diff options
1145 files changed, 182111 insertions, 31418 deletions
diff --git a/Documentation/core-api/printk-formats.rst b/Documentation/core-api/printk-formats.rst index 9be6de402cb9..f063a384c7c8 100644 --- a/Documentation/core-api/printk-formats.rst +++ b/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 ====== diff --git a/Documentation/devicetree/bindings/display/allwinner,sun8i-a83t-dw-hdmi.yaml b/Documentation/devicetree/bindings/display/allwinner,sun8i-a83t-dw-hdmi.yaml index b3e9992525c2..907fb47cc84a 100644 --- a/Documentation/devicetree/bindings/display/allwinner,sun8i-a83t-dw-hdmi.yaml +++ b/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 - the following device-specific properties. + in bridge/synopsys,dw-hdmi.yaml with the following device-specific + properties. maintainers: - Chen-Yu Tsai <wens@csie.org> diff --git a/Documentation/devicetree/bindings/display/brcm,bcm2711-hdmi.yaml b/Documentation/devicetree/bindings/display/brcm,bcm2711-hdmi.yaml index a1d5a32660e0..57324a5f0271 100644 --- a/Documentation/devicetree/bindings/display/brcm,bcm2711-hdmi.yaml +++ b/Documentation/devicetree/bindings/display/brcm,bcm2711-hdmi.yaml @@ -109,7 +109,7 @@ required: - resets - ddc -additionalProperties: false +unevaluatedProperties: false examples: - | diff --git a/Documentation/devicetree/bindings/display/bridge/analogix,anx7625.yaml b/Documentation/devicetree/bindings/display/bridge/analogix,anx7625.yaml index c789784efe30..ab48ab2f4240 100644 --- a/Documentation/devicetree/bindings/display/bridge/analogix,anx7625.yaml +++ b/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: $ref: /schemas/graph.yaml#/properties/ports @@ -55,6 +64,9 @@ properties: required: - compatible - reg + - vdd10-supply + - vdd18-supply + - vdd33-supply - ports additionalProperties: false @@ -72,6 +84,9 @@ examples: reg = <0x58>; enable-gpios = <&pio 45 GPIO_ACTIVE_HIGH>; reset-gpios = <&pio 73 GPIO_ACTIVE_HIGH>; + vdd10-supply = <&pp1000_mipibrdg>; + vdd18-supply = <&pp1800_mipibrdg>; + vdd33-supply = <&pp3300_mipibrdg>; ports { #address-cells = <1>; diff --git a/Documentation/devicetree/bindings/display/bridge/chipone,icn6211.yaml b/Documentation/devicetree/bindings/display/bridge/chipone,icn6211.yaml new file mode 100644 index 000000000000..62c3bd4cb28d --- /dev/null +++ b/Documentation/devicetree/bindings/display/bridge/chipone,icn6211.yaml @@ -0,0 +1,99 @@ +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +%YAML 1.2 +--- +$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>; + }; + }; + }; + }; + }; diff --git a/Documentation/devicetree/bindings/display/bridge/dw_hdmi.txt b/Documentation/devicetree/bindings/display/bridge/dw_hdmi.txt deleted file mode 100644 index 33bf981fbe33..000000000000 --- a/Documentation/devicetree/bindings/display/bridge/dw_hdmi.txt +++ /dev/null @@ -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. diff --git a/Documentation/devicetree/bindings/display/bridge/lontium,lt8912b.yaml b/Documentation/devicetree/bindings/display/bridge/lontium,lt8912b.yaml new file mode 100644 index 000000000000..735d0233a7d6 --- /dev/null +++ b/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>; + }; + }; + }; + }; + }; + +... diff --git a/Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.txt b/Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.txt deleted file mode 100644 index 3f6072651182..000000000000 --- a/Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.txt +++ /dev/null @@ -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>; - }; - }; - }; diff --git a/Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.yaml b/Documentation/devicetree/bindings/display/bridge/renesas,dw-hdmi.yaml new file mode 100644 index 000000000000..0c9785c8db51 --- /dev/null +++ b/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: + $ref: /schemas/graph.yaml#/properties/port + description: Sound input port + + required: + - port@0 + - port@1 + - port@2 + + power-domains: + maxItems: 1 + +required: + - compatible + - reg + - clocks + - clock-names + - interrupts + - ports + +unevaluatedProperties: false + +examples: + - | + #include <dt-bindings/clock/r8a7795-cpg-mssr.h> + #include <dt-bindings/interrupt-controller/irq.h> + #include <dt-bindings/power/r8a7795-sysc.h> + + hdmi@fead0000 { + compatible = "renesas,r8a7795-hdmi", "renesas,rcar-gen3-hdmi"; + reg = <0xfead0000 0x10000>; + interrupts = <0 389 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&cpg CPG_CORE R8A7795_CLK_S0D4>, <&cpg CPG_MOD 729>; + clock-names = "iahb", "isfr"; + power-domains = <&sysc R8A7795_PD_ALWAYS_ON>; + + ports { + #address-cells = <1>; + #size-cells = <0>; + port@0 { + reg = <0>; + dw_hdmi0_in: endpoint { + remote-endpoint = <&du_out_hdmi0>; + }; + }; + port@1 { + reg = <1>; + rcar_dw_hdmi0_out: endpoint { + remote-endpoint = <&hdmi0_con>; + }; + }; + port@2 { + reg = <2>; + rcar_dw_hdmi0_sound_in: endpoint { + remote-endpoint = <&hdmi_sound_out>; + }; + }; + }; + }; + + hdmi0-out { + compatible = "hdmi-connector"; + label = "HDMI0 OUT"; + type = "a"; + + port { + hdmi0_con: endpoint { + remote-endpoint = <&rcar_dw_hdmi0_out>; + }; + }; + }; + +... diff --git a/Documentation/devicetree/bindings/display/bridge/synopsys,dw-hdmi.yaml b/Documentation/devicetree/bindings/display/bridge/synopsys,dw-hdmi.yaml new file mode 100644 index 000000000000..9be44a682e67 --- /dev/null +++ b/Documentation/devicetree/bindings/display/bridge/synopsys,dw-hdmi.yaml @@ -0,0 +1,55 @@ +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +%YAML 1.2 +--- +$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: + - $ref: /schemas/types.yaml#/definitions/uint32 + - enum: [1, 4] + default: 1 + + clocks: + minItems: 2 + maxItems: 5 + items: + - description: The bus clock for either AHB and APB + - description: The internal register configuration clock + additionalItems: true + + clock-names: + minItems: 2 + maxItems: 5 + items: + - const: iahb + - const: isfr + additionalItems: true + + interrupts: + maxItems: 1 + +additionalProperties: true + +... diff --git a/Documentation/devicetree/bindings/display/fsl,lcdif.yaml b/Documentation/devicetree/bindings/display/fsl,lcdif.yaml new file mode 100644 index 000000000000..a4c3064c778c --- /dev/null +++ b/Documentation/devicetree/bindings/display/fsl,lcdif.yaml @@ -0,0 +1,110 @@ +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +%YAML 1.2 +--- +$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: + $ref: /schemas/graph.yaml#/properties/port + description: The LCDIF output port + +required: + - compatible + - reg + - clocks + - interrupts + - port + +additionalProperties: false + +allOf: + - if: + properties: + compatible: + contains: + const: fsl,imx6sx-lcdif + then: + properties: + clocks: + minItems: 2 + maxItems: 3 + clock-names: + minItems: 2 + maxItems: 3 + required: + - clock-names + else: + properties: + clocks: + maxItems: 1 + clock-names: + maxItems: 1 + +examples: + - | + #include <dt-bindings/clock/imx6sx-clock.h> + #include <dt-bindings/interrupt-controller/arm-gic.h> + + display-controller@2220000 { + compatible = "fsl,imx6sx-lcdif"; + reg = <0x02220000 0x4000>; + interrupts = <GIC_SPI 5 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&clks IMX6SX_CLK_LCDIF1_PIX>, + <&clks IMX6SX_CLK_LCDIF_APB>, + <&clks IMX6SX_CLK_DISPLAY_AXI>; + clock-names = "pix", "axi", "disp_axi"; + + port { + endpoint { + remote-endpoint = <&panel_in>; + }; + }; + }; + +... diff --git a/Documentation/devicetree/bindings/display/imx/fsl,imx6-hdmi.yaml b/Documentation/devicetree/bindings/display/imx/fsl,imx6-hdmi.yaml new file mode 100644 index 000000000000..af7fe9c4d196 --- /dev/null +++ b/Documentation/devicetree/bindings/display/imx/fsl,imx6-hdmi.yaml @@ -0,0 +1,126 @@ +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) +%YAML 1.2 +--- +$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>; + }; + }; + }; + }; + +... diff --git a/Documentation/devicetree/bindings/display/imx/hdmi.txt b/Documentation/devicetree/bindings/display/imx/hdmi.txt deleted file mode 100644 index 6d021e71c9cf..000000000000 --- a/Documentation/devicetree/bindings/display/imx/hdmi.txt +++ /dev/null @@ -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>; - }; - }; - }; diff --git a/Documentation/devicetree/bindings/display/mediatek/mediatek,dpi.yaml b/Documentation/devicetree/bindings/display/mediatek/mediatek,dpi.yaml index 6cdb734c91a9..dd2896a40ff0 100644 --- a/Documentation/devicetree/bindings/display/mediatek/mediatek,dpi.yaml +++ b/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 - Documentation/devicetree/bindings/graph.txt. This port should be connected - to the input port of an attached HDMI or LVDS encoder chip. - - properties: - endpoint: - type: object + Output port node. This port should be connected to the input port of an + attached HDMI or LVDS encoder chip. required: - compatible diff --git a/Documentation/devicetree/bindings/display/mxsfb.txt b/Documentation/devicetree/bindings/display/mxsfb.txt deleted file mode 100644 index c985871c46b3..000000000000 --- a/Documentation/devicetree/bindings/display/mxsfb.txt +++ /dev/null @@ -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>; - }; - }; - }; -}; diff --git a/Documentation/devicetree/bindings/display/panel/panel-simple.yaml b/Documentation/devicetree/bindings/display/panel/panel-simple.yaml index 62b0d54d87b7..b3797ba2698b 100644 --- a/Documentation/devicetree/bindings/display/panel/panel-simple.yaml +++ b/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 diff --git a/Documentation/devicetree/bindings/display/renesas,du.txt b/Documentation/devicetree/bindings/display/renesas,du.txt deleted file mode 100644 index 7d65c24fcda8..000000000000 --- a/Documentation/devicetree/bindings/display/renesas,du.txt +++ /dev/null @@ -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 { - }; - }; - }; - }; diff --git a/Documentation/devicetree/bindings/display/renesas,du.yaml b/Documentation/devicetree/bindings/display/renesas,du.yaml new file mode 100644 index 000000000000..552a99ce4f12 --- /dev/null +++ b/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>; + }; + }; + }; + }; + +... diff --git a/Documentation/devicetree/bindings/display/rockchip/dw_hdmi-rockchip.txt b/Documentation/devicetree/bindings/display/rockchip/dw_hdmi-rockchip.txt deleted file mode 100644 index 3d32ce137e7f..000000000000 --- a/Documentation/devicetree/bindings/display/rockchip/dw_hdmi-rockchip.txt +++ /dev/null @@ -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>; - }; - }; - }; -}; diff --git a/Documentation/devicetree/bindings/display/rockchip/rockchip,dw-hdmi.yaml b/Documentation/devicetree/bindings/display/rockchip/rockchip,dw-hdmi.yaml new file mode 100644 index 000000000000..75cd9c686e98 --- /dev/null +++ b/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>; + }; + }; + }; + }; + +... diff --git a/Documentation/driver-api/dma-buf.rst b/Documentation/driver-api/dma-buf.rst index a2133d69872c..7f37ec30d9fd 100644 --- a/Documentation/driver-api/dma-buf.rst +++ b/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. diff --git a/Documentation/gpu/drm-kms-helpers.rst b/Documentation/gpu/drm-kms-helpers.rst index b89ddd06dabb..389892f36185 100644 --- a/Documentation/gpu/drm-kms-helpers.rst +++ b/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 =========================== diff --git a/Documentation/gpu/index.rst b/Documentation/gpu/index.rst index c9a51e3bfb5a..ec4bc72438e4 100644 --- a/Documentation/gpu/index.rst +++ b/Documentation/gpu/index.rst @@ -16,6 +16,7 @@ Linux GPU Driver Developer's Guide vga-switcheroo vgaarbiter todo + rfc/index .. only:: subproject and html diff --git a/Documentation/gpu/rfc/index.rst b/Documentation/gpu/rfc/index.rst new file mode 100644 index 000000000000..a8621f7dab8b --- /dev/null +++ b/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. diff --git a/Documentation/gpu/todo.rst b/Documentation/gpu/todo.rst index 22ce801e3a8d..7ff9fac10d8b 100644 --- a/Documentation/gpu/todo.rst +++ b/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 +--------------------- + +There's two related issues here + +- Cleanup up the various ->destroy callbacks, which often are all the same + simple code. -Some of these date from the very introduction of KMS in 2008 ... +- Lots of drivers erroneously allocate DRM modeset objects using devm_kzalloc, + which results in use-after free issues on driver unload. This can be serious + trouble even for drivers for hardware integrated on the SoC due to + EPROBE_DEFERRED backoff. -- Make ->funcs and ->helper_private vtables optional. There's a bunch of empty - function tables in drivers, but before we can remove them we need to make sure - that all the users in helpers and drivers do correctly check for a NULL - vtable. +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. -- Cleanup up the various ->destroy callbacks. A lot of them just wrapt the - drm_*_cleanup implementations and can be removed. Some tack a kfree() at the - end, for which we could add drm_*_cleanup_kfree(). And then there's the (for - historical reasons) misnamed drm_primary_helper_destroy() function. +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. diff --git a/MAINTAINERS b/MAINTAINERS index ccc31233f9ee..e32c844a154a 100644 --- a/MAINTAINERS +++ b/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 diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index a4b5af03dcc1..6edd1e2ee8af 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/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); diff --git a/drivers/clk/clk-mux.c b/drivers/clk/clk-mux.c index e54e79714818..20582aae7a35 100644 --- a/drivers/clk/clk-mux.c +++ b/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, diff --git a/drivers/dma-buf/dma-fence.c b/drivers/dma-buf/dma-fence.c index d64fc03929be..ce0f5eff575d 100644 --- a/drivers/dma-buf/dma-fence.c +++ b/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) { @@ -142,6 +144,29 @@ 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 * diff --git a/drivers/dma-buf/dma-heap.c b/drivers/dma-buf/dma-heap.c index 6b5db954569f..56bf5ad01ad5 100644 --- a/drivers/dma-buf/dma-heap.c +++ b/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; diff --git a/drivers/dma-buf/heaps/cma_heap.c b/drivers/dma-buf/heaps/cma_heap.c index 5d64eccd21d6..0c05b79870f9 100644 --- a/drivers/dma-buf/heaps/cma_heap.c +++ b/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; diff --git a/drivers/dma-buf/heaps/system_heap.c b/drivers/dma-buf/heaps/system_heap.c index 29e49ac17251..23a7e74ef966 100644 --- a/drivers/dma-buf/heaps/system_heap.c +++ b/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; diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index 85b79a7fee63..3c16bd1afd87 100644 --- a/drivers/gpu/drm/Kconfig +++ b/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 diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index 926adef289db..5279db4392df 100644 --- a/drivers/gpu/drm/Makefile +++ b/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/ diff --git a/drivers/gpu/drm/amd/amdgpu/Kconfig b/drivers/gpu/drm/amd/amdgpu/Kconfig index 9375e7f12420..74a8105fd2c0 100644 --- a/drivers/gpu/drm/amd/amdgpu/Kconfig +++ b/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" diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index 13ebb1f71e49..ee85e8aba636 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/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 diff --git a/drivers/gpu/drm/amd/amdgpu/aldebaran.c b/drivers/gpu/drm/amd/amdgpu/aldebaran.c new file mode 100644 index 000000000000..65b1dca4b02e --- /dev/null +++ b/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; +} diff --git a/drivers/gpu/drm/amd/amdgpu/aldebaran.h b/drivers/gpu/drm/amd/amdgpu/aldebaran.h new file mode 100644 index 000000000000..a07db5454d49 --- /dev/null +++ b/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 diff --git a/drivers/gpu/drm/amd/amdgpu/aldebaran_reg_init.c b/drivers/gpu/drm/amd/amdgpu/aldebaran_reg_init.c new file mode 100644 index 000000000000..28e6c9ab8767 --- /dev/null +++ b/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; +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 29885febc0b0..dc3a69296321 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/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]; - unsigned debugfs_count; -#if defined(CONFIG_DEBUG_FS) - struct dentry *debugfs_preempt; - struct dentry *debugfs_regs[AMDGPU_DEBUGFS_MAX_COMPONENTS]; -#endif + struct debugfs_blob_wrapper debugfs_vbios_blob; 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) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index c5343a5eecbe..5f6696a3c778 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/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); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index a81d9cacf9b8..14f68c028126 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/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 { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_aldebaran.c new file mode 100644 index 000000000000..a5434b713856 --- /dev/null +++ b/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, +}; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c index 604757a1e440..9ef9f3ddad48 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.c +++ b/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, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_arcturus.h new file mode 100644 index 000000000000..ce08131b7b5f --- /dev/null +++ b/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); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c index 3107b9575929..5af464933976 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_fence.c +++ b/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. * diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index ac0a432a9bf7..e93850f2f3b1 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/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, - sizeof(struct amdgpu_bo)); + acc_size = amdgpu_amdkfd_acc_size(size); 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, - sizeof(struct amdgpu_bo)); + acc_size = amdgpu_amdkfd_acc_size(size); 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 diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c index 86add0f4ea4d..494b2e1717d5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c +++ b/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; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h index 1321ec09c734..8cc0222dba19 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h +++ b/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); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c index 6107ac91db25..60716b35444b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c +++ b/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 && crev >= 1) || (frev > 3)) { + if (frev == 3) { umc_info = (union umc_info *) (mode_info->atom_context->bios + data_offset); - ecc_default_enabled = - (le32_to_cpu(umc_info->v31.umc_config) & - UMC_CONFIG__DEFAULT_MEM_ECC_ENABLE) ? true : false; + switch (crev) { + case 1: + 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; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c index d9b35df33806..313517f7cf10 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c +++ b/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) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c index cfb1a9a04477..27b19503773b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c +++ b/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: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index 3e240b952e79..b5c766998045 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/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) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index 43059ead733b..bcaf271b39bf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/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 drm_device *dev = node->minor->dev; - struct amdgpu_device *adev = drm_to_adev(dev); + struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + struct drm_device *dev = adev_to_drm(adev); 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 drm_device *dev = node->minor->dev; - 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); + struct amdgpu_device *adev = (struct amdgpu_device *)data; + struct drm_device *dev = adev_to_drm(adev); 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 drm_device *dev = node->minor->dev; - struct amdgpu_device *adev = drm_to_adev(dev); + struct amdgpu_device *adev = (struct amdgpu_device *)data; + struct drm_device *dev = adev_to_drm(adev); 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); - seq_printf(m, "(%d)\n", r); + *val = ttm_resource_manager_evict_all(&adev->mman.bdev, man); 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 drm_device *dev = node->minor->dev; + struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + 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[] = { - {"amdgpu_vbios", amdgpu_debugfs_get_vbios_dump}, - {"amdgpu_test_ib", &amdgpu_debugfs_test_ib}, - {"amdgpu_evict_vram", &amdgpu_debugfs_evict_vram}, - {"amdgpu_evict_gtt", &amdgpu_debugfs_evict_gtt}, - {"amdgpu_vm_info", &amdgpu_debugfs_vm_info}, -}; +DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_test_ib); +DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_vm_info); +DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_evict_vram_fops, amdgpu_debugfs_evict_vram, + NULL, "%lld\n"); +DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_evict_gtt_fops, amdgpu_debugfs_evict_gtt, + 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, - &fops_ib_preempt); - if (!(adev->debugfs_preempt)) { + + + ent = debugfs_create_file("amdgpu_preempt_ib", 0600, root, adev, + &fops_ib_preempt); + if (!ent) { DRM_ERROR("unable to create amdgpu_preempt_ib debugsfs file\n"); return -EIO; } - adev->smu.debugfs_sclk = - debugfs_create_file("amdgpu_force_sclk", 0200, - adev_to_drm(adev)->primary->debugfs_root, adev, - &fops_sclk_set); - if (!(adev->smu.debugfs_sclk)) { + ent = debugfs_create_file("amdgpu_force_sclk", 0200, root, adev, + &fops_sclk_set); + if (!ent) { 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); - if (r) { - DRM_ERROR("Failed to init debugfs\n"); - return r; - } - - 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); + amdgpu_ttm_debugfs_init(adev); + amdgpu_debugfs_pm_init(adev); + amdgpu_debugfs_sa_init(adev); + amdgpu_debugfs_fence_init(adev); + amdgpu_debugfs_gem_init(adev); r = amdgpu_debugfs_regs_init(adev); if (r) DRM_ERROR("registering register debugfs failed (%d).\n", r); - r = amdgpu_debugfs_firmware_init(adev); - if (r) - DRM_ERROR("registering firmware debugfs failed (%d).\n", r); + amdgpu_debugfs_firmware_init(adev); #if defined(CONFIG_DRM_AMD_DC) - if (amdgpu_device_has_dc_support(adev)) { - if (dtn_debugfs_init(adev)) - DRM_ERROR("amdgpu: failed initialize dtn debugfs support.\n"); - } + if (amdgpu_device_has_dc_support(adev)) + dtn_debugfs_init(adev); #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, - ARRAY_SIZE(amdgpu_debugfs_list)); + debugfs_create_file("amdgpu_evict_vram", 0444, root, adev, + &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 diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h index 2803884d338d..141a8474e24f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.h +++ b/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, - const struct drm_info_list *files, - unsigned nfiles); -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); +void amdgpu_debugfs_fence_init(struct amdgpu_device *adev); +void amdgpu_debugfs_firmware_init(struct amdgpu_device *adev); +void amdgpu_debugfs_gem_init(struct amdgpu_device *adev); int amdgpu_debugfs_wait_dump(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 8a5a8ff5d362..b4ad1c055c70 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -65,6 +65,7 @@ #include "amdgpu_ras.h" #include "amdgpu_pmu.h" #include "amdgpu_fru_eeprom.h" +#include "amdgpu_reset.h" #include <linux/suspend.h> #include <drm/task_barrier.h> @@ -110,6 +111,7 @@ const char *amdgpu_asic_name[] = { "RAVEN", "ARCTURUS", "RENOIR", + "ALDEBARAN", "NAVI10", "NAVI14", "NAVI12", @@ -136,7 +138,7 @@ static ssize_t amdgpu_device_get_pcie_replay_count(struct device *dev, struct amdgpu_device *adev = drm_to_adev(ddev); uint64_t cnt = amdgpu_asic_get_pcie_replay_count(adev); - return snprintf(buf, PAGE_SIZE, "%llu\n", cnt); + return sysfs_emit(buf, "%llu\n", cnt); } static DEVICE_ATTR(pcie_replay_count, S_IRUGO, @@ -160,7 +162,7 @@ static ssize_t amdgpu_device_get_product_name(struct device *dev, struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = drm_to_adev(ddev); - return snprintf(buf, PAGE_SIZE, "%s\n", adev->product_name); + return sysfs_emit(buf, "%s\n", adev->product_name); } static DEVICE_ATTR(product_name, S_IRUGO, @@ -182,7 +184,7 @@ static ssize_t amdgpu_device_get_product_number(struct device *dev, struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = drm_to_adev(ddev); - return snprintf(buf, PAGE_SIZE, "%s\n", adev->product_number); + return sysfs_emit(buf, "%s\n", adev->product_number); } static DEVICE_ATTR(product_number, S_IRUGO, @@ -204,25 +206,25 @@ static ssize_t amdgpu_device_get_serial_number(struct device *dev, struct drm_device *ddev = dev_get_drvdata(dev); struct amdgpu_device *adev = drm_to_adev(ddev); - return snprintf(buf, PAGE_SIZE, "%s\n", adev->serial); + return sysfs_emit(buf, "%s\n", adev->serial); } static DEVICE_ATTR(serial_number, S_IRUGO, amdgpu_device_get_serial_number, NULL); /** - * amdgpu_device_supports_atpx - Is the device a dGPU with HG/PX power control + * amdgpu_device_supports_px - Is the device a dGPU with ATPX power control * * @dev: drm_device pointer * - * Returns true if the device is a dGPU with HG/PX power control, + * Returns true if the device is a dGPU with ATPX power control, * otherwise return false. */ -bool amdgpu_device_supports_atpx(struct drm_device *dev) +bool amdgpu_device_supports_px(struct drm_device *dev) { struct amdgpu_device *adev = drm_to_adev(dev); - if (adev->flags & AMD_IS_PX) + if ((adev->flags & AMD_IS_PX) && !amdgpu_is_atpx_hybrid()) return true; return false; } @@ -232,14 +234,15 @@ bool amdgpu_device_supports_atpx(struct drm_device *dev) * * @dev: drm_device pointer * - * Returns true if the device is a dGPU with HG/PX power control, + * Returns true if the device is a dGPU with ACPI power control, * otherwise return false. */ bool amdgpu_device_supports_boco(struct drm_device *dev) { struct amdgpu_device *adev = drm_to_adev(dev); - if (adev->has_pr3) + if (adev->has_pr3 || + ((adev->flags & AMD_IS_PX) && amdgpu_is_atpx_hybrid())) return true; return false; } @@ -325,6 +328,35 @@ void amdgpu_device_vram_access(struct amdgpu_device *adev, loff_t pos, /* * register access helper functions. */ + +/* Check if hw access should be skipped because of hotplug or device error */ +bool amdgpu_device_skip_hw_access(struct amdgpu_device *adev) +{ + if (adev->in_pci_err_recovery) + return true; + +#ifdef CONFIG_LOCKDEP + /* + * This is a bit complicated to understand, so worth a comment. What we assert + * here is that the GPU reset is not running on another thread in parallel. + * + * For this we trylock the read side of the reset semaphore, if that succeeds + * we know that the reset is not running in paralell. + * + * If the trylock fails we assert that we are either already holding the read + * side of the lock or are the reset thread itself and hold the write side of + * the lock. + */ + if (in_task()) { + if (down_read_trylock(&adev->reset_sem)) + up_read(&adev->reset_sem); + else + lockdep_assert_held(&adev->reset_sem); + } +#endif + return false; +} + /** * amdgpu_device_rreg - read a memory mapped IO or indirect register * @@ -339,7 +371,7 @@ uint32_t amdgpu_device_rreg(struct amdgpu_device *adev, { uint32_t ret; - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return 0; if ((reg * 4) < adev->rmmio_size) { @@ -376,7 +408,7 @@ uint32_t amdgpu_device_rreg(struct amdgpu_device *adev, */ uint8_t amdgpu_mm_rreg8(struct amdgpu_device *adev, uint32_t offset) { - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return 0; if (offset < adev->rmmio_size) @@ -401,7 +433,7 @@ uint8_t amdgpu_mm_rreg8(struct amdgpu_device *adev, uint32_t offset) */ void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value) { - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return; if (offset < adev->rmmio_size) @@ -424,7 +456,7 @@ void amdgpu_device_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v, uint32_t acc_flags) { - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return; if ((reg * 4) < adev->rmmio_size) { @@ -451,63 +483,20 @@ void amdgpu_device_wreg(struct amdgpu_device *adev, void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev, uint32_t reg, uint32_t v) { - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return; if (amdgpu_sriov_fullaccess(adev) && adev->gfx.rlc.funcs && adev->gfx.rlc.funcs->is_rlcg_access_range) { if (adev->gfx.rlc.funcs->is_rlcg_access_range(adev, reg)) - return adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, v); + return adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, v, 0); } else { writel(v, ((void __iomem *)adev->rmmio) + (reg * 4)); } } /** - * amdgpu_io_rreg - read an IO register - * - * @adev: amdgpu_device pointer - * @reg: dword aligned register offset - * - * Returns the 32 bit value from the offset specified. - */ -u32 amdgpu_io_rreg(struct amdgpu_device *adev, u32 reg) -{ - if (adev->in_pci_err_recovery) - return 0; - - if ((reg * 4) < adev->rio_mem_size) - return ioread32(adev->rio_mem + (reg * 4)); - else { - iowrite32((reg * 4), adev->rio_mem + (mmMM_INDEX * 4)); - return ioread32(adev->rio_mem + (mmMM_DATA * 4)); - } -} - -/** - * amdgpu_io_wreg - write to an IO register - * - * @adev: amdgpu_device pointer - * @reg: dword aligned register offset - * @v: 32 bit value to write to the register - * - * Writes the value specified to the offset specified. - */ -void amdgpu_io_wreg(struct amdgpu_device *adev, u32 reg, u32 v) -{ - if (adev->in_pci_err_recovery) - return; - - if ((reg * 4) < adev->rio_mem_size) - iowrite32(v, adev->rio_mem + (reg * 4)); - else { - iowrite32((reg * 4), adev->rio_mem + (mmMM_INDEX * 4)); - iowrite32(v, adev->rio_mem + (mmMM_DATA * 4)); - } -} - -/** * amdgpu_mm_rdoorbell - read a doorbell dword * * @adev: amdgpu_device pointer @@ -518,7 +507,7 @@ void amdgpu_io_wreg(struct amdgpu_device *adev, u32 reg, u32 v) */ u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index) { - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return 0; if (index < adev->doorbell.num_doorbells) { @@ -541,7 +530,7 @@ u32 amdgpu_mm_rdoorbell(struct amdgpu_device *adev, u32 index) */ void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v) { - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return; if (index < adev->doorbell.num_doorbells) { @@ -562,7 +551,7 @@ void amdgpu_mm_wdoorbell(struct amdgpu_device *adev, u32 index, u32 v) */ u64 amdgpu_mm_rdoorbell64(struct amdgpu_device *adev, u32 index) { - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return 0; if (index < adev->doorbell.num_doorbells) { @@ -585,7 +574,7 @@ u64 amdgpu_mm_rdoorbell64(struct amdgpu_device *adev, u32 index) */ void amdgpu_mm_wdoorbell64(struct amdgpu_device *adev, u32 index, u64 v) { - if (adev->in_pci_err_recovery) + if (amdgpu_device_skip_hw_access(adev)) return; if (index < adev->doorbell.num_doorbells) { @@ -1223,6 +1212,10 @@ bool amdgpu_device_need_post(struct amdgpu_device *adev) } } + /* Don't post if we need to reset whole hive on init */ + if (adev->gmc.xgmi.pending_reset) + return false; + if (adev->has_hw_reset) { adev->has_hw_reset = false; return true; @@ -1429,7 +1422,7 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev, struct drm_device *dev = pci_get_drvdata(pdev); int r; - if (amdgpu_device_supports_atpx(dev) && state == VGA_SWITCHEROO_OFF) + if (amdgpu_device_supports_px(dev) && state == VGA_SWITCHEROO_OFF) return; if (state == VGA_SWITCHEROO_ON) { @@ -1810,6 +1803,7 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev) case CHIP_CARRIZO: case CHIP_STONEY: case CHIP_VEGA20: + case CHIP_ALDEBARAN: case CHIP_SIENNA_CICHLID: case CHIP_NAVY_FLOUNDER: case CHIP_DIMGREY_CAVEFISH: @@ -2010,6 +2004,7 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) case CHIP_RAVEN: case CHIP_ARCTURUS: case CHIP_RENOIR: + case CHIP_ALDEBARAN: if (adev->flags & AMD_IS_APU) adev->family = AMDGPU_FAMILY_RV; else @@ -2045,6 +2040,8 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) adev->pm.pp_feature = amdgpu_pp_feature_mask; if (amdgpu_sriov_vf(adev) || sched_policy == KFD_SCHED_POLICY_NO_HWS) adev->pm.pp_feature &= ~PP_GFXOFF_MASK; + if (amdgpu_sriov_vf(adev) && adev->asic_type == CHIP_SIENNA_CICHLID) + adev->pm.pp_feature &= ~PP_OVERDRIVE_MASK; for (i = 0; i < adev->num_ip_blocks; i++) { if ((amdgpu_ip_block_mask & (1 << i)) == 0) { @@ -2083,6 +2080,11 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev) amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, 0, 0); return r; } + + /*get pf2vf msg info at it's earliest time*/ + if (amdgpu_sriov_vf(adev)) + amdgpu_virt_init_data_exchange(adev); + } } @@ -2149,6 +2151,9 @@ static int amdgpu_device_fw_loading(struct amdgpu_device *adev) if (adev->ip_blocks[i].version->type != AMD_IP_BLOCK_TYPE_PSP) continue; + if (!adev->ip_blocks[i].status.sw) + continue; + /* no need to do the fw loading again if already done*/ if (adev->ip_blocks[i].status.hw == true) break; @@ -2289,7 +2294,10 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev) if (adev->gmc.xgmi.num_physical_nodes > 1) amdgpu_xgmi_add_device(adev); - amdgpu_amdkfd_device_init(adev); + + /* Don't init kfd if whole hive need to be reset during init */ + if (!adev->gmc.xgmi.pending_reset) + amdgpu_amdkfd_device_init(adev); amdgpu_fru_get_product_info(adev); @@ -2359,8 +2367,8 @@ static bool amdgpu_device_check_vram_lost(struct amdgpu_device *adev) * Returns 0 on success, negative error code on failure. */ -static int amdgpu_device_set_cg_state(struct amdgpu_device *adev, - enum amd_clockgating_state state) +int amdgpu_device_set_cg_state(struct amdgpu_device *adev, + enum amd_clockgating_state state) { int i, j, r; @@ -2395,7 +2403,8 @@ static int amdgpu_device_set_cg_state(struct amdgpu_device *adev, return 0; } -static int amdgpu_device_set_pg_state(struct amdgpu_device *adev, enum amd_powergating_state state) +int amdgpu_device_set_pg_state(struct amdgpu_device *adev, + enum amd_powergating_state state) { int i, j, r; @@ -2506,6 +2515,11 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev) if (r) DRM_ERROR("enable mgpu fan boost failed (%d).\n", r); + /* For XGMI + passthrough configuration on arcturus, enable light SBR */ + if (adev->asic_type == CHIP_ARCTURUS && + amdgpu_passthrough(adev) && + adev->gmc.xgmi.num_physical_nodes > 1) + smu_set_light_sbr(&adev->smu, true); if (adev->gmc.xgmi.num_physical_nodes > 1) { mutex_lock(&mgpu_info.mutex); @@ -2743,6 +2757,16 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev) continue; } + /* skip unnecessary suspend if we do not initialize them yet */ + if (adev->gmc.xgmi.pending_reset && + !(adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH)) { + adev->ip_blocks[i].status.hw = false; + continue; + } + /* skip suspend of gfx and psp for S0ix * gfx is in gfxoff state, so on resume it will exit gfxoff just * like at runtime. PSP is also part of the always on hardware @@ -2772,7 +2796,6 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev) } } } - adev->ip_blocks[i].status.hw = false; } return 0; @@ -2793,8 +2816,10 @@ int amdgpu_device_ip_suspend(struct amdgpu_device *adev) { int r; - if (amdgpu_sriov_vf(adev)) + if (amdgpu_sriov_vf(adev)) { + amdgpu_virt_fini_data_exchange(adev); amdgpu_virt_request_full_gpu(adev, false); + } r = amdgpu_device_ip_suspend_phase1(adev); if (r) @@ -3117,8 +3142,9 @@ static void amdgpu_device_xgmi_reset_func(struct work_struct *__work) if (adev->asic_reset_res) goto fail; - if (adev->mmhub.funcs && adev->mmhub.funcs->reset_ras_error_count) - adev->mmhub.funcs->reset_ras_error_count(adev); + if (adev->mmhub.ras_funcs && + adev->mmhub.ras_funcs->reset_ras_error_count) + adev->mmhub.ras_funcs->reset_ras_error_count(adev); } else { task_barrier_full(&hive->tb); @@ -3228,7 +3254,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, struct drm_device *ddev = adev_to_drm(adev); struct pci_dev *pdev = adev->pdev; int r, i; - bool atpx = false; + bool px = false; u32 max_MBps; adev->shutdown = false; @@ -3276,7 +3302,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, /* mutex initialization are all done here so we * can recall function without having locking issues */ - atomic_set(&adev->irq.ih.lock, 0); mutex_init(&adev->firmware.mutex); mutex_init(&adev->pm.mutex); mutex_init(&adev->gfx.gpu_clock_mutex); @@ -3309,6 +3334,8 @@ int amdgpu_device_init(struct amdgpu_device *adev, INIT_LIST_HEAD(&adev->shadow_list); mutex_init(&adev->shadow_list_lock); + INIT_LIST_HEAD(&adev->reset_list); + INIT_DELAYED_WORK(&adev->delayed_init_work, amdgpu_device_delayed_init_work_handler); INIT_DELAYED_WORK(&adev->gfx.gfx_off_delay_work, @@ -3347,17 +3374,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, DRM_INFO("register mmio base: 0x%08X\n", (uint32_t)adev->rmmio_base); DRM_INFO("register mmio size: %u\n", (unsigned)adev->rmmio_size); - /* io port mapping */ - for (i = 0; i < DEVICE_COUNT_RESOURCE; i++) { - if (pci_resource_flags(adev->pdev, i) & IORESOURCE_IO) { - adev->rio_mem_size = pci_resource_len(adev->pdev, i); - adev->rio_mem = pci_iomap(adev->pdev, i, adev->rio_mem_size); - break; - } - } - if (adev->rio_mem == NULL) - DRM_INFO("PCI I/O BAR is not found.\n"); - /* enable PCIE atomic ops */ r = pci_enable_atomic_ops_to_root(adev->pdev, PCI_EXP_DEVCAP2_ATOMIC_COMP32 | @@ -3400,16 +3416,12 @@ int amdgpu_device_init(struct amdgpu_device *adev, if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA) vga_client_register(adev->pdev, adev, NULL, amdgpu_device_vga_set_decode); - if (amdgpu_device_supports_atpx(ddev)) - atpx = true; - if (amdgpu_has_atpx() && - (amdgpu_is_atpx_hybrid() || - amdgpu_has_atpx_dgpu_power_cntl()) && - !pci_is_thunderbolt_attached(adev->pdev)) + if (amdgpu_device_supports_px(ddev)) { + px = true; vga_switcheroo_register_client(adev->pdev, - &amdgpu_switcheroo_ops, atpx); - if (atpx) + &amdgpu_switcheroo_ops, px); vga_switcheroo_init_domain_pm_ops(adev->dev, &adev->vga_pm_domain); + } if (amdgpu_emu_mode == 1) { /* post the asic on emulation mode */ @@ -3417,6 +3429,8 @@ int amdgpu_device_init(struct amdgpu_device *adev, goto fence_driver_init; } + amdgpu_reset_init(adev); + /* detect if we are with an SRIOV vbios */ amdgpu_device_detect_sriov_bios(adev); @@ -3424,10 +3438,28 @@ int amdgpu_device_init(struct amdgpu_device *adev, * E.g., driver was not cleanly unloaded previously, etc. */ if (!amdgpu_sriov_vf(adev) && amdgpu_asic_need_reset_on_init(adev)) { - r = amdgpu_asic_reset(adev); - if (r) { - dev_err(adev->dev, "asic reset on init failed\n"); - goto failed; + if (adev->gmc.xgmi.num_physical_nodes) { + dev_info(adev->dev, "Pending hive reset.\n"); + adev->gmc.xgmi.pending_reset = true; + /* Only need to init necessary block for SMU to handle the reset */ + for (i = 0; i < adev->num_ip_blocks; i++) { + if (!adev->ip_blocks[i].status.valid) + continue; + if (!(adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GMC || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_IH || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC)) { + DRM_DEBUG("IP %s disabled for hw_init.\n", + adev->ip_blocks[i].version->funcs->name); + adev->ip_blocks[i].status.hw = true; + } + } + } else { + r = amdgpu_asic_reset(adev); + if (r) { + dev_err(adev->dev, "asic reset on init failed\n"); + goto failed; + } } } @@ -3493,11 +3525,11 @@ fence_driver_init: adev->virt.caps &= ~AMDGPU_SRIOV_CAPS_RUNTIME; adev->virt.ops = NULL; r = -EAGAIN; - goto failed; + goto release_ras_con; } dev_err(adev->dev, "amdgpu_device_ip_init failed\n"); amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_AMDGPU_INIT_FAIL, 0, 0); - goto failed; + goto release_ras_con; } dev_info(adev->dev, @@ -3558,19 +3590,19 @@ fence_driver_init: /* enable clockgating, etc. after ib tests, etc. since some blocks require * explicit gating rather than handling it automatically. */ - r = amdgpu_device_ip_late_init(adev); - if (r) { - dev_err(adev->dev, "amdgpu_device_ip_late_init failed\n"); - amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_AMDGPU_LATE_INIT_FAIL, 0, r); - goto failed; + if (!adev->gmc.xgmi.pending_reset) { + r = amdgpu_device_ip_late_init(adev); + if (r) { + dev_err(adev->dev, "amdgpu_device_ip_late_init failed\n"); + amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_AMDGPU_LATE_INIT_FAIL, 0, r); + goto release_ras_con; + } + /* must succeed. */ + amdgpu_ras_resume(adev); + queue_delayed_work(system_wq, &adev->delayed_init_work, + msecs_to_jiffies(AMDGPU_RESUME_MS)); } - /* must succeed. */ - amdgpu_ras_resume(adev); - - queue_delayed_work(system_wq, &adev->delayed_init_work, - msecs_to_jiffies(AMDGPU_RESUME_MS)); - if (amdgpu_sriov_vf(adev)) flush_delayed_work(&adev->delayed_init_work); @@ -3587,11 +3619,18 @@ fence_driver_init: if (amdgpu_device_cache_pci_state(adev->pdev)) pci_restore_state(pdev); + if (adev->gmc.xgmi.pending_reset) + queue_delayed_work(system_wq, &mgpu_info.delayed_reset_work, + msecs_to_jiffies(AMDGPU_RESUME_MS)); + return 0; +release_ras_con: + amdgpu_release_ras_context(adev); + failed: amdgpu_vf_error_trans_all(adev); - if (atpx) + if (px) vga_switcheroo_fini_domain_pm_ops(adev->dev); failed_unmap: @@ -3613,6 +3652,7 @@ void amdgpu_device_fini(struct amdgpu_device *adev) { dev_info(adev->dev, "amdgpu: finishing device.\n"); flush_delayed_work(&adev->delayed_init_work); + ttm_bo_lock_delayed_workqueue(&adev->mman.bdev); adev->shutdown = true; kfree(adev->pci_state); @@ -3641,6 +3681,9 @@ void amdgpu_device_fini(struct amdgpu_device *adev) release_firmware(adev->firmware.gpu_info_fw); adev->firmware.gpu_info_fw = NULL; adev->accel_working = false; + + amdgpu_reset_fini(adev); + /* free i2c buses */ if (!amdgpu_device_has_dc_support(adev)) amdgpu_i2c_fini(adev); @@ -3650,18 +3693,12 @@ void amdgpu_device_fini(struct amdgpu_device *adev) kfree(adev->bios); adev->bios = NULL; - if (amdgpu_has_atpx() && - (amdgpu_is_atpx_hybrid() || - amdgpu_has_atpx_dgpu_power_cntl()) && - !pci_is_thunderbolt_attached(adev->pdev)) + if (amdgpu_device_supports_px(adev_to_drm(adev))) { vga_switcheroo_unregister_client(adev->pdev); - if (amdgpu_device_supports_atpx(adev_to_drm(adev))) vga_switcheroo_fini_domain_pm_ops(adev->dev); + } if ((adev->pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA) vga_client_register(adev->pdev, NULL, NULL, NULL); - if (adev->rio_mem) - pci_iounmap(adev->pdev, adev->rio_mem); - adev->rio_mem = NULL; iounmap(adev->rmmio); adev->rmmio = NULL; amdgpu_device_doorbell_fini(adev); @@ -4079,11 +4116,11 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev, amdgpu_amdkfd_post_reset(adev); error: - amdgpu_virt_release_full_gpu(adev, true); if (!r && adev->virt.gim_feature & AMDGIM_FEATURE_GIM_FLR_VRAMLOST) { amdgpu_inc_vram_lost(adev); r = amdgpu_device_recover_vram(adev); } + amdgpu_virt_release_full_gpu(adev, true); return r; } @@ -4160,6 +4197,8 @@ bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev) case CHIP_SIENNA_CICHLID: case CHIP_NAVY_FLOUNDER: case CHIP_DIMGREY_CAVEFISH: + case CHIP_VANGOGH: + case CHIP_ALDEBARAN: break; default: goto disabled; @@ -4173,15 +4212,60 @@ disabled: return false; } +int amdgpu_device_mode1_reset(struct amdgpu_device *adev) +{ + u32 i; + int ret = 0; + + amdgpu_atombios_scratch_regs_engine_hung(adev, true); + + dev_info(adev->dev, "GPU mode1 reset\n"); + + /* disable BM */ + pci_clear_master(adev->pdev); + + amdgpu_device_cache_pci_state(adev->pdev); + + if (amdgpu_dpm_is_mode1_reset_supported(adev)) { + dev_info(adev->dev, "GPU smu mode1 reset\n"); + ret = amdgpu_dpm_mode1_reset(adev); + } else { + dev_info(adev->dev, "GPU psp mode1 reset\n"); + ret = psp_gpu_reset(adev); + } + + if (ret) + dev_err(adev->dev, "GPU mode1 reset failed\n"); + + amdgpu_device_load_pci_state(adev->pdev); + + /* wait for asic to come out of reset */ + for (i = 0; i < adev->usec_timeout; i++) { + u32 memsize = adev->nbio.funcs->get_memsize(adev); + + if (memsize != 0xffffffff) + break; + udelay(1); + } + + amdgpu_atombios_scratch_regs_engine_hung(adev, false); + return ret; +} -static int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, - struct amdgpu_job *job, - bool *need_full_reset_arg) +int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, + struct amdgpu_reset_context *reset_context) { int i, r = 0; - bool need_full_reset = *need_full_reset_arg; + struct amdgpu_job *job = NULL; + bool need_full_reset = + test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); - amdgpu_debugfs_wait_dump(adev); + if (reset_context->reset_req_dev == adev) + job = reset_context->job; + + /* no need to dump if device is not in good state during probe period */ + if (!adev->gmc.xgmi.pending_reset) + amdgpu_debugfs_wait_dump(adev); if (amdgpu_sriov_vf(adev)) { /* stop the data exchange thread */ @@ -4202,6 +4286,13 @@ static int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, if(job) drm_sched_increase_karma(&job->base); + r = amdgpu_reset_prepare_hwcontext(adev, reset_context); + /* If reset handler not implemented, continue; otherwise return */ + if (r == -ENOSYS) + r = 0; + else + return r; + /* Don't suspend on bare metal if we are not going to HW reset the ASIC */ if (!amdgpu_sriov_vf(adev)) { @@ -4220,30 +4311,47 @@ static int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev, if (need_full_reset) r = amdgpu_device_ip_suspend(adev); - - *need_full_reset_arg = need_full_reset; + if (need_full_reset) + set_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); + else + clear_bit(AMDGPU_NEED_FULL_RESET, + &reset_context->flags); } return r; } -static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, - struct list_head *device_list_handle, - bool *need_full_reset_arg, - bool skip_hw_reset) +int amdgpu_do_asic_reset(struct list_head *device_list_handle, + struct amdgpu_reset_context *reset_context) { struct amdgpu_device *tmp_adev = NULL; - bool need_full_reset = *need_full_reset_arg, vram_lost = false; + bool need_full_reset, skip_hw_reset, vram_lost = false; int r = 0; + /* Try reset handler method first */ + tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device, + reset_list); + r = amdgpu_reset_perform_reset(tmp_adev, reset_context); + /* If reset handler not implemented, continue; otherwise return */ + if (r == -ENOSYS) + r = 0; + else + return r; + + /* Reset handler not implemented, use the default method */ + need_full_reset = + test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); + skip_hw_reset = test_bit(AMDGPU_SKIP_HW_RESET, &reset_context->flags); + /* - * ASIC reset has to be done on all HGMI hive nodes ASAP + * ASIC reset has to be done on all XGMI hive nodes ASAP * to allow proper links negotiation in FW (within 1 sec) */ if (!skip_hw_reset && need_full_reset) { - list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) { + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { /* For XGMI run all resets in parallel to speed up the process */ if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) { + tmp_adev->gmc.xgmi.pending_reset = false; if (!queue_work(system_unbound_wq, &tmp_adev->xgmi_reset_work)) r = -EALREADY; } else @@ -4258,8 +4366,7 @@ static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, /* For XGMI wait for all resets to complete before proceed */ if (!r) { - list_for_each_entry(tmp_adev, device_list_handle, - gmc.xgmi.head) { + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { if (tmp_adev->gmc.xgmi.num_physical_nodes > 1) { flush_work(&tmp_adev->xgmi_reset_work); r = tmp_adev->asic_reset_res; @@ -4271,22 +4378,22 @@ static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, } if (!r && amdgpu_ras_intr_triggered()) { - list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) { - if (tmp_adev->mmhub.funcs && - tmp_adev->mmhub.funcs->reset_ras_error_count) - tmp_adev->mmhub.funcs->reset_ras_error_count(tmp_adev); + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + if (tmp_adev->mmhub.ras_funcs && + tmp_adev->mmhub.ras_funcs->reset_ras_error_count) + tmp_adev->mmhub.ras_funcs->reset_ras_error_count(tmp_adev); } amdgpu_ras_intr_cleared(); } - list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) { + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { if (need_full_reset) { /* post card */ - if (amdgpu_device_asic_init(tmp_adev)) + r = amdgpu_device_asic_init(tmp_adev); + if (r) { dev_warn(tmp_adev->dev, "asic atom init failed!"); - - if (!r) { + } else { dev_info(tmp_adev->dev, "GPU reset succeeded, trying to resume\n"); r = amdgpu_device_ip_resume_phase1(tmp_adev); if (r) @@ -4319,6 +4426,10 @@ static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, */ amdgpu_register_gpu_instance(tmp_adev); + if (!reset_context->hive && + tmp_adev->gmc.xgmi.num_physical_nodes > 1) + amdgpu_xgmi_add_device(tmp_adev); + r = amdgpu_device_ip_late_init(tmp_adev); if (r) goto out; @@ -4335,7 +4446,7 @@ static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, * bad_page_threshold value to fix this once * probing driver again. */ - if (!amdgpu_ras_check_err_threshold(tmp_adev)) { + if (!amdgpu_ras_eeprom_check_err_threshold(tmp_adev)) { /* must succeed. */ amdgpu_ras_resume(tmp_adev); } else { @@ -4344,8 +4455,10 @@ static int amdgpu_do_asic_reset(struct amdgpu_hive_info *hive, } /* Update PSP FW topology after reset */ - if (hive && tmp_adev->gmc.xgmi.num_physical_nodes > 1) - r = amdgpu_xgmi_update_topology(hive, tmp_adev); + if (reset_context->hive && + tmp_adev->gmc.xgmi.num_physical_nodes > 1) + r = amdgpu_xgmi_update_topology( + reset_context->hive, tmp_adev); } } @@ -4369,7 +4482,10 @@ out: } end: - *need_full_reset_arg = need_full_reset; + if (need_full_reset) + set_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); + else + clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags); return r; } @@ -4385,7 +4501,6 @@ static bool amdgpu_device_lock_adev(struct amdgpu_device *adev, down_write(&adev->reset_sem); } - atomic_inc(&adev->gpu_reset_counter); switch (amdgpu_asic_reset_method(adev)) { case AMD_RESET_METHOD_MODE1: adev->mp1_state = PP_MP1_STATE_SHUTDOWN; @@ -4507,6 +4622,74 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev) return 0; } +void amdgpu_device_recheck_guilty_jobs( + struct amdgpu_device *adev, struct list_head *device_list_handle, + struct amdgpu_reset_context *reset_context) +{ + int i, r = 0; + + for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { + struct amdgpu_ring *ring = adev->rings[i]; + int ret = 0; + struct drm_sched_job *s_job; + + if (!ring || !ring->sched.thread) + continue; + + s_job = list_first_entry_or_null(&ring->sched.pending_list, + struct drm_sched_job, list); + if (s_job == NULL) + continue; + + /* clear job's guilty and depend the folowing step to decide the real one */ + drm_sched_reset_karma(s_job); + drm_sched_resubmit_jobs_ext(&ring->sched, 1); + + ret = dma_fence_wait_timeout(s_job->s_fence->parent, false, ring->sched.timeout); + if (ret == 0) { /* timeout */ + DRM_ERROR("Found the real bad job! ring:%s, job_id:%llx\n", + ring->sched.name, s_job->id); + + /* set guilty */ + drm_sched_increase_karma(s_job); +retry: + /* do hw reset */ + if (amdgpu_sriov_vf(adev)) { + amdgpu_virt_fini_data_exchange(adev); + r = amdgpu_device_reset_sriov(adev, false); + if (r) + adev->asic_reset_res = r; + } else { + clear_bit(AMDGPU_SKIP_HW_RESET, + &reset_context->flags); + r = amdgpu_do_asic_reset(device_list_handle, + reset_context); + if (r && r == -EAGAIN) + goto retry; + } + + /* + * add reset counter so that the following + * resubmitted job could flush vmid + */ + atomic_inc(&adev->gpu_reset_counter); + continue; + } + + /* got the hw fence, signal finished fence */ + atomic_dec(ring->sched.score); + dma_fence_get(&s_job->s_fence->finished); + dma_fence_signal(&s_job->s_fence->finished); + dma_fence_put(&s_job->s_fence->finished); + + /* remove node from list and free the job */ + spin_lock(&ring->sched.job_list_lock); + list_del_init(&s_job->list); + spin_unlock(&ring->sched.job_list_lock); + ring->sched.ops->free_job(s_job); + } +} + /** * amdgpu_device_gpu_recover - reset the asic and recover scheduler * @@ -4522,13 +4705,16 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, struct amdgpu_job *job) { struct list_head device_list, *device_list_handle = NULL; - bool need_full_reset = false; bool job_signaled = false; struct amdgpu_hive_info *hive = NULL; struct amdgpu_device *tmp_adev = NULL; int i, r = 0; bool need_emergency_restart = false; bool audio_suspended = false; + int tmp_vram_lost_counter; + struct amdgpu_reset_context reset_context; + + memset(&reset_context, 0, sizeof(reset_context)); /* * Special case: RAS triggered and full reset isn't supported @@ -4569,6 +4755,12 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, mutex_lock(&hive->hive_lock); } + reset_context.method = AMD_RESET_METHOD_NONE; + reset_context.reset_req_dev = adev; + reset_context.job = job; + reset_context.hive = hive; + clear_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); + /* * lock the device before we try to operate the linked list * if didn't get the device lock, don't touch the linked list since @@ -4592,16 +4784,18 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, */ INIT_LIST_HEAD(&device_list); if (adev->gmc.xgmi.num_physical_nodes > 1) { - if (!list_is_first(&adev->gmc.xgmi.head, &hive->device_list)) - list_rotate_to_front(&adev->gmc.xgmi.head, &hive->device_list); - device_list_handle = &hive->device_list; + list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) + list_add_tail(&tmp_adev->reset_list, &device_list); + if (!list_is_first(&adev->reset_list, &device_list)) + list_rotate_to_front(&adev->reset_list, &device_list); + device_list_handle = &device_list; } else { - list_add_tail(&adev->gmc.xgmi.head, &device_list); + list_add_tail(&adev->reset_list, &device_list); device_list_handle = &device_list; } /* block all schedulers and reset given job's ring */ - list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) { + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { /* * Try to put the audio codec into suspend state * before gpu reset started. @@ -4646,6 +4840,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, if (need_emergency_restart) amdgpu_job_stop_all_jobs_on_sched(&ring->sched); } + atomic_inc(&tmp_adev->gpu_reset_counter); } if (need_emergency_restart) @@ -4665,10 +4860,8 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev, } retry: /* Rest of adevs pre asic reset from XGMI hive. */ - list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) { - r = amdgpu_device_pre_asic_reset(tmp_adev, - (tmp_adev == adev) ? job : NULL, - &need_full_reset); + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + r = amdgpu_device_pre_asic_reset(tmp_adev, &reset_context); /*TODO Should we stop ?*/ if (r) { dev_err(tmp_adev->dev, "GPU pre asic reset failed with err, %d for drm dev, %s ", @@ -4677,6 +4870,7 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ } } + tmp_vram_lost_counter = atomic_read(&((adev)->vram_lost_counter)); /* Actual ASIC resets if needed.*/ /* TODO Implement XGMI hive reset logic for SRIOV */ if (amdgpu_sriov_vf(adev)) { @@ -4684,7 +4878,7 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ if (r) adev->asic_reset_res = r; } else { - r = amdgpu_do_asic_reset(hive, device_list_handle, &need_full_reset, false); + r = amdgpu_do_asic_reset(device_list_handle, &reset_context); if (r && r == -EAGAIN) goto retry; } @@ -4692,7 +4886,19 @@ retry: /* Rest of adevs pre asic reset from XGMI hive. */ skip_hw_reset: /* Post ASIC reset for all devs .*/ - list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) { + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + + /* + * Sometimes a later bad compute job can block a good gfx job as gfx + * and compute ring share internal GC HW mutually. We add an additional + * guilty jobs recheck step to find the real guilty job, it synchronously + * submits and pends for the first job being signaled. If it gets timeout, + * we identify it as a real guilty job. + */ + if (amdgpu_gpu_recovery == 2 && + !(tmp_vram_lost_counter < atomic_read(&adev->vram_lost_counter))) + amdgpu_device_recheck_guilty_jobs( + tmp_adev, device_list_handle, &reset_context); for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = tmp_adev->rings[i]; @@ -4723,10 +4929,17 @@ skip_hw_reset: } skip_sched_resume: - list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) { - /*unlock kfd: SRIOV would do it separately */ + list_for_each_entry(tmp_adev, device_list_handle, reset_list) { + /* unlock kfd: SRIOV would do it separately */ if (!need_emergency_restart && !amdgpu_sriov_vf(tmp_adev)) amdgpu_amdkfd_post_reset(tmp_adev); + + /* kfd_post_reset will do nothing if kfd device is not initialized, + * need to bring up kfd here if it's not be initialized before + */ + if (!adev->kfd.init_complete) + amdgpu_amdkfd_device_init(adev); + if (audio_suspended) amdgpu_device_resume_display_audio(tmp_adev); amdgpu_device_unlock_adev(tmp_adev); @@ -4988,6 +5201,7 @@ pci_ers_result_t amdgpu_pci_error_detected(struct pci_dev *pdev, pci_channel_sta drm_sched_stop(&ring->sched, NULL); } + atomic_inc(&adev->gpu_reset_counter); return PCI_ERS_RESULT_NEED_RESET; case pci_channel_io_perm_failure: /* Permanent error, prepare for device removal */ @@ -5029,14 +5243,16 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev) struct drm_device *dev = pci_get_drvdata(pdev); struct amdgpu_device *adev = drm_to_adev(dev); int r, i; - bool need_full_reset = true; + struct amdgpu_reset_context reset_context; u32 memsize; struct list_head device_list; DRM_INFO("PCI error: slot reset callback!!\n"); + memset(&reset_context, 0, sizeof(reset_context)); + INIT_LIST_HEAD(&device_list); - list_add_tail(&adev->gmc.xgmi.head, &device_list); + list_add_tail(&adev->reset_list, &device_list); /* wait for asic to come out of reset */ msleep(500); @@ -5057,13 +5273,18 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev) goto out; } + reset_context.method = AMD_RESET_METHOD_NONE; + reset_context.reset_req_dev = adev; + set_bit(AMDGPU_NEED_FULL_RESET, &reset_context.flags); + set_bit(AMDGPU_SKIP_HW_RESET, &reset_context.flags); + adev->in_pci_err_recovery = true; - r = amdgpu_device_pre_asic_reset(adev, NULL, &need_full_reset); + r = amdgpu_device_pre_asic_reset(adev, &reset_context); adev->in_pci_err_recovery = false; if (r) goto out; - r = amdgpu_do_asic_reset(NULL, &device_list, &need_full_reset, true); + r = amdgpu_do_asic_reset(&device_list, &reset_context); out: if (!r) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index f753e04fee99..9a2f811450ed 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/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; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c index 47e0b48dc26f..e0c4f7c7f1b9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c +++ b/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, - 0, ttm_bo_type_sg, resv, &gobj); + AMDGPU_GEM_DOMAIN_CPU, flags, + ttm_bo_type_sg, resv, &gobj); if (ret) goto error; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index e92e7dea71da..d8f131ed10cb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/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); /** @@ -528,6 +543,20 @@ MODULE_PARM_DESC(ras_mask, "Mask of RAS features to enable (default 0xffffffff), 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, * set value 0 to use radeon driver, while set value 1 to use amdgpu driver. The default is using radeon driver when it available, @@ -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 /** @@ -792,10 +828,21 @@ 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); - pci_disable_device(pdev); - pci_ignore_hotplug(pdev); - pci_set_power_state(pdev, PCI_D3cold); - } + amdgpu_device_cache_pci_state(pdev); + pci_disable_device(pdev); + pci_ignore_hotplug(pdev); + 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); - amdgpu_device_load_pci_state(pdev); - ret = pci_enable_device(pdev); - if (ret) - return ret; - } + pci_set_power_state(pdev, PCI_D0); + amdgpu_device_load_pci_state(pdev); + ret = pci_enable_device(pdev); + if (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, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c index 24010cacf7d0..4c5c19820d37 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fb.c +++ b/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, - &mode_cmd, gobj); + ret = amdgpu_display_gem_fb_init(adev_to_drm(adev), &rfbdev->rfb, + &mode_cmd, gobj); if (ret) { DRM_ERROR("failed to initialize framebuffer %d\n", ret); goto out; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index d56f4023ebb3..47ea46859618 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/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) { - switch (ring->funcs->type) { - 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; - } + if (ring->no_scheduler) + return 0; - r = drm_sched_init(&ring->sched, &amdgpu_sched_ops, - num_hw_submission, amdgpu_job_hang_limit, - timeout, ring->name); - if (r) { - DRM_ERROR("Failed to create scheduler on ring %s.\n", - ring->name); - return r; - } + switch (ring->funcs->type) { + 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, + 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 drm_device *dev = node->minor->dev; - struct amdgpu_device *adev = drm_to_adev(dev); + struct amdgpu_device *adev = (struct amdgpu_device *)m->private; 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 drm_device *dev = node->minor->dev; - struct amdgpu_device *adev = drm_to_adev(dev); + struct amdgpu_device *adev = (struct amdgpu_device *)data; + struct drm_device *dev = adev_to_drm(adev); 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"); - amdgpu_device_gpu_recover(adev, NULL); + *val = 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[] = { - {"amdgpu_fence_info", &amdgpu_debugfs_fence_info, 0, NULL}, - {"amdgpu_gpu_recover", &amdgpu_debugfs_gpu_recover, 0, NULL} -}; +DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_fence_info); +DEFINE_DEBUGFS_ATTRIBUTE(amdgpu_debugfs_gpu_recover_fops, gpu_recover_get, 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)) - return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_fence_list_sriov, - ARRAY_SIZE(amdgpu_debugfs_fence_list_sriov)); - return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_fence_list, - ARRAY_SIZE(amdgpu_debugfs_fence_list)); -#else - return 0; + struct drm_minor *minor = adev_to_drm(adev)->primary; + struct dentry *root = minor->debugfs_root; + + debugfs_create_file("amdgpu_fence_info", 0444, root, adev, + &amdgpu_debugfs_fence_info_fops); + + if (!amdgpu_sriov_vf(adev)) + debugfs_create_file("amdgpu_gpu_recover", 0444, root, adev, + &amdgpu_debugfs_gpu_recover_fops); #endif } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c index 0db933026722..c5a9a4fb10d2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.c +++ b/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); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h index afa2e2877d87..a25fe97b0196 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gart.h +++ b/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 */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index b443907afcea..311bcdc59eda 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/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 drm_device *dev = node->minor->dev; + struct amdgpu_device *adev = (struct amdgpu_device *)m->private; + 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[] = { - {"amdgpu_gem_info", &amdgpu_debugfs_gem_info, 0, NULL}, -}; +DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_gem_info); + #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, - ARRAY_SIZE(amdgpu_debugfs_gem_list)); + struct drm_minor *minor = adev_to_drm(adev)->primary; + struct dentry *root = minor->debugfs_root; + + debugfs_create_file("amdgpu_gem_info", 0444, root, adev, + &amdgpu_debugfs_gem_info_fops); #endif - return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c index 8e0a6c62322e..95d4f43a03df 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c +++ b/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, - irq, AMDGPU_CP_KIQ_IRQ_DRIVER0, - AMDGPU_RING_PRIO_DEFAULT); + r = amdgpu_ring_init(adev, ring, 1024, irq, AMDGPU_CP_KIQ_IRQ_DRIVER0, + AMDGPU_RING_PRIO_DEFAULT, NULL); 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) - adev->gfx.funcs->query_ras_error_count(adev, err_data); + if (adev->gfx.ras_funcs && + 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)) { - smu_gfx_state_change_set(&adev->smu, state); - } else { - mutex_lock(&adev->pm.mutex); - if (adev->powerplay.pp_funcs && - adev->powerplay.pp_funcs->gfx_state_change_set) - ((adev)->powerplay.pp_funcs->gfx_state_change_set( - (adev)->powerplay.pp_handle, state)); - mutex_unlock(&adev->pm.mutex); - } + mutex_lock(&adev->pm.mutex); + if (adev->powerplay.pp_funcs && + adev->powerplay.pp_funcs->gfx_state_change_set) + ((adev)->powerplay.pp_funcs->gfx_state_change_set( + (adev)->powerplay.pp_handle, state)); + mutex_unlock(&adev->pm.mutex); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h index 72dbcd2bc6a6..d43fe2ed8116 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.h +++ b/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)) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c index fe1a39ffda72..4d32233cde92 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c @@ -32,6 +32,59 @@ #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 * * @bo: the BO to get the PDE for @@ -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) { - r = adev->umc.funcs->ras_late_init(adev); + if (adev->umc.ras_funcs && + 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) { - r = adev->mmhub.funcs->ras_late_init(adev); + if (adev->mmhub.ras_funcs && + 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); - amdgpu_mmhub_ras_fini(adev); - amdgpu_xgmi_ras_fini(adev); + if (adev->umc.ras_funcs && + adev->umc.ras_funcs->ras_fini) + 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); +} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h index aa0c83776ce0..cbb7735c6988 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h +++ b/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) - * and calculate vram_start of this local device by adding an - * offset inside the XGMI hive. - * Under VMID0, logical address == MC address + * If driver uses FB aperture to access FB, driver get fb_start from + * MC_VM_FB_LOCATION_BASE (set by vbios) and calculate vram_start + * of this local device by adding an offset inside the XGMI hive. + * 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 diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c index 8980329cded0..540c01052b21 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gtt_mgr.c +++ b/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", - man->size * PAGE_SIZE); + return sysfs_emit(buf, "%llu\n", 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", - amdgpu_gtt_mgr_usage(man)); + return sysfs_emit(buf, "%llu\n", amdgpu_gtt_mgr_usage(man)); } static DEVICE_ATTR(mem_info_gtt_total, S_IRUGO, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index 7645223ea0ef..148a3b481b12 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/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 drm_device *dev = node->minor->dev; - struct amdgpu_device *adev = drm_to_adev(dev); + struct amdgpu_device *adev = (struct amdgpu_device *)m->private; 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[] = { - {"amdgpu_sa_info", &amdgpu_debugfs_sa_info, 0, NULL}, -}; +DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_sa_info); #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, - ARRAY_SIZE(amdgpu_debugfs_sa_list)); -#else - return 0; + struct drm_minor *minor = adev_to_drm(adev)->primary; + struct dentry *root = minor->debugfs_root; + + debugfs_create_file("amdgpu_sa_info", 0444, root, adev, + &amdgpu_debugfs_sa_info_fops); + #endif } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c index dc852af4f3b7..faaa6aa2faaf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c +++ b/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); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h index 6ed4a85fc7c3..0649b59830a5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h +++ b/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, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c index afbbec82a289..90f50561b43a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.c +++ b/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 && - adev->nbio.funcs->handle_ras_controller_intr_no_bifring) - adev->nbio.funcs->handle_ras_controller_intr_no_bifring(adev); + if (adev->nbio.ras_funcs && + adev->nbio.ras_funcs->handle_ras_controller_intr_no_bifring) + adev->nbio.ras_funcs->handle_ras_controller_intr_no_bifring(adev); - if (adev->nbio.funcs && - adev->nbio.funcs->handle_ras_err_event_athub_intr_no_bifring) - adev->nbio.funcs->handle_ras_err_event_athub_intr_no_bifring(adev); + if (adev->nbio.ras_funcs && + adev->nbio.ras_funcs->handle_ras_err_event_athub_intr_no_bifring) + 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); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h index ac527e5deae6..cf6116648322 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_irq.h +++ b/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 { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c index ff48101bab55..759b34799221 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c +++ b/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; } } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index a4e2cf7cada1..39ee88d29cca 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/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) && - !amdgpu_is_atpx_hybrid()) + if (amdgpu_device_supports_px(dev)) 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 drm_device *dev = node->minor->dev; - struct amdgpu_device *adev = drm_to_adev(dev); + struct amdgpu_device *adev = (struct amdgpu_device *)m->private; 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", - "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; - } + + seq_printf(m, "TA %s feature version: 0x%08x, firmware version: 0x%08x\n", + ta_fw_name[i], fw_info.feature, fw_info.ver); } /* 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[] = { - {"amdgpu_firmware_info", amdgpu_debugfs_firmware_info, 0, NULL}, -}; +DEFINE_SHOW_ATTRIBUTE(amdgpu_debugfs_firmware_info); + #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, - ARRAY_SIZE(amdgpu_firmware_info_list)); -#else - return 0; + struct drm_minor *minor = adev_to_drm(adev)->primary; + struct dentry *root = minor->debugfs_root; + + debugfs_create_file("amdgpu_firmware_info", 0444, root, + adev, &amdgpu_debugfs_firmware_info_fops); + #endif } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h index 1ae9bdae7311..11aa29933c1f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mmhub.h @@ -21,12 +21,16 @@ #ifndef __AMDGPU_MMHUB_H__ #define __AMDGPU_MMHUB_H__ -struct amdgpu_mmhub_funcs { - void (*ras_init)(struct amdgpu_device *adev); +struct amdgpu_mmhub_ras_funcs { 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); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h index 319cb19e1b99..cb0b581bbce7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h +++ b/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, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h index 7c11bce4514b..25ee53545837 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_nbio.h +++ b/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); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 072050429a2f..1345f7eba011 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/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; - *bo_ptr = NULL; + BUG_ON(bp->bo_ptr_size < sizeof(struct amdgpu_bo)); - acc_size = ttm_bo_dma_acc_size(&adev->mman.bdev, size, - sizeof(struct amdgpu_bo)); - - bo = kzalloc(sizeof(struct amdgpu_bo), GFP_KERNEL); + *bo_ptr = NULL; + bo = kzalloc(bp->bo_ptr_size, 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, - NULL, bp->resv, &amdgpu_bo_destroy); + &bo->placement, page_align, &ctx, NULL, + 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; @@ -695,6 +700,34 @@ int amdgpu_bo_create(struct amdgpu_device *adev, } /** + * 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 */ - arch_io_reserve_memtype_wc(adev->gmc.aper_base, - adev->gmc.aper_size); + /* On A+A platform, VRAM can be mapped as WB */ + if (!adev->gmc.xgmi.connected_to_cpu) { + /* 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); - 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); + if (!adev->gmc.xgmi.connected_to_cpu) { + arch_phys_wc_del(adev->gmc.vram_mtrr); + arch_io_free_memtype_wc(adev->gmc.aper_base, adev->gmc.aper_size); + } } /** @@ -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) { - kfree(bo->metadata); - bo->metadata = NULL; - bo->metadata_size = 0; + if (ubo->metadata_size) { + kfree(ubo->metadata); + ubo->metadata = NULL; + 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); - bo->metadata_flags = flags; - bo->metadata = buffer; - bo->metadata_size = metadata_size; + kfree(ubo->metadata); + ubo->metadata_flags = flags; + ubo->metadata = buffer; + 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) - memcpy(buffer, bo->metadata, bo->metadata_size); + if (ubo->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; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h index 9ac37569823f..2d1fefbe1e99 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/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); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index 839917eb7bc3..9e769cf6095b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/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. |