Merge tag 'spi-v5.20' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi

Pull spi updates from Mark Brown:
"The big update this time around is some excellent work from David
Jander who went through the fast path and really eliminated overheads,
meaning that we are seeing a huge reduction in the time spent between
transfers for single threaded clients.

Benchmarking has been coming out at about a halving of overhead which
is clearly visible in system level usage that stresses SPI like some
CAN and IIO applications, especially with small transfers. Thanks to
David for taking the time to drill down into this and push the work
upstream.

Otherwise there's been a bunch of new device support and the usual
updates.

- Optimisation of the fast path, particularly around the number and
types of locking operations, from David Jander.

- Support for Arbel NPCM845, HP GXP, Intel Meteor Lake and Thunder
Bay, MediaTek MT8188 and MT8365, Microchip FPGAs, nVidia Tegra 241
and Samsung Exynos Auto v9 and 4210"

* tag 'spi-v5.20' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi: (97 commits)
MAINTAINERS: add spi support to GXP
spi: dt-bindings: add documentation for hpe,gxp-spifi
spi: spi-gxp: Add support for HPE GXP SoCs
spi: a3700: support BE for AC5 SPI driver
spi/panel: dt-bindings: drop CPHA and CPOL from common properties
spi: bcm2835: enable shared interrupt support
spi: dt-bindings: spi-controller: correct example indentation
spi: dt-bindings: qcom,spi-geni-qcom: allow three interconnects
spi: npcm-fiu: Add NPCM8XX support
dt-binding: spi: Add npcm845 compatible to npcm-fiu document
spi: npcm-fiu: Modify direct read dummy configuration
spi: atmel: remove #ifdef CONFIG_{PM, SLEEP}
spi: dt-bindings: Add compatible for MediaTek MT8188
spi: dt-bindings: mediatek,spi-mtk-nor: Update bindings for nor flash
spi: dt-bindings: atmel,at91rm9200-spi: convert to json-schema
spi: tegra20-slink: fix UAF in tegra_slink_remove()
spi: Fix simplification of devm_spi_register_controller
spi: microchip-core: switch to use dev_err_probe()
spi: microchip-core: switch to use devm_spi_alloc_master()
spi: microchip-core: fix UAF in mchp_corespi_remove()
...

Documentation/devicetree/bindings/display/panel/lgphilips,lb035q02.yaml

···
   enable-gpios: true
   port: true
 
+  spi-cpha: true
+  spi-cpol: true
+
 required:
   - compatible
   - enable-gpios

Documentation/devicetree/bindings/display/panel/samsung,ld9040.yaml

···
   panel-height-mm:
     description: physical panel height [mm]
 
+  spi-cpha: true
+  spi-cpol: true
+
 required:
   - compatible
   - reg

Documentation/devicetree/bindings/display/panel/sitronix,st7789v.yaml

···
   backlight: true
   port: true
 
+  spi-cpha: true
+  spi-cpol: true
+
 required:
   - compatible
   - reg

Documentation/devicetree/bindings/display/panel/tpo,td.yaml

···
   backlight: true
   port: true
 
+  spi-cpha: true
+  spi-cpol: true
+
 required:
   - compatible
   - port

Documentation/devicetree/bindings/sound/renesas,rz-ssi.yaml

···
   compatible:
     items:
       - enum:
+          - renesas,r9a07g043-ssi  # RZ/G2UL
           - renesas,r9a07g044-ssi  # RZ/G2{L,LC}
           - renesas,r9a07g054-ssi  # RZ/V2L
       - const: renesas,rz-ssi
···
     minItems: 1
     maxItems: 2
     description:
-      The first cell represents a phandle to dmac
+      The first cell represents a phandle to dmac.
       The second cell specifies the encoded MID/RID values of the SSI port
       connected to the DMA client and the slave channel configuration
       parameters.

Documentation/devicetree/bindings/spi/atmel,at91rm9200-spi.yaml

···
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+# Copyright (C) 2022 Microchip Technology, Inc. and its subsidiaries
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/spi/atmel,at91rm9200-spi.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Atmel SPI device
+
+maintainers:
+  - Tudor Ambarus <[e-mail]>
+
+allOf:
+  - $ref: spi-controller.yaml#
+
+properties:
+  compatible:
+    oneOf:
+      - const: atmel,at91rm9200-spi
+      - items:
+          - const: microchip,sam9x60-spi
+          - const: atmel,at91rm9200-spi
+
+  reg:
+    maxItems: 1
+
+  interrupts:
+    maxItems: 1
+
+  clock-names:
+    contains:
+      const: spi_clk
+
+  clocks:
+    maxItems: 1
+
+  atmel,fifo-size:
+    $ref: /schemas/types.yaml#/definitions/uint32
+    description: |
+      Maximum number of data the RX and TX FIFOs can store for FIFO
+      capable SPI controllers.
+    enum: [ 16, 32 ]
+
+required:
+  - compatible
+  - reg
+  - interrupts
+  - clock-names
+  - clocks
+
+unevaluatedProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/gpio/gpio.h>
+    #include <dt-bindings/interrupt-controller/irq.h>
+
+    spi1: spi@fffcc000 {
+        compatible = "atmel,at91rm9200-spi";
+        reg = <0xfffcc000 0x4000>;
+        interrupts = <13 IRQ_TYPE_LEVEL_HIGH 5>;
+        #address-cells = <1>;
+        #size-cells = <0>;
+        clocks = <&spi1_clk>;
+        clock-names = "spi_clk";
+        cs-gpios = <&pioB 3 GPIO_ACTIVE_HIGH>;
+        atmel,fifo-size = <32>;
+
+        mmc@0 {
+            compatible = "mmc-spi-slot";
+            reg = <0>;
+            gpios = <&pioC 4 GPIO_ACTIVE_HIGH>;    /* CD */
+            spi-max-frequency = <25000000>;
+        };
+    };

Documentation/devicetree/bindings/spi/hpe,gxp-spifi.yaml

···
+# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/spi/hpe,gxp-spifi.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: HPE GXP spi controller flash interface
+
+maintainers:
+  - Nick Hawkins <[e-mail]>
+  - Jean-Marie Verdun <[e-mail]>
+
+allOf:
+  - $ref: spi-controller.yaml#
+
+properties:
+  compatible:
+    const: hpe,gxp-spifi
+
+  reg:
+    items:
+      - description: cfg registers
+      - description: data registers
+      - description: mapped memory
+
+  interrupts:
+    maxItems: 1
+
+required:
+  - compatible
+  - reg
+  - interrupts
+
+unevaluatedProperties: false
+
+examples:
+  - |
+
+    spi@200 {
+      compatible = "hpe,gxp-spifi";
+      reg = <0x200 0x80>, <0xc000 0x100>, <0x38000000 0x800000>;
+      interrupts = <20>;
+      interrupt-parent = <&vic0>;
+      #address-cells = <1>;
+      #size-cells = <0>;
+
+      flash@0 {
+        reg = <0>;
+        compatible = "jedec,spi-nor";
+      };
+
+      flash@1 {
+        reg = <1>;
+        compatible = "jedec,spi-nor";
+      };
+    };

Documentation/devicetree/bindings/spi/mediatek,spi-mt65xx.yaml

···
       - items:
           - enum:
               - mediatek,mt7629-spi
+              - mediatek,mt8365-spi
           - const: mediatek,mt7622-spi
       - items:
           - enum:
···
       - items:
           - enum:
               - mediatek,mt7986-spi-ipm
+              - mediatek,mt8188-spi-ipm
           - const: mediatek,spi-ipm
       - items:
           - enum:

Documentation/devicetree/bindings/spi/mediatek,spi-mtk-nor.yaml

···
 properties:
   compatible:
     oneOf:
+      - enum:
+          - mediatek,mt8173-nor
+          - mediatek,mt8186-nor
+          - mediatek,mt8192-nor
       - items:
           - enum:
               - mediatek,mt2701-nor
···
               - mediatek,mt7622-nor
               - mediatek,mt7623-nor
               - mediatek,mt7629-nor
-              - mediatek,mt8186-nor
-              - mediatek,mt8192-nor
               - mediatek,mt8195-nor
-          - enum:
-              - mediatek,mt8173-nor
-      - items:
           - const: mediatek,mt8173-nor
+      - items:
+          - enum:
+              - mediatek,mt8188-nor
+          - const: mediatek,mt8186-nor
+
   reg:
     maxItems: 1
 
···
 required:
   - compatible
   - reg
-  - interrupts
   - clocks
   - clock-names
 

Documentation/devicetree/bindings/spi/nuvoton,npcm-fiu.txt

···
 FIU0 and FIUx supports two chip selects,
 FIU3 support four chip select.
 
+The NPCM8XX supports four FIU modules,
+FIU0 and FIUx supports two chip selects,
+FIU1 and FIU3 supports four chip selects.
+
 Required properties:
-  - compatible : "nuvoton,npcm750-fiu" for the NPCM7XX BMC
+  - compatible : "nuvoton,npcm750-fiu" for Poleg NPCM7XX BMC
+			     "nuvoton,npcm845-fiu" for Arbel NPCM8XX BMC
   - #address-cells : should be 1.
   - #size-cells : should be 0.
   - reg : the first contains the register location and length,
···
   		fiu0 represent fiu 0 controller
   		fiu1 represent fiu 3 controller
   		fiu2 represent fiu x controller
+
+  In the NPCM8XX BMC:
+  		fiu0 represent fiu 0 controller
+  		fiu1 represent fiu 1 controller
+  		fiu2 represent fiu 3 controller
+  		fiu3 represent fiu x controller
 
 Example:
 fiu3: spi@c00000000 {

Documentation/devicetree/bindings/spi/nvidia,tegra210-quad-peripheral-props.yaml

···
+# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/spi/nvidia,tegra210-quad-peripheral-props.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Peripheral properties for Tegra Quad SPI Controller
+
+maintainers:
+  - Thierry Reding <[e-mail]>
+  - Jonathan Hunter <[e-mail]>
+
+properties:
+  nvidia,tx-clk-tap-delay:
+    description:
+      Delays the clock going out to device with this tap value.
+      Tap value varies based on platform design trace lengths from Tegra
+      QSPI to corresponding slave device.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    minimum: 0
+    maximum: 31
+
+  nvidia,rx-clk-tap-delay:
+    description:
+      Delays the clock coming in from the device with this tap value.
+      Tap value varies based on platform design trace lengths from Tegra
+      QSPI to corresponding slave device.
+    $ref: /schemas/types.yaml#/definitions/uint32
+    minimum: 0
+    maximum: 255
+
+unevaluatedProperties: true
+

Documentation/devicetree/bindings/spi/nvidia,tegra210-quad.yaml

···
       - nvidia,tegra186-qspi
       - nvidia,tegra194-qspi
       - nvidia,tegra234-qspi
+      - nvidia,tegra241-qspi
 
   reg:
     maxItems: 1
···
 
       spi-tx-bus-width:
         enum: [1, 2, 4]
-
-      nvidia,tx-clk-tap-delay:
-        description:
-          Delays the clock going out to device with this tap value.
-          Tap value varies based on platform design trace lengths from Tegra
-          QSPI to corresponding slave device.
-        $ref: /schemas/types.yaml#/definitions/uint32
-        minimum: 0
-        maximum: 31
-
-      nvidia,rx-clk-tap-delay:
-        description:
-          Delays the clock coming in from the device with this tap value.
-          Tap value varies based on platform design trace lengths from Tegra
-          QSPI to corresponding slave device.
-        $ref: /schemas/types.yaml#/definitions/uint32
-        minimum: 0
-        maximum: 255
-
-    required:
-      - reg
 
 required:
   - compatible

Documentation/devicetree/bindings/spi/qcom,spi-geni-qcom.yaml

···
       - const: rx
 
   interconnects:
-    maxItems: 2
+    minItems: 2
+    maxItems: 3
 
   interconnect-names:
+    minItems: 2
     items:
       - const: qup-core
       - const: qup-config
+      - const: qup-memory
 
   interrupts:
     maxItems: 1

Documentation/devicetree/bindings/spi/samsung,spi.yaml

···
           - samsung,s3c2443-spi # for S3C2443, S3C2416 and S3C2450
           - samsung,s3c6410-spi
           - samsung,s5pv210-spi # for S5PV210 and S5PC110
+          - samsung,exynos4210-spi
           - samsung,exynos5433-spi
+          - samsung,exynosautov9-spi
           - tesla,fsd-spi
       - const: samsung,exynos7-spi
         deprecated: true
···
       properties:
         compatible:
           contains:
-            const: samsung,exynos5433-spi
+            enum:
+              - samsung,exynos5433-spi
+              - samsung,exynosautov9-spi
     then:
       properties:
         clocks:

Documentation/devicetree/bindings/spi/snps,dw-apb-ssi.yaml

···
           - const: snps,dw-apb-ssi
       - description: Intel Keem Bay SPI Controller
         const: intel,keembay-ssi
+      - description: Intel Thunder Bay SPI Controller
+        const: intel,thunderbay-ssi
       - description: Baikal-T1 SPI Controller
         const: baikal,bt1-ssi
       - description: Baikal-T1 System Boot SPI Controller
···
 
   rx-sample-delay-ns:
     default: 0
-    description: Default value of the rx-sample-delay-ns property.
+    description: |
+      Default value of the rx-sample-delay-ns property.
       This value will be used if the property is not explicitly defined
-      for a SPI slave device. See below.
+      for a SPI slave device.
+
+      SPI Rx sample delay offset, unit is nanoseconds.
+      The delay from the default sample time before the actual sample of the
+      rxd input signal occurs. The "rx_sample_delay" is an optional feature
+      of the designware controller, and the upper limit is also subject to
+      controller configuration.
 
 patternProperties:
   "^.*@[0-9a-f]+$":
···
       reg:
         minimum: 0
         maximum: 3
-
-      spi-rx-bus-width:
-        const: 1
-
-      spi-tx-bus-width:
-        const: 1
-
-      rx-sample-delay-ns:
-        description: SPI Rx sample delay offset, unit is nanoseconds.
-          The delay from the default sample time before the actual
-          sample of the rxd input signal occurs. The "rx_sample_delay"
-          is an optional feature of the designware controller, and the
-          upper limit is also subject to controller configuration.
 
 unevaluatedProperties: false
 

Documentation/devicetree/bindings/spi/spi-cadence.yaml

···
     enum: [ 0, 1 ]
     default: 0
 
+required:
+  - compatible
+  - reg
+  - interrupts
+  - clock-names
+  - clocks
+
 unevaluatedProperties: false
 
 examples:

Documentation/devicetree/bindings/spi/spi-controller.yaml

···
     type: object
     $ref: spi-peripheral-props.yaml
 
+    properties:
+      spi-cpha:
+        $ref: /schemas/types.yaml#/definitions/flag
+        description:
+          The device requires shifted clock phase (CPHA) mode.
+
+      spi-cpol:
+        $ref: /schemas/types.yaml#/definitions/flag
+        description:
+          The device requires inverse clock polarity (CPOL) mode.
+
     required:
       - compatible
       - reg
···
         };
 
         flash@2 {
-          compatible = "jedec,spi-nor";
-          spi-max-frequency = <50000000>;
-          reg = <2>, <3>;
-          stacked-memories = /bits/ 64 <0x10000000 0x10000000>;
+            compatible = "jedec,spi-nor";
+            spi-max-frequency = <50000000>;
+            reg = <2>, <3>;
+            stacked-memories = /bits/ 64 <0x10000000 0x10000000>;
         };
     };

Documentation/devicetree/bindings/spi/spi-peripheral-props.yaml

···
     description:
       The device requires 3-wire mode.
 
-  spi-cpha:
-    $ref: /schemas/types.yaml#/definitions/flag
-    description:
-      The device requires shifted clock phase (CPHA) mode.
-
-  spi-cpol:
-    $ref: /schemas/types.yaml#/definitions/flag
-    description:
-      The device requires inverse clock polarity (CPOL) mode.
-
   spi-cs-high:
     $ref: /schemas/types.yaml#/definitions/flag
     description:
···
   spi-rx-delay-us:
     description:
       Delay, in microseconds, after a read transfer.
+
+  rx-sample-delay-ns:
+    description: SPI Rx sample delay offset, unit is nanoseconds.
+      The delay from the default sample time before the actual
+      sample of the rxd input signal occurs.
 
   spi-tx-bus-width:
     description:
···
 allOf:
   - $ref: cdns,qspi-nor-peripheral-props.yaml#
   - $ref: samsung,spi-peripheral-props.yaml#
+  - $ref: nvidia,tegra210-quad-peripheral-props.yaml#
 
 additionalProperties: true

Documentation/devicetree/bindings/spi/spi-zynqmp-qspi.yaml

···
   clocks:
     maxItems: 2
 
+required:
+  - compatible
+  - reg
+  - interrupts
+  - clock-names
+  - clocks
+
 unevaluatedProperties: false
 
 examples:

Documentation/devicetree/bindings/spi/spi_atmel.txt

···
-Atmel SPI device
-
-Required properties:
-- compatible : should be "atmel,at91rm9200-spi" or "microchip,sam9x60-spi".
-- reg: Address and length of the register set for the device
-- interrupts: Should contain spi interrupt
-- cs-gpios: chipselects (optional for SPI controller version >= 2 with the
-  Chip Select Active After Transfer feature).
-- clock-names: tuple listing input clock names.
-	Required elements: "spi_clk"
-- clocks: phandles to input clocks.
-
-Optional properties:
-- atmel,fifo-size: maximum number of data the RX and TX FIFOs can store for FIFO
-  capable SPI controllers.
-
-Example:
-
-spi1: spi@fffcc000 {
-	compatible = "atmel,at91rm9200-spi";
-	reg = <0xfffcc000 0x4000>;
-	interrupts = <13 4 5>;
-	#address-cells = <1>;
-	#size-cells = <0>;
-	clocks = <&spi1_clk>;
-	clock-names = "spi_clk";
-	cs-gpios = <&pioB 3 0>;
-	atmel,fifo-size = <32>;
-
-	mmc-slot@0 {
-		compatible = "mmc-spi-slot";
-		reg = <0>;
-		gpios = <&pioC 4 0>;	/* CD */
-		spi-max-frequency = <25000000>;
-	};
-};

MAINTAINERS

···
 M:	Nick Hawkins <[e-mail]>
 S:	Maintained
 F:	Documentation/devicetree/bindings/arm/hpe,gxp.yaml
+F:	Documentation/devicetree/bindings/spi/hpe,gxp-spi.yaml
 F:	Documentation/devicetree/bindings/timer/hpe,gxp-timer.yaml
 F:	arch/arm/boot/dts/hpe-bmc*
 F:	arch/arm/boot/dts/hpe-gxp*
 F:	arch/arm/mach-hpe/
 F:	drivers/clocksource/timer-gxp.c
+F:	drivers/spi/spi-gxp.c
 F:	drivers/watchdog/gxp-wdt.c
 
 ARM/IGEP MACHINE SUPPORT
···
 F:	drivers/mailbox/mailbox-mpfs.c
 F:	drivers/pci/controller/pcie-microchip-host.c
 F:	drivers/soc/microchip/
+F:	drivers/spi/spi-microchip-core.c
 F:	include/soc/microchip/mpfs.h
 
 RNBD BLOCK DRIVERS

drivers/spi/Kconfig

···
 	  This controller does not support generic SPI messages. It only
 	  supports the high-level SPI memory interface.
 
+config SPI_GXP
+	tristate "GXP SPI driver"
+	depends on ARCH_HPE || COMPILE_TEST
+	help
+	  This enables support for the driver for GXP bus attached SPI
+	  controllers.
+
 config SPI_HISI_KUNPENG
 	tristate "HiSilicon SPI Controller for Kunpeng SoCs"
 	depends on (ARM64 && ACPI) || COMPILE_TEST
···
 	help
 	  This enables master mode support for the SPIFC (SPI flash
 	  controller) available in Amlogic Meson SoCs.
+
+config SPI_MICROCHIP_CORE
+	tristate "Microchip FPGA SPI controllers"
+	depends on SPI_MASTER
+	help
+	  This enables the SPI driver for Microchip FPGA SPI controllers.
+	  Say Y or M here if you want to use the "hard" controllers on
+	  PolarFire SoC.
+	  If built as a module, it will be called spi-microchip-core.
 
 config SPI_MT65XX
 	tristate "MediaTek SPI controller"

drivers/spi/Makefile

···
 obj-$(CONFIG_SPI_FSL_QUADSPI)		+= spi-fsl-qspi.o
 obj-$(CONFIG_SPI_FSL_SPI)		+= spi-fsl-spi.o
 obj-$(CONFIG_SPI_GPIO)			+= spi-gpio.o
+obj-$(CONFIG_SPI_GXP)			+= spi-gxp.o
 obj-$(CONFIG_SPI_HISI_KUNPENG)		+= spi-hisi-kunpeng.o
 obj-$(CONFIG_SPI_HISI_SFC_V3XX)		+= spi-hisi-sfc-v3xx.o
 obj-$(CONFIG_SPI_IMG_SPFI)		+= spi-img-spfi.o
···
 obj-$(CONFIG_SPI_LP8841_RTC)		+= spi-lp8841-rtc.o
 obj-$(CONFIG_SPI_MESON_SPICC)		+= spi-meson-spicc.o
 obj-$(CONFIG_SPI_MESON_SPIFC)		+= spi-meson-spifc.o
+obj-$(CONFIG_SPI_MICROCHIP_CORE)	+= spi-microchip-core.o
 obj-$(CONFIG_SPI_MPC512x_PSC)		+= spi-mpc512x-psc.o
 obj-$(CONFIG_SPI_MPC52xx_PSC)		+= spi-mpc52xx-psc.o
 obj-$(CONFIG_SPI_MPC52xx)		+= spi-mpc52xx.o

drivers/spi/atmel-quadspi.c

···
 #include <linux/of.h>
 #include <linux/of_platform.h>
 #include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
 #include <linux/spi/spi-mem.h>
 
 /* QSPI register offsets */
···
 
 	/* special case not supported by hardware */
 	if (op->addr.nbytes == 2 && op->cmd.buswidth != op->addr.buswidth &&
-		op->dummy.nbytes == 0)
+	    op->dummy.nbytes == 0)
 		return false;
 
 	return true;
···
 	if (op->addr.val + op->data.nbytes > aq->mmap_size)
 		return -ENOTSUPP;
 
+	err = pm_runtime_resume_and_get(&aq->pdev->dev);
+	if (err < 0)
+		return err;
+
 	err = atmel_qspi_set_cfg(aq, op, &offset);
 	if (err)
-		return err;
+		goto pm_runtime_put;
 
 	/* Skip to the final steps if there is no data */
 	if (op->data.nbytes) {
···
 	/* Poll INSTRuction End status */
 	sr = atmel_qspi_read(aq, QSPI_SR);
 	if ((sr & QSPI_SR_CMD_COMPLETED) == QSPI_SR_CMD_COMPLETED)
-		return err;
+		goto pm_runtime_put;
 
 	/* Wait for INSTRuction End interrupt */
 	reinit_completion(&aq->cmd_completion);
···
 		err = -ETIMEDOUT;
 	atmel_qspi_write(QSPI_SR_CMD_COMPLETED, aq, QSPI_IDR);
 
+pm_runtime_put:
+	pm_runtime_mark_last_busy(&aq->pdev->dev);
+	pm_runtime_put_autosuspend(&aq->pdev->dev);
 	return err;
 }
 
···
 	struct atmel_qspi *aq = spi_controller_get_devdata(ctrl);
 	unsigned long src_rate;
 	u32 scbr;
+	int ret;
 
 	if (ctrl->busy)
 		return -EBUSY;
···
 	if (scbr > 0)
 		scbr--;
 
+	ret = pm_runtime_resume_and_get(ctrl->dev.parent);
+	if (ret < 0)
+		return ret;
+
 	aq->scr = QSPI_SCR_SCBR(scbr);
 	atmel_qspi_write(aq->scr, aq, QSPI_SCR);
+
+	pm_runtime_mark_last_busy(ctrl->dev.parent);
+	pm_runtime_put_autosuspend(ctrl->dev.parent);
 
 	return 0;
 }
···
 	if (err)
 		goto disable_qspick;
 
+	pm_runtime_set_autosuspend_delay(&pdev->dev, 500);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_runtime_set_active(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+	pm_runtime_get_noresume(&pdev->dev);
+
 	atmel_qspi_init(aq);
 
 	err = spi_register_controller(ctrl);
-	if (err)
+	if (err) {
+		pm_runtime_put_noidle(&pdev->dev);
+		pm_runtime_disable(&pdev->dev);
+		pm_runtime_set_suspended(&pdev->dev);
+		pm_runtime_dont_use_autosuspend(&pdev->dev);
 		goto disable_qspick;
+	}
+	pm_runtime_mark_last_busy(&pdev->dev);
+	pm_runtime_put_autosuspend(&pdev->dev);
 
 	return 0;
 
···
 {
 	struct spi_controller *ctrl = platform_get_drvdata(pdev);
 	struct atmel_qspi *aq = spi_controller_get_devdata(ctrl);
+	int ret;
+
+	ret = pm_runtime_resume_and_get(&pdev->dev);
+	if (ret < 0)
+		return ret;
 
 	spi_unregister_controller(ctrl);
 	atmel_qspi_write(QSPI_CR_QSPIDIS, aq, QSPI_CR);
+
+	pm_runtime_disable(&pdev->dev);
+	pm_runtime_put_noidle(&pdev->dev);
+
 	clk_disable_unprepare(aq->qspick);
 	clk_disable_unprepare(aq->pclk);
 	return 0;
···
 {
 	struct spi_controller *ctrl = dev_get_drvdata(dev);
 	struct atmel_qspi *aq = spi_controller_get_devdata(ctrl);
+	int ret;
+
+	ret = pm_runtime_resume_and_get(dev);
+	if (ret < 0)
+		return ret;
 
 	atmel_qspi_write(QSPI_CR_QSPIDIS, aq, QSPI_CR);
-	clk_disable_unprepare(aq->qspick);
-	clk_disable_unprepare(aq->pclk);
+
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_force_suspend(dev);
+
+	clk_unprepare(aq->qspick);
+	clk_unprepare(aq->pclk);
 
 	return 0;
 }
···
 {
 	struct spi_controller *ctrl = dev_get_drvdata(dev);
 	struct atmel_qspi *aq = spi_controller_get_devdata(ctrl);
+	int ret;
 
-	clk_prepare_enable(aq->pclk);
-	clk_prepare_enable(aq->qspick);
+	clk_prepare(aq->pclk);
+	clk_prepare(aq->qspick);
+
+	ret = pm_runtime_force_resume(dev);
+	if (ret < 0)
+		return ret;
 
 	atmel_qspi_init(aq);
 
 	atmel_qspi_write(aq->scr, aq, QSPI_SCR);
 
+	pm_runtime_mark_last_busy(dev);
+	pm_runtime_put_autosuspend(dev);
+
 	return 0;
 }
 
-static SIMPLE_DEV_PM_OPS(atmel_qspi_pm_ops, atmel_qspi_suspend,
-			 atmel_qspi_resume);
+static int __maybe_unused atmel_qspi_runtime_suspend(struct device *dev)
+{
+	struct spi_controller *ctrl = dev_get_drvdata(dev);
+	struct atmel_qspi *aq = spi_controller_get_devdata(ctrl);
+
+	clk_disable(aq->qspick);
+	clk_disable(aq->pclk);
+
+	return 0;
+}
+
+static int __maybe_unused atmel_qspi_runtime_resume(struct device *dev)
+{
+	struct spi_controller *ctrl = dev_get_drvdata(dev);
+	struct atmel_qspi *aq = spi_controller_get_devdata(ctrl);
+	int ret;
+
+	ret = clk_enable(aq->pclk);
+	if (ret)
+		return ret;
+
+	return clk_enable(aq->qspick);
+}
+
+static const struct dev_pm_ops __maybe_unused atmel_qspi_pm_ops = {
+	SET_SYSTEM_SLEEP_PM_OPS(atmel_qspi_suspend, atmel_qspi_resume)
+	SET_RUNTIME_PM_OPS(atmel_qspi_runtime_suspend,
+			   atmel_qspi_runtime_resume, NULL)
+};
 
 static const struct atmel_qspi_caps atmel_sama5d2_qspi_caps = {};
 
···
 	.driver = {
 		.name	= "atmel_qspi",
 		.of_match_table	= atmel_qspi_dt_ids,
-		.pm	= &atmel_qspi_pm_ops,
+		.pm	= pm_ptr(&atmel_qspi_pm_ops),
 	},
 	.probe		= atmel_qspi_probe,
 	.remove		= atmel_qspi_remove,

drivers/spi/spi-altera-dfl.c

···
 	struct spi_master *master;
 	struct altera_spi *hw;
 	void __iomem *base;
-	int err = -ENODEV;
+	int err;
 
-	master = spi_alloc_master(dev, sizeof(struct altera_spi));
+	master = devm_spi_alloc_master(dev, sizeof(struct altera_spi));
 	if (!master)
 		return -ENOMEM;
 
···
 	altera_spi_init_master(master);
 
 	err = devm_spi_register_master(dev, master);
-	if (err) {
-		dev_err(dev, "%s failed to register spi master %d\n", __func__, err);
-		goto exit;
-	}
+	if (err)
+		return dev_err_probe(dev, err, "%s failed to register spi master\n",
+				     __func__);
 
 	if (dfl_dev->revision == FME_FEATURE_REV_MAX10_SPI_N5010)
 		strscpy(board_info.modalias, "m10-n5010", SPI_NAME_SIZE);
···
 	}
 
 	return 0;
-exit:
-	spi_master_put(master);
-	return err;
 }
 
 static const struct dfl_device_id dfl_spi_altera_ids[] = {

drivers/spi/spi-amd.c

···
 #define AMD_SPI_XFER_TX		1
 #define AMD_SPI_XFER_RX		2
 
+/**
+ * enum amd_spi_versions - SPI controller versions
+ * @AMD_SPI_V1:		AMDI0061 hardware version
+ * @AMD_SPI_V2:		AMDI0062 hardware version
+ */
 enum amd_spi_versions {
-	AMD_SPI_V1 = 1,	/* AMDI0061 */
-	AMD_SPI_V2,	/* AMDI0062 */
+	AMD_SPI_V1 = 1,
+	AMD_SPI_V2,
 };
 
+/**
+ * struct amd_spi - SPI driver instance
+ * @io_remap_addr:	Start address of the SPI controller registers
+ * @version:		SPI controller hardware version
+ */
 struct amd_spi {
 	void __iomem *io_remap_addr;
-	unsigned long io_base_addr;
 	enum amd_spi_versions version;
 };
 
···
 	struct device *dev = &pdev->dev;
 	struct spi_master *master;
 	struct amd_spi *amd_spi;
-	int err = 0;
+	int err;
 
 	/* Allocate storage for spi_master and driver private data */
-	master = spi_alloc_master(dev, sizeof(struct amd_spi));
-	if (!master) {
-		dev_err(dev, "Error allocating SPI master\n");
-		return -ENOMEM;
-	}
+	master = devm_spi_alloc_master(dev, sizeof(struct amd_spi));
+	if (!master)
+		return dev_err_probe(dev, -ENOMEM, "Error allocating SPI master\n");
 
 	amd_spi = spi_master_get_devdata(master);
 	amd_spi->io_remap_addr = devm_platform_ioremap_resource(pdev, 0);
-	if (IS_ERR(amd_spi->io_remap_addr)) {
-		err = PTR_ERR(amd_spi->io_remap_addr);
-		dev_err(dev, "error %d ioremap of SPI registers failed\n", err);
-		goto err_free_master;
-	}
+	if (IS_ERR(amd_spi->io_remap_addr))
+		return dev_err_probe(dev, PTR_ERR(amd_spi->io_remap_addr),
+				     "ioremap of SPI registers failed\n");
+
 	dev_dbg(dev, "io_remap_address: %p\n", amd_spi->io_remap_addr);
 
 	amd_spi->version = (enum amd_spi_versions) device_get_match_data(dev);
···
 
 	/* Register the controller with SPI framework */
 	err = devm_spi_register_master(dev, master);
-	if (err) {
-		dev_err(dev, "error %d registering SPI controller\n", err);
-		goto err_free_master;
-	}
+	if (err)
+		return dev_err_probe(dev, err, "error registering SPI controller\n");
 
 	return 0;
-
-err_free_master:
-	spi_master_put(master);
-
-	return err;
 }
 
 #ifdef CONFIG_ACPI

drivers/spi/spi-armada-3700.c

···
 
 	while (!a3700_is_wfifo_full(a3700_spi) && a3700_spi->buf_len) {
 		val = *(u32 *)a3700_spi->tx_buf;
-		spireg_write(a3700_spi, A3700_SPI_DATA_OUT_REG, val);
+		spireg_write(a3700_spi, A3700_SPI_DATA_OUT_REG, cpu_to_le32(val));
 		a3700_spi->buf_len -= 4;
 		a3700_spi->tx_buf += 4;
 	}
···
 	while (!a3700_is_rfifo_empty(a3700_spi) && a3700_spi->buf_len) {
 		val = spireg_read(a3700_spi, A3700_SPI_DATA_IN_REG);
 		if (a3700_spi->buf_len >= 4) {
-
+			val = le32_to_cpu(val);
 			memcpy(a3700_spi->rx_buf, &val, 4);
 
 			a3700_spi->buf_len -= 4;

drivers/spi/spi-atmel.c

···
 	return 0;
 }
 
-#ifdef CONFIG_PM
 static int atmel_spi_runtime_suspend(struct device *dev)
 {
 	struct spi_master *master = dev_get_drvdata(dev);
···
 	return clk_prepare_enable(as->clk);
 }
 
-#ifdef CONFIG_PM_SLEEP
 static int atmel_spi_suspend(struct device *dev)
 {
 	struct spi_master *master = dev_get_drvdata(dev);
···
 	/* Start the queue running */
 	return spi_master_resume(master);
 }
-#endif
 
 static const struct dev_pm_ops atmel_spi_pm_ops = {
-	SET_SYSTEM_SLEEP_PM_OPS(atmel_spi_suspend, atmel_spi_resume)
-	SET_RUNTIME_PM_OPS(atmel_spi_runtime_suspend,
-			   atmel_spi_runtime_resume, NULL)
+	SYSTEM_SLEEP_PM_OPS(atmel_spi_suspend, atmel_spi_resume)
+	RUNTIME_PM_OPS(atmel_spi_runtime_suspend,
+		       atmel_spi_runtime_resume, NULL)
 };
-#define ATMEL_SPI_PM_OPS	(&atmel_spi_pm_ops)
-#else
-#define ATMEL_SPI_PM_OPS	NULL
-#endif
 
 static const struct of_device_id atmel_spi_dt_ids[] = {
 	{ .compatible = "atmel,at91rm9200-spi" },
···
 static struct platform_driver atmel_spi_driver = {
 	.driver		= {
 		.name	= "atmel_spi",
-		.pm	= ATMEL_SPI_PM_OPS,
+		.pm	= pm_ptr(&atmel_spi_pm_ops),
 		.of_match_table	= atmel_spi_dt_ids,
 	},
 	.probe		= atmel_spi_probe,

drivers/spi/spi-bcm2835.c

···
 	struct bcm2835_spi *bs = dev_id;
 	u32 cs = bcm2835_rd(bs, BCM2835_SPI_CS);
 
+	/* Bail out early if interrupts are not enabled */
+	if (!(cs & BCM2835_SPI_CS_INTR))
+		return IRQ_NONE;
+
 	/*
 	 * An interrupt is signaled either if DONE is set (TX FIFO empty)
 	 * or if RXR is set (RX FIFO >= ¾ full).
···
 	bcm2835_wr(bs, BCM2835_SPI_CS,
 		   BCM2835_SPI_CS_CLEAR_RX | BCM2835_SPI_CS_CLEAR_TX);
 
-	err = devm_request_irq(&pdev->dev, bs->irq, bcm2835_spi_interrupt, 0,
-			       dev_name(&pdev->dev), bs);
+	err = devm_request_irq(&pdev->dev, bs->irq, bcm2835_spi_interrupt,
+			       IRQF_SHARED, dev_name(&pdev->dev), bs);
 	if (err) {
 		dev_err(&pdev->dev, "could not request IRQ: %d\n", err);
 		goto out_dma_release;

drivers/spi/spi-dw-core.c

···
 		if (spi->mode & SPI_LOOP)
 			cr0 |= DW_HSSI_CTRLR0_SRL;
 
-		if (dws->caps & DW_SPI_CAP_KEEMBAY_MST)
-			cr0 |= DW_HSSI_CTRLR0_KEEMBAY_MST;
+		/* CTRLR0[31] MST */
+		if (dw_spi_ver_is_ge(dws, HSSI, 102A))
+			cr0 |= DW_HSSI_CTRLR0_MST;
 	}
 
 	return cr0;
···
 
 	if (dws->dma_ops && dws->dma_ops->dma_init) {
 		ret = dws->dma_ops->dma_init(dev, dws);
-		if (ret) {
+		if (ret == -EPROBE_DEFER) {
+			goto err_free_irq;
+		} else if (ret) {
 			dev_warn(dev, "DMA init failed\n");
 		} else {
 			master->can_dma = dws->dma_ops->can_dma;
···
 	if (dws->dma_ops && dws->dma_ops->dma_exit)
 		dws->dma_ops->dma_exit(dws);
 	dw_spi_enable_chip(dws, 0);
+err_free_irq:
 	free_irq(dws->irq, master);
 err_free_master:
 	spi_controller_put(master);

drivers/spi/spi-dw-dma.c

···
 
 static int dw_spi_dma_init_generic(struct device *dev, struct dw_spi *dws)
 {
-	dws->rxchan = dma_request_slave_channel(dev, "rx");
-	if (!dws->rxchan)
-		return -ENODEV;
+	int ret;
 
-	dws->txchan = dma_request_slave_channel(dev, "tx");
-	if (!dws->txchan) {
-		dma_release_channel(dws->rxchan);
+	dws->rxchan = dma_request_chan(dev, "rx");
+	if (IS_ERR(dws->rxchan)) {
+		ret = PTR_ERR(dws->rxchan);
 		dws->rxchan = NULL;
-		return -ENODEV;
+		goto err_exit;
+	}
+
+	dws->txchan = dma_request_chan(dev, "tx");
+	if (IS_ERR(dws->txchan)) {
+		ret = PTR_ERR(dws->txchan);
+		dws->txchan = NULL;
+		goto free_rxchan;
 	}
 
 	dws->master->dma_rx = dws->rxchan;
···
 	dw_spi_dma_sg_burst_init(dws);
 
 	return 0;
+
+free_rxchan:
+	dma_release_channel(dws->rxchan);
+	dws->rxchan = NULL;
+err_exit:
+	return ret;
 }
 
 static void dw_spi_dma_exit(struct dw_spi *dws)

drivers/spi/spi-dw-mmio.c

···
 	return 0;
 }
 
-static int dw_spi_keembay_init(struct platform_device *pdev,
-			       struct dw_spi_mmio *dwsmmio)
+static int dw_spi_intel_init(struct platform_device *pdev,
+			     struct dw_spi_mmio *dwsmmio)
 {
 	dwsmmio->dws.ip = DW_HSSI_ID;
-	dwsmmio->dws.caps = DW_SPI_CAP_KEEMBAY_MST;
 
 	return 0;
 }
···
 	{ .compatible = "amazon,alpine-dw-apb-ssi", .data = dw_spi_alpine_init},
 	{ .compatible = "renesas,rzn1-spi", .data = dw_spi_pssi_init},
 	{ .compatible = "snps,dwc-ssi-1.01a", .data = dw_spi_hssi_init},
-	{ .compatible = "intel,keembay-ssi", .data = dw_spi_keembay_init},
+	{ .compatible = "intel,keembay-ssi", .data = dw_spi_intel_init},
+	{ .compatible = "intel,thunderbay-ssi", .data = dw_spi_intel_init},
 	{ .compatible = "microchip,sparx5-spi", dw_spi_mscc_sparx5_init},
 	{ .compatible = "canaan,k210-spi", dw_spi_canaan_k210_init},
 	{ /* end of table */}

drivers/spi/spi-dw.h

···
 	((_dws)->ip == DW_ ## _ip ## _ID)
 
 #define __dw_spi_ver_cmp(_dws, _ip, _ver, _op) \
-	(dw_spi_ip_is(_dws, _ip) && (_dws)->ver _op DW_ ## _ip ## _ver)
+	(dw_spi_ip_is(_dws, _ip) && (_dws)->ver _op DW_ ## _ip ## _ ## _ver)
 
 #define dw_spi_ver_is(_dws, _ip, _ver) __dw_spi_ver_cmp(_dws, _ip, _ver, ==)
 
···
 
 /* DW SPI controller capabilities */
 #define DW_SPI_CAP_CS_OVERRIDE		BIT(0)
-#define DW_SPI_CAP_KEEMBAY_MST		BIT(1)
-#define DW_SPI_CAP_DFS32		BIT(2)
+#define DW_SPI_CAP_DFS32		BIT(1)
 
 /* Register offsets (Generic for both DWC APB SSI and DWC SSI IP-cores) */
 #define DW_SPI_CTRLR0			0x00
···
 #define DW_HSSI_CTRLR0_SCPOL			BIT(9)
 #define DW_HSSI_CTRLR0_TMOD_MASK		GENMASK(11, 10)
 #define DW_HSSI_CTRLR0_SRL			BIT(13)
-
-/*
- * For Keem Bay, CTRLR0[31] is used to select controller mode.
- * 0: SSI is slave
- * 1: SSI is master
- */
-#define DW_HSSI_CTRLR0_KEEMBAY_MST		BIT(31)
+#define DW_HSSI_CTRLR0_MST			BIT(31)
 
 /* Bit fields in CTRLR1 */
 #define DW_SPI_NDF_MASK				GENMASK(15, 0)

drivers/spi/spi-fsi.c

···
 #define FSI2SPI_IRQ			0x20
 
 #define SPI_FSI_BASE			0x70000
-#define SPI_FSI_INIT_TIMEOUT_MS		1000
-#define SPI_FSI_STATUS_TIMEOUT_MS	100
+#define SPI_FSI_TIMEOUT_MS		1000
 #define SPI_FSI_MAX_RX_SIZE		8
 #define SPI_FSI_MAX_TX_SIZE		40
 
···
 static int fsi_spi_transfer_data(struct fsi_spi *ctx,
 				 struct spi_transfer *transfer)
 {
+	int loops;
 	int rc = 0;
 	unsigned long end;
 	u64 status = 0ULL;
···
 			if (rc)
 				return rc;
 
-			end = jiffies + msecs_to_jiffies(SPI_FSI_STATUS_TIMEOUT_MS);
+			loops = 0;
+			end = jiffies + msecs_to_jiffies(SPI_FSI_TIMEOUT_MS);
 			do {
-				if (time_after(jiffies, end))
+				if (loops++ && time_after(jiffies, end))
 					return -ETIMEDOUT;
 
 				rc = fsi_spi_status(ctx, &status, "TX");
···
 		u8 *rx = transfer->rx_buf;
 
 		while (transfer->len > recv) {
-			end = jiffies + msecs_to_jiffies(SPI_FSI_STATUS_TIMEOUT_MS);
+			loops = 0;
+			end = jiffies + msecs_to_jiffies(SPI_FSI_TIMEOUT_MS);
 			do {
-				if (time_after(jiffies, end))
+				if (loops++ && time_after(jiffies, end))
 					return -ETIMEDOUT;
 
 				rc = fsi_spi_status(ctx, &status, "RX");
···
 
 static int fsi_spi_transfer_init(struct fsi_spi *ctx)
 {
+	int loops = 0;
 	int rc;
 	bool reset = false;
 	unsigned long end;
···
 		SPI_FSI_CLOCK_CFG_SCK_NO_DEL |
 		FIELD_PREP(SPI_FSI_CLOCK_CFG_SCK_DIV, 19);
 
-	end = jiffies + msecs_to_jiffies(SPI_FSI_INIT_TIMEOUT_MS);
+	end = jiffies + msecs_to_jiffies(SPI_FSI_TIMEOUT_MS);
 	do {
-		if (time_after(jiffies, end))
+		if (loops++ && time_after(jiffies, end))
 			return -ETIMEDOUT;
 
 		rc = fsi_spi_read_reg(ctx, SPI_FSI_STATUS, &status);

drivers/spi/spi-gxp.c

···
+// SPDX-License-Identifier: GPL-2.0=or-later
+/* Copyright (C) 2022 Hewlett-Packard Development Company, L.P. */
+
+#include <linux/iopoll.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/spi/spi.h>
+#include <linux/spi/spi-mem.h>
+
+#define GXP_SPI0_MAX_CHIPSELECT	2
+#define GXP_SPI_SLEEP_TIME	1
+#define GXP_SPI_TIMEOUT (130 * 1000000 / GXP_SPI_SLEEP_TIME)
+
+#define MANUAL_MODE		0
+#define DIRECT_MODE		1
+#define SPILDAT_LEN		256
+
+#define OFFSET_SPIMCFG		0x0
+#define OFFSET_SPIMCTRL		0x4
+#define OFFSET_SPICMD		0x5
+#define OFFSET_SPIDCNT		0x6
+#define OFFSET_SPIADDR		0x8
+#define OFFSET_SPIINTSTS	0xc
+
+#define SPIMCTRL_START		0x01
+#define SPIMCTRL_BUSY		0x02
+#define SPIMCTRL_DIR		0x08
+
+struct gxp_spi;
+
+struct gxp_spi_chip {
+	struct gxp_spi *spifi;
+	u32 cs;
+};
+
+struct gxp_spi_data {
+	u32 max_cs;
+	u32 mode_bits;
+};
+
+struct gxp_spi {
+	const struct gxp_spi_data *data;
+	void __iomem *reg_base;
+	void __iomem *dat_base;
+	void __iomem *dir_base;
+	struct device *dev;
+	struct gxp_spi_chip chips[GXP_SPI0_MAX_CHIPSELECT];
+};
+
+static void gxp_spi_set_mode(struct gxp_spi *spifi, int mode)
+{
+	u8 value;
+	void __iomem *reg_base = spifi->reg_base;
+
+	value = readb(reg_base + OFFSET_SPIMCTRL);
+
+	if (mode == MANUAL_MODE) {
+		writeb(0x55, reg_base + OFFSET_SPICMD);
+		writeb(0xaa, reg_base + OFFSET_SPICMD);
+		value &= ~0x30;
+	} else {
+		value |= 0x30;
+	}
+	writeb(value, reg_base + OFFSET_SPIMCTRL);
+}
+
+static int gxp_spi_read_reg(struct gxp_spi_chip *chip, const struct spi_mem_op *op)
+{
+	int ret;
+	struct gxp_spi *spifi = chip->spifi;
+	void __iomem *reg_base = spifi->reg_base;
+	u32 value;
+
+	value = readl(reg_base + OFFSET_SPIMCFG);
+	value &= ~(1 << 24);
+	value |= (chip->cs << 24);
+	value &= ~(0x07 << 16);
+	value &= ~(0x1f << 19);
+	writel(value, reg_base + OFFSET_SPIMCFG);
+
+	writel(0, reg_base + OFFSET_SPIADDR);
+
+	writeb(op->cmd.opcode, reg_base + OFFSET_SPICMD);
+
+	writew(op->data.nbytes, reg_base + OFFSET_SPIDCNT);
+
+	value = readb(reg_base + OFFSET_SPIMCTRL);
+	value &= ~SPIMCTRL_DIR;
+	value |= SPIMCTRL_START;
+
+	writeb(value, reg_base + OFFSET_SPIMCTRL);
+
+	ret = readb_poll_timeout(reg_base + OFFSET_SPIMCTRL, value,
+				 !(value & SPIMCTRL_BUSY),
+				 GXP_SPI_SLEEP_TIME, GXP_SPI_TIMEOUT);
+	if (ret) {
+		dev_warn(spifi->dev, "read reg busy time out\n");
+		return ret;
+	}
+
+	memcpy_fromio(op->data.buf.in, spifi->dat_base, op->data.nbytes);
+	return ret;
+}
+
+static int gxp_spi_write_reg(struct gxp_spi_chip *chip, const struct spi_mem_op *op)
+{
+	int ret;
+	struct gxp_spi *spifi = chip->spifi;
+	void __iomem *reg_base = spifi->reg_base;
+	u32 value;
+
+	value = readl(reg_base + OFFSET_SPIMCFG);
+	value &= ~(1 << 24);
+	value |= (chip->cs << 24);
+	value &= ~(0x07 << 16);
+	value &= ~(0x1f << 19);
+	writel(value, reg_base + OFFSET_SPIMCFG);
+
+	writel(0, reg_base + OFFSET_SPIADDR);
+
+	writeb(op->cmd.opcode, reg_base + OFFSET_SPICMD);
+
+	memcpy_toio(spifi->dat_base, op->data.buf.in, op->data.nbytes);
+
+	writew(op->data.nbytes, reg_base + OFFSET_SPIDCNT);
+
+	value = readb(reg_base + OFFSET_SPIMCTRL);
+	value |= SPIMCTRL_DIR;
+	value |= SPIMCTRL_START;
+
+	writeb(value, reg_base + OFFSET_SPIMCTRL);
+
+	ret = readb_poll_timeout(reg_base + OFFSET_SPIMCTRL, value,
+				 !(value & SPIMCTRL_BUSY),
+				 GXP_SPI_SLEEP_TIME, GXP_SPI_TIMEOUT);
+	if (ret)
+		dev_warn(spifi->dev, "write reg busy time out\n");
+
+	return ret;
+}
+
+static ssize_t gxp_spi_read(struct gxp_spi_chip *chip, const struct spi_mem_op *op)
+{
+	struct gxp_spi *spifi = chip->spifi;
+	u32 offset = op->addr.val;
+
+	if (chip->cs == 0)
+		offset += 0x4000000;
+
+	memcpy_fromio(op->data.buf.in, spifi->dir_base + offset, op->data.nbytes);
+
+	return 0;
+}
+
+static ssize_t gxp_spi_write(struct gxp_spi_chip *chip, const struct spi_mem_op *op)
+{
+	struct gxp_spi *spifi = chip->spifi;
+	void __iomem *reg_base = spifi->reg_base;
+	u32 write_len;
+	u32 value;
+	int ret;
+
+	write_len = op->data.nbytes;
+	if (write_len > SPILDAT_LEN)
+		write_len = SPILDAT_LEN;
+
+	value = readl(reg_base + OFFSET_SPIMCFG);
+	value &= ~(1 << 24);
+	value |= (chip->cs << 24);
+	value &= ~(0x07 << 16);
+	value |= (op->addr.nbytes << 16);
+	value &= ~(0x1f << 19);
+	writel(value, reg_base + OFFSET_SPIMCFG);
+
+	writel(op->addr.val, reg_base + OFFSET_SPIADDR);
+
+	writeb(op->cmd.opcode, reg_base + OFFSET_SPICMD);
+
+	writew(write_len, reg_base + OFFSET_SPIDCNT);
+
+	memcpy_toio(spifi->dat_base, op->data.buf.in, write_len);
+
+	value = readb(reg_base + OFFSET_SPIMCTRL);
+	value |= SPIMCTRL_DIR;
+	value |= SPIMCTRL_START;
+
+	writeb(value, reg_base + OFFSET_SPIMCTRL);
+
+	ret = readb_poll_timeout(reg_base + OFFSET_SPIMCTRL, value,
+				 !(value & SPIMCTRL_BUSY),
+				 GXP_SPI_SLEEP_TIME, GXP_SPI_TIMEOUT);
+	if (ret) {
+		dev_warn(spifi->dev, "write busy time out\n");
+		return ret;
+	}
+
+	return write_len;
+}
+
+static int do_gxp_exec_mem_op(struct spi_mem *mem, const struct spi_mem_op *op)
+{
+	struct gxp_spi *spifi = spi_controller_get_devdata(mem->spi->master);
+	struct gxp_spi_chip *chip = &spifi->chips[mem->spi->chip_select];
+	int ret;
+
+	if (op->data.dir == SPI_MEM_DATA_IN) {
+		if (!op->addr.nbytes)
+			ret = gxp_spi_read_reg(chip, op);
+		else
+			ret = gxp_spi_read(chip, op);
+	} else {
+		if (!op->addr.nbytes)
+			ret = gxp_spi_write_reg(chip, op);
+		else
+			ret = gxp_spi_write(chip, op);
+	}
+
+	return ret;
+}
+
+static int gxp_exec_mem_op(struct spi_mem *mem, const struct spi_mem_op *op)
+{
+	int ret;
+
+	ret = do_gxp_exec_mem_op(mem, op);
+	if (ret)
+		dev_err(&mem->spi->dev, "operation failed: %d", ret);
+
+	return ret;
+}
+
+static const struct spi_controller_mem_ops gxp_spi_mem_ops = {
+	.exec_op = gxp_exec_mem_op,
+};
+
+static int gxp_spi_setup(struct spi_device *spi)
+{
+	struct gxp_spi *spifi = spi_controller_get_devdata(spi->master);
+	unsigned int cs = spi->chip_select;
+	struct gxp_spi_chip *chip = &spifi->chips[cs];
+
+	chip->spifi = spifi;
+	chip->cs = cs;
+
+	gxp_spi_set_mode(spifi, MANUAL_MODE);
+
+	return 0;
+}
+
+static int gxp_spifi_probe(struct platform_device *pdev)
+{
+	struct device *dev = &pdev->dev;
+	const struct gxp_spi_data *data;
+	struct spi_controller *ctlr;
+	struct gxp_spi *spifi;
+	struct resource *res;
+	int ret;
+
+	data = of_device_get_match_data(&pdev->dev);
+
+	ctlr = devm_spi_alloc_master(dev, sizeof(*spifi));
+	if (!ctlr)
+		return -ENOMEM;
+
+	spifi = spi_controller_get_devdata(ctlr);
+
+	platform_set_drvdata(pdev, spifi);
+	spifi->data = data;
+	spifi->dev = dev;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	spifi->reg_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(spifi->reg_base))
+		return PTR_ERR(spifi->reg_base);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	spifi->dat_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(spifi->dat_base))
+		return PTR_ERR(spifi->dat_base);
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 2);
+	spifi->dir_base = devm_ioremap_resource(&pdev->dev, res);
+	if (IS_ERR(spifi->dir_base))
+		return PTR_ERR(spifi->dir_base);
+
+	ctlr->mode_bits = data->mode_bits;
+	ctlr->bus_num = pdev->id;
+	ctlr->mem_ops = &gxp_spi_mem_ops;
+	ctlr->setup = gxp_spi_setup;
+	ctlr->num_chipselect = data->max_cs;
+	ctlr->dev.of_node = dev->of_node;
+
+	ret = devm_spi_register_controller(dev, ctlr);
+	if (ret) {
+		return dev_err_probe(&pdev->dev, ret,
+				     "failed to register spi controller\n");
+	}
+
+	return 0;
+}
+
+static const struct gxp_spi_data gxp_spifi_data = {
+	.max_cs	= 2,
+	.mode_bits = 0,
+};
+
+static const struct of_device_id gxp_spifi_match[] = {
+	{.compatible = "hpe,gxp-spifi", .data = &gxp_spifi_data },
+	{ /* null */ }
+};
+MODULE_DEVICE_TABLE(of, gxp_spifi_match);
+
+static struct platform_driver gxp_spifi_driver = {
+	.probe = gxp_spifi_probe,
+	.driver = {
+		.name = "gxp-spifi",
+		.of_match_table = gxp_spifi_match,
+	},
+};
+module_platform_driver(gxp_spifi_driver);
+
+MODULE_DESCRIPTION("HPE GXP SPI Flash Interface driver");
+MODULE_AUTHOR("Nick Hawkins <[e-mail]>");
+MODULE_LICENSE("GPL");

drivers/spi/spi-intel-pci.c

···
 	{ PCI_VDEVICE(INTEL, 0x54a4), (unsigned long)&cnl_info },
 	{ PCI_VDEVICE(INTEL, 0x7a24), (unsigned long)&cnl_info },
 	{ PCI_VDEVICE(INTEL, 0x7aa4), (unsigned long)&cnl_info },
+	{ PCI_VDEVICE(INTEL, 0x7e23), (unsigned long)&cnl_info },
 	{ PCI_VDEVICE(INTEL, 0xa0a4), (unsigned long)&bxt_info },
 	{ PCI_VDEVICE(INTEL, 0xa1a4), (unsigned long)&bxt_info },
 	{ PCI_VDEVICE(INTEL, 0xa224), (unsigned long)&bxt_info },

drivers/spi/spi-intel.c

···
 		return -ENOMEM;
 
 	pdata->nr_parts = 1;
-	pdata->parts = devm_kcalloc(ispi->dev, sizeof(*pdata->parts),
-				    pdata->nr_parts, GFP_KERNEL);
+	pdata->parts = devm_kcalloc(ispi->dev, pdata->nr_parts,
+				    sizeof(*pdata->parts), GFP_KERNEL);
 	if (!pdata->parts)
 		return -ENOMEM;
 

drivers/spi/spi-microchip-core.c

···
+// SPDX-License-Identifier: (GPL-2.0)
+/*
+ * Microchip CoreSPI SPI controller driver
+ *
+ * Copyright (c) 2018-2022 Microchip Technology Inc. and its subsidiaries
+ *
+ * Author: Daire McNamara <[e-mail]>
+ * Author: Conor Dooley <[e-mail]>
+ *
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/io.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/spi/spi.h>
+
+#define MAX_LEN				(0xffff)
+#define MAX_CS				(8)
+#define DEFAULT_FRAMESIZE		(8)
+#define FIFO_DEPTH			(32)
+#define CLK_GEN_MODE1_MAX		(255)
+#define CLK_GEN_MODE0_MAX		(15)
+#define CLK_GEN_MIN			(0)
+#define MODE_X_MASK_SHIFT		(24)
+
+#define CONTROL_ENABLE			BIT(0)
+#define CONTROL_MASTER			BIT(1)
+#define CONTROL_RX_DATA_INT		BIT(4)
+#define CONTROL_TX_DATA_INT		BIT(5)
+#define CONTROL_RX_OVER_INT		BIT(6)
+#define CONTROL_TX_UNDER_INT		BIT(7)
+#define CONTROL_SPO			BIT(24)
+#define CONTROL_SPH			BIT(25)
+#define CONTROL_SPS			BIT(26)
+#define CONTROL_FRAMEURUN		BIT(27)
+#define CONTROL_CLKMODE			BIT(28)
+#define CONTROL_BIGFIFO			BIT(29)
+#define CONTROL_OENOFF			BIT(30)
+#define CONTROL_RESET			BIT(31)
+
+#define CONTROL_MODE_MASK		GENMASK(3, 2)
+#define  MOTOROLA_MODE			(0)
+#define CONTROL_FRAMECNT_MASK		GENMASK(23, 8)
+#define CONTROL_FRAMECNT_SHIFT		(8)
+
+#define STATUS_ACTIVE			BIT(14)
+#define STATUS_SSEL			BIT(13)
+#define STATUS_FRAMESTART		BIT(12)
+#define STATUS_TXFIFO_EMPTY_NEXT_READ	BIT(11)
+#define STATUS_TXFIFO_EMPTY		BIT(10)
+#define STATUS_TXFIFO_FULL_NEXT_WRITE	BIT(9)
+#define STATUS_TXFIFO_FULL		BIT(8)
+#define STATUS_RXFIFO_EMPTY_NEXT_READ	BIT(7)
+#define STATUS_RXFIFO_EMPTY		BIT(6)
+#define STATUS_RXFIFO_FULL_NEXT_WRITE	BIT(5)
+#define STATUS_RXFIFO_FULL		BIT(4)
+#define STATUS_TX_UNDERRUN		BIT(3)
+#define STATUS_RX_OVERFLOW		BIT(2)
+#define STATUS_RXDAT_RXED		BIT(1)
+#define STATUS_TXDAT_SENT		BIT(0)
+
+#define INT_TXDONE			BIT(0)
+#define INT_RXRDY			BIT(1)
+#define INT_RX_CHANNEL_OVERFLOW		BIT(2)
+#define INT_TX_CHANNEL_UNDERRUN		BIT(3)
+
+#define INT_ENABLE_MASK (CONTROL_RX_DATA_INT | CONTROL_TX_DATA_INT | \
+			 CONTROL_RX_OVER_INT | CONTROL_TX_UNDER_INT)
+
+#define REG_CONTROL		(0x00)
+#define REG_FRAME_SIZE		(0x04)
+#define REG_STATUS		(0x08)
+#define REG_INT_CLEAR		(0x0c)
+#define REG_RX_DATA		(0x10)
+#define REG_TX_DATA		(0x14)
+#define REG_CLK_GEN		(0x18)
+#define REG_SLAVE_SELECT	(0x1c)
+#define  SSEL_MASK		GENMASK(7, 0)
+#define  SSEL_DIRECT		BIT(8)
+#define  SSELOUT_SHIFT		9
+#define  SSELOUT		BIT(SSELOUT_SHIFT)
+#define REG_MIS			(0x20)
+#define REG_RIS			(0x24)
+#define REG_CONTROL2		(0x28)
+#define REG_COMMAND		(0x2c)
+#define REG_PKTSIZE		(0x30)
+#define REG_CMD_SIZE		(0x34)
+#define REG_HWSTATUS		(0x38)
+#define REG_STAT8		(0x3c)
+#define REG_CTRL2		(0x48)
+#define REG_FRAMESUP		(0x50)
+
+struct mchp_corespi {
+	void __iomem *regs;
+	struct clk *clk;
+	const u8 *tx_buf;
+	u8 *rx_buf;
+	u32 clk_gen; /* divider for spi output clock generated by the controller */
+	u32 clk_mode;
+	int irq;
+	int tx_len;
+	int rx_len;
+	int pending;
+};
+
+static inline u32 mchp_corespi_read(struct mchp_corespi *spi, unsigned int reg)
+{
+	return readl(spi->regs + reg);
+}
+
+static inline void mchp_corespi_write(struct mchp_corespi *spi, unsigned int reg, u32 val)
+{
+	writel(val, spi->regs + reg);
+}
+
+static inline void mchp_corespi_enable(struct mchp_corespi *spi)
+{
+	u32 control = mchp_corespi_read(spi, REG_CONTROL);
+
+	control |= CONTROL_ENABLE;
+
+	mchp_corespi_write(spi, REG_CONTROL, control);
+}
+
+static inline void mchp_corespi_disable(struct mchp_corespi *spi)
+{
+	u32 control = mchp_corespi_read(spi, REG_CONTROL);
+
+	control &= ~CONTROL_ENABLE;
+
+	mchp_corespi_write(spi, REG_CONTROL, control);
+}
+
+static inline void mchp_corespi_read_fifo(struct mchp_corespi *spi)
+{
+	u8 data;
+	int fifo_max, i = 0;
+
+	fifo_max = min(spi->rx_len, FIFO_DEPTH);
+
+	while ((i < fifo_max) && !(mchp_corespi_read(spi, REG_STATUS) & STATUS_RXFIFO_EMPTY)) {
+		data = mchp_corespi_read(spi, REG_RX_DATA);
+
+		if (spi->rx_buf)
+			*spi->rx_buf++ = data;
+		i++;
+	}
+	spi->rx_len -= i;
+	spi->pending -= i;
+}
+
+static void mchp_corespi_enable_ints(struct mchp_corespi *spi)
+{
+	u32 control, mask = INT_ENABLE_MASK;
+
+	mchp_corespi_disable(spi);
+
+	control = mchp_corespi_read(spi, REG_CONTROL);
+
+	control |= mask;
+	mchp_corespi_write(spi, REG_CONTROL, control);
+
+	control |= CONTROL_ENABLE;
+	mchp_corespi_write(spi, REG_CONTROL, control);
+}
+
+static void mchp_corespi_disable_ints(struct mchp_corespi *spi)
+{
+	u32 control, mask = INT_ENABLE_MASK;
+
+	mchp_corespi_disable(spi);
+
+	control = mchp_corespi_read(spi, REG_CONTROL);
+	control &= ~mask;
+	mchp_corespi_write(spi, REG_CONTROL, control);
+
+	control |= CONTROL_ENABLE;
+	mchp_corespi_write(spi, REG_CONTROL, control);
+}
+
+static inline void mchp_corespi_set_xfer_size(struct mchp_corespi *spi, int len)
+{
+	u32 control;
+	u16 lenpart;
+
+	/*
+	 * Disable the SPI controller. Writes to transfer length have
+	 * no effect when the controller is enabled.
+	 */
+	mchp_corespi_disable(spi);
+
+	/*
+	 * The lower 16 bits of the frame count are stored in the control reg
+	 * for legacy reasons, but the upper 16 written to a different register:
+	 * FRAMESUP. While both the upper and lower bits can be *READ* from the
+	 * FRAMESUP register, writing to the lower 16 bits is a NOP
+	 */
+	lenpart = len & 0xffff;
+
+	control = mchp_corespi_read(spi, REG_CONTROL);
+	control &= ~CONTROL_FRAMECNT_MASK;
+	control |= lenpart << CONTROL_FRAMECNT_SHIFT;
+	mchp_corespi_write(spi, REG_CONTROL, control);
+
+	lenpart = len & 0xffff0000;
+	mchp_corespi_write(spi, REG_FRAMESUP, lenpart);
+
+	control |= CONTROL_ENABLE;
+	mchp_corespi_write(spi, REG_CONTROL, control);
+}
+
+static inline void mchp_corespi_write_fifo(struct mchp_corespi *spi)
+{
+	u8 byte;
+	int fifo_max, i = 0;
+
+	fifo_max = min(spi->tx_len, FIFO_DEPTH);
+	mchp_corespi_set_xfer_size(spi, fifo_max);
+
+	while ((i < fifo_max) && !(mchp_corespi_read(spi, REG_STATUS) & STATUS_TXFIFO_FULL)) {
+		byte = spi->tx_buf ? *spi->tx_buf++ : 0xaa;
+		mchp_corespi_write(spi, REG_TX_DATA, byte);
+		i++;
+	}
+
+	spi->tx_len -= i;
+	spi->pending += i;
+}
+
+static inline void mchp_corespi_set_framesize(struct mchp_corespi *spi, int bt)
+{
+	u32 control;
+
+	/*
+	 * Disable the SPI controller. Writes to the frame size have
+	 * no effect when the controller is enabled.
+	 */
+	mchp_corespi_disable(spi);
+
+	mchp_corespi_write(spi, REG_FRAME_SIZE, bt);
+
+	control = mchp_corespi_read(spi, REG_CONTROL);
+	control |= CONTROL_ENABLE;
+	mchp_corespi_write(spi, REG_CONTROL, control);
+}
+
+static void mchp_corespi_set_cs(struct spi_device *spi, bool disable)
+{
+	u32 reg;
+	struct mchp_corespi *corespi = spi_master_get_devdata(spi->master);
+
+	reg = mchp_corespi_read(corespi, REG_SLAVE_SELECT);
+	reg &= ~BIT(spi->chip_select);
+	reg |= !disable << spi->chip_select;
+
+	mchp_corespi_write(corespi, REG_SLAVE_SELECT, reg);
+}
+
+static int mchp_corespi_setup(struct spi_device *spi)
+{
+	struct mchp_corespi *corespi = spi_master_get_devdata(spi->master);
+	u32 reg;
+
+	/*
+	 * Active high slaves need to be specifically set to their inactive
+	 * states during probe by adding them to the "control group" & thus
+	 * driving their select line low.
+	 */
+	if (spi->mode & SPI_CS_HIGH) {
+		reg = mchp_corespi_read(corespi, REG_SLAVE_SELECT);
+		reg |= BIT(spi->chip_select);
+		mchp_corespi_write(corespi, REG_SLAVE_SELECT, reg);
+	}
+	return 0;
+}
+
+static void mchp_corespi_init(struct spi_master *master, struct mchp_corespi *spi)
+{
+	unsigned long clk_hz;
+	u32 control = mchp_corespi_read(spi, REG_CONTROL);
+
+	control |= CONTROL_MASTER;
+
+	control &= ~CONTROL_MODE_MASK;
+	control |= MOTOROLA_MODE;
+
+	mchp_corespi_set_framesize(spi, DEFAULT_FRAMESIZE);
+
+	/* max. possible spi clock rate is the apb clock rate */
+	clk_hz = clk_get_rate(spi->clk);
+	master->max_speed_hz = clk_hz;
+
+	/*
+	 * The controller must be configured so that it doesn't remove Chip
+	 * Select until the entire message has been transferred, even if at
+	 * some points TX FIFO becomes empty.
+	 *
+	 * BIGFIFO mode is also enabled, which sets the fifo depth to 32 frames
+	 * for the 8 bit transfers that this driver uses.
+	 */
+	control = mchp_corespi_read(spi, REG_CONTROL);
+	control |= CONTROL_SPS | CONTROL_BIGFIFO;
+
+	mchp_corespi_write(spi, REG_CONTROL, control);
+
+	mchp_corespi_enable_ints(spi);
+
+	/*
+	 * It is required to enable direct mode, otherwise control over the chip
+	 * select is relinquished to the hardware. SSELOUT is enabled too so we
+	 * can deal with active high slaves.
+	 */
+	mchp_corespi_write(spi, REG_SLAVE_SELECT, SSELOUT | SSEL_DIRECT);
+
+	control = mchp_corespi_read(spi, REG_CONTROL);
+
+	control &= ~CONTROL_RESET;
+	control |= CONTROL_ENABLE;
+
+	mchp_corespi_write(spi, REG_CONTROL, control);
+}
+
+static inline void mchp_corespi_set_clk_gen(struct mchp_corespi *spi)
+{
+	u32 control;
+
+	mchp_corespi_disable(spi);
+
+	control = mchp_corespi_read(spi, REG_CONTROL);
+	if (spi->clk_mode)
+		control |= CONTROL_CLKMODE;
+	else
+		control &= ~CONTROL_CLKMODE;
+
+	mchp_corespi_write(spi, REG_CLK_GEN, spi->clk_gen);
+	mchp_corespi_write(spi, REG_CONTROL, control);
+	mchp_corespi_write(spi, REG_CONTROL, control | CONTROL_ENABLE);
+}
+
+static inline void mchp_corespi_set_mode(struct mchp_corespi *spi, unsigned int mode)
+{
+	u32 control, mode_val;
+
+	switch (mode & SPI_MODE_X_MASK) {
+	case SPI_MODE_0:
+		mode_val = 0;
+		break;
+	case SPI_MODE_1:
+		mode_val = CONTROL_SPH;
+		break;
+	case SPI_MODE_2:
+		mode_val = CONTROL_SPO;
+		break;
+	case SPI_MODE_3:
+		mode_val = CONTROL_SPH | CONTROL_SPO;
+		break;
+	}
+
+	/*
+	 * Disable the SPI controller. Writes to the frame size have
+	 * no effect when the controller is enabled.
+	 */
+	mchp_corespi_disable(spi);
+
+	control = mchp_corespi_read(spi, REG_CONTROL);
+	control &= ~(SPI_MODE_X_MASK << MODE_X_MASK_SHIFT);
+	control |= mode_val;
+
+	mchp_corespi_write(spi, REG_CONTROL, control);
+
+	control |= CONTROL_ENABLE;
+	mchp_corespi_write(spi, REG_CONTROL, control);
+}
+
+static irqreturn_t mchp_corespi_interrupt(int irq, void *dev_id)
+{
+	struct spi_master *master = dev_id;
+	struct mchp_corespi *spi = spi_master_get_devdata(master);
+	u32 intfield = mchp_corespi_read(spi, REG_MIS) & 0xf;
+	bool finalise = false;
+
+	/* Interrupt line may be shared and not for us at all */
+	if (intfield == 0)
+		return IRQ_NONE;
+
+	if (intfield & INT_TXDONE) {
+		mchp_corespi_write(spi, REG_INT_CLEAR, INT_TXDONE);
+
+		if (spi->rx_len)
+			mchp_corespi_read_fifo(spi);
+
+		if (spi->tx_len)
+			mchp_corespi_write_fifo(spi);
+
+		if (!spi->rx_len)
+			finalise = true;
+	}
+
+	if (intfield & INT_RXRDY)
+		mchp_corespi_write(spi, REG_INT_CLEAR, INT_RXRDY);
+
+	if (intfield & INT_RX_CHANNEL_OVERFLOW) {
+		mchp_corespi_write(spi, REG_INT_CLEAR, INT_RX_CHANNEL_OVERFLOW);
+		finalise = true;
+		dev_err(&master->dev,
+			"%s: RX OVERFLOW: rxlen: %d, txlen: %d\n", __func__,
+			spi->rx_len, spi->tx_len);
+	}
+
+	if (intfield & INT_TX_CHANNEL_UNDERRUN) {
+		mchp_corespi_write(spi, REG_INT_CLEAR, INT_TX_CHANNEL_UNDERRUN);
+		finalise = true;
+		dev_err(&master->dev,
+			"%s: TX UNDERFLOW: rxlen: %d, txlen: %d\n", __func__,
+			spi->rx_len, spi->tx_len);
+	}
+
+	if (finalise)
+		spi_finalize_current_transfer(master);
+
+	return IRQ_HANDLED;
+}
+
+static int mchp_corespi_calculate_clkgen(struct mchp_corespi *spi,
+					 unsigned long target_hz)
+{
+	unsigned long clk_hz, spi_hz, clk_gen;
+
+	clk_hz = clk_get_rate(spi->clk);
+	if (!clk_hz)
+		return -EINVAL;
+	spi_hz = min(target_hz, clk_hz);
+
+	/*
+	 * There are two possible clock modes for the controller generated
+	 * clock's division ratio:
+	 * CLK_MODE = 0: 1 / (2^(CLK_GEN + 1)) where CLK_GEN = 0 to 15.
+	 * CLK_MODE = 1: 1 / (2 * CLK_GEN + 1) where CLK_GEN = 0 to 255.
+	 * First try mode 1, fall back to 0 and if we have tried both modes and
+	 * we /still/ can't get a good setting, we then throw the toys out of
+	 * the pram and give up
+	 * clk_gen is the register name for the clock divider on MPFS.
+	 */
+	clk_gen = DIV_ROUND_UP(clk_hz, 2 * spi_hz) - 1;
+	if (clk_gen > CLK_GEN_MODE1_MAX || clk_gen <= CLK_GEN_MIN) {
+		clk_gen = DIV_ROUND_UP(clk_hz, spi_hz);
+		clk_gen = fls(clk_gen) - 1;
+
+		if (clk_gen > CLK_GEN_MODE0_MAX)
+			return -EINVAL;
+
+		spi->clk_mode = 0;
+	} else {
+		spi->clk_mode = 1;
+	}
+
+	spi->clk_gen = clk_gen;
+	return 0;
+}
+
+static int mchp_corespi_transfer_one(struct spi_master *master,
+				     struct spi_device *spi_dev,
+				     struct spi_transfer *xfer)
+{
+	struct mchp_corespi *spi = spi_master_get_devdata(master);
+	int ret;
+
+	ret = mchp_corespi_calculate_clkgen(spi, (unsigned long)xfer->speed_hz);
+	if (ret) {
+		dev_err(&master->dev, "failed to set clk_gen for target %u Hz\n", xfer->speed_hz);
+		return ret;
+	}
+
+	mchp_corespi_set_clk_gen(spi);
+
+	spi->tx_buf = xfer->tx_buf;
+	spi->rx_buf = xfer->rx_buf;
+	spi->tx_len = xfer->len;
+	spi->rx_len = xfer->len;
+	spi->pending = 0;
+
+	mchp_corespi_set_xfer_size(spi, (spi->tx_len > FIFO_DEPTH)
+				   ? FIFO_DEPTH : spi->tx_len);
+
+	if (spi->tx_len)
+		mchp_corespi_write_fifo(spi);
+	return 1;
+}
+
+static int mchp_corespi_prepare_message(struct spi_master *master,
+					struct spi_message *msg)
+{
+	struct spi_device *spi_dev = msg->spi;
+	struct mchp_corespi *spi = spi_master_get_devdata(master);
+
+	mchp_corespi_set_framesize(spi, DEFAULT_FRAMESIZE);
+	mchp_corespi_set_mode(spi, spi_dev->mode);
+
+	return 0;
+}
+
+static int mchp_corespi_probe(struct platform_device *pdev)
+{
+	struct spi_master *master;
+	struct mchp_corespi *spi;
+	struct resource *res;
+	u32 num_cs;
+	int ret = 0;
+
+	master = devm_spi_alloc_master(&pdev->dev, sizeof(*spi));
+	if (!master)
+		return dev_err_probe(&pdev->dev, -ENOMEM,
+				     "unable to allocate master for SPI controller\n");
+
+	platform_set_drvdata(pdev, master);
+
+	if (of_property_read_u32(pdev->dev.of_node, "num-cs", &num_cs))
+		num_cs = MAX_CS;
+
+	master->num_chipselect = num_cs;
+	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_CS_HIGH;
+	master->setup = mchp_corespi_setup;
+	master->bits_per_word_mask = SPI_BPW_MASK(8);
+	master->transfer_one = mchp_corespi_transfer_one;
+	master->prepare_message = mchp_corespi_prepare_message;
+	master->set_cs = mchp_corespi_set_cs;
+	master->dev.of_node = pdev->dev.of_node;
+
+	spi = spi_master_get_devdata(master);
+
+	spi->regs = devm_platform_get_and_ioremap_resource(pdev, 0, &res);
+	if (IS_ERR(spi->regs))
+		return PTR_ERR(spi->regs);
+
+	spi->irq = platform_get_irq(pdev, 0);
+	if (spi->irq <= 0)
+		return dev_err_probe(&pdev->dev, -ENXIO,
+				     "invalid IRQ %d for SPI controller\n",
+				     spi->irq);
+
+	ret = devm_request_irq(&pdev->dev, spi->irq, mchp_corespi_interrupt,
+			       IRQF_SHARED, dev_name(&pdev->dev), master);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret,
+				     "could not request irq: %d\n", ret);
+
+	spi->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(spi->clk))
+		return dev_err_probe(&pdev->dev, PTR_ERR(spi->clk),
+				     "could not get clk: %d\n", ret);
+
+	ret = clk_prepare_enable(spi->clk);
+	if (ret)
+		return dev_err_probe(&pdev->dev, ret,
+				     "failed to enable clock\n");
+
+	mchp_corespi_init(master, spi);
+
+	ret = devm_spi_register_master(&pdev->dev, master);
+	if (ret) {
+		mchp_corespi_disable(spi);
+		clk_disable_unprepare(spi->clk);
+		return dev_err_probe(&pdev->dev, ret,
+				     "unable to register master for SPI controller\n");
+	}
+
+	dev_info(&pdev->dev, "Registered SPI controller %d\n", master->bus_num);
+
+	return 0;
+}
+
+static int mchp_corespi_remove(struct platform_device *pdev)
+{
+	struct spi_master *master  = platform_get_drvdata(pdev);
+	struct mchp_corespi *spi = spi_master_get_devdata(master);
+
+	mchp_corespi_disable_ints(spi);
+	clk_disable_unprepare(spi->clk);
+	mchp_corespi_disable(spi);
+
+	return 0;
+}
+
+#define MICROCHIP_SPI_PM_OPS (NULL)
+
+/*
+ * Platform driver data structure
+ */
+
+#if defined(CONFIG_OF)
+static const struct of_device_id mchp_corespi_dt_ids[] = {
+	{ .compatible = "microchip,mpfs-spi" },
+	{ /* sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, mchp_corespi_dt_ids);
+#endif
+
+static struct platform_driver mchp_corespi_driver = {
+	.probe = mchp_corespi_probe,
+	.driver = {
+		.name = "microchip-corespi",
+		.pm = MICROCHIP_SPI_PM_OPS,
+		.of_match_table = of_match_ptr(mchp_corespi_dt_ids),
+	},
+	.remove = mchp_corespi_remove,
+};
+module_platform_driver(mchp_corespi_driver);
+MODULE_DESCRIPTION("Microchip coreSPI SPI controller driver");
+MODULE_AUTHOR("Daire McNamara <[e-mail]>");
+MODULE_AUTHOR("Conor Dooley <[e-mail]>");
+MODULE_LICENSE("GPL");

drivers/spi/spi-mpc52xx-psc.c

···
 	struct mpc52xx_psc_fifo __iomem *fifo;
 	unsigned int irq;
 	u8 bits_per_word;
-	u8 busy;
-
-	struct work_struct work;
-
-	struct list_head queue;
-	spinlock_t lock;
 
 	struct completion done;
 };
···
 	return 0;
 }
 
-static void mpc52xx_psc_spi_work(struct work_struct *work)
+int mpc52xx_psc_spi_transfer_one_message(struct spi_controller *ctlr,
+					 struct spi_message *m)
 {
-	struct mpc52xx_psc_spi *mps =
-		container_of(work, struct mpc52xx_psc_spi, work);
+	struct spi_device *spi;
+	struct spi_transfer *t = NULL;
+	unsigned cs_change;
+	int status;
 
-	spin_lock_irq(&mps->lock);
-	mps->busy = 1;
-	while (!list_empty(&mps->queue)) {
-		struct spi_message *m;
-		struct spi_device *spi;
-		struct spi_transfer *t = NULL;
-		unsigned cs_change;
-		int status;
-
-		m = container_of(mps->queue.next, struct spi_message, queue);
-		list_del_init(&m->queue);
-		spin_unlock_irq(&mps->lock);
-
-		spi = m->spi;
-		cs_change = 1;
-		status = 0;
-		list_for_each_entry (t, &m->transfers, transfer_list) {
-			if (t->bits_per_word || t->speed_hz) {
-				status = mpc52xx_psc_spi_transfer_setup(spi, t);
-				if (status < 0)
-					break;
-			}
-
-			if (cs_change)
-				mpc52xx_psc_spi_activate_cs(spi);
-			cs_change = t->cs_change;
-
-			status = mpc52xx_psc_spi_transfer_rxtx(spi, t);
-			if (status)
+	spi = m->spi;
+	cs_change = 1;
+	status = 0;
+	list_for_each_entry (t, &m->transfers, transfer_list) {
+		if (t->bits_per_word || t->speed_hz) {
+			status = mpc52xx_psc_spi_transfer_setup(spi, t);
+			if (status < 0)
 				break;
-			m->actual_length += t->len;
-
-			spi_transfer_delay_exec(t);
-
-			if (cs_change)
-				mpc52xx_psc_spi_deactivate_cs(spi);
 		}
 
-		m->status = status;
-		if (m->complete)
-			m->complete(m->context);
+		if (cs_change)
+			mpc52xx_psc_spi_activate_cs(spi);
+		cs_change = t->cs_change;
 
-		if (status || !cs_change)
+		status = mpc52xx_psc_spi_transfer_rxtx(spi, t);
+		if (status)
+			break;
+		m->actual_length += t->len;
+
+		spi_transfer_delay_exec(t);
+
+		if (cs_change)
 			mpc52xx_psc_spi_deactivate_cs(spi);
-
-		mpc52xx_psc_spi_transfer_setup(spi, NULL);
-
-		spin_lock_irq(&mps->lock);
 	}
-	mps->busy = 0;
-	spin_unlock_irq(&mps->lock);
+
+	m->status = status;
+	if (status || !cs_change)
+		mpc52xx_psc_spi_deactivate_cs(spi);
+
+	mpc52xx_psc_spi_transfer_setup(spi, NULL);
+
+	spi_finalize_current_message(ctlr);
+
+	return 0;
 }
 
 static int mpc52xx_psc_spi_setup(struct spi_device *spi)
 {
-	struct mpc52xx_psc_spi *mps = spi_master_get_devdata(spi->master);
 	struct mpc52xx_psc_spi_cs *cs = spi->controller_state;
-	unsigned long flags;
 
 	if (spi->bits_per_word%8)
 		return -EINVAL;
···
 
 	cs->bits_per_word = spi->bits_per_word;
 	cs->speed_hz = spi->max_speed_hz;
-
-	spin_lock_irqsave(&mps->lock, flags);
-	if (!mps->busy)
-		mpc52xx_psc_spi_deactivate_cs(spi);
-	spin_unlock_irqrestore(&mps->lock, flags);
-
-	return 0;
-}
-
-static int mpc52xx_psc_spi_transfer(struct spi_device *spi,
-		struct spi_message *m)
-{
-	struct mpc52xx_psc_spi *mps = spi_master_get_devdata(spi->master);
-	unsigned long flags;
-
-	m->actual_length = 0;
-	m->status = -EINPROGRESS;
-
-	spin_lock_irqsave(&mps->lock, flags);
-	list_add_tail(&m->queue, &mps->queue);
-	schedule_work(&mps->work);
-	spin_unlock_irqrestore(&mps->lock, flags);
 
 	return 0;
 }
···
 		master->num_chipselect = pdata->max_chipselect;
 	}
 	master->setup = mpc52xx_psc_spi_setup;
-	master->transfer = mpc52xx_psc_spi_transfer;
+	master->transfer_one_message = mpc52xx_psc_spi_transfer_one_message;
 	master->cleanup = mpc52xx_psc_spi_cleanup;
 	master->dev.of_node = dev->of_node;
 
···
 		goto free_irq;
 	}
 
-	spin_lock_init(&mps->lock);
 	init_completion(&mps->done);
-	INIT_WORK(&mps->work, mpc52xx_psc_spi_work);
-	INIT_LIST_HEAD(&mps->queue);
 
 	ret = spi_register_master(master);
 	if (ret < 0)
···
 	struct spi_master *master = spi_master_get(platform_get_drvdata(op));
 	struct mpc52xx_psc_spi *mps = spi_master_get_devdata(master);
 
-	flush_work(&mps->work);
 	spi_unregister_master(master);
 	free_irq(mps->irq, mps);
 	if (mps->psc)

drivers/spi/spi-npcm-fiu.c

···
 #define NPCM_FIU_UMA_DR1		0x34
 #define NPCM_FIU_UMA_DR2		0x38
 #define NPCM_FIU_UMA_DR3		0x3C
+#define NPCM_FIU_CFG			0x78
 #define NPCM_FIU_MAX_REG_LIMIT		0x80
 
 /* FIU Direct Read Configuration Register */
···
 #define NPCM_FIU_UMA_DR3_RB13		GENMASK(15, 8)
 #define NPCM_FIU_UMA_DR3_RB12		GENMASK(7, 0)
 
+/* FIU Configuration Register */
+#define NPCM_FIU_CFG_FIU_FIX		BIT(31)
+
 /* FIU Read Mode */
 enum {
 	DRD_SINGLE_WIRE_MODE	= 0,
···
 	FIU0 = 0,
 	FIU3,
 	FIUX,
+	FIU1,
 };
 
 struct npcm_fiu_info {
···
 static const struct fiu_data npcm7xx_fiu_data = {
 	.npcm_fiu_data_info = npcm7xx_fiu_info,
 	.fiu_max = 3,
+};
+
+static const struct npcm_fiu_info npxm8xx_fiu_info[] = {
+	{.name = "FIU0", .fiu_id = FIU0,
+		.max_map_size = MAP_SIZE_128MB, .max_cs = 2},
+	{.name = "FIU3", .fiu_id = FIU3,
+		.max_map_size = MAP_SIZE_128MB, .max_cs = 4},
+	{.name = "FIUX", .fiu_id = FIUX,
+		.max_map_size = MAP_SIZE_16MB, .max_cs = 2},
+	{.name = "FIU1", .fiu_id = FIU1,
+		.max_map_size = MAP_SIZE_16MB, .max_cs = 4} };
+
+static const struct fiu_data npxm8xx_fiu_data = {
+	.npcm_fiu_data_info = npxm8xx_fiu_info,
+	.fiu_max = 4,
 };
 
 struct npcm_fiu_spi;
···
 	fiu->drd_op.addr.buswidth = op->addr.buswidth;
 	regmap_update_bits(fiu->regmap, NPCM_FIU_DRD_CFG,
 			   NPCM_FIU_DRD_CFG_DBW,
-			   ((op->dummy.nbytes * ilog2(op->addr.buswidth)) / BITS_PER_BYTE)
-			   << NPCM_FIU_DRD_DBW_SHIFT);
+			   op->dummy.nbytes << NPCM_FIU_DRD_DBW_SHIFT);
 	fiu->drd_op.dummy.nbytes = op->dummy.nbytes;
 	regmap_update_bits(fiu->regmap, NPCM_FIU_DRD_CFG,
 			   NPCM_FIU_DRD_CFG_RDCMD, op->cmd.opcode);
···
 		regmap_update_bits(gcr_regmap, NPCM7XX_INTCR3_OFFSET,
 				   NPCM7XX_INTCR3_FIU_FIX,
 				   NPCM7XX_INTCR3_FIU_FIX);
+	} else {
+		regmap_update_bits(fiu->regmap, NPCM_FIU_CFG,
+				   NPCM_FIU_CFG_FIU_FIX,
+				   NPCM_FIU_CFG_FIU_FIX);
 	}
 
 	if (desc->info.op_tmpl.data.dir == SPI_MEM_DATA_IN) {
···
 
 static const struct of_device_id npcm_fiu_dt_ids[] = {
 	{ .compatible = "nuvoton,npcm750-fiu", .data = &npcm7xx_fiu_data  },
+	{ .compatible = "nuvoton,npcm845-fiu", .data = &npxm8xx_fiu_data  },
 	{ /* sentinel */ }
 };
 

drivers/spi/spi-pxa2xx.c

···
 	{ PCI_VDEVICE(INTEL, 0x7aab), LPSS_CNL_SSP },
 	{ PCI_VDEVICE(INTEL, 0x7af9), LPSS_CNL_SSP },
 	{ PCI_VDEVICE(INTEL, 0x7afb), LPSS_CNL_SSP },
+	/* MTL-P */
+	{ PCI_VDEVICE(INTEL, 0x7e27), LPSS_CNL_SSP },
+	{ PCI_VDEVICE(INTEL, 0x7e30), LPSS_CNL_SSP },
+	{ PCI_VDEVICE(INTEL, 0x7e46), LPSS_CNL_SSP },
 	/* CNL-LP */
 	{ PCI_VDEVICE(INTEL, 0x9daa), LPSS_CNL_SSP },
 	{ PCI_VDEVICE(INTEL, 0x9dab), LPSS_CNL_SSP },

drivers/spi/spi-s3c64xx.c

···
 
 #include <linux/platform_data/spi-s3c64xx.h>
 
-#define MAX_SPI_PORTS		6
+#define MAX_SPI_PORTS		12
 #define S3C64XX_SPI_QUIRK_POLL		(1 << 0)
 #define S3C64XX_SPI_QUIRK_CS_AUTO	(1 << 1)
 #define AUTOSUSPEND_TIMEOUT	2000
···
 #define S3C64XX_SPI_MODE_BUS_TSZ_HALFWORD	(1<<17)
 #define S3C64XX_SPI_MODE_BUS_TSZ_WORD		(2<<17)
 #define S3C64XX_SPI_MODE_BUS_TSZ_MASK		(3<<17)
+#define S3C64XX_SPI_MODE_SELF_LOOPBACK		(1<<3)
 #define S3C64XX_SPI_MODE_RXDMA_ON		(1<<2)
 #define S3C64XX_SPI_MODE_TXDMA_ON		(1<<1)
 #define S3C64XX_SPI_MODE_4BURST			(1<<0)
···
  * @fifo_lvl_mask: Bit-mask for {TX|RX}_FIFO_LVL bits in SPI_STATUS register.
  * @rx_lvl_offset: Bit offset of RX_FIFO_LVL bits in SPI_STATUS regiter.
  * @tx_st_done: Bit offset of TX_DONE bit in SPI_STATUS regiter.
+ * @clk_div: Internal clock divider
  * @quirks: Bitmask of known quirks
  * @high_speed: True, if the controller supports HIGH_SPEED_EN bit.
  * @clk_from_cmu: True, if the controller does not include a clock mux and
  *	prescaler unit.
  * @clk_ioclk: True if clock is present on this device
+ * @has_loopback: True if loopback mode can be supported
  *
  * The Samsung s3c64xx SPI controller are used on various Samsung SoC's but
  * differ in some aspects such as the size of the fifo and spi bus clock
···
 	int	rx_lvl_offset;
 	int	tx_st_done;
 	int	quirks;
+	int	clk_div;
 	bool	high_speed;
 	bool	clk_from_cmu;
 	bool	clk_ioclk;
+	bool	has_loopback;
 };
 
 /**
···
 	if (is_polling(sdd))
 		return 0;
 
+	/* Requests DMA channels */
+	sdd->rx_dma.ch = dma_request_chan(&sdd->pdev->dev, "rx");
+	if (IS_ERR(sdd->rx_dma.ch)) {
+		dev_err(&sdd->pdev->dev, "Failed to get RX DMA channel\n");
+		sdd->rx_dma.ch = NULL;
+		return 0;
+	}
+
+	sdd->tx_dma.ch = dma_request_chan(&sdd->pdev->dev, "tx");
+	if (IS_ERR(sdd->tx_dma.ch)) {
+		dev_err(&sdd->pdev->dev, "Failed to get TX DMA channel\n");
+		dma_release_channel(sdd->rx_dma.ch);
+		sdd->tx_dma.ch = NULL;
+		sdd->rx_dma.ch = NULL;
+		return 0;
+	}
+
 	spi->dma_rx = sdd->rx_dma.ch;
 	spi->dma_tx = sdd->tx_dma.ch;
+
+	return 0;
+}
+
+static int s3c64xx_spi_unprepare_transfer(struct spi_master *spi)
+{
+	struct s3c64xx_spi_driver_data *sdd = spi_master_get_devdata(spi);
+
+	if (is_polling(sdd))
+		return 0;
+
+	/* Releases DMA channels if they are allocated */
+	if (sdd->rx_dma.ch && sdd->tx_dma.ch) {
+		dma_release_channel(sdd->rx_dma.ch);
+		dma_release_channel(sdd->tx_dma.ch);
+		sdd->rx_dma.ch = 0;
+		sdd->tx_dma.ch = 0;
+	}
 
 	return 0;
 }
···
 {
 	struct s3c64xx_spi_driver_data *sdd = spi_master_get_devdata(master);
 
-	return xfer->len > (FIFO_LVL_MASK(sdd) >> 1) + 1;
+	if (sdd->rx_dma.ch && sdd->tx_dma.ch) {
+		return xfer->len > (FIFO_LVL_MASK(sdd) >> 1) + 1;
+	} else {
+		return false;
+	}
+
 }
 
 static int s3c64xx_enable_datapath(struct s3c64xx_spi_driver_data *sdd,
···
 	void __iomem *regs = sdd->regs;
 	int ret;
 	u32 val;
+	int div = sdd->port_conf->clk_div;
 
 	/* Disable Clock */
 	if (!sdd->port_conf->clk_from_cmu) {
···
 		break;
 	}
 
+	if ((sdd->cur_mode & SPI_LOOP) && sdd->port_conf->has_loopback)
+		val |= S3C64XX_SPI_MODE_SELF_LOOPBACK;
+
 	writel(val, regs + S3C64XX_SPI_MODE_CFG);
 
 	if (sdd->port_conf->clk_from_cmu) {
-		/* The src_clk clock is divided internally by 2 */
-		ret = clk_set_rate(sdd->src_clk, sdd->cur_speed * 2);
+		ret = clk_set_rate(sdd->src_clk, sdd->cur_speed * div);
 		if (ret)
 			return ret;
-		sdd->cur_speed = clk_get_rate(sdd->src_clk) / 2;
+		sdd->cur_speed = clk_get_rate(sdd->src_clk) / div;
 	} else {
 		/* Configure Clock */
 		val = readl(regs + S3C64XX_SPI_CLK_CFG);
 		val &= ~S3C64XX_SPI_PSR_MASK;
-		val |= ((clk_get_rate(sdd->src_clk) / sdd->cur_speed / 2 - 1)
+		val |= ((clk_get_rate(sdd->src_clk) / sdd->cur_speed / div - 1)
 				& S3C64XX_SPI_PSR_MASK);
 		writel(val, regs + S3C64XX_SPI_CLK_CFG);
 
···
 	    sdd->rx_dma.ch && sdd->tx_dma.ch) {
 		use_dma = 1;
 
-	} else if (is_polling(sdd) && xfer->len > fifo_len) {
+	} else if (xfer->len > fifo_len) {
 		tx_buf = xfer->tx_buf;
 		rx_buf = xfer->rx_buf;
 		origin_len = xfer->len;
···
 	struct s3c64xx_spi_csinfo *cs = spi->controller_data;
 	struct s3c64xx_spi_driver_data *sdd;
 	int err;
+	int div;
 
 	sdd = spi_master_get_devdata(spi->master);
 	if (spi->dev.of_node) {
···
 
 	pm_runtime_get_sync(&sdd->pdev->dev);
 
+	div = sdd->port_conf->clk_div;
+
 	/* Check if we can provide the requested rate */
 	if (!sdd->port_conf->clk_from_cmu) {
 		u32 psr, speed;
 
 		/* Max possible */
-		speed = clk_get_rate(sdd->src_clk) / 2 / (0 + 1);
+		speed = clk_get_rate(sdd->src_clk) / div / (0 + 1);
 
 		if (spi->max_speed_hz > speed)
 			spi->max_speed_hz = speed;
 
-		psr = clk_get_rate(sdd->src_clk) / 2 / spi->max_speed_hz - 1;
+		psr = clk_get_rate(sdd->src_clk) / div / spi->max_speed_hz - 1;
 		psr &= S3C64XX_SPI_PSR_MASK;
 		if (psr == S3C64XX_SPI_PSR_MASK)
 			psr--;
 
-		speed = clk_get_rate(sdd->src_clk) / 2 / (psr + 1);
+		speed = clk_get_rate(sdd->src_clk) / div / (psr + 1);
 		if (spi->max_speed_hz < speed) {
 			if (psr+1 < S3C64XX_SPI_PSR_MASK) {
 				psr++;
···
 			}
 		}
 
-		speed = clk_get_rate(sdd->src_clk) / 2 / (psr + 1);
+		speed = clk_get_rate(sdd->src_clk) / div / (psr + 1);
 		if (spi->max_speed_hz >= speed) {
 			spi->max_speed_hz = speed;
 		} else {
···
 	master->setup = s3c64xx_spi_setup;
 	master->cleanup = s3c64xx_spi_cleanup;
 	master->prepare_transfer_hardware = s3c64xx_spi_prepare_transfer;
+	master->unprepare_transfer_hardware = s3c64xx_spi_unprepare_transfer;
 	master->prepare_message = s3c64xx_spi_prepare_message;
 	master->transfer_one = s3c64xx_spi_transfer_one;
 	master->num_chipselect = sci->num_cs;
···
 					SPI_BPW_MASK(8);
 	/* the spi->mode bits understood by this driver: */
 	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_CS_HIGH;
+	if (sdd->port_conf->has_loopback)
+		master->mode_bits |= SPI_LOOP;
 	master->auto_runtime_pm = true;
 	if (!is_polling(sdd))
 		master->can_dma = s3c64xx_spi_can_dma;
···
 		}
 	}
 
-	if (!is_polling(sdd)) {
-		/* Acquire DMA channels */
-		sdd->rx_dma.ch = dma_request_chan(&pdev->dev, "rx");
-		if (IS_ERR(sdd->rx_dma.ch)) {
-			dev_err(&pdev->dev, "Failed to get RX DMA channel\n");
-			ret = PTR_ERR(sdd->rx_dma.ch);
-			goto err_disable_io_clk;
-		}
-		sdd->tx_dma.ch = dma_request_chan(&pdev->dev, "tx");
-		if (IS_ERR(sdd->tx_dma.ch)) {
-			dev_err(&pdev->dev, "Failed to get TX DMA channel\n");
-			ret = PTR_ERR(sdd->tx_dma.ch);
-			goto err_release_rx_dma;
-		}
-	}
-
 	pm_runtime_set_autosuspend_delay(&pdev->dev, AUTOSUSPEND_TIMEOUT);
 	pm_runtime_use_autosuspend(&pdev->dev);
 	pm_runtime_set_active(&pdev->dev);
···
 	pm_runtime_disable(&pdev->dev);
 	pm_runtime_set_suspended(&pdev->dev);
 
-	if (!is_polling(sdd))
-		dma_release_channel(sdd->tx_dma.ch);
-err_release_rx_dma:
-	if (!is_polling(sdd))
-		dma_release_channel(sdd->rx_dma.ch);
-err_disable_io_clk:
 	clk_disable_unprepare(sdd->ioclk);
 err_disable_src_clk:
 	clk_disable_unprepare(sdd->src_clk);
···
 	.fifo_lvl_mask	= { 0x7f },
 	.rx_lvl_offset	= 13,
 	.tx_st_done	= 21,
+	.clk_div	= 2,
 	.high_speed	= true,
 };
 
···
 	.fifo_lvl_mask	= { 0x7f, 0x7F },
 	.rx_lvl_offset	= 13,
 	.tx_st_done	= 21,
+	.clk_div	= 2,
 };
 
 static const struct s3c64xx_spi_port_config s5pv210_spi_port_config = {
 	.fifo_lvl_mask	= { 0x1ff, 0x7F },
 	.rx_lvl_offset	= 15,
 	.tx_st_done	= 25,
+	.clk_div	= 2,
 	.high_speed	= true,
 };
 
···
 	.fifo_lvl_mask	= { 0x1ff, 0x7F, 0x7F },
 	.rx_lvl_offset	= 15,
 	.tx_st_done	= 25,
+	.clk_div	= 2,
 	.high_speed	= true,
 	.clk_from_cmu	= true,
 	.quirks		= S3C64XX_SPI_QUIRK_CS_AUTO,
···
 	.fifo_lvl_mask	= { 0x1ff, 0x7F, 0x7F, 0x7F, 0x7F, 0x1ff},
 	.rx_lvl_offset	= 15,
 	.tx_st_done	= 25,
+	.clk_div	= 2,
 	.high_speed	= true,
 	.clk_from_cmu	= true,
 	.quirks		= S3C64XX_SPI_QUIRK_CS_AUTO,
···
 	.fifo_lvl_mask	= { 0x1ff, 0x7f, 0x7f, 0x7f, 0x7f, 0x1ff},
 	.rx_lvl_offset	= 15,
 	.tx_st_done	= 25,
+	.clk_div	= 2,
 	.high_speed	= true,
 	.clk_from_cmu	= true,
 	.clk_ioclk	= true,
 	.quirks		= S3C64XX_SPI_QUIRK_CS_AUTO,
 };
 
-static struct s3c64xx_spi_port_config fsd_spi_port_config = {
+static const struct s3c64xx_spi_port_config exynosautov9_spi_port_config = {
+	.fifo_lvl_mask	= { 0x1ff, 0x1ff, 0x7f, 0x7f, 0x7f, 0x7f, 0x1ff, 0x7f,
+			    0x7f, 0x7f, 0x7f, 0x7f},
+	.rx_lvl_offset	= 15,
+	.tx_st_done	= 25,
+	.clk_div	= 4,
+	.high_speed	= true,
+	.clk_from_cmu	= true,
+	.clk_ioclk	= true,
+	.has_loopback	= true,
+	.quirks		= S3C64XX_SPI_QUIRK_CS_AUTO,
+};
+
+static const struct s3c64xx_spi_port_config fsd_spi_port_config = {
 	.fifo_lvl_mask	= { 0x7f, 0x7f, 0x7f, 0x7f, 0x7f},
 	.rx_lvl_offset	= 15,
 	.tx_st_done	= 25,
+	.clk_div	= 2,
 	.high_speed	= true,
 	.clk_from_cmu	= true,
 	.clk_ioclk	= false,
···
 	},
 	{ .compatible = "samsung,exynos5433-spi",
 			.data = (void *)&exynos5433_spi_port_config,
+	},
+	{ .compatible = "samsung,exynosautov9-spi",
+			.data = (void *)&exynosautov9_spi_port_config,
 	},
 	{ .compatible = "tesla,fsd-spi",
 			.data = (void *)&fsd_spi_port_config,

drivers/spi/spi-sh.c

···
 	void __iomem *addr;
 	int irq;
 	struct spi_master *master;
-	struct list_head queue;
-	struct work_struct ws;
 	unsigned long cr1;
 	wait_queue_head_t wait;
-	spinlock_t lock;
 	int width;
 };
 
···
 	return 0;
 }
 
-static void spi_sh_work(struct work_struct *work)
+static int spi_sh_transfer_one_message(struct spi_controller *ctlr,
+					struct spi_message *mesg)
 {
-	struct spi_sh_data *ss = container_of(work, struct spi_sh_data, ws);
-	struct spi_message *mesg;
+	struct spi_sh_data *ss = spi_controller_get_devdata(ctlr);
 	struct spi_transfer *t;
-	unsigned long flags;
 	int ret;
 
 	pr_debug("%s: enter\n", __func__);
 
-	spin_lock_irqsave(&ss->lock, flags);
-	while (!list_empty(&ss->queue)) {
-		mesg = list_entry(ss->queue.next, struct spi_message, queue);
-		list_del_init(&mesg->queue);
+	spi_sh_clear_bit(ss, SPI_SH_SSA, SPI_SH_CR1);
 
-		spin_unlock_irqrestore(&ss->lock, flags);
-		list_for_each_entry(t, &mesg->transfers, transfer_list) {
-			pr_debug("tx_buf = %p, rx_buf = %p\n",
-					t->tx_buf, t->rx_buf);
-			pr_debug("len = %d, delay.value = %d\n",
-					t->len, t->delay.value);
+	list_for_each_entry(t, &mesg->transfers, transfer_list) {
+		pr_debug("tx_buf = %p, rx_buf = %p\n",
+			 t->tx_buf, t->rx_buf);
+		pr_debug("len = %d, delay.value = %d\n",
+			 t->len, t->delay.value);
 
-			if (t->tx_buf) {
-				ret = spi_sh_send(ss, mesg, t);
-				if (ret < 0)
-					goto error;
-			}
-			if (t->rx_buf) {
-				ret = spi_sh_receive(ss, mesg, t);
-				if (ret < 0)
-					goto error;
-			}
-			mesg->actual_length += t->len;
+		if (t->tx_buf) {
+			ret = spi_sh_send(ss, mesg, t);
+			if (ret < 0)
+				goto error;
 		}
-		spin_lock_irqsave(&ss->lock, flags);
-
-		mesg->status = 0;
-		if (mesg->complete)
-			mesg->complete(mesg->context);
+		if (t->rx_buf) {
+			ret = spi_sh_receive(ss, mesg, t);
+			if (ret < 0)
+				goto error;
+		}
+		mesg->actual_length += t->len;
 	}
+
+	mesg->status = 0;
+	spi_finalize_current_message(ctlr);
 
 	clear_fifo(ss);
 	spi_sh_set_bit(ss, SPI_SH_SSD, SPI_SH_CR1);
···
 
 	clear_fifo(ss);
 
-	spin_unlock_irqrestore(&ss->lock, flags);
-
-	return;
+	return 0;
 
  error:
 	mesg->status = ret;
+	spi_finalize_current_message(ctlr);
 	if (mesg->complete)
 		mesg->complete(mesg->context);
 
···
 			 SPI_SH_CR1);
 	clear_fifo(ss);
 
+	return ret;
 }
 
 static int spi_sh_setup(struct spi_device *spi)
···
 	/* 1/8 clock */
 	spi_sh_write(ss, spi_sh_read(ss, SPI_SH_CR2) | 0x07, SPI_SH_CR2);
 	udelay(10);
-
-	return 0;
-}
-
-static int spi_sh_transfer(struct spi_device *spi, struct spi_message *mesg)
-{
-	struct spi_sh_data *ss = spi_master_get_devdata(spi->master);
-	unsigned long flags;
-
-	pr_debug("%s: enter\n", __func__);
-	pr_debug("\tmode = %02x\n", spi->mode);
-
-	spin_lock_irqsave(&ss->lock, flags);
-
-	mesg->actual_length = 0;
-	mesg->status = -EINPROGRESS;
-
-	spi_sh_clear_bit(ss, SPI_SH_SSA, SPI_SH_CR1);
-
-	list_add_tail(&mesg->queue, &ss->queue);
-	schedule_work(&ss->ws);
-
-	spin_unlock_irqrestore(&ss->lock, flags);
 
 	return 0;
 }
···
 	struct spi_sh_data *ss = platform_get_drvdata(pdev);
 
 	spi_unregister_master(ss->master);
-	flush_work(&ss->ws);
 	free_irq(ss->irq, ss);
 
 	return 0;
···
 		dev_err(&pdev->dev, "ioremap error.\n");
 		return -ENOMEM;
 	}
-	INIT_LIST_HEAD(&ss->queue);
-	spin_lock_init(&ss->lock);
-	INIT_WORK(&ss->ws, spi_sh_work);
 	init_waitqueue_head(&ss->wait);
 
 	ret = request_irq(irq, spi_sh_irq, 0, "spi_sh", ss);
···
 	master->num_chipselect = 2;
 	master->bus_num = pdev->id;
 	master->setup = spi_sh_setup;
-	master->transfer = spi_sh_transfer;
+	master->transfer_one_message = spi_sh_transfer_one_message;
 	master->cleanup = spi_sh_cleanup;
 
 	ret = spi_register_master(master);

drivers/spi/spi-sifive.c

···
 	return 0;
 }
 
+static int sifive_spi_suspend(struct device *dev)
+{
+	struct spi_master *master = dev_get_drvdata(dev);
+	struct sifive_spi *spi = spi_master_get_devdata(master);
+	int ret;
+
+	ret = spi_master_suspend(master);
+	if (ret)
+		return ret;
+
+	/* Disable all the interrupts just in case */
+	sifive_spi_write(spi, SIFIVE_SPI_REG_IE, 0);
+
+	clk_disable_unprepare(spi->clk);
+
+	return ret;
+}
+
+static int sifive_spi_resume(struct device *dev)
+{
+	struct spi_master *master = dev_get_drvdata(dev);
+	struct sifive_spi *spi = spi_master_get_devdata(master);
+	int ret;
+
+	ret = clk_prepare_enable(spi->clk);
+	if (ret)
+		return ret;
+	ret = spi_master_resume(master);
+	if (ret)
+		clk_disable_unprepare(spi->clk);
+
+	return ret;
+}
+
+static DEFINE_SIMPLE_DEV_PM_OPS(sifive_spi_pm_ops,
+				sifive_spi_suspend, sifive_spi_resume);
+
+
 static const struct of_device_id sifive_spi_of_match[] = {
 	{ .compatible = "sifive,spi0", },
 	{}
···
 	.remove = sifive_spi_remove,
 	.driver = {
 		.name = SIFIVE_SPI_DRIVER_NAME,
+		.pm = &sifive_spi_pm_ops,
 		.of_match_table = sifive_spi_of_match,
 	},
 };

drivers/spi/spi-stm32-qspi.c

···
 						 STM32_BUSY_TIMEOUT_US);
 }
 
-static int stm32_qspi_wait_cmd(struct stm32_qspi *qspi,
-			       const struct spi_mem_op *op)
+static int stm32_qspi_wait_cmd(struct stm32_qspi *qspi)
 {
 	u32 cr, sr;
 	int err = 0;
···
 	return err;
 }
 
-static int stm32_qspi_wait_poll_status(struct stm32_qspi *qspi,
-				       const struct spi_mem_op *op)
+static int stm32_qspi_wait_poll_status(struct stm32_qspi *qspi)
 {
 	u32 cr;
 
···
 	return 0;
 }
 
-static int stm32_qspi_get_mode(struct stm32_qspi *qspi, u8 buswidth)
+static int stm32_qspi_get_mode(u8 buswidth)
 {
 	if (buswidth == 4)
 		return CCR_BUSWIDTH_4;
···
 	ccr = qspi->fmode;
 	ccr |= FIELD_PREP(CCR_INST_MASK, op->cmd.opcode);
 	ccr |= FIELD_PREP(CCR_IMODE_MASK,
-			  stm32_qspi_get_mode(qspi, op->cmd.buswidth));
+			  stm32_qspi_get_mode(op->cmd.buswidth));
 
 	if (op->addr.nbytes) {
 		ccr |= FIELD_PREP(CCR_ADMODE_MASK,
-				  stm32_qspi_get_mode(qspi, op->addr.buswidth));
+				  stm32_qspi_get_mode(op->addr.buswidth));
 		ccr |= FIELD_PREP(CCR_ADSIZE_MASK, op->addr.nbytes - 1);
 	}
 
···
 
 	if (op->data.nbytes) {
 		ccr |= FIELD_PREP(CCR_DMODE_MASK,
-				  stm32_qspi_get_mode(qspi, op->data.buswidth));
+				  stm32_qspi_get_mode(op->data.buswidth));
 	}
 
 	writel_relaxed(ccr, qspi->io_base + QSPI_CCR);
···
 		writel_relaxed(op->addr.val, qspi->io_base + QSPI_AR);
 
 	if (qspi->fmode == CCR_FMODE_APM)
-		err_poll_status = stm32_qspi_wait_poll_status(qspi, op);
+		err_poll_status = stm32_qspi_wait_poll_status(qspi);
 
 	err = stm32_qspi_tx(qspi, op);
 
···
 		goto abort;
 
 	/* wait end of tx in indirect mode */
-	err = stm32_qspi_wait_cmd(qspi, op);
+	err = stm32_qspi_wait_cmd(qspi);
 	if (err)
 		goto abort;
 

drivers/spi/spi-synquacer.c

···
 
 		ret = synquacer_spi_enable(master);
 		if (ret) {
+			clk_disable_unprepare(sspi->clk);
 			dev_err(dev, "failed to enable spi (%d)\n", ret);
 			return ret;
 		}

drivers/spi/spi-tegra20-slink.c

···
 
 static int tegra_slink_remove(struct platform_device *pdev)
 {
-	struct spi_master *master = platform_get_drvdata(pdev);
+	struct spi_master *master = spi_master_get(platform_get_drvdata(pdev));
 	struct tegra_slink_data	*tspi = spi_master_get_devdata(master);
 
 	spi_unregister_master(master);
···
 	if (tspi->rx_dma_chan)
 		tegra_slink_deinit_dma_param(tspi, true);
 
+	spi_master_put(master);
 	return 0;
 }
 

drivers/spi/spi-tegra210-quad.c

···
 #define QSPI_RX_EN				BIT(12)
 #define QSPI_CS_SW_VAL				BIT(20)
 #define QSPI_CS_SW_HW				BIT(21)
+
+#define QSPI_CS_POL_INACTIVE(n)			(1 << (22 + (n)))
+#define QSPI_CS_POL_INACTIVE_MASK		(0xF << 22)
+#define QSPI_CS_SEL_0				(0 << 26)
+#define QSPI_CS_SEL_1				(1 << 26)
+#define QSPI_CS_SEL_2				(2 << 26)
+#define QSPI_CS_SEL_3				(3 << 26)
+#define QSPI_CS_SEL_MASK			(3 << 26)
+#define QSPI_CS_SEL(x)				(((x) & 0x3) << 26)
+
 #define QSPI_CONTROL_MODE_0			(0 << 28)
 #define QSPI_CONTROL_MODE_3			(3 << 28)
 #define QSPI_CONTROL_MODE_MASK			(3 << 28)
···
 struct tegra_qspi_soc_data {
 	bool has_dma;
 	bool cmb_xfer_capable;
+	unsigned int cs_count;
 };
 
 struct tegra_qspi_client_data {
···
 		tegra_qspi_mask_clear_irq(tqspi);
 
 		command1 = tqspi->def_command1_reg;
+		command1 |= QSPI_CS_SEL(spi->chip_select);
 		command1 |= QSPI_BIT_LENGTH(bits_per_word - 1);
 
 		command1 &= ~QSPI_CONTROL_MODE_MASK;
···
 
 	/* keep default cs state to inactive */
 	val = tqspi->def_command1_reg;
+	val |= QSPI_CS_SEL(spi->chip_select);
 	if (spi->mode & SPI_CS_HIGH)
-		val &= ~QSPI_CS_SW_VAL;
+		val &= ~QSPI_CS_POL_INACTIVE(spi->chip_select);
 	else
-		val |= QSPI_CS_SW_VAL;
+		val |= QSPI_CS_POL_INACTIVE(spi->chip_select);
 
 	tqspi->def_command1_reg = val;
 	tegra_qspi_writel(tqspi, tqspi->def_command1_reg, QSPI_COMMAND1);
···
 static struct tegra_qspi_soc_data tegra210_qspi_soc_data = {
 	.has_dma = true,
 	.cmb_xfer_capable = false,
+	.cs_count = 1,
 };
 
 static struct tegra_qspi_soc_data tegra186_qspi_soc_data = {
 	.has_dma = true,
 	.cmb_xfer_capable = true,
+	.cs_count = 1,
 };
 
 static struct tegra_qspi_soc_data tegra234_qspi_soc_data = {
 	.has_dma = false,
 	.cmb_xfer_capable = true,
+	.cs_count = 1,
+};
+
+static struct tegra_qspi_soc_data tegra241_qspi_soc_data = {
+	.has_dma = false,
+	.cmb_xfer_capable = true,
+	.cs_count = 4,
 };
 
 static const struct of_device_id tegra_qspi_of_match[] = {
···
 	}, {
 		.compatible = "nvidia,tegra234-qspi",
 		.data	    = &tegra234_qspi_soc_data,
+	}, {
+		.compatible = "nvidia,tegra241-qspi",
+		.data	    = &tegra241_qspi_soc_data,
 	},
 	{}
 };
···
 	}, {
 		.id = "NVDA1413",
 		.driver_data = (kernel_ulong_t)&tegra234_qspi_soc_data,
+	}, {
+		.id = "NVDA1513",
+		.driver_data = (kernel_ulong_t)&tegra241_qspi_soc_data,
 	},
 	{}
 };
···
 	spin_lock_init(&tqspi->lock);
 
 	tqspi->soc_data = device_get_match_data(&pdev->dev);
+	master->num_chipselect = tqspi->soc_data->cs_count;
 	r = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	tqspi->base = devm_ioremap_resource(&pdev->dev, r);
 	if (IS_ERR(tqspi->base))

drivers/spi/spi-ti-qspi.c

···
 	void			*rx_bb_addr;
 	struct dma_chan		*rx_chan;
 
-	u32 spi_max_frequency;
 	u32 cmd;
 	u32 dc;
 
···
 static int ti_qspi_setup(struct spi_device *spi)
 {
 	struct ti_qspi	*qspi = spi_master_get_devdata(spi->master);
-	struct ti_qspi_regs *ctx_reg = &qspi->ctx_reg;
-	int clk_div = 0, ret;
-	u32 clk_ctrl_reg, clk_rate, clk_mask;
+	int ret;
 
 	if (spi->master->busy) {
 		dev_dbg(qspi->dev, "master busy doing other transfers\n");
 		return -EBUSY;
 	}
 
-	if (!qspi->spi_max_frequency) {
+	if (!qspi->master->max_speed_hz) {
 		dev_err(qspi->dev, "spi max frequency not defined\n");
 		return -EINVAL;
 	}
 
-	clk_rate = clk_get_rate(qspi->fclk);
-
-	clk_div = DIV_ROUND_UP(clk_rate, qspi->spi_max_frequency) - 1;
-
-	if (clk_div < 0) {
-		dev_dbg(qspi->dev, "clock divider < 0, using /1 divider\n");
-		return -EINVAL;
-	}
-
-	if (clk_div > QSPI_CLK_DIV_MAX) {
-		dev_dbg(qspi->dev, "clock divider >%d , using /%d divider\n",
-				QSPI_CLK_DIV_MAX, QSPI_CLK_DIV_MAX + 1);
-		return -EINVAL;
-	}
-
-	dev_dbg(qspi->dev, "hz: %d, clock divider %d\n",
-			qspi->spi_max_frequency, clk_div);
+	spi->max_speed_hz = min(spi->max_speed_hz, qspi->master->max_speed_hz);
 
 	ret = pm_runtime_resume_and_get(qspi->dev);
 	if (ret < 0) {
 		dev_err(qspi->dev, "pm_runtime_get_sync() failed\n");
 		return ret;
 	}
-
-	clk_ctrl_reg = ti_qspi_read(qspi, QSPI_SPI_CLOCK_CNTRL_REG);
-
-	clk_ctrl_reg &= ~QSPI_CLK_EN;
-
-	/* disable SCLK */
-	ti_qspi_write(qspi, clk_ctrl_reg, QSPI_SPI_CLOCK_CNTRL_REG);
-
-	/* enable SCLK */
-	clk_mask = QSPI_CLK_EN | clk_div;
-	ti_qspi_write(qspi, clk_mask, QSPI_SPI_CLOCK_CNTRL_REG);
-	ctx_reg->clkctrl = clk_mask;
 
 	pm_runtime_mark_last_busy(qspi->dev);
 	ret = pm_runtime_put_autosuspend(qspi->dev);
···
 	}
 
 	return 0;
+}
+
+static void ti_qspi_setup_clk(struct ti_qspi *qspi, u32 speed_hz)
+{
+	struct ti_qspi_regs *ctx_reg = &qspi->ctx_reg;
+	int clk_div;
+	u32 clk_ctrl_reg, clk_rate, clk_ctrl_new;
+
+	clk_rate = clk_get_rate(qspi->fclk);
+	clk_div = DIV_ROUND_UP(clk_rate, speed_hz) - 1;
+	clk_div = clamp(clk_div, 0, QSPI_CLK_DIV_MAX);
+	dev_dbg(qspi->dev, "hz: %d, clock divider %d\n", speed_hz, clk_div);
+
+	pm_runtime_resume_and_get(qspi->dev);
+
+	clk_ctrl_new = QSPI_CLK_EN | clk_div;
+	if (ctx_reg->clkctrl != clk_ctrl_new) {
+		clk_ctrl_reg = ti_qspi_read(qspi, QSPI_SPI_CLOCK_CNTRL_REG);
+
+		clk_ctrl_reg &= ~QSPI_CLK_EN;
+
+		/* disable SCLK */
+		ti_qspi_write(qspi, clk_ctrl_reg, QSPI_SPI_CLOCK_CNTRL_REG);
+
+		/* enable SCLK */
+		ti_qspi_write(qspi, clk_ctrl_new, QSPI_SPI_CLOCK_CNTRL_REG);
+		ctx_reg->clkctrl = clk_ctrl_new;
+	}
+
+	pm_runtime_mark_last_busy(qspi->dev);
+	pm_runtime_put_autosuspend(qspi->dev);
 }
 
 static void ti_qspi_restore_ctx(struct ti_qspi *qspi)
···
 
 	mutex_lock(&qspi->list_lock);
 
-	if (!qspi->mmap_enabled || qspi->current_cs != mem->spi->chip_select)
+	if (!qspi->mmap_enabled || qspi->current_cs != mem->spi->chip_select) {
+		ti_qspi_setup_clk(qspi, mem->spi->max_speed_hz);
 		ti_qspi_enable_memory_map(mem->spi);
+	}
 	ti_qspi_setup_mmap_read(mem->spi, op->cmd.opcode, op->data.buswidth,
 				op->addr.nbytes, op->dummy.nbytes);
 
···
 		wlen = t->bits_per_word >> 3;
 		transfer_len_words = min(t->len / wlen, frame_len_words);
 
+		ti_qspi_setup_clk(qspi, t->speed_hz);
 		ret = qspi_transfer_msg(qspi, t, transfer_len_words * wlen);
 		if (ret) {
 			dev_dbg(qspi->dev, "transfer message failed\n");
···
 	pm_runtime_enable(&pdev->dev);
 
 	if (!of_property_read_u32(np, "spi-max-frequency", &max_freq))
-		qspi->spi_max_frequency = max_freq;
+		master->max_speed_hz = max_freq;
 
 	dma_cap_zero(mask);
 	dma_cap_set(DMA_MEMCPY, mask);

drivers/spi/spi-topcliff-pch.c

···
 
 static int pch_spi_transfer(struct spi_device *pspi, struct spi_message *pmsg)
 {
-
-	struct spi_transfer *transfer;
 	struct pch_spi_data *data = spi_master_get_devdata(pspi->master);
 	int retval;
 	unsigned long flags;
-
-	spin_lock_irqsave(&data->lock, flags);
-	/* validate Tx/Rx buffers and Transfer length */
-	list_for_each_entry(transfer, &pmsg->transfers, transfer_list) {
-		if (!transfer->tx_buf && !transfer->rx_buf) {
-			dev_err(&pspi->dev,
-				"%s Tx and Rx buffer NULL\n", __func__);
-			retval = -EINVAL;
-			goto err_return_spinlock;
-		}
-
-		if (!transfer->len) {
-			dev_err(&pspi->dev, "%s Transfer length invalid\n",
-				__func__);
-			retval = -EINVAL;
-			goto err_return_spinlock;
-		}
-
-		dev_dbg(&pspi->dev,
-			"%s Tx/Rx buffer valid. Transfer length valid\n",
-			__func__);
-	}
-	spin_unlock_irqrestore(&data->lock, flags);
 
 	/* We won't process any messages if we have been asked to terminate */
 	if (data->status == STATUS_EXITING) {
···
 
 err_out:
 	dev_dbg(&pspi->dev, "%s RETURN=%d\n", __func__, retval);
-	return retval;
-err_return_spinlock:
-	dev_dbg(&pspi->dev, "%s RETURN=%d\n", __func__, retval);
-	spin_unlock_irqrestore(&data->lock, flags);
 	return retval;
 }
 
···
 	master->mode_bits = SPI_CPOL | SPI_CPHA | SPI_LSB_FIRST;
 	master->bits_per_word_mask = SPI_BPW_MASK(8) | SPI_BPW_MASK(16);
 	master->max_speed_hz = PCH_MAX_BAUDRATE;
+	master->flags = SPI_CONTROLLER_MUST_RX | SPI_CONTROLLER_MUST_TX;
 
 	data->board_dat = board_dat;
 	data->plat_dev = plat_dev;

drivers/spi/spi-zynqmp-gqspi.c

···
 #define GQSPI_DMA_UNALIGN		0x3
 #define GQSPI_DEFAULT_NUM_CS	1	/* Default number of chip selects */
 
+#define GQSPI_MAX_NUM_CS	2	/* Maximum number of chip selects */
+
 #define SPI_AUTOSUSPEND_TIMEOUT		3000
 enum mode_type {GQSPI_MODE_IO, GQSPI_MODE_DMA};
 
···
 	genfifoentry |= GQSPI_GENFIFO_MODE_SPI;
 
 	if (!is_high) {
-		xqspi->genfifobus = GQSPI_GENFIFO_BUS_LOWER;
-		xqspi->genfifocs = GQSPI_GENFIFO_CS_LOWER;
+		if (!qspi->chip_select) {
+			xqspi->genfifobus = GQSPI_GENFIFO_BUS_LOWER;
+			xqspi->genfifocs = GQSPI_GENFIFO_CS_LOWER;
+		} else {
+			xqspi->genfifobus = GQSPI_GENFIFO_BUS_UPPER;
+			xqspi->genfifocs = GQSPI_GENFIFO_CS_UPPER;
+		}
 		genfifoentry |= xqspi->genfifobus;
 		genfifoentry |= xqspi->genfifocs;
 		genfifoentry |= GQSPI_GENFIFO_CS_SETUP;
···
 	struct zynqmp_qspi *xqspi;
 	struct device *dev = &pdev->dev;
 	struct device_node *np = dev->of_node;
+	u32 num_cs;
 
 	ctlr = spi_alloc_master(&pdev->dev, sizeof(*xqspi));
 	if (!ctlr)
···
 	if (ret)
 		goto clk_dis_all;
 
+	ret = of_property_read_u32(np, "num-cs", &num_cs);
+	if (ret < 0) {
+		ctlr->num_chipselect = GQSPI_DEFAULT_NUM_CS;
+	} else if (num_cs > GQSPI_MAX_NUM_CS) {
+		ret = -EINVAL;
+		dev_err(&pdev->dev, "only %d chip selects are available\n",
+			GQSPI_MAX_NUM_CS);
+		goto clk_dis_all;
+	} else {
+		ctlr->num_chipselect = num_cs;
+	}
+
 	ctlr->bits_per_word_mask = SPI_BPW_MASK(8);
-	ctlr->num_chipselect = GQSPI_DEFAULT_NUM_CS;
 	ctlr->mem_ops = &zynqmp_qspi_mem_ops;
 	ctlr->setup = zynqmp_qspi_setup_op;
 	ctlr->max_speed_hz = clk_get_rate(xqspi->refclk) / 2;

drivers/spi/spi.c

···
 #include <linux/idr.h>
 #include <linux/platform_data/x86/apple.h>
 #include <linux/ptp_clock_kernel.h>
+#include <linux/percpu.h>
 
 #define CREATE_TRACE_POINTS
 #include <trace/events/spi.h>
···
 
 	spi_controller_put(spi->controller);
 	kfree(spi->driver_override);
+	free_percpu(spi->pcpu_statistics);
 	kfree(spi);
 }
 
···
 }
 static DEVICE_ATTR_RW(driver_override);
 
+static struct spi_statistics *spi_alloc_pcpu_stats(struct device *dev)
+{
+	struct spi_statistics __percpu *pcpu_stats;
+
+	if (dev)
+		pcpu_stats = devm_alloc_percpu(dev, struct spi_statistics);
+	else
+		pcpu_stats = alloc_percpu_gfp(struct spi_statistics, GFP_KERNEL);
+
+	if (pcpu_stats) {
+		int cpu;
+
+		for_each_possible_cpu(cpu) {
+			struct spi_statistics *stat;
+
+			stat = per_cpu_ptr(pcpu_stats, cpu);
+			u64_stats_init(&stat->syncp);
+		}
+	}
+	return pcpu_stats;
+}
+
+#define spi_pcpu_stats_totalize(ret, in, field)				\
+do {									\
+	int i;								\
+	ret = 0;							\
+	for_each_possible_cpu(i) {					\
+		const struct spi_statistics *pcpu_stats;		\
+		u64 inc;						\
+		unsigned int start;					\
+		pcpu_stats = per_cpu_ptr(in, i);			\
+		do {							\
+			start = u64_stats_fetch_begin_irq(		\
+					&pcpu_stats->syncp);		\
+			inc = u64_stats_read(&pcpu_stats->field);	\
+		} while (u64_stats_fetch_retry_irq(			\
+					&pcpu_stats->syncp, start));	\
+		ret += inc;						\
+	}								\
+} while (0)
+
 #define SPI_STATISTICS_ATTRS(field, file)				\
 static ssize_t spi_controller_##field##_show(struct device *dev,	\
 					     struct device_attribute *attr, \
···
 {									\
 	struct spi_controller *ctlr = container_of(dev,			\
 					 struct spi_controller, dev);	\
-	return spi_statistics_##field##_show(&ctlr->statistics, buf);	\
+	return spi_statistics_##field##_show(ctlr->pcpu_statistics, buf); \
 }									\
 static struct device_attribute dev_attr_spi_controller_##field = {	\
 	.attr = { .name = file, .mode = 0444 },				\
···
 					char *buf)			\
 {									\
 	struct spi_device *spi = to_spi_device(dev);			\
-	return spi_statistics_##field##_show(&spi->statistics, buf);	\
+	return spi_statistics_##field##_show(spi->pcpu_statistics, buf); \
 }									\
 static struct device_attribute dev_attr_spi_device_##field = {		\
 	.attr = { .name = file, .mode = 0444 },				\
 	.show = spi_device_##field##_show,				\
 }
 
-#define SPI_STATISTICS_SHOW_NAME(name, file, field, format_string)	\
+#define SPI_STATISTICS_SHOW_NAME(name, file, field)			\
 static ssize_t spi_statistics_##name##_show(struct spi_statistics *stat, \
 					    char *buf)			\
 {									\
-	unsigned long flags;						\
 	ssize_t len;							\
-	spin_lock_irqsave(&stat->lock, flags);				\
-	len = sysfs_emit(buf, format_string "\n", stat->field);		\
-	spin_unlock_irqrestore(&stat->lock, flags);			\
+	u64 val;							\
+	spi_pcpu_stats_totalize(val, stat, field);			\
+	len = sysfs_emit(buf, "%llu\n", val);				\
 	return len;							\
 }									\
 SPI_STATISTICS_ATTRS(name, file)
 
-#define SPI_STATISTICS_SHOW(field, format_string)			\
+#define SPI_STATISTICS_SHOW(field)					\
 	SPI_STATISTICS_SHOW_NAME(field, __stringify(field),		\
-				 field, format_string)
+				 field)
 
-SPI_STATISTICS_SHOW(messages, "%lu");
-SPI_STATISTICS_SHOW(transfers, "%lu");
-SPI_STATISTICS_SHOW(errors, "%lu");
-SPI_STATISTICS_SHOW(timedout, "%lu");
+SPI_STATISTICS_SHOW(messages);
+SPI_STATISTICS_SHOW(transfers);
+SPI_STATISTICS_SHOW(errors);
+SPI_STATISTICS_SHOW(timedout);
 
-SPI_STATISTICS_SHOW(spi_sync, "%lu");
-SPI_STATISTICS_SHOW(spi_sync_immediate, "%lu");
-SPI_STATISTICS_SHOW(spi_async, "%lu");
+SPI_STATISTICS_SHOW(spi_sync);
+SPI_STATISTICS_SHOW(spi_sync_immediate);
+SPI_STATISTICS_SHOW(spi_async);
 
-SPI_STATISTICS_SHOW(bytes, "%llu");
-SPI_STATISTICS_SHOW(bytes_rx, "%llu");
-SPI_STATISTICS_SHOW(bytes_tx, "%llu");
+SPI_STATISTICS_SHOW(bytes);
+SPI_STATISTICS_SHOW(bytes_rx);
+SPI_STATISTICS_SHOW(bytes_tx);
 
 #define SPI_STATISTICS_TRANSFER_BYTES_HISTO(index, number)		\
 	SPI_STATISTICS_SHOW_NAME(transfer_bytes_histo##index,		\
 				 "transfer_bytes_histo_" number,	\
-				 transfer_bytes_histo[index],  "%lu")
+				 transfer_bytes_histo[index])
 SPI_STATISTICS_TRANSFER_BYTES_HISTO(0,  "0-1");
 SPI_STATISTICS_TRANSFER_BYTES_HISTO(1,  "2-3");
 SPI_STATISTICS_TRANSFER_BYTES_HISTO(2,  "4-7");
···
 SPI_STATISTICS_TRANSFER_BYTES_HISTO(15, "32768-65535");
 SPI_STATISTICS_TRANSFER_BYTES_HISTO(16, "65536+");
 
-SPI_STATISTICS_SHOW(transfers_split_maxsize, "%lu");
+SPI_STATISTICS_SHOW(transfers_split_maxsize);
 
 static struct attribute *spi_dev_attrs[] = {
 	&dev_attr_modalias.attr,
···
 	NULL,
 };
 
-static void spi_statistics_add_transfer_stats(struct spi_statistics *stats,
+static void spi_statistics_add_transfer_stats(struct spi_statistics *pcpu_stats,
 					      struct spi_transfer *xfer,
 					      struct spi_controller *ctlr)
 {
-	unsigned long flags;
 	int l2len = min(fls(xfer->len), SPI_STATISTICS_HISTO_SIZE) - 1;
+	struct spi_statistics *stats;
 
 	if (l2len < 0)
 		l2len = 0;
 
-	spin_lock_irqsave(&stats->lock, flags);
+	get_cpu();
+	stats = this_cpu_ptr(pcpu_stats);
+	u64_stats_update_begin(&stats->syncp);
 
-	stats->transfers++;
-	stats->transfer_bytes_histo[l2len]++;
+	u64_stats_inc(&stats->transfers);
+	u64_stats_inc(&stats->transfer_bytes_histo[l2len]);
 
-	stats->bytes += xfer->len;
+	u64_stats_add(&stats->bytes, xfer->len);
 	if ((xfer->tx_buf) &&
 	    (xfer->tx_buf != ctlr->dummy_tx))
-		stats->bytes_tx += xfer->len;
+		u64_stats_add(&stats->bytes_tx, xfer->len);
 	if ((xfer->rx_buf) &&
 	    (xfer->rx_buf != ctlr->dummy_rx))
-		stats->bytes_rx += xfer->len;
+		u64_stats_add(&stats->bytes_rx, xfer->len);
 
-	spin_unlock_irqrestore(&stats->lock, flags);
+	u64_stats_update_end(&stats->syncp);
+	put_cpu();
 }
 
 /*
···
 		return NULL;
 	}
 
+	spi->pcpu_statistics = spi_alloc_pcpu_stats(NULL);
+	if (!spi->pcpu_statistics) {
+		kfree(spi);
+		spi_controller_put(ctlr);
+		return NULL;
+	}
+
 	spi->master = spi->controller = ctlr;
 	spi->dev.parent = &ctlr->dev;
 	spi->dev.bus = &spi_bus_type;
 	spi->dev.release = spidev_release;
 	spi->mode = ctlr->buswidth_override_bits;
-
-	spin_lock_init(&spi->statistics.lock);
 
 	device_initialize(&spi->dev);
 	return spi;
···
 			     struct spi_message *msg,
 			     struct spi_transfer *xfer)
 {
-	struct spi_statistics *statm = &ctlr->statistics;
-	struct spi_statistics *stats = &msg->spi->statistics;
+	struct spi_statistics *statm = ctlr->pcpu_statistics;
+	struct spi_statistics *stats = msg->spi->pcpu_statistics;
 	u32 speed_hz = xfer->speed_hz;
 	unsigned long long ms;
 
···
 		/* Nothing to do here */
 		break;
 	case SPI_DELAY_UNIT_SCK:
-		/* clock cycles need to be obtained from spi_transfer */
+		/* Clock cycles need to be obtained from spi_transfer */
 		if (!xfer)
 			return -EINVAL;
 		/*
···
 	u32 unit = xfer->cs_change_delay.unit;
 	int ret;
 
-	/* return early on "fast" mode - for everything but USECS */
+	/* Return early on "fast" mode - for everything but USECS */
 	if (!delay) {
 		if (unit == SPI_DELAY_UNIT_USECS)
 			_spi_transfer_delay_ns(default_delay_ns);
···
 	struct spi_transfer *xfer;
 	bool keep_cs = false;
 	int ret = 0;
-	struct spi_statistics *statm = &ctlr->statistics;
-	struct spi_statistics *stats = &msg->spi->statistics;
+	struct spi_statistics *statm = ctlr->pcpu_statistics;
+	struct spi_statistics *stats = msg->spi->pcpu_statistics;
 
 	spi_set_cs(msg->spi, true, false);
 
···
 	}
 }
 
-/**
- * __spi_pump_messages - function which processes spi message queue
- * @ctlr: controller to process queue for
- * @in_kthread: true if we are in the context of the message pump thread
- *
- * This function checks if there is any spi message in the queue that
- * needs processing and if so call out to the driver to initialize hardware
- * and transfer each message.
- *
- * Note that it is called both from the kthread itself and also from
- * inside spi_sync(); the queue extraction handling at the top of the
- * function should deal with this safely.
- */
-static void __spi_pump_messages(struct spi_controller *ctlr, bool in_kthread)
+static int __spi_pump_transfer_message(struct spi_controller *ctlr,
+		struct spi_message *msg, bool was_busy)
 {
 	struct spi_transfer *xfer;
-	struct spi_message *msg;
-	bool was_busy = false;
-	unsigned long flags;
 	int ret;
 
-	/* Lock queue */
-	spin_lock_irqsave(&ctlr->queue_lock, flags);
-
-	/* Make sure we are not already running a message */
-	if (ctlr->cur_msg) {
-		spin_unlock_irqrestore(&ctlr->queue_lock, flags);
-		return;
-	}
-
-	/* If another context is idling the device then defer */
-	if (ctlr->idling) {
-		kthread_queue_work(ctlr->kworker, &ctlr->pump_messages);
-		spin_unlock_irqrestore(&ctlr->queue_lock, flags);
-		return;
-	}
-
-	/* Check if the queue is idle */
-	if (list_empty(&ctlr->queue) || !ctlr->running) {
-		if (!ctlr->busy) {
-			spin_unlock_irqrestore(&ctlr->queue_lock, flags);
-			return;
-		}
-
-		/* Defer any non-atomic teardown to the thread */
-		if (!in_kthread) {
-			if (!ctlr->dummy_rx && !ctlr->dummy_tx &&
-			    !ctlr->unprepare_transfer_hardware) {
-				spi_idle_runtime_pm(ctlr);
-				ctlr->busy = false;
-				trace_spi_controller_idle(ctlr);
-			} else {
-				kthread_queue_work(ctlr->kworker,
-						   &ctlr->pump_messages);
-			}
-			spin_unlock_irqrestore(&ctlr->queue_lock, flags);
-			return;
-		}
-
-		ctlr->busy = false;
-		ctlr->idling = true;
-		spin_unlock_irqrestore(&ctlr->queue_lock, flags);
-
-		kfree(ctlr->dummy_rx);
-		ctlr->dummy_rx = NULL;
-		kfree(ctlr->dummy_tx);
-		ctlr->dummy_tx = NULL;
-		if (ctlr->unprepare_transfer_hardware &&
-		    ctlr->unprepare_transfer_hardware(ctlr))
-			dev_err(&ctlr->dev,
-				"failed to unprepare transfer hardware\n");
-		spi_idle_runtime_pm(ctlr);
-		trace_spi_controller_idle(ctlr);
-
-		spin_lock_irqsave(&ctlr->queue_lock, flags);
-		ctlr->idling = false;
-		spin_unlock_irqrestore(&ctlr->queue_lock, flags);
-		return;
-	}
-
-	/* Extract head of queue */
-	msg = list_first_entry(&ctlr->queue, struct spi_message, queue);
-	ctlr->cur_msg = msg;
-
-	list_del_init(&msg->queue);
-	if (ctlr->busy)
-		was_busy = true;
-	else
-		ctlr->busy = true;
-	spin_unlock_irqrestore(&ctlr->queue_lock, flags);
-
-	mutex_lock(&ctlr->io_mutex);
-
 	if (!was_busy && ctlr->auto_runtime_pm) {
-		ret = pm_runtime_resume_and_get(ctlr->dev.parent);
+		ret = pm_runtime_get_sync(ctlr->dev.parent);
 		if (ret < 0) {
+			pm_runtime_put_noidle(ctlr->dev.parent);
 			dev_err(&ctlr->dev, "Failed to power device: %d\n",
 				ret);
-			mutex_unlock(&ctlr->io_mutex);
-			return;
+			return ret;
 		}
 	}
 
···
 			msg->status = ret;
 			spi_finalize_current_message(ctlr);
 
-			mutex_unlock(&ctlr->io_mutex);
-			return;
+			return ret;
 		}
 	}
 
···
 				ret);
 			msg->status = ret;
 			spi_finalize_current_message(ctlr);
-			goto out;
+			return ret;
 		}
-		ctlr->cur_msg_prepared = true;
+		msg->prepared = true;
 	}
 
 	ret = spi_map_msg(ctlr, msg);
 	if (ret) {
 		msg->status = ret;
 		spi_finalize_current_message(ctlr);
-		goto out;
+		return ret;
 	}
 
 	if (!ctlr->ptp_sts_supported && !ctlr->transfer_one) {
···
 		}
 	}
 
+	/*
+	 * Drivers implementation of transfer_one_message() must arrange for
+	 * spi_finalize_current_message() to get called. Most drivers will do
+	 * this in the calling context, but some don't. For those cases, a
+	 * completion is used to guarantee that this function does not return
+	 * until spi_finalize_current_message() is done accessing
+	 * ctlr->cur_msg.
+	 * Use of the following two flags enable to opportunistically skip the
+	 * use of the completion since its use involves expensive spin locks.
+	 * In case of a race with the context that calls
+	 * spi_finalize_current_message() the completion will always be used,
+	 * due to strict ordering of these flags using barriers.
+	 */
+	WRITE_ONCE(ctlr->cur_msg_incomplete, true);
+	WRITE_ONCE(ctlr->cur_msg_need_completion, false);
+	reinit_completion(&ctlr->cur_msg_completion);
+	smp_wmb(); /* Make these available to spi_finalize_current_message() */
+
 	ret = ctlr->transfer_one_message(ctlr, msg);
 	if (ret) {
 		dev_err(&ctlr->dev,
-			"failed to transfer one message from queue: %d\n",
-			ret);
-		goto out;
+			"failed to transfer one message from queue\n");
+		return ret;
 	}
 
-out:
+	WRITE_ONCE(ctlr->cur_msg_need_completion, true);
+	smp_mb(); /* See spi_finalize_current_message()... */
+	if (READ_ONCE(ctlr->cur_msg_incomplete))
+		wait_for_completion(&ctlr->cur_msg_completion);
+
+	return 0;
+}
+
+/**
+ * __spi_pump_messages - function which processes spi message queue
+ * @ctlr: controller to process queue for
+ * @in_kthread: true if we are in the context of the message pump thread
+ *
+ * This function checks if there is any spi message in the queue that
+ * needs processing and if so call out to the driver to initialize hardware
+ * and transfer each message.
+ *
+ * Note that it is called both from the kthread itself and also from
+ * inside spi_sync(); the queue extraction handling at the top of the
+ * function should deal with this safely.
+ */
+static void __spi_pump_messages(struct spi_controller *ctlr, bool in_kthread)
+{
+	struct spi_message *msg;
+	bool was_busy = false;
+	unsigned long flags;
+	int ret;
+
+	/* Take the IO mutex */
+	mutex_lock(&ctlr->io_mutex);
+
+	/* Lock queue */
+	spin_lock_irqsave(&ctlr->queue_lock, flags);
+
+	/* Make sure we are not already running a message */
+	if (ctlr->cur_msg)
+		goto out_unlock;
+
+	/* Check if the queue is idle */
+	if (list_empty(&ctlr->queue) || !ctlr->running) {
+		if (!ctlr->busy)
+			goto out_unlock;
+
+		/* Defer any non-atomic teardown to the thread */
+		if (!in_kthread) {
+			if (!ctlr->dummy_rx && !ctlr->dummy_tx &&
+			    !ctlr->unprepare_transfer_hardware) {
+				spi_idle_runtime_pm(ctlr);
+				ctlr->busy = false;
+				ctlr->queue_empty = true;
+				trace_spi_controller_idle(ctlr);
+			} else {
+				kthread_queue_work(ctlr->kworker,
+						   &ctlr->pump_messages);
+			}
+			goto out_unlock;
+		}
+
+		ctlr->busy = false;
+		spin_unlock_irqrestore(&ctlr->queue_lock, flags);
+
+		kfree(ctlr->dummy_rx);
+		ctlr->dummy_rx = NULL;
+		kfree(ctlr->dummy_tx);
+		ctlr->dummy_tx = NULL;
+		if (ctlr->unprepare_transfer_hardware &&
+		    ctlr->unprepare_transfer_hardware(ctlr))
+			dev_err(&ctlr->dev,
+				"failed to unprepare transfer hardware\n");
+		spi_idle_runtime_pm(ctlr);
+		trace_spi_controller_idle(ctlr);
+
+		spin_lock_irqsave(&ctlr->queue_lock, flags);
+		ctlr->queue_empty = true;
+		goto out_unlock;
+	}
+
+	/* Extract head of queue */
+	msg = list_first_entry(&ctlr->queue, struct spi_message, queue);
+	ctlr->cur_msg = msg;
+
+	list_del_init(&msg->queue);
+	if (ctlr->busy)
+		was_busy = true;
+	else
+		ctlr->busy = true;
+	spin_unlock_irqrestore(&ctlr->queue_lock, flags);
+
+	ret = __spi_pump_transfer_message(ctlr, msg, was_busy);
+	if (!ret)
+		kthread_queue_work(ctlr->kworker, &ctlr->pump_messages);
+
+	ctlr->cur_msg = NULL;
+	ctlr->fallback = false;
+
 	mutex_unlock(&ctlr->io_mutex);
 
 	/* Prod the scheduler in case transfer_one() was busy waiting */
 	if (!ret)
 		cond_resched();
+	return;
+
+out_unlock:
+	spin_unlock_irqrestore(&ctlr->queue_lock, flags);
+	mutex_unlock(&ctlr->io_mutex);
 }
 
 /**
···
 {
 	ctlr->running = false;
 	ctlr->busy = false;
+	ctlr->queue_empty = true;
 
 	ctlr->kworker = kthread_create_worker(0, dev_name(&ctlr->dev));
 	if (IS_ERR(ctlr->kworker)) {
···
 	struct spi_message *next;
 	unsigned long flags;
 
-	/* get a pointer to the next message, if any */
+	/* Get a pointer to the next message, if any */
 	spin_lock_irqsave(&ctlr->queue_lock, flags);
 	next = list_first_entry_or_null(&ctlr->queue, struct spi_message,
 					queue);
···
 {
 	struct spi_transfer *xfer;
 	struct spi_message *mesg;
-	unsigned long flags;
 	int ret;
 
-	spin_lock_irqsave(&ctlr->queue_lock, flags);
 	mesg = ctlr->cur_msg;
-	spin_unlock_irqrestore(&ctlr->queue_lock, flags);
 
 	if (!ctlr->ptp_sts_supported && !ctlr->transfer_one) {
 		list_for_each_entry(xfer, &mesg->transfers, transfer_list) {
···
 	 */
 	spi_res_release(ctlr, mesg);
 
-	if (ctlr->cur_msg_prepared && ctlr->unprepare_message) {
+	if (mesg->prepared && ctlr->unprepare_message) {
 		ret = ctlr->unprepare_message(ctlr, mesg);
 		if (ret) {
 			dev_err(&ctlr->dev, "failed to unprepare message: %d\n",
···
 		}
 	}
 
-	spin_lock_irqsave(&ctlr->queue_lock, flags);
-	ctlr->cur_msg = NULL;
-	ctlr->cur_msg_prepared = false;
-	ctlr->fallback = false;
-	kthread_queue_work(ctlr->kworker, &ctlr->pump_messages);
-	spin_unlock_irqrestore(&ctlr->queue_lock, flags);
+	mesg->prepared = false;
+
+	WRITE_ONCE(ctlr->cur_msg_incomplete, false);
+	smp_mb(); /* See __spi_pump_transfer_message()... */
+	if (READ_ONCE(ctlr->cur_msg_need_completion))
+		complete(&ctlr->cur_msg_completion);
 
 	trace_spi_message_done(mesg);
 
···
 	msg->status = -EINPROGRESS;
 
 	list_add_tail(&msg->queue, &ctlr->queue);
+	ctlr->queue_empty = false;
 	if (!ctlr->busy && need_pump)
 		kthread_queue_work(ctlr->kworker, &ctlr->pump_messages);
 
···
 			if (lookup->index != -1 && lookup->n++ != lookup->index)
 				return 1;
 
-			if (lookup->index == -1 && !ctlr)
-				return -ENODEV;
-
 			status = acpi_get_handle(NULL,
 						 sb->resource_source.string_ptr,
 						 &parent_handle);
···
 
 				ctlr = acpi_spi_find_controller_by_adev(adev);
 				if (!ctlr)
-					return -ENODEV;
+					return -EPROBE_DEFER;
 
 				lookup->ctlr = ctlr;
 			}
···
 	acpi_dev_free_resource_list(&resource_list);
 
 	if (ret < 0)
-		/* found SPI in _CRS but it points to another controller */
-		return ERR_PTR(-ENODEV);
+		/* Found SPI in _CRS but it points to another controller */
+		return ERR_PTR(ret);
 
 	if (!lookup.max_speed_hz &&
 	    ACPI_SUCCESS(acpi_get_parent(adev->handle, &parent_handle)) &&
···
 		return status;
 
 	if (ctlr->bus_num >= 0) {
-		/* devices with a fixed bus num must check-in with the num */
+		/* Devices with a fixed bus num must check-in with the num */
 		mutex_lock(&board_lock);
 		id = idr_alloc(&spi_master_idr, ctlr, ctlr->bus_num,
 			ctlr->bus_num + 1, GFP_KERNEL);
···
 			return id == -ENOSPC ? -EBUSY : id;
 		ctlr->bus_num = id;
 	} else if (ctlr->dev.of_node) {
-		/* allocate dynamic bus number using Linux idr */
+		/* Allocate dynamic bus number using Linux idr */
 		id = of_alias_get_id(ctlr->dev.of_node, "spi");
 		if (id >= 0) {
 			ctlr->bus_num = id;
···
 	}
 	ctlr->bus_lock_flag = 0;
 	init_completion(&ctlr->xfer_completion);
+	init_completion(&ctlr->cur_msg_completion);
 	if (!ctlr->max_dma_len)
 		ctlr->max_dma_len = INT_MAX;
 
···
 		goto free_bus_id;
 	}
 
-	/* setting last_cs to -1 means no chip selected */
+	/* Setting last_cs to -1 means no chip selected */
 	ctlr->last_cs = -1;
 
 	status = device_add(&ctlr->dev);
···
 			goto free_bus_id;
 		}
 	}
-	/* add statistics */
-	spin_lock_init(&ctlr->statistics.lock);
+	/* Add statistics */
+	ctlr->pcpu_statistics = spi_alloc_pcpu_stats(dev);
+	if (!ctlr->pcpu_statistics) {
+		dev_err(dev, "Error allocating per-cpu statistics\n");
+		status = -ENOMEM;
+		goto destroy_queue;
+	}
 
 	mutex_lock(&board_lock);
 	list_add_tail(&ctlr->list, &spi_controller_list);
···
 	acpi_register_spi_devices(ctlr);
 	return status;
 
+destroy_queue:
+	spi_destroy_queue(ctlr);
 free_bus_id:
 	mutex_lock(&board_lock);
 	idr_remove(&spi_master_idr, ctlr->bus_num);
···
 }
 EXPORT_SYMBOL_GPL(spi_register_controller);
 
-static void devm_spi_unregister(void *ctlr)
+static void devm_spi_unregister(struct device *dev, void *res)
 {
-	spi_unregister_controller(ctlr);
+	spi_unregister_controller(*(struct spi_controller **)res);
 }
 
 /**
···
 int devm_spi_register_controller(struct device *dev,
 				 struct spi_controller *ctlr)
 {
+	struct spi_controller **ptr;
 	int ret;
 
-	ret = spi_register_controller(ctlr);
-	if (ret)
-		return ret;
+	ptr = devres_alloc(devm_spi_unregister, sizeof(*ptr), GFP_KERNEL);
+	if (!ptr)
+		return -ENOMEM;
 
-	return devm_add_action_or_reset(dev, devm_spi_unregister, ctlr);
+	ret = spi_register_controller(ctlr);
+	if (!ret) {
+		*ptr = ctlr;
+		devres_add(dev, ptr);
+	} else {
+		devres_free(ptr);
+	}
+
+	return ret;
 }
 EXPORT_SYMBOL_GPL(devm_spi_register_controller);
 
···
 
 	device_del(&ctlr->dev);
 
-	/* free bus id */
+	/* Free bus id */
 	mutex_lock(&board_lock);
 	if (found == ctlr)
 		idr_remove(&spi_master_idr, id);
···
 	struct spi_replaced_transfers *rxfer = res;
 	size_t i;
 
-	/* call extra callback if requested */
+	/* Call extra callback if requested */
 	if (rxfer->release)
 		rxfer->release(ctlr, msg, res);
 
-	/* insert replaced transfers back into the message */
+	/* Insert replaced transfers back into the message */
 	list_splice(&rxfer->replaced_transfers, rxfer->replaced_after);
 
-	/* remove the formerly inserted entries */
+	/* Remove the formerly inserted entries */
 	for (i = 0; i < rxfer->inserted; i++)
 		list_del(&rxfer->inserted_transfers[i].transfer_list);
 }
···
 	struct spi_transfer *xfer;
 	size_t i;
 
-	/* allocate the structure using spi_res */
+	/* Allocate the structure using spi_res */
 	rxfer = spi_res_alloc(msg->spi, __spi_replace_transfers_release,
 			      struct_size(rxfer, inserted_transfers, insert)
 			      + extradatasize,
···
 	if (!rxfer)
 		return ERR_PTR(-ENOMEM);
 
-	/* the release code to invoke before running the generic release */
+	/* The release code to invoke before running the generic release */
 	rxfer->release = release;
 
-	/* assign extradata */
+	/* Assign extradata */
 	if (extradatasize)
 		rxfer->extradata =
 			&rxfer->inserted_transfers[insert];
 
-	/* init the replaced_transfers list */
+	/* Init the replaced_transfers list */
 	INIT_LIST_HEAD(&rxfer->replaced_transfers);
 
 	/*
···
 	 */
 	rxfer->replaced_after = xfer_first->transfer_list.prev;
 
-	/* remove the requested number of transfers */
+	/* Remove the requested number of transfers */
 	for (i = 0; i < remove; i++) {
 		/*
 		 * If the entry after replaced_after it is msg->transfers
···
 		if (rxfer->replaced_after->next == &msg->transfers) {
 			dev_err(&msg->spi->dev,
 				"requested to remove more spi_transfers than are available\n");
-			/* insert replaced transfers back into the message */
+			/* Insert replaced transfers back into the message */
 			list_splice(&rxfer->replaced_transfers,
 				    rxfer->replaced_after);
 
-			/* free the spi_replace_transfer structure */
+			/* Free the spi_replace_transfer structure... */
 			spi_res_free(rxfer);
 
-			/* and return with an error */
+			/* ...and return with an error */
 			return ERR_PTR(-EINVAL);
 		}
 
···
 	 * based on the first transfer to get removed.
 	 */
 	for (i = 0; i < insert; i++) {
-		/* we need to run in reverse order */
+		/* We need to run in reverse order */
 		xfer = &rxfer->inserted_transfers[insert - 1 - i];
 
-		/* copy all spi_transfer data */
+		/* Copy all spi_transfer data */
 		memcpy(xfer, xfer_first, sizeof(*xfer));
 
-		/* add to list */
+		/* Add to list */
 		list_add(&xfer->transfer_list, rxfer->replaced_after);
 
-		/* clear cs_change and delay for all but the last */
+		/* Clear cs_change and delay for all but the last */
 		if (i) {
 			xfer->cs_change = false;
 			xfer->delay.value = 0;
 		}
 	}
 
-	/* set up inserted */
+	/* Set up inserted... */
 	rxfer->inserted = insert;
 
-	/* and register it with spi_res/spi_message */
+	/* ...and register it with spi_res/spi_message */
 	spi_res_add(msg, rxfer);
 
 	return rxfer;
···
 	size_t offset;
 	size_t count, i;
 
-	/* calculate how many we have to replace */
+	/* Calculate how many we have to replace */
 	count = DIV_ROUND_UP(xfer->len, maxsize);
 
-	/* create replacement */
+	/* Create replacement */
 	srt = spi_replace_transfers(msg, xfer, 1, count, NULL, 0, gfp);
 	if (IS_ERR(srt))
 		return PTR_ERR(srt);
···
 	 */
 	xfers[0].len = min_t(size_t, maxsize, xfer[0].len);
 
-	/* all the others need rx_buf/tx_buf also set */
+	/* All the others need rx_buf/tx_buf also set */
 	for (i = 1, offset = maxsize; i < count; offset += maxsize, i++) {
-		/* update rx_buf, tx_buf and dma */
+		/* Update rx_buf, tx_buf and dma */
 		if (xfers[i].rx_buf)
 			xfers[i].rx_buf += offset;
 		if (xfers[i].rx_dma)
···
 		if (xfers[i].tx_dma)
 			xfers[i].tx_dma += offset;
 
-		/* update length */
+		/* Update length */
 		xfers[i].len = min(maxsize, xfers[i].len - offset);
 	}
 
···
 	 */
 	*xferp = &xfers[count - 1];
 
-	/* increment statistics counters */
-	SPI_STATISTICS_INCREMENT_FIELD(&ctlr->statistics,
+	/* Increment statistics counters */
+	SPI_STATISTICS_INCREMENT_FIELD(ctlr->pcpu_statistics,
 				       transfers_split_maxsize);
-	SPI_STATISTICS_INCREMENT_FIELD(&msg->spi->statistics,
+	SPI_STATISTICS_INCREMENT_FIELD(msg->spi->pcpu_statistics,
 				       transfers_split_maxsize);
 
 	return 0;
···
 			return ret;
 
 		list_for_each_entry(xfer, &message->transfers, transfer_list) {
-			/* don't change cs_change on the last entry in the list */
+			/* Don't change cs_change on the last entry in the list */
 			if (list_is_last(&xfer->transfer_list, &message->transfers))
 				break;
 			xfer->cs_change = 1;
···
 				!(spi->mode & SPI_TX_QUAD))
 				return -EINVAL;
 		}
-		/* check transfer rx_nbits */
+		/* Check transfer rx_nbits */
 		if (xfer->rx_buf) {
 			if (spi->mode & SPI_NO_RX)
 				return -EINVAL;
···
 
 	message->spi = spi;
 
-	SPI_STATISTICS_INCREMENT_FIELD(&ctlr->statistics, spi_async);
-	SPI_STATISTICS_INCREMENT_FIELD(&spi->statistics, spi_async);
+	SPI_STATISTICS_INCREMENT_FIELD(ctlr->pcpu_statistics, spi_async);
+	SPI_STATISTICS_INCREMENT_FIELD(spi->pcpu_statistics, spi_async);
 
 	trace_spi_message_submit(message);
 
···
 
 }
 
+static void __spi_transfer_message_noqueue(struct spi_controller *ctlr, struct spi_message *msg)
+{
+	bool was_busy;
+	int ret;
+
+	mutex_lock(&ctlr->io_mutex);
+
+	was_busy = ctlr->busy;
+
+	ctlr->cur_msg = msg;
+	ret = __spi_pump_transfer_message(ctlr, msg, was_busy);
+	if (ret)
+		goto out;
+
+	ctlr->cur_msg = NULL;
+	ctlr->fallback = false;
+
+	if (!was_busy) {
+		kfree(ctlr->dummy_rx);
+		ctlr->dummy_rx = NULL;
+		kfree(ctlr->dummy_tx);
+		ctlr->dummy_tx = NULL;
+		if (ctlr->unprepare_transfer_hardware &&
+		    ctlr->unprepare_transfer_hardware(ctlr))
+			dev_err(&ctlr->dev,
+				"failed to unprepare transfer hardware\n");
+		spi_idle_runtime_pm(ctlr);
+	}
+
+out:
+	mutex_unlock(&ctlr->io_mutex);
+}
+
 /*-------------------------------------------------------------------------*/
 
 /*
···
 	DECLARE_COMPLETION_ONSTACK(done);
 	int status;
 	struct spi_controller *ctlr = spi->controller;
-	unsigned long flags;
 
 	status = __spi_validate(spi, message);
 	if (status != 0)
 		return status;
 
-	message->complete = spi_complete;
-	message->context = &done;
 	message->spi = spi;
 
-	SPI_STATISTICS_INCREMENT_FIELD(&ctlr->statistics, spi_sync);
-	SPI_STATISTICS_INCREMENT_FIELD(&spi->statistics, spi_sync);
+	SPI_STATISTICS_INCREMENT_FIELD(ctlr->pcpu_statistics, spi_sync);
+	SPI_STATISTICS_INCREMENT_FIELD(spi->pcpu_statistics, spi_sync);
 
 	/*
-	 * If we're not using the legacy transfer method then we will
-	 * try to transfer in the calling context so special case.
-	 * This code would be less tricky if we could remove the
-	 * support for driver implemented message queues.
+	 * Checking queue_empty here only guarantees async/sync message
+	 * ordering when coming from the same context. It does not need to
+	 * guard against reentrancy from a different context. The io_mutex
+	 * will catch those cases.
 	 */
-	if (ctlr->transfer == spi_queued_transfer) {
-		spin_lock_irqsave(&ctlr->bus_lock_spinlock, flags);
+	if (READ_ONCE(ctlr->queue_empty)) {
+		message->actual_length = 0;
+		message->status = -EINPROGRESS;
 
 		trace_spi_message_submit(message);
 
-		status = __spi_queued_transfer(spi, message, false);
+		SPI_STATISTICS_INCREMENT_FIELD(ctlr->pcpu_statistics, spi_sync_immediate);
+		SPI_STATISTICS_INCREMENT_FIELD(spi->pcpu_statistics, spi_sync_immediate);
 
-		spin_unlock_irqrestore(&ctlr->bus_lock_spinlock, flags);
-	} else {
-		status = spi_async_locked(spi, message);
+		__spi_transfer_message_noqueue(ctlr, message);
+
+		return message->status;
 	}
 
+	/*
+	 * There are messages in the async queue that could have originated
+	 * from the same context, so we need to preserve ordering.
+	 * Therefor we send the message to the async queue and wait until they
+	 * are completed.
+	 */
+	message->complete = spi_complete;
+	message->context = &done;
+	status = spi_async_locked(spi, message);
 	if (status == 0) {
-		/* Push out the messages in the calling context if we can */
-		if (ctlr->transfer == spi_queued_transfer) {
-			SPI_STATISTICS_INCREMENT_FIELD(&ctlr->statistics,
-						       spi_sync_immediate);
-			SPI_STATISTICS_INCREMENT_FIELD(&spi->statistics,
-						       spi_sync_immediate);
-			__spi_pump_messages(ctlr, false);
-		}
-
 		wait_for_completion(&done);
 		status = message->status;
 	}
 	message->context = NULL;
+
 	return status;
 }
 
···
 	ctlr->bus_lock_flag = 1;
 	spin_unlock_irqrestore(&ctlr->bus_lock_spinlock, flags);
 
-	/* mutex remains locked until spi_bus_unlock is called */
+	/* Mutex remains locked until spi_bus_unlock() is called */
 
 	return 0;
 }
···
 }
 EXPORT_SYMBOL_GPL(spi_bus_unlock);
 
-/* portable code must never pass more than 32 bytes */
+/* Portable code must never pass more than 32 bytes */
 #define	SPI_BUFSIZ	max(32, SMP_CACHE_BYTES)
 
 static u8	*buf;
···
 	x[0].tx_buf = local_buf;
 	x[1].rx_buf = local_buf + n_tx;
 
-	/* do the i/o */
+	/* Do the i/o */
 	status = spi_sync(spi, &message);
 	if (status == 0)
 		memcpy(rxbuf, x[1].rx_buf, n_rx);
···
 /*-------------------------------------------------------------------------*/
 
 #if IS_ENABLED(CONFIG_OF_DYNAMIC)
-/* must call put_device() when done with returned spi_device device */
+/* Must call put_device() when done with returned spi_device device */
 static struct spi_device *of_find_spi_device_by_node(struct device_node *node)
 {
 	struct device *dev = bus_find_device_by_of_node(&spi_bus_type, node);
···
 	return dev ? to_spi_device(dev) : NULL;
 }
 
-/* the spi controllers are not using spi_bus, so we find it with another way */
+/* The spi controllers are not using spi_bus, so we find it with another way */
 static struct spi_controller *of_find_spi_controller_by_node(struct device_node *node)
 {
 	struct device *dev;
···
 	if (!dev)
 		return NULL;
 
-	/* reference got in class_find_device */
+	/* Reference got in class_find_device */
 	return container_of(dev, struct spi_controller, dev);
 }
 
···
 	case OF_RECONFIG_CHANGE_ADD:
 		ctlr = of_find_spi_controller_by_node(rd->dn->parent);
 		if (ctlr == NULL)
-			return NOTIFY_OK;	/* not for us */
+			return NOTIFY_OK;	/* Not for us */
 
 		if (of_node_test_and_set_flag(rd->dn, OF_POPULATED)) {
 			put_device(&ctlr->dev);
···
 		break;
 
 	case OF_RECONFIG_CHANGE_REMOVE:
-		/* already depopulated? */
+		/* Already depopulated? */
 		if (!of_node_check_flag(rd->dn, OF_POPULATED))
 			return NOTIFY_OK;
 
-		/* find our device by node */
+		/* Find our device by node */
 		spi = of_find_spi_device_by_node(rd->dn);
 		if (spi == NULL)
-			return NOTIFY_OK;	/* no? not meant for us */
+			return NOTIFY_OK;	/* No? not meant for us */
 
-		/* unregister takes one ref away */
+		/* Unregister takes one ref away */
 		spi_unregister_device(spi);
 
-		/* and put the reference of the find */
+		/* And put the reference of the find */
 		put_device(&spi->dev);
 		break;
 	}

include/linux/spi/spi.h

···
 
 #include <uapi/linux/spi/spi.h>
 #include <linux/acpi.h>
+#include <linux/u64_stats_sync.h>
 
 struct dma_chan;
 struct software_node;
···
 
 /**
  * struct spi_statistics - statistics for spi transfers
- * @lock:          lock protecting this structure
+ * @syncp:         seqcount to protect members in this struct for per-cpu udate
+ *                 on 32-bit systems
  *
  * @messages:      number of spi-messages handled
  * @transfers:     number of spi_transfers handled
···
  *                 maxsize limit
  */
 struct spi_statistics {
-	spinlock_t		lock; /* lock for the whole structure */
+	struct u64_stats_sync	syncp;
 
-	unsigned long		messages;
-	unsigned long		transfers;
-	unsigned long		errors;
-	unsigned long		timedout;
+	u64_stats_t		messages;
+	u64_stats_t		transfers;
+	u64_stats_t		errors;
+	u64_stats_t		timedout;
 
-	unsigned long		spi_sync;
-	unsigned long		spi_sync_immediate;
-	unsigned long		spi_async;
+	u64_stats_t		spi_sync;
+	u64_stats_t		spi_sync_immediate;
+	u64_stats_t		spi_async;
 
-	unsigned long long	bytes;
-	unsigned long long	bytes_rx;
-	unsigned long long	bytes_tx;
+	u64_stats_t		bytes;
+	u64_stats_t		bytes_rx;
+	u64_stats_t		bytes_tx;
 
 #define SPI_STATISTICS_HISTO_SIZE 17
-	unsigned long transfer_bytes_histo[SPI_STATISTICS_HISTO_SIZE];
+	u64_stats_t	transfer_bytes_histo[SPI_STATISTICS_HISTO_SIZE];
 
-	unsigned long transfers_split_maxsize;
+	u64_stats_t	transfers_split_maxsize;
 };
 
-#define SPI_STATISTICS_ADD_TO_FIELD(stats, field, count)	\
-	do {							\
-		unsigned long flags;				\
-		spin_lock_irqsave(&(stats)->lock, flags);	\
-		(stats)->field += count;			\
-		spin_unlock_irqrestore(&(stats)->lock, flags);	\
+#define SPI_STATISTICS_ADD_TO_FIELD(pcpu_stats, field, count)		\
+	do {								\
+		struct spi_statistics *__lstats;			\
+		get_cpu();						\
+		__lstats = this_cpu_ptr(pcpu_stats);			\
+		u64_stats_update_begin(&__lstats->syncp);		\
+		u64_stats_add(&__lstats->field, count);			\
+		u64_stats_update_end(&__lstats->syncp);			\
+		put_cpu();						\
 	} while (0)
 
-#define SPI_STATISTICS_INCREMENT_FIELD(stats, field)	\
-	SPI_STATISTICS_ADD_TO_FIELD(stats, field, 1)
+#define SPI_STATISTICS_INCREMENT_FIELD(pcpu_stats, field)		\
+	do {								\
+		struct spi_statistics *__lstats;			\
+		get_cpu();						\
+		__lstats = this_cpu_ptr(pcpu_stats);			\
+		u64_stats_update_begin(&__lstats->syncp);		\
+		u64_stats_inc(&__lstats->field);			\
+		u64_stats_update_end(&__lstats->syncp);			\
+		put_cpu();						\
+	} while (0)
 
 /**
  * struct spi_delay - SPI delay information
···
  * @cs_inactive: delay to be introduced by the controller after CS is
  *	deasserted. If @cs_change_delay is used from @spi_transfer, then the
  *	two delays will be added up.
- * @statistics: statistics for the spi_device
+ * @pcpu_statistics: statistics for the spi_device
  *
  * A @spi_device is used to interchange data between an SPI slave
  * (usually a discrete chip) and CPU memory.
···
 struct spi_device {
 	struct device		dev;
 	struct spi_controller	*controller;
-	struct spi_controller	*master;	/* compatibility layer */
+	struct spi_controller	*master;	/* Compatibility layer */
 	u32			max_speed_hz;
 	u8			chip_select;
 	u8			bits_per_word;
 	bool			rt;
-#define SPI_NO_TX	BIT(31)		/* no transmit wire */
-#define SPI_NO_RX	BIT(30)		/* no receive wire */
+#define SPI_NO_TX	BIT(31)		/* No transmit wire */
+#define SPI_NO_RX	BIT(30)		/* No receive wire */
 	/*
 	 * All bits defined above should be covered by SPI_MODE_KERNEL_MASK.
 	 * The SPI_MODE_KERNEL_MASK has the SPI_MODE_USER_MASK counterpart,
···
 	void			*controller_data;
 	char			modalias[SPI_NAME_SIZE];
 	const char		*driver_override;
-	struct gpio_desc	*cs_gpiod;	/* chip select gpio desc */
-	struct spi_delay	word_delay; /* inter-word delay */
+	struct gpio_desc	*cs_gpiod;	/* Chip select gpio desc */
+	struct spi_delay	word_delay; /* Inter-word delay */
 	/* CS delays */
 	struct spi_delay	cs_setup;
 	struct spi_delay	cs_hold;
 	struct spi_delay	cs_inactive;
 
-	/* the statistics */
-	struct spi_statistics	statistics;
+	/* The statistics */
+	struct spi_statistics __percpu	*pcpu_statistics;
 
 	/*
 	 * likely need more hooks for more protocol options affecting how
···
 	return dev ? container_of(dev, struct spi_device, dev) : NULL;
 }
 
-/* most drivers won't need to care about device refcounting */
+/* Most drivers won't need to care about device refcounting */
 static inline struct spi_device *spi_dev_get(struct spi_device *spi)
 {
 	return (spi && get_device(&spi->dev)) ? spi : NULL;
···
 	spi->controller_state = state;
 }
 
-/* device driver data */
+/* Device driver data */
 
 static inline void spi_set_drvdata(struct spi_device *spi, void *data)
 {
···
 
 extern struct spi_device *spi_new_ancillary_device(struct spi_device *spi, u8 chip_select);
 
-/* use a define to avoid include chaining to get THIS_MODULE */
+/* Use a define to avoid include chaining to get THIS_MODULE */
 #define spi_register_driver(driver) \
 	__spi_register_driver(THIS_MODULE, driver)
 
···
  * @pump_messages: work struct for scheduling work to the message pump
  * @queue_lock: spinlock to syncronise access to message queue
  * @queue: message queue
- * @idling: the device is entering idle state
  * @cur_msg: the currently in-flight message
- * @cur_msg_prepared: spi_prepare_message was called for the currently
- *                    in-flight message
+ * @cur_msg_completion: a completion for the current in-flight message
+ * @cur_msg_incomplete: Flag used internally to opportunistically skip
+ *	the @cur_msg_completion. This flag is used to check if the driver has
+ *	already called spi_finalize_current_message().
+ * @cur_msg_need_completion: Flag used internally to opportunistically skip
+ *	the @cur_msg_completion. This flag is used to signal the context that
+ *	is running spi_finalize_current_message() that it needs to complete()
  * @cur_msg_mapped: message has been mapped for DMA
  * @last_cs: the last chip_select that is recorded by set_cs, -1 on non chip
  *           selected
···
  * @max_native_cs: When cs_gpiods is used, and this field is filled in,
  *	spi_register_controller() will validate all native CS (including the
  *	unused native CS) against this value.
- * @statistics: statistics for the spi_controller
+ * @pcpu_statistics: statistics for the spi_controller
  * @dma_tx: DMA transmit channel
  * @dma_rx: DMA receive channel
  * @dummy_rx: dummy receive buffer for full-duplex devices
···
  * @irq_flags: Interrupt enable state during PTP system timestamping
  * @fallback: fallback to pio if dma transfer return failure with
  *	SPI_TRANS_FAIL_NO_START.
+ * @queue_empty: signal green light for opportunistically skipping the queue
+ *	for spi_sync transfers.
  *
  * Each SPI controller can communicate with one or more @spi_device
  * children.  These make a small bus, sharing MOSI, MISO and SCK signals
···
 
 	struct list_head list;
 
-	/* other than negative (== assign one dynamically), bus_num is fully
+	/* Other than negative (== assign one dynamically), bus_num is fully
 	 * board-specific.  usually that simplifies to being SOC-specific.
 	 * example:  one SOC has three SPI controllers, numbered 0..2,
 	 * and one board's schematics might show it using SPI-2.  software
···
 	 */
 	u16			num_chipselect;
 
-	/* some SPI controllers pose alignment requirements on DMAable
+	/* Some SPI controllers pose alignment requirements on DMAable
 	 * buffers; let protocol drivers know about these requirements.
 	 */
 	u16			dma_alignment;
···
 	/* spi_device.mode flags override flags for this controller */
 	u32			buswidth_override_bits;
 
-	/* bitmask of supported bits_per_word for transfers */
+	/* Bitmask of supported bits_per_word for transfers */
 	u32			bits_per_word_mask;
 #define SPI_BPW_MASK(bits) BIT((bits) - 1)
 #define SPI_BPW_RANGE_MASK(min, max) GENMASK((max) - 1, (min) - 1)
 
-	/* limits on transfer speed */
+	/* Limits on transfer speed */
 	u32			min_speed_hz;
 	u32			max_speed_hz;
 
-	/* other constraints relevant to this driver */
+	/* Other constraints relevant to this driver */
 	u16			flags;
-#define SPI_CONTROLLER_HALF_DUPLEX	BIT(0)	/* can't do full duplex */
-#define SPI_CONTROLLER_NO_RX		BIT(1)	/* can't do buffer read */
-#define SPI_CONTROLLER_NO_TX		BIT(2)	/* can't do buffer write */
-#define SPI_CONTROLLER_MUST_RX		BIT(3)	/* requires rx */
-#define SPI_CONTROLLER_MUST_TX		BIT(4)	/* requires tx */
+#define SPI_CONTROLLER_HALF_DUPLEX	BIT(0)	/* Can't do full duplex */
+#define SPI_CONTROLLER_NO_RX		BIT(1)	/* Can't do buffer read */
+#define SPI_CONTROLLER_NO_TX		BIT(2)	/* Can't do buffer write */
+#define SPI_CONTROLLER_MUST_RX		BIT(3)	/* Requires rx */
+#define SPI_CONTROLLER_MUST_TX		BIT(4)	/* Requires tx */
 
 #define SPI_MASTER_GPIO_SS		BIT(5)	/* GPIO CS must select slave */
 
-	/* flag indicating if the allocation of this struct is devres-managed */
+	/* Flag indicating if the allocation of this struct is devres-managed */
 	bool			devm_allocated;
 
-	/* flag indicating this is an SPI slave controller */
+	/* Flag indicating this is an SPI slave controller */
 	bool			slave;
 
 	/*
···
 	/* Used to avoid adding the same CS twice */
 	struct mutex		add_lock;
 
-	/* lock and mutex for SPI bus locking */
+	/* Lock and mutex for SPI bus locking */
 	spinlock_t		bus_lock_spinlock;
 	struct mutex		bus_lock_mutex;
 
-	/* flag indicating that the SPI bus is locked for exclusive use */
+	/* Flag indicating that the SPI bus is locked for exclusive use */
 	bool			bus_lock_flag;
 
 	/* Setup mode and clock, etc (spi driver may call many times).
···
 	 */
 	int (*set_cs_timing)(struct spi_device *spi);
 
-	/* bidirectional bulk transfers
+	/* Bidirectional bulk transfers
 	 *
 	 * + The transfer() method may not sleep; its main role is
 	 *   just to add the message to the queue.
···
 	int			(*transfer)(struct spi_device *spi,
 						struct spi_message *mesg);
 
-	/* called on release() to free memory provided by spi_controller */
+	/* Called on release() to free memory provided by spi_controller */
 	void			(*cleanup)(struct spi_device *spi);
 
 	/*
···
 	spinlock_t			queue_lock;
 	struct list_head		queue;
 	struct spi_message		*cur_msg;
-	bool				idling;
+	struct completion               cur_msg_completion;
+	bool				cur_msg_incomplete;
+	bool				cur_msg_need_completion;
 	bool				busy;
 	bool				running;
 	bool				rt;
 	bool				auto_runtime_pm;
-	bool                            cur_msg_prepared;
 	bool				cur_msg_mapped;
 	char				last_cs;
 	bool				last_cs_mode_high;
···
 	s8			unused_native_cs;
 	s8			max_native_cs;
 
-	/* statistics */
-	struct spi_statistics	statistics;
+	/* Statistics */
+	struct spi_statistics __percpu	*pcpu_statistics;
 
 	/* DMA channels for use with core dmaengine helpers */
 	struct dma_chan		*dma_tx;
 	struct dma_chan		*dma_rx;
 
-	/* dummy data for full duplex devices */
+	/* Dummy data for full duplex devices */
 	void			*dummy_rx;
 	void			*dummy_tx;
 
···
 
 	/* Interrupt enable state during PTP system timestamping */
 	unsigned long		irq_flags;
+
+	/* Flag for enabling opportunistic skipping of the queue in spi_sync */
+	bool			queue_empty;
 };
 
 static inline void *spi_controller_get_devdata(struct spi_controller *ctlr)
···
 			     struct spi_transfer *xfer,
 			     size_t progress, bool irqs_off);
 
-/* the spi driver core manages memory for the spi_controller classdev */
+/* The spi driver core manages memory for the spi_controller classdev */
 extern struct spi_controller *__spi_alloc_controller(struct device *host,
 						unsigned int size, bool slave);
 
···
 struct spi_res {
 	struct list_head        entry;
 	spi_res_release_t       release;
-	unsigned long long      data[]; /* guarantee ull alignment */
+	unsigned long long      data[]; /* Guarantee ull alignment */
 };
 
 /*---------------------------------------------------------------------------*/
···
  * and its transfers, ignore them until its completion callback.
  */
 struct spi_transfer {
-	/* it's ok if tx_buf == rx_buf (right?)
+	/* It's ok if tx_buf == rx_buf (right?)
 	 * for MicroWire, one buffer must be null
 	 * buffers must work with dma_*map_single() calls, unless
 	 *   spi_message.is_dma_mapped reports a pre-existing mapping
···
  * @queue: for use by whichever driver currently owns the message
  * @state: for use by whichever driver currently owns the message
  * @resources: for resource management when the spi message is processed
+ * @prepared: spi_prepare_message was called for the this message
  *
  * A @spi_message is used to execute an atomic sequence of data transfers,
  * each represented by a struct spi_transfer.  The sequence is "atomic"
···
 	 * tell them about such special cases.
 	 */
 
-	/* completion is reported through a callback */
+	/* Completion is reported through a callback */
 	void			(*complete)(void *context);
 	void			*context;
 	unsigned		frame_length;
 	unsigned		actual_length;
 	int			status;
 
-	/* for optional use by whatever driver currently owns the
+	/* For optional use by whatever driver currently owns the
 	 * spi_message ...  between calls to spi_async and then later
 	 * complete(), that's the spi_controller controller driver.
 	 */
 	struct list_head	queue;
 	void			*state;
 
-	/* list of spi_res reources when the spi message is processed */
+	/* List of spi_res reources when the spi message is processed */
 	struct list_head        resources;
+
+	/* spi_prepare_message() was called for this message */
+	bool			prepared;
 };
 
 static inline void spi_message_init_no_memset(struct spi_message *m)
···
 	if (ctlr->max_transfer_size)
 		tr_max = ctlr->max_transfer_size(spi);
 
-	/* transfer size limit must not be greater than messsage size limit */
+	/* Transfer size limit must not be greater than message size limit */
 	return min(tr_max, msg_max);
 }
 
···
 	return spi_sync_transfer(spi, &t, 1);
 }
 
-/* this copies txbuf and rxbuf data; for small transfers only! */
+/* This copies txbuf and rxbuf data; for small transfers only! */
 extern int spi_write_then_read(struct spi_device *spi,
 		const void *txbuf, unsigned n_tx,
 		void *rxbuf, unsigned n_rx);
···
 
 	status = spi_write_then_read(spi, &cmd, 1, &result, 1);
 
-	/* return negative errno or unsigned value */
+	/* Return negative errno or unsigned value */
 	return (status < 0) ? status : result;
 }
 
···
 
 	status = spi_write_then_read(spi, &cmd, 1, &result, 2);
 
-	/* return negative errno or unsigned value */
+	/* Return negative errno or unsigned value */
 	return (status < 0) ? status : result;
 }
 
···
  * are active in some dynamic board configuration models.
  */
 struct spi_board_info {
-	/* the device name and module name are coupled, like platform_bus;
+	/* The device name and module name are coupled, like platform_bus;
 	 * "modalias" is normally the driver name.
 	 *
 	 * platform_data goes to spi_device.dev.platform_data,
···
 	void		*controller_data;
 	int		irq;
 
-	/* slower signaling on noisy or low voltage boards */
+	/* Slower signaling on noisy or low voltage boards */
 	u32		max_speed_hz;
 
 
···
 extern int
 spi_register_board_info(struct spi_board_info const *info, unsigned n);
 #else
-/* board init code may ignore whether SPI is configured or not */
+/* Board init code may ignore whether SPI is configured or not */
 static inline int
 spi_register_board_info(struct spi_board_info const *info, unsigned n)
 	{ return 0; }

tools/spi/spidev_test.c

···
 {
 	int ret = 0;
 	int fd;
+	uint32_t request;
 
 	parse_opts(argc, argv);
 
···
 	/*
 	 * spi mode
 	 */
+	/* WR is make a request to assign 'mode' */
+	request = mode;
 	ret = ioctl(fd, SPI_IOC_WR_MODE32, &mode);
 	if (ret == -1)
 		pabort("can't set spi mode");
 
+	/* RD is read what mode the device actually is in */
 	ret = ioctl(fd, SPI_IOC_RD_MODE32, &mode);
 	if (ret == -1)
 		pabort("can't get spi mode");
+	/* Drivers can reject some mode bits without returning an error.
+	 * Read the current value to identify what mode it is in, and if it
+	 * differs from the requested mode, warn the user.
+	 */
+	if (request != mode)
+		printf("WARNING device does not support requested mode 0x%x\n",
+			request);
 
 	/*
 	 * bits per word