_state state);
+
 #include "amdgpu_object.h"
 
 static inline bool amdgpu_is_tmz(struct amdgpu_device *adev)

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

@@ -44,7 +44,7 @@ int amdgpu_amdkfd_init(void)
 	int ret;
 
 	si_meminfo(&si);
-	amdgpu_amdkfd_total_mem_size = si.totalram - si.totalhigh;
+	amdgpu_amdkfd_total_mem_size = si.freeram - si.freehigh;
 	amdgpu_amdkfd_total_mem_size *= si.mem_unit;
 
 	ret = kgd2kfd_init();
@@ -165,7 +165,8 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
 					adev->doorbell_index.last_non_cp;
 		}
 
-		kgd2kfd_device_init(adev->kfd.dev, adev_to_drm(adev), &gpu_resources);
+		adev->kfd.init_complete = kgd2kfd_device_init(adev->kfd.dev,
+						adev_to_drm(adev), &gpu_resources);
 	}
 }
 
@@ -245,6 +246,7 @@ int amdgpu_amdkfd_alloc_gtt_mem(struct kgd_dev *kgd, size_t size,
 	bp.flags = AMDGPU_GEM_CREATE_CPU_GTT_USWC;
 	bp.type = ttm_bo_type_kernel;
 	bp.resv = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
 
 	if (cp_mqd_gfx9)
 		bp.flags |= AMDGPU_GEM_CREATE_CP_MQD_GFX9;
@@ -316,6 +318,7 @@ int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size,
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 	struct amdgpu_bo *bo = NULL;
+	struct amdgpu_bo_user *ubo;
 	struct amdgpu_bo_param bp;
 	int r;
 
@@ -326,14 +329,16 @@ int amdgpu_amdkfd_alloc_gws(struct kgd_dev *kgd, size_t size,
 	bp.flags = AMDGPU_GEM_CREATE_NO_CPU_ACCESS;
 	bp.type = ttm_bo_type_device;
 	bp.resv = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
 
-	r = amdgpu_bo_create(adev, &bp, &bo);
+	r = amdgpu_bo_create_user(adev, &bp, &ubo);
 	if (r) {
 		dev_err(adev->dev,
 			"failed to allocate gws BO for amdkfd (%d)\n", r);
 		return r;
 	}
 
+	bo = &ubo->bo;
 	*mem_obj = bo;
 	return 0;
 }
@@ -494,8 +499,6 @@ int amdgpu_amdkfd_get_dmabuf_info(struct kgd_dev *kgd, int dma_buf_fd,
 		*dma_buf_kgd = (struct kgd_dev *)adev;
 	if (bo_size)
 		*bo_size = amdgpu_bo_size(bo);
-	if (metadata_size)
-		*metadata_size = bo->metadata_size;
 	if (metadata_buffer)
 		r = amdgpu_bo_get_metadata(bo, metadata_buffer, buffer_size,
 					   metadata_size, &metadata_flags);
@@ -638,13 +641,6 @@ void amdgpu_amdkfd_set_compute_idle(struct kgd_dev *kgd, bool idle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)kgd;
 
-	/* Temp workaround to fix the soft hang observed in certain compute
-	 * applications if GFXOFF is enabled.
-	 */
-	if (adev->asic_type == CHIP_SIENNA_CICHLID) {
-		pr_debug("GFXOFF is %s\n", idle ? "enabled" : "disabled");
-		amdgpu_gfx_off_ctrl(adev, idle);
-	}
 	amdgpu_dpm_switch_power_profile(adev,
 					PP_SMC_POWER_PROFILE_COMPUTE,
 					!idle);

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

@@ -80,6 +80,7 @@ struct amdgpu_amdkfd_fence {
 struct amdgpu_kfd_dev {
 	struct kfd_dev *dev;
 	uint64_t vram_used;
+	bool init_complete;
 };
 
 enum kgd_engine_type {

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

@@ -0,0 +1,47 @@
+/*
+ * Copyright 2020 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 "amdgpu.h"
+#include "amdgpu_amdkfd.h"
+#include "amdgpu_amdkfd_arcturus.h"
+#include "amdgpu_amdkfd_gfx_v9.h"
+
+const struct kfd2kgd_calls aldebaran_kfd2kgd = {
+	.program_sh_mem_settings = kgd_gfx_v9_program_sh_mem_settings,
+	.set_pasid_vmid_mapping = kgd_gfx_v9_set_pasid_vmid_mapping,
+	.init_interrupts = kgd_gfx_v9_init_interrupts,
+	.hqd_load = kgd_gfx_v9_hqd_load,
+	.hiq_mqd_load = kgd_gfx_v9_hiq_mqd_load,
+	.hqd_sdma_load = kgd_arcturus_hqd_sdma_load,
+	.hqd_dump = kgd_gfx_v9_hqd_dump,
+	.hqd_sdma_dump = kgd_arcturus_hqd_sdma_dump,
+	.hqd_is_occupied = kgd_gfx_v9_hqd_is_occupied,
+	.hqd_sdma_is_occupied = kgd_arcturus_hqd_sdma_is_occupied,
+	.hqd_destroy = kgd_gfx_v9_hqd_destroy,
+	.hqd_sdma_destroy = kgd_arcturus_hqd_sdma_destroy,
+	.address_watch_disable = kgd_gfx_v9_address_watch_disable,
+	.address_watch_execute = kgd_gfx_v9_address_watch_execute,
+	.wave_control_execute = kgd_gfx_v9_wave_control_execute,
+	.address_watch_get_offset = kgd_gfx_v9_address_watch_get_offset,
+	.get_atc_vmid_pasid_mapping_info =
+				kgd_gfx_v9_get_atc_vmid_pasid_mapping_info,
+	.set_vm_context_page_table_base = kgd_gfx_v9_set_vm_context_page_table_base,
+};

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

@@ -122,7 +122,7 @@ static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
 	return sdma_rlc_reg_offset;
 }
 
-static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 			     uint32_t __user *wptr, struct mm_struct *mm)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
@@ -192,7 +192,7 @@ static int kgd_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
 	return 0;
 }
 
-static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
+int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd,
 			     uint32_t engine_id, uint32_t queue_id,
 			     uint32_t (**dump)[2], uint32_t *n_regs)
 {
@@ -224,7 +224,7 @@ static int kgd_hqd_sdma_dump(struct kgd_dev *kgd,
 	return 0;
 }
 
-static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
+bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
 	struct v9_sdma_mqd *m;
@@ -243,7 +243,7 @@ static bool kgd_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd)
 	return false;
 }
 
-static int kgd_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
 				unsigned int utimeout)
 {
 	struct amdgpu_device *adev = get_amdgpu_device(kgd);
@@ -289,13 +289,13 @@ const struct kfd2kgd_calls arcturus_kfd2kgd = {
 	.init_interrupts = kgd_gfx_v9_init_interrupts,
 	.hqd_load = kgd_gfx_v9_hqd_load,
 	.hiq_mqd_load = kgd_gfx_v9_hiq_mqd_load,
-	.hqd_sdma_load = kgd_hqd_sdma_load,
+	.hqd_sdma_load = kgd_arcturus_hqd_sdma_load,
 	.hqd_dump = kgd_gfx_v9_hqd_dump,
-	.hqd_sdma_dump = kgd_hqd_sdma_dump,
+	.hqd_sdma_dump = kgd_arcturus_hqd_sdma_dump,
 	.hqd_is_occupied = kgd_gfx_v9_hqd_is_occupied,
-	.hqd_sdma_is_occupied = kgd_hqd_sdma_is_occupied,
+	.hqd_sdma_is_occupied = kgd_arcturus_hqd_sdma_is_occupied,
 	.hqd_destroy = kgd_gfx_v9_hqd_destroy,
-	.hqd_sdma_destroy = kgd_hqd_sdma_destroy,
+	.hqd_sdma_destroy = kgd_arcturus_hqd_sdma_destroy,
 	.address_watch_disable = kgd_gfx_v9_address_watch_disable,
 	.address_watch_execute = kgd_gfx_v9_address_watch_execute,
 	.wave_control_execute = kgd_gfx_v9_wave_control_execute,

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

@@ -0,0 +1,30 @@
+/*
+ * Copyright 2020 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.
+ */
+
+int kgd_arcturus_hqd_sdma_load(struct kgd_dev *kgd, void *mqd,
+			     uint32_t __user *wptr, struct mm_struct *mm);
+int kgd_arcturus_hqd_sdma_dump(struct kgd_dev *kgd,
+			     uint32_t engine_id, uint32_t queue_id,
+			     uint32_t (**dump)[2], uint32_t *n_regs);
+bool kgd_arcturus_hqd_sdma_is_occupied(struct kgd_dev *kgd, void *mqd);
+int kgd_arcturus_hqd_sdma_destroy(struct kgd_dev *kgd, void *mqd,
+				unsigned int utimeout);

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

@@ -40,13 +40,13 @@ static atomic_t fence_seq = ATOMIC_INIT(0);
  * All the BOs in a process share an eviction fence. When process X wants
  * to map VRAM memory but TTM can't find enough space, TTM will attempt to
  * evict BOs from its LRU list. TTM checks if the BO is valuable to evict
- * by calling ttm_bo_driver->eviction_valuable().
+ * by calling ttm_device_funcs->eviction_valuable().
  *
- * ttm_bo_driver->eviction_valuable() - will return false if the BO belongs
+ * ttm_device_funcs->eviction_valuable() - will return false if the BO belongs
  *  to process X. Otherwise, it will return true to indicate BO can be
  *  evicted by TTM.
  *
- * If ttm_bo_driver->eviction_valuable returns true, then TTM will continue
+ * If ttm_device_funcs->eviction_valuable returns true, then TTM will continue
  * the evcition process for that BO by calling ttm_bo_evict --> amdgpu_bo_move
  * --> amdgpu_copy_buffer(). This sets up job in GPU scheduler.
  *

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

@@ -31,6 +31,7 @@
 #include "amdgpu_amdkfd.h"
 #include "amdgpu_dma_buf.h"
 #include <uapi/linux/kfd_ioctl.h>
+#include "amdgpu_xgmi.h"
 
 /* BO flag to indicate a KFD userptr BO */
 #define AMDGPU_AMDKFD_USERPTR_BO (1ULL << 63)
@@ -96,7 +97,7 @@ void amdgpu_amdkfd_gpuvm_init_mem_limits(void)
 	uint64_t mem;
 
 	si_meminfo(&si);
-	mem = si.totalram - si.totalhigh;
+	mem = si.freeram - si.freehigh;
 	mem *= si.mem_unit;
 
 	spin_lock_init(&kfd_mem_limit.mem_limit_lock);
@@ -119,6 +120,16 @@ void amdgpu_amdkfd_gpuvm_init_mem_limits(void)
  */
 #define ESTIMATE_PT_SIZE(mem_size) ((mem_size) >> 14)
 
+static size_t amdgpu_amdkfd_acc_size(uint64_t size)
+{
+	size >>= PAGE_SHIFT;
+	size *= sizeof(dma_addr_t) + sizeof(void *);
+
+	return __roundup_pow_of_two(sizeof(struct amdgpu_bo)) +
+		__roundup_pow_of_two(sizeof(struct ttm_tt)) +
+		PAGE_ALIGN(size);
+}
+
 static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
 		uint64_t size, u32 domain, bool sg)
 {
@@ -127,8 +138,7 @@ static int amdgpu_amdkfd_reserve_mem_limit(struct amdgpu_device *adev,
 	size_t acc_size, system_mem_needed, ttm_mem_needed, vram_needed;
 	int ret = 0;
 
-	acc_size = ttm_bo_dma_acc_size(&adev->mman.bdev, size,
+	acc_size = amdgpu_amdkfd_acc_size(size);
-				       sizeof(struct amdgpu_bo));
 
 	vram_needed = 0;
 	if (domain == AMDGPU_GEM_DOMAIN_GTT) {
@@ -175,8 +185,7 @@ static void unreserve_mem_limit(struct amdgpu_device *adev,
 {
 	size_t acc_size;
 
-	acc_size = ttm_bo_dma_acc_size(&adev->mman.bdev, size,
+	acc_size = amdgpu_amdkfd_acc_size(size);
-				       sizeof(struct amdgpu_bo));
 
 	spin_lock(&kfd_mem_limit.mem_limit_lock);
 	if (domain == AMDGPU_GEM_DOMAIN_GTT) {
@@ -404,7 +413,10 @@ static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
 {
 	struct amdgpu_device *bo_adev = amdgpu_ttm_adev(mem->bo->tbo.bdev);
 	bool coherent = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_COHERENT;
+	bool uncached = mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_UNCACHED;
 	uint32_t mapping_flags;
+	uint64_t pte_flags;
+	bool snoop = false;
 
 	mapping_flags = AMDGPU_VM_PAGE_READABLE;
 	if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_WRITABLE)
@@ -425,12 +437,41 @@ static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
 				AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
 		}
 		break;
+	case CHIP_ALDEBARAN:
+		if (coherent && uncached) {
+			if (adev->gmc.xgmi.connected_to_cpu ||
+				!(mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM))
+				snoop = true;
+			mapping_flags |= AMDGPU_VM_MTYPE_UC;
+		} else if (mem->alloc_flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+			if (bo_adev == adev) {
+				mapping_flags |= AMDGPU_VM_MTYPE_RW;
+				if (adev->gmc.xgmi.connected_to_cpu)
+					snoop = true;
+			} else {
+				mapping_flags |= AMDGPU_VM_MTYPE_NC;
+				if (amdgpu_xgmi_same_hive(adev, bo_adev))
+					snoop = true;
+			}
+		} else {
+			snoop = true;
+			if (adev->gmc.xgmi.connected_to_cpu)
+				/* system memory uses NC on A+A */
+				mapping_flags |= AMDGPU_VM_MTYPE_NC;
+			else
+				mapping_flags |= coherent ?
+					AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
+		}
+		break;
 	default:
 		mapping_flags |= coherent ?
 			AMDGPU_VM_MTYPE_UC : AMDGPU_VM_MTYPE_NC;
 	}
 
-	return amdgpu_gem_va_map_flags(adev, mapping_flags);
+	pte_flags = amdgpu_gem_va_map_flags(adev, mapping_flags);
+	pte_flags |= snoop ? AMDGPU_PTE_SNOOPED : 0;
+
+	return pte_flags;
 }
 
 /* add_bo_to_vm - Add a BO to a VM

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

@@ -1232,157 +1232,6 @@ int amdgpu_atombios_get_leakage_vddc_based_on_leakage_idx(struct amdgpu_device *
 	return amdgpu_atombios_get_max_vddc(adev, VOLTAGE_TYPE_VDDC, leakage_idx, voltage);
 }
 
-int amdgpu_atombios_get_leakage_id_from_vbios(struct amdgpu_device *adev,
-					      u16 *leakage_id)
-{
-	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 3:
-	case 4:
-		args.v3.ucVoltageType = 0;
-		args.v3.ucVoltageMode = ATOM_GET_LEAKAGE_ID;
-		args.v3.usVoltageLevel = 0;
-
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
-
-		*leakage_id = 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_params(struct amdgpu_device *adev,
-							     u16 *vddc, u16 *vddci,
-							     u16 virtual_voltage_id,
-							     u16 vbios_voltage_id)
-{
-	int index = GetIndexIntoMasterTable(DATA, ASIC_ProfilingInfo);
-	u8 frev, crev;
-	u16 data_offset, size;
-	int i, j;
-	ATOM_ASIC_PROFILING_INFO_V2_1 *profile;
-	u16 *leakage_bin, *vddc_id_buf, *vddc_buf, *vddci_id_buf, *vddci_buf;
-
-	*vddc = 0;
-	*vddci = 0;
-
-	if (!amdgpu_atom_parse_data_header(adev->mode_info.atom_context, index, &size,
-				    &frev, &crev, &data_offset))
-		return -EINVAL;
-
-	profile = (ATOM_ASIC_PROFILING_INFO_V2_1 *)
-		(adev->mode_info.atom_context->bios + data_offset);
-
-	switch (frev) {
-	case 1:
-		return -EINVAL;
-	case 2:
-		switch (crev) {
-		case 1:
-			if (size < sizeof(ATOM_ASIC_PROFILING_INFO_V2_1))
-				return -EINVAL;
-			leakage_bin = (u16 *)
-				(adev->mode_info.atom_context->bios + data_offset +
-				 le16_to_cpu(profile->usLeakageBinArrayOffset));
-			vddc_id_buf = (u16 *)
-				(adev->mode_info.atom_context->bios + data_offset +
-				 le16_to_cpu(profile->usElbVDDC_IdArrayOffset));
-			vddc_buf = (u16 *)
-				(adev->mode_info.atom_context->bios + data_offset +
-				 le16_to_cpu(profile->usElbVDDC_LevelArrayOffset));
-			vddci_id_buf = (u16 *)
-				(adev->mode_info.atom_context->bios + data_offset +
-				 le16_to_cpu(profile->usElbVDDCI_IdArrayOffset));
-			vddci_buf = (u16 *)
-				(adev->mode_info.atom_context->bios + data_offset +
-				 le16_to_cpu(profile->usElbVDDCI_LevelArrayOffset));
-
-			if (profile->ucElbVDDC_Num > 0) {
-				for (i = 0; i < profile->ucElbVDDC_Num; i++) {
-					if (vddc_id_buf[i] == virtual_voltage_id) {
-						for (j = 0; j < profile->ucLeakageBinNum; j++) {
-							if (vbios_voltage_id <= leakage_bin[j]) {
-								*vddc = vddc_buf[j * profile->ucElbVDDC_Num + i];
-								break;
-							}
-						}
-						break;
-					}
-				}
-			}
-			if (profile->ucElbVDDCI_Num > 0) {
-				for (i = 0; i < profile->ucElbVDDCI_Num; i++) {
-					if (vddci_id_buf[i] == virtual_voltage_id) {
-						for (j = 0; j < profile->ucLeakageBinNum; j++) {
-							if (vbios_voltage_id <= leakage_bin[j]) {
-								*vddci = vddci_buf[j * profile->ucElbVDDCI_Num + i];
-								break;
-							}
-						}
-						break;
-					}
-				}
-			}
-			break;
-		default:
-			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-			return -EINVAL;
-		}
-		break;
-	default:
-		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-union get_voltage_info {
-	struct _GET_VOLTAGE_INFO_INPUT_PARAMETER_V1_2 in;
-	struct _GET_EVV_VOLTAGE_INFO_OUTPUT_PARAMETER_V1_2 evv_out;
-};
-
-int amdgpu_atombios_get_voltage_evv(struct amdgpu_device *adev,
-				    u16 virtual_voltage_id,
-				    u16 *voltage)
-{
-	int index = GetIndexIntoMasterTable(COMMAND, GetVoltageInfo);
-	u32 entry_id;
-	u32 count = adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.count;
-	union get_voltage_info args;
-
-	for (entry_id = 0; entry_id < count; entry_id++) {
-		if (adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[entry_id].v ==
-		    virtual_voltage_id)
-			break;
-	}
-
-	if (entry_id >= count)
-		return -EINVAL;
-
-	args.in.ucVoltageType = VOLTAGE_TYPE_VDDC;
-	args.in.ucVoltageMode = ATOM_GET_VOLTAGE_EVV_VOLTAGE;
-	args.in.usVoltageLevel = cpu_to_le16(virtual_voltage_id);
-	args.in.ulSCLKFreq =
-		cpu_to_le32(adev->pm.dpm.dyn_state.vddc_dependency_on_sclk.entries[entry_id].clk);
-
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
-
-	*voltage = le16_to_cpu(args.evv_out.usVoltageLevel);
-
-	return 0;
-}
-
 union voltage_object_info {
 	struct _ATOM_VOLTAGE_OBJECT_INFO v1;
 	struct _ATOM_VOLTAGE_OBJECT_INFO_V2 v2;
@@ -1905,40 +1754,6 @@ static uint32_t cail_reg_read(struct card_info *info, uint32_t reg)
 	return r;
 }
 
-/**
- * cail_ioreg_write - write IO register
- *
- * @info: atom card_info pointer
- * @reg: IO register offset
- * @val: value to write to the pll register
- *
- * Provides a IO register accessor for the atom interpreter (r4xx+).
- */
-static void cail_ioreg_write(struct card_info *info, uint32_t reg, uint32_t val)
-{
-	struct amdgpu_device *adev = drm_to_adev(info->dev);
-
-	WREG32_IO(reg, val);
-}
-
-/**
- * cail_ioreg_read - read IO register
- *
- * @info: atom card_info pointer
- * @reg: IO register offset
- *
- * Provides an IO register accessor for the atom interpreter (r4xx+).
- * Returns the value of the IO register.
- */
-static uint32_t cail_ioreg_read(struct card_info *info, uint32_t reg)
-{
-	struct amdgpu_device *adev = drm_to_adev(info->dev);
-	uint32_t r;
-
-	r = RREG32_IO(reg);
-	return r;
-}
-
 static ssize_t amdgpu_atombios_get_vbios_version(struct device *dev,
 						 struct device_attribute *attr,
 						 char *buf)
@@ -1947,7 +1762,7 @@ static ssize_t amdgpu_atombios_get_vbios_version(struct device *dev,
 	struct amdgpu_device *adev = drm_to_adev(ddev);
 	struct atom_context *ctx = adev->mode_info.atom_context;
 
-	return snprintf(buf, PAGE_SIZE, "%s\n", ctx->vbios_version);
+	return sysfs_emit(buf, "%s\n", ctx->vbios_version);
 }
 
 static DEVICE_ATTR(vbios_version, 0444, amdgpu_atombios_get_vbios_version,
@@ -1998,15 +1813,6 @@ int amdgpu_atombios_init(struct amdgpu_device *adev)
 	atom_card_info->dev = adev_to_drm(adev);
 	atom_card_info->reg_read = cail_reg_read;
 	atom_card_info->reg_write = cail_reg_write;
-	/* needed for iio ops */
-	if (adev->rio_mem) {
-		atom_card_info->ioreg_read = cail_ioreg_read;
-		atom_card_info->ioreg_write = cail_ioreg_write;
-	} else {
-		DRM_DEBUG("PCI I/O BAR is not found. Using MMIO to access ATOM BIOS\n");
-		atom_card_info->ioreg_read = cail_reg_read;
-		atom_card_info->ioreg_write = cail_reg_write;
-	}
 	atom_card_info->mc_read = cail_mc_read;
 	atom_card_info->mc_write = cail_mc_write;
 	atom_card_info->pll_read = cail_pll_read;

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

@@ -168,18 +168,6 @@ int amdgpu_atombios_get_memory_pll_dividers(struct amdgpu_device *adev,
 void amdgpu_atombios_set_engine_dram_timings(struct amdgpu_device *adev,
 					     u32 eng_clock, u32 mem_clock);
 
-int amdgpu_atombios_get_leakage_id_from_vbios(struct amdgpu_device *adev,
-					      u16 *leakage_id);
-
-int amdgpu_atombios_get_leakage_vddc_based_on_leakage_params(struct amdgpu_device *adev,
-							     u16 *vddc, u16 *vddci,
-							     u16 virtual_voltage_id,
-							     u16 vbios_voltage_id);
-
-int amdgpu_atombios_get_voltage_evv(struct amdgpu_device *adev,
-				    u16 virtual_voltage_id,
-				    u16 *voltage);
-
 bool
 amdgpu_atombios_is_voltage_gpio(struct amdgpu_device *adev,
 				u8 voltage_type, u8 voltage_mode);

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

@@ -117,12 +117,15 @@ union igp_info {
 
 union umc_info {
 	struct atom_umc_info_v3_1 v31;
+	struct atom_umc_info_v3_2 v32;
+	struct atom_umc_info_v3_3 v33;
 };
 
 union vram_info {
 	struct atom_vram_info_header_v2_3 v23;
 	struct atom_vram_info_header_v2_4 v24;
 	struct atom_vram_info_header_v2_5 v25;
+	struct atom_vram_info_header_v2_6 v26;
 };
 
 union vram_module {
@@ -164,6 +167,7 @@ static int convert_atom_mem_type_to_vram_type(struct amdgpu_device *adev,
 			vram_type = AMDGPU_VRAM_TYPE_GDDR5;
 			break;
 		case ATOM_DGPU_VRAM_TYPE_HBM2:
+		case ATOM_DGPU_VRAM_TYPE_HBM2E:
 			vram_type = AMDGPU_VRAM_TYPE_HBM;
 			break;
 		case ATOM_DGPU_VRAM_TYPE_GDDR6:
@@ -315,6 +319,26 @@ amdgpu_atomfirmware_get_vram_info(struct amdgpu_device *adev,
 				if (vram_vendor)
 					*vram_vendor = mem_vendor;
 				break;
+			case 6:
+				if (module_id > vram_info->v26.vram_module_num)
+					module_id = 0;
+				vram_module = (union vram_module *)vram_info->v26.vram_module;
+				while (i < module_id) {
+					vram_module = (union vram_module *)
+						((u8 *)vram_module + vram_module->v9.vram_module_size);
+					i++;
+				}
+				mem_type = vram_module->v9.memory_type;
+				if (vram_type)
+					*vram_type = convert_atom_mem_type_to_vram_type(adev, mem_type);
+				mem_channel_number = vram_module->v9.channel_num;
+				mem_channel_width = vram_module->v9.channel_width;
+				if (vram_width)
+					*vram_width = mem_channel_number * (1 << mem_channel_width);
+				mem_vendor = (vram_module->v9.vender_rev_id) & 0xF;
+				if (vram_vendor)
+					*vram_vendor = mem_vendor;
+				break;
 			default:
 				return -EINVAL;
 			}
@@ -337,19 +361,39 @@ bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev)
 	union umc_info *umc_info;
 	u8 frev, crev;
 	bool ecc_default_enabled = false;
+	u8 umc_config;
+	u32 umc_config1;
 
 	index = get_index_into_master_table(atom_master_list_of_data_tables_v2_1,
 			umc_info);
 
 	if (amdgpu_atom_parse_data_header(mode_info->atom_context,
 				index, &size, &frev, &crev, &data_offset)) {
-		/* support umc_info 3.1+ */
+		if (frev == 3) {
-		if ((frev == 3 && crev >= 1) || (frev > 3)) {
 			umc_info = (union umc_info *)
 				(mode_info->atom_context->bios + data_offset);
-			ecc_default_enabled =
+			switch (crev) {
-				(le32_to_cpu(umc_info->v31.umc_config) &
+			case 1:
-				 UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false;
+				umc_config = le32_to_cpu(umc_info->v31.umc_config);
+				ecc_default_enabled =
+					(umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false;
+				break;
+			case 2:
+				umc_config = le32_to_cpu(umc_info->v32.umc_config);
+				ecc_default_enabled =
+					(umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false;
+				break;
+			case 3:
+				umc_config = le32_to_cpu(umc_info->v33.umc_config);
+				umc_config1 = le32_to_cpu(umc_info->v33.umc_config1);
+				ecc_default_enabled =
+					((umc_config & UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ||
+					 (umc_config1 & UMC_CONFIG1__ENABLE_ECC_CAPABLE)) ? true : false;
+				break;
+			default:
+				/* unsupported crev */
+				return false;
+			}
 		}
 	}
 
@@ -479,7 +523,8 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev)
 }
 
 union gfx_info {
-	struct  atom_gfx_info_v2_4 v24;
+	struct atom_gfx_info_v2_4 v24;
+	struct atom_gfx_info_v2_7 v27;
 };
 
 int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev)
@@ -514,6 +559,22 @@ int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev)
 			adev->gfx.cu_info.max_scratch_slots_per_cu = gfx_info->v24.gc_max_scratch_slots_per_cu;
 			adev->gfx.cu_info.lds_size = le16_to_cpu(gfx_info->v24.gc_lds_size);
 			return 0;
+		case 7:
+			adev->gfx.config.max_shader_engines = gfx_info->v27.max_shader_engines;
+			adev->gfx.config.max_cu_per_sh = gfx_info->v27.max_cu_per_sh;
+			adev->gfx.config.max_sh_per_se = gfx_info->v27.max_sh_per_se;
+			adev->gfx.config.max_backends_per_se = gfx_info->v27.max_backends_per_se;
+			adev->gfx.config.max_texture_channel_caches = gfx_info->v27.max_texture_channel_caches;
+			adev->gfx.config.max_gprs = le16_to_cpu(gfx_info->v27.gc_num_gprs);
+			adev->gfx.config.max_gs_threads = gfx_info->v27.gc_num_max_gs_thds;
+			adev->gfx.config.gs_vgt_table_depth = gfx_info->v27.gc_gs_table_depth;
+			adev->gfx.config.gs_prim_buffer_depth = le16_to_cpu(gfx_info->v27.gc_gsprim_buff_depth);
+			adev->gfx.config.double_offchip_lds_buf = gfx_info->v27.gc_double_offchip_lds_buffer;
+			adev->gfx.cu_info.wave_front_size = le16_to_cpu(gfx_info->v27.gc_wave_size);
+			adev->gfx.cu_info.max_waves_per_simd = le16_to_cpu(gfx_info->v27.gc_max_waves_per_simd);
+			adev->gfx.cu_info.max_scratch_slots_per_cu = gfx_info->v27.gc_max_scratch_slots_per_cu;
+			adev->gfx.cu_info.lds_size = le16_to_cpu(gfx_info->v27.gc_lds_size);
+			return 0;
 		default:
 			return -EINVAL;
 		}

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

@@ -85,6 +85,8 @@ static void amdgpu_benchmark_move(struct amdgpu_device *adev, unsigned size,
 	bp.flags = 0;
 	bp.type = ttm_bo_type_kernel;
 	bp.resv = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+
 	n = AMDGPU_BENCHMARK_ITERATIONS;
 	r = amdgpu_bo_create(adev, &bp, &sobj);
 	if (r) {

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

@@ -97,6 +97,10 @@ static bool igp_read_bios_from_vram(struct amdgpu_device *adev)
 		if (amdgpu_device_need_post(adev))
 			return false;
 
+	/* FB BAR not enabled */
+	if (pci_resource_len(adev->pdev, 0) == 0)
+		return false;
+
 	adev->bios = NULL;
 	vram_base = pci_resource_start(adev->pdev, 0);
 	bios = ioremap_wc(vram_base, size);
@@ -316,7 +320,7 @@ static bool amdgpu_atrm_get_bios(struct amdgpu_device *adev)
 
 	adev->bios = kmalloc(size, GFP_KERNEL);
 	if (!adev->bios) {
-		DRM_ERROR("Unable to allocate bios\n");
+		dev_err(adev->dev, "Unable to allocate bios\n");
 		return false;
 	}
 
@@ -364,7 +368,7 @@ static bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev)
 		return false;
 	tbl_size = hdr->length;
 	if (tbl_size < sizeof(UEFI_ACPI_VFCT)) {
-		DRM_ERROR("ACPI VFCT table present but broken (too short #1)\n");
+		dev_info(adev->dev, "ACPI VFCT table present but broken (too short #1),skipping\n");
 		return false;
 	}
 
@@ -377,13 +381,13 @@ static bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev)
 
 		offset += sizeof(VFCT_IMAGE_HEADER);
 		if (offset > tbl_size) {
-			DRM_ERROR("ACPI VFCT image header truncated\n");
+			dev_info(adev->dev, "ACPI VFCT image header truncated,skipping\n");
 			return false;
 		}
 
 		offset += vhdr->ImageLength;
 		if (offset > tbl_size) {
-			DRM_ERROR("ACPI VFCT image truncated\n");
+			dev_info(adev->dev, "ACPI VFCT image truncated,skipping\n");
 			return false;
 		}
 
@@ -406,7 +410,7 @@ static bool amdgpu_acpi_vfct_bios(struct amdgpu_device *adev)
 		}
 	}
 
-	DRM_ERROR("ACPI VFCT table present but broken (too short #2)\n");
+	dev_info(adev->dev, "ACPI VFCT table present but broken (too short #2),skipping\n");
 	return false;
 }
 #else
@@ -453,7 +457,7 @@ bool amdgpu_get_bios(struct amdgpu_device *adev)
 		goto success;
 	}
 
-	DRM_ERROR("Unable to locate a BIOS ROM\n");
+	dev_err(adev->dev, "Unable to locate a BIOS ROM\n");
 	return false;
 
 success:

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

@@ -117,7 +117,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 	if (cs->in.num_chunks == 0)
 		return 0;
 
-	chunk_array = kmalloc_array(cs->in.num_chunks, sizeof(uint64_t), GFP_KERNEL);
+	chunk_array = kvmalloc_array(cs->in.num_chunks, sizeof(uint64_t), GFP_KERNEL);
 	if (!chunk_array)
 		return -ENOMEM;
 
@@ -144,7 +144,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 	}
 
 	p->nchunks = cs->in.num_chunks;
-	p->chunks = kmalloc_array(p->nchunks, sizeof(struct amdgpu_cs_chunk),
+	p->chunks = kvmalloc_array(p->nchunks, sizeof(struct amdgpu_cs_chunk),
 			    GFP_KERNEL);
 	if (!p->chunks) {
 		ret = -ENOMEM;
@@ -238,7 +238,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 
 	if (p->uf_entry.tv.bo)
 		p->job->uf_addr = uf_offset;
-	kfree(chunk_array);
+	kvfree(chunk_array);
 
 	/* Use this opportunity to fill in task info for the vm */
 	amdgpu_vm_set_task_info(vm);
@@ -250,11 +250,11 @@ free_all_kdata:
 free_partial_kdata:
 	for (; i >= 0; i--)
 		kvfree(p->chunks[i].kdata);
-	kfree(p->chunks);
+	kvfree(p->chunks);
 	p->chunks = NULL;
 	p->nchunks = 0;
 free_chunk:
-	kfree(chunk_array);
+	kvfree(chunk_array);
 
 	return ret;
 }
@@ -559,7 +559,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 					sizeof(struct page *),
 					GFP_KERNEL | __GFP_ZERO);
 		if (!e->user_pages) {
-			DRM_ERROR("calloc failure\n");
+			DRM_ERROR("kvmalloc_array failure\n");
 			return -ENOMEM;
 		}
 
@@ -706,7 +706,7 @@ static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser, int error,
 
 	for (i = 0; i < parser->nchunks; i++)
 		kvfree(parser->chunks[i].kdata);
-	kfree(parser->chunks);
+	kvfree(parser->chunks);
 	if (parser->job)
 		amdgpu_job_free(parser->job);
 	if (parser->uf_entry.tv.bo) {

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

@@ -28,7 +28,6 @@
 #include <linux/uaccess.h>
 #include <linux/pm_runtime.h>
 #include <linux/poll.h>
-#include <drm/drm_debugfs.h>
 
 #include "amdgpu.h"
 #include "amdgpu_pm.h"
@@ -38,45 +37,6 @@
 #include "amdgpu_securedisplay.h"
 #include "amdgpu_fw_attestation.h"
 
-/**
- * amdgpu_debugfs_add_files - Add simple debugfs entries
- *
- * @adev:  Device to attach debugfs entries to
- * @files:  Array of function callbacks that respond to reads
- * @nfiles: Number of callbacks to register
- *
- */
-int amdgpu_debugfs_add_files(struct amdgpu_device *adev,
-			     const struct drm_info_list *files,
-			     unsigned nfiles)
-{
-	unsigned i;
-
-	for (i = 0; i < adev->debugfs_count; i++) {
-		if (adev->debugfs[i].files == files) {
-			/* Already registered */
-			return 0;
-		}
-	}
-
-	i = adev->debugfs_count + 1;
-	if (i > AMDGPU_DEBUGFS_MAX_COMPONENTS) {
-		DRM_ERROR("Reached maximum number of debugfs components.\n");
-		DRM_ERROR("Report so we increase "
-			  "AMDGPU_DEBUGFS_MAX_COMPONENTS.\n");
-		return -EINVAL;
-	}
-	adev->debugfs[adev->debugfs_count].files = files;
-	adev->debugfs[adev->debugfs_count].num_files = nfiles;
-	adev->debugfs_count = i;
-#if defined(CONFIG_DEBUG_FS)
-	drm_debugfs_create_files(files, nfiles,
-				 adev_to_drm(adev)->primary->debugfs_root,
-				 adev_to_drm(adev)->primary);
-#endif
-	return 0;
-}
-
 int amdgpu_debugfs_wait_dump(struct amdgpu_device *adev)
 {
 #if defined(CONFIG_DEBUG_FS)
@@ -1228,22 +1188,20 @@ int amdgpu_debugfs_regs_init(struct amdgpu_device *adev)
 					  adev, debugfs_regs[i]);
 		if (!i && !IS_ERR_OR_NULL(ent))
 			i_size_write(ent->d_inode, adev->rmmio_size);
-		adev->debugfs_regs[i] = ent;
 	}
 
 	return 0;
 }
 
-static int amdgpu_debugfs_test_ib(struct seq_file *m, void *data)
+static int amdgpu_debugfs_test_ib_show(struct seq_file *m, void *unused)
 {
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct drm_device *dev = node->minor->dev;
+	struct drm_device *dev = adev_to_drm(adev);
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	int r = 0, i;
 
 	r = pm_runtime_get_sync(dev->dev);
 	if (r < 0) {
-		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		pm_runtime_put_autosuspend(dev->dev);
 		return r;
 	}
 
@@ -1285,30 +1243,19 @@ static int amdgpu_debugfs_test_ib(struct seq_file *m, void *data)
 	return 0;
 }
 
-static int amdgpu_debugfs_get_vbios_dump(struct seq_file *m, void *data)
+static int amdgpu_debugfs_evict_vram(void *data, u64 *val)
 {
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
-	struct drm_device *dev = node->minor->dev;
+	struct drm_device *dev = adev_to_drm(adev);
-	struct amdgpu_device *adev = drm_to_adev(dev);
-
-	seq_write(m, adev->bios, adev->bios_size);
-	return 0;
-}
-
-static int amdgpu_debugfs_evict_vram(struct seq_file *m, void *data)
-{
-	struct drm_info_node *node = (struct drm_info_node *)m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	int r;
 
 	r = pm_runtime_get_sync(dev->dev);
 	if (r < 0) {
-		pm_runtime_put_autosuspend(adev_to_drm(adev)->dev);
+		pm_runtime_put_autosuspend(dev->dev);
 		return r;
 	}
 
-	seq_printf(m, "(%d)\n", amdgpu_bo_evict_vram(adev));
+	*val = amdgpu_bo_evict_vram(adev);
 
 	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
@@ -1316,11 +1263,11 @@ static int amdgpu_debugfs_evict_vram(struct seq_file *m, void *data)
 	return 0;
 }
 
-static int amdgpu_debugfs_evict_gtt(struct seq_file *m, void *data)
+
+static int amdgpu_debugfs_evict_gtt(void *data, u64 *val)
 {
-	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
-	struct drm_device *dev = node->minor->dev;
+	struct drm_device *dev = adev_to_drm(adev);
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct ttm_resource_manager *man;
 	int r;
 
@@ -1331,8 +1278,7 @@ static int amdgpu_debugfs_evict_gtt(struct seq_file *m, void *data)
 	}
 
 	man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
-	r = ttm_resource_manager_evict_all(&adev->mman.bdev, man);
+	*val = ttm_resource_manager_evict_all(&adev->mman.bdev, man);
-	seq_printf(m, "(%d)\n", r);
 
 	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
@@ -1340,10 +1286,11 @@ static int amdgpu_debugfs_evict_gtt(struct seq_file *m, void *data)
 	return 0;
 }
 
-static int amdgpu_debugfs_vm_info(struct seq_file *m, void *data)
+
+static int amdgpu_debugfs_vm_info_show(struct seq_file *m, void *unused)
 {
-	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct drm_device *dev = node->minor->dev;
+	struct drm_device *dev = adev_to_drm(adev);
 	struct drm_file *file;
 	int r;
 
@@ -1369,13 +1316,12 @@ static int amdgpu_debugfs_vm_info(struct seq_file *m, void *data)
 	return r;
 }
 
-static const struct drm_info_list amdgpu_debugfs_list[] = {
+DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_test_ib);
-	{"amdgpu_vbios", amdgpu_debugfs_get_vbios_dump},
+DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_vm_info);
-	{"amdgpu_test_ib", &amdgpu_debugfs_test_ib},
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_evict_vram_fops, amdgpu_debugfs_evict_vram,
-	{"amdgpu_evict_vram", &amdgpu_debugfs_evict_vram},
+			 NULL, "%lld\n");
-	{"amdgpu_evict_gtt", &amdgpu_debugfs_evict_gtt},
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_evict_gtt_fops, amdgpu_debugfs_evict_gtt,
-	{"amdgpu_vm_info", &amdgpu_debugfs_vm_info},
+			 NULL, "%lld\n");
-};
 
 static void amdgpu_ib_preempt_fences_swap(struct amdgpu_ring *ring,
 					  struct dma_fence **fences)
@@ -1586,71 +1532,50 @@ static int amdgpu_debugfs_sclk_set(void *data, u64 val)
 	return 0;
 }
 
-DEFINE_SIMPLE_ATTRIBUTE(fops_ib_preempt, NULL,
+DEFINE_DEBUGFS_ATTRIBUTE(fops_ib_preempt, NULL,
 			amdgpu_debugfs_ib_preempt, "%llu\n");
 
-DEFINE_SIMPLE_ATTRIBUTE(fops_sclk_set, NULL,
+DEFINE_DEBUGFS_ATTRIBUTE(fops_sclk_set, NULL,
 			amdgpu_debugfs_sclk_set, "%llu\n");
 
 int amdgpu_debugfs_init(struct amdgpu_device *adev)
 {
+	struct dentry *root = adev_to_drm(adev)->primary->debugfs_root;
+	struct dentry *ent;
 	int r, i;
 
-	adev->debugfs_preempt =
+
-		debugfs_create_file("amdgpu_preempt_ib", 0600,
+
-				    adev_to_drm(adev)->primary->debugfs_root, adev,
+	ent = debugfs_create_file("amdgpu_preempt_ib", 0600, root, adev,
-				    &fops_ib_preempt);
+				  &fops_ib_preempt);
-	if (!(adev->debugfs_preempt)) {
+	if (!ent) {
 		DRM_ERROR("unable to create amdgpu_preempt_ib debugsfs file\n");
 		return -EIO;
 	}
 
-	adev->smu.debugfs_sclk =
+	ent = debugfs_create_file("amdgpu_force_sclk", 0200, root, adev,
-		debugfs_create_file("amdgpu_force_sclk", 0200,
+				  &fops_sclk_set);
-				    adev_to_drm(adev)->primary->debugfs_root, adev,
+	if (!ent) {
-				    &fops_sclk_set);
-	if (!(adev->smu.debugfs_sclk)) {
 		DRM_ERROR("unable to create amdgpu_set_sclk debugsfs file\n");
 		return -EIO;
 	}
 
 	/* Register debugfs entries for amdgpu_ttm */
-	r = amdgpu_ttm_debugfs_init(adev);
+	amdgpu_ttm_debugfs_init(adev);
-	if (r) {
+	amdgpu_debugfs_pm_init(adev);
-		DRM_ERROR("Failed to init debugfs\n");
+	amdgpu_debugfs_sa_init(adev);
-		return r;
+	amdgpu_debugfs_fence_init(adev);
-	}
+	amdgpu_debugfs_gem_init(adev);
-
-	r = amdgpu_debugfs_pm_init(adev);
-	if (r) {
-		DRM_ERROR("Failed to register debugfs file for dpm!\n");
-		return r;
-	}
-
-	if (amdgpu_debugfs_sa_init(adev)) {
-		dev_err(adev->dev, "failed to register debugfs file for SA\n");
-	}
-
-	if (amdgpu_debugfs_fence_init(adev))
-		dev_err(adev->dev, "fence debugfs file creation failed\n");
-
-	r = amdgpu_debugfs_gem_init(adev);
-	if (r)
-		DRM_ERROR("registering gem debugfs failed (%d).\n", r);
 
 	r = amdgpu_debugfs_regs_init(adev);
 	if (r)
 		DRM_ERROR("registering register debugfs failed (%d).\n", r);
 
-	r = amdgpu_debugfs_firmware_init(adev);
+	amdgpu_debugfs_firmware_init(adev);
-	if (r)
-		DRM_ERROR("registering firmware debugfs failed (%d).\n", r);
 
 #if defined(CONFIG_DRM_AMD_DC)
-	if (amdgpu_device_has_dc_support(adev)) {
+	if (amdgpu_device_has_dc_support(adev))
-		if (dtn_debugfs_init(adev))
+		dtn_debugfs_init(adev);
-			DRM_ERROR("amdgpu: failed initialize dtn debugfs support.\n");
-	}
 #endif
 
 	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
@@ -1665,17 +1590,26 @@ int amdgpu_debugfs_init(struct amdgpu_device *adev)
 	}
 
 	amdgpu_ras_debugfs_create_all(adev);
-
 	amdgpu_debugfs_autodump_init(adev);
-
 	amdgpu_rap_debugfs_init(adev);
-
 	amdgpu_securedisplay_debugfs_init(adev);
-
 	amdgpu_fw_attestation_debugfs_init(adev);
 
-	return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_list,
+	debugfs_create_file("amdgpu_evict_vram", 0444, root, adev,
-					ARRAY_SIZE(amdgpu_debugfs_list));
+			    &amdgpu_evict_vram_fops);
+	debugfs_create_file("amdgpu_evict_gtt", 0444, root, adev,
+			    &amdgpu_evict_gtt_fops);
+	debugfs_create_file("amdgpu_test_ib", 0444, root, adev,
+			    &amdgpu_debugfs_test_ib_fops);
+	debugfs_create_file("amdgpu_vm_info", 0444, root, adev,
+			    &amdgpu_debugfs_vm_info_fops);
+
+	adev->debugfs_vbios_blob.data = adev->bios;
+	adev->debugfs_vbios_blob.size = adev->bios_size;
+	debugfs_create_blob("amdgpu_vbios", 0444, root,
+			    &adev->debugfs_vbios_blob);
+
+	return 0;
 }
 
 #else

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

@@ -26,11 +26,6 @@
 /*
  * Debugfs
  */
-struct amdgpu_debugfs {
-	const struct drm_info_list	*files;
-	unsigned		num_files;
-};
-
 struct amdgpu_autodump {
 	struct completion		dumping;
 	struct wait_queue_head		gpu_hang;
@@ -39,10 +34,7 @@ struct amdgpu_autodump {
 int amdgpu_debugfs_regs_init(struct amdgpu_device *adev);
 int amdgpu_debugfs_init(struct amdgpu_device *adev);
 void amdgpu_debugfs_fini(struct amdgpu_device *adev);
-int amdgpu_debugfs_add_files(struct amdgpu_device *adev,
+void amdgpu_debugfs_fence_init(struct amdgpu_device *adev);
-			     const struct drm_info_list *files,
+void amdgpu_debugfs_firmware_init(struct amdgpu_device *adev);
-			     unsigned nfiles);
+void amdgpu_debugfs_gem_init(struct amdgpu_device *adev);
-int amdgpu_debugfs_fence_init(struct amdgpu_device *adev);
-int amdgpu_debugfs_firmware_init(struct amdgpu_device *adev);
-int amdgpu_debugfs_gem_init(struct amdgpu_device *adev);
 int amdgpu_debugfs_wait_dump(struct amdgpu_device *adev);

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

@@ -870,17 +870,62 @@ static int amdgpu_display_get_fb_info(const struct amdgpu_framebuffer *amdgpu_fb
 	return r;
 }
 
+int amdgpu_display_gem_fb_init(struct drm_device *dev,
+			       struct amdgpu_framebuffer *rfb,
+			       const struct drm_mode_fb_cmd2 *mode_cmd,
+			       struct drm_gem_object *obj)
+{
+	int ret;
+
+	rfb->base.obj[0] = obj;
+	drm_helper_mode_fill_fb_struct(dev, &rfb->base, mode_cmd);
+	ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs);
+	if (ret)
+		goto err;
+
+	ret = amdgpu_display_framebuffer_init(dev, rfb, mode_cmd, obj);
+	if (ret)
+		goto err;
+
+	return 0;
+err:
+	drm_err(dev, "Failed to init gem fb: %d\n", ret);
+	rfb->base.obj[0] = NULL;
+	return ret;
+}
+
+int amdgpu_display_gem_fb_verify_and_init(
+	struct drm_device *dev, struct amdgpu_framebuffer *rfb,
+	struct drm_file *file_priv, const struct drm_mode_fb_cmd2 *mode_cmd,
+	struct drm_gem_object *obj)
+{
+	int ret;
+
+	rfb->base.obj[0] = obj;
+
+	/* Verify that bo size can fit the fb size. */
+	ret = drm_gem_fb_init_with_funcs(dev, &rfb->base, file_priv, mode_cmd,
+					 &amdgpu_fb_funcs);
+	if (ret)
+		goto err;
+
+	ret = amdgpu_display_framebuffer_init(dev, rfb, mode_cmd, obj);
+	if (ret)
+		goto err;
+
+	return 0;
+err:
+	drm_err(dev, "Failed to verify and init gem fb: %d\n", ret);
+	rfb->base.obj[0] = NULL;
+	return ret;
+}
+
 int amdgpu_display_framebuffer_init(struct drm_device *dev,
 				    struct amdgpu_framebuffer *rfb,
 				    const struct drm_mode_fb_cmd2 *mode_cmd,
 				    struct drm_gem_object *obj)
 {
 	int ret, i;
-	rfb->base.obj[0] = obj;
-	drm_helper_mode_fill_fb_struct(dev, &rfb->base, mode_cmd);
-	ret = drm_framebuffer_init(dev, &rfb->base, &amdgpu_fb_funcs);
-	if (ret)
-		goto fail;
 
 	/*
 	 * This needs to happen before modifier conversion as that might change
@@ -891,13 +936,13 @@ int amdgpu_display_framebuffer_init(struct drm_device *dev,
 			drm_dbg_kms(dev, "Plane 0 and %d have different BOs: %u vs. %u\n",
 				    i, mode_cmd->handles[0], mode_cmd->handles[i]);
 			ret = -EINVAL;
-			goto fail;
+			return ret;
 		}
 	}
 
 	ret = amdgpu_display_get_fb_info(rfb, &rfb->tiling_flags, &rfb->tmz_surface);
 	if (ret)
-		goto fail;
+		return ret;
 
 	if (dev->mode_config.allow_fb_modifiers &&
 	    !(rfb->base.flags & DRM_MODE_FB_MODIFIERS)) {
@@ -905,20 +950,17 @@ int amdgpu_display_framebuffer_init(struct drm_device *dev,
 		if (ret) {
 			drm_dbg_kms(dev, "Failed to convert tiling flags 0x%llX to a modifier",
 				    rfb->tiling_flags);
-			goto fail;
+			return ret;
 		}
 	}
 
 	for (i = 1; i < rfb->base.format->num_planes; ++i) {
+		drm_gem_object_get(rfb->base.obj[0]);
+		drm_gem_object_put(rfb->base.obj[i]);
 		rfb->base.obj[i] = rfb->base.obj[0];
-		drm_gem_object_get(rfb->base.obj[i]);
 	}
 
 	return 0;
-
-fail:
-	rfb->base.obj[0] = NULL;
-	return ret;
 }
 
 struct drm_framebuffer *
@@ -953,13 +995,15 @@ amdgpu_display_user_framebuffer_create(struct drm_device *dev,
 		return ERR_PTR(-ENOMEM);
 	}
 
-	ret = amdgpu_display_framebuffer_init(dev, amdgpu_fb, mode_cmd, obj);
+	ret = amdgpu_display_gem_fb_verify_and_init(dev, amdgpu_fb, file_priv,
+						    mode_cmd, obj);
 	if (ret) {
 		kfree(amdgpu_fb);
 		drm_gem_object_put(obj);
 		return ERR_PTR(ret);
 	}
 
+	drm_gem_object_put(obj);
 	return &amdgpu_fb->base;
 }
 

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

@@ -321,17 +321,12 @@ static void amdgpu_dma_buf_unmap(struct dma_buf_attachment *attach,
 				 struct sg_table *sgt,
 				 enum dma_data_direction dir)
 {
-	struct dma_buf *dma_buf = attach->dmabuf;
-	struct drm_gem_object *obj = dma_buf->priv;
-	struct amdgpu_bo *bo = gem_to_amdgpu_bo(obj);
-	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
-
 	if (sgt->sgl->page_link) {
 		dma_unmap_sgtable(attach->dev, sgt, dir, 0);
 		sg_free_table(sgt);
 		kfree(sgt);
 	} else {
-		amdgpu_vram_mgr_free_sgt(adev, attach->dev, dir, sgt);
+		amdgpu_vram_mgr_free_sgt(attach->dev, dir, sgt);
 	}
 }
 
@@ -434,22 +429,22 @@ amdgpu_dma_buf_create_obj(struct drm_device *dev, struct dma_buf *dma_buf)
 {
 	struct dma_resv *resv = dma_buf->resv;
 	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct amdgpu_bo *bo;
-	struct amdgpu_bo_param bp;
 	struct drm_gem_object *gobj;
+	struct amdgpu_bo *bo;
+	uint64_t flags = 0;
 	int ret;
 
-	memset(&bp, 0, sizeof(bp));
-	bp.size = dma_buf->size;
-	bp.byte_align = PAGE_SIZE;
-	bp.domain = AMDGPU_GEM_DOMAIN_CPU;
-	bp.flags = 0;
-	bp.type = ttm_bo_type_sg;
-	bp.resv = resv;
 	dma_resv_lock(resv, NULL);
+
+	if (dma_buf->ops == &amdgpu_dmabuf_ops) {
+		struct amdgpu_bo *other = gem_to_amdgpu_bo(dma_buf->priv);
+
+		flags |= other->flags & AMDGPU_GEM_CREATE_CPU_GTT_USWC;
+	}
+
 	ret = amdgpu_gem_object_create(adev, dma_buf->size, PAGE_SIZE,
-			AMDGPU_GEM_DOMAIN_CPU,
+				       AMDGPU_GEM_DOMAIN_CPU, flags,
-			0, ttm_bo_type_sg, resv, &gobj);
+				       ttm_bo_type_sg, resv, &gobj);
 	if (ret)
 		goto error;
 

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

@@ -36,6 +36,7 @@
 #include <linux/vga_switcheroo.h>
 #include <drm/drm_probe_helper.h>
 #include <linux/mmu_notifier.h>
+#include <linux/suspend.h>
 
 #include "amdgpu.h"
 #include "amdgpu_irq.h"
@@ -45,6 +46,8 @@
 #include "amdgpu_amdkfd.h"
 
 #include "amdgpu_ras.h"
+#include "amdgpu_xgmi.h"
+#include "amdgpu_reset.h"
 
 /*
  * KMS wrapper.
@@ -90,9 +93,10 @@
  * - 3.38.0 - Add AMDGPU_IB_FLAG_EMIT_MEM_SYNC
  * - 3.39.0 - DMABUF implicit sync does a full pipeline sync
  * - 3.40.0 - Add AMDGPU_IDS_FLAGS_TMZ
+ * - 3.41.0 - Add video codec query
  */
 #define KMS_DRIVER_MAJOR	3
-#define KMS_DRIVER_MINOR	40
+#define KMS_DRIVER_MINOR	41
 #define KMS_DRIVER_PATCHLEVEL	0
 
 int amdgpu_vram_limit;
@@ -145,6 +149,7 @@ int amdgpu_compute_multipipe = -1;
 int amdgpu_gpu_recovery = -1; /* auto */
 int amdgpu_emu_mode;
 uint amdgpu_smu_memory_pool_size;
+int amdgpu_smu_pptable_id = -1;
 /*
  * FBC (bit 0) disabled by default
  * MULTI_MON_PP_MCLK_SWITCH (bit 1) enabled by default
@@ -162,16 +167,26 @@ int amdgpu_discovery = -1;
 int amdgpu_mes;
 int amdgpu_noretry = -1;
 int amdgpu_force_asic_type = -1;
-int amdgpu_tmz;
+int amdgpu_tmz = -1; /* auto */
+uint amdgpu_freesync_vid_mode;
 int amdgpu_reset_method = -1; /* auto */
 int amdgpu_num_kcq = -1;
 
+static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work);
+
 struct amdgpu_mgpu_info mgpu_info = {
 	.mutex = __MUTEX_INITIALIZER(mgpu_info.mutex),
+	.delayed_reset_work = __DELAYED_WORK_INITIALIZER(
+			mgpu_info.delayed_reset_work,
+			amdgpu_drv_delayed_reset_work_handler, 0),
 };
 int amdgpu_ras_enable = -1;
 uint amdgpu_ras_mask = 0xffffffff;
 int amdgpu_bad_page_threshold = -1;
+struct amdgpu_watchdog_timer amdgpu_watchdog_timer = {
+	.timeout_fatal_disable = false,
+	.period = 0x23, /* default to max. timeout = 1 << 0x23 cycles */
+};
 
 /**
  * DOC: vramlimit (int)
@@ -502,7 +517,7 @@ module_param_named(compute_multipipe, amdgpu_compute_multipipe, int, 0444);
  * DOC: gpu_recovery (int)
  * Set to enable GPU recovery mechanism (1 = enable, 0 = disable). The default is -1 (auto, disabled except SRIOV).
  */
-MODULE_PARM_DESC(gpu_recovery, "Enable GPU recovery mechanism, (1 = enable, 0 = disable, -1 = auto)");
+MODULE_PARM_DESC(gpu_recovery, "Enable GPU recovery mechanism, (2 = advanced tdr mode, 1 = enable, 0 = disable, -1 = auto)");
 module_param_named(gpu_recovery, amdgpu_gpu_recovery, int, 0444);
 
 /**
@@ -527,6 +542,20 @@ module_param_named(ras_enable, amdgpu_ras_enable, int, 0444);
 MODULE_PARM_DESC(ras_mask, "Mask of RAS features to enable (default 0xffffffff), only valid when ras_enable == 1");
 module_param_named(ras_mask, amdgpu_ras_mask, uint, 0444);
 
+/**
+ * DOC: timeout_fatal_disable (bool)
+ * Disable Watchdog timeout fatal error event
+ */
+MODULE_PARM_DESC(timeout_fatal_disable, "disable watchdog timeout fatal error (false = default)");
+module_param_named(timeout_fatal_disable, amdgpu_watchdog_timer.timeout_fatal_disable, bool, 0644);
+
+/**
+ * DOC: timeout_period (uint)
+ * Modify the watchdog timeout max_cycles as (1 << period)
+ */
+MODULE_PARM_DESC(timeout_period, "watchdog timeout period (1 to 0x23(default), timeout maxCycles = (1 << period)");
+module_param_named(timeout_period, amdgpu_watchdog_timer.period, uint, 0644);
+
 /**
  * DOC: si_support (int)
  * Set SI support driver. This parameter works after set config CONFIG_DRM_AMDGPU_SI. For SI asic, when radeon driver is enabled,
@@ -748,6 +777,13 @@ bool no_system_mem_limit;
 module_param(no_system_mem_limit, bool, 0644);
 MODULE_PARM_DESC(no_system_mem_limit, "disable system memory limit (false = default)");
 
+/**
+ * DOC: no_queue_eviction_on_vm_fault (int)
+ * If set, process queues will not be evicted on gpuvm fault. This is to keep the wavefront context for debugging (0 = queue eviction, 1 = no queue eviction). The default is 0 (queue eviction).
+ */
+int amdgpu_no_queue_eviction_on_vm_fault = 0;
+MODULE_PARM_DESC(no_queue_eviction_on_vm_fault, "No queue eviction on VM fault (0 = queue eviction, 1 = no queue eviction)");
+module_param_named(no_queue_eviction_on_vm_fault, amdgpu_no_queue_eviction_on_vm_fault, int, 0444);
 #endif
 
 /**
@@ -792,9 +828,20 @@ module_param_named(backlight, amdgpu_backlight, bint, 0444);
  *
  * The default value: 0 (off).  TODO: change to auto till it is completed.
  */
-MODULE_PARM_DESC(tmz, "Enable TMZ feature (-1 = auto, 0 = off (default), 1 = on)");
+MODULE_PARM_DESC(tmz, "Enable TMZ feature (-1 = auto (default), 0 = off, 1 = on)");
 module_param_named(tmz, amdgpu_tmz, int, 0444);
 
+/**
+ * DOC: freesync_video (uint)
+ * Enabled the optimization to adjust front porch timing to achieve seamless mode change experience
+ * when setting a freesync supported mode for which full modeset is not needed.
+ * The default value: 0 (off).
+ */
+MODULE_PARM_DESC(
+	freesync_video,
+	"Enable freesync modesetting optimization feature (0 = off (default), 1 = on)");
+module_param_named(freesync_video, amdgpu_freesync_vid_mode, uint, 0444);
+
 /**
  * DOC: reset_method (int)
  * GPU reset method (-1 = auto (default), 0 = legacy, 1 = mode0, 2 = mode1, 3 = mode2, 4 = baco, 5 = pci)
@@ -815,6 +862,15 @@ module_param_named(bad_page_threshold, amdgpu_bad_page_threshold, int, 0444);
 MODULE_PARM_DESC(num_kcq, "number of kernel compute queue user want to setup (8 if set to greater than 8 or less than 0, only affect gfx 8+)");
 module_param_named(num_kcq, amdgpu_num_kcq, int, 0444);
 
+/**
+ * DOC: smu_pptable_id (int)
+ * Used to override pptable id. id = 0 use VBIOS pptable.
+ * id > 0 use the soft pptable with specicfied id.
+ */
+MODULE_PARM_DESC(smu_pptable_id,
+	"specify pptable id to be used (-1 = auto(default) value, 0 = use pptable from vbios, > 0 = soft pptable id)");
+module_param_named(smu_pptable_id, amdgpu_smu_pptable_id, int, 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},
@@ -1125,6 +1181,11 @@ static const struct pci_device_id pciidlist[] = {
 	{0x1002, 0x73E2, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_DIMGREY_CAVEFISH},
 	{0x1002, 0x73FF, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_DIMGREY_CAVEFISH},
 
+	/* Aldebaran */
+	{0x1002, 0x7408, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN|AMD_EXP_HW_SUPPORT},
+	{0x1002, 0x740C, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN|AMD_EXP_HW_SUPPORT},
+	{0x1002, 0x740F, PCI_ANY_ID, PCI_ANY_ID, 0, 0, CHIP_ALDEBARAN|AMD_EXP_HW_SUPPORT},
+
 	{0, 0, 0}
 };
 
@@ -1279,6 +1340,98 @@ amdgpu_pci_shutdown(struct pci_dev *pdev)
 	adev->mp1_state = PP_MP1_STATE_NONE;
 }
 
+/**
+ * amdgpu_drv_delayed_reset_work_handler - work handler for reset
+ *
+ * @work: work_struct.
+ */
+static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work)
+{
+	struct list_head device_list;
+	struct amdgpu_device *adev;
+	int i, r;
+	struct amdgpu_reset_context reset_context;
+
+	memset(&reset_context, 0, sizeof(reset_context));
+
+	mutex_lock(&mgpu_info.mutex);
+	if (mgpu_info.pending_reset == true) {
+		mutex_unlock(&mgpu_info.mutex);
+		return;
+	}
+	mgpu_info.pending_reset = true;
+	mutex_unlock(&mgpu_info.mutex);
+
+	/* Use a common context, just need to make sure full reset is done */
+	reset_context.method = AMD_RESET_METHOD_NONE;
+	set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags);
+
+	for (i = 0; i < mgpu_info.num_dgpu; i++) {
+		adev = mgpu_info.gpu_ins[i].adev;
+		reset_context.reset_req_dev = adev;
+		r = amdgpu_device_pre_asic_reset(adev, &reset_context);
+		if (r) {
+			dev_err(adev->dev, "GPU pre asic reset failed with err, %d for drm dev, %s ",
+				r, adev_to_drm(adev)->unique);
+		}
+		if (!queue_work(system_unbound_wq, &adev->xgmi_reset_work))
+			r = -EALREADY;
+	}
+	for (i = 0; i < mgpu_info.num_dgpu; i++) {
+		adev = mgpu_info.gpu_ins[i].adev;
+		flush_work(&adev->xgmi_reset_work);
+		adev->gmc.xgmi.pending_reset = false;
+	}
+
+	/* reset function will rebuild the xgmi hive info , clear it now */
+	for (i = 0; i < mgpu_info.num_dgpu; i++)
+		amdgpu_xgmi_remove_device(mgpu_info.gpu_ins[i].adev);
+
+	INIT_LIST_HEAD(&device_list);
+
+	for (i = 0; i < mgpu_info.num_dgpu; i++)
+		list_add_tail(&mgpu_info.gpu_ins[i].adev->reset_list, &device_list);
+
+	/* unregister the GPU first, reset function will add them back */
+	list_for_each_entry(adev, &device_list, reset_list)
+		amdgpu_unregister_gpu_instance(adev);
+
+	/* Use a common context, just need to make sure full reset is done */
+	set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags);
+	r = amdgpu_do_asic_reset(&device_list, &reset_context);
+
+	if (r) {
+		DRM_ERROR("reinit gpus failure");
+		return;
+	}
+	for (i = 0; i < mgpu_info.num_dgpu; i++) {
+		adev = mgpu_info.gpu_ins[i].adev;
+		if (!adev->kfd.init_complete)
+			amdgpu_amdkfd_device_init(adev);
+		amdgpu_ttm_set_buffer_funcs_status(adev, true);
+	}
+	return;
+}
+
+static int amdgpu_pmops_prepare(struct device *dev)
+{
+	struct drm_device *drm_dev = dev_get_drvdata(dev);
+
+	/* Return a positive number here so
+	 * DPM_FLAG_SMART_SUSPEND works properly
+	 */
+	if (amdgpu_device_supports_boco(drm_dev))
+		return pm_runtime_suspended(dev) &&
+			pm_suspend_via_firmware();
+
+	return 0;
+}
+
+static void amdgpu_pmops_complete(struct device *dev)
+{
+	/* nothing to do */
+}
+
 static int amdgpu_pmops_suspend(struct device *dev)
 {
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
@@ -1364,7 +1517,7 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 	}
 
 	adev->in_runpm = true;
-	if (amdgpu_device_supports_atpx(drm_dev))
+	if (amdgpu_device_supports_px(drm_dev))
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
 	ret = amdgpu_device_suspend(drm_dev, false);
@@ -1373,16 +1526,14 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 		return ret;
 	}
 
-	if (amdgpu_device_supports_atpx(drm_dev)) {
+	if (amdgpu_device_supports_px(drm_dev)) {
 		/* Only need to handle PCI state in the driver for ATPX
 		 * PCI core handles it for _PR3.
 		 */
-		if (!amdgpu_is_atpx_hybrid()) {
+		amdgpu_device_cache_pci_state(pdev);
-			amdgpu_device_cache_pci_state(pdev);
+		pci_disable_device(pdev);
-			pci_disable_device(pdev);
+		pci_ignore_hotplug(pdev);
-			pci_ignore_hotplug(pdev);
+		pci_set_power_state(pdev, PCI_D3cold);
-			pci_set_power_state(pdev, PCI_D3cold);
-		}
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_DYNAMIC_OFF;
 	} else if (amdgpu_device_supports_baco(drm_dev)) {
 		amdgpu_device_baco_enter(drm_dev);
@@ -1401,19 +1552,17 @@ static int amdgpu_pmops_runtime_resume(struct device *dev)
 	if (!adev->runpm)
 		return -EINVAL;
 
-	if (amdgpu_device_supports_atpx(drm_dev)) {
+	if (amdgpu_device_supports_px(drm_dev)) {
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
 		/* Only need to handle PCI state in the driver for ATPX
 		 * PCI core handles it for _PR3.
 		 */
-		if (!amdgpu_is_atpx_hybrid()) {
+		pci_set_power_state(pdev, PCI_D0);
-			pci_set_power_state(pdev, PCI_D0);
+		amdgpu_device_load_pci_state(pdev);
-			amdgpu_device_load_pci_state(pdev);
+		ret = pci_enable_device(pdev);
-			ret = pci_enable_device(pdev);
+		if (ret)
-			if (ret)
+			return ret;
-				return ret;
-		}
 		pci_set_master(pdev);
 	} else if (amdgpu_device_supports_boco(drm_dev)) {
 		/* Only need to handle PCI state in the driver for ATPX
@@ -1424,7 +1573,7 @@ static int amdgpu_pmops_runtime_resume(struct device *dev)
 		amdgpu_device_baco_exit(drm_dev);
 	}
 	ret = amdgpu_device_resume(drm_dev, false);
-	if (amdgpu_device_supports_atpx(drm_dev))
+	if (amdgpu_device_supports_px(drm_dev))
 		drm_dev->switch_power_state = DRM_SWITCH_POWER_ON;
 	adev->in_runpm = false;
 	return 0;
@@ -1505,6 +1654,8 @@ out:
 }
 
 static const struct dev_pm_ops amdgpu_pm_ops = {
+	.prepare = amdgpu_pmops_prepare,
+	.complete = amdgpu_pmops_complete,
 	.suspend = amdgpu_pmops_suspend,
 	.resume = amdgpu_pmops_resume,
 	.freeze = amdgpu_pmops_freeze,

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

@@ -232,8 +232,8 @@ static int amdgpufb_create(struct drm_fb_helper *helper,
 		goto out;
 	}
 
-	ret = amdgpu_display_framebuffer_init(adev_to_drm(adev), &rfbdev->rfb,
+	ret = amdgpu_display_gem_fb_init(adev_to_drm(adev), &rfbdev->rfb,
-					      &mode_cmd, gobj);
+					 &mode_cmd, gobj);
 	if (ret) {
 		DRM_ERROR("failed to initialize framebuffer %d\n", ret);
 		goto out;

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

@@ -36,8 +36,6 @@
 #include <linux/firmware.h>
 #include <linux/pm_runtime.h>
 
-#include <drm/drm_debugfs.h>
-
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
 
@@ -441,7 +439,8 @@ int amdgpu_fence_driver_start_ring(struct amdgpu_ring *ring,
  * Helper function for amdgpu_fence_driver_init().
  */
 int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
-				  unsigned num_hw_submission)
+				  unsigned num_hw_submission,
+				  atomic_t *sched_score)
 {
 	struct amdgpu_device *adev = ring->adev;
 	long timeout;
@@ -469,30 +468,31 @@ int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
 		return -ENOMEM;
 
 	/* No need to setup the GPU scheduler for rings that don't need it */
-	if (!ring->no_scheduler) {
+	if (ring->no_scheduler)
-		switch (ring->funcs->type) {
+		return 0;
-		case AMDGPU_RING_TYPE_GFX:
-			timeout = adev->gfx_timeout;
-			break;
-		case AMDGPU_RING_TYPE_COMPUTE:
-			timeout = adev->compute_timeout;
-			break;
-		case AMDGPU_RING_TYPE_SDMA:
-			timeout = adev->sdma_timeout;
-			break;
-		default:
-			timeout = adev->video_timeout;
-			break;
-		}
 
-		r = drm_sched_init(&ring->sched, &amdgpu_sched_ops,
+	switch (ring->funcs->type) {
-				   num_hw_submission, amdgpu_job_hang_limit,
+	case AMDGPU_RING_TYPE_GFX:
-				   timeout, ring->name);
+		timeout = adev->gfx_timeout;
-		if (r) {
+		break;
-			DRM_ERROR("Failed to create scheduler on ring %s.\n",
+	case AMDGPU_RING_TYPE_COMPUTE:
-				  ring->name);
+		timeout = adev->compute_timeout;
-			return r;
+		break;
-		}
+	case AMDGPU_RING_TYPE_SDMA:
+		timeout = adev->sdma_timeout;
+		break;
+	default:
+		timeout = adev->video_timeout;
+		break;
+	}
+
+	r = drm_sched_init(&ring->sched, &amdgpu_sched_ops,
+			   num_hw_submission, amdgpu_job_hang_limit,
+			   timeout, sched_score, ring->name);
+	if (r) {
+		DRM_ERROR("Failed to create scheduler on ring %s.\n",
+			  ring->name);
+		return r;
 	}
 
 	return 0;
@@ -533,6 +533,8 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev)
 
 		if (!ring || !ring->fence_drv.initialized)
 			continue;
+		if (!ring->no_scheduler)
+			drm_sched_fini(&ring->sched);
 		r = amdgpu_fence_wait_empty(ring);
 		if (r) {
 			/* no need to trigger GPU reset as we are unloading */
@@ -541,8 +543,7 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev)
 		if (ring->fence_drv.irq_src)
 			amdgpu_irq_put(adev, ring->fence_drv.irq_src,
 				       ring->fence_drv.irq_type);
-		if (!ring->no_scheduler)
+
-			drm_sched_fini(&ring->sched);
 		del_timer_sync(&ring->fence_drv.fallback_timer);
 		for (j = 0; j <= ring->fence_drv.num_fences_mask; ++j)
 			dma_fence_put(ring->fence_drv.fences[j]);
@@ -697,11 +698,9 @@ static const struct dma_fence_ops amdgpu_fence_ops = {
  * Fence debugfs
  */
 #if defined(CONFIG_DEBUG_FS)
-static int amdgpu_debugfs_fence_info(struct seq_file *m, void *data)
+static int amdgpu_debugfs_fence_info_show(struct seq_file *m, void *unused)
 {
-	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	int i;
 
 	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
@@ -746,11 +745,10 @@ static int amdgpu_debugfs_fence_info(struct seq_file *m, void *data)
  *
  * Manually trigger a gpu reset at the next fence wait.
  */
-static int amdgpu_debugfs_gpu_recover(struct seq_file *m, void *data)
+static int gpu_recover_get(void *data, u64 *val)
 {
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
-	struct drm_device *dev = node->minor->dev;
+	struct drm_device *dev = adev_to_drm(adev);
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	int r;
 
 	r = pm_runtime_get_sync(dev->dev);
@@ -759,8 +757,7 @@ static int amdgpu_debugfs_gpu_recover(struct seq_file *m, void *data)
 		return 0;
 	}
 
-	seq_printf(m, "gpu recover\n");
+	*val = amdgpu_device_gpu_recover(adev, NULL);
-	amdgpu_device_gpu_recover(adev, NULL);
 
 	pm_runtime_mark_last_busy(dev->dev);
 	pm_runtime_put_autosuspend(dev->dev);
@@ -768,26 +765,24 @@ static int amdgpu_debugfs_gpu_recover(struct seq_file *m, void *data)
 	return 0;
 }
 
-static const struct drm_info_list amdgpu_debugfs_fence_list[] = {
+DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_fence_info);
-	{"amdgpu_fence_info", &amdgpu_debugfs_fence_info, 0, NULL},
+DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_gpu_recover_fops, gpu_recover_get, NULL,
-	{"amdgpu_gpu_recover", &amdgpu_debugfs_gpu_recover, 0, NULL}
+			 "%lld\n");
-};
 
-static const struct drm_info_list amdgpu_debugfs_fence_list_sriov[] = {
-	{"amdgpu_fence_info", &amdgpu_debugfs_fence_info, 0, NULL},
-};
 #endif
 
-int amdgpu_debugfs_fence_init(struct amdgpu_device *adev)
+void amdgpu_debugfs_fence_init(struct amdgpu_device *adev)
 {
 #if defined(CONFIG_DEBUG_FS)
-	if (amdgpu_sriov_vf(adev))
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
-		return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_fence_list_sriov,
+	struct dentry *root = minor->debugfs_root;
-						ARRAY_SIZE(amdgpu_debugfs_fence_list_sriov));
+
-	return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_fence_list,
+	debugfs_create_file("amdgpu_fence_info", 0444, root, adev,
-					ARRAY_SIZE(amdgpu_debugfs_fence_list));
+			    &amdgpu_debugfs_fence_info_fops);
-#else
+
-	return 0;
+	if (!amdgpu_sriov_vf(adev))
+		debugfs_create_file("amdgpu_gpu_recover", 0444, root, adev,
+				    &amdgpu_debugfs_gpu_recover_fops);
 #endif
 }
 

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

@@ -71,7 +71,7 @@
  */
 static int amdgpu_gart_dummy_page_init(struct amdgpu_device *adev)
 {
-	struct page *dummy_page = ttm_bo_glob.dummy_read_page;
+	struct page *dummy_page = ttm_glob.dummy_read_page;
 
 	if (adev->dummy_page_addr)
 		return 0;
@@ -126,6 +126,8 @@ int amdgpu_gart_table_vram_alloc(struct amdgpu_device *adev)
 			AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 		bp.type = ttm_bo_type_kernel;
 		bp.resv = NULL;
+		bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+
 		r = amdgpu_bo_create(adev, &bp, &adev->gart.bo);
 		if (r) {
 			return r;
@@ -202,6 +204,7 @@ void amdgpu_gart_table_vram_free(struct amdgpu_device *adev)
 		return;
 	}
 	amdgpu_bo_unref(&adev->gart.bo);
+	adev->gart.ptr = NULL;
 }
 
 /*
@@ -236,9 +239,6 @@ int amdgpu_gart_unbind(struct amdgpu_device *adev, uint64_t offset,
 	t = offset / AMDGPU_GPU_PAGE_SIZE;
 	p = t / AMDGPU_GPU_PAGES_IN_CPU_PAGE;
 	for (i = 0; i < pages; i++, p++) {
-#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
-		adev->gart.pages[p] = NULL;
-#endif
 		page_base = adev->dummy_page_addr;
 		if (!adev->gart.ptr)
 			continue;
@@ -312,9 +312,6 @@ int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
 		     int pages, struct page **pagelist, dma_addr_t *dma_addr,
 		     uint64_t flags)
 {
-#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
-	unsigned t,p;
-#endif
 	int r, i;
 
 	if (!adev->gart.ready) {
@@ -322,13 +319,6 @@ int amdgpu_gart_bind(struct amdgpu_device *adev, uint64_t offset,
 		return -EINVAL;
 	}
 
-#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
-	t = offset / AMDGPU_GPU_PAGE_SIZE;
-	p = t / AMDGPU_GPU_PAGES_IN_CPU_PAGE;
-	for (i = 0; i < pages; i++, p++)
-		adev->gart.pages[p] = pagelist ? pagelist[i] : NULL;
-#endif
-
 	if (!adev->gart.ptr)
 		return 0;
 
@@ -373,14 +363,6 @@ 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_DRM_AMDGPU_GART_DEBUGFS
-	/* Allocate pages table */
-	adev->gart.pages = vzalloc(array_size(sizeof(void *),
-					      adev->gart.num_cpu_pages));
-	if (adev->gart.pages == NULL)
-		return -ENOMEM;
-#endif
-
 	return 0;
 }
 
@@ -393,9 +375,5 @@ int amdgpu_gart_init(struct amdgpu_device *adev)
  */
 void amdgpu_gart_fini(struct amdgpu_device *adev)
 {
-#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
-	vfree(adev->gart.pages);
-	adev->gart.pages = NULL;
-#endif
 	amdgpu_gart_dummy_page_fini(adev);
 }

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

@@ -46,9 +46,6 @@ struct amdgpu_gart {
 	unsigned			num_gpu_pages;
 	unsigned			num_cpu_pages;
 	unsigned			table_size;
-#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
-	struct page			**pages;
-#endif
 	bool				ready;
 
 	/* Asic default pte flags */

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

@@ -32,7 +32,6 @@
 #include <linux/dma-buf.h>
 
 #include <drm/amdgpu_drm.h>
-#include <drm/drm_debugfs.h>
 #include <drm/drm_gem_ttm_helper.h>
 
 #include "amdgpu.h"
@@ -59,6 +58,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 			     struct drm_gem_object **obj)
 {
 	struct amdgpu_bo *bo;
+	struct amdgpu_bo_user *ubo;
 	struct amdgpu_bo_param bp;
 	int r;
 
@@ -72,10 +72,13 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
 	bp.preferred_domain = initial_domain;
 	bp.flags = flags;
 	bp.domain = initial_domain;
-	r = amdgpu_bo_create(adev, &bp, &bo);
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+
+	r = amdgpu_bo_create_user(adev, &bp, &ubo);
 	if (r)
 		return r;
 
+	bo = &ubo->bo;
 	*obj = &bo->tbo.base;
 	(*obj)->funcs = &amdgpu_gem_object_funcs;
 
@@ -855,10 +858,10 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv,
 }
 
 #if defined(CONFIG_DEBUG_FS)
-static int amdgpu_debugfs_gem_info(struct seq_file *m, void *data)
+static int amdgpu_debugfs_gem_info_show(struct seq_file *m, void *unused)
 {
-	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct drm_device *dev = node->minor->dev;
+	struct drm_device *dev = adev_to_drm(adev);
 	struct drm_file *file;
 	int r;
 
@@ -896,16 +899,17 @@ static int amdgpu_debugfs_gem_info(struct seq_file *m, void *data)
 	return 0;
 }
 
-static const struct drm_info_list amdgpu_debugfs_gem_list[] = {
+DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_gem_info);
-	{"amdgpu_gem_info", &amdgpu_debugfs_gem_info, 0, NULL},
+
-};
 #endif
 
-int amdgpu_debugfs_gem_init(struct amdgpu_device *adev)
+void amdgpu_debugfs_gem_init(struct amdgpu_device *adev)
 {
 #if defined(CONFIG_DEBUG_FS)
-	return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_gem_list,
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
-					ARRAY_SIZE(amdgpu_debugfs_gem_list));
+	struct dentry *root = minor->debugfs_root;
+
+	debugfs_create_file("amdgpu_gem_info", 0444, root, adev,
+			    &amdgpu_debugfs_gem_info_fops);
 #endif
-	return 0;
 }

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

@@ -310,9 +310,8 @@ int amdgpu_gfx_kiq_init_ring(struct amdgpu_device *adev,
 	ring->eop_gpu_addr = kiq->eop_gpu_addr;
 	ring->no_scheduler = true;
 	sprintf(ring->name, "kiq_%d.%d.%d", ring->me, ring->pipe, ring->queue);
-	r = amdgpu_ring_init(adev, ring, 1024,
+	r = amdgpu_ring_init(adev, ring, 1024, irq, AMDGPU_CP_KIQ_IRQ_DRIVER0,
-			     irq, AMDGPU_CP_KIQ_IRQ_DRIVER0,
+			     AMDGPU_RING_PRIO_DEFAULT, NULL);
-			     AMDGPU_RING_PRIO_DEFAULT);
 	if (r)
 		dev_warn(adev->dev, "(%d) failed to init kiq ring\n", r);
 
@@ -463,20 +462,25 @@ int amdgpu_gfx_disable_kcq(struct amdgpu_device *adev)
 {
 	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
 	struct amdgpu_ring *kiq_ring = &kiq->ring;
-	int i;
+	int i, r;
 
 	if (!kiq->pmf || !kiq->pmf->kiq_unmap_queues)
 		return -EINVAL;
 
+	spin_lock(&adev->gfx.kiq.ring_lock);
 	if (amdgpu_ring_alloc(kiq_ring, kiq->pmf->unmap_queues_size *
-					adev->gfx.num_compute_rings))
+					adev->gfx.num_compute_rings)) {
+		spin_unlock(&adev->gfx.kiq.ring_lock);
 		return -ENOMEM;
+	}
 
 	for (i = 0; i < adev->gfx.num_compute_rings; i++)
 		kiq->pmf->kiq_unmap_queues(kiq_ring, &adev->gfx.compute_ring[i],
 					   RESET_QUEUES, 0, 0);
+	r = amdgpu_ring_test_helper(kiq_ring);
+	spin_unlock(&adev->gfx.kiq.ring_lock);
 
-	return amdgpu_ring_test_helper(kiq_ring);
+	return r;
 }
 
 int amdgpu_queue_mask_bit_to_set_resource_bit(struct amdgpu_device *adev,
@@ -519,12 +523,13 @@ int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev)
 
 	DRM_INFO("kiq ring mec %d pipe %d q %d\n", kiq_ring->me, kiq_ring->pipe,
 							kiq_ring->queue);
-
+	spin_lock(&adev->gfx.kiq.ring_lock);
 	r = amdgpu_ring_alloc(kiq_ring, kiq->pmf->map_queues_size *
 					adev->gfx.num_compute_rings +
 					kiq->pmf->set_resources_size);
 	if (r) {
 		DRM_ERROR("Failed to lock KIQ (%d).\n", r);
+		spin_unlock(&adev->gfx.kiq.ring_lock);
 		return r;
 	}
 
@@ -533,6 +538,7 @@ int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev)
 		kiq->pmf->kiq_map_queues(kiq_ring, &adev->gfx.compute_ring[i]);
 
 	r = amdgpu_ring_test_helper(kiq_ring);
+	spin_unlock(&adev->gfx.kiq.ring_lock);
 	if (r)
 		DRM_ERROR("KCQ enable failed\n");
 
@@ -601,6 +607,7 @@ int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev)
 	struct ras_ih_if ih_info = {
 		.cb = amdgpu_gfx_process_ras_data_cb,
 	};
+	struct ras_query_if info = { 0 };
 
 	if (!adev->gfx.ras_if) {
 		adev->gfx.ras_if = kmalloc(sizeof(struct ras_common_if), GFP_KERNEL);
@@ -612,13 +619,19 @@ int amdgpu_gfx_ras_late_init(struct amdgpu_device *adev)
 		strcpy(adev->gfx.ras_if->name, "gfx");
 	}
 	fs_info.head = ih_info.head = *adev->gfx.ras_if;
-
 	r = amdgpu_ras_late_init(adev, adev->gfx.ras_if,
 				 &fs_info, &ih_info);
 	if (r)
 		goto free;
 
 	if (amdgpu_ras_is_supported(adev, adev->gfx.ras_if->block)) {
+		if (adev->gmc.xgmi.connected_to_cpu) {
+			info.head = *adev->gfx.ras_if;
+			amdgpu_ras_query_error_status(adev, &info);
+		} else {
+			amdgpu_ras_reset_error_status(adev, AMDGPU_RAS_BLOCK__GFX);
+		}
+
 		r = amdgpu_irq_get(adev, &adev->gfx.cp_ecc_error_irq, 0);
 		if (r)
 			goto late_fini;
@@ -664,8 +677,9 @@ int amdgpu_gfx_process_ras_data_cb(struct amdgpu_device *adev,
 	 */
 	if (!amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__GFX)) {
 		kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
-		if (adev->gfx.funcs->query_ras_error_count)
+		if (adev->gfx.ras_funcs &&
-			adev->gfx.funcs->query_ras_error_count(adev, err_data);
+		    adev->gfx.ras_funcs->query_ras_error_count)
+			adev->gfx.ras_funcs->query_ras_error_count(adev, err_data);
 		amdgpu_ras_reset_gpu(adev);
 	}
 	return AMDGPU_RAS_SUCCESS;
@@ -698,7 +712,7 @@ uint32_t amdgpu_kiq_rreg(struct amdgpu_device *adev, uint32_t reg)
 	struct amdgpu_kiq *kiq = &adev->gfx.kiq;
 	struct amdgpu_ring *ring = &kiq->ring;
 
-	if (adev->in_pci_err_recovery)
+	if (amdgpu_device_skip_hw_access(adev))
 		return 0;
 
 	BUG_ON(!ring->funcs->emit_rreg);
@@ -765,7 +779,7 @@ void amdgpu_kiq_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v)
 
 	BUG_ON(!ring->funcs->emit_wreg);
 
-	if (adev->in_pci_err_recovery)
+	if (amdgpu_device_skip_hw_access(adev))
 		return;
 
 	spin_lock_irqsave(&kiq->ring_lock, flags);
@@ -829,14 +843,10 @@ int amdgpu_gfx_get_num_kcq(struct amdgpu_device *adev)
 
 void amdgpu_gfx_state_change_set(struct amdgpu_device *adev, enum gfx_change_state state)
 {
-	if (is_support_sw_smu(adev)) {
+	mutex_lock(&adev->pm.mutex);
-		smu_gfx_state_change_set(&adev->smu, state);
+	if (adev->powerplay.pp_funcs &&
-	} else {
+	    adev->powerplay.pp_funcs->gfx_state_change_set)
-		mutex_lock(&adev->pm.mutex);
+		((adev)->powerplay.pp_funcs->gfx_state_change_set(
-		if (adev->powerplay.pp_funcs &&
+			(adev)->powerplay.pp_handle, state));
-		    adev->powerplay.pp_funcs->gfx_state_change_set)
+	mutex_unlock(&adev->pm.mutex);
-			((adev)->powerplay.pp_funcs->gfx_state_change_set(
-				(adev)->powerplay.pp_handle, state));
-		mutex_unlock(&adev->pm.mutex);
-	}
 }

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

@@ -30,6 +30,7 @@
 #include "clearstate_defs.h"
 #include "amdgpu_ring.h"
 #include "amdgpu_rlc.h"
+#include "soc15.h"
 
 /* GFX current status */
 #define AMDGPU_GFX_NORMAL_MODE			0x00000000L
@@ -204,6 +205,19 @@ struct amdgpu_cu_info {
 	uint32_t bitmap[4][4];
 };
 
+struct amdgpu_gfx_ras_funcs {
+	int (*ras_late_init)(struct amdgpu_device *adev);
+	void (*ras_fini)(struct amdgpu_device *adev);
+	int (*ras_error_inject)(struct amdgpu_device *adev,
+				void *inject_if);
+	int (*query_ras_error_count)(struct amdgpu_device *adev,
+				     void *ras_error_status);
+	void (*reset_ras_error_count)(struct amdgpu_device *adev);
+	void (*query_ras_error_status)(struct amdgpu_device *adev);
+	void (*reset_ras_error_status)(struct amdgpu_device *adev);
+	void (*enable_watchdog_timer)(struct amdgpu_device *adev);
+};
+
 struct amdgpu_gfx_funcs {
 	/* get the gpu clock counter */
 	uint64_t (*get_gpu_clock_counter)(struct amdgpu_device *adev);
@@ -219,11 +233,7 @@ struct amdgpu_gfx_funcs {
 				uint32_t *dst);
 	void (*select_me_pipe_q)(struct amdgpu_device *adev, u32 me, u32 pipe,
 				 u32 queue, u32 vmid);
-	int (*ras_error_inject)(struct amdgpu_device *adev, void *inject_if);
-	int (*query_ras_error_count) (struct amdgpu_device *adev, void *ras_error_status);
-	void (*reset_ras_error_count) (struct amdgpu_device *adev);
 	void (*init_spm_golden)(struct amdgpu_device *adev);
-	void (*query_ras_error_status) (struct amdgpu_device *adev);
 	void (*update_perfmon_mgcg)(struct amdgpu_device *adev, bool enable);
 };
 
@@ -327,7 +337,8 @@ struct amdgpu_gfx {
 	DECLARE_BITMAP			(pipe_reserve_bitmap, AMDGPU_MAX_COMPUTE_QUEUES);
 
 	/*ras */
-	struct ras_common_if		*ras_if;
+	struct ras_common_if			*ras_if;
+	const struct amdgpu_gfx_ras_funcs	*ras_funcs;
 };
 
 #define amdgpu_gfx_get_gpu_clock_counter(adev) (adev)->gfx.funcs->get_gpu_clock_counter((adev))

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

@@ -31,6 +31,59 @@
 #include "amdgpu_ras.h"
 #include "amdgpu_xgmi.h"
 
+/**
+ * amdgpu_gmc_pdb0_alloc - allocate vram for pdb0
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Allocate video memory for pdb0 and map it for CPU access
+ * Returns 0 for success, error for failure.
+ */
+int amdgpu_gmc_pdb0_alloc(struct amdgpu_device *adev)
+{
+	int r;
+	struct amdgpu_bo_param bp;
+	u64 vram_size = adev->gmc.xgmi.node_segment_size * adev->gmc.xgmi.num_physical_nodes;
+	uint32_t pde0_page_shift = adev->gmc.vmid0_page_table_block_size + 21;
+	uint32_t npdes = (vram_size + (1ULL << pde0_page_shift) -1) >> pde0_page_shift;
+
+	memset(&bp, 0, sizeof(bp));
+	bp.size = PAGE_ALIGN((npdes + 1) * 8);
+	bp.byte_align = PAGE_SIZE;
+	bp.domain = AMDGPU_GEM_DOMAIN_VRAM;
+	bp.flags = AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED |
+		AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
+	bp.type = ttm_bo_type_kernel;
+	bp.resv = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
+
+	r = amdgpu_bo_create(adev, &bp, &adev->gmc.pdb0_bo);
+	if (r)
+		return r;
+
+	r = amdgpu_bo_reserve(adev->gmc.pdb0_bo, false);
+	if (unlikely(r != 0))
+		goto bo_reserve_failure;
+
+	r = amdgpu_bo_pin(adev->gmc.pdb0_bo, AMDGPU_GEM_DOMAIN_VRAM);
+	if (r)
+		goto bo_pin_failure;
+	r = amdgpu_bo_kmap(adev->gmc.pdb0_bo, &adev->gmc.ptr_pdb0);
+	if (r)
+		goto bo_kmap_failure;
+
+	amdgpu_bo_unreserve(adev->gmc.pdb0_bo);
+	return 0;
+
+bo_kmap_failure:
+	amdgpu_bo_unpin(adev->gmc.pdb0_bo);
+bo_pin_failure:
+	amdgpu_bo_unreserve(adev->gmc.pdb0_bo);
+bo_reserve_failure:
+	amdgpu_bo_unref(&adev->gmc.pdb0_bo);
+	return r;
+}
+
 /**
  * amdgpu_gmc_get_pde_for_bo - get the PDE for a BO
  *
@@ -158,6 +211,39 @@ void amdgpu_gmc_vram_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc,
 			mc->vram_end, mc->real_vram_size >> 20);
 }
 
+/** amdgpu_gmc_sysvm_location - place vram and gart in sysvm aperture
+ *
+ * @adev: amdgpu device structure holding all necessary information
+ * @mc: memory controller structure holding memory information
+ *
+ * This function is only used if use GART for FB translation. In such
+ * case, we use sysvm aperture (vmid0 page tables) for both vram
+ * and gart (aka system memory) access.
+ *
+ * GPUVM (and our organization of vmid0 page tables) require sysvm
+ * aperture to be placed at a location aligned with 8 times of native
+ * page size. For example, if vm_context0_cntl.page_table_block_size
+ * is 12, then native page size is 8G (2M*2^12), sysvm should start
+ * with a 64G aligned address. For simplicity, we just put sysvm at
+ * address 0. So vram start at address 0 and gart is right after vram.
+ */
+void amdgpu_gmc_sysvm_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
+{
+	u64 hive_vram_start = 0;
+	u64 hive_vram_end = mc->xgmi.node_segment_size * mc->xgmi.num_physical_nodes - 1;
+	mc->vram_start = mc->xgmi.node_segment_size * mc->xgmi.physical_node_id;
+	mc->vram_end = mc->vram_start + mc->xgmi.node_segment_size - 1;
+	mc->gart_start = hive_vram_end + 1;
+	mc->gart_end = mc->gart_start + mc->gart_size - 1;
+	mc->fb_start = hive_vram_start;
+	mc->fb_end = hive_vram_end;
+	dev_info(adev->dev, "VRAM: %lluM 0x%016llX - 0x%016llX (%lluM used)\n",
+			mc->mc_vram_size >> 20, mc->vram_start,
+			mc->vram_end, mc->real_vram_size >> 20);
+	dev_info(adev->dev, "GART: %lluM 0x%016llX - 0x%016llX\n",
+			mc->gart_size >> 20, mc->gart_start, mc->gart_end);
+}
+
 /**
  * amdgpu_gmc_gart_location - try to find GART location
  *
@@ -165,7 +251,6 @@ void amdgpu_gmc_vram_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc,
  * @mc: memory controller structure holding memory information
  *
  * Function will place try to place GART before or after VRAM.
- *
  * If GART size is bigger than space left then we ajust GART size.
  * Thus function will never fails.
  */
@@ -176,8 +261,6 @@ void amdgpu_gmc_gart_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc)
 	/*To avoid the hole, limit the max mc address to AMDGPU_GMC_HOLE_START*/
 	u64 max_mc_address = min(adev->gmc.mc_mask, AMDGPU_GMC_HOLE_START - 1);
 
-	mc->gart_size += adev->pm.smu_prv_buffer_size;
-
 	/* VCE doesn't like it when BOs cross a 4GB segment, so align
 	 * the GART base on a 4GB boundary as well.
 	 */
@@ -308,26 +391,46 @@ int amdgpu_gmc_ras_late_init(struct amdgpu_device *adev)
 {
 	int r;
 
-	if (adev->umc.funcs && adev->umc.funcs->ras_late_init) {
+	if (adev->umc.ras_funcs &&
-		r = adev->umc.funcs->ras_late_init(adev);
+	    adev->umc.ras_funcs->ras_late_init) {
+		r = adev->umc.ras_funcs->ras_late_init(adev);
 		if (r)
 			return r;
 	}
 
-	if (adev->mmhub.funcs && adev->mmhub.funcs->ras_late_init) {
+	if (adev->mmhub.ras_funcs &&
-		r = adev->mmhub.funcs->ras_late_init(adev);
+	    adev->mmhub.ras_funcs->ras_late_init) {
+		r = adev->mmhub.ras_funcs->ras_late_init(adev);
 		if (r)
 			return r;
 	}
 
-	return amdgpu_xgmi_ras_late_init(adev);
+	if (!adev->gmc.xgmi.connected_to_cpu)
+		adev->gmc.xgmi.ras_funcs = &xgmi_ras_funcs;
+
+	if (adev->gmc.xgmi.ras_funcs &&
+	    adev->gmc.xgmi.ras_funcs->ras_late_init) {
+		r = adev->gmc.xgmi.ras_funcs->ras_late_init(adev);
+		if (r)
+			return r;
+	}
+
+	return 0;
 }
 
 void amdgpu_gmc_ras_fini(struct amdgpu_device *adev)
 {
-	amdgpu_umc_ras_fini(adev);
+	if (adev->umc.ras_funcs &&
-	amdgpu_mmhub_ras_fini(adev);
+	    adev->umc.ras_funcs->ras_fini)
-	amdgpu_xgmi_ras_fini(adev);
+		adev->umc.ras_funcs->ras_fini(adev);
+
+	if (adev->mmhub.ras_funcs &&
+	    adev->mmhub.ras_funcs->ras_fini)
+		amdgpu_mmhub_ras_fini(adev);
+
+	if (adev->gmc.xgmi.ras_funcs &&
+	    adev->gmc.xgmi.ras_funcs->ras_fini)
+		adev->gmc.xgmi.ras_funcs->ras_fini(adev);
 }
 
 	/*
@@ -384,6 +487,16 @@ void amdgpu_gmc_tmz_set(struct amdgpu_device *adev)
 {
 	switch (adev->asic_type) {
 	case CHIP_RAVEN:
+		if (amdgpu_tmz == 0) {
+			adev->gmc.tmz_enabled = false;
+			dev_info(adev->dev,
+				 "Trusted Memory Zone (TMZ) feature disabled (cmd line)\n");
+		} else {
+			adev->gmc.tmz_enabled = true;
+			dev_info(adev->dev,
+				 "Trusted Memory Zone (TMZ) feature enabled\n");
+		}
+		break;
 	case CHIP_RENOIR:
 	case CHIP_NAVI10:
 	case CHIP_NAVI14:
@@ -423,6 +536,8 @@ void amdgpu_gmc_noretry_set(struct amdgpu_device *adev)
 	switch (adev->asic_type) {
 	case CHIP_VEGA10:
 	case CHIP_VEGA20:
+	case CHIP_ARCTURUS:
+	case CHIP_ALDEBARAN:
 		/*
 		 * noretry = 0 will cause kfd page fault tests fail
 		 * for some ASICs, so set default to 1 for these ASICs.
@@ -518,3 +633,55 @@ void amdgpu_gmc_get_vbios_allocations(struct amdgpu_device *adev)
 		adev->mman.stolen_extended_size = 0;
 	}
 }
+
+/**
+ * amdgpu_gmc_init_pdb0 - initialize PDB0
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * This function is only used when GART page table is used
+ * for FB address translatioin. In such a case, we construct
+ * a 2-level system VM page table: PDB0->PTB, to cover both
+ * VRAM of the hive and system memory.
+ *
+ * PDB0 is static, initialized once on driver initialization.
+ * The first n entries of PDB0 are used as PTE by setting
+ * P bit to 1, pointing to VRAM. The n+1'th entry points
+ * to a big PTB covering system memory.
+ *
+ */
+void amdgpu_gmc_init_pdb0(struct amdgpu_device *adev)
+{
+	int i;
+	uint64_t flags = adev->gart.gart_pte_flags; //TODO it is UC. explore NC/RW?
+	/* Each PDE0 (used as PTE) covers (2^vmid0_page_table_block_size)*2M
+	 */
+	u64 vram_size = adev->gmc.xgmi.node_segment_size * adev->gmc.xgmi.num_physical_nodes;
+	u64 pde0_page_size = (1ULL<<adev->gmc.vmid0_page_table_block_size)<<21;
+	u64 vram_addr = adev->vm_manager.vram_base_offset -
+		adev->gmc.xgmi.physical_node_id * adev->gmc.xgmi.node_segment_size;
+	u64 vram_end = vram_addr + vram_size;
+	u64 gart_ptb_gpu_pa = amdgpu_bo_gpu_offset(adev->gart.bo) +
+		adev->vm_manager.vram_base_offset - adev->gmc.vram_start;
+
+	flags |= AMDGPU_PTE_VALID | AMDGPU_PTE_READABLE;
+	flags |= AMDGPU_PTE_WRITEABLE;
+	flags |= AMDGPU_PTE_SNOOPED;
+	flags |= AMDGPU_PTE_FRAG((adev->gmc.vmid0_page_table_block_size + 9*1));
+	flags |= AMDGPU_PDE_PTE;
+
+	/* The first n PDE0 entries are used as PTE,
+	 * pointing to vram
+	 */
+	for (i = 0; vram_addr < vram_end; i++, vram_addr += pde0_page_size)
+		amdgpu_gmc_set_pte_pde(adev, adev->gmc.ptr_pdb0, i, vram_addr, flags);
+
+	/* The n+1'th PDE0 entry points to a huge
+	 * PTB who has more than 512 entries each
+	 * pointing to a 4K system page
+	 */
+	flags = AMDGPU_PTE_VALID;
+	flags |= AMDGPU_PDE_BFS(0) | AMDGPU_PTE_SNOOPED;
+	/* Requires gart_ptb_gpu_pa to be 4K aligned */
+	amdgpu_gmc_set_pte_pde(adev, adev->gmc.ptr_pdb0, i, gart_ptb_gpu_pa, flags);
+}

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

@@ -135,6 +135,14 @@ struct amdgpu_gmc_funcs {
 	unsigned int (*get_vbios_fb_size)(struct amdgpu_device *adev);
 };
 
+struct amdgpu_xgmi_ras_funcs {
+	int (*ras_late_init)(struct amdgpu_device *adev);
+	void (*ras_fini)(struct amdgpu_device *adev);
+	int (*query_ras_error_count)(struct amdgpu_device *adev,
+				     void *ras_error_status);
+	void (*reset_ras_error_count)(struct amdgpu_device *adev);
+};
+
 struct amdgpu_xgmi {
 	/* from psp */
 	u64 node_id;
@@ -149,6 +157,9 @@ struct amdgpu_xgmi {
 	struct list_head head;
 	bool supported;
 	struct ras_common_if *ras_if;
+	bool connected_to_cpu;
+	bool pending_reset;
+	const struct amdgpu_xgmi_ras_funcs *ras_funcs;
 };
 
 struct amdgpu_gmc {
@@ -189,10 +200,13 @@ struct amdgpu_gmc {
 	u64			gart_end;
 	/* Frame buffer aperture of this GPU device. Different from
 	 * fb_start (see below), this only covers the local GPU device.
-	 * Driver get fb_start from MC_VM_FB_LOCATION_BASE (set by vbios)
+	 * If driver uses FB aperture to access FB, driver get fb_start from
-	 * and calculate vram_start of this local device by adding an
+	 * MC_VM_FB_LOCATION_BASE (set by vbios) and calculate vram_start
-	 * offset inside the XGMI hive.
+	 * of this local device by adding an offset inside the XGMI hive.
-	 * Under VMID0, logical address == MC address
+	 * If driver uses GART table for VMID0 FB access, driver finds a hole in
+	 * VMID0's virtual address space to place the SYSVM aperture inside
+	 * which the first part is vram and the second part is gart (covering
+	 * system ram).
 	 */
 	u64			vram_start;
 	u64			vram_end;
@@ -204,6 +218,15 @@ struct amdgpu_gmc {
 	 */
 	u64			fb_start;
 	u64			fb_end;
+	/* In the case of use GART table for vmid0 FB access, [fb_start, fb_end]
+	 * will be squeezed to GART aperture. But we have a PSP FW issue to fix
+	 * for now. To temporarily workaround the PSP FW issue, added below two
+	 * variables to remember the original fb_start/end to re-enable FB
+	 * aperture to workaround the PSP FW issue. Will delete it after we
+	 * get a proper PSP FW fix.
+	 */
+	u64			fb_start_original;
+	u64			fb_end_original;
 	unsigned		vram_width;
 	u64			real_vram_size;
 	int			vram_mtrr;
@@ -240,6 +263,12 @@ struct amdgpu_gmc {
 	struct amdgpu_xgmi xgmi;
 	struct amdgpu_irq_src	ecc_irq;
 	int noretry;
+
+	uint32_t	vmid0_page_table_block_size;
+	uint32_t	vmid0_page_table_depth;
+	struct amdgpu_bo		*pdb0_bo;
+	/* CPU kmapped address of pdb0*/
+	void				*ptr_pdb0;
 };
 
 #define amdgpu_gmc_flush_gpu_tlb(adev, vmid, vmhub, type) ((adev)->gmc.gmc_funcs->flush_gpu_tlb((adev), (vmid), (vmhub), (type)))
@@ -281,6 +310,7 @@ static inline uint64_t amdgpu_gmc_sign_extend(uint64_t addr)
 	return addr;
 }
 
+int amdgpu_gmc_pdb0_alloc(struct amdgpu_device *adev);
 void amdgpu_gmc_get_pde_for_bo(struct amdgpu_bo *bo, int level,
 			       uint64_t *addr, uint64_t *flags);
 int amdgpu_gmc_set_pte_pde(struct amdgpu_device *adev, void *cpu_pt_addr,
@@ -288,6 +318,7 @@ int amdgpu_gmc_set_pte_pde(struct amdgpu_device *adev, void *cpu_pt_addr,
 				uint64_t flags);
 uint64_t amdgpu_gmc_pd_addr(struct amdgpu_bo *bo);
 uint64_t amdgpu_gmc_agp_addr(struct ttm_buffer_object *bo);
+void amdgpu_gmc_sysvm_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc);
 void amdgpu_gmc_vram_location(struct amdgpu_device *adev, struct amdgpu_gmc *mc,
 			      u64 base);
 void amdgpu_gmc_gart_location(struct amdgpu_device *adev,
@@ -309,4 +340,5 @@ amdgpu_gmc_set_vm_fault_masks(struct amdgpu_device *adev, int hub_type,
 
 void amdgpu_gmc_get_vbios_allocations(struct amdgpu_device *adev);
 
+void amdgpu_gmc_init_pdb0(struct amdgpu_device *adev);
 #endif

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

@@ -49,8 +49,7 @@ static ssize_t amdgpu_mem_info_gtt_total_show(struct device *dev,
 	struct amdgpu_device *adev = drm_to_adev(ddev);
 	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
 
-	return snprintf(buf, PAGE_SIZE, "%llu\n",
+	return sysfs_emit(buf, "%llu\n", man->size * PAGE_SIZE);
-			man->size * PAGE_SIZE);
 }
 
 /**
@@ -68,8 +67,7 @@ static ssize_t amdgpu_mem_info_gtt_used_show(struct device *dev,
 	struct amdgpu_device *adev = drm_to_adev(ddev);
 	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT);
 
-	return snprintf(buf, PAGE_SIZE, "%llu\n",
+	return sysfs_emit(buf, "%llu\n", amdgpu_gtt_mgr_usage(man));
-			amdgpu_gtt_mgr_usage(man));
 }
 
 static DEVICE_ATTR(mem_info_gtt_total, S_IRUGO,

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

@@ -30,7 +30,6 @@
 #include <linux/slab.h>
 
 #include <drm/amdgpu_drm.h>
-#include <drm/drm_debugfs.h>
 
 #include "amdgpu.h"
 #include "atom.h"
@@ -453,11 +452,9 @@ int amdgpu_ib_ring_tests(struct amdgpu_device *adev)
  */
 #if defined(CONFIG_DEBUG_FS)
 
-static int amdgpu_debugfs_sa_info(struct seq_file *m, void *data)
+static int amdgpu_debugfs_sa_info_show(struct seq_file *m, void *unused)
 {
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
 
 	seq_printf(m, "--------------------- DELAYED --------------------- \n");
 	amdgpu_sa_bo_dump_debug_info(&adev->ib_pools[AMDGPU_IB_POOL_DELAYED],
@@ -471,18 +468,18 @@ static int amdgpu_debugfs_sa_info(struct seq_file *m, void *data)
 	return 0;
 }
 
-static const struct drm_info_list amdgpu_debugfs_sa_list[] = {
+DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_sa_info);
-	{"amdgpu_sa_info", &amdgpu_debugfs_sa_info, 0, NULL},
-};
 
 #endif
 
-int amdgpu_debugfs_sa_init(struct amdgpu_device *adev)
+void amdgpu_debugfs_sa_init(struct amdgpu_device *adev)
 {
 #if defined(CONFIG_DEBUG_FS)
-	return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_sa_list,
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
-					ARRAY_SIZE(amdgpu_debugfs_sa_list));
+	struct dentry *root = minor->debugfs_root;
-#else
+
-	return 0;
+	debugfs_create_file("amdgpu_sa_info", 0444, root, adev,
+			    &amdgpu_debugfs_sa_info_fops);
+
 #endif
 }

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

@@ -99,6 +99,8 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
 		ih->rptr_addr = adev->wb.gpu_addr + rptr_offs * 4;
 		ih->rptr_cpu = &adev->wb.wb[rptr_offs];
 	}
+
+	init_waitqueue_head(&ih->wait_process);
 	return 0;
 }
 
@@ -160,6 +162,52 @@ void amdgpu_ih_ring_write(struct amdgpu_ih_ring *ih, const uint32_t *iv,
 	}
 }
 
+/* Waiter helper that checks current rptr matches or passes checkpoint wptr */
+static bool amdgpu_ih_has_checkpoint_processed(struct amdgpu_device *adev,
+					struct amdgpu_ih_ring *ih,
+					uint32_t checkpoint_wptr,
+					uint32_t *prev_rptr)
+{
+	uint32_t cur_rptr = ih->rptr | (*prev_rptr & ~ih->ptr_mask);
+
+	/* rptr has wrapped. */
+	if (cur_rptr < *prev_rptr)
+		cur_rptr += ih->ptr_mask + 1;
+	*prev_rptr = cur_rptr;
+
+	return cur_rptr >= checkpoint_wptr;
+}
+
+/**
+ * amdgpu_ih_wait_on_checkpoint_process - wait to process IVs up to checkpoint
+ *
+ * @adev: amdgpu_device pointer
+ * @ih: ih ring to process
+ *
+ * Used to ensure ring has processed IVs up to the checkpoint write pointer.
+ */
+int amdgpu_ih_wait_on_checkpoint_process(struct amdgpu_device *adev,
+					struct amdgpu_ih_ring *ih)
+{
+	uint32_t checkpoint_wptr, rptr;
+
+	if (!ih->enabled || adev->shutdown)
+		return -ENODEV;
+
+	checkpoint_wptr = amdgpu_ih_get_wptr(adev, ih);
+	/* Order wptr with rptr. */
+	rmb();
+	rptr = READ_ONCE(ih->rptr);
+
+	/* wptr has wrapped. */
+	if (rptr > checkpoint_wptr)
+		checkpoint_wptr += ih->ptr_mask + 1;
+
+	return wait_event_interruptible(ih->wait_process,
+				amdgpu_ih_has_checkpoint_processed(adev, ih,
+						checkpoint_wptr, &rptr));
+}
+
 /**
  * amdgpu_ih_process - interrupt handler
  *
@@ -180,10 +228,6 @@ int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih)
 	wptr = amdgpu_ih_get_wptr(adev, ih);
 
 restart_ih:
-	/* is somebody else already processing irqs? */
-	if (atomic_xchg(&ih->lock, 1))
-		return IRQ_NONE;
-
 	DRM_DEBUG("%s: rptr %d, wptr %d\n", __func__, ih->rptr, wptr);
 
 	/* Order reading of wptr vs. reading of IH ring data */
@@ -195,7 +239,7 @@ restart_ih:
 	}
 
 	amdgpu_ih_set_rptr(adev, ih);
-	atomic_set(&ih->lock, 0);
+	wake_up_all(&ih->wait_process);
 
 	/* make sure wptr hasn't changed while processing */
 	wptr = amdgpu_ih_get_wptr(adev, ih);

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

@@ -64,8 +64,10 @@ struct amdgpu_ih_ring {
 
 	bool                    enabled;
 	unsigned		rptr;
-	atomic_t		lock;
 	struct amdgpu_ih_regs	ih_regs;
+
+	/* For waiting on IH processing at checkpoint. */
+	wait_queue_head_t wait_process;
 };
 
 /* provided by the ih block */
@@ -87,6 +89,8 @@ int amdgpu_ih_ring_init(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih,
 void amdgpu_ih_ring_fini(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
 void amdgpu_ih_ring_write(struct amdgpu_ih_ring *ih, const uint32_t *iv,
 			  unsigned int num_dw);
+int amdgpu_ih_wait_on_checkpoint_process(struct amdgpu_device *adev,
+					struct amdgpu_ih_ring *ih);
 int amdgpu_ih_process(struct amdgpu_device *adev, struct amdgpu_ih_ring *ih);
 void amdgpu_ih_decode_iv_helper(struct amdgpu_device *adev,
 				struct amdgpu_ih_ring *ih,

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

@@ -65,6 +65,41 @@
 
 #define AMDGPU_WAIT_IDLE_TIMEOUT 200
 
+const char *soc15_ih_clientid_name[] = {
+	"IH",
+	"SDMA2 or ACP",
+	"ATHUB",
+	"BIF",
+	"SDMA3 or DCE",
+	"SDMA4 or ISP",
+	"VMC1 or PCIE0",
+	"RLC",
+	"SDMA0",
+	"SDMA1",
+	"SE0SH",
+	"SE1SH",
+	"SE2SH",
+	"SE3SH",
+	"VCN1 or UVD1",
+	"THM",
+	"VCN or UVD",
+	"SDMA5 or VCE0",
+	"VMC",
+	"SDMA6 or XDMA",
+	"GRBM_CP",
+	"ATS",
+	"ROM_SMUIO",
+	"DF",
+	"SDMA7 or VCE1",
+	"PWR",
+	"reserved",
+	"UTCL2",
+	"EA",
+	"UTCL2LOG",
+	"MP0",
+	"MP1"
+};
+
 /**
  * amdgpu_hotplug_work_func - work handler for display hotplug event
  *
@@ -164,13 +199,13 @@ irqreturn_t amdgpu_irq_handler(int irq, void *arg)
 	 * ack the interrupt if it is there
 	 */
 	if (amdgpu_ras_is_supported(adev, AMDGPU_RAS_BLOCK__PCIE_BIF)) {
-		if (adev->nbio.funcs &&
+		if (adev->nbio.ras_funcs &&
-		    adev->nbio.funcs->handle_ras_controller_intr_no_bifring)
+		    adev->nbio.ras_funcs->handle_ras_controller_intr_no_bifring)
-			adev->nbio.funcs->handle_ras_controller_intr_no_bifring(adev);
+			adev->nbio.ras_funcs->handle_ras_controller_intr_no_bifring(adev);
 
-		if (adev->nbio.funcs &&
+		if (adev->nbio.ras_funcs &&
-		    adev->nbio.funcs->handle_ras_err_event_athub_intr_no_bifring)
+		    adev->nbio.ras_funcs->handle_ras_err_event_athub_intr_no_bifring)
-			adev->nbio.funcs->handle_ras_err_event_athub_intr_no_bifring(adev);
+			adev->nbio.ras_funcs->handle_ras_err_event_athub_intr_no_bifring(adev);
 	}
 
 	return ret;
@@ -347,11 +382,6 @@ void amdgpu_irq_fini(struct amdgpu_device *adev)
 
 			kfree(src->enabled_types);
 			src->enabled_types = NULL;
-			if (src->data) {
-				kfree(src->data);
-				kfree(src);
-				adev->irq.client[i].sources[j] = NULL;
-			}
 		}
 		kfree(adev->irq.client[i].sources);
 		adev->irq.client[i].sources = NULL;
@@ -535,7 +565,7 @@ void amdgpu_irq_gpu_reset_resume_helper(struct amdgpu_device *adev)
 		for (j = 0; j < AMDGPU_MAX_IRQ_SRC_ID; ++j) {
 			struct amdgpu_irq_src *src = adev->irq.client[i].sources[j];
 
-			if (!src)
+			if (!src || !src->funcs || !src->funcs->set)
 				continue;
 			for (k = 0; k < src->num_types; k++)
 				amdgpu_irq_update(adev, src, k);

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

@@ -62,7 +62,6 @@ struct amdgpu_irq_src {
 	unsigned				num_types;
 	atomic_t				*enabled_types;
 	const struct amdgpu_irq_src_funcs	*funcs;
-	void *data;
 };
 
 struct amdgpu_irq_client {

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

@@ -28,7 +28,7 @@
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
 
-static void amdgpu_job_timedout(struct drm_sched_job *s_job)
+static enum drm_gpu_sched_stat amdgpu_job_timedout(struct drm_sched_job *s_job)
 {
 	struct amdgpu_ring *ring = to_amdgpu_ring(s_job->sched);
 	struct amdgpu_job *job = to_amdgpu_job(s_job);
@@ -41,7 +41,7 @@ static void amdgpu_job_timedout(struct drm_sched_job *s_job)
 	    amdgpu_ring_soft_recovery(ring, job->vmid, s_job->s_fence->parent)) {
 		DRM_ERROR("ring %s timeout, but soft recovered\n",
 			  s_job->sched->name);
-		return;
+		return DRM_GPU_SCHED_STAT_NOMINAL;
 	}
 
 	amdgpu_vm_get_task_info(ring->adev, job->pasid, &ti);
@@ -53,10 +53,12 @@ static void amdgpu_job_timedout(struct drm_sched_job *s_job)
 
 	if (amdgpu_device_should_recover_gpu(ring->adev)) {
 		amdgpu_device_gpu_recover(ring->adev, job);
+		return DRM_GPU_SCHED_STAT_NOMINAL;
 	} else {
 		drm_sched_suspend_timeout(&ring->sched);
 		if (amdgpu_sriov_vf(adev))
 			adev->virt.tdr_debug = true;
+		return DRM_GPU_SCHED_STAT_NOMINAL;
 	}
 }
 

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

@@ -27,7 +27,6 @@
  */
 
 #include "amdgpu.h"
-#include <drm/drm_debugfs.h>
 #include <drm/amdgpu_drm.h>
 #include "amdgpu_uvd.h"
 #include "amdgpu_vce.h"
@@ -160,7 +159,7 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags)
 		goto out;
 	}
 
-	if (amdgpu_device_supports_atpx(dev) &&
+	if (amdgpu_device_supports_px(dev) &&
 	    (amdgpu_runtime_pm != 0)) { /* enable runpm by default for atpx */
 		adev->runpm = true;
 		dev_info(adev->dev, "Using ATPX for runtime pm\n");
@@ -201,9 +200,13 @@ int amdgpu_driver_load_kms(struct amdgpu_device *adev, unsigned long flags)
 
 	if (adev->runpm) {
 		/* only need to skip on ATPX */
-		if (amdgpu_device_supports_atpx(dev) &&
+		if (amdgpu_device_supports_px(dev))
-		    !amdgpu_is_atpx_hybrid())
 			dev_pm_set_driver_flags(dev->dev, DPM_FLAG_NO_DIRECT_COMPLETE);
+		/* we want direct complete for BOCO */
+		if (amdgpu_device_supports_boco(dev))
+			dev_pm_set_driver_flags(dev->dev, DPM_FLAG_SMART_PREPARE |
+						DPM_FLAG_SMART_SUSPEND |
+						DPM_FLAG_MAY_SKIP_RESUME);
 		pm_runtime_use_autosuspend(dev->dev);
 		pm_runtime_set_autosuspend_delay(dev->dev, 5000);
 		pm_runtime_allow(dev->dev);
@@ -287,22 +290,30 @@ static int amdgpu_firmware_info(struct drm_amdgpu_info_firmware *fw_info,
 		break;
 	case AMDGPU_INFO_FW_TA:
 		switch (query_fw->index) {
-		case 0:
+		case TA_FW_TYPE_PSP_XGMI:
 			fw_info->ver = adev->psp.ta_fw_version;
 			fw_info->feature = adev->psp.ta_xgmi_ucode_version;
 			break;
-		case 1:
+		case TA_FW_TYPE_PSP_RAS:
 			fw_info->ver = adev->psp.ta_fw_version;
 			fw_info->feature = adev->psp.ta_ras_ucode_version;
 			break;
-		case 2:
+		case TA_FW_TYPE_PSP_HDCP:
 			fw_info->ver = adev->psp.ta_fw_version;
 			fw_info->feature = adev->psp.ta_hdcp_ucode_version;
 			break;
-		case 3:
+		case TA_FW_TYPE_PSP_DTM:
 			fw_info->ver = adev->psp.ta_fw_version;
 			fw_info->feature = adev->psp.ta_dtm_ucode_version;
 			break;
+		case TA_FW_TYPE_PSP_RAP:
+			fw_info->ver = adev->psp.ta_fw_version;
+			fw_info->feature = adev->psp.ta_rap_ucode_version;
+			break;
+		case TA_FW_TYPE_PSP_SECUREDISPLAY:
+			fw_info->ver = adev->psp.ta_fw_version;
+			fw_info->feature = adev->psp.ta_securedisplay_ucode_version;
+			break;
 		default:
 			return -EINVAL;
 		}
@@ -981,6 +992,63 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 				min_t(u64, size, sizeof(ras_mask))) ?
 			-EFAULT : 0;
 	}
+	case AMDGPU_INFO_VIDEO_CAPS: {
+		const struct amdgpu_video_codecs *codecs;
+		struct drm_amdgpu_info_video_caps *caps;
+		int r;
+
+		switch (info->video_cap.type) {
+		case AMDGPU_INFO_VIDEO_CAPS_DECODE:
+			r = amdgpu_asic_query_video_codecs(adev, false, &codecs);
+			if (r)
+				return -EINVAL;
+			break;
+		case AMDGPU_INFO_VIDEO_CAPS_ENCODE:
+			r = amdgpu_asic_query_video_codecs(adev, true, &codecs);
+			if (r)
+				return -EINVAL;
+			break;
+		default:
+			DRM_DEBUG_KMS("Invalid request %d\n",
+				      info->video_cap.type);
+			return -EINVAL;
+		}
+
+		caps = kzalloc(sizeof(*caps), GFP_KERNEL);
+		if (!caps)
+			return -ENOMEM;
+
+		for (i = 0; i < codecs->codec_count; i++) {
+			int idx = codecs->codec_array[i].codec_type;
+
+			switch (idx) {
+			case AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2:
+			case AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4:
+			case AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1:
+			case AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC:
+			case AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC:
+			case AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG:
+			case AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9:
+			case AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1:
+				caps->codec_info[idx].valid = 1;
+				caps->codec_info[idx].max_width =
+					codecs->codec_array[i].max_width;
+				caps->codec_info[idx].max_height =
+					codecs->codec_array[i].max_height;
+				caps->codec_info[idx].max_pixels_per_frame =
+					codecs->codec_array[i].max_pixels_per_frame;
+				caps->codec_info[idx].max_level =
+					codecs->codec_array[i].max_level;
+				break;
+			default:
+				break;
+			}
+		}
+		r = copy_to_user(out, caps,
+				 min((size_t)size, sizeof(*caps))) ? -EFAULT : 0;
+		kfree(caps);
+		return r;
+	}
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->query);
 		return -EINVAL;
@@ -1262,16 +1330,25 @@ void amdgpu_disable_vblank_kms(struct drm_crtc *crtc)
  */
 #if defined(CONFIG_DEBUG_FS)
 
-static int amdgpu_debugfs_firmware_info(struct seq_file *m, void *data)
+static int amdgpu_debugfs_firmware_info_show(struct seq_file *m, void *unused)
 {
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
 	struct drm_amdgpu_info_firmware fw_info;
 	struct drm_amdgpu_query_fw query_fw;
 	struct atom_context *ctx = adev->mode_info.atom_context;
 	int ret, i;
 
+	static const char *ta_fw_name[TA_FW_TYPE_MAX_INDEX] = {
+#define TA_FW_NAME(type) [TA_FW_TYPE_PSP_##type] = #type
+		TA_FW_NAME(XGMI),
+		TA_FW_NAME(RAS),
+		TA_FW_NAME(HDCP),
+		TA_FW_NAME(DTM),
+		TA_FW_NAME(RAP),
+		TA_FW_NAME(SECUREDISPLAY),
+#undef TA_FW_NAME
+	};
+
 	/* VCE */
 	query_fw.fw_type = AMDGPU_INFO_FW_VCE;
 	ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
@@ -1389,31 +1466,14 @@ static int amdgpu_debugfs_firmware_info(struct seq_file *m, void *data)
 		   fw_info.feature, fw_info.ver);
 
 	query_fw.fw_type = AMDGPU_INFO_FW_TA;
-	for (i = 0; i < 4; i++) {
+	for (i = TA_FW_TYPE_PSP_XGMI; i < TA_FW_TYPE_MAX_INDEX; i++) {
 		query_fw.index = i;
 		ret = amdgpu_firmware_info(&fw_info, &query_fw, adev);
 		if (ret)
 			continue;
-		switch (query_fw.index) {
+
-		case 0:
+		seq_printf(m, "TA %s feature version: 0x%08x, firmware version: 0x%08x\n",
-			seq_printf(m, "TA %s feature version: 0x%08x, firmware version: 0x%08x\n",
+			   ta_fw_name[i], fw_info.feature, fw_info.ver);
-					"RAS", fw_info.feature, fw_info.ver);
-			break;
-		case 1:
-			seq_printf(m, "TA %s feature version: 0x%08x, firmware version: 0x%08x\n",
-					"XGMI", fw_info.feature, fw_info.ver);
-			break;
-		case 2:
-			seq_printf(m, "TA %s feature version: 0x%08x, firmware version: 0x%08x\n",
-					"HDCP", fw_info.feature, fw_info.ver);
-			break;
-		case 3:
-			seq_printf(m, "TA %s feature version: 0x%08x, firmware version: 0x%08x\n",
-					"DTM", fw_info.feature, fw_info.ver);
-			break;
-		default:
-			return -EINVAL;
-		}
 	}
 
 	/* SMC */
@@ -1472,17 +1532,18 @@ static int amdgpu_debugfs_firmware_info(struct seq_file *m, void *data)
 	return 0;
 }
 
-static const struct drm_info_list amdgpu_firmware_info_list[] = {
+DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_firmware_info);
-	{"amdgpu_firmware_info", amdgpu_debugfs_firmware_info, 0, NULL},
+
-};
 #endif
 
-int amdgpu_debugfs_firmware_init(struct amdgpu_device *adev)
+void amdgpu_debugfs_firmware_init(struct amdgpu_device *adev)
 {
 #if defined(CONFIG_DEBUG_FS)
-	return amdgpu_debugfs_add_files(adev, amdgpu_firmware_info_list,
+	struct drm_minor *minor = adev_to_drm(adev)->primary;
-					ARRAY_SIZE(amdgpu_firmware_info_list));
+	struct dentry *root = minor->debugfs_root;
-#else
+
-	return 0;
+	debugfs_create_file("amdgpu_firmware_info", 0444, root,
+			    adev, &amdgpu_debugfs_firmware_info_fops);
+
 #endif
 }

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

@@ -21,12 +21,16 @@
 #ifndef __AMDGPU_MMHUB_H__
 #define __AMDGPU_MMHUB_H__
 
-struct amdgpu_mmhub_funcs {
+struct amdgpu_mmhub_ras_funcs {
-	void (*ras_init)(struct amdgpu_device *adev);
 	int (*ras_late_init)(struct amdgpu_device *adev);
+	void (*ras_fini)(struct amdgpu_device *adev);
 	void (*query_ras_error_count)(struct amdgpu_device *adev,
-					void *ras_error_status);
+				      void *ras_error_status);
+	void (*query_ras_error_status)(struct amdgpu_device *adev);
 	void (*reset_ras_error_count)(struct amdgpu_device *adev);
+};
+
+struct amdgpu_mmhub_funcs {
 	u64 (*get_fb_location)(struct amdgpu_device *adev);
 	void (*init)(struct amdgpu_device *adev);
 	int (*gart_enable)(struct amdgpu_device *adev);
@@ -40,12 +44,12 @@ struct amdgpu_mmhub_funcs {
 				uint64_t page_table_base);
 	void (*update_power_gating)(struct amdgpu_device *adev,
                                 bool enable);
-	void (*query_ras_error_status)(struct amdgpu_device *adev);
 };
 
 struct amdgpu_mmhub {
 	struct ras_common_if *ras_if;
 	const struct amdgpu_mmhub_funcs *funcs;
+	const struct amdgpu_mmhub_ras_funcs *ras_funcs;
 };
 
 int amdgpu_mmhub_ras_late_init(struct amdgpu_device *adev);

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

@@ -602,6 +602,14 @@ int amdgpu_display_get_crtc_scanoutpos(struct drm_device *dev,
 			int *hpos, ktime_t *stime, ktime_t *etime,
 			const struct drm_display_mode *mode);
 
+int amdgpu_display_gem_fb_init(struct drm_device *dev,
+			       struct amdgpu_framebuffer *rfb,
+			       const struct drm_mode_fb_cmd2 *mode_cmd,
+			       struct drm_gem_object *obj);
+int amdgpu_display_gem_fb_verify_and_init(
+	struct drm_device *dev, struct amdgpu_framebuffer *rfb,
+	struct drm_file *file_priv, const struct drm_mode_fb_cmd2 *mode_cmd,
+	struct drm_gem_object *obj);
 int amdgpu_display_framebuffer_init(struct drm_device *dev,
 				    struct amdgpu_framebuffer *rfb,
 				    const struct drm_mode_fb_cmd2 *mode_cmd,

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

@@ -47,6 +47,17 @@ struct nbio_hdp_flush_reg {
 	u32 ref_and_mask_sdma7;
 };
 
+struct amdgpu_nbio_ras_funcs {
+	void (*handle_ras_controller_intr_no_bifring)(struct amdgpu_device *adev);
+	void (*handle_ras_err_event_athub_intr_no_bifring)(struct amdgpu_device *adev);
+	int (*init_ras_controller_interrupt)(struct amdgpu_device *adev);
+	int (*init_ras_err_event_athub_interrupt)(struct amdgpu_device *adev);
+	void (*query_ras_error_count)(struct amdgpu_device *adev,
+				      void *ras_error_status);
+	int (*ras_late_init)(struct amdgpu_device *adev);
+	void (*ras_fini)(struct amdgpu_device *adev);
+};
+
 struct amdgpu_nbio_funcs {
 	const struct nbio_hdp_flush_reg *hdp_flush_reg;
 	u32 (*get_hdp_flush_req_offset)(struct amdgpu_device *adev);
@@ -79,13 +90,6 @@ struct amdgpu_nbio_funcs {
 	void (*ih_control)(struct amdgpu_device *adev);
 	void (*init_registers)(struct amdgpu_device *adev);
 	void (*remap_hdp_registers)(struct amdgpu_device *adev);
-	void (*handle_ras_controller_intr_no_bifring)(struct amdgpu_device *adev);
-	void (*handle_ras_err_event_athub_intr_no_bifring)(struct amdgpu_device *adev);
-	int (*init_ras_controller_interrupt)(struct amdgpu_device *adev);
-	int (*init_ras_err_event_athub_interrupt)(struct amdgpu_device *adev);
-	void (*query_ras_error_count)(struct amdgpu_device *adev,
-					void *ras_error_status);
-	int (*ras_late_init)(struct amdgpu_device *adev);
 	void (*enable_aspm)(struct amdgpu_device *adev,
 			    bool enable);
 	void (*program_aspm)(struct amdgpu_device *adev);
@@ -97,6 +101,7 @@ struct amdgpu_nbio {
 	struct amdgpu_irq_src ras_err_event_athub_irq;
 	struct ras_common_if *ras_if;
 	const struct amdgpu_nbio_funcs *funcs;
+	const struct amdgpu_nbio_ras_funcs *ras_funcs;
 };
 
 int amdgpu_nbio_ras_late_init(struct amdgpu_device *adev);

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

@@ -77,6 +77,7 @@ static void amdgpu_bo_destroy(struct ttm_buffer_object *tbo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev);
 	struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo);
+	struct amdgpu_bo_user *ubo;
 
 	if (bo->tbo.pin_count > 0)
 		amdgpu_bo_subtract_pin_size(bo);
@@ -94,7 +95,11 @@ static void amdgpu_bo_destroy(struct ttm_buffer_object *tbo)
 	}
 	amdgpu_bo_unref(&bo->parent);
 
-	kfree(bo->metadata);
+	if (bo->tbo.type == ttm_bo_type_device) {
+		ubo = to_amdgpu_bo_user(bo);
+		kfree(ubo->metadata);
+	}
+
 	kfree(bo);
 }
 
@@ -248,6 +253,7 @@ int amdgpu_bo_create_reserved(struct amdgpu_device *adev,
 	bp.flags |= AMDGPU_GEM_CREATE_VRAM_CONTIGUOUS;
 	bp.type = ttm_bo_type_kernel;
 	bp.resv = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
 
 	if (!*bo_ptr) {
 		r = amdgpu_bo_create(adev, &bp, bo_ptr);
@@ -523,7 +529,6 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
 	};
 	struct amdgpu_bo *bo;
 	unsigned long page_align, size = bp->size;
-	size_t acc_size;
 	int r;
 
 	/* Note that GDS/GWS/OA allocates 1 page per byte/resource. */
@@ -544,12 +549,10 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
 	if (!amdgpu_bo_validate_size(adev, size, bp->domain))
 		return -ENOMEM;
 
+	BUG_ON(bp->bo_ptr_size < sizeof(struct amdgpu_bo));
+
 	*bo_ptr = NULL;
-
+	bo = kzalloc(bp->bo_ptr_size, GFP_KERNEL);
-	acc_size = ttm_bo_dma_acc_size(&adev->mman.bdev, size,
-				       sizeof(struct amdgpu_bo));
-
-	bo = kzalloc(sizeof(struct amdgpu_bo), GFP_KERNEL);
 	if (bo == NULL)
 		return -ENOMEM;
 	drm_gem_private_object_init(adev_to_drm(adev), &bo->tbo.base, size);
@@ -577,8 +580,8 @@ static int amdgpu_bo_do_create(struct amdgpu_device *adev,
 		bo->tbo.priority = 1;
 
 	r = ttm_bo_init_reserved(&adev->mman.bdev, &bo->tbo, size, bp->type,
-				 &bo->placement, page_align, &ctx, acc_size,
+				 &bo->placement, page_align, &ctx,  NULL,
-				 NULL, bp->resv, &amdgpu_bo_destroy);
+				 bp->resv, &amdgpu_bo_destroy);
 	if (unlikely(r != 0))
 		return r;
 
@@ -639,6 +642,7 @@ static int amdgpu_bo_create_shadow(struct amdgpu_device *adev,
 		AMDGPU_GEM_CREATE_SHADOW;
 	bp.type = ttm_bo_type_kernel;
 	bp.resv = bo->tbo.base.resv;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
 
 	r = amdgpu_bo_do_create(adev, &bp, &bo->shadow);
 	if (!r) {
@@ -673,6 +677,7 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 	int r;
 
 	bp->flags = bp->flags & ~AMDGPU_GEM_CREATE_SHADOW;
+
 	r = amdgpu_bo_do_create(adev, bp, bo_ptr);
 	if (r)
 		return r;
@@ -694,6 +699,34 @@ int amdgpu_bo_create(struct amdgpu_device *adev,
 	return r;
 }
 
+/**
+ * amdgpu_bo_create_user - create an &amdgpu_bo_user buffer object
+ * @adev: amdgpu device object
+ * @bp: parameters to be used for the buffer object
+ * @ubo_ptr: pointer to the buffer object pointer
+ *
+ * Create a BO to be used by user application;
+ *
+ * Returns:
+ * 0 for success or a negative error code on failure.
+ */
+
+int amdgpu_bo_create_user(struct amdgpu_device *adev,
+			  struct amdgpu_bo_param *bp,
+			  struct amdgpu_bo_user **ubo_ptr)
+{
+	struct amdgpu_bo *bo_ptr;
+	int r;
+
+	bp->flags = bp->flags & ~AMDGPU_GEM_CREATE_SHADOW;
+	bp->bo_ptr_size = sizeof(struct amdgpu_bo_user);
+	r = amdgpu_bo_do_create(adev, bp, &bo_ptr);
+	if (r)
+		return r;
+
+	*ubo_ptr = to_amdgpu_bo_user(bo_ptr);
+	return r;
+}
 /**
  * amdgpu_bo_validate - validate an &amdgpu_bo buffer object
  * @bo: pointer to the buffer object
@@ -1062,13 +1095,17 @@ static const char *amdgpu_vram_names[] = {
  */
 int amdgpu_bo_init(struct amdgpu_device *adev)
 {
-	/* reserve PAT memory space to WC for VRAM */
+	/* On A+A platform, VRAM can be mapped as WB */
-	arch_io_reserve_memtype_wc(adev->gmc.aper_base,
+	if (!adev->gmc.xgmi.connected_to_cpu) {
-				   adev->gmc.aper_size);
+		/* reserve PAT memory space to WC for VRAM */
+		arch_io_reserve_memtype_wc(adev->gmc.aper_base,
+				adev->gmc.aper_size);
+
+		/* Add an MTRR for the VRAM */
+		adev->gmc.vram_mtrr = arch_phys_wc_add(adev->gmc.aper_base,
+				adev->gmc.aper_size);
+	}
 
-	/* Add an MTRR for the VRAM */
-	adev->gmc.vram_mtrr = arch_phys_wc_add(adev->gmc.aper_base,
-					      adev->gmc.aper_size);
 	DRM_INFO("Detected VRAM RAM=%lluM, BAR=%lluM\n",
 		 adev->gmc.mc_vram_size >> 20,
 		 (unsigned long long)adev->gmc.aper_size >> 20);
@@ -1086,27 +1123,10 @@ int amdgpu_bo_init(struct amdgpu_device *adev)
 void amdgpu_bo_fini(struct amdgpu_device *adev)
 {
 	amdgpu_ttm_fini(adev);
-	arch_phys_wc_del(adev->gmc.vram_mtrr);
+	if (!adev->gmc.xgmi.connected_to_cpu) {
-	arch_io_free_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size);
+		arch_phys_wc_del(adev->gmc.vram_mtrr);
-}
+		arch_io_free_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size);
-
+	}
-/**
- * amdgpu_bo_fbdev_mmap - mmap fbdev memory
- * @bo: &amdgpu_bo buffer object
- * @vma: vma as input from the fbdev mmap method
- *
- * Calls ttm_fbdev_mmap() to mmap fbdev memory if it is backed by a bo.
- *
- * Returns:
- * 0 for success or a negative error code on failure.
- */
-int amdgpu_bo_fbdev_mmap(struct amdgpu_bo *bo,
-			     struct vm_area_struct *vma)
-{
-	if (vma->vm_pgoff != 0)
-		return -EACCES;
-
-	return ttm_bo_mmap_obj(vma, &bo->tbo);
 }
 
 /**
@@ -1123,12 +1143,15 @@ int amdgpu_bo_fbdev_mmap(struct amdgpu_bo *bo,
 int amdgpu_bo_set_tiling_flags(struct amdgpu_bo *bo, u64 tiling_flags)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
+	struct amdgpu_bo_user *ubo;
 
+	BUG_ON(bo->tbo.type == ttm_bo_type_kernel);
 	if (adev->family <= AMDGPU_FAMILY_CZ &&
 	    AMDGPU_TILING_GET(tiling_flags, TILE_SPLIT) > 6)
 		return -EINVAL;
 
-	bo->tiling_flags = tiling_flags;
+	ubo = to_amdgpu_bo_user(bo);
+	ubo->tiling_flags = tiling_flags;
 	return 0;
 }
 
@@ -1142,10 +1165,14 @@ int amdgpu_bo_set_tiling_flags(struct amdgpu_bo *bo, u64 tiling_flags)
  */
 void amdgpu_bo_get_tiling_flags(struct amdgpu_bo *bo, u64 *tiling_flags)
 {
+	struct amdgpu_bo_user *ubo;
+
+	BUG_ON(bo->tbo.type == ttm_bo_type_kernel);
 	dma_resv_assert_held(bo->tbo.base.resv);
+	ubo = to_amdgpu_bo_user(bo);
 
 	if (tiling_flags)
-		*tiling_flags = bo->tiling_flags;
+		*tiling_flags = ubo->tiling_flags;
 }
 
 /**
@@ -1164,13 +1191,16 @@ void amdgpu_bo_get_tiling_flags(struct amdgpu_bo *bo, u64 *tiling_flags)
 int amdgpu_bo_set_metadata (struct amdgpu_bo *bo, void *metadata,
 			    uint32_t metadata_size, uint64_t flags)
 {
+	struct amdgpu_bo_user *ubo;
 	void *buffer;
 
+	BUG_ON(bo->tbo.type == ttm_bo_type_kernel);
+	ubo = to_amdgpu_bo_user(bo);
 	if (!metadata_size) {
-		if (bo->metadata_size) {
+		if (ubo->metadata_size) {
-			kfree(bo->metadata);
+			kfree(ubo->metadata);
-			bo->metadata = NULL;
+			ubo->metadata = NULL;
-			bo->metadata_size = 0;
+			ubo->metadata_size = 0;
 		}
 		return 0;
 	}
@@ -1182,10 +1212,10 @@ int amdgpu_bo_set_metadata (struct amdgpu_bo *bo, void *metadata,
 	if (buffer == NULL)
 		return -ENOMEM;
 
-	kfree(bo->metadata);
+	kfree(ubo->metadata);
-	bo->metadata_flags = flags;
+	ubo->metadata_flags = flags;
-	bo->metadata = buffer;
+	ubo->metadata = buffer;
-	bo->metadata_size = metadata_size;
+	ubo->metadata_size = metadata_size;
 
 	return 0;
 }
@@ -1209,21 +1239,25 @@ int amdgpu_bo_get_metadata(struct amdgpu_bo *bo, void *buffer,
 			   size_t buffer_size, uint32_t *metadata_size,
 			   uint64_t *flags)
 {
+	struct amdgpu_bo_user *ubo;
+
 	if (!buffer && !metadata_size)
 		return -EINVAL;
 
+	BUG_ON(bo->tbo.type == ttm_bo_type_kernel);
+	ubo = to_amdgpu_bo_user(bo);
 	if (buffer) {
-		if (buffer_size < bo->metadata_size)
+		if (buffer_size < ubo->metadata_size)
 			return -EINVAL;
 
-		if (bo->metadata_size)
+		if (ubo->metadata_size)
-			memcpy(buffer, bo->metadata, bo->metadata_size);
+			memcpy(buffer, ubo->metadata, ubo->metadata_size);
 	}
 
 	if (metadata_size)
-		*metadata_size = bo->metadata_size;
+		*metadata_size = ubo->metadata_size;
 	if (flags)
-		*flags = bo->metadata_flags;
+		*flags = ubo->metadata_flags;
 
 	return 0;
 }

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

@@ -37,9 +37,12 @@
 #define AMDGPU_BO_INVALID_OFFSET	LONG_MAX
 #define AMDGPU_BO_MAX_PLACEMENTS	3
 
+#define to_amdgpu_bo_user(abo) container_of((abo), struct amdgpu_bo_user, bo)
+
 struct amdgpu_bo_param {
 	unsigned long			size;
 	int				byte_align;
+	u32				bo_ptr_size;
 	u32				domain;
 	u32				preferred_domain;
 	u64				flags;
@@ -89,10 +92,6 @@ struct amdgpu_bo {
 	struct ttm_buffer_object	tbo;
 	struct ttm_bo_kmap_obj		kmap;
 	u64				flags;
-	u64				tiling_flags;
-	u64				metadata_flags;
-	void				*metadata;
-	u32				metadata_size;
 	unsigned			prime_shared_count;
 	/* per VM structure for page tables and with virtual addresses */
 	struct amdgpu_vm_bo_base	*vm_bo;
@@ -100,7 +99,6 @@ struct amdgpu_bo {
 	struct amdgpu_bo		*parent;
 	struct amdgpu_bo		*shadow;
 
-	struct amdgpu_mn		*mn;
 
 
 #ifdef CONFIG_MMU_NOTIFIER
@@ -112,6 +110,15 @@ struct amdgpu_bo {
 	struct kgd_mem                  *kfd_bo;
 };
 
+struct amdgpu_bo_user {
+	struct amdgpu_bo		bo;
+	u64				tiling_flags;
+	u64				metadata_flags;
+	void				*metadata;
+	u32				metadata_size;
+
+};
+
 static inline struct amdgpu_bo *ttm_to_amdgpu_bo(struct ttm_buffer_object *tbo)
 {
 	return container_of(tbo, struct amdgpu_bo, tbo);
@@ -255,6 +262,9 @@ int amdgpu_bo_create_kernel(struct amdgpu_device *adev,
 int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev,
 			       uint64_t offset, uint64_t size, uint32_t domain,
 			       struct amdgpu_bo **bo_ptr, void **cpu_addr);
+int amdgpu_bo_create_user(struct amdgpu_device *adev,
+			  struct amdgpu_bo_param *bp,
+			  struct amdgpu_bo_user **ubo_ptr);
 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);
@@ -269,8 +279,6 @@ void amdgpu_bo_unpin(struct amdgpu_bo *bo);
 int amdgpu_bo_evict_vram(struct amdgpu_device *adev);
 int amdgpu_bo_init(struct amdgpu_device *adev);
 void amdgpu_bo_fini(struct amdgpu_device *adev);
-int amdgpu_bo_fbdev_mmap(struct amdgpu_bo *bo,
-				struct vm_area_struct *vma);
 int amdgpu_bo_set_tiling_flags(struct amdgpu_bo *bo, u64 tiling_flags);
 void amdgpu_bo_get_tiling_flags(struct amdgpu_bo *bo, u64 *tiling_flags);
 int amdgpu_bo_set_metadata (struct amdgpu_bo *bo, void *metadata,
@@ -329,7 +337,7 @@ void amdgpu_sa_bo_dump_debug_info(struct amdgpu_sa_manager *sa_manager,
 					 struct seq_file *m);
 u64 amdgpu_bo_print_info(int id, struct amdgpu_bo *bo, struct seq_file *m);
 #endif
-int amdgpu_debugfs_sa_init(struct amdgpu_device *adev);
+void amdgpu_debugfs_sa_init(struct amdgpu_device *adev);
 
 bool amdgpu_bo_support_uswc(u64 bo_flags);
 

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

@@ -34,6 +34,7 @@
 #include "psp_v10_0.h"
 #include "psp_v11_0.h"
 #include "psp_v12_0.h"
+#include "psp_v13_0.h"
 
 #include "amdgpu_ras.h"
 #include "amdgpu_securedisplay.h"
@@ -56,7 +57,7 @@ static int psp_load_smu_fw(struct psp_context *psp);
  *   - Load XGMI/RAS/HDCP/DTM TA if any
  *
  * This new sequence is required for
- *   - Arcturus
+ *   - Arcturus and onwards
  *   - Navi12 and onwards
  */
 static void psp_check_pmfw_centralized_cstate_management(struct psp_context *psp)
@@ -71,7 +72,7 @@ static void psp_check_pmfw_centralized_cstate_management(struct psp_context *psp
 	if (adev->flags & AMD_IS_APU)
 		return;
 
-	if ((adev->asic_type == CHIP_ARCTURUS) ||
+	if ((adev->asic_type >= CHIP_ARCTURUS) ||
 	    (adev->asic_type >= CHIP_NAVI12))
 		psp->pmfw_centralized_cstate_management = true;
 }
@@ -109,6 +110,9 @@ static int psp_early_init(void *handle)
 	case CHIP_RENOIR:
 		psp_v12_0_set_psp_funcs(psp);
 		break;
+	case CHIP_ALDEBARAN:
+		psp_v13_0_set_psp_funcs(psp);
+		break;
 	default:
 		return -EINVAL;
 	}
@@ -383,7 +387,7 @@ static int psp_tmr_init(struct psp_context *psp)
 	 * Note: this memory need be reserved till the driver
 	 * uninitializes.
 	 */
-	tmr_size = PSP_TMR_SIZE;
+	tmr_size = PSP_TMR_SIZE(psp->adev);
 
 	/* For ASICs support RLC autoload, psp will parse the toc
 	 * and calculate the total size of TMR needed */
@@ -399,10 +403,20 @@ static int psp_tmr_init(struct psp_context *psp)
 	}
 
 	pptr = amdgpu_sriov_vf(psp->adev) ? &tmr_buf : NULL;
-	ret = amdgpu_bo_create_kernel(psp->adev, tmr_size, PSP_TMR_SIZE,
+	ret = amdgpu_bo_create_kernel(psp->adev, tmr_size, PSP_TMR_SIZE(psp->adev),
 				      AMDGPU_GEM_DOMAIN_VRAM,
 				      &psp->tmr_bo, &psp->tmr_mc_addr, pptr);
 
+	/* workaround the tmr_mc_addr:
+	 * PSP requires an address in FB aperture. Right now driver produce
+	 * tmr_mc_addr in the GART aperture. Convert it back to FB aperture
+	 * for PSP. Will revert it after we get a fix from PSP FW.
+	 */
+	if (psp->adev->asic_type == CHIP_ALDEBARAN) {
+		psp->tmr_mc_addr -= psp->adev->gmc.fb_start;
+		psp->tmr_mc_addr += psp->adev->gmc.fb_start_original;
+	}
+
 	return ret;
 }
 
@@ -542,6 +556,46 @@ int psp_get_fw_attestation_records_addr(struct psp_context *psp,
 	return ret;
 }
 
+static int psp_boot_config_set(struct amdgpu_device *adev)
+{
+	struct psp_context *psp = &adev->psp;
+	struct psp_gfx_cmd_resp *cmd = psp->cmd;
+
+	if (adev->asic_type != CHIP_SIENNA_CICHLID)
+		return 0;
+
+	memset(cmd, 0, sizeof(struct psp_gfx_cmd_resp));
+
+	cmd->cmd_id = GFX_CMD_ID_BOOT_CFG;
+	cmd->cmd.boot_cfg.sub_cmd = BOOTCFG_CMD_SET;
+	cmd->cmd.boot_cfg.boot_config = BOOT_CONFIG_GECC;
+	cmd->cmd.boot_cfg.boot_config_valid = BOOT_CONFIG_GECC;
+
+	return psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
+}
+
+static int psp_rl_load(struct amdgpu_device *adev)
+{
+	struct psp_context *psp = &adev->psp;
+	struct psp_gfx_cmd_resp *cmd = psp->cmd;
+
+	if (psp->rl_bin_size == 0)
+		return 0;
+
+	memset(psp->fw_pri_buf, 0, PSP_1_MEG);
+	memcpy(psp->fw_pri_buf, psp->rl_start_addr, psp->rl_bin_size);
+
+	memset(cmd, 0, sizeof(struct psp_gfx_cmd_resp));
+
+	cmd->cmd_id = GFX_CMD_ID_LOAD_IP_FW;
+	cmd->cmd.cmd_load_ip_fw.fw_phy_addr_lo = lower_32_bits(psp->fw_pri_mc_addr);
+	cmd->cmd.cmd_load_ip_fw.fw_phy_addr_hi = upper_32_bits(psp->fw_pri_mc_addr);
+	cmd->cmd.cmd_load_ip_fw.fw_size = psp->rl_bin_size;
+	cmd->cmd.cmd_load_ip_fw.fw_type = GFX_FW_TYPE_REG_LIST;
+
+	return psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
+}
+
 static void psp_prep_asd_load_cmd_buf(struct psp_gfx_cmd_resp *cmd,
 				uint64_t asd_mc, uint32_t size)
 {
@@ -755,8 +809,9 @@ static int psp_xgmi_unload(struct psp_context *psp)
 	struct psp_gfx_cmd_resp *cmd;
 	struct amdgpu_device *adev = psp->adev;
 
-	/* XGMI TA unload currently is not supported on Arcturus */
+	/* XGMI TA unload currently is not supported on Arcturus/Aldebaran A+A */
-	if (adev->asic_type == CHIP_ARCTURUS)
+	if (adev->asic_type == CHIP_ARCTURUS ||
+		(adev->asic_type == CHIP_ALDEBARAN && adev->gmc.xgmi.connected_to_cpu))
 		return 0;
 
 	/*
@@ -1561,6 +1616,7 @@ static int psp_rap_unload(struct psp_context *psp)
 static int psp_rap_initialize(struct psp_context *psp)
 {
 	int ret;
+	enum ta_rap_status status = TA_RAP_STATUS__SUCCESS;
 
 	/*
 	 * TODO: bypass the initialize in sriov for now
@@ -1584,8 +1640,8 @@ static int psp_rap_initialize(struct psp_context *psp)
 	if (ret)
 		return ret;
 
-	ret = psp_rap_invoke(psp, TA_CMD_RAP__INITIALIZE);
+	ret = psp_rap_invoke(psp, TA_CMD_RAP__INITIALIZE, &status);
-	if (ret != TA_RAP_STATUS__SUCCESS) {
+	if (ret || status != TA_RAP_STATUS__SUCCESS) {
 		psp_rap_unload(psp);
 
 		amdgpu_bo_free_kernel(&psp->rap_context.rap_shared_bo,
@@ -1594,8 +1650,10 @@ static int psp_rap_initialize(struct psp_context *psp)
 
 		psp->rap_context.rap_initialized = false;
 
-		dev_warn(psp->adev->dev, "RAP TA initialize fail.\n");
+		dev_warn(psp->adev->dev, "RAP TA initialize fail (%d) status %d.\n",
-		return -EINVAL;
+			 ret, status);
+
+		return ret;
 	}
 
 	return 0;
@@ -1620,13 +1678,13 @@ static int psp_rap_terminate(struct psp_context *psp)
 	return ret;
 }
 
-int psp_rap_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
+int psp_rap_invoke(struct psp_context *psp, uint32_t ta_cmd_id, enum ta_rap_status *status)
 {
 	struct ta_rap_shared_memory *rap_cmd;
-	int ret;
+	int ret = 0;
 
 	if (!psp->rap_context.rap_initialized)
-		return -EINVAL;
+		return 0;
 
 	if (ta_cmd_id != TA_CMD_RAP__INITIALIZE &&
 	    ta_cmd_id != TA_CMD_RAP__VALIDATE_L0)
@@ -1642,14 +1700,16 @@ int psp_rap_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 	rap_cmd->validation_method_id = METHOD_A;
 
 	ret = psp_ta_invoke(psp, rap_cmd->cmd_id, psp->rap_context.session_id);
-	if (ret) {
+	if (ret)
-		mutex_unlock(&psp->rap_context.mutex);
+		goto out_unlock;
-		return ret;
-	}
 
+	if (status)
+		*status = rap_cmd->rap_status;
+
+out_unlock:
 	mutex_unlock(&psp->rap_context.mutex);
 
-	return rap_cmd->rap_status;
+	return ret;
 }
 // RAP end
 
@@ -1870,6 +1930,11 @@ static int psp_hw_start(struct psp_context *psp)
 		return ret;
 	}
 
+	ret = psp_boot_config_set(adev);
+	if (ret) {
+		DRM_WARN("PSP set boot config@\n");
+	}
+
 	ret = psp_tmr_init(psp);
 	if (ret) {
 		DRM_ERROR("PSP tmr init failed!\n");
@@ -2104,9 +2169,13 @@ static int psp_load_smu_fw(struct psp_context *psp)
 	if (!ucode->fw || amdgpu_sriov_vf(psp->adev))
 		return 0;
 
-
+	if ((amdgpu_in_reset(adev) &&
-	if (amdgpu_in_reset(adev) && ras && ras->supported &&
+	     ras && ras->supported &&
-		adev->asic_type == CHIP_ARCTURUS) {
+	     (adev->asic_type == CHIP_ARCTURUS ||
+	      adev->asic_type == CHIP_VEGA20)) ||
+	     (adev->in_runpm &&
+	      adev->asic_type >= CHIP_NAVI10 &&
+	      adev->asic_type <= CHIP_NAVI12)) {
 		ret = amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_UNLOAD);
 		if (ret) {
 			DRM_WARN("Failed to set MP1 state prepare for reload\n");
@@ -2159,6 +2228,22 @@ static bool fw_load_skip_check(struct psp_context *psp,
 	return false;
 }
 
+int psp_load_fw_list(struct psp_context *psp,
+		     struct amdgpu_firmware_info **ucode_list, int ucode_count)
+{
+	int ret = 0, i;
+	struct amdgpu_firmware_info *ucode;
+
+	for (i = 0; i < ucode_count; ++i) {
+		ucode = ucode_list[i];
+		psp_print_fw_hdr(psp, ucode);
+		ret = psp_execute_np_fw_load(psp, ucode);
+		if (ret)
+			return ret;
+	}
+	return ret;
+}
+
 static int psp_np_fw_load(struct psp_context *psp)
 {
 	int i, ret;
@@ -2276,6 +2361,12 @@ skip_memalloc:
 		return ret;
 	}
 
+	ret = psp_rl_load(adev);
+	if (ret) {
+		DRM_ERROR("PSP load RL failed!\n");
+		return ret;
+	}
+
 	if (psp->adev->psp.ta_fw) {
 		ret = psp_ras_initialize(psp);
 		if (ret)
@@ -2751,6 +2842,9 @@ int psp_init_sos_microcode(struct psp_context *psp,
 			adev->psp.spl_bin_size = le32_to_cpu(sos_hdr_v1_3->spl_size_bytes);
 			adev->psp.spl_start_addr = (uint8_t *)adev->psp.sys_start_addr +
 				le32_to_cpu(sos_hdr_v1_3->spl_offset_bytes);
+			adev->psp.rl_bin_size = le32_to_cpu(sos_hdr_v1_3->rl_size_bytes);
+			adev->psp.rl_start_addr = (uint8_t *)adev->psp.sys_start_addr +
+				le32_to_cpu(sos_hdr_v1_3->rl_offset_bytes);
 		}
 		break;
 	default:
@@ -2916,7 +3010,7 @@ static ssize_t psp_usbc_pd_fw_sysfs_read(struct device *dev,
 		return ret;
 	}
 
-	return snprintf(buf, PAGE_SIZE, "%x\n", fw_ver);
+	return sysfs_emit(buf, "%x\n", fw_ver);
 }
 
 static ssize_t psp_usbc_pd_fw_sysfs_write(struct device *dev,
@@ -3052,3 +3146,11 @@ const struct amdgpu_ip_block_version psp_v12_0_ip_block =
 	.rev = 0,
 	.funcs = &psp_ip_funcs,
 };
+
+const struct amdgpu_ip_block_version psp_v13_0_ip_block = {
+	.type = AMD_IP_BLOCK_TYPE_PSP,
+	.major = 13,
+	.minor = 0,
+	.rev = 0,
+	.funcs = &psp_ip_funcs,
+};

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

@@ -37,7 +37,7 @@
 #define PSP_XGMI_SHARED_MEM_SIZE 0x4000
 #define PSP_RAS_SHARED_MEM_SIZE 0x4000
 #define PSP_1_MEG		0x100000
-#define PSP_TMR_SIZE	0x400000
+#define PSP_TMR_SIZE(adev)	((adev)->asic_type == CHIP_ALDEBARAN ? 0x800000 : 0x400000)
 #define PSP_HDCP_SHARED_MEM_SIZE	0x4000
 #define PSP_DTM_SHARED_MEM_SIZE	0x4000
 #define PSP_RAP_SHARED_MEM_SIZE	0x4000
@@ -248,11 +248,13 @@ struct psp_context
 	uint32_t			toc_bin_size;
 	uint32_t			kdb_bin_size;
 	uint32_t			spl_bin_size;
+	uint32_t			rl_bin_size;
 	uint8_t				*sys_start_addr;
 	uint8_t				*sos_start_addr;
 	uint8_t				*toc_start_addr;
 	uint8_t				*kdb_start_addr;
 	uint8_t				*spl_start_addr;
+	uint8_t				*rl_start_addr;
 
 	/* tmr buffer */
 	struct amdgpu_bo		*tmr_bo;
@@ -365,12 +367,14 @@ struct amdgpu_psp_funcs {
 extern const struct amd_ip_funcs psp_ip_funcs;
 
 extern const struct amdgpu_ip_block_version psp_v3_1_ip_block;
+extern const struct amdgpu_ip_block_version psp_v10_0_ip_block;
+extern const struct amdgpu_ip_block_version psp_v11_0_ip_block;
+extern const struct amdgpu_ip_block_version psp_v12_0_ip_block;
+extern const struct amdgpu_ip_block_version psp_v13_0_ip_block;
+
 extern int psp_wait_for(struct psp_context *psp, uint32_t reg_index,
 			uint32_t field_val, uint32_t mask, bool check_changed);
 
-extern const struct amdgpu_ip_block_version psp_v10_0_ip_block;
-extern const struct amdgpu_ip_block_version psp_v12_0_ip_block;
-
 int psp_gpu_reset(struct amdgpu_device *adev);
 int psp_update_vcn_sram(struct amdgpu_device *adev, int inst_idx,
 			uint64_t cmd_gpu_addr, int cmd_size);
@@ -395,12 +399,11 @@ int psp_ras_trigger_error(struct psp_context *psp,
 
 int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
 int psp_dtm_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
-int psp_rap_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
+int psp_rap_invoke(struct psp_context *psp, uint32_t ta_cmd_id, enum ta_rap_status *status);
 int psp_securedisplay_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
 
 int psp_rlc_autoload_start(struct psp_context *psp);
 
-extern const struct amdgpu_ip_block_version psp_v11_0_ip_block;
 int psp_reg_program(struct psp_context *psp, enum psp_reg_prog_id reg,
 		uint32_t value);
 int psp_ring_cmd_submit(struct psp_context *psp,
@@ -417,4 +420,7 @@ int psp_init_ta_microcode(struct psp_context *psp,
 			  const char *chip_name);
 int psp_get_fw_attestation_records_addr(struct psp_context *psp,
 					uint64_t *output_ptr);
+
+int psp_load_fw_list(struct psp_context *psp,
+		     struct amdgpu_firmware_info **ucode_list, int ucode_count);
 #endif

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

@@ -48,6 +48,7 @@ static ssize_t amdgpu_rap_debugfs_write(struct file *f, const char __user *buf,
 	struct ta_rap_cmd_output_data *rap_cmd_output;
 	struct drm_device *dev = adev_to_drm(adev);
 	uint32_t op;
+	enum ta_rap_status status;
 	int ret;
 
 	if (*pos || size != 2)
@@ -70,9 +71,8 @@ static ssize_t amdgpu_rap_debugfs_write(struct file *f, const char __user *buf,
 
 	switch (op) {
 	case 2:
-		ret = psp_rap_invoke(&adev->psp, op);
+		ret = psp_rap_invoke(&adev->psp, op, &status);
-
+		if (!ret && status == TA_RAP_STATUS__SUCCESS) {
-		if (ret == TA_RAP_STATUS__SUCCESS) {
 			dev_info(adev->dev, "RAP L0 validate test success.\n");
 		} else {
 			rap_shared_mem = (struct ta_rap_shared_memory *)
@@ -97,6 +97,7 @@ static ssize_t amdgpu_rap_debugfs_write(struct file *f, const char __user *buf,
 	default:
 		dev_info(adev->dev, "Unsupported op id: %d, ", op);
 		dev_info(adev->dev, "Only support op 2(L0 validate test).\n");
+		break;
 	}
 
 	amdgpu_gfx_off_ctrl(adev, true);

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

@@ -99,6 +99,49 @@ static bool amdgpu_ras_get_error_query_ready(struct amdgpu_device *adev)
 	return false;
 }
 
+static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t address)
+{
+	struct ras_err_data err_data = {0, 0, 0, NULL};
+	struct eeprom_table_record err_rec;
+
+	if ((address >= adev->gmc.mc_vram_size) ||
+	    (address >= RAS_UMC_INJECT_ADDR_LIMIT)) {
+		dev_warn(adev->dev,
+		         "RAS WARN: input address 0x%llx is invalid.\n",
+		         address);
+		return -EINVAL;
+	}
+
+	if (amdgpu_ras_check_bad_page(adev, address)) {
+		dev_warn(adev->dev,
+			 "RAS WARN: 0x%llx has been marked as bad page!\n",
+			 address);
+		return 0;
+	}
+
+	memset(&err_rec, 0x0, sizeof(struct eeprom_table_record));
+
+	err_rec.address = address;
+	err_rec.retired_page = address >> AMDGPU_GPU_PAGE_SHIFT;
+	err_rec.ts = (uint64_t)ktime_get_real_seconds();
+	err_rec.err_type = AMDGPU_RAS_EEPROM_ERR_NON_RECOVERABLE;
+
+	err_data.err_addr = &err_rec;
+	err_data.err_addr_cnt = 1;
+
+	if (amdgpu_bad_page_threshold != 0) {
+		amdgpu_ras_add_bad_pages(adev, err_data.err_addr,
+					 err_data.err_addr_cnt);
+		amdgpu_ras_save_bad_pages(adev);
+	}
+
+	dev_warn(adev->dev, "WARNING: THIS IS ONLY FOR TEST PURPOSES AND WILL CORRUPT RAS EEPROM\n");
+	dev_warn(adev->dev, "Clear EEPROM:\n");
+	dev_warn(adev->dev, "    echo 1 > /sys/kernel/debug/dri/0/ras/ras_eeprom_reset\n");
+
+	return 0;
+}
+
 static ssize_t amdgpu_ras_debugfs_read(struct file *f, char __user *buf,
 					size_t size, loff_t *pos)
 {
@@ -109,7 +152,7 @@ static ssize_t amdgpu_ras_debugfs_read(struct file *f, char __user *buf,
 	ssize_t s;
 	char val[128];
 
-	if (amdgpu_ras_error_query(obj->adev, &info))
+	if (amdgpu_ras_query_error_status(obj->adev, &info))
 		return -EINVAL;
 
 	s = snprintf(val, sizeof(val), "%s: %lu\n%s: %lu\n",
@@ -178,11 +221,25 @@ static int amdgpu_ras_debugfs_ctrl_parse_data(struct file *f,
 		op = 1;
 	else if (sscanf(str, "inject %32s %8s", block_name, err) == 2)
 		op = 2;
+	else if (sscanf(str, "retire_page") == 0)
+		op = 3;
 	else if (str[0] && str[1] && str[2] && str[3])
 		/* ascii string, but commands are not matched. */
 		return -EINVAL;
 
 	if (op != -1) {
+
+		if (op == 3) {
+			if (sscanf(str, "%*s %llu", &address) != 1)
+				if (sscanf(str, "%*s 0x%llx", &address) != 1)
+					return -EINVAL;
+
+			data->op = op;
+			data->inject.address = address;
+
+			return 0;
+		}
+
 		if (amdgpu_ras_find_block_id_by_name(block_name, &block_id))
 			return -EINVAL;
 
@@ -310,6 +367,16 @@ static ssize_t amdgpu_ras_debugfs_ctrl_write(struct file *f, const char __user *
 	if (ret)
 		return -EINVAL;
 
+	if (data.op == 3)
+	{
+		ret = amdgpu_reserve_page_direct(adev, data.inject.address);
+
+		if (ret)
+			return size;
+		else
+			return ret;
+	}
+
 	if (!amdgpu_ras_is_supported(adev, data.head.block))
 		return -EINVAL;
 
@@ -431,15 +498,13 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev,
 	};
 
 	if (!amdgpu_ras_get_error_query_ready(obj->adev))
-		return snprintf(buf, PAGE_SIZE,
+		return sysfs_emit(buf, "Query currently inaccessible\n");
-				"Query currently inaccessible\n");
 
-	if (amdgpu_ras_error_query(obj->adev, &info))
+	if (amdgpu_ras_query_error_status(obj->adev, &info))
 		return -EINVAL;
 
-	return snprintf(buf, PAGE_SIZE, "%s: %lu\n%s: %lu\n",
+	return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count,
-			"ue", info.ue_count,
+			  "ce", info.ce_count);
-			"ce", info.ce_count);
 }
 
 /* obj begin */
@@ -449,11 +514,10 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev,
 
 static inline void put_obj(struct ras_manager *obj)
 {
-	if (obj && --obj->use == 0)
+	if (obj && (--obj->use == 0))
 		list_del(&obj->node);
-	if (obj && obj->use < 0) {
+	if (obj && (obj->use < 0))
-		 DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", obj->head.name);
+		DRM_ERROR("RAS ERROR: Unbalance obj(%s) use\n", obj->head.name);
-	}
 }
 
 /* make one obj and return it. */
@@ -463,7 +527,7 @@ static struct ras_manager *amdgpu_ras_create_obj(struct amdgpu_device *adev,
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_manager *obj;
 
-	if (!con)
+	if (!adev->ras_features || !con)
 		return NULL;
 
 	if (head->block >= AMDGPU_RAS_BLOCK_COUNT)
@@ -490,7 +554,7 @@ struct ras_manager *amdgpu_ras_find_obj(struct amdgpu_device *adev,
 	struct ras_manager *obj;
 	int i;
 
-	if (!con)
+	if (!adev->ras_features || !con)
 		return NULL;
 
 	if (head) {
@@ -590,7 +654,11 @@ static int __amdgpu_ras_feature_enable(struct amdgpu_device *adev,
 		con->features |= BIT(head->block);
 	} else {
 		if (obj && amdgpu_ras_is_feature_enabled(adev, head)) {
-			con->features &= ~BIT(head->block);
+			/* skip clean gfx ras context feature for VEGA20 Gaming.
+			 * will clean later
+			 */
+			if (!(!adev->ras_features && con->features & BIT(AMDGPU_RAS_BLOCK__GFX)))
+				con->features &= ~BIT(head->block);
 			put_obj(obj);
 		}
 	}
@@ -693,6 +761,10 @@ int amdgpu_ras_feature_enable_on_boot(struct amdgpu_device *adev,
 			if (ret)
 				return ret;
 
+			/* gfx block ras dsiable cmd must send to ras-ta */
+			if (head->block == AMDGPU_RAS_BLOCK__GFX)
+				con->features |= BIT(head->block);
+
 			ret = amdgpu_ras_feature_enable(adev, head, 0);
 		}
 	} else
@@ -757,8 +829,8 @@ static int amdgpu_ras_enable_all_features(struct amdgpu_device *adev,
 /* feature ctl end */
 
 /* query/inject/cure begin */
-int amdgpu_ras_error_query(struct amdgpu_device *adev,
+int amdgpu_ras_query_error_status(struct amdgpu_device *adev,
-		struct ras_query_if *info)
+	struct ras_query_if *info)
 {
 	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &info->head);
 	struct ras_err_data err_data = {0, 0, 0, NULL};
@@ -769,13 +841,15 @@ int amdgpu_ras_error_query(struct amdgpu_device *adev,
 
 	switch (info->head.block) {
 	case AMDGPU_RAS_BLOCK__UMC:
-		if (adev->umc.funcs->query_ras_error_count)
+		if (adev->umc.ras_funcs &&
-			adev->umc.funcs->query_ras_error_count(adev, &err_data);
+		    adev->umc.ras_funcs->query_ras_error_count)
+			adev->umc.ras_funcs->query_ras_error_count(adev, &err_data);
 		/* umc query_ras_error_address is also responsible for clearing
 		 * error status
 		 */
-		if (adev->umc.funcs->query_ras_error_address)
+		if (adev->umc.ras_funcs &&
-			adev->umc.funcs->query_ras_error_address(adev, &err_data);
+		    adev->umc.ras_funcs->query_ras_error_address)
+			adev->umc.ras_funcs->query_ras_error_address(adev, &err_data);
 		break;
 	case AMDGPU_RAS_BLOCK__SDMA:
 		if (adev->sdma.funcs->query_ras_error_count) {
@@ -785,19 +859,32 @@ int amdgpu_ras_error_query(struct amdgpu_device *adev,
 		}
 		break;
 	case AMDGPU_RAS_BLOCK__GFX:
-		if (adev->gfx.funcs->query_ras_error_count)
+		if (adev->gfx.ras_funcs &&
-			adev->gfx.funcs->query_ras_error_count(adev, &err_data);
+		    adev->gfx.ras_funcs->query_ras_error_count)
+			adev->gfx.ras_funcs->query_ras_error_count(adev, &err_data);
+
+		if (adev->gfx.ras_funcs &&
+		    adev->gfx.ras_funcs->query_ras_error_status)
+			adev->gfx.ras_funcs->query_ras_error_status(adev);
 		break;
 	case AMDGPU_RAS_BLOCK__MMHUB:
-		if (adev->mmhub.funcs->query_ras_error_count)
+		if (adev->mmhub.ras_funcs &&
-			adev->mmhub.funcs->query_ras_error_count(adev, &err_data);
+		    adev->mmhub.ras_funcs->query_ras_error_count)
+			adev->mmhub.ras_funcs->query_ras_error_count(adev, &err_data);
+
+		if (adev->mmhub.ras_funcs &&
+		    adev->mmhub.ras_funcs->query_ras_error_status)
+			adev->mmhub.ras_funcs->query_ras_error_status(adev);
 		break;
 	case AMDGPU_RAS_BLOCK__PCIE_BIF:
-		if (adev->nbio.funcs->query_ras_error_count)
+		if (adev->nbio.ras_funcs &&
-			adev->nbio.funcs->query_ras_error_count(adev, &err_data);
+		    adev->nbio.ras_funcs->query_ras_error_count)
+			adev->nbio.ras_funcs->query_ras_error_count(adev, &err_data);
 		break;
 	case AMDGPU_RAS_BLOCK__XGMI_WAFL:
-		amdgpu_xgmi_query_ras_error_count(adev, &err_data);
+		if (adev->gmc.xgmi.ras_funcs &&
+		    adev->gmc.xgmi.ras_funcs->query_ras_error_count)
+			adev->gmc.xgmi.ras_funcs->query_ras_error_count(adev, &err_data);
 		break;
 	default:
 		break;
@@ -826,6 +913,38 @@ int amdgpu_ras_error_query(struct amdgpu_device *adev,
 	return 0;
 }
 
+int amdgpu_ras_reset_error_status(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block)
+{
+	if (!amdgpu_ras_is_supported(adev, block))
+		return -EINVAL;
+
+	switch (block) {
+	case AMDGPU_RAS_BLOCK__GFX:
+		if (adev->gfx.ras_funcs &&
+		    adev->gfx.ras_funcs->reset_ras_error_count)
+			adev->gfx.ras_funcs->reset_ras_error_count(adev);
+
+		if (adev->gfx.ras_funcs &&
+		    adev->gfx.ras_funcs->reset_ras_error_status)
+			adev->gfx.ras_funcs->reset_ras_error_status(adev);
+		break;
+	case AMDGPU_RAS_BLOCK__MMHUB:
+		if (adev->mmhub.ras_funcs &&
+		    adev->mmhub.ras_funcs->reset_ras_error_count)
+			adev->mmhub.ras_funcs->reset_ras_error_count(adev);
+		break;
+	case AMDGPU_RAS_BLOCK__SDMA:
+		if (adev->sdma.funcs->reset_ras_error_count)
+			adev->sdma.funcs->reset_ras_error_count(adev);
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+}
+
 /* Trigger XGMI/WAFL error */
 static int amdgpu_ras_error_inject_xgmi(struct amdgpu_device *adev,
 				 struct ta_ras_trigger_error_input *block_info)
@@ -878,12 +997,14 @@ int amdgpu_ras_error_inject(struct amdgpu_device *adev,
 
 	switch (info->head.block) {
 	case AMDGPU_RAS_BLOCK__GFX:
-		if (adev->gfx.funcs->ras_error_inject)
+		if (adev->gfx.ras_funcs &&
-			ret = adev->gfx.funcs->ras_error_inject(adev, info);
+		    adev->gfx.ras_funcs->ras_error_inject)
+			ret = adev->gfx.ras_funcs->ras_error_inject(adev, info);
 		else
 			ret = -EINVAL;
 		break;
 	case AMDGPU_RAS_BLOCK__UMC:
+	case AMDGPU_RAS_BLOCK__SDMA:
 	case AMDGPU_RAS_BLOCK__MMHUB:
 	case AMDGPU_RAS_BLOCK__PCIE_BIF:
 		ret = psp_ras_trigger_error(&adev->psp, &block_info);
@@ -913,7 +1034,7 @@ unsigned long amdgpu_ras_query_error_count(struct amdgpu_device *adev,
 	struct ras_manager *obj;
 	struct ras_err_data data = {0, 0};
 
-	if (!con)
+	if (!adev->ras_features || !con)
 		return 0;
 
 	list_for_each_entry(obj, &con->head, node) {
@@ -921,7 +1042,7 @@ unsigned long amdgpu_ras_query_error_count(struct amdgpu_device *adev,
 			.head = obj->head,
 		};
 
-		if (amdgpu_ras_error_query(adev, &info))
+		if (amdgpu_ras_query_error_status(adev, &info))
 			return 0;
 
 		data.ce_count += info.ce_count;
@@ -1137,16 +1258,17 @@ static int amdgpu_ras_sysfs_remove_all(struct amdgpu_device *adev)
  *
  */
 /* debugfs begin */
-static void amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device *adev)
+static struct dentry *amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct dentry *dir;
 	struct drm_minor *minor = adev_to_drm(adev)->primary;
 
-	con->dir = debugfs_create_dir(RAS_FS_NAME, minor->debugfs_root);
+	dir = debugfs_create_dir(RAS_FS_NAME, minor->debugfs_root);
-	debugfs_create_file("ras_ctrl", S_IWUGO | S_IRUGO, con->dir,
+	debugfs_create_file("ras_ctrl", S_IWUGO | S_IRUGO, dir, adev,
-				adev, &amdgpu_ras_debugfs_ctrl_ops);
+			    &amdgpu_ras_debugfs_ctrl_ops);
-	debugfs_create_file("ras_eeprom_reset", S_IWUGO | S_IRUGO, con->dir,
+	debugfs_create_file("ras_eeprom_reset", S_IWUGO | S_IRUGO, dir, adev,
-				adev, &amdgpu_ras_debugfs_eeprom_ops);
+			    &amdgpu_ras_debugfs_eeprom_ops);
 
 	/*
 	 * After one uncorrectable error happens, usually GPU recovery will
@@ -1156,24 +1278,24 @@ static void amdgpu_ras_debugfs_create_ctrl_node(struct amdgpu_device *adev)
 	 * ERREVENT_ATHUB_INTERRUPT generated. Normal GPU recovery routine
 	 * will never be called.
 	 */
-	debugfs_create_bool("auto_reboot", S_IWUGO | S_IRUGO, con->dir,
+	debugfs_create_bool("auto_reboot", S_IWUGO | S_IRUGO, dir, &con->reboot);
-				&con->reboot);
 
 	/*
 	 * User could set this not to clean up hardware's error count register
 	 * of RAS IPs during ras recovery.
 	 */
-	debugfs_create_bool("disable_ras_err_cnt_harvest", 0644,
+	debugfs_create_bool("disable_ras_err_cnt_harvest", 0644, dir,
-			con->dir, &con->disable_ras_err_cnt_harvest);
+			    &con->disable_ras_err_cnt_harvest);
+	return dir;
 }
 
 static void amdgpu_ras_debugfs_create(struct amdgpu_device *adev,
-		struct ras_fs_if *head)
+				      struct ras_fs_if *head,
+				      struct dentry *dir)
 {
-	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head->head);
 
-	if (!obj || obj->ent)
+	if (!obj || !dir)
 		return;
 
 	get_obj(obj);
@@ -1182,14 +1304,14 @@ static void amdgpu_ras_debugfs_create(struct amdgpu_device *adev,
 			head->debugfs_name,
 			sizeof(obj->fs_data.debugfs_name));
 
-	obj->ent = debugfs_create_file(obj->fs_data.debugfs_name,
+	debugfs_create_file(obj->fs_data.debugfs_name, S_IWUGO | S_IRUGO, dir,
-				       S_IWUGO | S_IRUGO, con->dir, obj,
+			    obj, &amdgpu_ras_debugfs_ops);
-				       &amdgpu_ras_debugfs_ops);
 }
 
 void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct dentry *dir;
 	struct ras_manager *obj;
 	struct ras_fs_if fs_info;
 
@@ -1200,7 +1322,7 @@ void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev)
 	if (!IS_ENABLED(CONFIG_DEBUG_FS) || !con)
 		return;
 
-	amdgpu_ras_debugfs_create_ctrl_node(adev);
+	dir = amdgpu_ras_debugfs_create_ctrl_node(adev);
 
 	list_for_each_entry(obj, &con->head, node) {
 		if (amdgpu_ras_is_supported(adev, obj->head.block) &&
@@ -1208,34 +1330,11 @@ void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev)
 			sprintf(fs_info.debugfs_name, "%s_err_inject",
 					ras_block_str(obj->head.block));
 			fs_info.head = obj->head;
-			amdgpu_ras_debugfs_create(adev, &fs_info);
+			amdgpu_ras_debugfs_create(adev, &fs_info, dir);
 		}
 	}
 }
 
-static void amdgpu_ras_debugfs_remove(struct amdgpu_device *adev,
-		struct ras_common_if *head)
-{
-	struct ras_manager *obj = amdgpu_ras_find_obj(adev, head);
-
-	if (!obj || !obj->ent)
-		return;
-
-	obj->ent = NULL;
-	put_obj(obj);
-}
-
-static void amdgpu_ras_debugfs_remove_all(struct amdgpu_device *adev)
-{
-	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
-	struct ras_manager *obj, *tmp;
-
-	list_for_each_entry_safe(obj, tmp, &con->head, node) {
-		amdgpu_ras_debugfs_remove(adev, &obj->head);
-	}
-
-	con->dir = NULL;
-}
 /* debugfs end */
 
 /* ras fs */
@@ -1282,8 +1381,17 @@ static int amdgpu_ras_fs_init(struct amdgpu_device *adev)
 
 static int amdgpu_ras_fs_fini(struct amdgpu_device *adev)
 {
-	if (IS_ENABLED(CONFIG_DEBUG_FS))
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
-		amdgpu_ras_debugfs_remove_all(adev);
+	struct ras_manager *con_obj, *ip_obj, *tmp;
+
+	if (IS_ENABLED(CONFIG_DEBUG_FS)) {
+		list_for_each_entry_safe(con_obj, tmp, &con->head, node) {
+			ip_obj = amdgpu_ras_find_obj(adev, &con_obj->head);
+			if (ip_obj)
+				put_obj(ip_obj);
+		}
+	}
+
 	amdgpu_ras_sysfs_remove_all(adev);
 	return 0;
 }
@@ -1447,7 +1555,7 @@ static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev)
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_manager *obj;
 
-	if (!con)
+	if (!adev->ras_features || !con)
 		return;
 
 	list_for_each_entry(obj, &con->head, node) {
@@ -1464,7 +1572,7 @@ static void amdgpu_ras_log_on_err_counter(struct amdgpu_device *adev)
 		if (info.head.block == AMDGPU_RAS_BLOCK__PCIE_BIF)
 			continue;
 
-		amdgpu_ras_error_query(adev, &info);
+		amdgpu_ras_query_error_status(adev, &info);
 	}
 }
 
@@ -1478,12 +1586,14 @@ static void amdgpu_ras_error_status_query(struct amdgpu_device *adev,
 	 */
 	switch (info->head.block) {
 	case AMDGPU_RAS_BLOCK__GFX:
-		if (adev->gfx.funcs->query_ras_error_status)
+		if (adev->gfx.ras_funcs &&
-			adev->gfx.funcs->query_ras_error_status(adev);
+		    adev->gfx.ras_funcs->query_ras_error_status)
+			adev->gfx.ras_funcs->query_ras_error_status(adev);
 		break;
 	case AMDGPU_RAS_BLOCK__MMHUB:
-		if (adev->mmhub.funcs->query_ras_error_status)
+		if (adev->mmhub.ras_funcs &&
-			adev->mmhub.funcs->query_ras_error_status(adev);
+		    adev->mmhub.ras_funcs->query_ras_error_status)
+			adev->mmhub.ras_funcs->query_ras_error_status(adev);
 		break;
 	default:
 		break;
@@ -1495,7 +1605,7 @@ static void amdgpu_ras_query_err_status(struct amdgpu_device *adev)
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_manager *obj;
 
-	if (!con)
+	if (!adev->ras_features || !con)
 		return;
 
 	list_for_each_entry(obj, &con->head, node) {
@@ -1809,7 +1919,7 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
 	bool exc_err_limit = false;
 	int ret;
 
-	if (con)
+	if (adev->ras_features && con)
 		data = &con->eh_data;
 	else
 		return 0;
@@ -1828,6 +1938,12 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
 	max_eeprom_records_len = amdgpu_ras_eeprom_get_record_max_length();
 	amdgpu_ras_validate_threshold(adev, max_eeprom_records_len);
 
+	/* Todo: During test the SMU might fail to read the eeprom through I2C
+	 * when the GPU is pending on XGMI reset during probe time
+	 * (Mostly after second bus reset), skip it now
+	 */
+	if (adev->gmc.xgmi.pending_reset)
+		return 0;
 	ret = amdgpu_ras_eeprom_init(&con->eeprom_control, &exc_err_limit);
 	/*
 	 * This calling fails when exc_err_limit is true or
@@ -1897,15 +2013,13 @@ int amdgpu_ras_request_reset_on_boot(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int amdgpu_ras_check_asic_type(struct amdgpu_device *adev)
+static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev)
 {
-	if (adev->asic_type != CHIP_VEGA10 &&
+	return adev->asic_type == CHIP_VEGA10 ||
-		adev->asic_type != CHIP_VEGA20 &&
+		adev->asic_type == CHIP_VEGA20 ||
-		adev->asic_type != CHIP_ARCTURUS &&
+		adev->asic_type == CHIP_ARCTURUS ||
-		adev->asic_type != CHIP_SIENNA_CICHLID)
+		adev->asic_type == CHIP_ALDEBARAN ||
-		return 1;
+		adev->asic_type == CHIP_SIENNA_CICHLID;
-	else
-		return 0;
 }
 
 /*
@@ -1924,22 +2038,32 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev,
 	*supported = 0;
 
 	if (amdgpu_sriov_vf(adev) || !adev->is_atom_fw ||
-		amdgpu_ras_check_asic_type(adev))
+	    !amdgpu_ras_asic_supported(adev))
 		return;
 
-	if (amdgpu_atomfirmware_mem_ecc_supported(adev)) {
+	if (!adev->gmc.xgmi.connected_to_cpu) {
-		dev_info(adev->dev, "HBM ECC is active.\n");
+		if (amdgpu_atomfirmware_mem_ecc_supported(adev)) {
-		*hw_supported |= (1 << AMDGPU_RAS_BLOCK__UMC |
+			dev_info(adev->dev, "MEM ECC is active.\n");
-				1 << AMDGPU_RAS_BLOCK__DF);
+			*hw_supported |= (1 << AMDGPU_RAS_BLOCK__UMC |
-	} else
+					1 << AMDGPU_RAS_BLOCK__DF);
-		dev_info(adev->dev, "HBM ECC is not presented.\n");
+		} else {
+			dev_info(adev->dev, "MEM ECC is not presented.\n");
+		}
 
-	if (amdgpu_atomfirmware_sram_ecc_supported(adev)) {
+		if (amdgpu_atomfirmware_sram_ecc_supported(adev)) {
-		dev_info(adev->dev, "SRAM ECC is active.\n");
+			dev_info(adev->dev, "SRAM ECC is active.\n");
-		*hw_supported |= ~(1 << AMDGPU_RAS_BLOCK__UMC |
+			*hw_supported |= ~(1 << AMDGPU_RAS_BLOCK__UMC |
-				1 << AMDGPU_RAS_BLOCK__DF);
+					1 << AMDGPU_RAS_BLOCK__DF);
-	} else
+		} else {
-		dev_info(adev->dev, "SRAM ECC is not presented.\n");
+			dev_info(adev->dev, "SRAM ECC is not presented.\n");
+		}
+	} else {
+		/* driver only manages a few IP blocks RAS feature
+		 * when GPU is connected cpu through XGMI */
+		*hw_supported |= (1 << AMDGPU_RAS_BLOCK__GFX |
+				1 << AMDGPU_RAS_BLOCK__SDMA |
+				1 << AMDGPU_RAS_BLOCK__MMHUB);
+	}
 
 	/* hw_supported needs to be aligned with RAS block mask. */
 	*hw_supported &= AMDGPU_RAS_BLOCK_MASK;
@@ -1970,6 +2094,15 @@ int amdgpu_ras_init(struct amdgpu_device *adev)
 	amdgpu_ras_check_supported(adev, &con->hw_supported,
 			&con->supported);
 	if (!con->hw_supported || (adev->asic_type == CHIP_VEGA10)) {
+		/* set gfx block ras context feature for VEGA20 Gaming
+		 * send ras disable cmd to ras ta during ras late init.
+		 */
+		if (!adev->ras_features && adev->asic_type == CHIP_VEGA20) {
+			con->features |= BIT(AMDGPU_RAS_BLOCK__GFX);
+
+			return 0;
+		}
+
 		r = 0;
 		goto release_con;
 	}
@@ -1979,14 +2112,31 @@ int amdgpu_ras_init(struct amdgpu_device *adev)
 	/* Might need get this flag from vbios. */
 	con->flags = RAS_DEFAULT_FLAGS;
 
-	if (adev->nbio.funcs->init_ras_controller_interrupt) {
+	/* initialize nbio ras function ahead of any other
-		r = adev->nbio.funcs->init_ras_controller_interrupt(adev);
+	 * ras functions so hardware fatal error interrupt
+	 * can be enabled as early as possible */
+	switch (adev->asic_type) {
+	case CHIP_VEGA20:
+	case CHIP_ARCTURUS:
+	case CHIP_ALDEBARAN:
+		if (!adev->gmc.xgmi.connected_to_cpu)
+			adev->nbio.ras_funcs = &nbio_v7_4_ras_funcs;
+		break;
+	default:
+		/* nbio ras is not available */
+		break;
+	}
+
+	if (adev->nbio.ras_funcs &&
+	    adev->nbio.ras_funcs->init_ras_controller_interrupt) {
+		r = adev->nbio.ras_funcs->init_ras_controller_interrupt(adev);
 		if (r)
 			goto release_con;
 	}
 
-	if (adev->nbio.funcs->init_ras_err_event_athub_interrupt) {
+	if (adev->nbio.ras_funcs &&
-		r = adev->nbio.funcs->init_ras_err_event_athub_interrupt(adev);
+	    adev->nbio.ras_funcs->init_ras_err_event_athub_interrupt) {
+		r = adev->nbio.ras_funcs->init_ras_err_event_athub_interrupt(adev);
 		if (r)
 			goto release_con;
 	}
@@ -2007,6 +2157,32 @@ release_con:
 	return r;
 }
 
+static int amdgpu_persistent_edc_harvesting_supported(struct amdgpu_device *adev)
+{
+	if (adev->gmc.xgmi.connected_to_cpu)
+		return 1;
+	return 0;
+}
+
+static int amdgpu_persistent_edc_harvesting(struct amdgpu_device *adev,
+					struct ras_common_if *ras_block)
+{
+	struct ras_query_if info = {
+		.head = *ras_block,
+	};
+
+	if (!amdgpu_persistent_edc_harvesting_supported(adev))
+		return 0;
+
+	if (amdgpu_ras_query_error_status(adev, &info) != 0)
+		DRM_WARN("RAS init harvest failure");
+
+	if (amdgpu_ras_reset_error_status(adev, ras_block->block) != 0)
+		DRM_WARN("RAS init harvest reset failure");
+
+	return 0;
+}
+
 /* helper function to handle common stuff in ip late init phase */
 int amdgpu_ras_late_init(struct amdgpu_device *adev,
 			 struct ras_common_if *ras_block,
@@ -2036,6 +2212,9 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev,
 			return r;
 	}
 
+	/* check for errors on warm reset edc persisant supported ASIC */
+	amdgpu_persistent_edc_harvesting(adev, ras_block);
+
 	/* in resume phase, no need to create ras fs node */
 	if (adev->in_suspend || amdgpu_in_reset(adev))
 		return 0;
@@ -2083,8 +2262,12 @@ void amdgpu_ras_resume(struct amdgpu_device *adev)
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_manager *obj, *tmp;
 
-	if (!con)
+	if (!adev->ras_features || !con) {
+		/* clean ras context for VEGA20 Gaming after send ras disable cmd */
+		amdgpu_release_ras_context(adev);
+
 		return;
+	}
 
 	if (con->flags & AMDGPU_RAS_FLAG_INIT_BY_VBIOS) {
 		/* Set up all other IPs which are not implemented. There is a
@@ -2125,7 +2308,7 @@ void amdgpu_ras_suspend(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
-	if (!con)
+	if (!adev->ras_features || !con)
 		return;
 
 	amdgpu_ras_disable_all_features(adev, 0);
@@ -2139,7 +2322,7 @@ int amdgpu_ras_pre_fini(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
-	if (!con)
+	if (!adev->ras_features || !con)
 		return 0;
 
 	/* Need disable ras on all IPs here before ip [hw/sw]fini */
@@ -2152,7 +2335,7 @@ int amdgpu_ras_fini(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
-	if (!con)
+	if (!adev->ras_features || !con)
 		return 0;
 
 	amdgpu_ras_fs_fini(adev);
@@ -2196,18 +2379,16 @@ bool amdgpu_ras_need_emergency_restart(struct amdgpu_device *adev)
 	return false;
 }
 
-bool amdgpu_ras_check_err_threshold(struct amdgpu_device *adev)
+void amdgpu_release_ras_context(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
-	bool exc_err_limit = false;
 
-	if (con && (amdgpu_bad_page_threshold != 0))
+	if (!con)
-		amdgpu_ras_eeprom_check_err_threshold(&con->eeprom_control,
+		return;
-						&exc_err_limit);
 
-	/*
+	if (!adev->ras_features && con->features & BIT(AMDGPU_RAS_BLOCK__GFX)) {
-	 * We are only interested in variable exc_err_limit,
+		con->features &= ~BIT(AMDGPU_RAS_BLOCK__GFX);
-	 * as it says if GPU is in bad state or not.
+		amdgpu_ras_set_context(adev, NULL);
-	 */
+		kfree(con);
-	return exc_err_limit;
+	}
 }

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

@@ -318,8 +318,6 @@ struct amdgpu_ras {
 	uint32_t supported;
 	uint32_t features;
 	struct list_head head;
-	/* debugfs */
-	struct dentry *dir;
 	/* sysfs */
 	struct device_attribute features_attr;
 	struct bin_attribute badpages_attr;
@@ -395,8 +393,6 @@ struct ras_manager {
 	struct list_head node;
 	/* the device */
 	struct amdgpu_device *adev;
-	/* debugfs */
-	struct dentry *ent;
 	/* sysfs */
 	struct device_attribute sysfs_attr;
 	int attr_inuse;
@@ -495,8 +491,6 @@ void amdgpu_ras_suspend(struct amdgpu_device *adev);
 unsigned long amdgpu_ras_query_error_count(struct amdgpu_device *adev,
 		bool is_ce);
 
-bool amdgpu_ras_check_err_threshold(struct amdgpu_device *adev);
-
 /* error handling functions */
 int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev,
 		struct eeprom_table_record *bps, int pages);
@@ -594,9 +588,12 @@ int amdgpu_ras_sysfs_remove(struct amdgpu_device *adev,
 
 void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev);
 
-int amdgpu_ras_error_query(struct amdgpu_device *adev,
+int amdgpu_ras_query_error_status(struct amdgpu_device *adev,
 		struct ras_query_if *info);
 
+int amdgpu_ras_reset_error_status(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block);
+
 int amdgpu_ras_error_inject(struct amdgpu_device *adev,
 		struct ras_inject_if *info);
 
@@ -629,4 +626,6 @@ void amdgpu_ras_global_ras_isr(struct amdgpu_device *adev);
 void amdgpu_ras_set_error_query_ready(struct amdgpu_device *adev, bool ready);
 
 bool amdgpu_ras_need_emergency_restart(struct amdgpu_device *adev);
+
+void amdgpu_release_ras_context(struct amdgpu_device *adev);
 #endif

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

@@ -31,6 +31,7 @@
 #define EEPROM_I2C_TARGET_ADDR_ARCTURUS		0xA8
 #define EEPROM_I2C_TARGET_ADDR_ARCTURUS_D342	0xA0
 #define EEPROM_I2C_TARGET_ADDR_SIENNA_CICHLID   0xA0
+#define EEPROM_I2C_TARGET_ADDR_ALDEBARAN        0xA0
 
 /*
  * The 2 macros bellow represent the actual size in bytes that
@@ -64,7 +65,8 @@ static bool __is_ras_eeprom_supported(struct amdgpu_device *adev)
 {
 	if ((adev->asic_type == CHIP_VEGA20) ||
 	    (adev->asic_type == CHIP_ARCTURUS) ||
-	    (adev->asic_type == CHIP_SIENNA_CICHLID))
+	    (adev->asic_type == CHIP_SIENNA_CICHLID) ||
+	    (adev->asic_type == CHIP_ALDEBARAN))
 		return true;
 
 	return false;
@@ -106,6 +108,10 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev,
 		*i2c_addr = EEPROM_I2C_TARGET_ADDR_SIENNA_CICHLID;
 		break;
 
+	case CHIP_ALDEBARAN:
+		*i2c_addr = EEPROM_I2C_TARGET_ADDR_ALDEBARAN;
+		break;
+
 	default:
 		return false;
 	}
@@ -434,47 +440,28 @@ static uint32_t __correct_eeprom_dest_address(uint32_t curr_address)
 	return curr_address;
 }
 
-int amdgpu_ras_eeprom_check_err_threshold(
+bool amdgpu_ras_eeprom_check_err_threshold(struct amdgpu_device *adev)
-				struct amdgpu_ras_eeprom_control *control,
-				bool *exceed_err_limit)
 {
-	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
-	unsigned char buff[EEPROM_ADDRESS_SIZE +
-			EEPROM_TABLE_HEADER_SIZE] = { 0 };
-	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
-	struct i2c_msg msg = {
-			.addr = control->i2c_address,
-			.flags = I2C_M_RD,
-			.len = EEPROM_ADDRESS_SIZE + EEPROM_TABLE_HEADER_SIZE,
-			.buf = buff,
-	};
-	int ret;
-
-	*exceed_err_limit = false;
 
 	if (!__is_ras_eeprom_supported(adev))
-		return 0;
+		return false;
 
-	/* read EEPROM table header */
+	/* skip check eeprom table for VEGA20 Gaming */
-	mutex_lock(&control->tbl_mutex);
+	if (!con)
-	ret = i2c_transfer(&adev->pm.smu_i2c, &msg, 1);
+		return false;
-	if (ret < 1) {
+	else
-		dev_err(adev->dev, "Failed to read EEPROM table header.\n");
+		if (!(con->features & BIT(AMDGPU_RAS_BLOCK__UMC)))
-		goto err;
+			return false;
-	}
 
-	__decode_table_header_from_buff(hdr, &buff[2]);
+	if (con->eeprom_control.tbl_hdr.header == EEPROM_TABLE_HDR_BAD) {
-
-	if (hdr->header == EEPROM_TABLE_HDR_BAD) {
 		dev_warn(adev->dev, "This GPU is in BAD status.");
 		dev_warn(adev->dev, "Please retire it or setting one bigger "
 				"threshold value when reloading driver.\n");
-		*exceed_err_limit = true;
+		return true;
 	}
 
-err:
+	return false;
-	mutex_unlock(&control->tbl_mutex);
-	return 0;
 }
 
 int amdgpu_ras_eeprom_process_recods(struct amdgpu_ras_eeprom_control *control,

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

@@ -80,9 +80,7 @@ int amdgpu_ras_eeprom_init(struct amdgpu_ras_eeprom_control *control,
 			bool *exceed_err_limit);
 int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control);
 
-int amdgpu_ras_eeprom_check_err_threshold(
+bool amdgpu_ras_eeprom_check_err_threshold(struct amdgpu_device *adev);
-				struct amdgpu_ras_eeprom_control *control,
-				bool *exceed_err_limit);
 
 int amdgpu_ras_eeprom_process_recods(struct amdgpu_ras_eeprom_control *control,
 					    struct eeprom_table_record *records,

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

@@ -0,0 +1,105 @@
+// SPDX-License-Identifier: GPL-2.0 OR MIT
+/*
+ * Copyright 2020 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
+ */
+
+#ifndef __AMDGPU_RES_CURSOR_H__
+#define __AMDGPU_RES_CURSOR_H__
+
+#include <drm/drm_mm.h>
+#include <drm/ttm/ttm_resource.h>
+
+/* state back for walking over vram_mgr and gtt_mgr allocations */
+struct amdgpu_res_cursor {
+	uint64_t		start;
+	uint64_t		size;
+	uint64_t		remaining;
+	struct drm_mm_node	*node;
+};
+
+/**
+ * amdgpu_res_first - initialize a amdgpu_res_cursor
+ *
+ * @res: TTM resource object to walk
+ * @start: Start of the range
+ * @size: Size of the range
+ * @cur: cursor object to initialize
+ *
+ * Start walking over the range of allocations between @start and @size.
+ */
+static inline void amdgpu_res_first(struct ttm_resource *res,
+				    uint64_t start, uint64_t size,
+				    struct amdgpu_res_cursor *cur)
+{
+	struct drm_mm_node *node;
+
+	if (!res || !res->mm_node) {
+		cur->start = start;
+		cur->size = size;
+		cur->remaining = size;
+		cur->node = NULL;
+		return;
+	}
+
+	BUG_ON(start + size > res->num_pages << PAGE_SHIFT);
+
+	node = res->mm_node;
+	while (start >= node->size << PAGE_SHIFT)
+		start -= node++->size << PAGE_SHIFT;
+
+	cur->start = (node->start << PAGE_SHIFT) + start;
+	cur->size = min((node->size << PAGE_SHIFT) - start, size);
+	cur->remaining = size;
+	cur->node = node;
+}
+
+/**
+ * amdgpu_res_next - advance the cursor
+ *
+ * @cur: the cursor to advance
+ * @size: number of bytes to move forward
+ *
+ * Move the cursor @size bytes forwrad, walking to the next node if necessary.
+ */
+static inline void amdgpu_res_next(struct amdgpu_res_cursor *cur, uint64_t size)
+{
+	struct drm_mm_node *node = cur->node;
+
+	BUG_ON(size > cur->remaining);
+
+	cur->remaining -= size;
+	if (!cur->remaining)
+		return;
+
+	cur->size -= size;
+	if (cur->size) {
+		cur->start += size;
+		return;
+	}
+
+	cur->node = ++node;
+	cur->start = node->start << PAGE_SHIFT;
+	cur->size = min(node->size << PAGE_SHIFT, cur->remaining);
+}
+
+#endif

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

@@ -0,0 +1,98 @@
+/*
+ * Copyright 2021 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 "amdgpu_reset.h"
+#include "aldebaran.h"
+
+int amdgpu_reset_add_handler(struct amdgpu_reset_control *reset_ctl,
+			     struct amdgpu_reset_handler *handler)
+{
+	/* TODO: Check if handler exists? */
+	list_add_tail(&handler->handler_list, &reset_ctl->reset_handlers);
+	return 0;
+}
+
+int amdgpu_reset_init(struct amdgpu_device *adev)
+{
+	int ret = 0;
+
+	switch (adev->asic_type) {
+	case CHIP_ALDEBARAN:
+		ret = aldebaran_reset_init(adev);
+		break;
+	default:
+		break;
+	}
+
+	return ret;
+}
+
+int amdgpu_reset_fini(struct amdgpu_device *adev)
+{
+	int ret = 0;
+
+	switch (adev->asic_type) {
+	case CHIP_ALDEBARAN:
+		ret = aldebaran_reset_fini(adev);
+		break;
+	default:
+		break;
+	}
+
+	return ret;
+}
+
+int amdgpu_reset_prepare_hwcontext(struct amdgpu_device *adev,
+				   struct amdgpu_reset_context *reset_context)
+{
+	struct amdgpu_reset_handler *reset_handler = NULL;
+
+	if (adev->reset_cntl && adev->reset_cntl->get_reset_handler)
+		reset_handler = adev->reset_cntl->get_reset_handler(
+			adev->reset_cntl, reset_context);
+	if (!reset_handler)
+		return -ENOSYS;
+
+	return reset_handler->prepare_hwcontext(adev->reset_cntl,
+						reset_context);
+}
+
+int amdgpu_reset_perform_reset(struct amdgpu_device *adev,
+			       struct amdgpu_reset_context *reset_context)
+{
+	int ret;
+	struct amdgpu_reset_handler *reset_handler = NULL;
+
+	if (adev->reset_cntl)
+		reset_handler = adev->reset_cntl->get_reset_handler(
+			adev->reset_cntl, reset_context);
+	if (!reset_handler)
+		return -ENOSYS;
+
+	ret = reset_handler->perform_reset(adev->reset_cntl, reset_context);
+	if (ret)
+		return ret;
+
+	return reset_handler->restore_hwcontext(adev->reset_cntl,
+						reset_context);
+}

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

@@ -0,0 +1,85 @@
+/*
+ * Copyright 2021 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_RESET_H__
+#define __AMDGPU_RESET_H__
+
+#include "amdgpu.h"
+
+enum AMDGPU_RESET_FLAGS {
+
+	AMDGPU_NEED_FULL_RESET = 0,
+	AMDGPU_SKIP_HW_RESET = 1,
+};
+
+struct amdgpu_reset_context {
+	enum amd_reset_method method;
+	struct amdgpu_device *reset_req_dev;
+	struct amdgpu_job *job;
+	struct amdgpu_hive_info *hive;
+	unsigned long flags;
+};
+
+struct amdgpu_reset_handler {
+	enum amd_reset_method reset_method;
+	struct list_head handler_list;
+	int (*prepare_env)(struct amdgpu_reset_control *reset_ctl,
+			   struct amdgpu_reset_context *context);
+	int (*prepare_hwcontext)(struct amdgpu_reset_control *reset_ctl,
+				 struct amdgpu_reset_context *context);
+	int (*perform_reset)(struct amdgpu_reset_control *reset_ctl,
+			     struct amdgpu_reset_context *context);
+	int (*restore_hwcontext)(struct amdgpu_reset_control *reset_ctl,
+				 struct amdgpu_reset_context *context);
+	int (*restore_env)(struct amdgpu_reset_control *reset_ctl,
+			   struct amdgpu_reset_context *context);
+
+	int (*do_reset)(struct amdgpu_device *adev);
+};
+
+struct amdgpu_reset_control {
+	void *handle;
+	struct work_struct reset_work;
+	struct mutex reset_lock;
+	struct list_head reset_handlers;
+	atomic_t in_reset;
+	enum amd_reset_method active_reset;
+	struct amdgpu_reset_handler *(*get_reset_handler)(
+		struct amdgpu_reset_control *reset_ctl,
+		struct amdgpu_reset_context *context);
+	void (*async_reset)(struct work_struct *work);
+};
+
+int amdgpu_reset_init(struct amdgpu_device *adev);
+int amdgpu_reset_fini(struct amdgpu_device *adev);
+
+int amdgpu_reset_prepare_hwcontext(struct amdgpu_device *adev,
+				   struct amdgpu_reset_context *reset_context);
+
+int amdgpu_reset_perform_reset(struct amdgpu_device *adev,
+			       struct amdgpu_reset_context *reset_context);
+
+int amdgpu_reset_add_handler(struct amdgpu_reset_control *reset_ctl,
+			     struct amdgpu_reset_handler *handler);
+
+#endif

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

@@ -164,7 +164,8 @@ void amdgpu_ring_undo(struct amdgpu_ring *ring)
  */
 int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 		     unsigned int max_dw, struct amdgpu_irq_src *irq_src,
-		     unsigned int irq_type, unsigned int hw_prio)
+		     unsigned int irq_type, unsigned int hw_prio,
+		     atomic_t *sched_score)
 {
 	int r;
 	int sched_hw_submission = amdgpu_sched_hw_submission;
@@ -189,7 +190,8 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 		ring->adev = adev;
 		ring->idx = adev->num_rings++;
 		adev->rings[ring->idx] = ring;
-		r = amdgpu_fence_driver_init_ring(ring, sched_hw_submission);
+		r = amdgpu_fence_driver_init_ring(ring, sched_hw_submission,
+						  sched_score);
 		if (r)
 			return r;
 	}

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

@@ -111,7 +111,8 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev);
 void amdgpu_fence_driver_force_completion(struct amdgpu_ring *ring);
 
 int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring,
-				  unsigned num_hw_submission);
+				  unsigned num_hw_submission,
+				  atomic_t *sched_score);
 int amdgpu_fence_driver_start_ring(struct amdgpu_ring *ring,
 				   struct amdgpu_irq_src *irq_src,
 				   unsigned irq_type);
@@ -282,7 +283,8 @@ void amdgpu_ring_commit(struct amdgpu_ring *ring);
 void amdgpu_ring_undo(struct amdgpu_ring *ring);
 int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 		     unsigned int ring_size, struct amdgpu_irq_src *irq_src,
-		     unsigned int irq_type, unsigned int prio);
+		     unsigned int irq_type, unsigned int prio,
+		     atomic_t *sched_score);
 void amdgpu_ring_fini(struct amdgpu_ring *ring);
 void amdgpu_ring_emit_reg_write_reg_wait_helper(struct amdgpu_ring *ring,
 						uint32_t reg0, uint32_t val0,

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

@@ -127,7 +127,8 @@ struct amdgpu_rlc_funcs {
 	void (*reset)(struct amdgpu_device *adev);
 	void (*start)(struct amdgpu_device *adev);
 	void (*update_spm_vmid)(struct amdgpu_device *adev, unsigned vmid);
-	void (*rlcg_wreg)(struct amdgpu_device *adev, u32 offset, u32 v);
+	void (*rlcg_wreg)(struct amdgpu_device *adev, u32 offset, u32 v, u32 flag);
+	u32 (*rlcg_rreg)(struct amdgpu_device *adev, u32 offset, u32 flag);
 	bool (*is_rlcg_access_range)(struct amdgpu_device *adev, uint32_t reg);
 };
 

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

@@ -64,6 +64,11 @@ struct amdgpu_sdma {
 	struct amdgpu_irq_src	trap_irq;
 	struct amdgpu_irq_src	illegal_inst_irq;
 	struct amdgpu_irq_src	ecc_irq;
+	struct amdgpu_irq_src	vm_hole_irq;
+	struct amdgpu_irq_src	doorbell_invalid_irq;
+	struct amdgpu_irq_src	pool_timeout_irq;
+	struct amdgpu_irq_src	srbm_write_irq;
+
 	int			num_instances;
 	uint32_t                    srbm_soft_reset;
 	bool			has_page_queue;

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

@@ -69,6 +69,9 @@ void psp_securedisplay_parse_resp_status(struct psp_context *psp,
 	case TA_SECUREDISPLAY_STATUS__READ_CRC_ERROR:
 		dev_err(psp->adev->dev, "Secure display: Failed to Read CRC");
 		break;
+	case TA_SECUREDISPLAY_STATUS__I2C_INIT_ERROR:
+		dev_err(psp->adev->dev, "Secure display: Failed to initialize I2C.");
+		break;
 	default:
 		dev_err(psp->adev->dev, "Secure display: Failed to parse status: %d\n", status);
 	}
@@ -92,9 +95,7 @@ static ssize_t amdgpu_securedisplay_debugfs_write(struct file *f, const char __u
 	struct drm_device *dev = adev_to_drm(adev);
 	uint32_t phy_id;
 	uint32_t op;
-	int i;
 	char str[64];
-	char i2c_output[256];
 	int ret;
 
 	if (*pos || size > sizeof(str) - 1)
@@ -136,11 +137,9 @@ static ssize_t amdgpu_securedisplay_debugfs_write(struct file *f, const char __u
 		ret = psp_securedisplay_invoke(psp, TA_SECUREDISPLAY_COMMAND__SEND_ROI_CRC);
 		if (!ret) {
 			if (securedisplay_cmd->status == TA_SECUREDISPLAY_STATUS__SUCCESS) {
-				memset(i2c_output,  0, sizeof(i2c_output));
+				dev_info(adev->dev, "SECUREDISPLAY: I2C buffer out put is: %*ph\n",
-				for (i = 0; i < TA_SECUREDISPLAY_I2C_BUFFER_SIZE; i++)
+					 TA_SECUREDISPLAY_I2C_BUFFER_SIZE,
-					sprintf(i2c_output, "%s 0x%X", i2c_output,
+					 securedisplay_cmd->securedisplay_out_message.send_roi_crc.i2c_buf);
-						securedisplay_cmd->securedisplay_out_message.send_roi_crc.i2c_buf[i]);
-				dev_info(adev->dev, "SECUREDISPLAY: I2C buffer out put is :%s\n", i2c_output);
 			} else {
 				psp_securedisplay_parse_resp_status(psp, securedisplay_cmd->status);
 			}

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

@@ -28,6 +28,8 @@ struct amdgpu_smuio_funcs {
 	u32 (*get_rom_data_offset)(struct amdgpu_device *adev);
 	void (*update_rom_clock_gating)(struct amdgpu_device *adev, bool enable);
 	void (*get_clock_gating_state)(struct amdgpu_device *adev, u32 *flags);
+	u32 (*get_die_id)(struct amdgpu_device *adev);
+	bool (*is_host_gpu_xgmi_supported)(struct amdgpu_device *adev);
 };
 
 struct amdgpu_smuio {

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

@@ -62,6 +62,7 @@ static void amdgpu_do_test_moves(struct amdgpu_device *adev)
 	bp.flags = 0;
 	bp.type = ttm_bo_type_kernel;
 	bp.resv = NULL;
+	bp.bo_ptr_size = sizeof(struct amdgpu_bo);
 
 	r = amdgpu_bo_create(adev, &bp, &vram_obj);
 	if (r) {

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

@@ -47,7 +47,6 @@
 #include <drm/ttm/ttm_bo_driver.h>
 #include <drm/ttm/ttm_placement.h>
 
-#include <drm/drm_debugfs.h>
 #include <drm/amdgpu_drm.h>
 
 #include "amdgpu.h"
@@ -57,14 +56,15 @@
 #include "amdgpu_sdma.h"
 #include "amdgpu_ras.h"
 #include "amdgpu_atomfirmware.h"
+#include "amdgpu_res_cursor.h"
 #include "bif/bif_4_1_d.h"
 
 #define AMDGPU_TTM_VRAM_MAX_DW_READ	(size_t)128
 
-static int amdgpu_ttm_backend_bind(struct ttm_bo_device *bdev,
+static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
 				   struct ttm_tt *ttm,
 				   struct ttm_resource *bo_mem);
-static void amdgpu_ttm_backend_unbind(struct ttm_bo_device *bdev,
+static void amdgpu_ttm_backend_unbind(struct ttm_device *bdev,
 				      struct ttm_tt *ttm);
 
 static int amdgpu_ttm_init_on_chip(struct amdgpu_device *adev,
@@ -178,55 +178,12 @@ static int amdgpu_verify_access(struct ttm_buffer_object *bo, struct file *filp)
 					  filp->private_data);
 }
 
-/**
- * amdgpu_mm_node_addr - Compute the GPU relative offset of a GTT buffer.
- *
- * @bo: The bo to assign the memory to.
- * @mm_node: Memory manager node for drm allocator.
- * @mem: The region where the bo resides.
- *
- */
-static uint64_t amdgpu_mm_node_addr(struct ttm_buffer_object *bo,
-				    struct drm_mm_node *mm_node,
-				    struct ttm_resource *mem)
-{
-	uint64_t addr = 0;
-
-	if (mm_node->start != AMDGPU_BO_INVALID_OFFSET) {
-		addr = mm_node->start << PAGE_SHIFT;
-		addr += amdgpu_ttm_domain_start(amdgpu_ttm_adev(bo->bdev),
-						mem->mem_type);
-	}
-	return addr;
-}
-
-/**
- * amdgpu_find_mm_node - Helper function finds the drm_mm_node corresponding to
- * @offset. It also modifies the offset to be within the drm_mm_node returned
- *
- * @mem: The region where the bo resides.
- * @offset: The offset that drm_mm_node is used for finding.
- *
- */
-static struct drm_mm_node *amdgpu_find_mm_node(struct ttm_resource *mem,
-					       uint64_t *offset)
-{
-	struct drm_mm_node *mm_node = mem->mm_node;
-
-	while (*offset >= (mm_node->size << PAGE_SHIFT)) {
-		*offset -= (mm_node->size << PAGE_SHIFT);
-		++mm_node;
-	}
-	return mm_node;
-}
-
 /**
  * amdgpu_ttm_map_buffer - Map memory into the GART windows
  * @bo: buffer object to map
  * @mem: memory object to map
- * @mm_node: drm_mm node object to map
+ * @mm_cur: range to map
  * @num_pages: number of pages to map
- * @offset: offset into @mm_node where to start
  * @window: which GART window to use
  * @ring: DMA ring to use for the copy
  * @tmz: if we should setup a TMZ enabled mapping
@@ -237,10 +194,10 @@ static struct drm_mm_node *amdgpu_find_mm_node(struct ttm_resource *mem,
  */
 static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
 				 struct ttm_resource *mem,
-				 struct drm_mm_node *mm_node,
+				 struct amdgpu_res_cursor *mm_cur,
-				 unsigned num_pages, uint64_t offset,
+				 unsigned num_pages, unsigned window,
-				 unsigned window, struct amdgpu_ring *ring,
+				 struct amdgpu_ring *ring, bool tmz,
-				 bool tmz, uint64_t *addr)
+				 uint64_t *addr)
 {
 	struct amdgpu_device *adev = ring->adev;
 	struct amdgpu_job *job;
@@ -257,14 +214,15 @@ static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
 
 	/* Map only what can't be accessed directly */
 	if (!tmz && mem->start != AMDGPU_BO_INVALID_OFFSET) {
-		*addr = amdgpu_mm_node_addr(bo, mm_node, mem) + offset;
+		*addr = amdgpu_ttm_domain_start(adev, mem->mem_type) +
+			mm_cur->start;
 		return 0;
 	}
 
 	*addr = adev->gmc.gart_start;
 	*addr += (u64)window * AMDGPU_GTT_MAX_TRANSFER_SIZE *
 		AMDGPU_GPU_PAGE_SIZE;
-	*addr += offset & ~PAGE_MASK;
+	*addr += mm_cur->start & ~PAGE_MASK;
 
 	num_dw = ALIGN(adev->mman.buffer_funcs->copy_num_dw, 8);
 	num_bytes = num_pages * 8;
@@ -292,17 +250,17 @@ static int amdgpu_ttm_map_buffer(struct ttm_buffer_object *bo,
 	cpu_addr = &job->ibs[0].ptr[num_dw];
 
 	if (mem->mem_type == TTM_PL_TT) {
-		dma_addr_t *dma_address;
+		dma_addr_t *dma_addr;
 
-		dma_address = &bo->ttm->dma_address[offset >> PAGE_SHIFT];
+		dma_addr = &bo->ttm->dma_address[mm_cur->start >> PAGE_SHIFT];
-		r = amdgpu_gart_map(adev, 0, num_pages, dma_address, flags,
+		r = amdgpu_gart_map(adev, 0, num_pages, dma_addr, flags,
 				    cpu_addr);
 		if (r)
 			goto error_free;
 	} else {
 		dma_addr_t dma_address;
 
-		dma_address = (mm_node->start << PAGE_SHIFT) + offset;
+		dma_address = mm_cur->start;
 		dma_address += adev->vm_manager.vram_base_offset;
 
 		for (i = 0; i < num_pages; ++i) {
@@ -354,9 +312,8 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
 	const uint32_t GTT_MAX_BYTES = (AMDGPU_GTT_MAX_TRANSFER_SIZE *
 					AMDGPU_GPU_PAGE_SIZE);
 
-	uint64_t src_node_size, dst_node_size, src_offset, dst_offset;
 	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
-	struct drm_mm_node *src_mm, *dst_mm;
+	struct amdgpu_res_cursor src_mm, dst_mm;
 	struct dma_fence *fence = NULL;
 	int r = 0;
 
@@ -365,29 +322,13 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
 		return -EINVAL;
 	}
 
-	src_offset = src->offset;
+	amdgpu_res_first(src->mem, src->offset, size, &src_mm);
-	if (src->mem->mm_node) {
+	amdgpu_res_first(dst->mem, dst->offset, size, &dst_mm);
-		src_mm = amdgpu_find_mm_node(src->mem, &src_offset);
-		src_node_size = (src_mm->size << PAGE_SHIFT) - src_offset;
-	} else {
-		src_mm = NULL;
-		src_node_size = ULLONG_MAX;
-	}
-
-	dst_offset = dst->offset;
-	if (dst->mem->mm_node) {
-		dst_mm = amdgpu_find_mm_node(dst->mem, &dst_offset);
-		dst_node_size = (dst_mm->size << PAGE_SHIFT) - dst_offset;
-	} else {
-		dst_mm = NULL;
-		dst_node_size = ULLONG_MAX;
-	}
 
 	mutex_lock(&adev->mman.gtt_window_lock);
-
+	while (src_mm.remaining) {
-	while (size) {
+		uint32_t src_page_offset = src_mm.start & ~PAGE_MASK;
-		uint32_t src_page_offset = src_offset & ~PAGE_MASK;
+		uint32_t dst_page_offset = dst_mm.start & ~PAGE_MASK;
-		uint32_t dst_page_offset = dst_offset & ~PAGE_MASK;
 		struct dma_fence *next;
 		uint32_t cur_size;
 		uint64_t from, to;
@@ -396,19 +337,19 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
 		 * begins at an offset, then adjust the size accordingly
 		 */
 		cur_size = max(src_page_offset, dst_page_offset);
-		cur_size = min(min3(src_node_size, dst_node_size, size),
+		cur_size = min(min3(src_mm.size, dst_mm.size, size),
 			       (uint64_t)(GTT_MAX_BYTES - cur_size));
 
 		/* Map src to window 0 and dst to window 1. */
-		r = amdgpu_ttm_map_buffer(src->bo, src->mem, src_mm,
+		r = amdgpu_ttm_map_buffer(src->bo, src->mem, &src_mm,
 					  PFN_UP(cur_size + src_page_offset),
-					  src_offset, 0, ring, tmz, &from);
+					  0, ring, tmz, &from);
 		if (r)
 			goto error;
 
-		r = amdgpu_ttm_map_buffer(dst->bo, dst->mem, dst_mm,
+		r = amdgpu_ttm_map_buffer(dst->bo, dst->mem, &dst_mm,
 					  PFN_UP(cur_size + dst_page_offset),
-					  dst_offset, 1, ring, tmz, &to);
+					  1, ring, tmz, &to);
 		if (r)
 			goto error;
 
@@ -420,27 +361,8 @@ int amdgpu_ttm_copy_mem_to_mem(struct amdgpu_device *adev,
 		dma_fence_put(fence);
 		fence = next;
 
-		size -= cur_size;
+		amdgpu_res_next(&src_mm, cur_size);
-		if (!size)
+		amdgpu_res_next(&dst_mm, cur_size);
-			break;
-
-		src_node_size -= cur_size;
-		if (!src_node_size) {
-			++src_mm;
-			src_node_size = src_mm->size << PAGE_SHIFT;
-			src_offset = 0;
-		} else {
-			src_offset += cur_size;
-		}
-
-		dst_node_size -= cur_size;
-		if (!dst_node_size) {
-			++dst_mm;
-			dst_node_size = dst_mm->size << PAGE_SHIFT;
-			dst_offset = 0;
-		} else {
-			dst_offset += cur_size;
-		}
 	}
 error:
 	mutex_unlock(&adev->mman.gtt_window_lock);
@@ -519,7 +441,8 @@ error:
 static bool amdgpu_mem_visible(struct amdgpu_device *adev,
 			       struct ttm_resource *mem)
 {
-	struct drm_mm_node *nodes = mem->mm_node;
+	uint64_t mem_size = (u64)mem->num_pages << PAGE_SHIFT;
+	struct amdgpu_res_cursor cursor;
 
 	if (mem->mem_type == TTM_PL_SYSTEM ||
 	    mem->mem_type == TTM_PL_TT)
@@ -527,12 +450,13 @@ static bool amdgpu_mem_visible(struct amdgpu_device *adev,
 	if (mem->mem_type != TTM_PL_VRAM)
 		return false;
 
+	amdgpu_res_first(mem, 0, mem_size, &cursor);
+
 	/* ttm_resource_ioremap only supports contiguous memory */
-	if (nodes->size != mem->num_pages)
+	if (cursor.size != mem_size)
 		return false;
 
-	return ((nodes->start + nodes->size) << PAGE_SHIFT)
+	return cursor.start + cursor.size <= adev->gmc.visible_vram_size;
-		<= adev->gmc.visible_vram_size;
 }
 
 /*
@@ -646,7 +570,7 @@ out:
  *
  * Called by ttm_mem_io_reserve() ultimately via ttm_bo_vm_fault()
  */
-static int amdgpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resource *mem)
+static int amdgpu_ttm_io_mem_reserve(struct ttm_device *bdev, struct ttm_resource *mem)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct drm_mm_node *mm_node = mem->mm_node;
@@ -674,7 +598,10 @@ static int amdgpu_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_reso
 
 		mem->bus.offset += adev->gmc.aper_base;
 		mem->bus.is_iomem = true;
-		mem->bus.caching = ttm_write_combined;
+		if (adev->gmc.xgmi.connected_to_cpu)
+			mem->bus.caching = ttm_cached;
+		else
+			mem->bus.caching = ttm_write_combined;
 		break;
 	default:
 		return -EINVAL;
@@ -686,12 +613,10 @@ static unsigned long amdgpu_ttm_io_mem_pfn(struct ttm_buffer_object *bo,
 					   unsigned long page_offset)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev);
-	uint64_t offset = (page_offset << PAGE_SHIFT);
+	struct amdgpu_res_cursor cursor;
-	struct drm_mm_node *mm;
 
-	mm = amdgpu_find_mm_node(&bo->mem, &offset);
+	amdgpu_res_first(&bo->mem, (u64)page_offset << PAGE_SHIFT, 0, &cursor);
-	offset += adev->gmc.aper_base;
+	return (adev->gmc.aper_base + cursor.start) >> PAGE_SHIFT;
-	return mm->start + (offset >> PAGE_SHIFT);
 }
 
 /**
@@ -893,16 +818,15 @@ void amdgpu_ttm_tt_set_user_pages(struct ttm_tt *ttm, struct page **pages)
  *
  * Called by amdgpu_ttm_backend_bind()
  **/
-static int amdgpu_ttm_tt_pin_userptr(struct ttm_bo_device *bdev,
+static int amdgpu_ttm_tt_pin_userptr(struct ttm_device *bdev,
 				     struct ttm_tt *ttm)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-	int r;
-
 	int write = !(gtt->userflags & AMDGPU_GEM_USERPTR_READONLY);
 	enum dma_data_direction direction = write ?
 		DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
+	int r;
 
 	/* Allocate an SG array and squash pages into it */
 	r = sg_alloc_table_from_pages(ttm->sg, ttm->pages, ttm->num_pages, 0,
@@ -931,18 +855,17 @@ release_sg:
 /*
  * amdgpu_ttm_tt_unpin_userptr - Unpin and unmap userptr pages
  */
-static void amdgpu_ttm_tt_unpin_userptr(struct ttm_bo_device *bdev,
+static void amdgpu_ttm_tt_unpin_userptr(struct ttm_device *bdev,
 					struct ttm_tt *ttm)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
-
 	int write = !(gtt->userflags & AMDGPU_GEM_USERPTR_READONLY);
 	enum dma_data_direction direction = write ?
 		DMA_BIDIRECTIONAL : DMA_TO_DEVICE;
 
 	/* double check that we don't free the table twice */
-	if (!ttm->sg->sgl)
+	if (!ttm->sg || !ttm->sg->sgl)
 		return;
 
 	/* unmap the pages mapped to the device */
@@ -1015,7 +938,7 @@ gart_bind_fail:
  * Called by ttm_tt_bind() on behalf of ttm_bo_handle_move_mem().
  * This handles binding GTT memory to the device address space.
  */
-static int amdgpu_ttm_backend_bind(struct ttm_bo_device *bdev,
+static int amdgpu_ttm_backend_bind(struct ttm_device *bdev,
 				   struct ttm_tt *ttm,
 				   struct ttm_resource *bo_mem)
 {
@@ -1155,20 +1078,20 @@ int amdgpu_ttm_recover_gart(struct ttm_buffer_object *tbo)
  * Called by ttm_tt_unbind() on behalf of ttm_bo_move_ttm() and
  * ttm_tt_destroy().
  */
-static void amdgpu_ttm_backend_unbind(struct ttm_bo_device *bdev,
+static void amdgpu_ttm_backend_unbind(struct ttm_device *bdev,
 				      struct ttm_tt *ttm)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bdev);
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
 	int r;
 
-	if (!gtt->bound)
-		return;
-
 	/* if the pages have userptr pinning then clear that first */
 	if (gtt->userptr)
 		amdgpu_ttm_tt_unpin_userptr(bdev, ttm);
 
+	if (!gtt->bound)
+		return;
+
 	if (gtt->offset == AMDGPU_BO_INVALID_OFFSET)
 		return;
 
@@ -1180,7 +1103,7 @@ static void amdgpu_ttm_backend_unbind(struct ttm_bo_device *bdev,
 	gtt->bound = false;
 }
 
-static void amdgpu_ttm_backend_destroy(struct ttm_bo_device *bdev,
+static void amdgpu_ttm_backend_destroy(struct ttm_device *bdev,
 				       struct ttm_tt *ttm)
 {
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
@@ -1234,7 +1157,7 @@ static struct ttm_tt *amdgpu_ttm_tt_create(struct ttm_buffer_object *bo,
  * Map the pages of a ttm_tt object to an address space visible
  * to the underlying device.
  */
-static int amdgpu_ttm_tt_populate(struct ttm_bo_device *bdev,
+static int amdgpu_ttm_tt_populate(struct ttm_device *bdev,
 				  struct ttm_tt *ttm,
 				  struct ttm_operation_ctx *ctx)
 {
@@ -1278,7 +1201,7 @@ static int amdgpu_ttm_tt_populate(struct ttm_bo_device *bdev,
  * Unmaps pages of a ttm_tt object from the device address space and
  * unpopulates the page array backing it.
  */
-static void amdgpu_ttm_tt_unpopulate(struct ttm_bo_device *bdev,
+static void amdgpu_ttm_tt_unpopulate(struct ttm_device *bdev,
 				     struct ttm_tt *ttm)
 {
 	struct amdgpu_ttm_tt *gtt = (void *)ttm;
@@ -1430,6 +1353,10 @@ uint64_t amdgpu_ttm_tt_pde_flags(struct ttm_tt *ttm, struct ttm_resource *mem)
 			flags |= AMDGPU_PTE_SNOOPED;
 	}
 
+	if (mem && mem->mem_type == TTM_PL_VRAM &&
+			mem->bus.caching == ttm_cached)
+		flags |= AMDGPU_PTE_SNOOPED;
+
 	return flags;
 }
 
@@ -1469,7 +1396,7 @@ static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
 					    const struct ttm_place *place)
 {
 	unsigned long num_pages = bo->mem.num_pages;
-	struct drm_mm_node *node = bo->mem.mm_node;
+	struct amdgpu_res_cursor cursor;
 	struct dma_resv_list *flist;
 	struct dma_fence *f;
 	int i;
@@ -1501,13 +1428,15 @@ static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
 
 	case TTM_PL_VRAM:
 		/* Check each drm MM node individually */
-		while (num_pages) {
+		amdgpu_res_first(&bo->mem, 0, (u64)num_pages << PAGE_SHIFT,
-			if (place->fpfn < (node->start + node->size) &&
+				 &cursor);
-			    !(place->lpfn && place->lpfn <= node->start))
+		while (cursor.remaining) {
+			if (place->fpfn < PFN_DOWN(cursor.start + cursor.size)
+			    && !(place->lpfn &&
+				 place->lpfn <= PFN_DOWN(cursor.start)))
 				return true;
 
-			num_pages -= node->size;
+			amdgpu_res_next(&cursor, cursor.size);
-			++node;
 		}
 		return false;
 
@@ -1531,41 +1460,36 @@ static bool amdgpu_ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
  * access for debugging purposes.
  */
 static int amdgpu_ttm_access_memory(struct ttm_buffer_object *bo,
-				    unsigned long offset,
+				    unsigned long offset, void *buf, int len,
-				    void *buf, int len, int write)
+				    int write)
 {
 	struct amdgpu_bo *abo = ttm_to_amdgpu_bo(bo);
 	struct amdgpu_device *adev = amdgpu_ttm_adev(abo->tbo.bdev);
-	struct drm_mm_node *nodes;
+	struct amdgpu_res_cursor cursor;
+	unsigned long flags;
 	uint32_t value = 0;
 	int ret = 0;
-	uint64_t pos;
-	unsigned long flags;
 
 	if (bo->mem.mem_type != TTM_PL_VRAM)
 		return -EIO;
 
-	pos = offset;
+	amdgpu_res_first(&bo->mem, offset, len, &cursor);
-	nodes = amdgpu_find_mm_node(&abo->tbo.mem, &pos);
+	while (cursor.remaining) {
-	pos += (nodes->start << PAGE_SHIFT);
+		uint64_t aligned_pos = cursor.start & ~(uint64_t)3;
-
+		uint64_t bytes = 4 - (cursor.start & 3);
-	while (len && pos < adev->gmc.mc_vram_size) {
+		uint32_t shift = (cursor.start & 3) * 8;
-		uint64_t aligned_pos = pos & ~(uint64_t)3;
-		uint64_t bytes = 4 - (pos & 3);
-		uint32_t shift = (pos & 3) * 8;
 		uint32_t mask = 0xffffffff << shift;
 
-		if (len < bytes) {
+		if (cursor.size < bytes) {
-			mask &= 0xffffffff >> (bytes - len) * 8;
+			mask &= 0xffffffff >> (bytes - cursor.size) * 8;
-			bytes = len;
+			bytes = cursor.size;
 		}
 
 		if (mask != 0xffffffff) {
 			spin_lock_irqsave(&adev->mmio_idx_lock, flags);
 			WREG32_NO_KIQ(mmMM_INDEX, ((uint32_t)aligned_pos) | 0x80000000);
 			WREG32_NO_KIQ(mmMM_INDEX_HI, aligned_pos >> 31);
-			if (!write || mask != 0xffffffff)
+			value = RREG32_NO_KIQ(mmMM_DATA);
-				value = RREG32_NO_KIQ(mmMM_DATA);
 			if (write) {
 				value &= ~mask;
 				value |= (*(uint32_t *)buf << shift) & mask;
@@ -1577,21 +1501,15 @@ static int amdgpu_ttm_access_memory(struct ttm_buffer_object *bo,
 				memcpy(buf, &value, bytes);
 			}
 		} else {
-			bytes = (nodes->start + nodes->size) << PAGE_SHIFT;
+			bytes = cursor.size & ~0x3ULL;
-			bytes = min(bytes - pos, (uint64_t)len & ~0x3ull);
+			amdgpu_device_vram_access(adev, cursor.start,
-
+						  (uint32_t *)buf, bytes,
-			amdgpu_device_vram_access(adev, pos, (uint32_t *)buf,
+						  write);
-						  bytes, write);
 		}
 
 		ret += bytes;
 		buf = (uint8_t *)buf + bytes;
-		pos += bytes;
+		amdgpu_res_next(&cursor, bytes);
-		len -= bytes;
-		if (pos >= (nodes->start + nodes->size) << PAGE_SHIFT) {
-			++nodes;
-			pos = (nodes->start << PAGE_SHIFT);
-		}
 	}
 
 	return ret;
@@ -1603,7 +1521,7 @@ amdgpu_bo_delete_mem_notify(struct ttm_buffer_object *bo)
 	amdgpu_bo_move_notify(bo, false, NULL);
 }
 
-static struct ttm_bo_driver amdgpu_bo_driver = {
+static struct ttm_device_funcs amdgpu_bo_driver = {
 	.ttm_tt_create = &amdgpu_ttm_tt_create,
 	.ttm_tt_populate = &amdgpu_ttm_tt_populate,
 	.ttm_tt_unpopulate = &amdgpu_ttm_tt_unpopulate,
@@ -1696,7 +1614,7 @@ static void amdgpu_ttm_training_data_block_init(struct amdgpu_device *adev)
 		(adev->gmc.mc_vram_size - GDDR6_MEM_TRAINING_OFFSET);
 	ctx->train_data_size =
 		GDDR6_MEM_TRAINING_DATA_SIZE_IN_BYTES;
-	
+
 	DRM_DEBUG("train_data_size:%llx,p2c_train_data_offset:%llx,c2p_train_data_offset:%llx.\n",
 			ctx->train_data_size,
 			ctx->p2c_train_data_offset,
@@ -1785,7 +1703,7 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	mutex_init(&adev->mman.gtt_window_lock);
 
 	/* No others user of address space so set it to 0 */
-	r = ttm_bo_device_init(&adev->mman.bdev, &amdgpu_bo_driver, adev->dev,
+	r = ttm_device_init(&adev->mman.bdev, &amdgpu_bo_driver, adev->dev,
 			       adev_to_drm(adev)->anon_inode->i_mapping,
 			       adev_to_drm(adev)->vma_offset_manager,
 			       adev->need_swiotlb,
@@ -1812,8 +1730,15 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	/* Change the size here instead of the init above so only lpfn is affected */
 	amdgpu_ttm_set_buffer_funcs_status(adev, false);
 #ifdef CONFIG_64BIT
-	adev->mman.aper_base_kaddr = ioremap_wc(adev->gmc.aper_base,
+#ifdef CONFIG_X86
-						adev->gmc.visible_vram_size);
+	if (adev->gmc.xgmi.connected_to_cpu)
+		adev->mman.aper_base_kaddr = ioremap_cache(adev->gmc.aper_base,
+				adev->gmc.visible_vram_size);
+
+	else
+#endif
+		adev->mman.aper_base_kaddr = ioremap_wc(adev->gmc.aper_base,
+				adev->gmc.visible_vram_size);
 #endif
 
 	/*
@@ -1926,7 +1851,7 @@ void amdgpu_ttm_fini(struct amdgpu_device *adev)
 	ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_GDS);
 	ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_GWS);
 	ttm_range_man_fini(&adev->mman.bdev, AMDGPU_PL_OA);
-	ttm_bo_device_release(&adev->mman.bdev);
+	ttm_device_fini(&adev->mman.bdev);
 	adev->mman.initialized = false;
 	DRM_INFO("amdgpu: ttm finalized\n");
 }
@@ -2002,7 +1927,7 @@ unlock:
 	return ret;
 }
 
-static struct vm_operations_struct amdgpu_ttm_vm_ops = {
+static const struct vm_operations_struct amdgpu_ttm_vm_ops = {
 	.fault = amdgpu_ttm_fault,
 	.open = ttm_bo_vm_open,
 	.close = ttm_bo_vm_close,
@@ -2053,7 +1978,8 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, uint64_t src_offset,
 		return r;
 
 	if (vm_needs_flush) {
-		job->vm_pd_addr = amdgpu_gmc_pd_addr(adev->gart.bo);
+		job->vm_pd_addr = amdgpu_gmc_pd_addr(adev->gmc.pdb0_bo ?
+					adev->gmc.pdb0_bo : adev->gart.bo);
 		job->vm_needs_flush = true;
 	}
 	if (resv) {
@@ -2104,9 +2030,9 @@ int amdgpu_fill_buffer(struct amdgpu_bo *bo,
 	uint32_t max_bytes = adev->mman.buffer_funcs->fill_max_bytes;
 	struct amdgpu_ring *ring = adev->mman.buffer_funcs_ring;
 
-	struct drm_mm_node *mm_node;
+	struct amdgpu_res_cursor cursor;
-	unsigned long num_pages;
 	unsigned int num_loops, num_dw;
+	uint64_t num_bytes;
 
 	struct amdgpu_job *job;
 	int r;
@@ -2122,15 +2048,13 @@ int amdgpu_fill_buffer(struct amdgpu_bo *bo,
 			return r;
 	}
 
-	num_pages = bo->tbo.mem.num_pages;
+	num_bytes = bo->tbo.mem.num_pages << PAGE_SHIFT;
-	mm_node = bo->tbo.mem.mm_node;
 	num_loops = 0;
-	while (num_pages) {
-		uint64_t byte_count = mm_node->size << PAGE_SHIFT;
 
-		num_loops += DIV_ROUND_UP_ULL(byte_count, max_bytes);
+	amdgpu_res_first(&bo->tbo.mem, 0, num_bytes, &cursor);
-		num_pages -= mm_node->size;
+	while (cursor.remaining) {
-		++mm_node;
+		num_loops += DIV_ROUND_UP_ULL(cursor.size, max_bytes);
+		amdgpu_res_next(&cursor, cursor.size);
 	}
 	num_dw = num_loops * adev->mman.buffer_funcs->fill_num_dw;
 
@@ -2152,27 +2076,16 @@ int amdgpu_fill_buffer(struct amdgpu_bo *bo,
 		}
 	}
 
-	num_pages = bo->tbo.mem.num_pages;
+	amdgpu_res_first(&bo->tbo.mem, 0, num_bytes, &cursor);
-	mm_node = bo->tbo.mem.mm_node;
+	while (cursor.remaining) {
+		uint32_t cur_size = min_t(uint64_t, cursor.size, max_bytes);
+		uint64_t dst_addr = cursor.start;
 
-	while (num_pages) {
+		dst_addr += amdgpu_ttm_domain_start(adev, bo->tbo.mem.mem_type);
-		uint64_t byte_count = mm_node->size << PAGE_SHIFT;
+		amdgpu_emit_fill_buffer(adev, &job->ibs[0], src_data, dst_addr,
-		uint64_t dst_addr;
+					cur_size);
 
-		dst_addr = amdgpu_mm_node_addr(&bo->tbo, mm_node, &bo->tbo.mem);
+		amdgpu_res_next(&cursor, cur_size);
-		while (byte_count) {
-			uint32_t cur_size_in_bytes = min_t(uint64_t, byte_count,
-							   max_bytes);
-
-			amdgpu_emit_fill_buffer(adev, &job->ibs[0], src_data,
-						dst_addr, cur_size_in_bytes);
-
-			dst_addr += cur_size_in_bytes;
-			byte_count -= cur_size_in_bytes;
-		}
-
-		num_pages -= mm_node->size;
-		++mm_node;
 	}
 
 	amdgpu_ring_pad_ib(ring, &job->ibs[0]);
@@ -2191,36 +2104,74 @@ error_free:
 
 #if defined(CONFIG_DEBUG_FS)
 
-static int amdgpu_mm_dump_table(struct seq_file *m, void *data)
+static int amdgpu_mm_vram_table_show(struct seq_file *m, void *unused)
 {
-	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	unsigned ttm_pl = (uintptr_t)node->info_ent->data;
+	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
-	struct drm_device *dev = node->minor->dev;
+							    TTM_PL_VRAM);
-	struct amdgpu_device *adev = drm_to_adev(dev);
-	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev, ttm_pl);
 	struct drm_printer p = drm_seq_file_printer(m);
 
 	man->func->debug(man, &p);
 	return 0;
 }
 
-static int amdgpu_ttm_pool_debugfs(struct seq_file *m, void *data)
+static int amdgpu_ttm_page_pool_show(struct seq_file *m, void *unused)
 {
-	struct drm_info_node *node = (struct drm_info_node *)m->private;
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct amdgpu_device *adev = drm_to_adev(dev);
 
 	return ttm_pool_debugfs(&adev->mman.bdev.pool, m);
 }
 
-static const struct drm_info_list amdgpu_ttm_debugfs_list[] = {
+static int amdgpu_mm_tt_table_show(struct seq_file *m, void *unused)
-	{"amdgpu_vram_mm", amdgpu_mm_dump_table, 0, (void *)TTM_PL_VRAM},
+{
-	{"amdgpu_gtt_mm", amdgpu_mm_dump_table, 0, (void *)TTM_PL_TT},
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
-	{"amdgpu_gds_mm", amdgpu_mm_dump_table, 0, (void *)AMDGPU_PL_GDS},
+	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
-	{"amdgpu_gws_mm", amdgpu_mm_dump_table, 0, (void *)AMDGPU_PL_GWS},
+							    TTM_PL_TT);
-	{"amdgpu_oa_mm", amdgpu_mm_dump_table, 0, (void *)AMDGPU_PL_OA},
+	struct drm_printer p = drm_seq_file_printer(m);
-	{"ttm_page_pool", amdgpu_ttm_pool_debugfs, 0, NULL},
+
-};
+	man->func->debug(man, &p);
+	return 0;
+}
+
+static int amdgpu_mm_gds_table_show(struct seq_file *m, void *unused)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
+							    AMDGPU_PL_GDS);
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	man->func->debug(man, &p);
+	return 0;
+}
+
+static int amdgpu_mm_gws_table_show(struct seq_file *m, void *unused)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
+							    AMDGPU_PL_GWS);
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	man->func->debug(man, &p);
+	return 0;
+}
+
+static int amdgpu_mm_oa_table_show(struct seq_file *m, void *unused)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct ttm_resource_manager *man = ttm_manager_type(&adev->mman.bdev,
+							    AMDGPU_PL_OA);
+	struct drm_printer p = drm_seq_file_printer(m);
+
+	man->func->debug(man, &p);
+	return 0;
+}
+
+DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_vram_table);
+DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_tt_table);
+DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_gds_table);
+DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_gws_table);
+DEFINE_SHOW_ATTRIBUTE(amdgpu_mm_oa_table);
+DEFINE_SHOW_ATTRIBUTE(amdgpu_ttm_page_pool);
 
 /*
  * amdgpu_ttm_vram_read - Linear read access to VRAM
@@ -2308,58 +2259,6 @@ static const struct file_operations amdgpu_ttm_vram_fops = {
 	.llseek = default_llseek,
 };
 
-#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
-
-/*
- * amdgpu_ttm_gtt_read - Linear read access to GTT memory
- */
-static ssize_t amdgpu_ttm_gtt_read(struct file *f, char __user *buf,
-				   size_t size, loff_t *pos)
-{
-	struct amdgpu_device *adev = file_inode(f)->i_private;
-	ssize_t result = 0;
-	int r;
-
-	while (size) {
-		loff_t p = *pos / PAGE_SIZE;
-		unsigned off = *pos & ~PAGE_MASK;
-		size_t cur_size = min_t(size_t, size, PAGE_SIZE - off);
-		struct page *page;
-		void *ptr;
-
-		if (p >= adev->gart.num_cpu_pages)
-			return result;
-
-		page = adev->gart.pages[p];
-		if (page) {
-			ptr = kmap(page);
-			ptr += off;
-
-			r = copy_to_user(buf, ptr, cur_size);
-			kunmap(adev->gart.pages[p]);
-		} else
-			r = clear_user(buf, cur_size);
-
-		if (r)
-			return -EFAULT;
-
-		result += cur_size;
-		buf += cur_size;
-		*pos += cur_size;
-		size -= cur_size;
-	}
-
-	return result;
-}
-
-static const struct file_operations amdgpu_ttm_gtt_fops = {
-	.owner = THIS_MODULE,
-	.read = amdgpu_ttm_gtt_read,
-	.llseek = default_llseek
-};
-
-#endif
-
 /*
  * amdgpu_iomem_read - Virtual read access to GPU mapped memory
  *
@@ -2474,46 +2373,29 @@ static const struct file_operations amdgpu_ttm_iomem_fops = {
 	.llseek = default_llseek
 };
 
-static const struct {
-	char *name;
-	const struct file_operations *fops;
-	int domain;
-} ttm_debugfs_entries[] = {
-	{ "amdgpu_vram", &amdgpu_ttm_vram_fops, TTM_PL_VRAM },
-#ifdef CONFIG_DRM_AMDGPU_GART_DEBUGFS
-	{ "amdgpu_gtt", &amdgpu_ttm_gtt_fops, TTM_PL_TT },
-#endif
-	{ "amdgpu_iomem", &amdgpu_ttm_iomem_fops, TTM_PL_SYSTEM },
-};
-
 #endif
 
-int amdgpu_ttm_debugfs_init(struct amdgpu_device *adev)
+void amdgpu_ttm_debugfs_init(struct amdgpu_device *adev)
 {
 #if defined(CONFIG_DEBUG_FS)
-	unsigned count;
-
 	struct drm_minor *minor = adev_to_drm(adev)->primary;
-	struct dentry *ent, *root = minor->debugfs_root;
+	struct dentry *root = minor->debugfs_root;
 
-	for (count = 0; count < ARRAY_SIZE(ttm_debugfs_entries); count++) {
+	debugfs_create_file_size("amdgpu_vram", 0444, root, adev,
-		ent = debugfs_create_file(
+				 &amdgpu_ttm_vram_fops, adev->gmc.mc_vram_size);
-				ttm_debugfs_entries[count].name,
+	debugfs_create_file("amdgpu_iomem", 0444, root, adev,
-				S_IFREG | S_IRUGO, root,
+			    &amdgpu_ttm_iomem_fops);
-				adev,
+	debugfs_create_file("amdgpu_vram_mm", 0444, root, adev,
-				ttm_debugfs_entries[count].fops);
+			    &amdgpu_mm_vram_table_fops);
-		if (IS_ERR(ent))
+	debugfs_create_file("amdgpu_gtt_mm", 0444, root, adev,
-			return PTR_ERR(ent);
+			    &amdgpu_mm_tt_table_fops);
-		if (ttm_debugfs_entries[count].domain == TTM_PL_VRAM)
+	debugfs_create_file("amdgpu_gds_mm", 0444, root, adev,
-			i_size_write(ent->d_inode, adev->gmc.mc_vram_size);
+			    &amdgpu_mm_gds_table_fops);
-		else if (ttm_debugfs_entries[count].domain == TTM_PL_TT)
+	debugfs_create_file("amdgpu_gws_mm", 0444, root, adev,
-			i_size_write(ent->d_inode, adev->gmc.gart_size);
+			    &amdgpu_mm_gws_table_fops);
-		adev->mman.debugfs_entries[count] = ent;
+	debugfs_create_file("amdgpu_oa_mm", 0444, root, adev,
-	}
+			    &amdgpu_mm_oa_table_fops);
-
+	debugfs_create_file("ttm_page_pool", 0444, root, adev,
-	count = ARRAY_SIZE(amdgpu_ttm_debugfs_list);
+			    &amdgpu_ttm_page_pool_fops);
-	return amdgpu_debugfs_add_files(adev, amdgpu_ttm_debugfs_list, count);
-#else
-	return 0;
 #endif
 }

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

@@ -60,14 +60,10 @@ struct amdgpu_gtt_mgr {
 };
 
 struct amdgpu_mman {
-	struct ttm_bo_device		bdev;
+	struct ttm_device		bdev;
 	bool				initialized;
 	void __iomem			*aper_base_kaddr;
 
-#if defined(CONFIG_DEBUG_FS)
-	struct dentry			*debugfs_entries[8];
-#endif
-
 	/* buffer handling */
 	const struct amdgpu_buffer_funcs	*buffer_funcs;
 	struct amdgpu_ring			*buffer_funcs_ring;
@@ -119,8 +115,7 @@ int amdgpu_vram_mgr_alloc_sgt(struct amdgpu_device *adev,
 			      struct device *dev,
 			      enum dma_data_direction dir,
 			      struct sg_table **sgt);
-void amdgpu_vram_mgr_free_sgt(struct amdgpu_device *adev,
+void amdgpu_vram_mgr_free_sgt(struct device *dev,
-			      struct device *dev,
 			      enum dma_data_direction dir,
 			      struct sg_table *sgt);
 uint64_t amdgpu_vram_mgr_usage(struct ttm_resource_manager *man);
@@ -186,6 +181,6 @@ uint64_t amdgpu_ttm_tt_pde_flags(struct ttm_tt *ttm, struct ttm_resource *mem);
 uint64_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm,
 				 struct ttm_resource *mem);
 
-int amdgpu_ttm_debugfs_init(struct amdgpu_device *adev);
+void amdgpu_ttm_debugfs_init(struct amdgpu_device *adev);
 
 #endif