Skip to contentSkip to conversation

Enable IPQ5332 USB2

PATCH Superseded Varadarajan Narayanan 3 participants 14 messages Jul 21, 2023 – Jul 27, 2023

Subsystems (5)

v5 of 12 Superseded Latest is v12 →

lore
Share & tools
lore
#0 lore raw

v5 cover Enable IPQ5332 USB2

This patch series adds the relevant phy and controller configurations for enabling USB2 on IPQ5332

v5:
Binding and dts:-
Fix email id
Removed 'Reviewed-by: Krzysztof Kozlowski <[email protected]>' as had to change bindings file to be able to use generic phy instead of usb-phy

Driver:-
Remove unused definition Use generic phy instead of usb-phy
v4:
Binding and dts:-
Change node name (bindings & dts)
Driver:-
Remove unused enum
static const for '.data' Error handling for devm_clk_get
v3:
Fix bindings file based on review comments

v1:
Cleanup DTS
Combine driver, kconfig and makefile patches Remove unused functions from M31 driver Drop the clock driver changes

Varadarajan Narayanan (6):
dt-bindings: usb: dwc3: Add IPQ5332 compatible
dt-bindings: phy: qcom,m31: Document qcom,m31 USB phy
phy: qcom-m31: Introduce qcom,m31 USB phy driver
arm64: dts: qcom: ipq5332: Add USB related nodes
arm64: dts: qcom: ipq5332: Enable USB
arm64: defconfig: Enable M31 USB phy driver

 .../bindings/phy/qcom,ipq5332-usb-hsphy.yaml       |  51 ++++
 .../devicetree/bindings/usb/qcom,dwc3.yaml         |   3 +
 arch/arm64/boot/dts/qcom/ipq5332-rdp468.dts        |  12 +
 arch/arm64/boot/dts/qcom/ipq5332.dtsi              |  56 +++++
 arch/arm64/configs/defconfig                       |   1 +
 drivers/phy/qualcomm/Kconfig                       |  11 +
 drivers/phy/qualcomm/Makefile                      |   1 +
 drivers/phy/qualcomm/phy-qcom-m31.c                | 264 +++++++++++++++++++++
 8 files changed, 399 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml
 create mode 100644 drivers/phy/qualcomm/phy-qcom-m31.c
Signature · 7 lines
2.7.4


-- 
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy
#1 lore raw

PATCH v5 1/6 dt-bindings: usb: dwc3: Add IPQ5332 compatible

Document the IPQ5332 dwc3 compatible.

---
v1:
Add ipq5332 to interrupts sections

---
 Documentation/devicetree/bindings/usb/qcom,dwc3.yaml | 3 +++
 1 file changed, 3 insertions(+)
Documentation/devicetree/bindings/usb/qcom,dwc3.yaml +3 -0
index ae24dac..9447b54 100644
@@ -14,6 +14,7 @@ properties:
     items:
       - enum:
           - qcom,ipq4019-dwc3
+          - qcom,ipq5332-dwc3
           - qcom,ipq6018-dwc3
           - qcom,ipq8064-dwc3
           - qcom,ipq8074-dwc3
@@ -246,6 +247,7 @@ allOf:
         compatible:
           contains:
             enum:
+              - qcom,ipq5332-dwc3
               - qcom,msm8994-dwc3
               - qcom,qcs404-dwc3
     then:
@@ -410,6 +412,7 @@ allOf:
         compatible:
           contains:
             enum:
+              - qcom,ipq5332-dwc3
               - qcom,sdm660-dwc3
     then:
       properties:
Signature · 7 lines
2.7.4


-- 
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy
#2 lore raw

PATCH v5 2/6 dt-bindings: phy: qcom,m31: Document qcom,m31 USB phy

In reply to Varadarajan Narayanan

Document the M31 USB2 phy present in IPQ5332.

---
v5:
Add '#phy-cells', to be able to use generic phy Remove 'Reviewed-by: Krzysztof Kozlowski' due to above change
v4:
Move M31 URL to description
Remove maxItems and relevant content from clock-names Change node name to generic name
'make dt_binding_check DT_SCHEMA_FILES=qcom' passed
v3:
Incorporate review comments. Will bring in ipq5018 compatible string while posting ipq5018 usb patchset.

v1:
Rename qcom,m31.yaml -> qcom,ipq5332-usb-hsphy.yaml Drop default binding "m31,usb-hsphy"
Add clock
Remove 'oneOf' from compatible
Remove 'qscratch' region from register space as it is not needed Remove reset-names
Fix the example definition

---
 .../bindings/phy/qcom,ipq5332-usb-hsphy.yaml       | 51 ++++++++++++++++++++++
 1 file changed, 51 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +51 -0
new file mode 100644
index 0000000..e0b282b
@@ -0,0 +1,51 @@
+# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/phy/qcom,ipq5332-usb-hsphy.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: M31 USB PHY
+
+maintainers:
+  - Sricharan Ramabadhran <[email protected]>
+  - Varadarajan Narayanan <[email protected]>
+
+description:
+  USB M31 PHY (https://www.m31tech.com) found in Qualcomm
+  IPQ5018, IPQ5332 SoCs.
+
+properties:
+  "#phy-cells":
+    const: 0
+
+  compatible:
+    enum:
+      - qcom,ipq5332-usb-hsphy
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    maxItems: 1
+
+  clock-names:
+    items:
+      - const: cfg_ahb
+
+  resets:
+    maxItems: 1
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/clock/qcom,ipq5332-gcc.h>
+    usbphy0: usb-phy@7b000 {
+        compatible = "qcom,ipq5332-usb-hsphy";
+        reg = <0x0007b000 0x12c>;
+
+        clocks = <&gcc GCC_USB0_PHY_CFG_AHB_CLK>;
+        clock-names = "cfg_ahb";
+
+        resets = <&gcc GCC_QUSB2_0_PHY_BCR>;
+    };
Signature · 7 lines
2.7.4


-- 
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy
#7 lore raw

PATCH v5 3/6 phy: qcom-m31: Introduce qcom,m31 USB phy driver

In reply to Varadarajan Narayanan

Add the M31 USB2 phy driver.

---
v5:
Kconfig and Makefile:- place snippet according to sorted order Use generic phy instead of usb-phy
Use ARRAY_SIZE for reg init instead of blank last entry Fix copyright year

v4:
Remove unused enum
Error handling for devm_clk_get
v1:
Combine driver, makefile and kconfig into 1 patch
Remove 'qscratch' region and its usage. The controller driver takes care of those settings
Use compatible/data to handle ipq5332 init
Drop the default case
Get resources by index instead of name as there is only one resource Add clock
Fix review comments in the driver

---
 drivers/phy/qualcomm/Kconfig        |  11 ++
 drivers/phy/qualcomm/Makefile       |   1 +
 drivers/phy/qualcomm/phy-qcom-m31.c | 264 ++++++++++++++++++++++++++++++++++++
 3 files changed, 276 insertions(+)
 create mode 100644 drivers/phy/qualcomm/phy-qcom-m31.c
drivers/phy/qualcomm/Kconfig +11 -0
index ced6038..d638906 100644
@@ -143,6 +143,17 @@ config PHY_QCOM_EUSB2_REPEATER
 	  PMICs. The repeater is paired with a Synopsys eUSB2 Phy
 	  on Qualcomm SOCs.
 
+config PHY_QCOM_M31_USB
+	tristate "Qualcomm M31 HS PHY driver support"
+	depends on (USB || USB_GADGET) && ARCH_QCOM
+	select USB_PHY
+	help
+	  Enable this to support M31 HS PHY transceivers on Qualcomm chips
+	  with DWC3 USB core. It handles PHY initialization, clock
+	  management required after resetting the hardware and power
+	  management. This driver is required even for peripheral only or
+	  host only mode configurations.
+
 config PHY_QCOM_USB_HS
 	tristate "Qualcomm USB HS PHY module"
 	depends on USB_ULPI_BUS
drivers/phy/qualcomm/Makefile +1 -0
index df94581..ffd609a 100644
@@ -4,6 +4,7 @@ obj-$(CONFIG_PHY_QCOM_APQ8064_SATA)	+= phy-qcom-apq8064-sata.o
 obj-$(CONFIG_PHY_QCOM_EDP)		+= phy-qcom-edp.o
 obj-$(CONFIG_PHY_QCOM_IPQ4019_USB)	+= phy-qcom-ipq4019-usb.o
 obj-$(CONFIG_PHY_QCOM_IPQ806X_SATA)	+= phy-qcom-ipq806x-sata.o
+obj-$(CONFIG_PHY_QCOM_M31_USB)		+= phy-qcom-m31.o
 obj-$(CONFIG_PHY_QCOM_PCIE2)		+= phy-qcom-pcie2.o
 
 obj-$(CONFIG_PHY_QCOM_QMP_COMBO)	+= phy-qcom-qmp-combo.o
drivers/phy/qualcomm/phy-qcom-m31.c new file +264 -0

264 lines not on this page · raw message

Signature · 7 lines
2.7.4


-- 
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy
#9 lore raw

PATCH v5 4/6 arm64: dts: qcom: ipq5332: Add USB related nodes

In reply to Varadarajan Narayanan

Add USB phy and controller nodes.

---
v5:
Use generic phy instead of usb-phy
'make CHECK_DTBS=y DT_SCHEMA_FILES=qcom dtbs_check' passed 'DT_CHECKER_FLAGS='-v -m' DT_SCHEMA_FILES=qcom dt_binding_check' passed
v4:
Change node name
Remove blank line
'make CHECK_DTBS=y DT_SCHEMA_FILES=qcom qcom/ipq5332-rdp441.dtb' passed
v1:
Rename phy node
Change compatible from m31,ipq5332-usb-hsphy -> qcom,ipq5332-usb-hsphy Remove 'qscratch' from phy node
Fix alignment and upper-case hex no.s
Add clock definition for the phy
Remove snps,ref-clock-period-ns as it is not used. dwc3_ref_clk_period() in dwc3/core.c takes the frequency from ref clock and calculates fladj as appropriate.

---
 arch/arm64/boot/dts/qcom/ipq5332.dtsi | 56 +++++++++++++++++++++++++++++++++++
 1 file changed, 56 insertions(+)
arch/arm64/boot/dts/qcom/ipq5332.dtsi +56 -0
index 8bfc2db..a22f340 100644
@@ -405,6 +405,62 @@
 				status = "disabled";
 			};
 		};
+
+		usbphy0: usb-phy@7b000 {
+			compatible = "qcom,ipq5332-usb-hsphy";
+			reg = <0x0007b000 0x12c>;
+
+			clocks = <&gcc GCC_USB0_PHY_CFG_AHB_CLK>;
+			clock-names = "cfg_ahb";
+
+			resets = <&gcc GCC_QUSB2_0_PHY_BCR>;
+
+			#phy-cells = <0>;
+
+			status = "disabled";
+		};
+
+		usb2: usb2@8a00000 {
+			compatible = "qcom,ipq5332-dwc3", "qcom,dwc3";
+			reg = <0x08af8800 0x400>;
+
+			interrupts = <GIC_SPI 62 IRQ_TYPE_LEVEL_HIGH>;
+			interrupt-names = "hs_phy_irq";
+
+			clocks = <&gcc GCC_USB0_MASTER_CLK>,
+				 <&gcc GCC_SNOC_USB_CLK>,
+				 <&gcc GCC_USB0_SLEEP_CLK>,
+				 <&gcc GCC_USB0_MOCK_UTMI_CLK>;
+			clock-names = "core",
+				      "iface",
+				      "sleep",
+				      "mock_utmi";
+
+			resets = <&gcc GCC_USB_BCR>;
+
+			qcom,select-utmi-as-pipe-clk;
+
+			#address-cells = <1>;
+			#size-cells = <1>;
+			ranges;
+
+			status = "disabled";
+
+			usb2_0_dwc: usb@8a00000 {
+				compatible = "snps,dwc3";
+				reg = <0x08a00000 0xe000>;
+				clocks = <&gcc GCC_USB0_MOCK_UTMI_CLK>;
+				clock-names = "ref";
+				interrupts = <GIC_SPI 64 IRQ_TYPE_LEVEL_HIGH>;
+				phy-names = "usb2-phy";
+				phys = <&usbphy0>;
+				tx-fifo-resize;
+				snps,is-utmi-l1-suspend;
+				snps,hird-threshold = /bits/ 8 <0x0>;
+				snps,dis_u2_susphy_quirk;
+				snps,dis_u3_susphy_quirk;
+			};
+		};
 	};
 
 	timer {
Signature · 7 lines
2.7.4


-- 
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy
#11 lore raw

PATCH v5 5/6 arm64: dts: qcom: ipq5332: Enable USB

In reply to Varadarajan Narayanan

Enable USB2 in host mode.

---
v1:
Enable usb-phy node

---
 arch/arm64/boot/dts/qcom/ipq5332-rdp468.dts | 12 ++++++++++++
 1 file changed, 12 insertions(+)
arch/arm64/boot/dts/qcom/ipq5332-rdp468.dts +12 -0
index f96b0c8..f5dea16 100644
@@ -79,3 +79,15 @@
 		bias-pull-up;
 	};
 };
+
+&usbphy0 {
+	status = "okay";
+};
+
+&usb2 {
+	status = "okay";
+};
+
+&usb2_0_dwc {
+	dr_mode = "host";
+};
Signature · 7 lines
2.7.4


-- 
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy
#13 lore raw

PATCH v5 6/6 arm64: defconfig: Enable M31 USB phy driver

In reply to Varadarajan Narayanan

Enable M31 USB phy driver present in IPQ5332.

---
v2:
Add full stop to commit log.

---
 arch/arm64/configs/defconfig | 1 +
 1 file changed, 1 insertion(+)
arch/arm64/configs/defconfig +1 -0
index 6cbf6eb..e6d2012 100644
@@ -1389,6 +1389,7 @@ CONFIG_PHY_QCOM_USB_SNPS_FEMTO_V2=m
 CONFIG_PHY_QCOM_USB_HS_28NM=m
 CONFIG_PHY_QCOM_USB_SS=m
 CONFIG_PHY_QCOM_SGMII_ETH=m
+CONFIG_PHY_QCOM_M31_USB=m
 CONFIG_PHY_R8A779F0_ETHERNET_SERDES=y
 CONFIG_PHY_RCAR_GEN3_PCIE=y
 CONFIG_PHY_RCAR_GEN3_USB2=y
Signature · 7 lines
2.7.4


-- 
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy
maintainer
#8 lore raw

Re: [PATCH v5 3/6] phy: qcom-m31: Introduce qcom,m31 USB phy driver

In reply to Varadarajan Narayanan
Quoted 1 line · Varadarajan Narayanan

Add the M31 USB2 phy driver.

Is it possible to extend this description a little bit? E.g. mention where this PHY is used.

Quoted 33 lines

---
v5:
Kconfig and Makefile:- place snippet according to sorted order Use generic phy instead of usb-phy
Use ARRAY_SIZE for reg init instead of blank last entry Fix copyright year

v4:
Remove unused enum
Error handling for devm_clk_get
v1:
Combine driver, makefile and kconfig into 1 patch
Remove 'qscratch' region and its usage. The controller driver takes care of those settings
Use compatible/data to handle ipq5332 init
Drop the default case
Get resources by index instead of name as there is only one resource Add clock
Fix review comments in the driver

---
 drivers/phy/qualcomm/Kconfig        |  11 ++
 drivers/phy/qualcomm/Makefile       |   1 +
 drivers/phy/qualcomm/phy-qcom-m31.c | 264 ++++++++++++++++++++++++++++++++++++
 3 files changed, 276 insertions(+)
 create mode 100644 drivers/phy/qualcomm/phy-qcom-m31.c
drivers/phy/qualcomm/Kconfig +0 -0
index ced6038..d638906 100644

drivers/phy/qualcomm/Kconfig config PHY_QCOM_EUSB2_REPEATER · line 148

Quoted 6 lines
drivers/phy/qualcomm/Kconfig +3 -0
@@ -143,6 +143,17 @@ config PHY_QCOM_EUSB2_REPEATER
 	  PMICs. The repeater is paired with a Synopsys eUSB2 Phy
 	  on Qualcomm SOCs.
 
+config PHY_QCOM_M31_USB
+	tristate "Qualcomm M31 HS PHY driver support"
+	depends on (USB || USB_GADGET) && ARCH_QCOM

The USB part seems reasonable, but it's not clear to me why you depend on USB_GADGET. And this is the only option in this directory that depends on it's consumer subsystem...

It is reasonable help the user not choose the driver if ARCH_QCOM is disabled, but most other options in this file allow compilation during COMPILE_TEST. This provides additional build coverage, so perhaps you could adopt this as well?

drivers/phy/qualcomm/Kconfig config PHY_QCOM_EUSB2_REPEATER · line 149

Quoted 1 line
drivers/phy/qualcomm/Kconfig +1 -0
+	select USB_PHY

USB_PHY is for drivers in drivers/usb/phy, but you're using the generic PHY framework. Is this really needed, or just a remnant from downstream?

You should however select GENERIC_PHY, like all other options in this file.

Quoted 28 lines
drivers/phy/qualcomm/Kconfig +7 -0
+	help
+	  Enable this to support M31 HS PHY transceivers on Qualcomm chips
+	  with DWC3 USB core. It handles PHY initialization, clock
+	  management required after resetting the hardware and power
+	  management. This driver is required even for peripheral only or
+	  host only mode configurations.
+
 config PHY_QCOM_USB_HS
 	tristate "Qualcomm USB HS PHY module"
 	depends on USB_ULPI_BUS
drivers/phy/qualcomm/Makefile +1 -0
index df94581..ffd609a 100644
@@ -4,6 +4,7 @@ obj-$(CONFIG_PHY_QCOM_APQ8064_SATA)	+= phy-qcom-apq8064-sata.o
 obj-$(CONFIG_PHY_QCOM_EDP)		+= phy-qcom-edp.o
 obj-$(CONFIG_PHY_QCOM_IPQ4019_USB)	+= phy-qcom-ipq4019-usb.o
 obj-$(CONFIG_PHY_QCOM_IPQ806X_SATA)	+= phy-qcom-ipq806x-sata.o
+obj-$(CONFIG_PHY_QCOM_M31_USB)		+= phy-qcom-m31.o
 obj-$(CONFIG_PHY_QCOM_PCIE2)		+= phy-qcom-pcie2.o
 
 obj-$(CONFIG_PHY_QCOM_QMP_COMBO)	+= phy-qcom-qmp-combo.o
drivers/phy/qualcomm/phy-qcom-m31.c new file +0 -0
new file mode 100644
index 0000000..4627425

drivers/phy/qualcomm/phy-qcom-m31.c line 3

Quoted 3 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +3 -0
@@ -0,0 +1,264 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (c) 2014-2016, 2020-2023, The Linux Foundation. All rights reserved.

Please fix this.

drivers/phy/qualcomm/phy-qcom-m31.c line 12

Quoted 9 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +9 -0
+ */
+
+#include <linux/clk.h>
+#include <linux/delay.h>
+#include <linux/err.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/of_device.h>

Rob Herring has posted a bunch of patches lately removing uses of of_device.h, so please don't introduce a new one.

of.h and of_platform.h are to be used instead.

drivers/phy/qualcomm/phy-qcom-m31.c line 18

Quoted 6 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +6 -0
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/reset.h>
+#include <linux/slab.h>
+#include <linux/usb/of.h>
+#include <linux/usb/phy.h>

I don't think you need the last two, in a generic phy driver.

Quoted 56 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +56 -0
+
+#define USB2PHY_PORT_UTMI_CTRL1		0x40
+
+#define USB2PHY_PORT_UTMI_CTRL2		0x44
+ #define UTMI_ULPI_SEL			BIT(7)
+ #define UTMI_TEST_MUX_SEL		BIT(6)
+
+#define HS_PHY_CTRL_REG			0x10
+ #define UTMI_OTG_VBUS_VALID		BIT(20)
+ #define SW_SESSVLD_SEL			BIT(28)
+
+#define USB_PHY_UTMI_CTRL0		0x3c
+
+#define USB_PHY_UTMI_CTRL5		0x50
+ #define POR_EN				BIT(1)
+
+#define USB_PHY_HS_PHY_CTRL_COMMON0	0x54
+ #define COMMONONN			BIT(7)
+ #define FSEL				BIT(4)
+ #define RETENABLEN			BIT(3)
+ #define FREQ_24MHZ			(GENMASK(6, 6) | GENMASK(4, 4))
+
+#define USB_PHY_HS_PHY_CTRL2		0x64
+ #define USB2_SUSPEND_N_SEL		BIT(3)
+ #define USB2_SUSPEND_N			BIT(2)
+ #define USB2_UTMI_CLK_EN		BIT(1)
+
+#define USB_PHY_CFG0			0x94
+ #define UTMI_PHY_OVERRIDE_EN		BIT(1)
+
+#define USB_PHY_REFCLK_CTRL		0xa0
+ #define CLKCORE			BIT(1)
+
+#define USB2PHY_PORT_POWERDOWN		0xa4
+ #define POWER_UP			BIT(0)
+ #define POWER_DOWN			0
+
+#define USB_PHY_FSEL_SEL		0xb8
+ #define FREQ_SEL			BIT(0)
+
+#define USB2PHY_USB_PHY_M31_XCFGI_1	0xbc
+ #define USB2_0_TX_ENABLE		BIT(2)
+
+#define USB2PHY_USB_PHY_M31_XCFGI_4	0xc8
+ #define HSTX_SLEW_RATE_565PS		GENMASK(1, 0)
+ #define PLL_CHARGING_PUMP_CURRENT_35UA	GENMASK(4, 3)
+ #define ODT_VALUE_38_02_OHM		GENMASK(7, 6)
+
+#define USB2PHY_USB_PHY_M31_XCFGI_5	0xcc
+ #define ODT_VALUE_45_02_OHM		BIT(2)
+ #define HSTX_PRE_EMPHASIS_LEVEL_0_55MA	BIT(0)
+
+#define USB2PHY_USB_PHY_M31_XCFGI_11	0xe4
+ #define XCFG_COARSE_TUNE_NUM		BIT(1)
+ #define XCFG_FINE_TUNE_NUM		BIT(3)
+

drivers/phy/qualcomm/phy-qcom-m31.c line 84

Quoted 10 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +10 -0
+struct m31_phy_regs {
+	u32 off;
+	u32 val;
+	u32 delay;
+};
+
+struct m31_priv_data {
+	bool				ulpi_mode;
+	const struct m31_phy_regs	*regs;
+	const int			nregs;

Doesn't hurt to spell out that this is an unsigned int, and you don't need the const on the nregs.

drivers/phy/qualcomm/phy-qcom-m31.c line 88

Quoted 4 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +4 -0
+};
+
+#define M31_REG(__o, __v, __d)	\
+	{ .off = __o, .val = __v, .delay = __d }

This macro doesn't add any value.

Compare:

	M31_REG(USB_PHY_CFG0, UTMI_PHY_OVERRIDE_EN, 0),
with:
	{ USB_PHY_CFG0, UTMI_PHY_OVERRIDE_EN, 0 },

drivers/phy/qualcomm/phy-qcom-m31.c line 95

Quoted 7 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +7 -0
+
+struct m31_phy_regs m31_ipq5332_regs[] = {
+	M31_REG(USB_PHY_CFG0, UTMI_PHY_OVERRIDE_EN, 0),
+	M31_REG(USB_PHY_UTMI_CTRL5, POR_EN, 15),
+	M31_REG(USB_PHY_FSEL_SEL, FREQ_SEL, 0),
+	M31_REG(USB_PHY_HS_PHY_CTRL_COMMON0,
+		COMMONONN | FREQ_24MHZ | RETENABLEN, 0),

The guideline is to keep under 80 characters, but if it improves readability you may use up to 100. So this line can be unwrapped.

drivers/phy/qualcomm/phy-qcom-m31.c line 98

Quoted 3 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +3 -0
+	M31_REG(USB_PHY_UTMI_CTRL5, POR_EN, 0),
+	M31_REG(USB_PHY_HS_PHY_CTRL2,
+		USB2_SUSPEND_N_SEL | USB2_SUSPEND_N | USB2_UTMI_CLK_EN, 0),

This one as well.

drivers/phy/qualcomm/phy-qcom-m31.c line 106

Quoted 8 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +8 -0
+	M31_REG(USB2PHY_USB_PHY_M31_XCFGI_11,
+		XCFG_COARSE_TUNE_NUM  | XCFG_FINE_TUNE_NUM, 0),
+	M31_REG(USB2PHY_USB_PHY_M31_XCFGI_4,
+		HSTX_SLEW_RATE_565PS | PLL_CHARGING_PUMP_CURRENT_35UA |
+		ODT_VALUE_38_02_OHM, 0),
+	M31_REG(USB2PHY_USB_PHY_M31_XCFGI_1, USB2_0_TX_ENABLE, 0),
+	M31_REG(USB2PHY_USB_PHY_M31_XCFGI_5,
+		ODT_VALUE_45_02_OHM | HSTX_PRE_EMPHASIS_LEVEL_0_55MA, 4),

And this one, if you drop the M32_REG macro.

drivers/phy/qualcomm/phy-qcom-m31.c line 117

Quoted 11 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +11 -0
+	M31_REG(USB_PHY_UTMI_CTRL5, 0x0, 0),
+	M31_REG(USB_PHY_HS_PHY_CTRL2, USB2_SUSPEND_N | USB2_UTMI_CLK_EN, 0),
+};
+
+struct m31usb_phy {
+	struct phy			*phy;
+	void __iomem			*base;
+	const struct m31_phy_regs	*regs;
+	int				nregs;
+
+	struct clk			*cfg_ahb_clk;

If you only have one clock, you can name it "clk" without causing confusion.

drivers/phy/qualcomm/phy-qcom-m31.c line 118

Quoted 1 line
drivers/phy/qualcomm/phy-qcom-m31.c new file +1 -0
+	struct reset_control		*phy_reset;

Same here, if there's only one reset, then "reset" is a sufficient name.

drivers/phy/qualcomm/phy-qcom-m31.c line 120

Quoted 2 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +2 -0
+
+	bool				cable_connected;

Unused

drivers/phy/qualcomm/phy-qcom-m31.c line 121

Quoted 1 line
drivers/phy/qualcomm/phy-qcom-m31.c new file +1 -0
+	bool				suspended;

Unused, are you missing/omitting some relevant code here?

drivers/phy/qualcomm/phy-qcom-m31.c line 125

Quoted 4 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +4 -0
+	bool				ulpi_mode;
+};
+
+static void m31usb_phy_enable_clock(struct m31usb_phy *qphy)

It wouldn't hurt to just move this code into m31usb_phy_init().

Quoted 25 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +25 -0
+{
+	const struct m31_phy_regs *regs = qphy->regs;
+	int i;
+
+	for (i = 0; i < qphy->nregs; i++) {
+		writel(regs[i].val, qphy->base + regs[i].off);
+		if (regs[i].delay)
+			udelay(regs[i].delay);
+	}
+}
+
+static int m31usb_phy_init(struct phy *phy)
+{
+	int ret;
+	struct m31usb_phy *qphy = phy_get_drvdata(phy);
+
+	ret = clk_prepare_enable(qphy->cfg_ahb_clk);
+	if (ret) {
+		dev_err(&phy->dev, "failed to enable cfg ahb clock, %d\n", ret);
+		return ret;
+	}
+
+	/* Perform phy reset */
+	reset_control_assert(qphy->phy_reset);
+	udelay(5);

drivers/phy/qualcomm/phy-qcom-m31.c line 160

Quoted 10 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +10 -0
+	reset_control_deassert(qphy->phy_reset);
+
+	/* configure for ULPI mode if requested */
+	if (qphy->ulpi_mode)
+		writel(0x0, qphy->base + USB2PHY_PORT_UTMI_CTRL2);
+
+	/* Enable the PHY */
+	writel(POWER_UP, qphy->base + USB2PHY_PORT_POWERDOWN);
+
+	/* Make sure above write completed */

That's not what wmb() does. It ensures that any writes prior to this line is completed before any writes after this.

But m31usb_phy_enable_clock() is using writel(), which does the same already.

If you really need the previous writes to have completed (get through the write buffer etc), then you need to readl() the same register back. But I see no reason for that.

drivers/phy/qualcomm/phy-qcom-m31.c line 161

Quoted 1 line
drivers/phy/qualcomm/phy-qcom-m31.c new file +1 -0
+	wmb();

So I think this should go.

Quoted 5 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +5 -0
+
+	/* Turn on phy ref clock */
+	m31usb_phy_enable_clock(qphy);
+
+	return 0;

drivers/phy/qualcomm/phy-qcom-m31.c line 176

Quoted 10 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +10 -0
+}
+
+static int m31usb_phy_shutdown(struct phy *phy)
+{
+	struct m31usb_phy *qphy = phy_get_drvdata(phy);
+
+	/* Disable the PHY */
+	writel_relaxed(POWER_DOWN, qphy->base + USB2PHY_PORT_POWERDOWN);
+	/* Make sure above write completed */
+	wmb();

Same as above.

Although here you would assume that clk_disable_unprepare() uses writel() and doesn't overtake your power-down write, but that sounds like reasonable assumptions to make.

Quoted 9 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +9 -0
+
+	clk_disable_unprepare(qphy->cfg_ahb_clk);
+
+	return 0;
+}
+
+static const struct phy_ops m31usb_phy_gen_ops = {
+	.power_on	= m31usb_phy_init,
+	.power_off	= m31usb_phy_shutdown,

drivers/phy/qualcomm/phy-qcom-m31.c line 195

Quoted 10 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +10 -0
+	.owner		= THIS_MODULE,
+};
+
+static int m31usb_phy_probe(struct platform_device *pdev)
+{
+	const struct m31_priv_data *data;
+	struct device *dev = &pdev->dev;
+	struct m31usb_phy *qphy;
+	struct phy_provider *phy_provider;
+	int ret;

Please sort these in descending length order (reverse xmas tree).

Quoted 4 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +4 -0
+
+	qphy = devm_kzalloc(dev, sizeof(*qphy), GFP_KERNEL);
+	if (!qphy)
+		return -ENOMEM;

drivers/phy/qualcomm/phy-qcom-m31.c line 209

Quoted 10 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +10 -0
+
+	qphy->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(qphy->base))
+		return PTR_ERR(qphy->base);
+
+	qphy->phy_reset = devm_reset_control_get_exclusive_by_index(dev, 0);
+	if (IS_ERR(qphy->phy_reset))
+		return PTR_ERR(qphy->phy_reset);
+
+	qphy->cfg_ahb_clk = devm_clk_get(dev, "cfg_ahb");

If you only have one clock, then omitting clock-names and do devm_clk_get(dev, NULL) works well.

drivers/phy/qualcomm/phy-qcom-m31.c line 213

Quoted 4 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +4 -0
+	if (IS_ERR(qphy->cfg_ahb_clk)) {
+		ret = PTR_ERR(qphy->cfg_ahb_clk);
+		dev_err(dev, "failed to get cfg_ahb clock, %d\n", ret);
+		return ret;

Replace these tree lines with:

return dev_err_probe(dev, PTR_ERR(qphy->cfg_ahb_clk), "failed to get cfg_ahb_clk\n");

drivers/phy/qualcomm/phy-qcom-m31.c line 216

Quoted 3 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +3 -0
+	}
+
+	platform_set_drvdata(pdev, qphy);

There are no platform_get_drvdata() calls in this driver, so you can omit this.

drivers/phy/qualcomm/phy-qcom-m31.c line 227

Quoted 11 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +11 -0
+
+	data = of_device_get_match_data(dev);
+	qphy->regs		= data->regs;
+	qphy->nregs		= data->nregs;
+	qphy->ulpi_mode		= data->ulpi_mode;
+
+	qphy->phy = devm_phy_create(dev, NULL, &m31usb_phy_gen_ops);
+	if (IS_ERR(qphy->phy)) {
+		ret = PTR_ERR(qphy->phy);
+		dev_err(dev, "failed to create phy, %d\n", ret);
+		qphy->phy = NULL;

qphy is inaccessible, and freed after the next line, no need to clear this.

drivers/phy/qualcomm/phy-qcom-m31.c line 231

Quoted 4 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +4 -0
+		return ret;
+	}
+
+	dev_set_drvdata(dev, qphy);

This is exactly he same as platform_set_drvdata() above, so per the same logic you can drop this as well.

Quoted 13 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +13 -0
+	phy_set_drvdata(qphy->phy, qphy);
+
+	phy_provider = devm_of_phy_provider_register(dev, of_phy_simple_xlate);
+	if (!IS_ERR(phy_provider))
+		dev_info(dev, "Registered M31 USB phy\n");
+
+	return PTR_ERR_OR_ZERO(phy_provider);
+}
+
+static const struct m31_priv_data m31_ipq5332_data = {
+	.ulpi_mode = false,
+	.regs = m31_ipq5332_regs,
+	.nregs = ARRAY_SIZE(m31_ipq5332_regs),

drivers/phy/qualcomm/phy-qcom-m31.c line 254

Quoted 10 lines
drivers/phy/qualcomm/phy-qcom-m31.c new file +10 -0
+};
+
+static const struct of_device_id m31usb_phy_id_table[] = {
+	{ .compatible = "qcom,ipq5332-usb-hsphy", .data = &m31_ipq5332_data },
+	{ },
+};
+MODULE_DEVICE_TABLE(of, m31usb_phy_id_table);
+
+static struct platform_driver m31usb_phy_driver = {
+	.probe		= m31usb_phy_probe,

Please drop the indentation in these.

Regards,
Bjorn

maintainer
#10 lore raw

Re: [PATCH v5 4/6] arm64: dts: qcom: ipq5332: Add USB related nodes

In reply to Varadarajan Narayanan

arch/arm64/boot/dts/qcom/ipq5332.dtsi line 409

Quoted 10 lines · Varadarajan Narayanan
arch/arm64/boot/dts/qcom/ipq5332.dtsi +2 -0
index 8bfc2db..a22f340 100644
@@ -405,6 +405,62 @@
 				status = "disabled";
 			};
 		};
+
+		usbphy0: usb-phy@7b000 {

s/usb-phy/phy/

arch/arm64/boot/dts/qcom/ipq5332.dtsi line 411

Quoted 2 lines
arch/arm64/boot/dts/qcom/ipq5332.dtsi +2 -0
+			compatible = "qcom,ipq5332-usb-hsphy";
+			reg = <0x0007b000 0x12c>;

Sort your nodes under /soc by their address.

arch/arm64/boot/dts/qcom/ipq5332.dtsi line 423

Quoted 12 lines
arch/arm64/boot/dts/qcom/ipq5332.dtsi +12 -0
+
+			clocks = <&gcc GCC_USB0_PHY_CFG_AHB_CLK>;
+			clock-names = "cfg_ahb";
+
+			resets = <&gcc GCC_QUSB2_0_PHY_BCR>;
+
+			#phy-cells = <0>;
+
+			status = "disabled";
+		};
+
+		usb2: usb2@8a00000 {

usb@... drop the '2' in the node name (not in the label). Although, why does usb2 use the phy labeled usb0?

Regards,
Bjorn

maintainer
#12 lore raw

Re: [PATCH v5 5/6] arm64: dts: qcom: ipq5332: Enable USB

In reply to Varadarajan Narayanan
Quoted 15 lines · Varadarajan Narayanan

Enable USB2 in host mode.

---
v1:
Enable usb-phy node

---
 arch/arm64/boot/dts/qcom/ipq5332-rdp468.dts | 12 ++++++++++++
 1 file changed, 12 insertions(+)
arch/arm64/boot/dts/qcom/ipq5332-rdp468.dts +0 -0
index f96b0c8..f5dea16 100644

arch/arm64/boot/dts/qcom/ipq5332-rdp468.dts line 84

Quoted 6 lines
arch/arm64/boot/dts/qcom/ipq5332-rdp468.dts +3 -0
@@ -79,3 +79,15 @@
 		bias-pull-up;
 	};
 };
+
+&usbphy0 {
+	status = "okay";

Is this phy not powered? Should there be some regulators controlled?

Regards,
Bjorn

maintainer
#3 lore raw

Re: [PATCH v5 2/6] dt-bindings: phy: qcom,m31: Document qcom,m31 USB phy

In reply to Varadarajan Narayanan
Quoted 3 lines · Varadarajan Narayanan

Document the M31 USB2 phy present in IPQ5332.

As Sricharan is the first one to certify the patch's origin, it seems likely that he's the author. Please add a Co-developed-by: Sricharan.

[..]

Quoted 18 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +12 -0
new file mode 100644
index 0000000..e0b282b
@@ -0,0 +1,51 @@
+# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/phy/qcom,ipq5332-usb-hsphy.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: M31 USB PHY
+
+maintainers:
+  - Sricharan Ramabadhran <[email protected]>
+  - Varadarajan Narayanan <[email protected]>
+

Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml line 22

Quoted 10 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +10 -0
+description:
+  USB M31 PHY (https://www.m31tech.com) found in Qualcomm
+  IPQ5018, IPQ5332 SoCs.
+
+properties:
+  "#phy-cells":
+    const: 0
+
+  compatible:
+    enum:

const?

Quoted 11 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +11 -0
+      - qcom,ipq5332-usb-hsphy
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    maxItems: 1
+
+  clock-names:
+    items:
+      - const: cfg_ahb

Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml line 43

Quoted 10 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +10 -0
+
+  resets:
+    maxItems: 1
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/clock/qcom,ipq5332-gcc.h>
+    usbphy0: usb-phy@7b000 {

usb@

And you don't need to give the node a label in the example.

Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml line 49

Quoted 6 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +6 -0
+        compatible = "qcom,ipq5332-usb-hsphy";
+        reg = <0x0007b000 0x12c>;
+
+        clocks = <&gcc GCC_USB0_PHY_CFG_AHB_CLK>;
+        clock-names = "cfg_ahb";
+

No #phy-cells?

Regards,
Bjorn

Quoted 5 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +2 -0
+        resets = <&gcc GCC_QUSB2_0_PHY_BCR>;
+    };
Signature · 1 line
2.7.4
maintainer
#4 lore raw

Re: [PATCH v5 2/6] dt-bindings: phy: qcom,m31: Document qcom,m31 USB phy

Quoted 55 lines · Bjorn Andersson
Quoted 3 lines · Varadarajan Narayanan

Document the M31 USB2 phy present in IPQ5332.

As Sricharan is the first one to certify the patch's origin, it seems likely that he's the author. Please add a Co-developed-by: Sricharan.

[..]

Quoted 18 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +12 -0
new file mode 100644
index 0000000..e0b282b
@@ -0,0 +1,51 @@
+# SPDX-License-Identifier: (GPL-2.0 OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/phy/qcom,ipq5332-usb-hsphy.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: M31 USB PHY
+
+maintainers:
+  - Sricharan Ramabadhran <[email protected]>
+  - Varadarajan Narayanan <[email protected]>
+

Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml line 22

Quoted 10 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +10 -0
+description:
+  USB M31 PHY (https://www.m31tech.com) found in Qualcomm
+  IPQ5018, IPQ5332 SoCs.
+
+properties:
+  "#phy-cells":
+    const: 0
+
+  compatible:
+    enum:

const?

Quoted 15 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +15 -0
+      - qcom,ipq5332-usb-hsphy
+
+  reg:
+    maxItems: 1
+
+  clocks:
+    maxItems: 1
+
+  clock-names:
+    items:
+      - const: cfg_ahb
+
+  resets:
+    maxItems: 1
+
Quoted 8 lines

Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml line 43

Quoted 6 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +6 -0
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/clock/qcom,ipq5332-gcc.h>
+    usbphy0: usb-phy@7b000 {

usb@

You mean phy@? But 'usb2-phy' is accepted too.

Quoted 20 lines

And you don't need to give the node a label in the example.

Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml line 49

Quoted 6 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +6 -0
+        compatible = "qcom,ipq5332-usb-hsphy";
+        reg = <0x0007b000 0x12c>;
+
+        clocks = <&gcc GCC_USB0_PHY_CFG_AHB_CLK>;
+        clock-names = "cfg_ahb";
+

No #phy-cells?

Regards,
Bjorn

Quoted 5 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml new file +2 -0
+        resets = <&gcc GCC_QUSB2_0_PHY_BCR>;
+    };
Signature · 1 line
2.7.4
Signature · 3 lines
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy
maintainer
#5 lore raw

Re: [PATCH v5 2/6] dt-bindings: phy: qcom,m31: Document qcom,m31 USB phy

Quoted 4 lines · Rob Herring
Quoted 3 lines · Bjorn Andersson

On Fri, Jul 21, 2023 at 10:05:27AM +0530, Varadarajan Narayanan wrote:
[..]

Quoted 1 line
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml +0 -0

[..]

Quoted 9 lines
Quoted 6 lines

Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml line 43

Quoted 4 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml +4 -0
+examples:
+  - |
+    #include <dt-bindings/clock/qcom,ipq5332-gcc.h>
+    usbphy0: usb-phy@7b000 {

usb@

You mean phy@? But 'usb2-phy' is accepted too.

Yes, had the controller in mind, sorry about that.

Thanks,
Bjorn

#6 lore raw

Re: [PATCH v5 2/6] dt-bindings: phy: qcom,m31: Document qcom,m31 USB phy

Bjorn & Rob,

Quoted 11 lines · Bjorn Andersson
Quoted 4 lines · Rob Herring
Quoted 3 lines · Bjorn Andersson

On Fri, Jul 21, 2023 at 10:05:27AM +0530, Varadarajan Narayanan wrote:
[..]

Quoted 1 line
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml +0 -0

[..]

Quoted 5 lines
Quoted 5 lines
Quoted 4 lines
Documentation/devicetree/bindings/phy/qcom,ipq5332-usb-hsphy.yaml +4 -0
+examples:
+  - |
+    #include <dt-bindings/clock/qcom,ipq5332-gcc.h>
+    usbphy0: usb-phy@7b000 {
Quoted 6 lines
Quoted 4 lines
Quoted 1 line

usb@

You mean phy@? But 'usb2-phy' is accepted too.

Yes, had the controller in mind, sorry about that.

Have posted a new patchset addressing the comments. Please provide your feedback.

Thanks
Varada

Signature · 3 lines
linux-phy mailing list
[email protected]
https://lists.infradead.org/mailman/listinfo/linux-phy