[PATCH 0/9] usb: Add support for Tegra238 and device support for Tegra264

[PATCH 0/9] usb: Add support for Tegra238 and device support for Tegra264

[Jon Hunter]

Add USB device and host support for the Tegra238 device and add
USB device support for the Tegra264 device. USB host support for
Tegra264 has already sent in a separate series [0] and this series
is based on top of that series.

The series adds the necessary device-tree binding updates for these
devices and adds the SoC data to the USB drivers. For the Tegra264 USB
device, the register offsets and default settings differ from previous
generations and so the driver has been updated accordingly.

I have also included one fix for the Tegra XUDC driver which appears to
be a long standing bug in the driver.

[0] https://lore.kernel.org/linux-tegra/20260903074539.613220-1-waynec@nvidia.com/

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
Jon Hunter (9):
      dt-bindings: usb: Add Tegra238 XUSB host controller support
      dt-bindings: usb: tegra-xudc: Add Tegra238 XUDC support
      dt-bindings: usb: tegra-xudc: Add Tegra264 XUDC support
      usb: xhci: tegra: Add support for Tegra238
      usb: gadget: tegra-xudc: Fix CNT65 configuration
      usb: gadget: tegra-xudc: Add support for Tegra238
      usb: gadget: tegra-xudc: Add SSPX register offset to SoC data
      usb: gadget: tegra-xudc: Add default settings to SoC data
      usb: gadget: tegra-xudc: Add support for Tegra264

 .../devicetree/bindings/usb/nvidia,tegra-xudc.yaml |   6 +
 .../bindings/usb/nvidia,tegra234-xusb.yaml         |  20 ++
 drivers/usb/gadget/udc/tegra-xudc.c                | 227 +++++++++++++++------
 drivers/usb/host/xhci-tegra.c                      |  32 +++
 4 files changed, 224 insertions(+), 61 deletions(-)
---
base-commit: e657a464aba873f5eaf3ca821b6e4016674492d6
change-id: 20260918-tegra-usb-58d1d80e4797
prerequisite-message-id: <20260903074539.613220-1-waynec@nvidia.com>
prerequisite-patch-id: 1056b5f21efaa7acac3291cfccda57616492420d
prerequisite-patch-id: adc0a0861d215b39273395d38088f0f972745d6f
prerequisite-patch-id: 07f5cde5e49f7d9e231827ba2382bb371715161b

Best regards,
--  
Jon Hunter <jonathanh@nvidia.com>

[PATCH 1/9] dt-bindings: usb: Add Tegra238 XUSB host controller support

[Jon Hunter]

Add a compatible string for the Tegra238 XUSB host controller. The
Tegra238 XUSB host controller is similar to that found on the Tegra234
device but only supports 3 PHYs compared to the 4 PHYs on the Tegra234.
Hence, the binding is also updated to limit the number of PHYs that can
be specified for the Tegra238 device.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 .../bindings/usb/nvidia,tegra234-xusb.yaml           | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml b/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
index 3f56110be915..9d605db7e77f 100644
--- a/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
+++ b/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
@@ -19,6 +19,7 @@ properties:
   compatible:
     enum:
       - nvidia,tegra234-xusb
+      - nvidia,tegra238-xusb
       - nvidia,tegra264-xusb
 
   reg:
@@ -154,6 +155,25 @@ properties:
 
 allOf:
   - $ref: usb-xhci.yaml
+  - if:
+      properties:
+        compatible:
+          contains:
+            const: nvidia,tegra238-xusb
+    then:
+      properties:
+        phys:
+          maxItems: 6
+        phy-names:
+          maxItems: 6
+          items:
+            enum:
+              - usb2-0
+              - usb2-1
+              - usb2-2
+              - usb3-0
+              - usb3-1
+              - usb3-2
 
   - if:
       properties:

-- 
2.43.0

[PATCH 2/9] dt-bindings: usb: tegra-xudc: Add Tegra238 XUDC support

[Jon Hunter]

Add a compatible string for the Tegra238 XUDC device. The Tegra238 XUDC
is similar to that found on the Tegra234 device but only support 3 PHYs
compared to the 4 PHYs on the Tegra234. The XUDC binding only defines
which of the available PHYs is used by the XUDC controller and the Tegra
XUSB pad controller binding [0] defines the PHYs that are available.

[0] Documentation/devicetree/bindings/phy/nvidia,tegra194-xusb-padctl.yaml

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml b/Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml
index c6e661e8915c..a57ad8815d12 100644
--- a/Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml
+++ b/Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml
@@ -23,6 +23,7 @@ properties:
           - nvidia,tegra186-xudc # For Tegra186
           - nvidia,tegra194-xudc # For Tegra194
           - nvidia,tegra234-xudc # For Tegra234
+          - nvidia,tegra238-xudc # For Tegra238
 
   reg:
     minItems: 2
@@ -152,6 +153,7 @@ allOf:
               - nvidia,tegra186-xudc
               - nvidia,tegra194-xudc
               - nvidia,tegra234-xudc
+              - nvidia,tegra238-xudc
     then:
       properties:
         reg:
@@ -170,6 +172,7 @@ allOf:
             enum:
               - nvidia,tegra194-xudc
               - nvidia,tegra234-xudc
+              - nvidia,tegra238-xudc
     then:
       required:
         - dma-coherent

-- 
2.43.0

[PATCH 3/9] dt-bindings: usb: tegra-xudc: Add Tegra264 XUDC support

[Jon Hunter]

Add a compatible string for the Tegra264 XUDC device. The Tegra264 XUDC
is similar to that found on the Tegra234 device supporting 4 PHYs,
however, the register offsets for the Tegra264 device are different to
previous Tegra devices.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml b/Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml
index a57ad8815d12..672b7ba0d89d 100644
--- a/Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml
+++ b/Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml
@@ -24,6 +24,7 @@ properties:
           - nvidia,tegra194-xudc # For Tegra194
           - nvidia,tegra234-xudc # For Tegra234
           - nvidia,tegra238-xudc # For Tegra238
+          - nvidia,tegra264-xudc # For Tegra264
 
   reg:
     minItems: 2
@@ -154,6 +155,7 @@ allOf:
               - nvidia,tegra194-xudc
               - nvidia,tegra234-xudc
               - nvidia,tegra238-xudc
+              - nvidia,tegra264-xudc
     then:
       properties:
         reg:
@@ -173,6 +175,7 @@ allOf:
               - nvidia,tegra194-xudc
               - nvidia,tegra234-xudc
               - nvidia,tegra238-xudc
+              - nvidia,tegra264-xudc
     then:
       required:
         - dma-coherent

-- 
2.43.0

[PATCH 4/9] usb: xhci: tegra: Add support for Tegra238

[Jon Hunter]

Add XHCI support for the Tegra238 SoC which is similar to the Tegra234
XHCI but only supports 3 USB2/3 ports compared to the 4 USB2/3 ports on
Tegra234.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 drivers/usb/host/xhci-tegra.c | 32 ++++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)

diff --git a/drivers/usb/host/xhci-tegra.c b/drivers/usb/host/xhci-tegra.c
index 50d733b2ad71..78ffe415ead4 100644
--- a/drivers/usb/host/xhci-tegra.c
+++ b/drivers/usb/host/xhci-tegra.c
@@ -2749,6 +2749,37 @@ static const struct tegra_xusb_soc tegra234_soc = {
 	.enable_firmware_messages = true,
 };
 
+static const struct tegra_xusb_phy_type tegra238_phy_types[] = {
+	{ .name = "usb3", .num = 3, },
+	{ .name = "usb2", .num = 3, },
+};
+
+static const struct tegra_xusb_soc tegra238_soc = {
+	.supply_names = tegra194_supply_names,
+	.num_supplies = ARRAY_SIZE(tegra194_supply_names),
+	.phy_types = tegra238_phy_types,
+	.num_types = ARRAY_SIZE(tegra238_phy_types),
+	.context = &tegra186_xusb_context,
+	.ports = {
+		.usb3 = { .offset = 0, .count = 3, },
+		.usb2 = { .offset = 3, .count = 3, },
+	},
+	.scale_ss_clock = false,
+	.has_ipfs = false,
+	.otg_reset_sspi = false,
+	.ops = &tegra234_ops,
+	.mbox = {
+		.cmd = XUSB_BAR2_ARU_MBOX_CMD,
+		.data_in = XUSB_BAR2_ARU_MBOX_DATA_IN,
+		.data_out = XUSB_BAR2_ARU_MBOX_DATA_OUT,
+		.owner = XUSB_BAR2_ARU_MBOX_OWNER,
+		.smi_intr = XUSB_BAR2_ARU_SMI_INTR,
+	},
+	.lpm_support = true,
+	.has_bar2 = true,
+	.enable_firmware_messages = true,
+};
+
 static const struct tegra_xusb_soc tegra264_soc = {
 	.supply_names = tegra194_supply_names,
 	.num_supplies = ARRAY_SIZE(tegra194_supply_names),
@@ -2782,6 +2813,7 @@ static const struct of_device_id tegra_xusb_of_match[] = {
 	{ .compatible = "nvidia,tegra186-xusb", .data = &tegra186_soc },
 	{ .compatible = "nvidia,tegra194-xusb", .data = &tegra194_soc },
 	{ .compatible = "nvidia,tegra234-xusb", .data = &tegra234_soc },
+	{ .compatible = "nvidia,tegra238-xusb", .data = &tegra238_soc },
 	{ .compatible = "nvidia,tegra264-xusb", .data = &tegra264_soc },
 	{ },
 };

-- 
2.43.0

[PATCH 5/9] usb: gadget: tegra-xudc: Fix CNT65 configuration

[Jon Hunter]

When configuring the SSPX_CORE_CNT65 register, the value is being
incorrectly written back to the SSPX_CORE_CNT66 register. Fix this by
ensuring the value intended for SSPX_CORE_CNT65 is written to this
register.

Fixes: 88607a821ffc ("usb: gadget: tegra-xudc: add port_speed_quirk")
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 drivers/usb/gadget/udc/tegra-xudc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/usb/gadget/udc/tegra-xudc.c b/drivers/usb/gadget/udc/tegra-xudc.c
index 5042c13f92be..02002221c1f2 100644
--- a/drivers/usb/gadget/udc/tegra-xudc.c
+++ b/drivers/usb/gadget/udc/tegra-xudc.c
@@ -642,7 +642,7 @@ static void tegra_xudc_limit_port_speed(struct tegra_xudc *xudc)
 	val = xudc_readl(xudc, SSPX_CORE_CNT65);
 	val &= ~(SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID_MASK);
 	val |= SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID(0x4B0);
-	xudc_writel(xudc, val, SSPX_CORE_CNT66);
+	xudc_writel(xudc, val, SSPX_CORE_CNT65);
 
 	val = xudc_readl(xudc, SSPX_CORE_CNT66);
 	val &= ~(SSPX_CORE_CNT66_TX_SCD_BIT0_TRPT_MID_MASK);

-- 
2.43.0

[PATCH 6/9] usb: gadget: tegra-xudc: Add support for Tegra238

[Jon Hunter]

Add XUDC support for the Tegra238 SoC which is similar to Tegra234 XUDC
but only supports 3 PHYs compared to the 4 PHYs on the Tegra234.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 drivers/usb/gadget/udc/tegra-xudc.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/drivers/usb/gadget/udc/tegra-xudc.c b/drivers/usb/gadget/udc/tegra-xudc.c
index 02002221c1f2..0b88c1610f27 100644
--- a/drivers/usb/gadget/udc/tegra-xudc.c
+++ b/drivers/usb/gadget/udc/tegra-xudc.c
@@ -3692,6 +3692,19 @@ static struct tegra_xudc_soc tegra234_xudc_soc_data = {
 	.has_ipfs = false,
 };
 
+static struct tegra_xudc_soc tegra238_xudc_soc_data = {
+	.clock_names = tegra186_xudc_clock_names,
+	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
+	.num_phys = 3,
+	.u1_enable = true,
+	.u2_enable = true,
+	.lpm_enable = true,
+	.invalid_seq_num = false,
+	.pls_quirk = false,
+	.port_reset_quirk = false,
+	.has_ipfs = false,
+};
+
 static const struct of_device_id tegra_xudc_of_match[] = {
 	{
 		.compatible = "nvidia,tegra210-xudc",
@@ -3709,6 +3722,10 @@ static const struct of_device_id tegra_xudc_of_match[] = {
 		.compatible = "nvidia,tegra234-xudc",
 		.data = &tegra234_xudc_soc_data
 	},
+	{
+		.compatible = "nvidia,tegra238-xudc",
+		.data = &tegra238_xudc_soc_data
+	},
 	{ }
 };
 MODULE_DEVICE_TABLE(of, tegra_xudc_of_match);

-- 
2.43.0

[PATCH 7/9] usb: gadget: tegra-xudc: Add SSPX register offset to SoC data

[Jon Hunter]

The SuperSpeed Port (SSPX) register offsets for Tegra264 (0x4000) are
different to the existing Tegra devices (0x600). In preparation for
adding support for Tegra264, add the SSPX register offset into the SoC
data and update the register definitions and register accesses to use
the offset from the SoC data.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 drivers/usb/gadget/udc/tegra-xudc.c | 90 ++++++++++++++++++++-----------------
 1 file changed, 48 insertions(+), 42 deletions(-)

diff --git a/drivers/usb/gadget/udc/tegra-xudc.c b/drivers/usb/gadget/udc/tegra-xudc.c
index 0b88c1610f27..c97cacb7e416 100644
--- a/drivers/usb/gadget/udc/tegra-xudc.c
+++ b/drivers/usb/gadget/udc/tegra-xudc.c
@@ -143,42 +143,42 @@
 #define  HSFSPI_COUNT13_U2_RESUME_K_DURATION(x) ((x) & \
 				HSFSPI_COUNT13_U2_RESUME_K_DURATION_MASK)
 #define BLCG 0x840
-#define SSPX_CORE_CNT0 0x610
+#define SSPX_CORE_CNT0 0x10
 #define  SSPX_CORE_CNT0_PING_TBURST_MASK GENMASK(7, 0)
 #define  SSPX_CORE_CNT0_PING_TBURST(x) ((x) & SSPX_CORE_CNT0_PING_TBURST_MASK)
-#define SSPX_CORE_CNT30 0x688
+#define SSPX_CORE_CNT30 0x88
 #define  SSPX_CORE_CNT30_LMPITP_TIMER_MASK GENMASK(19, 0)
 #define  SSPX_CORE_CNT30_LMPITP_TIMER(x) ((x) & \
 					SSPX_CORE_CNT30_LMPITP_TIMER_MASK)
-#define SSPX_CORE_CNT32 0x690
+#define SSPX_CORE_CNT32 0x90
 #define  SSPX_CORE_CNT32_POLL_TBURST_MAX_MASK GENMASK(7, 0)
 #define  SSPX_CORE_CNT32_POLL_TBURST_MAX(x) ((x) & \
 					SSPX_CORE_CNT32_POLL_TBURST_MAX_MASK)
-#define SSPX_CORE_CNT56 0x6fc
+#define SSPX_CORE_CNT56 0xfc
 #define  SSPX_CORE_CNT56_SCD_BIT0_TRPT_MAX_MASK GENMASK(19, 0)
 #define  SSPX_CORE_CNT56_SCD_BIT0_TRPT_MAX(x) ((x) & \
 				SSPX_CORE_CNT56_SCD_BIT0_TRPT_MAX_MASK)
-#define SSPX_CORE_CNT57 0x700
+#define SSPX_CORE_CNT57 0x100
 #define  SSPX_CORE_CNT57_SCD_BIT1_TRPT_MAX_MASK GENMASK(19, 0)
 #define  SSPX_CORE_CNT57_SCD_BIT1_TRPT_MAX(x) ((x) & \
 				SSPX_CORE_CNT57_SCD_BIT1_TRPT_MAX_MASK)
-#define SSPX_CORE_CNT65 0x720
+#define SSPX_CORE_CNT65 0x120
 #define  SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID_MASK GENMASK(19, 0)
 #define  SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID(x) ((x) & \
 				SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID_MASK)
-#define SSPX_CORE_CNT66 0x724
+#define SSPX_CORE_CNT66 0x124
 #define  SSPX_CORE_CNT66_TX_SCD_BIT0_TRPT_MID_MASK GENMASK(19, 0)
 #define  SSPX_CORE_CNT66_TX_SCD_BIT0_TRPT_MID(x) ((x) & \
 				SSPX_CORE_CNT66_TX_SCD_BIT0_TRPT_MID_MASK)
-#define SSPX_CORE_CNT67 0x728
+#define SSPX_CORE_CNT67 0x128
 #define  SSPX_CORE_CNT67_TX_SCD_BIT1_TRPT_MID_MASK GENMASK(19, 0)
 #define  SSPX_CORE_CNT67_TX_SCD_BIT1_TRPT_MID(x) ((x) & \
 				SSPX_CORE_CNT67_TX_SCD_BIT1_TRPT_MID_MASK)
-#define SSPX_CORE_CNT72 0x73c
+#define SSPX_CORE_CNT72 0x13c
 #define  SSPX_CORE_CNT72_SCD_LFPS_TIMEOUT_MASK GENMASK(19, 0)
 #define  SSPX_CORE_CNT72_SCD_LFPS_TIMEOUT(x) ((x) & \
 				SSPX_CORE_CNT72_SCD_LFPS_TIMEOUT_MASK)
-#define SSPX_CORE_PADCTL4 0x750
+#define SSPX_CORE_PADCTL4 0x150
 #define  SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3_MASK GENMASK(19, 0)
 #define  SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3(x) ((x) & \
 				SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3_MASK)
@@ -548,6 +548,7 @@ struct tegra_xudc_soc {
 	const char * const *clock_names;
 	unsigned int num_clks;
 	unsigned int num_phys;
+	unsigned int sspx_offset;
 	bool u1_enable;
 	bool u2_enable;
 	bool lpm_enable;
@@ -629,35 +630,35 @@ static void tegra_xudc_limit_port_speed(struct tegra_xudc *xudc)
 	u32 val;
 
 	/* limit port speed to gen 1 */
-	val = xudc_readl(xudc, SSPX_CORE_CNT56);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT56);
 	val &= ~(SSPX_CORE_CNT56_SCD_BIT0_TRPT_MAX_MASK);
 	val |= SSPX_CORE_CNT56_SCD_BIT0_TRPT_MAX(0x260);
-	xudc_writel(xudc, val, SSPX_CORE_CNT56);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT56);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT57);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT57);
 	val &= ~(SSPX_CORE_CNT57_SCD_BIT1_TRPT_MAX_MASK);
 	val |= SSPX_CORE_CNT57_SCD_BIT1_TRPT_MAX(0x6D6);
-	xudc_writel(xudc, val, SSPX_CORE_CNT57);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT57);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT65);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT65);
 	val &= ~(SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID_MASK);
 	val |= SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID(0x4B0);
-	xudc_writel(xudc, val, SSPX_CORE_CNT65);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT65);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT66);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT66);
 	val &= ~(SSPX_CORE_CNT66_TX_SCD_BIT0_TRPT_MID_MASK);
 	val |= SSPX_CORE_CNT66_TX_SCD_BIT0_TRPT_MID(0x4B0);
-	xudc_writel(xudc, val, SSPX_CORE_CNT66);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT66);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT67);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT67);
 	val &= ~(SSPX_CORE_CNT67_TX_SCD_BIT1_TRPT_MID_MASK);
 	val |= SSPX_CORE_CNT67_TX_SCD_BIT1_TRPT_MID(0x4B0);
-	xudc_writel(xudc, val, SSPX_CORE_CNT67);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT67);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT72);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT72);
 	val &= ~(SSPX_CORE_CNT72_SCD_LFPS_TIMEOUT_MASK);
 	val |= SSPX_CORE_CNT72_SCD_LFPS_TIMEOUT(0x10);
-	xudc_writel(xudc, val, SSPX_CORE_CNT72);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT72);
 }
 
 static void tegra_xudc_restore_port_speed(struct tegra_xudc *xudc)
@@ -665,35 +666,35 @@ static void tegra_xudc_restore_port_speed(struct tegra_xudc *xudc)
 	u32 val;
 
 	/* restore port speed to gen2 */
-	val = xudc_readl(xudc, SSPX_CORE_CNT56);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT56);
 	val &= ~(SSPX_CORE_CNT56_SCD_BIT0_TRPT_MAX_MASK);
 	val |= SSPX_CORE_CNT56_SCD_BIT0_TRPT_MAX(0x438);
-	xudc_writel(xudc, val, SSPX_CORE_CNT56);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT56);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT57);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT57);
 	val &= ~(SSPX_CORE_CNT57_SCD_BIT1_TRPT_MAX_MASK);
 	val |= SSPX_CORE_CNT57_SCD_BIT1_TRPT_MAX(0x528);
-	xudc_writel(xudc, val, SSPX_CORE_CNT57);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT57);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT65);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT65);
 	val &= ~(SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID_MASK);
 	val |= SSPX_CORE_CNT65_TX_SCD_END_TRPT_MID(0xE10);
-	xudc_writel(xudc, val, SSPX_CORE_CNT66);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT66);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT66);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT66);
 	val &= ~(SSPX_CORE_CNT66_TX_SCD_BIT0_TRPT_MID_MASK);
 	val |= SSPX_CORE_CNT66_TX_SCD_BIT0_TRPT_MID(0x348);
-	xudc_writel(xudc, val, SSPX_CORE_CNT66);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT66);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT67);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT67);
 	val &= ~(SSPX_CORE_CNT67_TX_SCD_BIT1_TRPT_MID_MASK);
 	val |= SSPX_CORE_CNT67_TX_SCD_BIT1_TRPT_MID(0x5a0);
-	xudc_writel(xudc, val, SSPX_CORE_CNT67);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT67);
 
-	val = xudc_readl(xudc, SSPX_CORE_CNT72);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT72);
 	val &= ~(SSPX_CORE_CNT72_SCD_LFPS_TIMEOUT_MASK);
 	val |= SSPX_CORE_CNT72_SCD_LFPS_TIMEOUT(0x1c21);
-	xudc_writel(xudc, val, SSPX_CORE_CNT72);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT72);
 }
 
 static void tegra_xudc_device_mode_on(struct tegra_xudc *xudc)
@@ -3409,22 +3410,22 @@ static void tegra_xudc_device_params_init(struct tegra_xudc *xudc)
 		tegra_xudc_limit_port_speed(xudc);
 
 	/* Set a reasonable U3 exit timer value. */
-	val = xudc_readl(xudc, SSPX_CORE_PADCTL4);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_PADCTL4);
 	val &= ~(SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3_MASK);
 	val |= SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3(0x5dc0);
-	xudc_writel(xudc, val, SSPX_CORE_PADCTL4);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_PADCTL4);
 
 	/* Default ping LFPS tBurst is too large. */
-	val = xudc_readl(xudc, SSPX_CORE_CNT0);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT0);
 	val &= ~(SSPX_CORE_CNT0_PING_TBURST_MASK);
 	val |= SSPX_CORE_CNT0_PING_TBURST(0xa);
-	xudc_writel(xudc, val, SSPX_CORE_CNT0);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT0);
 
 	/* Default tPortConfiguration timeout is too small. */
-	val = xudc_readl(xudc, SSPX_CORE_CNT30);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT30);
 	val &= ~(SSPX_CORE_CNT30_LMPITP_TIMER_MASK);
 	val |= SSPX_CORE_CNT30_LMPITP_TIMER(0x978);
-	xudc_writel(xudc, val, SSPX_CORE_CNT30);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT30);
 
 	if (xudc->soc->lpm_enable) {
 		/* Set L1 resume duration to 95 us. */
@@ -3438,10 +3439,10 @@ static void tegra_xudc_device_params_init(struct tegra_xudc *xudc)
 	 * Compliance suite appears to be violating polling LFPS tBurst max
 	 * of 1.4us.  Send 1.45us instead.
 	 */
-	val = xudc_readl(xudc, SSPX_CORE_CNT32);
+	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT32);
 	val &= ~(SSPX_CORE_CNT32_POLL_TBURST_MAX_MASK);
 	val |= SSPX_CORE_CNT32_POLL_TBURST_MAX(0xb0);
-	xudc_writel(xudc, val, SSPX_CORE_CNT32);
+	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT32);
 
 	/* Direct HS/FS port instance to RxDetect. */
 	val = xudc_readl(xudc, CFG_DEV_FE);
@@ -3641,6 +3642,7 @@ static struct tegra_xudc_soc tegra210_xudc_soc_data = {
 	.clock_names = tegra210_xudc_clock_names,
 	.num_clks = ARRAY_SIZE(tegra210_xudc_clock_names),
 	.num_phys = 4,
+	.sspx_offset = 0x600,
 	.u1_enable = false,
 	.u2_enable = true,
 	.lpm_enable = false,
@@ -3655,6 +3657,7 @@ static struct tegra_xudc_soc tegra186_xudc_soc_data = {
 	.clock_names = tegra186_xudc_clock_names,
 	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
 	.num_phys = 4,
+	.sspx_offset = 0x600,
 	.u1_enable = true,
 	.u2_enable = true,
 	.lpm_enable = false,
@@ -3669,6 +3672,7 @@ static struct tegra_xudc_soc tegra194_xudc_soc_data = {
 	.clock_names = tegra186_xudc_clock_names,
 	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
 	.num_phys = 4,
+	.sspx_offset = 0x600,
 	.u1_enable = true,
 	.u2_enable = true,
 	.lpm_enable = true,
@@ -3683,6 +3687,7 @@ static struct tegra_xudc_soc tegra234_xudc_soc_data = {
 	.clock_names = tegra186_xudc_clock_names,
 	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
 	.num_phys = 4,
+	.sspx_offset = 0x600,
 	.u1_enable = true,
 	.u2_enable = true,
 	.lpm_enable = true,
@@ -3696,6 +3701,7 @@ static struct tegra_xudc_soc tegra238_xudc_soc_data = {
 	.clock_names = tegra186_xudc_clock_names,
 	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
 	.num_phys = 3,
+	.sspx_offset = 0x600,
 	.u1_enable = true,
 	.u2_enable = true,
 	.lpm_enable = true,

-- 
2.43.0

[PATCH 8/9] usb: gadget: tegra-xudc: Add default settings to SoC data

[Jon Hunter]

Some of the default settings that need to be programmed into the
SuperSpeed Port (SSPX) registers differ for Tegra264. In preparation for
adding support for Tegra264 move the necessary register settings into
the SoC data so that different values can be programmed for Tegra264.
Note that for Tegra264 not all of the SSPX registers programmed for
existing Tegra devices are programmed and so only write these registers
if there is a non-zero value in the SoC data.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 drivers/usb/gadget/udc/tegra-xudc.c | 80 ++++++++++++++++++++++++++-----------
 1 file changed, 56 insertions(+), 24 deletions(-)

diff --git a/drivers/usb/gadget/udc/tegra-xudc.c b/drivers/usb/gadget/udc/tegra-xudc.c
index c97cacb7e416..56585c350cab 100644
--- a/drivers/usb/gadget/udc/tegra-xudc.c
+++ b/drivers/usb/gadget/udc/tegra-xudc.c
@@ -549,6 +549,10 @@ struct tegra_xudc_soc {
 	unsigned int num_clks;
 	unsigned int num_phys;
 	unsigned int sspx_offset;
+	unsigned int lfps_ping;
+	unsigned int lfps_poll;
+	unsigned int lmpitp_timer;
+	unsigned int u3_timeout;
 	bool u1_enable;
 	bool u2_enable;
 	bool lpm_enable;
@@ -3391,7 +3395,7 @@ static void tegra_xudc_fpci_ipfs_init(struct tegra_xudc *xudc)
 
 static void tegra_xudc_device_params_init(struct tegra_xudc *xudc)
 {
-	u32 val, imod;
+	u32 val, imod, offset = xudc->soc->sspx_offset;
 
 	val = xudc_readl(xudc, BLCG);
 	if (xudc->soc->has_ipfs) {
@@ -3409,23 +3413,29 @@ static void tegra_xudc_device_params_init(struct tegra_xudc *xudc)
 	if (xudc->soc->port_speed_quirk)
 		tegra_xudc_limit_port_speed(xudc);
 
-	/* Set a reasonable U3 exit timer value. */
-	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_PADCTL4);
-	val &= ~(SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3_MASK);
-	val |= SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3(0x5dc0);
-	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_PADCTL4);
+	if (xudc->soc->u3_timeout) {
+		/* Set U3 exit timer value. */
+		val = xudc_readl(xudc, offset + SSPX_CORE_PADCTL4);
+		val &= ~(SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3_MASK);
+		val |= SSPX_CORE_PADCTL4_RXDAT_VLD_TIMEOUT_U3(xudc->soc->u3_timeout);
+		xudc_writel(xudc, val, offset + SSPX_CORE_PADCTL4);
+	}
 
-	/* Default ping LFPS tBurst is too large. */
-	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT0);
-	val &= ~(SSPX_CORE_CNT0_PING_TBURST_MASK);
-	val |= SSPX_CORE_CNT0_PING_TBURST(0xa);
-	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT0);
+	if (xudc->soc->lfps_ping) {
+		/* Default ping LFPS tBurst is too large. */
+		val = xudc_readl(xudc, offset + SSPX_CORE_CNT0);
+		val &= ~(SSPX_CORE_CNT0_PING_TBURST_MASK);
+		val |= SSPX_CORE_CNT0_PING_TBURST(xudc->soc->lfps_ping);
+		xudc_writel(xudc, val, offset + SSPX_CORE_CNT0);
+	}
 
-	/* Default tPortConfiguration timeout is too small. */
-	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT30);
-	val &= ~(SSPX_CORE_CNT30_LMPITP_TIMER_MASK);
-	val |= SSPX_CORE_CNT30_LMPITP_TIMER(0x978);
-	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT30);
+	if (xudc->soc->lmpitp_timer) {
+		/* Default tPortConfiguration timeout is too small. */
+		val = xudc_readl(xudc, offset + SSPX_CORE_CNT30);
+		val &= ~(SSPX_CORE_CNT30_LMPITP_TIMER_MASK);
+		val |= SSPX_CORE_CNT30_LMPITP_TIMER(xudc->soc->lmpitp_timer);
+		xudc_writel(xudc, val, offset + SSPX_CORE_CNT30);
+	}
 
 	if (xudc->soc->lpm_enable) {
 		/* Set L1 resume duration to 95 us. */
@@ -3435,14 +3445,16 @@ static void tegra_xudc_device_params_init(struct tegra_xudc *xudc)
 		xudc_writel(xudc, val, HSFSPI_COUNT13);
 	}
 
-	/*
-	 * Compliance suite appears to be violating polling LFPS tBurst max
-	 * of 1.4us.  Send 1.45us instead.
-	 */
-	val = xudc_readl(xudc, xudc->soc->sspx_offset + SSPX_CORE_CNT32);
-	val &= ~(SSPX_CORE_CNT32_POLL_TBURST_MAX_MASK);
-	val |= SSPX_CORE_CNT32_POLL_TBURST_MAX(0xb0);
-	xudc_writel(xudc, val, xudc->soc->sspx_offset + SSPX_CORE_CNT32);
+	if (xudc->soc->lfps_poll) {
+		/*
+		 * Compliance suite appears to be violating polling LFPS tBurst
+		 * max of 1.4us.
+		 */
+		val = xudc_readl(xudc, offset + SSPX_CORE_CNT32);
+		val &= ~(SSPX_CORE_CNT32_POLL_TBURST_MAX_MASK);
+		val |= SSPX_CORE_CNT32_POLL_TBURST_MAX(xudc->soc->lfps_poll);
+		xudc_writel(xudc, val, offset + SSPX_CORE_CNT32);
+	}
 
 	/* Direct HS/FS port instance to RxDetect. */
 	val = xudc_readl(xudc, CFG_DEV_FE);
@@ -3643,6 +3655,10 @@ static struct tegra_xudc_soc tegra210_xudc_soc_data = {
 	.num_clks = ARRAY_SIZE(tegra210_xudc_clock_names),
 	.num_phys = 4,
 	.sspx_offset = 0x600,
+	.lfps_ping = 0xa,
+	.lfps_poll = 0xb0, /* 1.45us */
+	.lmpitp_timer = 0x978,
+	.u3_timeout = 0x5dc0,
 	.u1_enable = false,
 	.u2_enable = true,
 	.lpm_enable = false,
@@ -3658,6 +3674,10 @@ static struct tegra_xudc_soc tegra186_xudc_soc_data = {
 	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
 	.num_phys = 4,
 	.sspx_offset = 0x600,
+	.lfps_ping = 0xa,
+	.lfps_poll = 0xb0, /* 1.45us */
+	.lmpitp_timer = 0x978,
+	.u3_timeout = 0x5dc0,
 	.u1_enable = true,
 	.u2_enable = true,
 	.lpm_enable = false,
@@ -3673,6 +3693,10 @@ static struct tegra_xudc_soc tegra194_xudc_soc_data = {
 	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
 	.num_phys = 4,
 	.sspx_offset = 0x600,
+	.lfps_ping = 0xa,
+	.lfps_poll = 0xb0, /* 1.45us */
+	.lmpitp_timer = 0x978,
+	.u3_timeout = 0x5dc0,
 	.u1_enable = true,
 	.u2_enable = true,
 	.lpm_enable = true,
@@ -3688,6 +3712,10 @@ static struct tegra_xudc_soc tegra234_xudc_soc_data = {
 	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
 	.num_phys = 4,
 	.sspx_offset = 0x600,
+	.lfps_ping = 0xa,
+	.lfps_poll = 0xb0, /* 1.45us */
+	.lmpitp_timer = 0x978,
+	.u3_timeout = 0x5dc0,
 	.u1_enable = true,
 	.u2_enable = true,
 	.lpm_enable = true,
@@ -3702,6 +3730,10 @@ static struct tegra_xudc_soc tegra238_xudc_soc_data = {
 	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
 	.num_phys = 3,
 	.sspx_offset = 0x600,
+	.lfps_ping = 0xa,
+	.lfps_poll = 0xb0, /* 1.45us */
+	.lmpitp_timer = 0x978,
+	.u3_timeout = 0x5dc0,
 	.u1_enable = true,
 	.u2_enable = true,
 	.lpm_enable = true,

-- 
2.43.0

[PATCH 9/9] usb: gadget: tegra-xudc: Add support for Tegra264

[Jon Hunter]

Add support for the Tegra264 XUDC by adding the device-tree compatible
string, SoC data and additional registers that need to be configured for
Tegra264. The LMP ITP timer mask in the SSPX_CORE_CNT30 register is
different for Tegra264 and so move the mask into the SoC data.

Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
---
 drivers/usb/gadget/udc/tegra-xudc.c | 60 +++++++++++++++++++++++++++++++++----
 1 file changed, 55 insertions(+), 5 deletions(-)

diff --git a/drivers/usb/gadget/udc/tegra-xudc.c b/drivers/usb/gadget/udc/tegra-xudc.c
index 56585c350cab..cbbdcd00e9b9 100644
--- a/drivers/usb/gadget/udc/tegra-xudc.c
+++ b/drivers/usb/gadget/udc/tegra-xudc.c
@@ -146,10 +146,15 @@
 #define SSPX_CORE_CNT0 0x10
 #define  SSPX_CORE_CNT0_PING_TBURST_MASK GENMASK(7, 0)
 #define  SSPX_CORE_CNT0_PING_TBURST(x) ((x) & SSPX_CORE_CNT0_PING_TBURST_MASK)
+#define SSPX_CORE_CNT13 0x44
+#define  SSPX_CORE_CNT13_CRDTHP_TIMER_MASK GENMASK(11, 0)
+#define  SSPX_CORE_CNT13_CRDTHP_TIMER(x) ((x) & \
+				SSPX_CORE_CNT13_CRDTHP_TIMER_MASK)
+#define SSPX_CORE_CNT27 0x7c
+#define  SSPX_CORE_CNT27_PING_LFPS_TRPT_MASK GENMASK(29, 0)
+#define  SSPX_CORE_CNT27_PING_LFPS_TRPT(x) ((x) & \
+				SSPX_CORE_CNT27_PING_LFPS_TRPT_MASK)
 #define SSPX_CORE_CNT30 0x88
-#define  SSPX_CORE_CNT30_LMPITP_TIMER_MASK GENMASK(19, 0)
-#define  SSPX_CORE_CNT30_LMPITP_TIMER(x) ((x) & \
-					SSPX_CORE_CNT30_LMPITP_TIMER_MASK)
 #define SSPX_CORE_CNT32 0x90
 #define  SSPX_CORE_CNT32_POLL_TBURST_MAX_MASK GENMASK(7, 0)
 #define  SSPX_CORE_CNT32_POLL_TBURST_MAX(x) ((x) & \
@@ -549,9 +554,12 @@ struct tegra_xudc_soc {
 	unsigned int num_clks;
 	unsigned int num_phys;
 	unsigned int sspx_offset;
+	unsigned int crdthp_timer;
 	unsigned int lfps_ping;
+	unsigned int lfps_ping_trpt;
 	unsigned int lfps_poll;
 	unsigned int lmpitp_timer;
+	unsigned int lmpitp_timer_mask;
 	unsigned int u3_timeout;
 	bool u1_enable;
 	bool u2_enable;
@@ -3429,11 +3437,25 @@ static void tegra_xudc_device_params_init(struct tegra_xudc *xudc)
 		xudc_writel(xudc, val, offset + SSPX_CORE_CNT0);
 	}
 
+	if (xudc->soc->crdthp_timer) {
+		val = xudc_readl(xudc, offset + SSPX_CORE_CNT13);
+		val &= ~(SSPX_CORE_CNT13_CRDTHP_TIMER_MASK);
+		val |= SSPX_CORE_CNT13_CRDTHP_TIMER(xudc->soc->crdthp_timer);
+		xudc_writel(xudc, val, offset + SSPX_CORE_CNT13);
+	}
+
+	if (xudc->soc->lfps_ping_trpt) {
+		val = xudc_readl(xudc, offset + SSPX_CORE_CNT27);
+		val &= ~(SSPX_CORE_CNT27_PING_LFPS_TRPT_MASK);
+		val |= SSPX_CORE_CNT27_PING_LFPS_TRPT(xudc->soc->lfps_ping_trpt);
+		xudc_writel(xudc, val, offset + SSPX_CORE_CNT27);
+	}
+
 	if (xudc->soc->lmpitp_timer) {
 		/* Default tPortConfiguration timeout is too small. */
 		val = xudc_readl(xudc, offset + SSPX_CORE_CNT30);
-		val &= ~(SSPX_CORE_CNT30_LMPITP_TIMER_MASK);
-		val |= SSPX_CORE_CNT30_LMPITP_TIMER(xudc->soc->lmpitp_timer);
+		val &= ~(xudc->soc->lmpitp_timer_mask);
+		val |= xudc->soc->lmpitp_timer & xudc->soc->lmpitp_timer_mask;
 		xudc_writel(xudc, val, offset + SSPX_CORE_CNT30);
 	}
 
@@ -3658,6 +3680,7 @@ static struct tegra_xudc_soc tegra210_xudc_soc_data = {
 	.lfps_ping = 0xa,
 	.lfps_poll = 0xb0, /* 1.45us */
 	.lmpitp_timer = 0x978,
+	.lmpitp_timer_mask = GENMASK(19, 0),
 	.u3_timeout = 0x5dc0,
 	.u1_enable = false,
 	.u2_enable = true,
@@ -3677,6 +3700,7 @@ static struct tegra_xudc_soc tegra186_xudc_soc_data = {
 	.lfps_ping = 0xa,
 	.lfps_poll = 0xb0, /* 1.45us */
 	.lmpitp_timer = 0x978,
+	.lmpitp_timer_mask = GENMASK(19, 0),
 	.u3_timeout = 0x5dc0,
 	.u1_enable = true,
 	.u2_enable = true,
@@ -3696,6 +3720,7 @@ static struct tegra_xudc_soc tegra194_xudc_soc_data = {
 	.lfps_ping = 0xa,
 	.lfps_poll = 0xb0, /* 1.45us */
 	.lmpitp_timer = 0x978,
+	.lmpitp_timer_mask = GENMASK(19, 0),
 	.u3_timeout = 0x5dc0,
 	.u1_enable = true,
 	.u2_enable = true,
@@ -3715,6 +3740,7 @@ static struct tegra_xudc_soc tegra234_xudc_soc_data = {
 	.lfps_ping = 0xa,
 	.lfps_poll = 0xb0, /* 1.45us */
 	.lmpitp_timer = 0x978,
+	.lmpitp_timer_mask = GENMASK(19, 0),
 	.u3_timeout = 0x5dc0,
 	.u1_enable = true,
 	.u2_enable = true,
@@ -3733,6 +3759,7 @@ static struct tegra_xudc_soc tegra238_xudc_soc_data = {
 	.lfps_ping = 0xa,
 	.lfps_poll = 0xb0, /* 1.45us */
 	.lmpitp_timer = 0x978,
+	.lmpitp_timer_mask = GENMASK(19, 0),
 	.u3_timeout = 0x5dc0,
 	.u1_enable = true,
 	.u2_enable = true,
@@ -3743,6 +3770,25 @@ static struct tegra_xudc_soc tegra238_xudc_soc_data = {
 	.has_ipfs = false,
 };
 
+static struct tegra_xudc_soc tegra264_xudc_soc_data = {
+	.clock_names = tegra186_xudc_clock_names,
+	.num_clks = ARRAY_SIZE(tegra186_xudc_clock_names),
+	.num_phys = 4,
+	.sspx_offset = 0x4000,
+	.crdthp_timer = 0x33,
+	.lfps_ping_trpt = 0x7a1200,
+	.lfps_poll = 0x3b, /* 1.466us */
+	.lmpitp_timer = 0xc9,
+	.lmpitp_timer_mask = GENMASK(11, 0),
+	.u1_enable = true,
+	.u2_enable = true,
+	.lpm_enable = true,
+	.invalid_seq_num = false,
+	.pls_quirk = false,
+	.port_reset_quirk = false,
+	.has_ipfs = false,
+};
+
 static const struct of_device_id tegra_xudc_of_match[] = {
 	{
 		.compatible = "nvidia,tegra210-xudc",
@@ -3764,6 +3810,10 @@ static const struct of_device_id tegra_xudc_of_match[] = {
 		.compatible = "nvidia,tegra238-xudc",
 		.data = &tegra238_xudc_soc_data
 	},
+	{
+		.compatible = "nvidia,tegra264-xudc",
+		.data = &tegra264_xudc_soc_data
+	},
 	{ }
 };
 MODULE_DEVICE_TABLE(of, tegra_xudc_of_match);

-- 
2.43.0

Re: [PATCH 1/9] dt-bindings: usb: Add Tegra238 XUSB host controller support

[Krzysztof Kozlowski]

On Mon, Sep 21, 2026 at 01:19:01PM +0100, Jon Hunter wrote:
> Add a compatible string for the Tegra238 XUSB host controller. The
> Tegra238 XUSB host controller is similar to that found on the Tegra234
> device but only supports 3 PHYs compared to the 4 PHYs on the Tegra234.
> Hence, the binding is also updated to limit the number of PHYs that can
> be specified for the Tegra238 device.
> 
> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
> ---
>  .../bindings/usb/nvidia,tegra234-xusb.yaml           | 20 ++++++++++++++++++++
>  1 file changed, 20 insertions(+)
> 
> diff --git a/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml b/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
> index 3f56110be915..9d605db7e77f 100644
> --- a/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
> +++ b/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
> @@ -19,6 +19,7 @@ properties:
>    compatible:
>      enum:
>        - nvidia,tegra234-xusb
> +      - nvidia,tegra238-xusb
>        - nvidia,tegra264-xusb
>  
>    reg:
> @@ -154,6 +155,25 @@ properties:
>  
>  allOf:
>    - $ref: usb-xhci.yaml
> +  - if:
> +      properties:
> +        compatible:
> +          contains:
> +            const: nvidia,tegra238-xusb
> +    then:
> +      properties:
> +        phys:

minItems: to be explicit. I would expect minItems: 6, but your phy-names
is flexible, so probably wording in your commit msg should be "*up to* 3
PHYs compared to the...."

Best regards,
Krzysztof

Re: [PATCH 2/9] dt-bindings: usb: tegra-xudc: Add Tegra238 XUDC support

[Krzysztof Kozlowski]

On Mon, Sep 21, 2026 at 01:19:02PM +0100, Jon Hunter wrote:
> Add a compatible string for the Tegra238 XUDC device. The Tegra238 XUDC
> is similar to that found on the Tegra234 device but only support 3 PHYs
> compared to the 4 PHYs on the Tegra234. The XUDC binding only defines
> which of the available PHYs is used by the XUDC controller and the Tegra
> XUSB pad controller binding [0] defines the PHYs that are available.
> 
> [0] Documentation/devicetree/bindings/phy/nvidia,tegra194-xusb-padctl.yaml
> 
> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
> ---
>  Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml | 3 +++
>  1 file changed, 3 insertions(+)

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>

Best regards,
Krzysztof

Re: [PATCH 3/9] dt-bindings: usb: tegra-xudc: Add Tegra264 XUDC support

[Krzysztof Kozlowski]

On Mon, Sep 21, 2026 at 01:19:03PM +0100, Jon Hunter wrote:
> Add a compatible string for the Tegra264 XUDC device. The Tegra264 XUDC
> is similar to that found on the Tegra234 device supporting 4 PHYs,
> however, the register offsets for the Tegra264 device are different to
> previous Tegra devices.
> 
> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
> ---
>  Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml | 3 +++
>  1 file changed, 3 insertions(+)

Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>

Best regards,
Krzysztof

Re: [PATCH 1/9] dt-bindings: usb: Add Tegra238 XUSB host controller support

[Jon Hunter]

On 25/09/2026 14:27, Krzysztof Kozlowski wrote:
> On Mon, Sep 21, 2026 at 01:19:01PM +0100, Jon Hunter wrote:
>> Add a compatible string for the Tegra238 XUSB host controller. The
>> Tegra238 XUSB host controller is similar to that found on the Tegra234
>> device but only supports 3 PHYs compared to the 4 PHYs on the Tegra234.
>> Hence, the binding is also updated to limit the number of PHYs that can
>> be specified for the Tegra238 device.
>>
>> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
>> ---
>>   .../bindings/usb/nvidia,tegra234-xusb.yaml           | 20 ++++++++++++++++++++
>>   1 file changed, 20 insertions(+)
>>
>> diff --git a/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml b/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
>> index 3f56110be915..9d605db7e77f 100644
>> --- a/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
>> +++ b/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
>> @@ -19,6 +19,7 @@ properties:
>>     compatible:
>>       enum:
>>         - nvidia,tegra234-xusb
>> +      - nvidia,tegra238-xusb
>>         - nvidia,tegra264-xusb
>>   
>>     reg:
>> @@ -154,6 +155,25 @@ properties:
>>   
>>   allOf:
>>     - $ref: usb-xhci.yaml
>> +  - if:
>> +      properties:
>> +        compatible:
>> +          contains:
>> +            const: nvidia,tegra238-xusb
>> +    then:
>> +      properties:
>> +        phys:
> 
> minItems: to be explicit. I would expect minItems: 6, but your phy-names
> is flexible, so probably wording in your commit msg should be "*up to* 3
> PHYs compared to the...."
ACK. I will update the commit message to make this clear that this is
'up to 3 PHYs'.

Jon

-- 
nvpublic

Re: [PATCH 1/9] dt-bindings: usb: Add Tegra238 XUSB host controller support

[Jon Hunter]

On 28/09/2026 14:57, Jon Hunter wrote:
> 
> On 25/09/2026 14:27, Krzysztof Kozlowski wrote:
>> On Mon, Sep 21, 2026 at 01:19:01PM +0100, Jon Hunter wrote:
>>> Add a compatible string for the Tegra238 XUSB host controller. The
>>> Tegra238 XUSB host controller is similar to that found on the Tegra234
>>> device but only supports 3 PHYs compared to the 4 PHYs on the Tegra234.
>>> Hence, the binding is also updated to limit the number of PHYs that can
>>> be specified for the Tegra238 device.
>>>
>>> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
>>> ---
>>>   .../bindings/usb/nvidia,tegra234-xusb.yaml           | 20 ++++++++++++++++++++
>>>   1 file changed, 20 insertions(+)
>>>
>>> diff --git a/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml b/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
>>> index 3f56110be915..9d605db7e77f 100644
>>> --- a/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
>>> +++ b/Documentation/devicetree/bindings/usb/nvidia,tegra234-xusb.yaml
>>> @@ -19,6 +19,7 @@ properties:
>>>     compatible:
>>>       enum:
>>>         - nvidia,tegra234-xusb
>>> +      - nvidia,tegra238-xusb
>>>         - nvidia,tegra264-xusb
>>>     reg:
>>> @@ -154,6 +155,25 @@ properties:
>>>   allOf:
>>>     - $ref: usb-xhci.yaml
>>> +  - if:
>>> +      properties:
>>> +        compatible:
>>> +          contains:
>>> +            const: nvidia,tegra238-xusb
>>> +    then:
>>> +      properties:
>>> +        phys:
>>
>> minItems: to be explicit. I would expect minItems: 6, but your phy-names
>> is flexible, so probably wording in your commit msg should be "*up to* 3
>> PHYs compared to the...."
> ACK. I will update the commit message to make this clear that this is
> 'up to 3 PHYs'.

I have noticed that this binding does not have a 'required' section and it
should because for Tegra238 there are no power domains for USB and we only
need to control the resets. I will update the binding to add the required
section and then ensure the correct required properties are specified for
all devices. And of course I will add this detail to the commit message.

Jon

-- 
nvpublic

Re: [PATCH 2/9] dt-bindings: usb: tegra-xudc: Add Tegra238 XUDC support

[Jon Hunter]

On 25/09/2026 14:28, Krzysztof Kozlowski wrote:
> On Mon, Sep 21, 2026 at 01:19:02PM +0100, Jon Hunter wrote:
>> Add a compatible string for the Tegra238 XUDC device. The Tegra238 XUDC
>> is similar to that found on the Tegra234 device but only support 3 PHYs
>> compared to the 4 PHYs on the Tegra234. The XUDC binding only defines
>> which of the available PHYs is used by the XUDC controller and the Tegra
>> XUSB pad controller binding [0] defines the PHYs that are available.
>>
>> [0] Documentation/devicetree/bindings/phy/nvidia,tegra194-xusb-padctl.yaml
>>
>> Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
>> ---
>>   Documentation/devicetree/bindings/usb/nvidia,tegra-xudc.yaml | 3 +++
>>   1 file changed, 3 insertions(+)
> 
> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>

Similarly to my recent comment for the XHCI binding, Tegra238 does not have
power-domains for USB. Unlike the UDC binding this binding has a required
section and so I should have caught this before, but I missed it and that
is my fault. I will correct this.

Jon

-- 
nvpublic