[PATCH v5 net-next 00/13] add basic support for i.MX95 NETC

[PATCH v5 net-next 00/13] add basic support for i.MX95 NETC

[Wei Fang]

This is first time that the NETC IP is applied on i.MX MPU platform.
Its revision has been upgraded to 4.1, which is very different from
the NETC of LS1028A (its revision is 1.0). Therefore, some existing
drivers of NETC devices in the Linux kernel are not compatible with
the current hardware. For example, the fsl-enetc driver is used to
drive the ENETC PF of LS1028A, but for i.MX95 ENETC PF, its registers
and tables configuration are very different from those of LS1028A,
and only the station interface (SI) part remains basically the same.
For the SI part, Vladimir has separated the fsl-enetc-core driver, so
we can reuse this driver on i.MX95. However, for other parts of PF,
the fsl-enetc driver cannot be reused, so the nxp-enetc4 driver is
added to support revision 4.1 and later.

During the development process, we found that the two PF drivers have
some interfaces with basically the same logic, and the only difference
is the hardware configuration. So in order to reuse these interfaces
and reduce code redundancy, we extracted these interfaces and compiled
them into a separate nxp-enetc-pf-common driver for use by the two PF
drivers.

In addition, we have developed the nxp-netc-blk-ctrl driver, which
is used to control three blocks, namely Integrated Endpoint Register
Block (IERB), Privileged Register Block (PRB) and NETCMIX block. The
IERB contains registers that are used for pre-boot initialization,
debug, and non-customer configuration. The PRB controls global reset
and global error handling for NETC. The NETCMIX block is mainly used
to set MII protocol and PCS protocol of the links, it also contains
settings for some other functions.

---
v1 Link: https://lore.kernel.org/imx/20241009095116.147412-1-wei.fang@nxp.com/
v2 Link: https://lore.kernel.org/imx/20241015125841.1075560-1-wei.fang@nxp.com/
v3 Link: https://lore.kernel.org/imx/20241017074637.1265584-1-wei.fang@nxp.com/
v4 Link: https://lore.kernel.org/imx/20241022055223.382277-1-wei.fang@nxp.com/
---

Clark Wang (2):
  net: enetc: extract enetc_int_vector_init/destroy() from
    enetc_alloc_msix()
  net: enetc: optimize the allocation of tx_bdr

Vladimir Oltean (1):
  net: enetc: remove ERR050089 workaround for i.MX95

Wei Fang (10):
  dt-bindings: net: add compatible string for i.MX95 EMDIO
  dt-bindings: net: add i.MX95 ENETC support
  dt-bindings: net: add bindings for NETC blocks control
  net: enetc: add initial netc-blk-ctrl driver support
  net: enetc: extract common ENETC PF parts for LS1028A and i.MX95
    platforms
  net: enetc: build enetc_pf_common.c as a separate module
  PCI: Add NXP NETC vendor ID and device IDs
  net: enetc: add i.MX95 EMDIO support
  net: enetc: add preliminary support for i.MX95 ENETC PF
  MAINTAINERS: update ENETC driver files and maintainers

 .../bindings/net/fsl,enetc-mdio.yaml          |  11 +-
 .../devicetree/bindings/net/fsl,enetc.yaml    |  34 +-
 .../bindings/net/nxp,netc-blk-ctrl.yaml       | 104 +++
 MAINTAINERS                                   |   7 +
 drivers/net/ethernet/freescale/enetc/Kconfig  |  40 +
 drivers/net/ethernet/freescale/enetc/Makefile |   9 +
 drivers/net/ethernet/freescale/enetc/enetc.c  | 261 +++---
 drivers/net/ethernet/freescale/enetc/enetc.h  |  30 +-
 .../net/ethernet/freescale/enetc/enetc4_hw.h  | 152 ++++
 .../net/ethernet/freescale/enetc/enetc4_pf.c  | 759 ++++++++++++++++++
 .../ethernet/freescale/enetc/enetc_ethtool.c  |  35 +-
 .../net/ethernet/freescale/enetc/enetc_hw.h   |  53 +-
 .../ethernet/freescale/enetc/enetc_pci_mdio.c |  21 +
 .../net/ethernet/freescale/enetc/enetc_pf.c   | 310 +------
 .../net/ethernet/freescale/enetc/enetc_pf.h   |  42 +
 .../freescale/enetc/enetc_pf_common.c         | 338 ++++++++
 .../net/ethernet/freescale/enetc/enetc_qos.c  |   2 +-
 .../net/ethernet/freescale/enetc/enetc_vf.c   |   6 +
 .../ethernet/freescale/enetc/netc_blk_ctrl.c  | 438 ++++++++++
 include/linux/fsl/netc_global.h               |  19 +
 include/linux/pci_ids.h                       |   7 +
 21 files changed, 2263 insertions(+), 415 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/net/nxp,netc-blk-ctrl.yaml
 create mode 100644 drivers/net/ethernet/freescale/enetc/enetc4_hw.h
 create mode 100644 drivers/net/ethernet/freescale/enetc/enetc4_pf.c
 create mode 100644 drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
 create mode 100644 drivers/net/ethernet/freescale/enetc/netc_blk_ctrl.c
 create mode 100644 include/linux/fsl/netc_global.h

-- 
2.34.1

[PATCH v5 net-next 01/13] dt-bindings: net: add compatible string for i.MX95 EMDIO

[Wei Fang]

The EMDIO of i.MX95 has been upgraded to revision 4.1, and the vendor
ID and device ID have also changed, so add the new compatible strings
for i.MX95 EMDIO.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Rob Herring (Arm) <robh@kernel.org>
---
v5: no changes
---
 .../devicetree/bindings/net/fsl,enetc-mdio.yaml       | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/Documentation/devicetree/bindings/net/fsl,enetc-mdio.yaml b/Documentation/devicetree/bindings/net/fsl,enetc-mdio.yaml
index c1dd6aa04321..71c43ece8295 100644
--- a/Documentation/devicetree/bindings/net/fsl,enetc-mdio.yaml
+++ b/Documentation/devicetree/bindings/net/fsl,enetc-mdio.yaml
@@ -20,10 +20,13 @@ maintainers:
 
 properties:
   compatible:
-    items:
-      - enum:
-          - pci1957,ee01
-      - const: fsl,enetc-mdio
+    oneOf:
+      - items:
+          - enum:
+              - pci1957,ee01
+          - const: fsl,enetc-mdio
+      - items:
+          - const: pci1131,ee00
 
   reg:
     maxItems: 1
-- 
2.34.1

[PATCH v5 net-next 02/13] dt-bindings: net: add i.MX95 ENETC support

[Wei Fang]

The ENETC of i.MX95 has been upgraded to revision 4.1, and the vendor
ID and device ID have also changed, so add the new compatible strings
for i.MX95 ENETC. In addition, i.MX95 supports configuration of RGMII
or RMII reference clock.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
v2: Remove "nxp,imx95-enetc" compatible string.
v3:
1. Add restriction to "clcoks" and "clock-names" properties and rename
the clock, also remove the items from these two properties.
2. Remove unnecessary items for "pci1131,e101" compatible string.
v4: Move clocks and clock-names to top level.
v5: Add items to clocks and clock-names
---
 .../devicetree/bindings/net/fsl,enetc.yaml    | 34 +++++++++++++++++--
 1 file changed, 31 insertions(+), 3 deletions(-)

diff --git a/Documentation/devicetree/bindings/net/fsl,enetc.yaml b/Documentation/devicetree/bindings/net/fsl,enetc.yaml
index e152c93998fe..72d2d5d285cd 100644
--- a/Documentation/devicetree/bindings/net/fsl,enetc.yaml
+++ b/Documentation/devicetree/bindings/net/fsl,enetc.yaml
@@ -20,14 +20,23 @@ maintainers:
 
 properties:
   compatible:
-    items:
+    oneOf:
+      - items:
+          - enum:
+              - pci1957,e100
+          - const: fsl,enetc
       - enum:
-          - pci1957,e100
-      - const: fsl,enetc
+          - pci1131,e101
 
   reg:
     maxItems: 1
 
+  clocks:
+    maxItems: 1
+
+  clock-names:
+    maxItems: 1
+
   mdio:
     $ref: mdio.yaml
     unevaluatedProperties: false
@@ -40,6 +49,25 @@ required:
 allOf:
   - $ref: /schemas/pci/pci-device.yaml
   - $ref: ethernet-controller.yaml
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - pci1131,e101
+    then:
+      properties:
+        clocks:
+          items:
+            - description: MAC transmit/receiver reference clock
+
+        clock-names:
+          items:
+            - const: ref
+    else:
+      properties:
+        clocks: false
+        clock-names: false
 
 unevaluatedProperties: false
 
-- 
2.34.1

[PATCH v5 net-next 03/13] dt-bindings: net: add bindings for NETC blocks control

[Wei Fang]

Add bindings for NXP NETC blocks control. Usually, NETC has 2 blocks of
64KB registers, integrated endpoint register block (IERB) and privileged
register block (PRB). IERB is used for pre-boot initialization for all
NETC devices, such as ENETC, Timer, EMDIO and so on. And PRB controls
global reset and global error handling for NETC. Moreover, for the i.MX
platform, there is also a NETCMIX block for link configuration, such as
MII protocol, PCS protocol, etc.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
v2 changes:
1. Rephrase the commit message.
2. Change unevaluatedProperties to additionalProperties.
3. Remove the useless lables from examples.
v3 changes:
1. Remove the items from clocks and clock-names, add maxItems to clocks
and rename the clock.
v4 changes:
1. Reorder the required properties.
2. Add assigned-clocks, assigned-clock-parents and assigned-clock-rates.
v5 changes:
1. Remove assigned-clocks, assigned-clock-parents and assigned-clock-rates
2. Remove minItems from reg and reg-names
3. Rename the node in the examples to be more general
---
 .../bindings/net/nxp,netc-blk-ctrl.yaml       | 104 ++++++++++++++++++
 1 file changed, 104 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/nxp,netc-blk-ctrl.yaml

diff --git a/Documentation/devicetree/bindings/net/nxp,netc-blk-ctrl.yaml b/Documentation/devicetree/bindings/net/nxp,netc-blk-ctrl.yaml
new file mode 100644
index 000000000000..97389fd5dbbf
--- /dev/null
+++ b/Documentation/devicetree/bindings/net/nxp,netc-blk-ctrl.yaml
@@ -0,0 +1,104 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/nxp,netc-blk-ctrl.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: NETC Blocks Control
+
+description:
+  Usually, NETC has 2 blocks of 64KB registers, integrated endpoint register
+  block (IERB) and privileged register block (PRB). IERB is used for pre-boot
+  initialization for all NETC devices, such as ENETC, Timer, EMIDO and so on.
+  And PRB controls global reset and global error handling for NETC. Moreover,
+  for the i.MX platform, there is also a NETCMIX block for link configuration,
+  such as MII protocol, PCS protocol, etc.
+
+maintainers:
+  - Wei Fang <wei.fang@nxp.com>
+  - Clark Wang <xiaoning.wang@nxp.com>
+
+properties:
+  compatible:
+    enum:
+      - nxp,imx95-netc-blk-ctrl
+
+  reg:
+    maxItems: 3
+
+  reg-names:
+    items:
+      - const: ierb
+      - const: prb
+      - const: netcmix
+
+  "#address-cells":
+    const: 2
+
+  "#size-cells":
+    const: 2
+
+  ranges: true
+
+  clocks:
+    maxItems: 1
+
+  clock-names:
+    items:
+      - const: ipg
+
+  power-domains:
+    maxItems: 1
+
+patternProperties:
+  "^pcie@[0-9a-f]+$":
+    $ref: /schemas/pci/host-generic-pci.yaml#
+
+required:
+  - compatible
+  - reg
+  - reg-names
+  - "#address-cells"
+  - "#size-cells"
+  - ranges
+
+additionalProperties: false
+
+examples:
+  - |
+    bus {
+        #address-cells = <2>;
+        #size-cells = <2>;
+
+        system-controller@4cde0000 {
+            compatible = "nxp,imx95-netc-blk-ctrl";
+            reg = <0x0 0x4cde0000 0x0 0x10000>,
+                  <0x0 0x4cdf0000 0x0 0x10000>,
+                  <0x0 0x4c81000c 0x0 0x18>;
+            reg-names = "ierb", "prb", "netcmix";
+            #address-cells = <2>;
+            #size-cells = <2>;
+            ranges;
+            clocks = <&scmi_clk 98>;
+            clock-names = "ipg";
+            power-domains = <&scmi_devpd 18>;
+
+            pcie@4cb00000 {
+                compatible = "pci-host-ecam-generic";
+                reg = <0x0 0x4cb00000 0x0 0x100000>;
+                #address-cells = <3>;
+                #size-cells = <2>;
+                device_type = "pci";
+                bus-range = <0x1 0x1>;
+                ranges = <0x82000000 0x0 0x4cce0000  0x0 0x4cce0000  0x0 0x20000
+                          0xc2000000 0x0 0x4cd10000  0x0 0x4cd10000  0x0 0x10000>;
+
+                mdio@0,0 {
+                    compatible = "pci1131,ee00";
+                    reg = <0x010000 0 0 0 0>;
+                    #address-cells = <1>;
+                    #size-cells = <0>;
+                };
+            };
+        };
+    };
-- 
2.34.1

[PATCH v5 net-next 04/13] net: enetc: add initial netc-blk-ctrl driver support

[Wei Fang]

The netc-blk-ctrl driver is used to configure Integrated Endpoint
Register Block (IERB) and Privileged Register Block (PRB) of NETC.
For i.MX platforms, it is also used to configure the NETCMIX block.

The IERB contains registers that are used for pre-boot initialization,
debug, and non-customer configuration. The PRB controls global reset
and global error handling for NETC. The NETCMIX block is mainly used
to set MII protocol and PCS protocol of the links, it also contains
settings for some other functions.

Note the IERB configuration registers can only be written after being
unlocked by PRB, otherwise, all write operations are inhibited. A warm
reset is performed when the IERB is unlocked, and it results in an FLR
to all NETC devices. Therefore, all NETC device drivers must be probed
or initialized after the warm reset is finished.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
---
v5: no changes
---
 drivers/net/ethernet/freescale/enetc/Kconfig  |  14 +
 drivers/net/ethernet/freescale/enetc/Makefile |   3 +
 .../ethernet/freescale/enetc/netc_blk_ctrl.c  | 438 ++++++++++++++++++
 include/linux/fsl/netc_global.h               |  19 +
 4 files changed, 474 insertions(+)
 create mode 100644 drivers/net/ethernet/freescale/enetc/netc_blk_ctrl.c
 create mode 100644 include/linux/fsl/netc_global.h

diff --git a/drivers/net/ethernet/freescale/enetc/Kconfig b/drivers/net/ethernet/freescale/enetc/Kconfig
index 4d75e6807e92..51d80ea959d4 100644
--- a/drivers/net/ethernet/freescale/enetc/Kconfig
+++ b/drivers/net/ethernet/freescale/enetc/Kconfig
@@ -75,3 +75,17 @@ config FSL_ENETC_QOS
 	  enable/disable from user space via Qos commands(tc). In the kernel
 	  side, it can be loaded by Qos driver. Currently, it is only support
 	  taprio(802.1Qbv) and Credit Based Shaper(802.1Qbu).
+
+config NXP_NETC_BLK_CTRL
+	tristate "NETC blocks control driver"
+	help
+	  This driver configures Integrated Endpoint Register Block (IERB) and
+	  Privileged Register Block (PRB) of NETC. For i.MX platforms, it also
+	  includes the configuration of NETCMIX block.
+	  The IERB contains registers that are used for pre-boot initialization,
+	  debug, and non-customer configuration. The PRB controls global reset
+	  and global error handling for NETC. The NETCMIX block is mainly used
+	  to set MII protocol and PCS protocol of the links, it also contains
+	  settings for some other functions.
+
+	  If compiled as module (M), the module name is nxp-netc-blk-ctrl.
diff --git a/drivers/net/ethernet/freescale/enetc/Makefile b/drivers/net/ethernet/freescale/enetc/Makefile
index b13cbbabb2ea..737c32f83ea5 100644
--- a/drivers/net/ethernet/freescale/enetc/Makefile
+++ b/drivers/net/ethernet/freescale/enetc/Makefile
@@ -19,3 +19,6 @@ fsl-enetc-mdio-y := enetc_pci_mdio.o enetc_mdio.o
 
 obj-$(CONFIG_FSL_ENETC_PTP_CLOCK) += fsl-enetc-ptp.o
 fsl-enetc-ptp-y := enetc_ptp.o
+
+obj-$(CONFIG_NXP_NETC_BLK_CTRL) += nxp-netc-blk-ctrl.o
+nxp-netc-blk-ctrl-y := netc_blk_ctrl.o
diff --git a/drivers/net/ethernet/freescale/enetc/netc_blk_ctrl.c b/drivers/net/ethernet/freescale/enetc/netc_blk_ctrl.c
new file mode 100644
index 000000000000..9bdee15ef013
--- /dev/null
+++ b/drivers/net/ethernet/freescale/enetc/netc_blk_ctrl.c
@@ -0,0 +1,438 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause)
+/*
+ * NXP NETC Blocks Control Driver
+ *
+ * Copyright 2024 NXP
+ */
+
+#include <linux/bits.h>
+#include <linux/clk.h>
+#include <linux/debugfs.h>
+#include <linux/delay.h>
+#include <linux/fsl/netc_global.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_net.h>
+#include <linux/of_platform.h>
+#include <linux/phy.h>
+#include <linux/platform_device.h>
+#include <linux/seq_file.h>
+
+/* NETCMIX registers */
+#define IMX95_CFG_LINK_IO_VAR		0x0
+#define  IO_VAR_16FF_16G_SERDES		0x1
+#define  IO_VAR(port, var)		(((var) & 0xf) << ((port) << 2))
+
+#define IMX95_CFG_LINK_MII_PROT		0x4
+#define CFG_LINK_MII_PORT_0		GENMASK(3, 0)
+#define CFG_LINK_MII_PORT_1		GENMASK(7, 4)
+#define  MII_PROT_MII			0x0
+#define  MII_PROT_RMII			0x1
+#define  MII_PROT_RGMII			0x2
+#define  MII_PROT_SERIAL		0x3
+#define  MII_PROT(port, prot)		(((prot) & 0xf) << ((port) << 2))
+
+#define IMX95_CFG_LINK_PCS_PROT(a)	(0x8 + (a) * 4)
+#define PCS_PROT_1G_SGMII		BIT(0)
+#define PCS_PROT_2500M_SGMII		BIT(1)
+#define PCS_PROT_XFI			BIT(3)
+#define PCS_PROT_SFI			BIT(4)
+#define PCS_PROT_10G_SXGMII		BIT(6)
+
+/* NETC privileged register block register */
+#define PRB_NETCRR			0x100
+#define  NETCRR_SR			BIT(0)
+#define  NETCRR_LOCK			BIT(1)
+
+#define PRB_NETCSR			0x104
+#define  NETCSR_ERROR			BIT(0)
+#define  NETCSR_STATE			BIT(1)
+
+/* NETC integrated endpoint register block register */
+#define IERB_EMDIOFAUXR			0x344
+#define IERB_T0FAUXR			0x444
+#define IERB_EFAUXR(a)			(0x3044 + 0x100 * (a))
+#define IERB_VFAUXR(a)			(0x4004 + 0x40 * (a))
+#define FAUXR_LDID			GENMASK(3, 0)
+
+/* Platform information */
+#define IMX95_ENETC0_BUS_DEVFN		0x0
+#define IMX95_ENETC1_BUS_DEVFN		0x40
+#define IMX95_ENETC2_BUS_DEVFN		0x80
+
+/* Flags for different platforms */
+#define NETC_HAS_NETCMIX		BIT(0)
+
+struct netc_devinfo {
+	u32 flags;
+	int (*netcmix_init)(struct platform_device *pdev);
+	int (*ierb_init)(struct platform_device *pdev);
+};
+
+struct netc_blk_ctrl {
+	void __iomem *prb;
+	void __iomem *ierb;
+	void __iomem *netcmix;
+
+	const struct netc_devinfo *devinfo;
+	struct platform_device *pdev;
+	struct dentry *debugfs_root;
+};
+
+static void netc_reg_write(void __iomem *base, u32 offset, u32 val)
+{
+	netc_write(base + offset, val);
+}
+
+static u32 netc_reg_read(void __iomem *base, u32 offset)
+{
+	return netc_read(base + offset);
+}
+
+static int netc_of_pci_get_bus_devfn(struct device_node *np)
+{
+	u32 reg[5];
+	int error;
+
+	error = of_property_read_u32_array(np, "reg", reg, ARRAY_SIZE(reg));
+	if (error)
+		return error;
+
+	return (reg[0] >> 8) & 0xffff;
+}
+
+static int netc_get_link_mii_protocol(phy_interface_t interface)
+{
+	switch (interface) {
+	case PHY_INTERFACE_MODE_MII:
+		return MII_PROT_MII;
+	case PHY_INTERFACE_MODE_RMII:
+		return MII_PROT_RMII;
+	case PHY_INTERFACE_MODE_RGMII:
+	case PHY_INTERFACE_MODE_RGMII_ID:
+	case PHY_INTERFACE_MODE_RGMII_RXID:
+	case PHY_INTERFACE_MODE_RGMII_TXID:
+		return MII_PROT_RGMII;
+	case PHY_INTERFACE_MODE_SGMII:
+	case PHY_INTERFACE_MODE_2500BASEX:
+	case PHY_INTERFACE_MODE_10GBASER:
+	case PHY_INTERFACE_MODE_XGMII:
+	case PHY_INTERFACE_MODE_USXGMII:
+		return MII_PROT_SERIAL;
+	default:
+		return -EINVAL;
+	}
+}
+
+static int imx95_netcmix_init(struct platform_device *pdev)
+{
+	struct netc_blk_ctrl *priv = platform_get_drvdata(pdev);
+	struct device_node *np = pdev->dev.of_node;
+	phy_interface_t interface;
+	int bus_devfn, mii_proto;
+	u32 val;
+	int err;
+
+	/* Default setting of MII protocol */
+	val = MII_PROT(0, MII_PROT_RGMII) | MII_PROT(1, MII_PROT_RGMII) |
+	      MII_PROT(2, MII_PROT_SERIAL);
+
+	/* Update the link MII protocol through parsing phy-mode */
+	for_each_available_child_of_node_scoped(np, child) {
+		for_each_available_child_of_node_scoped(child, gchild) {
+			if (!of_device_is_compatible(gchild, "pci1131,e101"))
+				continue;
+
+			bus_devfn = netc_of_pci_get_bus_devfn(gchild);
+			if (bus_devfn < 0)
+				return -EINVAL;
+
+			if (bus_devfn == IMX95_ENETC2_BUS_DEVFN)
+				continue;
+
+			err = of_get_phy_mode(gchild, &interface);
+			if (err)
+				continue;
+
+			mii_proto = netc_get_link_mii_protocol(interface);
+			if (mii_proto < 0)
+				return -EINVAL;
+
+			switch (bus_devfn) {
+			case IMX95_ENETC0_BUS_DEVFN:
+				val = u32_replace_bits(val, mii_proto,
+						       CFG_LINK_MII_PORT_0);
+				break;
+			case IMX95_ENETC1_BUS_DEVFN:
+				val = u32_replace_bits(val, mii_proto,
+						       CFG_LINK_MII_PORT_1);
+				break;
+			default:
+				return -EINVAL;
+			}
+		}
+	}
+
+	/* Configure Link I/O variant */
+	netc_reg_write(priv->netcmix, IMX95_CFG_LINK_IO_VAR,
+		       IO_VAR(2, IO_VAR_16FF_16G_SERDES));
+	/* Configure Link 2 PCS protocol */
+	netc_reg_write(priv->netcmix, IMX95_CFG_LINK_PCS_PROT(2),
+		       PCS_PROT_10G_SXGMII);
+	netc_reg_write(priv->netcmix, IMX95_CFG_LINK_MII_PROT, val);
+
+	return 0;
+}
+
+static bool netc_ierb_is_locked(struct netc_blk_ctrl *priv)
+{
+	return !!(netc_reg_read(priv->prb, PRB_NETCRR) & NETCRR_LOCK);
+}
+
+static int netc_lock_ierb(struct netc_blk_ctrl *priv)
+{
+	u32 val;
+
+	netc_reg_write(priv->prb, PRB_NETCRR, NETCRR_LOCK);
+
+	return read_poll_timeout(netc_reg_read, val, !(val & NETCSR_STATE),
+				 100, 2000, false, priv->prb, PRB_NETCSR);
+}
+
+static int netc_unlock_ierb_with_warm_reset(struct netc_blk_ctrl *priv)
+{
+	u32 val;
+
+	netc_reg_write(priv->prb, PRB_NETCRR, 0);
+
+	return read_poll_timeout(netc_reg_read, val, !(val & NETCRR_LOCK),
+				 1000, 100000, true, priv->prb, PRB_NETCRR);
+}
+
+static int imx95_ierb_init(struct platform_device *pdev)
+{
+	struct netc_blk_ctrl *priv = platform_get_drvdata(pdev);
+
+	/* EMDIO : No MSI-X intterupt */
+	netc_reg_write(priv->ierb, IERB_EMDIOFAUXR, 0);
+	/* ENETC0 PF */
+	netc_reg_write(priv->ierb, IERB_EFAUXR(0), 0);
+	/* ENETC0 VF0 */
+	netc_reg_write(priv->ierb, IERB_VFAUXR(0), 1);
+	/* ENETC0 VF1 */
+	netc_reg_write(priv->ierb, IERB_VFAUXR(1), 2);
+	/* ENETC1 PF */
+	netc_reg_write(priv->ierb, IERB_EFAUXR(1), 3);
+	/* ENETC1 VF0 */
+	netc_reg_write(priv->ierb, IERB_VFAUXR(2), 5);
+	/* ENETC1 VF1 */
+	netc_reg_write(priv->ierb, IERB_VFAUXR(3), 6);
+	/* ENETC2 PF */
+	netc_reg_write(priv->ierb, IERB_EFAUXR(2), 4);
+	/* ENETC2 VF0 */
+	netc_reg_write(priv->ierb, IERB_VFAUXR(4), 5);
+	/* ENETC2 VF1 */
+	netc_reg_write(priv->ierb, IERB_VFAUXR(5), 6);
+	/* NETC TIMER */
+	netc_reg_write(priv->ierb, IERB_T0FAUXR, 7);
+
+	return 0;
+}
+
+static int netc_ierb_init(struct platform_device *pdev)
+{
+	struct netc_blk_ctrl *priv = platform_get_drvdata(pdev);
+	const struct netc_devinfo *devinfo = priv->devinfo;
+	int err;
+
+	if (netc_ierb_is_locked(priv)) {
+		err = netc_unlock_ierb_with_warm_reset(priv);
+		if (err) {
+			dev_err(&pdev->dev, "Unlock IERB failed.\n");
+			return err;
+		}
+	}
+
+	if (devinfo->ierb_init) {
+		err = devinfo->ierb_init(pdev);
+		if (err)
+			return err;
+	}
+
+	err = netc_lock_ierb(priv);
+	if (err) {
+		dev_err(&pdev->dev, "Lock IERB failed.\n");
+		return err;
+	}
+
+	return 0;
+}
+
+#if IS_ENABLED(CONFIG_DEBUG_FS)
+static int netc_prb_show(struct seq_file *s, void *data)
+{
+	struct netc_blk_ctrl *priv = s->private;
+	u32 val;
+
+	val = netc_reg_read(priv->prb, PRB_NETCRR);
+	seq_printf(s, "[PRB NETCRR] Lock:%d SR:%d\n",
+		   (val & NETCRR_LOCK) ? 1 : 0,
+		   (val & NETCRR_SR) ? 1 : 0);
+
+	val = netc_reg_read(priv->prb, PRB_NETCSR);
+	seq_printf(s, "[PRB NETCSR] State:%d Error:%d\n",
+		   (val & NETCSR_STATE) ? 1 : 0,
+		   (val & NETCSR_ERROR) ? 1 : 0);
+
+	return 0;
+}
+DEFINE_SHOW_ATTRIBUTE(netc_prb);
+
+static void netc_blk_ctrl_create_debugfs(struct netc_blk_ctrl *priv)
+{
+	struct dentry *root;
+
+	root = debugfs_create_dir("netc_blk_ctrl", NULL);
+	if (IS_ERR(root))
+		return;
+
+	priv->debugfs_root = root;
+
+	debugfs_create_file("prb", 0444, root, priv, &netc_prb_fops);
+}
+
+static void netc_blk_ctrl_remove_debugfs(struct netc_blk_ctrl *priv)
+{
+	debugfs_remove_recursive(priv->debugfs_root);
+	priv->debugfs_root = NULL;
+}
+
+#else
+
+static void netc_blk_ctrl_create_debugfs(struct netc_blk_ctrl *priv)
+{
+}
+
+static void netc_blk_ctrl_remove_debugfs(struct netc_blk_ctrl *priv)
+{
+}
+#endif
+
+static int netc_prb_check_error(struct netc_blk_ctrl *priv)
+{
+	if (netc_reg_read(priv->prb, PRB_NETCSR) & NETCSR_ERROR)
+		return -1;
+
+	return 0;
+}
+
+static const struct netc_devinfo imx95_devinfo = {
+	.flags = NETC_HAS_NETCMIX,
+	.netcmix_init = imx95_netcmix_init,
+	.ierb_init = imx95_ierb_init,
+};
+
+static const struct of_device_id netc_blk_ctrl_match[] = {
+	{ .compatible = "nxp,imx95-netc-blk-ctrl", .data = &imx95_devinfo },
+	{},
+};
+MODULE_DEVICE_TABLE(of, netc_blk_ctrl_match);
+
+static int netc_blk_ctrl_probe(struct platform_device *pdev)
+{
+	struct device_node *node = pdev->dev.of_node;
+	const struct netc_devinfo *devinfo;
+	struct device *dev = &pdev->dev;
+	const struct of_device_id *id;
+	struct netc_blk_ctrl *priv;
+	struct clk *ipg_clk;
+	void __iomem *regs;
+	int err;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->pdev = pdev;
+	ipg_clk = devm_clk_get_optional_enabled(dev, "ipg");
+	if (IS_ERR(ipg_clk))
+		return dev_err_probe(dev, PTR_ERR(ipg_clk),
+				     "Set ipg clock failed\n");
+
+	id = of_match_device(netc_blk_ctrl_match, dev);
+	if (!id)
+		return dev_err_probe(dev, -EINVAL, "Cannot match device\n");
+
+	devinfo = (struct netc_devinfo *)id->data;
+	if (!devinfo)
+		return dev_err_probe(dev, -EINVAL, "No device information\n");
+
+	priv->devinfo = devinfo;
+	regs = devm_platform_ioremap_resource_byname(pdev, "ierb");
+	if (IS_ERR(regs))
+		return dev_err_probe(dev, PTR_ERR(regs),
+				     "Missing IERB resource\n");
+
+	priv->ierb = regs;
+	regs = devm_platform_ioremap_resource_byname(pdev, "prb");
+	if (IS_ERR(regs))
+		return dev_err_probe(dev, PTR_ERR(regs),
+				     "Missing PRB resource\n");
+
+	priv->prb = regs;
+	if (devinfo->flags & NETC_HAS_NETCMIX) {
+		regs = devm_platform_ioremap_resource_byname(pdev, "netcmix");
+		if (IS_ERR(regs))
+			return dev_err_probe(dev, PTR_ERR(regs),
+					     "Missing NETCMIX resource\n");
+		priv->netcmix = regs;
+	}
+
+	platform_set_drvdata(pdev, priv);
+	if (devinfo->netcmix_init) {
+		err = devinfo->netcmix_init(pdev);
+		if (err)
+			return dev_err_probe(dev, err,
+					     "Initializing NETCMIX failed\n");
+	}
+
+	err = netc_ierb_init(pdev);
+	if (err)
+		return dev_err_probe(dev, err, "Initializing IERB failed\n");
+
+	if (netc_prb_check_error(priv) < 0)
+		dev_warn(dev, "The current IERB configuration is invalid\n");
+
+	netc_blk_ctrl_create_debugfs(priv);
+
+	err = of_platform_populate(node, NULL, NULL, dev);
+	if (err) {
+		netc_blk_ctrl_remove_debugfs(priv);
+		return dev_err_probe(dev, err, "of_platform_populate failed\n");
+	}
+
+	return 0;
+}
+
+static void netc_blk_ctrl_remove(struct platform_device *pdev)
+{
+	struct netc_blk_ctrl *priv = platform_get_drvdata(pdev);
+
+	of_platform_depopulate(&pdev->dev);
+	netc_blk_ctrl_remove_debugfs(priv);
+}
+
+static struct platform_driver netc_blk_ctrl_driver = {
+	.driver = {
+		.name = "nxp-netc-blk-ctrl",
+		.of_match_table = netc_blk_ctrl_match,
+	},
+	.probe = netc_blk_ctrl_probe,
+	.remove = netc_blk_ctrl_remove,
+};
+
+module_platform_driver(netc_blk_ctrl_driver);
+
+MODULE_DESCRIPTION("NXP NETC Blocks Control Driver");
+MODULE_LICENSE("Dual BSD/GPL");
diff --git a/include/linux/fsl/netc_global.h b/include/linux/fsl/netc_global.h
new file mode 100644
index 000000000000..fdecca8c90f0
--- /dev/null
+++ b/include/linux/fsl/netc_global.h
@@ -0,0 +1,19 @@
+/* SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause) */
+/* Copyright 2024 NXP
+ */
+#ifndef __NETC_GLOBAL_H
+#define __NETC_GLOBAL_H
+
+#include <linux/io.h>
+
+static inline u32 netc_read(void __iomem *reg)
+{
+	return ioread32(reg);
+}
+
+static inline void netc_write(void __iomem *reg, u32 val)
+{
+	iowrite32(val, reg);
+}
+
+#endif
-- 
2.34.1

[PATCH v5 net-next 05/13] net: enetc: extract common ENETC PF parts for LS1028A and i.MX95 platforms

[Wei Fang]

The ENETC PF driver of LS1028A (rev 1.0) is incompatible with the version
used on the i.MX95 platform (rev 4.1), except for the station interface
(SI) part. To reduce code redundancy and prepare for a new driver for rev
4.1 and later, extract shared interfaces from enetc_pf.c and move them to
enetc_pf_common.c. This refactoring lays the groundwork for compiling
enetc_pf_common.c into a shared driver for both platforms' PF drivers.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
---
v5: no changes
---
 drivers/net/ethernet/freescale/enetc/Makefile |   2 +-
 .../net/ethernet/freescale/enetc/enetc_pf.c   | 297 +-----------------
 .../net/ethernet/freescale/enetc/enetc_pf.h   |  13 +
 .../freescale/enetc/enetc_pf_common.c         | 295 +++++++++++++++++
 4 files changed, 313 insertions(+), 294 deletions(-)
 create mode 100644 drivers/net/ethernet/freescale/enetc/enetc_pf_common.c

diff --git a/drivers/net/ethernet/freescale/enetc/Makefile b/drivers/net/ethernet/freescale/enetc/Makefile
index 737c32f83ea5..8f4d8e9c37a0 100644
--- a/drivers/net/ethernet/freescale/enetc/Makefile
+++ b/drivers/net/ethernet/freescale/enetc/Makefile
@@ -4,7 +4,7 @@ obj-$(CONFIG_FSL_ENETC_CORE) += fsl-enetc-core.o
 fsl-enetc-core-y := enetc.o enetc_cbdr.o enetc_ethtool.o
 
 obj-$(CONFIG_FSL_ENETC) += fsl-enetc.o
-fsl-enetc-y := enetc_pf.o
+fsl-enetc-y := enetc_pf.o enetc_pf_common.o
 fsl-enetc-$(CONFIG_PCI_IOV) += enetc_msg.o
 fsl-enetc-$(CONFIG_FSL_ENETC_QOS) += enetc_qos.o
 
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.c b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
index 8f6b0bf48139..3cdd149056f9 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
@@ -2,11 +2,8 @@
 /* Copyright 2017-2019 NXP */
 
 #include <linux/unaligned.h>
-#include <linux/mdio.h>
 #include <linux/module.h>
-#include <linux/fsl/enetc_mdio.h>
 #include <linux/of_platform.h>
-#include <linux/of_mdio.h>
 #include <linux/of_net.h>
 #include <linux/pcs-lynx.h>
 #include "enetc_ierb.h"
@@ -14,7 +11,7 @@
 
 #define ENETC_DRV_NAME_STR "ENETC PF driver"
 
-static void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr)
+void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr)
 {
 	u32 upper = __raw_readl(hw->port + ENETC_PSIPMAR0(si));
 	u16 lower = __raw_readw(hw->port + ENETC_PSIPMAR1(si));
@@ -23,8 +20,8 @@ static void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr)
 	put_unaligned_le16(lower, addr + 4);
 }
 
-static void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
-					  const u8 *addr)
+void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
+				   const u8 *addr)
 {
 	u32 upper = get_unaligned_le32(addr);
 	u16 lower = get_unaligned_le16(addr + 4);
@@ -33,20 +30,6 @@ static void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
 	__raw_writew(lower, hw->port + ENETC_PSIPMAR1(si));
 }
 
-static int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr)
-{
-	struct enetc_ndev_priv *priv = netdev_priv(ndev);
-	struct sockaddr *saddr = addr;
-
-	if (!is_valid_ether_addr(saddr->sa_data))
-		return -EADDRNOTAVAIL;
-
-	eth_hw_addr_set(ndev, saddr->sa_data);
-	enetc_pf_set_primary_mac_addr(&priv->si->hw, 0, saddr->sa_data);
-
-	return 0;
-}
-
 static void enetc_set_vlan_promisc(struct enetc_hw *hw, char si_map)
 {
 	u32 val = enetc_port_rd(hw, ENETC_PSIPVMR);
@@ -393,56 +376,6 @@ static int enetc_pf_set_vf_spoofchk(struct net_device *ndev, int vf, bool en)
 	return 0;
 }
 
-static int enetc_setup_mac_address(struct device_node *np, struct enetc_pf *pf,
-				   int si)
-{
-	struct device *dev = &pf->si->pdev->dev;
-	struct enetc_hw *hw = &pf->si->hw;
-	u8 mac_addr[ETH_ALEN] = { 0 };
-	int err;
-
-	/* (1) try to get the MAC address from the device tree */
-	if (np) {
-		err = of_get_mac_address(np, mac_addr);
-		if (err == -EPROBE_DEFER)
-			return err;
-	}
-
-	/* (2) bootloader supplied MAC address */
-	if (is_zero_ether_addr(mac_addr))
-		enetc_pf_get_primary_mac_addr(hw, si, mac_addr);
-
-	/* (3) choose a random one */
-	if (is_zero_ether_addr(mac_addr)) {
-		eth_random_addr(mac_addr);
-		dev_info(dev, "no MAC address specified for SI%d, using %pM\n",
-			 si, mac_addr);
-	}
-
-	enetc_pf_set_primary_mac_addr(hw, si, mac_addr);
-
-	return 0;
-}
-
-static int enetc_setup_mac_addresses(struct device_node *np,
-				     struct enetc_pf *pf)
-{
-	int err, i;
-
-	/* The PF might take its MAC from the device tree */
-	err = enetc_setup_mac_address(np, pf, 0);
-	if (err)
-		return err;
-
-	for (i = 0; i < pf->total_vfs; i++) {
-		err = enetc_setup_mac_address(NULL, pf, i + 1);
-		if (err)
-			return err;
-	}
-
-	return 0;
-}
-
 static void enetc_port_assign_rfs_entries(struct enetc_si *si)
 {
 	struct enetc_pf *pf = enetc_si_priv(si);
@@ -775,187 +708,6 @@ static const struct net_device_ops enetc_ndev_ops = {
 	.ndo_xdp_xmit		= enetc_xdp_xmit,
 };
 
-static void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
-				  const struct net_device_ops *ndev_ops)
-{
-	struct enetc_ndev_priv *priv = netdev_priv(ndev);
-
-	SET_NETDEV_DEV(ndev, &si->pdev->dev);
-	priv->ndev = ndev;
-	priv->si = si;
-	priv->dev = &si->pdev->dev;
-	si->ndev = ndev;
-
-	priv->msg_enable = (NETIF_MSG_WOL << 1) - 1;
-	ndev->netdev_ops = ndev_ops;
-	enetc_set_ethtool_ops(ndev);
-	ndev->watchdog_timeo = 5 * HZ;
-	ndev->max_mtu = ENETC_MAX_MTU;
-
-	ndev->hw_features = NETIF_F_SG | NETIF_F_RXCSUM |
-			    NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX |
-			    NETIF_F_HW_VLAN_CTAG_FILTER | NETIF_F_LOOPBACK |
-			    NETIF_F_HW_CSUM | NETIF_F_TSO | NETIF_F_TSO6;
-	ndev->features = NETIF_F_HIGHDMA | NETIF_F_SG | NETIF_F_RXCSUM |
-			 NETIF_F_HW_VLAN_CTAG_TX |
-			 NETIF_F_HW_VLAN_CTAG_RX |
-			 NETIF_F_HW_CSUM | NETIF_F_TSO | NETIF_F_TSO6;
-	ndev->vlan_features = NETIF_F_SG | NETIF_F_HW_CSUM |
-			      NETIF_F_TSO | NETIF_F_TSO6;
-
-	if (si->num_rss)
-		ndev->hw_features |= NETIF_F_RXHASH;
-
-	ndev->priv_flags |= IFF_UNICAST_FLT;
-	ndev->xdp_features = NETDEV_XDP_ACT_BASIC | NETDEV_XDP_ACT_REDIRECT |
-			     NETDEV_XDP_ACT_NDO_XMIT | NETDEV_XDP_ACT_RX_SG |
-			     NETDEV_XDP_ACT_NDO_XMIT_SG;
-
-	if (si->hw_features & ENETC_SI_F_PSFP && !enetc_psfp_enable(priv)) {
-		priv->active_offloads |= ENETC_F_QCI;
-		ndev->features |= NETIF_F_HW_TC;
-		ndev->hw_features |= NETIF_F_HW_TC;
-	}
-
-	/* pick up primary MAC address from SI */
-	enetc_load_primary_mac_addr(&si->hw, ndev);
-}
-
-static int enetc_mdio_probe(struct enetc_pf *pf, struct device_node *np)
-{
-	struct device *dev = &pf->si->pdev->dev;
-	struct enetc_mdio_priv *mdio_priv;
-	struct mii_bus *bus;
-	int err;
-
-	bus = devm_mdiobus_alloc_size(dev, sizeof(*mdio_priv));
-	if (!bus)
-		return -ENOMEM;
-
-	bus->name = "Freescale ENETC MDIO Bus";
-	bus->read = enetc_mdio_read_c22;
-	bus->write = enetc_mdio_write_c22;
-	bus->read_c45 = enetc_mdio_read_c45;
-	bus->write_c45 = enetc_mdio_write_c45;
-	bus->parent = dev;
-	mdio_priv = bus->priv;
-	mdio_priv->hw = &pf->si->hw;
-	mdio_priv->mdio_base = ENETC_EMDIO_BASE;
-	snprintf(bus->id, MII_BUS_ID_SIZE, "%s", dev_name(dev));
-
-	err = of_mdiobus_register(bus, np);
-	if (err)
-		return dev_err_probe(dev, err, "cannot register MDIO bus\n");
-
-	pf->mdio = bus;
-
-	return 0;
-}
-
-static void enetc_mdio_remove(struct enetc_pf *pf)
-{
-	if (pf->mdio)
-		mdiobus_unregister(pf->mdio);
-}
-
-static int enetc_imdio_create(struct enetc_pf *pf)
-{
-	struct device *dev = &pf->si->pdev->dev;
-	struct enetc_mdio_priv *mdio_priv;
-	struct phylink_pcs *phylink_pcs;
-	struct mii_bus *bus;
-	int err;
-
-	bus = mdiobus_alloc_size(sizeof(*mdio_priv));
-	if (!bus)
-		return -ENOMEM;
-
-	bus->name = "Freescale ENETC internal MDIO Bus";
-	bus->read = enetc_mdio_read_c22;
-	bus->write = enetc_mdio_write_c22;
-	bus->read_c45 = enetc_mdio_read_c45;
-	bus->write_c45 = enetc_mdio_write_c45;
-	bus->parent = dev;
-	bus->phy_mask = ~0;
-	mdio_priv = bus->priv;
-	mdio_priv->hw = &pf->si->hw;
-	mdio_priv->mdio_base = ENETC_PM_IMDIO_BASE;
-	snprintf(bus->id, MII_BUS_ID_SIZE, "%s-imdio", dev_name(dev));
-
-	err = mdiobus_register(bus);
-	if (err) {
-		dev_err(dev, "cannot register internal MDIO bus (%d)\n", err);
-		goto free_mdio_bus;
-	}
-
-	phylink_pcs = lynx_pcs_create_mdiodev(bus, 0);
-	if (IS_ERR(phylink_pcs)) {
-		err = PTR_ERR(phylink_pcs);
-		dev_err(dev, "cannot create lynx pcs (%d)\n", err);
-		goto unregister_mdiobus;
-	}
-
-	pf->imdio = bus;
-	pf->pcs = phylink_pcs;
-
-	return 0;
-
-unregister_mdiobus:
-	mdiobus_unregister(bus);
-free_mdio_bus:
-	mdiobus_free(bus);
-	return err;
-}
-
-static void enetc_imdio_remove(struct enetc_pf *pf)
-{
-	if (pf->pcs)
-		lynx_pcs_destroy(pf->pcs);
-	if (pf->imdio) {
-		mdiobus_unregister(pf->imdio);
-		mdiobus_free(pf->imdio);
-	}
-}
-
-static bool enetc_port_has_pcs(struct enetc_pf *pf)
-{
-	return (pf->if_mode == PHY_INTERFACE_MODE_SGMII ||
-		pf->if_mode == PHY_INTERFACE_MODE_1000BASEX ||
-		pf->if_mode == PHY_INTERFACE_MODE_2500BASEX ||
-		pf->if_mode == PHY_INTERFACE_MODE_USXGMII);
-}
-
-static int enetc_mdiobus_create(struct enetc_pf *pf, struct device_node *node)
-{
-	struct device_node *mdio_np;
-	int err;
-
-	mdio_np = of_get_child_by_name(node, "mdio");
-	if (mdio_np) {
-		err = enetc_mdio_probe(pf, mdio_np);
-
-		of_node_put(mdio_np);
-		if (err)
-			return err;
-	}
-
-	if (enetc_port_has_pcs(pf)) {
-		err = enetc_imdio_create(pf);
-		if (err) {
-			enetc_mdio_remove(pf);
-			return err;
-		}
-	}
-
-	return 0;
-}
-
-static void enetc_mdiobus_destroy(struct enetc_pf *pf)
-{
-	enetc_mdio_remove(pf);
-	enetc_imdio_remove(pf);
-}
-
 static struct phylink_pcs *
 enetc_pl_mac_select_pcs(struct phylink_config *config, phy_interface_t iface)
 {
@@ -1101,47 +853,6 @@ static const struct phylink_mac_ops enetc_mac_phylink_ops = {
 	.mac_link_down = enetc_pl_mac_link_down,
 };
 
-static int enetc_phylink_create(struct enetc_ndev_priv *priv,
-				struct device_node *node)
-{
-	struct enetc_pf *pf = enetc_si_priv(priv->si);
-	struct phylink *phylink;
-	int err;
-
-	pf->phylink_config.dev = &priv->ndev->dev;
-	pf->phylink_config.type = PHYLINK_NETDEV;
-	pf->phylink_config.mac_capabilities = MAC_ASYM_PAUSE | MAC_SYM_PAUSE |
-		MAC_10 | MAC_100 | MAC_1000 | MAC_2500FD;
-
-	__set_bit(PHY_INTERFACE_MODE_INTERNAL,
-		  pf->phylink_config.supported_interfaces);
-	__set_bit(PHY_INTERFACE_MODE_SGMII,
-		  pf->phylink_config.supported_interfaces);
-	__set_bit(PHY_INTERFACE_MODE_1000BASEX,
-		  pf->phylink_config.supported_interfaces);
-	__set_bit(PHY_INTERFACE_MODE_2500BASEX,
-		  pf->phylink_config.supported_interfaces);
-	__set_bit(PHY_INTERFACE_MODE_USXGMII,
-		  pf->phylink_config.supported_interfaces);
-	phy_interface_set_rgmii(pf->phylink_config.supported_interfaces);
-
-	phylink = phylink_create(&pf->phylink_config, of_fwnode_handle(node),
-				 pf->if_mode, &enetc_mac_phylink_ops);
-	if (IS_ERR(phylink)) {
-		err = PTR_ERR(phylink);
-		return err;
-	}
-
-	priv->phylink = phylink;
-
-	return 0;
-}
-
-static void enetc_phylink_destroy(struct enetc_ndev_priv *priv)
-{
-	phylink_destroy(priv->phylink);
-}
-
 /* Initialize the entire shared memory for the flow steering entries
  * of this port (PF + VFs)
  */
@@ -1338,7 +1049,7 @@ static int enetc_pf_probe(struct pci_dev *pdev,
 	if (err)
 		goto err_mdiobus_create;
 
-	err = enetc_phylink_create(priv, node);
+	err = enetc_phylink_create(priv, node, &enetc_mac_phylink_ops);
 	if (err)
 		goto err_phylink_create;
 
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.h b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
index c26bd66e4597..92a26b09cf57 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
@@ -58,3 +58,16 @@ struct enetc_pf {
 int enetc_msg_psi_init(struct enetc_pf *pf);
 void enetc_msg_psi_free(struct enetc_pf *pf);
 void enetc_msg_handle_rxmsg(struct enetc_pf *pf, int mbox_id, u16 *status);
+
+void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr);
+void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
+				   const u8 *addr);
+int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr);
+int enetc_setup_mac_addresses(struct device_node *np, struct enetc_pf *pf);
+void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
+			   const struct net_device_ops *ndev_ops);
+int enetc_mdiobus_create(struct enetc_pf *pf, struct device_node *node);
+void enetc_mdiobus_destroy(struct enetc_pf *pf);
+int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node *node,
+			 const struct phylink_mac_ops *ops);
+void enetc_phylink_destroy(struct enetc_ndev_priv *priv);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
new file mode 100644
index 000000000000..bce81a4f6f88
--- /dev/null
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
@@ -0,0 +1,295 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause)
+/* Copyright 2024 NXP */
+
+#include <linux/fsl/enetc_mdio.h>
+#include <linux/of_mdio.h>
+#include <linux/of_net.h>
+#include <linux/pcs-lynx.h>
+
+#include "enetc_pf.h"
+
+int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr)
+{
+	struct enetc_ndev_priv *priv = netdev_priv(ndev);
+	struct sockaddr *saddr = addr;
+
+	if (!is_valid_ether_addr(saddr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	eth_hw_addr_set(ndev, saddr->sa_data);
+	enetc_pf_set_primary_mac_addr(&priv->si->hw, 0, saddr->sa_data);
+
+	return 0;
+}
+
+static int enetc_setup_mac_address(struct device_node *np, struct enetc_pf *pf,
+				   int si)
+{
+	struct device *dev = &pf->si->pdev->dev;
+	struct enetc_hw *hw = &pf->si->hw;
+	u8 mac_addr[ETH_ALEN] = { 0 };
+	int err;
+
+	/* (1) try to get the MAC address from the device tree */
+	if (np) {
+		err = of_get_mac_address(np, mac_addr);
+		if (err == -EPROBE_DEFER)
+			return err;
+	}
+
+	/* (2) bootloader supplied MAC address */
+	if (is_zero_ether_addr(mac_addr))
+		enetc_pf_get_primary_mac_addr(hw, si, mac_addr);
+
+	/* (3) choose a random one */
+	if (is_zero_ether_addr(mac_addr)) {
+		eth_random_addr(mac_addr);
+		dev_info(dev, "no MAC address specified for SI%d, using %pM\n",
+			 si, mac_addr);
+	}
+
+	enetc_pf_set_primary_mac_addr(hw, si, mac_addr);
+
+	return 0;
+}
+
+int enetc_setup_mac_addresses(struct device_node *np, struct enetc_pf *pf)
+{
+	int err, i;
+
+	/* The PF might take its MAC from the device tree */
+	err = enetc_setup_mac_address(np, pf, 0);
+	if (err)
+		return err;
+
+	for (i = 0; i < pf->total_vfs; i++) {
+		err = enetc_setup_mac_address(NULL, pf, i + 1);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
+			   const struct net_device_ops *ndev_ops)
+{
+	struct enetc_ndev_priv *priv = netdev_priv(ndev);
+
+	SET_NETDEV_DEV(ndev, &si->pdev->dev);
+	priv->ndev = ndev;
+	priv->si = si;
+	priv->dev = &si->pdev->dev;
+	si->ndev = ndev;
+
+	priv->msg_enable = (NETIF_MSG_WOL << 1) - 1;
+	ndev->netdev_ops = ndev_ops;
+	enetc_set_ethtool_ops(ndev);
+	ndev->watchdog_timeo = 5 * HZ;
+	ndev->max_mtu = ENETC_MAX_MTU;
+
+	ndev->hw_features = NETIF_F_SG | NETIF_F_RXCSUM |
+			    NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_CTAG_RX |
+			    NETIF_F_HW_VLAN_CTAG_FILTER | NETIF_F_LOOPBACK |
+			    NETIF_F_HW_CSUM | NETIF_F_TSO | NETIF_F_TSO6;
+	ndev->features = NETIF_F_HIGHDMA | NETIF_F_SG | NETIF_F_RXCSUM |
+			 NETIF_F_HW_VLAN_CTAG_TX |
+			 NETIF_F_HW_VLAN_CTAG_RX |
+			 NETIF_F_HW_CSUM | NETIF_F_TSO | NETIF_F_TSO6;
+	ndev->vlan_features = NETIF_F_SG | NETIF_F_HW_CSUM |
+			      NETIF_F_TSO | NETIF_F_TSO6;
+
+	if (si->num_rss)
+		ndev->hw_features |= NETIF_F_RXHASH;
+
+	ndev->priv_flags |= IFF_UNICAST_FLT;
+	ndev->xdp_features = NETDEV_XDP_ACT_BASIC | NETDEV_XDP_ACT_REDIRECT |
+			     NETDEV_XDP_ACT_NDO_XMIT | NETDEV_XDP_ACT_RX_SG |
+			     NETDEV_XDP_ACT_NDO_XMIT_SG;
+
+	if (si->hw_features & ENETC_SI_F_PSFP && !enetc_psfp_enable(priv)) {
+		priv->active_offloads |= ENETC_F_QCI;
+		ndev->features |= NETIF_F_HW_TC;
+		ndev->hw_features |= NETIF_F_HW_TC;
+	}
+
+	/* pick up primary MAC address from SI */
+	enetc_load_primary_mac_addr(&si->hw, ndev);
+}
+
+static int enetc_mdio_probe(struct enetc_pf *pf, struct device_node *np)
+{
+	struct device *dev = &pf->si->pdev->dev;
+	struct enetc_mdio_priv *mdio_priv;
+	struct mii_bus *bus;
+	int err;
+
+	bus = devm_mdiobus_alloc_size(dev, sizeof(*mdio_priv));
+	if (!bus)
+		return -ENOMEM;
+
+	bus->name = "Freescale ENETC MDIO Bus";
+	bus->read = enetc_mdio_read_c22;
+	bus->write = enetc_mdio_write_c22;
+	bus->read_c45 = enetc_mdio_read_c45;
+	bus->write_c45 = enetc_mdio_write_c45;
+	bus->parent = dev;
+	mdio_priv = bus->priv;
+	mdio_priv->hw = &pf->si->hw;
+	mdio_priv->mdio_base = ENETC_EMDIO_BASE;
+	snprintf(bus->id, MII_BUS_ID_SIZE, "%s", dev_name(dev));
+
+	err = of_mdiobus_register(bus, np);
+	if (err)
+		return dev_err_probe(dev, err, "cannot register MDIO bus\n");
+
+	pf->mdio = bus;
+
+	return 0;
+}
+
+static void enetc_mdio_remove(struct enetc_pf *pf)
+{
+	if (pf->mdio)
+		mdiobus_unregister(pf->mdio);
+}
+
+static int enetc_imdio_create(struct enetc_pf *pf)
+{
+	struct device *dev = &pf->si->pdev->dev;
+	struct enetc_mdio_priv *mdio_priv;
+	struct phylink_pcs *phylink_pcs;
+	struct mii_bus *bus;
+	int err;
+
+	bus = mdiobus_alloc_size(sizeof(*mdio_priv));
+	if (!bus)
+		return -ENOMEM;
+
+	bus->name = "Freescale ENETC internal MDIO Bus";
+	bus->read = enetc_mdio_read_c22;
+	bus->write = enetc_mdio_write_c22;
+	bus->read_c45 = enetc_mdio_read_c45;
+	bus->write_c45 = enetc_mdio_write_c45;
+	bus->parent = dev;
+	bus->phy_mask = ~0;
+	mdio_priv = bus->priv;
+	mdio_priv->hw = &pf->si->hw;
+	mdio_priv->mdio_base = ENETC_PM_IMDIO_BASE;
+	snprintf(bus->id, MII_BUS_ID_SIZE, "%s-imdio", dev_name(dev));
+
+	err = mdiobus_register(bus);
+	if (err) {
+		dev_err(dev, "cannot register internal MDIO bus (%d)\n", err);
+		goto free_mdio_bus;
+	}
+
+	phylink_pcs = lynx_pcs_create_mdiodev(bus, 0);
+	if (IS_ERR(phylink_pcs)) {
+		err = PTR_ERR(phylink_pcs);
+		dev_err(dev, "cannot create lynx pcs (%d)\n", err);
+		goto unregister_mdiobus;
+	}
+
+	pf->imdio = bus;
+	pf->pcs = phylink_pcs;
+
+	return 0;
+
+unregister_mdiobus:
+	mdiobus_unregister(bus);
+free_mdio_bus:
+	mdiobus_free(bus);
+	return err;
+}
+
+static void enetc_imdio_remove(struct enetc_pf *pf)
+{
+	if (pf->pcs)
+		lynx_pcs_destroy(pf->pcs);
+
+	if (pf->imdio) {
+		mdiobus_unregister(pf->imdio);
+		mdiobus_free(pf->imdio);
+	}
+}
+
+static bool enetc_port_has_pcs(struct enetc_pf *pf)
+{
+	return (pf->if_mode == PHY_INTERFACE_MODE_SGMII ||
+		pf->if_mode == PHY_INTERFACE_MODE_1000BASEX ||
+		pf->if_mode == PHY_INTERFACE_MODE_2500BASEX ||
+		pf->if_mode == PHY_INTERFACE_MODE_USXGMII);
+}
+
+int enetc_mdiobus_create(struct enetc_pf *pf, struct device_node *node)
+{
+	struct device_node *mdio_np;
+	int err;
+
+	mdio_np = of_get_child_by_name(node, "mdio");
+	if (mdio_np) {
+		err = enetc_mdio_probe(pf, mdio_np);
+
+		of_node_put(mdio_np);
+		if (err)
+			return err;
+	}
+
+	if (enetc_port_has_pcs(pf)) {
+		err = enetc_imdio_create(pf);
+		if (err) {
+			enetc_mdio_remove(pf);
+			return err;
+		}
+	}
+
+	return 0;
+}
+
+void enetc_mdiobus_destroy(struct enetc_pf *pf)
+{
+	enetc_mdio_remove(pf);
+	enetc_imdio_remove(pf);
+}
+
+int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node *node,
+			 const struct phylink_mac_ops *ops)
+{
+	struct enetc_pf *pf = enetc_si_priv(priv->si);
+	struct phylink *phylink;
+	int err;
+
+	pf->phylink_config.dev = &priv->ndev->dev;
+	pf->phylink_config.type = PHYLINK_NETDEV;
+	pf->phylink_config.mac_capabilities = MAC_ASYM_PAUSE | MAC_SYM_PAUSE |
+		MAC_10 | MAC_100 | MAC_1000 | MAC_2500FD;
+
+	__set_bit(PHY_INTERFACE_MODE_INTERNAL,
+		  pf->phylink_config.supported_interfaces);
+	__set_bit(PHY_INTERFACE_MODE_SGMII,
+		  pf->phylink_config.supported_interfaces);
+	__set_bit(PHY_INTERFACE_MODE_1000BASEX,
+		  pf->phylink_config.supported_interfaces);
+	__set_bit(PHY_INTERFACE_MODE_2500BASEX,
+		  pf->phylink_config.supported_interfaces);
+	__set_bit(PHY_INTERFACE_MODE_USXGMII,
+		  pf->phylink_config.supported_interfaces);
+	phy_interface_set_rgmii(pf->phylink_config.supported_interfaces);
+
+	phylink = phylink_create(&pf->phylink_config, of_fwnode_handle(node),
+				 pf->if_mode, ops);
+	if (IS_ERR(phylink)) {
+		err = PTR_ERR(phylink);
+		return err;
+	}
+
+	priv->phylink = phylink;
+
+	return 0;
+}
+
+void enetc_phylink_destroy(struct enetc_ndev_priv *priv)
+{
+	phylink_destroy(priv->phylink);
+}
-- 
2.34.1

[PATCH v5 net-next 06/13] net: enetc: build enetc_pf_common.c as a separate module

[Wei Fang]

Compile enetc_pf_common.c as a standalone module to allow shared usage
between ENETC v1 and v4 PF drivers. Add struct enetc_pf_ops to register
different hardware operation interfaces for both ENETC v1 and v4 PF
drivers.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
---
v5: no changes
---
 drivers/net/ethernet/freescale/enetc/Kconfig  |  9 ++++
 drivers/net/ethernet/freescale/enetc/Makefile |  5 +-
 .../net/ethernet/freescale/enetc/enetc_pf.c   | 26 ++++++++--
 .../net/ethernet/freescale/enetc/enetc_pf.h   | 21 ++++++--
 .../freescale/enetc/enetc_pf_common.c         | 50 ++++++++++++++++---
 5 files changed, 96 insertions(+), 15 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/Kconfig b/drivers/net/ethernet/freescale/enetc/Kconfig
index 51d80ea959d4..e1b151a98b41 100644
--- a/drivers/net/ethernet/freescale/enetc/Kconfig
+++ b/drivers/net/ethernet/freescale/enetc/Kconfig
@@ -7,6 +7,14 @@ config FSL_ENETC_CORE
 
 	  If compiled as module (M), the module name is fsl-enetc-core.
 
+config NXP_ENETC_PF_COMMON
+	tristate
+	help
+	  This module supports common functionality between drivers of
+	  different versions of NXP ENETC PF controllers.
+
+	  If compiled as module (M), the module name is nxp-enetc-pf-common.
+
 config FSL_ENETC
 	tristate "ENETC PF driver"
 	depends on PCI_MSI
@@ -14,6 +22,7 @@ config FSL_ENETC
 	select FSL_ENETC_CORE
 	select FSL_ENETC_IERB
 	select FSL_ENETC_MDIO
+	select NXP_ENETC_PF_COMMON
 	select PHYLINK
 	select PCS_LYNX
 	select DIMLIB
diff --git a/drivers/net/ethernet/freescale/enetc/Makefile b/drivers/net/ethernet/freescale/enetc/Makefile
index 8f4d8e9c37a0..ebe232673ed4 100644
--- a/drivers/net/ethernet/freescale/enetc/Makefile
+++ b/drivers/net/ethernet/freescale/enetc/Makefile
@@ -3,8 +3,11 @@
 obj-$(CONFIG_FSL_ENETC_CORE) += fsl-enetc-core.o
 fsl-enetc-core-y := enetc.o enetc_cbdr.o enetc_ethtool.o
 
+obj-$(CONFIG_NXP_ENETC_PF_COMMON) += nxp-enetc-pf-common.o
+nxp-enetc-pf-common-y := enetc_pf_common.o
+
 obj-$(CONFIG_FSL_ENETC) += fsl-enetc.o
-fsl-enetc-y := enetc_pf.o enetc_pf_common.o
+fsl-enetc-y := enetc_pf.o
 fsl-enetc-$(CONFIG_PCI_IOV) += enetc_msg.o
 fsl-enetc-$(CONFIG_FSL_ENETC_QOS) += enetc_qos.o
 
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.c b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
index 3cdd149056f9..7522316ddfea 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
@@ -11,7 +11,7 @@
 
 #define ENETC_DRV_NAME_STR "ENETC PF driver"
 
-void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr)
+static void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr)
 {
 	u32 upper = __raw_readl(hw->port + ENETC_PSIPMAR0(si));
 	u16 lower = __raw_readw(hw->port + ENETC_PSIPMAR1(si));
@@ -20,8 +20,8 @@ void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr)
 	put_unaligned_le16(lower, addr + 4);
 }
 
-void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
-				   const u8 *addr)
+static void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
+					  const u8 *addr)
 {
 	u32 upper = get_unaligned_le32(addr);
 	u16 lower = get_unaligned_le16(addr + 4);
@@ -30,6 +30,17 @@ void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
 	__raw_writew(lower, hw->port + ENETC_PSIPMAR1(si));
 }
 
+static struct phylink_pcs *enetc_pf_create_pcs(struct enetc_pf *pf,
+					       struct mii_bus *bus)
+{
+	return lynx_pcs_create_mdiodev(bus, 0);
+}
+
+static void enetc_pf_destroy_pcs(struct phylink_pcs *pcs)
+{
+	lynx_pcs_destroy(pcs);
+}
+
 static void enetc_set_vlan_promisc(struct enetc_hw *hw, char si_map)
 {
 	u32 val = enetc_port_rd(hw, ENETC_PSIPVMR);
@@ -970,6 +981,14 @@ static void enetc_psi_destroy(struct pci_dev *pdev)
 	enetc_pci_remove(pdev);
 }
 
+static const struct enetc_pf_ops enetc_pf_ops = {
+	.set_si_primary_mac = enetc_pf_set_primary_mac_addr,
+	.get_si_primary_mac = enetc_pf_get_primary_mac_addr,
+	.create_pcs = enetc_pf_create_pcs,
+	.destroy_pcs = enetc_pf_destroy_pcs,
+	.enable_psfp = enetc_psfp_enable,
+};
+
 static int enetc_pf_probe(struct pci_dev *pdev,
 			  const struct pci_device_id *ent)
 {
@@ -997,6 +1016,7 @@ static int enetc_pf_probe(struct pci_dev *pdev,
 	pf = enetc_si_priv(si);
 	pf->si = si;
 	pf->total_vfs = pci_sriov_get_totalvfs(pdev);
+	enetc_pf_ops_register(pf, &enetc_pf_ops);
 
 	err = enetc_setup_mac_addresses(node, pf);
 	if (err)
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.h b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
index 92a26b09cf57..39db9d5c2e50 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
@@ -28,6 +28,16 @@ struct enetc_vf_state {
 	enum enetc_vf_flags flags;
 };
 
+struct enetc_pf;
+
+struct enetc_pf_ops {
+	void (*set_si_primary_mac)(struct enetc_hw *hw, int si, const u8 *addr);
+	void (*get_si_primary_mac)(struct enetc_hw *hw, int si, u8 *addr);
+	struct phylink_pcs *(*create_pcs)(struct enetc_pf *pf, struct mii_bus *bus);
+	void (*destroy_pcs)(struct phylink_pcs *pcs);
+	int (*enable_psfp)(struct enetc_ndev_priv *priv);
+};
+
 struct enetc_pf {
 	struct enetc_si *si;
 	int num_vfs; /* number of active VFs, after sriov_init */
@@ -50,6 +60,8 @@ struct enetc_pf {
 
 	phy_interface_t if_mode;
 	struct phylink_config phylink_config;
+
+	const struct enetc_pf_ops *ops;
 };
 
 #define phylink_to_enetc_pf(config) \
@@ -59,9 +71,6 @@ int enetc_msg_psi_init(struct enetc_pf *pf);
 void enetc_msg_psi_free(struct enetc_pf *pf);
 void enetc_msg_handle_rxmsg(struct enetc_pf *pf, int mbox_id, u16 *status);
 
-void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr);
-void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
-				   const u8 *addr);
 int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr);
 int enetc_setup_mac_addresses(struct device_node *np, struct enetc_pf *pf);
 void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
@@ -71,3 +80,9 @@ void enetc_mdiobus_destroy(struct enetc_pf *pf);
 int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node *node,
 			 const struct phylink_mac_ops *ops);
 void enetc_phylink_destroy(struct enetc_ndev_priv *priv);
+
+static inline void enetc_pf_ops_register(struct enetc_pf *pf,
+					 const struct enetc_pf_ops *ops)
+{
+	pf->ops = ops;
+}
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
index bce81a4f6f88..94690ed92e3f 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
@@ -8,19 +8,37 @@
 
 #include "enetc_pf.h"
 
+static int enetc_set_si_hw_addr(struct enetc_pf *pf, int si, u8 *mac_addr)
+{
+	struct enetc_hw *hw = &pf->si->hw;
+
+	if (pf->ops->set_si_primary_mac)
+		pf->ops->set_si_primary_mac(hw, si, mac_addr);
+	else
+		return -EOPNOTSUPP;
+
+	return 0;
+}
+
 int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr)
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
+	struct enetc_pf *pf = enetc_si_priv(priv->si);
 	struct sockaddr *saddr = addr;
+	int err;
 
 	if (!is_valid_ether_addr(saddr->sa_data))
 		return -EADDRNOTAVAIL;
 
+	err = enetc_set_si_hw_addr(pf, 0, saddr->sa_data);
+	if (err)
+		return err;
+
 	eth_hw_addr_set(ndev, saddr->sa_data);
-	enetc_pf_set_primary_mac_addr(&priv->si->hw, 0, saddr->sa_data);
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(enetc_pf_set_mac_addr);
 
 static int enetc_setup_mac_address(struct device_node *np, struct enetc_pf *pf,
 				   int si)
@@ -38,8 +56,8 @@ static int enetc_setup_mac_address(struct device_node *np, struct enetc_pf *pf,
 	}
 
 	/* (2) bootloader supplied MAC address */
-	if (is_zero_ether_addr(mac_addr))
-		enetc_pf_get_primary_mac_addr(hw, si, mac_addr);
+	if (is_zero_ether_addr(mac_addr) && pf->ops->get_si_primary_mac)
+		pf->ops->get_si_primary_mac(hw, si, mac_addr);
 
 	/* (3) choose a random one */
 	if (is_zero_ether_addr(mac_addr)) {
@@ -48,7 +66,9 @@ static int enetc_setup_mac_address(struct device_node *np, struct enetc_pf *pf,
 			 si, mac_addr);
 	}
 
-	enetc_pf_set_primary_mac_addr(hw, si, mac_addr);
+	err = enetc_set_si_hw_addr(pf, si, mac_addr);
+	if (err)
+		return err;
 
 	return 0;
 }
@@ -70,11 +90,13 @@ int enetc_setup_mac_addresses(struct device_node *np, struct enetc_pf *pf)
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(enetc_setup_mac_addresses);
 
 void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
 			   const struct net_device_ops *ndev_ops)
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
+	struct enetc_pf *pf = enetc_si_priv(si);
 
 	SET_NETDEV_DEV(ndev, &si->pdev->dev);
 	priv->ndev = ndev;
@@ -107,7 +129,8 @@ void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
 			     NETDEV_XDP_ACT_NDO_XMIT | NETDEV_XDP_ACT_RX_SG |
 			     NETDEV_XDP_ACT_NDO_XMIT_SG;
 
-	if (si->hw_features & ENETC_SI_F_PSFP && !enetc_psfp_enable(priv)) {
+	if (si->hw_features & ENETC_SI_F_PSFP && pf->ops->enable_psfp &&
+	    !pf->ops->enable_psfp(priv)) {
 		priv->active_offloads |= ENETC_F_QCI;
 		ndev->features |= NETIF_F_HW_TC;
 		ndev->hw_features |= NETIF_F_HW_TC;
@@ -116,6 +139,7 @@ void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
 	/* pick up primary MAC address from SI */
 	enetc_load_primary_mac_addr(&si->hw, ndev);
 }
+EXPORT_SYMBOL_GPL(enetc_pf_netdev_setup);
 
 static int enetc_mdio_probe(struct enetc_pf *pf, struct device_node *np)
 {
@@ -162,6 +186,9 @@ static int enetc_imdio_create(struct enetc_pf *pf)
 	struct mii_bus *bus;
 	int err;
 
+	if (!pf->ops->create_pcs)
+		return -EOPNOTSUPP;
+
 	bus = mdiobus_alloc_size(sizeof(*mdio_priv));
 	if (!bus)
 		return -ENOMEM;
@@ -184,7 +211,7 @@ static int enetc_imdio_create(struct enetc_pf *pf)
 		goto free_mdio_bus;
 	}
 
-	phylink_pcs = lynx_pcs_create_mdiodev(bus, 0);
+	phylink_pcs = pf->ops->create_pcs(pf, bus);
 	if (IS_ERR(phylink_pcs)) {
 		err = PTR_ERR(phylink_pcs);
 		dev_err(dev, "cannot create lynx pcs (%d)\n", err);
@@ -205,8 +232,8 @@ static int enetc_imdio_create(struct enetc_pf *pf)
 
 static void enetc_imdio_remove(struct enetc_pf *pf)
 {
-	if (pf->pcs)
-		lynx_pcs_destroy(pf->pcs);
+	if (pf->pcs && pf->ops->destroy_pcs)
+		pf->ops->destroy_pcs(pf->pcs);
 
 	if (pf->imdio) {
 		mdiobus_unregister(pf->imdio);
@@ -246,12 +273,14 @@ int enetc_mdiobus_create(struct enetc_pf *pf, struct device_node *node)
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(enetc_mdiobus_create);
 
 void enetc_mdiobus_destroy(struct enetc_pf *pf)
 {
 	enetc_mdio_remove(pf);
 	enetc_imdio_remove(pf);
 }
+EXPORT_SYMBOL_GPL(enetc_mdiobus_destroy);
 
 int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node *node,
 			 const struct phylink_mac_ops *ops)
@@ -288,8 +317,13 @@ int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node *node,
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(enetc_phylink_create);
 
 void enetc_phylink_destroy(struct enetc_ndev_priv *priv)
 {
 	phylink_destroy(priv->phylink);
 }
+EXPORT_SYMBOL_GPL(enetc_phylink_destroy);
+
+MODULE_DESCRIPTION("NXP ENETC PF common functionality driver");
+MODULE_LICENSE("Dual BSD/GPL");
-- 
2.34.1

[PATCH v5 net-next 07/13] net: enetc: remove ERR050089 workaround for i.MX95

[Wei Fang]

From: Vladimir Oltean <vladimir.oltean@nxp.com>

The ERR050089 workaround causes performance degradation and potential
functional issues (e.g., RCU stalls) under certain workloads. Since
new SoCs like i.MX95 do not require this workaround, use a static key
to compile out enetc_lock_mdio() and enetc_unlock_mdio() at runtime,
improving performance and avoiding unnecessary logic.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
---
v5: no changes
---
 .../net/ethernet/freescale/enetc/enetc_hw.h   | 34 +++++++++++++------
 .../ethernet/freescale/enetc/enetc_pci_mdio.c | 17 ++++++++++
 2 files changed, 41 insertions(+), 10 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc_hw.h b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
index 1619943fb263..6a7b9b75d660 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_hw.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
@@ -396,18 +396,22 @@ struct enetc_hw {
  */
 extern rwlock_t enetc_mdio_lock;
 
+DECLARE_STATIC_KEY_FALSE(enetc_has_err050089);
+
 /* use this locking primitive only on the fast datapath to
  * group together multiple non-MDIO register accesses to
  * minimize the overhead of the lock
  */
 static inline void enetc_lock_mdio(void)
 {
-	read_lock(&enetc_mdio_lock);
+	if (static_branch_unlikely(&enetc_has_err050089))
+		read_lock(&enetc_mdio_lock);
 }
 
 static inline void enetc_unlock_mdio(void)
 {
-	read_unlock(&enetc_mdio_lock);
+	if (static_branch_unlikely(&enetc_has_err050089))
+		read_unlock(&enetc_mdio_lock);
 }
 
 /* use these accessors only on the fast datapath under
@@ -416,14 +420,16 @@ static inline void enetc_unlock_mdio(void)
  */
 static inline u32 enetc_rd_reg_hot(void __iomem *reg)
 {
-	lockdep_assert_held(&enetc_mdio_lock);
+	if (static_branch_unlikely(&enetc_has_err050089))
+		lockdep_assert_held(&enetc_mdio_lock);
 
 	return ioread32(reg);
 }
 
 static inline void enetc_wr_reg_hot(void __iomem *reg, u32 val)
 {
-	lockdep_assert_held(&enetc_mdio_lock);
+	if (static_branch_unlikely(&enetc_has_err050089))
+		lockdep_assert_held(&enetc_mdio_lock);
 
 	iowrite32(val, reg);
 }
@@ -452,9 +458,13 @@ static inline u32 _enetc_rd_mdio_reg_wa(void __iomem *reg)
 	unsigned long flags;
 	u32 val;
 
-	write_lock_irqsave(&enetc_mdio_lock, flags);
-	val = ioread32(reg);
-	write_unlock_irqrestore(&enetc_mdio_lock, flags);
+	if (static_branch_unlikely(&enetc_has_err050089)) {
+		write_lock_irqsave(&enetc_mdio_lock, flags);
+		val = ioread32(reg);
+		write_unlock_irqrestore(&enetc_mdio_lock, flags);
+	} else {
+		val = ioread32(reg);
+	}
 
 	return val;
 }
@@ -463,9 +473,13 @@ static inline void _enetc_wr_mdio_reg_wa(void __iomem *reg, u32 val)
 {
 	unsigned long flags;
 
-	write_lock_irqsave(&enetc_mdio_lock, flags);
-	iowrite32(val, reg);
-	write_unlock_irqrestore(&enetc_mdio_lock, flags);
+	if (static_branch_unlikely(&enetc_has_err050089)) {
+		write_lock_irqsave(&enetc_mdio_lock, flags);
+		iowrite32(val, reg);
+		write_unlock_irqrestore(&enetc_mdio_lock, flags);
+	} else {
+		iowrite32(val, reg);
+	}
 }
 
 #ifdef ioread64
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c b/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
index a1b595bd7993..2445e35a764a 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
@@ -9,6 +9,9 @@
 #define ENETC_MDIO_BUS_NAME	ENETC_MDIO_DEV_NAME " Bus"
 #define ENETC_MDIO_DRV_NAME	ENETC_MDIO_DEV_NAME " driver"
 
+DEFINE_STATIC_KEY_FALSE(enetc_has_err050089);
+EXPORT_SYMBOL_GPL(enetc_has_err050089);
+
 static int enetc_pci_mdio_probe(struct pci_dev *pdev,
 				const struct pci_device_id *ent)
 {
@@ -62,6 +65,12 @@ static int enetc_pci_mdio_probe(struct pci_dev *pdev,
 		goto err_pci_mem_reg;
 	}
 
+	if (pdev->vendor == PCI_VENDOR_ID_FREESCALE &&
+	    pdev->device == ENETC_MDIO_DEV_ID) {
+		static_branch_inc(&enetc_has_err050089);
+		dev_info(&pdev->dev, "Enabled ERR050089 workaround\n");
+	}
+
 	err = of_mdiobus_register(bus, dev->of_node);
 	if (err)
 		goto err_mdiobus_reg;
@@ -88,6 +97,14 @@ static void enetc_pci_mdio_remove(struct pci_dev *pdev)
 	struct enetc_mdio_priv *mdio_priv;
 
 	mdiobus_unregister(bus);
+
+	if (pdev->vendor == PCI_VENDOR_ID_FREESCALE &&
+	    pdev->device == ENETC_MDIO_DEV_ID) {
+		static_branch_dec(&enetc_has_err050089);
+		if (!static_key_enabled(&enetc_has_err050089.key))
+			dev_info(&pdev->dev, "Disabled ERR050089 workaround\n");
+	}
+
 	mdio_priv = bus->priv;
 	iounmap(mdio_priv->hw->port);
 	pci_release_region(pdev, 0);
-- 
2.34.1

[PATCH v5 net-next 08/13] PCI: Add NXP NETC vendor ID and device IDs

[Wei Fang]

NXP NETC is a multi-function RCiEP and it contains multiple functions,
such as EMDIO, PTP Timer, ENETC PF and VF. Therefore, add these device
IDs to pci_ids.h.

Below are the device IDs and corresponding drivers.
PCI_DEVICE_ID_NXP2_ENETC_PF: nxp-enetc4
PCI_DEVICE_ID_NXP2_NETC_EMDIO: fsl-enetc-mdio
PCI_DEVICE_ID_NXP2_NETC_TIMER: ptp_netc
PCI_DEVICE_ID_NXP2_ENETC_VF: fsl-enetc-vf

Signed-off-by: Wei Fang <wei.fang@nxp.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
---
v5: no changes
---
 include/linux/pci_ids.h | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h
index 4cf6aaed5f35..acd7ae774913 100644
--- a/include/linux/pci_ids.h
+++ b/include/linux/pci_ids.h
@@ -1556,6 +1556,13 @@
 #define PCI_DEVICE_ID_PHILIPS_SAA7146	0x7146
 #define PCI_DEVICE_ID_PHILIPS_SAA9730	0x9730
 
+/* NXP has two vendor IDs, the other one is 0x1957 */
+#define PCI_VENDOR_ID_NXP2		PCI_VENDOR_ID_PHILIPS
+#define PCI_DEVICE_ID_NXP2_ENETC_PF	0xe101
+#define PCI_DEVICE_ID_NXP2_NETC_EMDIO	0xee00
+#define PCI_DEVICE_ID_NXP2_NETC_TIMER	0xee02
+#define PCI_DEVICE_ID_NXP2_ENETC_VF	0xef00
+
 #define PCI_VENDOR_ID_EICON		0x1133
 #define PCI_DEVICE_ID_EICON_DIVA20	0xe002
 #define PCI_DEVICE_ID_EICON_DIVA20_U	0xe004
-- 
2.34.1

[PATCH v5 net-next 09/13] net: enetc: add i.MX95 EMDIO support

[Wei Fang]

The verdor ID and device ID of i.MX95 EMDIO are different from LS1028A
EMDIO, so add new vendor ID and device ID to pci_device_id table to
support i.MX95 EMDIO. And the i.MX95 EMDIO has two pins that need to be
controlled, namely MDC and MDIO.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
---
v5: no changes
---
 drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c b/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
index 2445e35a764a..9968a1e9b5ef 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
@@ -2,6 +2,7 @@
 /* Copyright 2019 NXP */
 #include <linux/fsl/enetc_mdio.h>
 #include <linux/of_mdio.h>
+#include <linux/pinctrl/consumer.h>
 #include "enetc_pf.h"
 
 #define ENETC_MDIO_DEV_ID	0xee01
@@ -71,6 +72,8 @@ static int enetc_pci_mdio_probe(struct pci_dev *pdev,
 		dev_info(&pdev->dev, "Enabled ERR050089 workaround\n");
 	}
 
+	pinctrl_pm_select_default_state(dev);
+
 	err = of_mdiobus_register(bus, dev->of_node);
 	if (err)
 		goto err_mdiobus_reg;
@@ -113,6 +116,7 @@ static void enetc_pci_mdio_remove(struct pci_dev *pdev)
 
 static const struct pci_device_id enetc_pci_mdio_id_table[] = {
 	{ PCI_DEVICE(PCI_VENDOR_ID_FREESCALE, ENETC_MDIO_DEV_ID) },
+	{ PCI_DEVICE(PCI_VENDOR_ID_NXP2, PCI_DEVICE_ID_NXP2_NETC_EMDIO) },
 	{ 0, } /* End of table. */
 };
 MODULE_DEVICE_TABLE(pci, enetc_pci_mdio_id_table);
-- 
2.34.1

[PATCH v5 net-next 10/13] net: enetc: extract enetc_int_vector_init/destroy() from enetc_alloc_msix()

[Wei Fang]

From: Clark Wang <xiaoning.wang@nxp.com>

Extract enetc_int_vector_init() and enetc_int_vector_destroy() from
enetc_alloc_msix() so that the code is more concise and readable. In
addition, slightly different from before, the cleanup helper function
is used to manage dynamically allocated memory resources.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
---
v5: no changes
---
 drivers/net/ethernet/freescale/enetc/enetc.c | 172 ++++++++++---------
 1 file changed, 87 insertions(+), 85 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
index 032d8eadd003..bd725561b8a2 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc.c
@@ -2965,6 +2965,87 @@ int enetc_ioctl(struct net_device *ndev, struct ifreq *rq, int cmd)
 }
 EXPORT_SYMBOL_GPL(enetc_ioctl);
 
+static int enetc_int_vector_init(struct enetc_ndev_priv *priv, int i,
+				 int v_tx_rings)
+{
+	struct enetc_int_vector *v __free(kfree);
+	struct enetc_bdr *bdr;
+	int j, err;
+
+	v = kzalloc(struct_size(v, tx_ring, v_tx_rings), GFP_KERNEL);
+	if (!v)
+		return -ENOMEM;
+
+	bdr = &v->rx_ring;
+	bdr->index = i;
+	bdr->ndev = priv->ndev;
+	bdr->dev = priv->dev;
+	bdr->bd_count = priv->rx_bd_count;
+	bdr->buffer_offset = ENETC_RXB_PAD;
+	priv->rx_ring[i] = bdr;
+
+	err = xdp_rxq_info_reg(&bdr->xdp.rxq, priv->ndev, i, 0);
+	if (err)
+		return err;
+
+	err = xdp_rxq_info_reg_mem_model(&bdr->xdp.rxq,
+					 MEM_TYPE_PAGE_SHARED, NULL);
+	if (err) {
+		xdp_rxq_info_unreg(&bdr->xdp.rxq);
+		return err;
+	}
+
+	/* init defaults for adaptive IC */
+	if (priv->ic_mode & ENETC_IC_RX_ADAPTIVE) {
+		v->rx_ictt = 0x1;
+		v->rx_dim_en = true;
+	}
+
+	INIT_WORK(&v->rx_dim.work, enetc_rx_dim_work);
+	netif_napi_add(priv->ndev, &v->napi, enetc_poll);
+	v->count_tx_rings = v_tx_rings;
+
+	for (j = 0; j < v_tx_rings; j++) {
+		int idx;
+
+		/* default tx ring mapping policy */
+		idx = priv->bdr_int_num * j + i;
+		__set_bit(idx, &v->tx_rings_map);
+		bdr = &v->tx_ring[j];
+		bdr->index = idx;
+		bdr->ndev = priv->ndev;
+		bdr->dev = priv->dev;
+		bdr->bd_count = priv->tx_bd_count;
+		priv->tx_ring[idx] = bdr;
+	}
+
+	priv->int_vector[i] = no_free_ptr(v);
+
+	return 0;
+}
+
+static void enetc_int_vector_destroy(struct enetc_ndev_priv *priv, int i)
+{
+	struct enetc_int_vector *v = priv->int_vector[i];
+	struct enetc_bdr *rx_ring = &v->rx_ring;
+	int j, tx_ring_index;
+
+	xdp_rxq_info_unreg_mem_model(&rx_ring->xdp.rxq);
+	xdp_rxq_info_unreg(&rx_ring->xdp.rxq);
+	netif_napi_del(&v->napi);
+	cancel_work_sync(&v->rx_dim.work);
+
+	priv->rx_ring[i] = NULL;
+
+	for (j = 0; j < v->count_tx_rings; j++) {
+		tx_ring_index = priv->bdr_int_num * j + i;
+		priv->tx_ring[tx_ring_index] = NULL;
+	}
+
+	kfree(v);
+	priv->int_vector[i] = NULL;
+}
+
 int enetc_alloc_msix(struct enetc_ndev_priv *priv)
 {
 	struct pci_dev *pdev = priv->si->pdev;
@@ -2987,62 +3068,9 @@ int enetc_alloc_msix(struct enetc_ndev_priv *priv)
 	v_tx_rings = priv->num_tx_rings / priv->bdr_int_num;
 
 	for (i = 0; i < priv->bdr_int_num; i++) {
-		struct enetc_int_vector *v;
-		struct enetc_bdr *bdr;
-		int j;
-
-		v = kzalloc(struct_size(v, tx_ring, v_tx_rings), GFP_KERNEL);
-		if (!v) {
-			err = -ENOMEM;
-			goto fail;
-		}
-
-		priv->int_vector[i] = v;
-
-		bdr = &v->rx_ring;
-		bdr->index = i;
-		bdr->ndev = priv->ndev;
-		bdr->dev = priv->dev;
-		bdr->bd_count = priv->rx_bd_count;
-		bdr->buffer_offset = ENETC_RXB_PAD;
-		priv->rx_ring[i] = bdr;
-
-		err = xdp_rxq_info_reg(&bdr->xdp.rxq, priv->ndev, i, 0);
-		if (err) {
-			kfree(v);
-			goto fail;
-		}
-
-		err = xdp_rxq_info_reg_mem_model(&bdr->xdp.rxq,
-						 MEM_TYPE_PAGE_SHARED, NULL);
-		if (err) {
-			xdp_rxq_info_unreg(&bdr->xdp.rxq);
-			kfree(v);
+		err = enetc_int_vector_init(priv, i, v_tx_rings);
+		if (err)
 			goto fail;
-		}
-
-		/* init defaults for adaptive IC */
-		if (priv->ic_mode & ENETC_IC_RX_ADAPTIVE) {
-			v->rx_ictt = 0x1;
-			v->rx_dim_en = true;
-		}
-		INIT_WORK(&v->rx_dim.work, enetc_rx_dim_work);
-		netif_napi_add(priv->ndev, &v->napi, enetc_poll);
-		v->count_tx_rings = v_tx_rings;
-
-		for (j = 0; j < v_tx_rings; j++) {
-			int idx;
-
-			/* default tx ring mapping policy */
-			idx = priv->bdr_int_num * j + i;
-			__set_bit(idx, &v->tx_rings_map);
-			bdr = &v->tx_ring[j];
-			bdr->index = idx;
-			bdr->ndev = priv->ndev;
-			bdr->dev = priv->dev;
-			bdr->bd_count = priv->tx_bd_count;
-			priv->tx_ring[idx] = bdr;
-		}
 	}
 
 	num_stack_tx_queues = enetc_num_stack_tx_queues(priv);
@@ -3062,16 +3090,8 @@ int enetc_alloc_msix(struct enetc_ndev_priv *priv)
 	return 0;
 
 fail:
-	while (i--) {
-		struct enetc_int_vector *v = priv->int_vector[i];
-		struct enetc_bdr *rx_ring = &v->rx_ring;
-
-		xdp_rxq_info_unreg_mem_model(&rx_ring->xdp.rxq);
-		xdp_rxq_info_unreg(&rx_ring->xdp.rxq);
-		netif_napi_del(&v->napi);
-		cancel_work_sync(&v->rx_dim.work);
-		kfree(v);
-	}
+	while (i--)
+		enetc_int_vector_destroy(priv, i);
 
 	pci_free_irq_vectors(pdev);
 
@@ -3083,26 +3103,8 @@ void enetc_free_msix(struct enetc_ndev_priv *priv)
 {
 	int i;
 
-	for (i = 0; i < priv->bdr_int_num; i++) {
-		struct enetc_int_vector *v = priv->int_vector[i];
-		struct enetc_bdr *rx_ring = &v->rx_ring;
-
-		xdp_rxq_info_unreg_mem_model(&rx_ring->xdp.rxq);
-		xdp_rxq_info_unreg(&rx_ring->xdp.rxq);
-		netif_napi_del(&v->napi);
-		cancel_work_sync(&v->rx_dim.work);
-	}
-
-	for (i = 0; i < priv->num_rx_rings; i++)
-		priv->rx_ring[i] = NULL;
-
-	for (i = 0; i < priv->num_tx_rings; i++)
-		priv->tx_ring[i] = NULL;
-
-	for (i = 0; i < priv->bdr_int_num; i++) {
-		kfree(priv->int_vector[i]);
-		priv->int_vector[i] = NULL;
-	}
+	for (i = 0; i < priv->bdr_int_num; i++)
+		enetc_int_vector_destroy(priv, i);
 
 	/* disable all MSIX for this device */
 	pci_free_irq_vectors(priv->si->pdev);
-- 
2.34.1

[PATCH v5 net-next 11/13] net: enetc: optimize the allocation of tx_bdr

[Wei Fang]

From: Clark Wang <xiaoning.wang@nxp.com>

There is a situation where num_tx_rings cannot be divided by bdr_int_num.
For example, num_tx_rings is 8 and bdr_int_num is 3. According to the
previous logic, this results in two tx_bdr corresponding memories not
being allocated, so when sending packets to tx ring 6 or 7, wild pointers
will be accessed. Of course, this issue doesn't exist on LS1028A, because
its num_tx_rings is 8, and bdr_int_num is either 1 or 2. However, there
is a risk for the upcoming i.MX95. Therefore, it is necessary to ensure
that each tx_bdr can be allocated to the corresponding memory.

Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
Signed-off-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
---
v5: no changes
---
 drivers/net/ethernet/freescale/enetc/enetc.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
index bd725561b8a2..bccbeb1f355c 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc.c
@@ -3049,10 +3049,10 @@ static void enetc_int_vector_destroy(struct enetc_ndev_priv *priv, int i)
 int enetc_alloc_msix(struct enetc_ndev_priv *priv)
 {
 	struct pci_dev *pdev = priv->si->pdev;
+	int v_tx_rings, v_remainder;
 	int num_stack_tx_queues;
 	int first_xdp_tx_ring;
 	int i, n, err, nvec;
-	int v_tx_rings;
 
 	nvec = ENETC_BDR_INT_BASE_IDX + priv->bdr_int_num;
 	/* allocate MSIX for both messaging and Rx/Tx interrupts */
@@ -3066,9 +3066,12 @@ int enetc_alloc_msix(struct enetc_ndev_priv *priv)
 
 	/* # of tx rings per int vector */
 	v_tx_rings = priv->num_tx_rings / priv->bdr_int_num;
+	v_remainder = priv->num_tx_rings % priv->bdr_int_num;
 
 	for (i = 0; i < priv->bdr_int_num; i++) {
-		err = enetc_int_vector_init(priv, i, v_tx_rings);
+		int num_tx_rings = i < v_remainder ? v_tx_rings + 1 : v_tx_rings;
+
+		err = enetc_int_vector_init(priv, i, num_tx_rings);
 		if (err)
 			goto fail;
 	}
-- 
2.34.1

[PATCH v5 net-next 12/13] net: enetc: add preliminary support for i.MX95 ENETC PF

[Wei Fang]

The i.MX95 ENETC has been upgraded to revision 4.1, which is different
from the LS1028A ENETC (revision 1.0) except for the SI part. Therefore,
the fsl-enetc driver is incompatible with i.MX95 ENETC PF. So add new
nxp-enetc4 driver to support i.MX95 ENETC PF, and this driver will be
used to support the ENETC PF with major revision 4 for other SoCs in the
future.

Currently, the nxp-enetc4 driver only supports basic transmission feature
for i.MX95 ENETC PF, the more basic and advanced features will be added
in the subsequent patches.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
v2 changes:
1. Refine the commit message.
2. Sort the header files.
3. Use dev_err_probe() in enetc4_pf_probe().
4. Remove unused variable 'pf' from enetc4_pf_remove().
v3 changes:
1. Remove is_enetc_rev1() from enetc_init_si_rings_params().
2. Use devm_add_action_or_reset() in enetc4_pf_probe().
3. Directly return dev_err_probe() in enetc4_pf_probe().
v4 changes:
1. Move clk_freq from enect_si to enetc_ndev_priv.
2. Remove is_enetc_rev4().
3. Add enetc4_pf_ethtool_ops for i.MX95 ENETC PF.
v5 changes:
1. Add driver data support, some information such as system clock
frequency, pMAC offset and so on are defined in driver data.
2. Add some macros to replace numbers in the driver
---
 drivers/net/ethernet/freescale/enetc/Kconfig  |  17 +
 drivers/net/ethernet/freescale/enetc/Makefile |   3 +
 drivers/net/ethernet/freescale/enetc/enetc.c  |  86 +-
 drivers/net/ethernet/freescale/enetc/enetc.h  |  30 +-
 .../net/ethernet/freescale/enetc/enetc4_hw.h  | 152 ++++
 .../net/ethernet/freescale/enetc/enetc4_pf.c  | 759 ++++++++++++++++++
 .../ethernet/freescale/enetc/enetc_ethtool.c  |  35 +-
 .../net/ethernet/freescale/enetc/enetc_hw.h   |  19 +-
 .../net/ethernet/freescale/enetc/enetc_pf.c   |   7 +
 .../net/ethernet/freescale/enetc/enetc_pf.h   |  14 +
 .../freescale/enetc/enetc_pf_common.c         |  11 +-
 .../net/ethernet/freescale/enetc/enetc_qos.c  |   2 +-
 .../net/ethernet/freescale/enetc/enetc_vf.c   |   6 +
 13 files changed, 1113 insertions(+), 28 deletions(-)
 create mode 100644 drivers/net/ethernet/freescale/enetc/enetc4_hw.h
 create mode 100644 drivers/net/ethernet/freescale/enetc/enetc4_pf.c

diff --git a/drivers/net/ethernet/freescale/enetc/Kconfig b/drivers/net/ethernet/freescale/enetc/Kconfig
index e1b151a98b41..6c2779047dcd 100644
--- a/drivers/net/ethernet/freescale/enetc/Kconfig
+++ b/drivers/net/ethernet/freescale/enetc/Kconfig
@@ -33,6 +33,23 @@ config FSL_ENETC
 
 	  If compiled as module (M), the module name is fsl-enetc.
 
+config NXP_ENETC4
+	tristate "ENETC4 PF driver"
+	depends on PCI_MSI
+	select MDIO_DEVRES
+	select FSL_ENETC_CORE
+	select FSL_ENETC_MDIO
+	select NXP_ENETC_PF_COMMON
+	select PHYLINK
+	select DIMLIB
+	help
+	  This driver supports NXP ENETC devices with major revision 4. ENETC is
+	  as the NIC functionality in NETC, it supports virtualization/isolation
+	  based on PCIe Single Root IO Virtualization (SR-IOV) and a full range
+	  of TSN standards and NIC offload capabilities.
+
+	  If compiled as module (M), the module name is nxp-enetc4.
+
 config FSL_ENETC_VF
 	tristate "ENETC VF driver"
 	depends on PCI_MSI
diff --git a/drivers/net/ethernet/freescale/enetc/Makefile b/drivers/net/ethernet/freescale/enetc/Makefile
index ebe232673ed4..6fd27ee4fcd1 100644
--- a/drivers/net/ethernet/freescale/enetc/Makefile
+++ b/drivers/net/ethernet/freescale/enetc/Makefile
@@ -11,6 +11,9 @@ fsl-enetc-y := enetc_pf.o
 fsl-enetc-$(CONFIG_PCI_IOV) += enetc_msg.o
 fsl-enetc-$(CONFIG_FSL_ENETC_QOS) += enetc_qos.o
 
+obj-$(CONFIG_NXP_ENETC4) += nxp-enetc4.o
+nxp-enetc4-y := enetc4_pf.o
+
 obj-$(CONFIG_FSL_ENETC_VF) += fsl-enetc-vf.o
 fsl-enetc-vf-y := enetc_vf.o
 
diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
index bccbeb1f355c..34b2f06b34c9 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc.c
@@ -3,6 +3,7 @@
 
 #include "enetc.h"
 #include <linux/bpf_trace.h>
+#include <linux/clk.h>
 #include <linux/tcp.h>
 #include <linux/udp.h>
 #include <linux/vmalloc.h>
@@ -21,7 +22,7 @@ void enetc_port_mac_wr(struct enetc_si *si, u32 reg, u32 val)
 {
 	enetc_port_wr(&si->hw, reg, val);
 	if (si->hw_features & ENETC_SI_F_QBU)
-		enetc_port_wr(&si->hw, reg + ENETC_PMAC_OFFSET, val);
+		enetc_port_wr(&si->hw, reg + si->drvdata->pmac_offset, val);
 }
 EXPORT_SYMBOL_GPL(enetc_port_mac_wr);
 
@@ -700,8 +701,9 @@ static void enetc_rx_dim_work(struct work_struct *w)
 		net_dim_get_rx_moderation(dim->mode, dim->profile_ix);
 	struct enetc_int_vector	*v =
 		container_of(dim, struct enetc_int_vector, rx_dim);
+	struct enetc_ndev_priv *priv = netdev_priv(v->rx_ring.ndev);
 
-	v->rx_ictt = enetc_usecs_to_cycles(moder.usec);
+	v->rx_ictt = enetc_usecs_to_cycles(moder.usec, priv->sysclk_freq);
 	dim->state = DIM_START_MEASURE;
 }
 
@@ -1726,9 +1728,15 @@ void enetc_get_si_caps(struct enetc_si *si)
 	si->num_rx_rings = (val >> 16) & 0xff;
 	si->num_tx_rings = val & 0xff;
 
-	val = enetc_rd(hw, ENETC_SIRFSCAPR);
-	si->num_fs_entries = ENETC_SIRFSCAPR_GET_NUM_RFS(val);
-	si->num_fs_entries = min(si->num_fs_entries, ENETC_MAX_RFS_SIZE);
+	val = enetc_rd(hw, ENETC_SIPCAPR0);
+	if (val & ENETC_SIPCAPR0_RFS) {
+		val = enetc_rd(hw, ENETC_SIRFSCAPR);
+		si->num_fs_entries = ENETC_SIRFSCAPR_GET_NUM_RFS(val);
+		si->num_fs_entries = min(si->num_fs_entries, ENETC_MAX_RFS_SIZE);
+	} else {
+		/* ENETC which not supports RFS */
+		si->num_fs_entries = 0;
+	}
 
 	si->num_rss = 0;
 	val = enetc_rd(hw, ENETC_SIPCAPR0);
@@ -2056,7 +2064,10 @@ int enetc_configure_si(struct enetc_ndev_priv *priv)
 	/* enable SI */
 	enetc_wr(hw, ENETC_SIMR, ENETC_SIMR_EN);
 
-	if (si->num_rss) {
+	/* TODO: RSS support for i.MX95 will be supported later, and the
+	 * is_enetc_rev1() condition will be removed
+	 */
+	if (si->num_rss && is_enetc_rev1(si)) {
 		err = enetc_setup_default_rss_table(si, priv->num_rx_rings);
 		if (err)
 			return err;
@@ -2080,9 +2091,9 @@ void enetc_init_si_rings_params(struct enetc_ndev_priv *priv)
 	 */
 	priv->num_rx_rings = min_t(int, cpus, si->num_rx_rings);
 	priv->num_tx_rings = si->num_tx_rings;
-	priv->bdr_int_num = cpus;
+	priv->bdr_int_num = priv->num_rx_rings;
 	priv->ic_mode = ENETC_IC_RX_ADAPTIVE | ENETC_IC_TX_MANUAL;
-	priv->tx_ictt = ENETC_TXIC_TIMETHR;
+	priv->tx_ictt = enetc_usecs_to_cycles(600, priv->sysclk_freq);
 }
 EXPORT_SYMBOL_GPL(enetc_init_si_rings_params);
 
@@ -2475,10 +2486,14 @@ int enetc_open(struct net_device *ndev)
 
 	extended = !!(priv->active_offloads & ENETC_F_RX_TSTAMP);
 
-	err = enetc_setup_irqs(priv);
+	err = clk_prepare_enable(priv->ref_clk);
 	if (err)
 		return err;
 
+	err = enetc_setup_irqs(priv);
+	if (err)
+		goto err_setup_irqs;
+
 	err = enetc_phylink_connect(ndev);
 	if (err)
 		goto err_phy_connect;
@@ -2510,6 +2525,8 @@ int enetc_open(struct net_device *ndev)
 		phylink_disconnect_phy(priv->phylink);
 err_phy_connect:
 	enetc_free_irqs(priv);
+err_setup_irqs:
+	clk_disable_unprepare(priv->ref_clk);
 
 	return err;
 }
@@ -2559,6 +2576,7 @@ int enetc_close(struct net_device *ndev)
 	enetc_assign_tx_resources(priv, NULL);
 
 	enetc_free_irqs(priv);
+	clk_disable_unprepare(priv->ref_clk);
 
 	return 0;
 }
@@ -3216,5 +3234,55 @@ void enetc_pci_remove(struct pci_dev *pdev)
 }
 EXPORT_SYMBOL_GPL(enetc_pci_remove);
 
+static const struct enetc_drvdata enetc_pf_data = {
+	.sysclk_freq = ENETC_CLK_400M,
+	.pmac_offset = ENETC_PMAC_OFFSET,
+	.eth_ops = &enetc_pf_ethtool_ops,
+};
+
+static const struct enetc_drvdata enetc4_pf_data = {
+	.sysclk_freq = ENETC_CLK_333M,
+	.pmac_offset = ENETC4_PMAC_OFFSET,
+	.eth_ops = &enetc4_pf_ethtool_ops,
+};
+
+static const struct enetc_drvdata enetc_vf_data = {
+	.sysclk_freq = ENETC_CLK_400M,
+	.eth_ops = &enetc_vf_ethtool_ops,
+};
+
+static const struct enetc_platform_info enetc_info[] = {
+	{ .revision = ENETC_REV_1_0,
+	  .dev_id = ENETC_DEV_ID_PF,
+	  .data = &enetc_pf_data,
+	},
+	{ .revision = ENETC_REV_4_1,
+	  .dev_id = PCI_DEVICE_ID_NXP2_ENETC_PF,
+	  .data = &enetc4_pf_data,
+	},
+	{ .revision = ENETC_REV_1_0,
+	  .dev_id = ENETC_DEV_ID_VF,
+	  .data = &enetc_vf_data,
+	},
+};
+
+int enetc_get_driver_data(struct enetc_si *si)
+{
+	u16 dev_id = si->pdev->device;
+	int i;
+
+	for (i = 0; i < ARRAY_SIZE(enetc_info); i++) {
+		if (si->revision == enetc_info[i].revision &&
+		    dev_id == enetc_info[i].dev_id) {
+			si->drvdata = enetc_info[i].data;
+
+			return 0;
+		}
+	}
+
+	return -ERANGE;
+}
+EXPORT_SYMBOL_GPL(enetc_get_driver_data);
+
 MODULE_DESCRIPTION("NXP ENETC Ethernet driver");
 MODULE_LICENSE("Dual BSD/GPL");
diff --git a/drivers/net/ethernet/freescale/enetc/enetc.h b/drivers/net/ethernet/freescale/enetc/enetc.h
index 97524dfa234c..3d7d49dcdbf4 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc.h
@@ -14,6 +14,7 @@
 #include <net/xdp.h>
 
 #include "enetc_hw.h"
+#include "enetc4_hw.h"
 
 #define ENETC_MAC_MAXFRM_SIZE	9600
 #define ENETC_MAX_MTU		(ENETC_MAC_MAXFRM_SIZE - \
@@ -231,6 +232,18 @@ enum enetc_errata {
 #define ENETC_SI_F_QBV  BIT(1)
 #define ENETC_SI_F_QBU  BIT(2)
 
+struct enetc_drvdata {
+	u32 pmac_offset; /* Only valid for PSI which supports 802.1Qbu */
+	u64 sysclk_freq;
+	const struct ethtool_ops *eth_ops;
+};
+
+struct enetc_platform_info {
+	u16 revision;
+	u16 dev_id;
+	const struct enetc_drvdata *data;
+};
+
 /* PCI IEP device data */
 struct enetc_si {
 	struct pci_dev *pdev;
@@ -246,11 +259,18 @@ struct enetc_si {
 	int num_fs_entries;
 	int num_rss; /* number of RSS buckets */
 	unsigned short pad;
+	u16 revision;
 	int hw_features;
+	const struct enetc_drvdata *drvdata;
 };
 
 #define ENETC_SI_ALIGN	32
 
+static inline bool is_enetc_rev1(struct enetc_si *si)
+{
+	return si->pdev->revision == ENETC_REV1;
+}
+
 static inline void *enetc_si_priv(const struct enetc_si *si)
 {
 	return (char *)si + ALIGN(sizeof(struct enetc_si), ENETC_SI_ALIGN);
@@ -302,7 +322,7 @@ struct enetc_cls_rule {
 	int used;
 };
 
-#define ENETC_MAX_BDR_INT	2 /* fixed to max # of available cpus */
+#define ENETC_MAX_BDR_INT	6 /* fixed to max # of available cpus */
 struct psfp_cap {
 	u32 max_streamid;
 	u32 max_psfp_filter;
@@ -340,7 +360,6 @@ enum enetc_ic_mode {
 
 #define ENETC_RXIC_PKTTHR	min_t(u32, 256, ENETC_RX_RING_DEFAULT_SIZE / 2)
 #define ENETC_TXIC_PKTTHR	min_t(u32, 128, ENETC_TX_RING_DEFAULT_SIZE / 2)
-#define ENETC_TXIC_TIMETHR	enetc_usecs_to_cycles(600)
 
 struct enetc_ndev_priv {
 	struct net_device *ndev;
@@ -388,6 +407,9 @@ struct enetc_ndev_priv {
 	 * and link state updates
 	 */
 	struct mutex		mm_lock;
+
+	struct clk *ref_clk; /* RGMII/RMII reference clock */
+	u64 sysclk_freq; /* NETC system clock frequency */
 };
 
 /* Messaging */
@@ -417,6 +439,7 @@ void enetc_init_si_rings_params(struct enetc_ndev_priv *priv);
 int enetc_alloc_si_resources(struct enetc_ndev_priv *priv);
 void enetc_free_si_resources(struct enetc_ndev_priv *priv);
 int enetc_configure_si(struct enetc_ndev_priv *priv);
+int enetc_get_driver_data(struct enetc_si *si);
 
 int enetc_open(struct net_device *ndev);
 int enetc_close(struct net_device *ndev);
@@ -433,6 +456,9 @@ int enetc_xdp_xmit(struct net_device *ndev, int num_frames,
 		   struct xdp_frame **frames, u32 flags);
 
 /* ethtool */
+extern const struct ethtool_ops enetc_pf_ethtool_ops;
+extern const struct ethtool_ops enetc4_pf_ethtool_ops;
+extern const struct ethtool_ops enetc_vf_ethtool_ops;
 void enetc_set_ethtool_ops(struct net_device *ndev);
 void enetc_mm_link_state_update(struct enetc_ndev_priv *priv, bool link);
 void enetc_mm_commit_preemptible_tcs(struct enetc_ndev_priv *priv);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc4_hw.h b/drivers/net/ethernet/freescale/enetc/enetc4_hw.h
new file mode 100644
index 000000000000..acaa51c6b891
--- /dev/null
+++ b/drivers/net/ethernet/freescale/enetc/enetc4_hw.h
@@ -0,0 +1,152 @@
+/* SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause) */
+/*
+ * This header file defines the register offsets and bit fields
+ * of ENETC4 PF and VFs. Note that the same registers as ENETC
+ * version 1.0 are defined in the enetc_hw.h file.
+ *
+ * Copyright 2024 NXP
+ */
+#ifndef __ENETC4_HW_H_
+#define __ENETC4_HW_H_
+
+/***************************ENETC port registers**************************/
+#define ENETC4_ECAPR0			0x0
+#define  ECAPR0_RFS			BIT(2)
+#define  ECAPR0_TSD			BIT(5)
+#define  ECAPR0_RSS			BIT(8)
+#define  ECAPR0_RSC			BIT(9)
+#define  ECAPR0_LSO			BIT(10)
+#define  ECAPR0_WO			BIT(13)
+
+#define ENETC4_ECAPR1			0x4
+#define  ECAPR1_NUM_TCS			GENMASK(6, 4)
+#define  ECAPR1_NUM_MCH			GENMASK(9, 8)
+#define  ECAPR1_NUM_UCH			GENMASK(11, 10)
+#define  ECAPR1_NUM_MSIX		GENMASK(22, 12)
+#define  ECAPR1_NUM_VSI			GENMASK(27, 24)
+#define  ECAPR1_NUM_IPV			BIT(31)
+
+#define ENETC4_ECAPR2			0x8
+#define  ECAPR2_NUM_TX_BDR		GENMASK(9, 0)
+#define  ECAPR2_NUM_RX_BDR		GENMASK(25, 16)
+
+#define ENETC4_PMR			0x10
+#define  PMR_SI_EN(a)			BIT((16 + (a)))
+
+/* Port Pause ON/OFF threshold register */
+#define ENETC4_PPAUONTR			0x108
+#define ENETC4_PPAUOFFTR		0x10c
+
+/* Port Station interface promiscuous MAC mode register */
+#define ENETC4_PSIPMMR			0x200
+#define  PSIPMMR_SI_MAC_UP(a)		BIT(a) /* a = SI index */
+#define  PSIPMMR_SI_MAC_MP(a)		BIT((a) + 16)
+
+/* Port Station interface promiscuous VLAN mode register */
+#define ENETC4_PSIPVMR			0x204
+
+/* Port RSS key register n. n = 0,1,2,...,9 */
+#define ENETC4_PRSSKR(n)		((n) * 0x4 + 0x250)
+
+/* Port station interface MAC address filtering capability register */
+#define ENETC4_PSIMAFCAPR		0x280
+#define  PSIMAFCAPR_NUM_MAC_AFTE	GENMASK(11, 0)
+
+/* Port station interface VLAN filtering capability register */
+#define ENETC4_PSIVLANFCAPR		0x2c0
+#define  PSIVLANFCAPR_NUM_VLAN_FTE	GENMASK(11, 0)
+
+/* Port station interface VLAN filtering mode register */
+#define ENETC4_PSIVLANFMR		0x2c4
+#define  PSIVLANFMR_VS			BIT(0)
+
+/* Port Station interface a primary MAC address registers */
+#define ENETC4_PSIPMAR0(a)		((a) * 0x80 + 0x2000)
+#define ENETC4_PSIPMAR1(a)		((a) * 0x80 + 0x2004)
+
+/* Port station interface a configuration register 0/2 */
+#define ENETC4_PSICFGR0(a)		((a) * 0x80 + 0x2010)
+#define  PSICFGR0_VASE			BIT(13)
+#define  PSICFGR0_ASE			BIT(15)
+#define  PSICFGR0_ANTI_SPOOFING		(PSICFGR0_VASE | PSICFGR0_ASE)
+
+#define ENETC4_PSICFGR2(a)		((a) * 0x80 + 0x2018)
+#define  PSICFGR2_NUM_MSIX		GENMASK(5, 0)
+
+#define ENETC4_PMCAPR			0x4004
+#define  PMCAPR_HD			BIT(8)
+#define  PMCAPR_FP			GENMASK(10, 9)
+
+/* Port configuration register */
+#define ENETC4_PCR			0x4010
+#define  PCR_HDR_FMT			BIT(0)
+#define  PCR_L2DOSE			BIT(4)
+#define  PCR_TIMER_CS			BIT(8)
+#define  PCR_PSPEED			GENMASK(29, 16)
+#define  PCR_PSPEED_VAL(speed)		(((speed) / 10 - 1) << 16)
+
+/* Port MAC address register 0/1 */
+#define ENETC4_PMAR0			0x4020
+#define ENETC4_PMAR1			0x4024
+
+/* Port operational register */
+#define ENETC4_POR			0x4100
+
+/* Port traffic class a transmit maximum SDU register */
+#define ENETC4_PTCTMSDUR(a)		((a) * 0x20 + 0x4208)
+#define  PTCTMSDUR_MAXSDU		GENMASK(15, 0)
+#define  PTCTMSDUR_SDU_TYPE		GENMASK(17, 16)
+#define   SDU_TYPE_PPDU			0
+#define   SDU_TYPE_MPDU			1
+#define   SDU_TYPE_MSDU			2
+
+#define ENETC4_PMAC_OFFSET		0x400
+#define ENETC4_PM_CMD_CFG(mac)		(0x5008 + (mac) * 0x400)
+#define  PM_CMD_CFG_TX_EN		BIT(0)
+#define  PM_CMD_CFG_RX_EN		BIT(1)
+#define  PM_CMD_CFG_PAUSE_FWD		BIT(7)
+#define  PM_CMD_CFG_PAUSE_IGN		BIT(8)
+#define  PM_CMD_CFG_TX_ADDR_INS		BIT(9)
+#define  PM_CMD_CFG_LOOP_EN		BIT(10)
+#define  PM_CMD_CFG_LPBK_MODE		GENMASK(12, 11)
+#define   LPBCK_MODE_EXT_TX_CLK		0
+#define   LPBCK_MODE_MAC_LEVEL		1
+#define   LPBCK_MODE_INT_TX_CLK		2
+#define  PM_CMD_CFG_CNT_FRM_EN		BIT(13)
+#define  PM_CMD_CFG_TXP			BIT(15)
+#define  PM_CMD_CFG_SEND_IDLE		BIT(16)
+#define  PM_CMD_CFG_HD_FCEN		BIT(18)
+#define  PM_CMD_CFG_SFD			BIT(21)
+#define  PM_CMD_CFG_TX_FLUSH		BIT(22)
+#define  PM_CMD_CFG_TX_LOWP_EN		BIT(23)
+#define  PM_CMD_CFG_RX_LOWP_EMPTY	BIT(24)
+#define  PM_CMD_CFG_SWR			BIT(26)
+#define  PM_CMD_CFG_TS_MODE		BIT(30)
+#define  PM_CMD_CFG_MG			BIT(31)
+
+/* Port MAC 0/1 Maximum Frame Length Register */
+#define ENETC4_PM_MAXFRM(mac)		(0x5014 + (mac) * 0x400)
+
+/* Port MAC 0/1 Pause Quanta Register */
+#define ENETC4_PM_PAUSE_QUANTA(mac)	(0x5054 + (mac) * 0x400)
+
+/* Port MAC 0/1 Pause Quanta Threshold Register */
+#define ENETC4_PM_PAUSE_THRESH(mac)	(0x5064 + (mac) * 0x400)
+
+/* Port MAC 0 Interface Mode Control Register */
+#define ENETC4_PM_IF_MODE(mac)		(0x5300 + (mac) * 0x400)
+#define  PM_IF_MODE_IFMODE		GENMASK(2, 0)
+#define   IFMODE_XGMII			0
+#define   IFMODE_RMII			3
+#define   IFMODE_RGMII			4
+#define   IFMODE_SGMII			5
+#define  PM_IF_MODE_REVMII		BIT(3)
+#define  PM_IF_MODE_M10			BIT(4)
+#define  PM_IF_MODE_HD			BIT(6)
+#define  PM_IF_MODE_SSP			GENMASK(14, 13)
+#define   SSP_100M			0
+#define   SSP_10M			1
+#define   SSP_1G			2
+#define  PM_IF_MODE_ENA			BIT(15)
+
+#endif
diff --git a/drivers/net/ethernet/freescale/enetc/enetc4_pf.c b/drivers/net/ethernet/freescale/enetc/enetc4_pf.c
new file mode 100644
index 000000000000..9562caba59b1
--- /dev/null
+++ b/drivers/net/ethernet/freescale/enetc/enetc4_pf.c
@@ -0,0 +1,759 @@
+// SPDX-License-Identifier: (GPL-2.0+ OR BSD-3-Clause)
+/* Copyright 2024 NXP */
+
+#include <linux/clk.h>
+#include <linux/fsl/netc_global.h>
+#include <linux/module.h>
+#include <linux/of_net.h>
+#include <linux/of_platform.h>
+#include <linux/pinctrl/consumer.h>
+#include <linux/unaligned.h>
+
+#include "enetc_pf.h"
+
+#define ENETC_SI_MAX_RING_NUM	8
+
+static void enetc4_get_port_caps(struct enetc_pf *pf)
+{
+	struct enetc_hw *hw = &pf->si->hw;
+	u32 val;
+
+	val = enetc_port_rd(hw, ENETC4_ECAPR1);
+	pf->caps.num_vsi = (val & ECAPR1_NUM_VSI) >> 24;
+	pf->caps.num_msix = ((val & ECAPR1_NUM_MSIX) >> 12) + 1;
+
+	val = enetc_port_rd(hw, ENETC4_ECAPR2);
+	pf->caps.num_rx_bdr = (val & ECAPR2_NUM_RX_BDR) >> 16;
+	pf->caps.num_tx_bdr = val & ECAPR2_NUM_TX_BDR;
+
+	val = enetc_port_rd(hw, ENETC4_PMCAPR);
+	pf->caps.half_duplex = (val & PMCAPR_HD) ? 1 : 0;
+}
+
+static void enetc4_pf_set_si_primary_mac(struct enetc_hw *hw, int si,
+					 const u8 *addr)
+{
+	u16 lower = get_unaligned_le16(addr + 4);
+	u32 upper = get_unaligned_le32(addr);
+
+	if (si != 0) {
+		__raw_writel(upper, hw->port + ENETC4_PSIPMAR0(si));
+		__raw_writew(lower, hw->port + ENETC4_PSIPMAR1(si));
+	} else {
+		__raw_writel(upper, hw->port + ENETC4_PMAR0);
+		__raw_writew(lower, hw->port + ENETC4_PMAR1);
+	}
+}
+
+static void enetc4_pf_get_si_primary_mac(struct enetc_hw *hw, int si,
+					 u8 *addr)
+{
+	u32 upper;
+	u16 lower;
+
+	upper = __raw_readl(hw->port + ENETC4_PSIPMAR0(si));
+	lower = __raw_readw(hw->port + ENETC4_PSIPMAR1(si));
+
+	put_unaligned_le32(upper, addr);
+	put_unaligned_le16(lower, addr + 4);
+}
+
+static const struct enetc_pf_ops enetc4_pf_ops = {
+	.set_si_primary_mac = enetc4_pf_set_si_primary_mac,
+	.get_si_primary_mac = enetc4_pf_get_si_primary_mac,
+};
+
+static int enetc4_pf_struct_init(struct enetc_si *si)
+{
+	struct enetc_pf *pf = enetc_si_priv(si);
+
+	pf->si = si;
+	pf->total_vfs = pci_sriov_get_totalvfs(si->pdev);
+
+	enetc4_get_port_caps(pf);
+	enetc_pf_ops_register(pf, &enetc4_pf_ops);
+
+	return 0;
+}
+
+static u32 enetc4_psicfgr0_val_construct(bool is_vf, u32 num_tx_bdr, u32 num_rx_bdr)
+{
+	u32 val;
+
+	val = ENETC_PSICFGR0_SET_TXBDR(num_tx_bdr);
+	val |= ENETC_PSICFGR0_SET_RXBDR(num_rx_bdr);
+	val |= ENETC_PSICFGR0_SIVC(ENETC_VLAN_TYPE_C | ENETC_VLAN_TYPE_S);
+
+	if (is_vf)
+		val |= ENETC_PSICFGR0_VTE | ENETC_PSICFGR0_SIVIE;
+
+	return val;
+}
+
+static void enetc4_default_rings_allocation(struct enetc_pf *pf)
+{
+	struct enetc_hw *hw = &pf->si->hw;
+	u32 num_rx_bdr, num_tx_bdr, val;
+	u32 vf_tx_bdr, vf_rx_bdr;
+	int i, rx_rem, tx_rem;
+
+	if (pf->caps.num_rx_bdr < ENETC_SI_MAX_RING_NUM + pf->caps.num_vsi)
+		num_rx_bdr = pf->caps.num_rx_bdr - pf->caps.num_vsi;
+	else
+		num_rx_bdr = ENETC_SI_MAX_RING_NUM;
+
+	if (pf->caps.num_tx_bdr < ENETC_SI_MAX_RING_NUM + pf->caps.num_vsi)
+		num_tx_bdr = pf->caps.num_tx_bdr - pf->caps.num_vsi;
+	else
+		num_tx_bdr = ENETC_SI_MAX_RING_NUM;
+
+	val = enetc4_psicfgr0_val_construct(false, num_tx_bdr, num_rx_bdr);
+	enetc_port_wr(hw, ENETC4_PSICFGR0(0), val);
+
+	num_rx_bdr = pf->caps.num_rx_bdr - num_rx_bdr;
+	rx_rem = num_rx_bdr % pf->caps.num_vsi;
+	num_rx_bdr = num_rx_bdr / pf->caps.num_vsi;
+
+	num_tx_bdr = pf->caps.num_tx_bdr - num_tx_bdr;
+	tx_rem = num_tx_bdr % pf->caps.num_vsi;
+	num_tx_bdr = num_tx_bdr / pf->caps.num_vsi;
+
+	for (i = 0; i < pf->caps.num_vsi; i++) {
+		vf_tx_bdr = (i < tx_rem) ? num_tx_bdr + 1 : num_tx_bdr;
+		vf_rx_bdr = (i < rx_rem) ? num_rx_bdr + 1 : num_rx_bdr;
+		val = enetc4_psicfgr0_val_construct(true, vf_tx_bdr, vf_rx_bdr);
+		enetc_port_wr(hw, ENETC4_PSICFGR0(i + 1), val);
+	}
+}
+
+static void enetc4_allocate_si_rings(struct enetc_pf *pf)
+{
+	enetc4_default_rings_allocation(pf);
+}
+
+static void enetc4_pf_set_si_vlan_promisc(struct enetc_hw *hw, int si, bool en)
+{
+	u32 val = enetc_port_rd(hw, ENETC4_PSIPVMR);
+
+	if (en)
+		val |= BIT(si);
+	else
+		val &= ~BIT(si);
+
+	enetc_port_wr(hw, ENETC4_PSIPVMR, val);
+}
+
+static void enetc4_set_default_si_vlan_promisc(struct enetc_pf *pf)
+{
+	struct enetc_hw *hw = &pf->si->hw;
+	int num_si = pf->caps.num_vsi + 1;
+	int i;
+
+	/* enforce VLAN promisc mode for all SIs */
+	for (i = 0; i < num_si; i++)
+		enetc4_pf_set_si_vlan_promisc(hw, i, true);
+}
+
+/* Allocate the number of MSI-X vectors for per SI. */
+static void enetc4_set_si_msix_num(struct enetc_pf *pf)
+{
+	struct enetc_hw *hw = &pf->si->hw;
+	int i, num_msix, total_si;
+	u32 val;
+
+	total_si = pf->caps.num_vsi + 1;
+
+	num_msix = pf->caps.num_msix / total_si +
+		   pf->caps.num_msix % total_si - 1;
+	val = num_msix & PSICFGR2_NUM_MSIX;
+	enetc_port_wr(hw, ENETC4_PSICFGR2(0), val);
+
+	num_msix = pf->caps.num_msix / total_si - 1;
+	val = num_msix & PSICFGR2_NUM_MSIX;
+	for (i = 0; i < pf->caps.num_vsi; i++)
+		enetc_port_wr(hw, ENETC4_PSICFGR2(i + 1), val);
+}
+
+static void enetc4_enable_all_si(struct enetc_pf *pf)
+{
+	struct enetc_hw *hw = &pf->si->hw;
+	int num_si = pf->caps.num_vsi + 1;
+	u32 si_bitmap = 0;
+	int i;
+
+	/* Master enable for all SIs */
+	for (i = 0; i < num_si; i++)
+		si_bitmap |= PMR_SI_EN(i);
+
+	enetc_port_wr(hw, ENETC4_PMR, si_bitmap);
+}
+
+static void enetc4_configure_port_si(struct enetc_pf *pf)
+{
+	struct enetc_hw *hw = &pf->si->hw;
+
+	enetc4_allocate_si_rings(pf);
+
+	/* Outer VLAN tag will be used for VLAN filtering */
+	enetc_port_wr(hw, ENETC4_PSIVLANFMR, PSIVLANFMR_VS);
+
+	enetc4_set_default_si_vlan_promisc(pf);
+
+	/* Disable SI MAC multicast & unicast promiscuous */
+	enetc_port_wr(hw, ENETC4_PSIPMMR, 0);
+
+	enetc4_set_si_msix_num(pf);
+
+	enetc4_enable_all_si(pf);
+}
+
+static void enetc4_pf_reset_tc_msdu(struct enetc_hw *hw)
+{
+	u32 val = ENETC_MAC_MAXFRM_SIZE;
+	int tc;
+
+	val = u32_replace_bits(val, SDU_TYPE_MPDU, PTCTMSDUR_SDU_TYPE);
+
+	for (tc = 0; tc < ENETC_NUM_TC; tc++)
+		enetc_port_wr(hw, ENETC4_PTCTMSDUR(tc), val);
+}
+
+static void enetc4_set_trx_frame_size(struct enetc_pf *pf)
+{
+	struct enetc_si *si = pf->si;
+
+	enetc_port_mac_wr(si, ENETC4_PM_MAXFRM(0),
+			  ENETC_SET_MAXFRM(ENETC_MAC_MAXFRM_SIZE));
+
+	enetc4_pf_reset_tc_msdu(&si->hw);
+}
+
+static void enetc4_set_rss_key(struct enetc_hw *hw, const u8 *bytes)
+{
+	int i;
+
+	for (i = 0; i < ENETC_RSSHASH_KEY_SIZE / 4; i++)
+		enetc_port_wr(hw, ENETC4_PRSSKR(i), ((u32 *)bytes)[i]);
+}
+
+static void enetc4_set_default_rss_key(struct enetc_pf *pf)
+{
+	u8 hash_key[ENETC_RSSHASH_KEY_SIZE] = {0};
+	struct enetc_hw *hw = &pf->si->hw;
+
+	/* set up hash key */
+	get_random_bytes(hash_key, ENETC_RSSHASH_KEY_SIZE);
+	enetc4_set_rss_key(hw, hash_key);
+}
+
+static void enetc4_enable_trx(struct enetc_pf *pf)
+{
+	struct enetc_hw *hw = &pf->si->hw;
+
+	/* Enable port transmit/receive */
+	enetc_port_wr(hw, ENETC4_POR, 0);
+}
+
+static void enetc4_configure_port(struct enetc_pf *pf)
+{
+	enetc4_configure_port_si(pf);
+	enetc4_set_trx_frame_size(pf);
+	enetc4_set_default_rss_key(pf);
+	enetc4_enable_trx(pf);
+}
+
+static int enetc4_pf_init(struct enetc_pf *pf)
+{
+	struct device *dev = &pf->si->pdev->dev;
+	int err;
+
+	/* Initialize the MAC address for PF and VFs */
+	err = enetc_setup_mac_addresses(dev->of_node, pf);
+	if (err) {
+		dev_err(dev, "Failed to set MAC addresses\n");
+		return err;
+	}
+
+	enetc4_configure_port(pf);
+
+	return 0;
+}
+
+static const struct net_device_ops enetc4_ndev_ops = {
+	.ndo_open		= enetc_open,
+	.ndo_stop		= enetc_close,
+	.ndo_start_xmit		= enetc_xmit,
+	.ndo_get_stats		= enetc_get_stats,
+	.ndo_set_mac_address	= enetc_pf_set_mac_addr,
+};
+
+static struct phylink_pcs *
+enetc4_pl_mac_select_pcs(struct phylink_config *config, phy_interface_t iface)
+{
+	struct enetc_pf *pf = phylink_to_enetc_pf(config);
+
+	return pf->pcs;
+}
+
+static void enetc4_mac_config(struct enetc_pf *pf, unsigned int mode,
+			      phy_interface_t phy_mode)
+{
+	struct enetc_ndev_priv *priv = netdev_priv(pf->si->ndev);
+	struct enetc_si *si = pf->si;
+	u32 val;
+
+	val = enetc_port_mac_rd(si, ENETC4_PM_IF_MODE(0));
+	val &= ~(PM_IF_MODE_IFMODE | PM_IF_MODE_ENA);
+
+	switch (phy_mode) {
+	case PHY_INTERFACE_MODE_RGMII:
+	case PHY_INTERFACE_MODE_RGMII_ID:
+	case PHY_INTERFACE_MODE_RGMII_RXID:
+	case PHY_INTERFACE_MODE_RGMII_TXID:
+		val |= IFMODE_RGMII;
+		/* We need to enable auto-negotiation for the MAC
+		 * if its RGMII interface support In-Band status.
+		 */
+		if (phylink_autoneg_inband(mode))
+			val |= PM_IF_MODE_ENA;
+		break;
+	case PHY_INTERFACE_MODE_RMII:
+		val |= IFMODE_RMII;
+		break;
+	case PHY_INTERFACE_MODE_SGMII:
+	case PHY_INTERFACE_MODE_2500BASEX:
+		val |= IFMODE_SGMII;
+		break;
+	case PHY_INTERFACE_MODE_10GBASER:
+	case PHY_INTERFACE_MODE_XGMII:
+	case PHY_INTERFACE_MODE_USXGMII:
+		val |= IFMODE_XGMII;
+		break;
+	default:
+		dev_err(priv->dev,
+			"Unsupported PHY mode:%d\n", phy_mode);
+		return;
+	}
+
+	enetc_port_mac_wr(si, ENETC4_PM_IF_MODE(0), val);
+}
+
+static void enetc4_pl_mac_config(struct phylink_config *config, unsigned int mode,
+				 const struct phylink_link_state *state)
+{
+	struct enetc_pf *pf = phylink_to_enetc_pf(config);
+
+	enetc4_mac_config(pf, mode, state->interface);
+}
+
+static void enetc4_set_port_speed(struct enetc_ndev_priv *priv, int speed)
+{
+	u32 old_speed = priv->speed;
+	u32 val;
+
+	if (speed == old_speed)
+		return;
+
+	val = enetc_port_rd(&priv->si->hw, ENETC4_PCR);
+	val &= ~PCR_PSPEED;
+
+	switch (speed) {
+	case SPEED_100:
+	case SPEED_1000:
+	case SPEED_2500:
+	case SPEED_10000:
+		val |= (PCR_PSPEED & PCR_PSPEED_VAL(speed));
+		break;
+	case SPEED_10:
+	default:
+		val |= (PCR_PSPEED & PCR_PSPEED_VAL(SPEED_10));
+	}
+
+	priv->speed = speed;
+	enetc_port_wr(&priv->si->hw, ENETC4_PCR, val);
+}
+
+static void enetc4_set_rgmii_mac(struct enetc_pf *pf, int speed, int duplex)
+{
+	struct enetc_si *si = pf->si;
+	u32 old_val, val;
+
+	old_val = enetc_port_mac_rd(si, ENETC4_PM_IF_MODE(0));
+	val = old_val & ~(PM_IF_MODE_ENA | PM_IF_MODE_M10 | PM_IF_MODE_REVMII);
+
+	switch (speed) {
+	case SPEED_1000:
+		val = u32_replace_bits(val, SSP_1G, PM_IF_MODE_SSP);
+		break;
+	case SPEED_100:
+		val = u32_replace_bits(val, SSP_100M, PM_IF_MODE_SSP);
+		break;
+	case SPEED_10:
+		val = u32_replace_bits(val, SSP_10M, PM_IF_MODE_SSP);
+	}
+
+	val = u32_replace_bits(val, duplex == DUPLEX_FULL ? 0 : 1,
+			       PM_IF_MODE_HD);
+
+	if (val == old_val)
+		return;
+
+	enetc_port_mac_wr(si, ENETC4_PM_IF_MODE(0), val);
+}
+
+static void enetc4_set_rmii_mac(struct enetc_pf *pf, int speed, int duplex)
+{
+	struct enetc_si *si = pf->si;
+	u32 old_val, val;
+
+	old_val = enetc_port_mac_rd(si, ENETC4_PM_IF_MODE(0));
+	val = old_val & ~(PM_IF_MODE_ENA | PM_IF_MODE_SSP);
+
+	switch (speed) {
+	case SPEED_100:
+		val &= ~PM_IF_MODE_M10;
+		break;
+	case SPEED_10:
+		val |= PM_IF_MODE_M10;
+	}
+
+	val = u32_replace_bits(val, duplex == DUPLEX_FULL ? 0 : 1,
+			       PM_IF_MODE_HD);
+
+	if (val == old_val)
+		return;
+
+	enetc_port_mac_wr(si, ENETC4_PM_IF_MODE(0), val);
+}
+
+static void enetc4_set_hd_flow_control(struct enetc_pf *pf, bool enable)
+{
+	struct enetc_si *si = pf->si;
+	u32 old_val, val;
+
+	if (!pf->caps.half_duplex)
+		return;
+
+	old_val = enetc_port_mac_rd(si, ENETC4_PM_CMD_CFG(0));
+	val = u32_replace_bits(old_val, enable ? 1 : 0, PM_CMD_CFG_HD_FCEN);
+	if (val == old_val)
+		return;
+
+	enetc_port_mac_wr(si, ENETC4_PM_CMD_CFG(0), val);
+}
+
+static void enetc4_set_rx_pause(struct enetc_pf *pf, bool rx_pause)
+{
+	struct enetc_si *si = pf->si;
+	u32 old_val, val;
+
+	old_val = enetc_port_mac_rd(si, ENETC4_PM_CMD_CFG(0));
+	val = u32_replace_bits(old_val, rx_pause ? 0 : 1, PM_CMD_CFG_PAUSE_IGN);
+	if (val == old_val)
+		return;
+
+	enetc_port_mac_wr(si, ENETC4_PM_CMD_CFG(0), val);
+}
+
+static void enetc4_set_tx_pause(struct enetc_pf *pf, int num_rxbdr, bool tx_pause)
+{
+	u32 pause_off_thresh = 0, pause_on_thresh = 0;
+	u32 init_quanta = 0, refresh_quanta = 0;
+	struct enetc_hw *hw = &pf->si->hw;
+	u32 rbmr, old_rbmr;
+	int i;
+
+	for (i = 0; i < num_rxbdr; i++) {
+		old_rbmr = enetc_rxbdr_rd(hw, i, ENETC_RBMR);
+		rbmr = u32_replace_bits(old_rbmr, tx_pause ? 1 : 0, ENETC_RBMR_CM);
+		if (rbmr == old_rbmr)
+			continue;
+
+		enetc_rxbdr_wr(hw, i, ENETC_RBMR, rbmr);
+	}
+
+	if (tx_pause) {
+		/* When the port first enters congestion, send a PAUSE request
+		 * with the maximum number of quanta. When the port exits
+		 * congestion, it will automatically send a PAUSE frame with
+		 * zero quanta.
+		 */
+		init_quanta = 0xffff;
+
+		/* Also, set up the refresh timer to send follow-up PAUSE
+		 * frames at half the quanta value, in case the congestion
+		 * condition persists.
+		 */
+		refresh_quanta = 0xffff / 2;
+
+		/* Start emitting PAUSE frames when 3 large frames (or more
+		 * smaller frames) have accumulated in the FIFO waiting to be
+		 * DMAed to the RX ring.
+		 */
+		pause_on_thresh = 3 * ENETC_MAC_MAXFRM_SIZE;
+		pause_off_thresh = 1 * ENETC_MAC_MAXFRM_SIZE;
+	}
+
+	enetc_port_mac_wr(pf->si, ENETC4_PM_PAUSE_QUANTA(0), init_quanta);
+	enetc_port_mac_wr(pf->si, ENETC4_PM_PAUSE_THRESH(0), refresh_quanta);
+	enetc_port_wr(hw, ENETC4_PPAUONTR, pause_on_thresh);
+	enetc_port_wr(hw, ENETC4_PPAUOFFTR, pause_off_thresh);
+}
+
+static void enetc4_enable_mac(struct enetc_pf *pf, bool en)
+{
+	struct enetc_si *si = pf->si;
+	u32 val;
+
+	val = enetc_port_mac_rd(si, ENETC4_PM_CMD_CFG(0));
+	val &= ~(PM_CMD_CFG_TX_EN | PM_CMD_CFG_RX_EN);
+	val |= en ? (PM_CMD_CFG_TX_EN | PM_CMD_CFG_RX_EN) : 0;
+
+	enetc_port_mac_wr(si, ENETC4_PM_CMD_CFG(0), val);
+}
+
+static void enetc4_pl_mac_link_up(struct phylink_config *config,
+				  struct phy_device *phy, unsigned int mode,
+				  phy_interface_t interface, int speed,
+				  int duplex, bool tx_pause, bool rx_pause)
+{
+	struct enetc_pf *pf = phylink_to_enetc_pf(config);
+	struct enetc_si *si = pf->si;
+	struct enetc_ndev_priv *priv;
+	bool hd_fc = false;
+
+	priv = netdev_priv(si->ndev);
+	enetc4_set_port_speed(priv, speed);
+
+	if (!phylink_autoneg_inband(mode) &&
+	    phy_interface_mode_is_rgmii(interface))
+		enetc4_set_rgmii_mac(pf, speed, duplex);
+
+	if (interface == PHY_INTERFACE_MODE_RMII)
+		enetc4_set_rmii_mac(pf, speed, duplex);
+
+	if (duplex == DUPLEX_FULL) {
+		/* When preemption is enabled, generation of PAUSE frames
+		 * must be disabled, as stated in the IEEE 802.3 standard.
+		 */
+		if (priv->active_offloads & ENETC_F_QBU)
+			tx_pause = false;
+	} else { /* DUPLEX_HALF */
+		if (tx_pause || rx_pause)
+			hd_fc = true;
+
+		/* As per 802.3 annex 31B, PAUSE frames are only supported
+		 * when the link is configured for full duplex operation.
+		 */
+		tx_pause = false;
+		rx_pause = false;
+	}
+
+	enetc4_set_hd_flow_control(pf, hd_fc);
+	enetc4_set_tx_pause(pf, priv->num_rx_rings, tx_pause);
+	enetc4_set_rx_pause(pf, rx_pause);
+	enetc4_enable_mac(pf, true);
+}
+
+static void enetc4_pl_mac_link_down(struct phylink_config *config,
+				    unsigned int mode,
+				    phy_interface_t interface)
+{
+	struct enetc_pf *pf = phylink_to_enetc_pf(config);
+
+	enetc4_enable_mac(pf, false);
+}
+
+static const struct phylink_mac_ops enetc_pl_mac_ops = {
+	.mac_select_pcs = enetc4_pl_mac_select_pcs,
+	.mac_config = enetc4_pl_mac_config,
+	.mac_link_up = enetc4_pl_mac_link_up,
+	.mac_link_down = enetc4_pl_mac_link_down,
+};
+
+static void enetc4_pci_remove(void *data)
+{
+	struct pci_dev *pdev = data;
+
+	enetc_pci_remove(pdev);
+}
+
+static int enetc4_link_init(struct enetc_ndev_priv *priv,
+			    struct device_node *node)
+{
+	struct enetc_pf *pf = enetc_si_priv(priv->si);
+	struct device *dev = priv->dev;
+	int err;
+
+	err = of_get_phy_mode(node, &pf->if_mode);
+	if (err) {
+		dev_err(dev, "Failed to get PHY mode\n");
+		return err;
+	}
+
+	err = enetc_mdiobus_create(pf, node);
+	if (err) {
+		dev_err(dev, "Failed to create MDIO bus\n");
+		return err;
+	}
+
+	err = enetc_phylink_create(priv, node, &enetc_pl_mac_ops);
+	if (err) {
+		dev_err(dev, "Failed to create phylink\n");
+		goto err_phylink_create;
+	}
+
+	return 0;
+
+err_phylink_create:
+	enetc_mdiobus_destroy(pf);
+
+	return err;
+}
+
+static void enetc4_link_deinit(struct enetc_ndev_priv *priv)
+{
+	struct enetc_pf *pf = enetc_si_priv(priv->si);
+
+	enetc_phylink_destroy(priv);
+	enetc_mdiobus_destroy(pf);
+}
+
+static int enetc4_pf_netdev_create(struct enetc_si *si)
+{
+	struct device *dev = &si->pdev->dev;
+	struct enetc_ndev_priv *priv;
+	struct net_device *ndev;
+	int err;
+
+	ndev = alloc_etherdev_mqs(sizeof(struct enetc_ndev_priv),
+				  si->num_tx_rings, si->num_rx_rings);
+	if (!ndev)
+		return  -ENOMEM;
+
+	priv = netdev_priv(ndev);
+	priv->ref_clk = devm_clk_get_optional(dev, "ref");
+	if (IS_ERR(priv->ref_clk)) {
+		dev_err(dev, "Get referencce clock failed\n");
+		err = PTR_ERR(priv->ref_clk);
+		goto err_clk_get;
+	}
+
+	enetc_pf_netdev_setup(si, ndev, &enetc4_ndev_ops);
+
+	enetc_init_si_rings_params(priv);
+
+	err = enetc_configure_si(priv);
+	if (err) {
+		dev_err(dev, "Failed to configure SI\n");
+		goto err_config_si;
+	}
+
+	err = enetc_alloc_msix(priv);
+	if (err) {
+		dev_err(dev, "Failed to alloc MSI-X\n");
+		goto err_alloc_msix;
+	}
+
+	err = enetc4_link_init(priv, dev->of_node);
+	if (err)
+		goto err_link_init;
+
+	err = register_netdev(ndev);
+	if (err) {
+		dev_err(dev, "Failed to register netdev\n");
+		goto err_reg_netdev;
+	}
+
+	return 0;
+
+err_reg_netdev:
+	enetc4_link_deinit(priv);
+err_link_init:
+	enetc_free_msix(priv);
+err_alloc_msix:
+err_config_si:
+err_clk_get:
+	mutex_destroy(&priv->mm_lock);
+	free_netdev(ndev);
+
+	return err;
+}
+
+static void enetc4_pf_netdev_destroy(struct enetc_si *si)
+{
+	struct enetc_ndev_priv *priv = netdev_priv(si->ndev);
+	struct net_device *ndev = si->ndev;
+
+	unregister_netdev(ndev);
+	enetc_free_msix(priv);
+	free_netdev(ndev);
+}
+
+static int enetc4_pf_probe(struct pci_dev *pdev,
+			   const struct pci_device_id *ent)
+{
+	struct device *dev = &pdev->dev;
+	struct enetc_si *si;
+	struct enetc_pf *pf;
+	int err;
+
+	err = enetc_pci_probe(pdev, KBUILD_MODNAME, sizeof(*pf));
+	if (err)
+		return dev_err_probe(dev, err, "PCIe probing failed\n");
+
+	err = devm_add_action_or_reset(dev, enetc4_pci_remove, pdev);
+	if (err)
+		return dev_err_probe(dev, err,
+				     "Add enetc4_pci_remove() action failed\n");
+
+	/* si is the private data. */
+	si = pci_get_drvdata(pdev);
+	if (!si->hw.port || !si->hw.global)
+		return dev_err_probe(dev, -ENODEV,
+				     "Couldn't map PF only space\n");
+
+	si->revision = enetc_get_ip_revision(&si->hw);
+	err = enetc_get_driver_data(si);
+	if (err)
+		return dev_err_probe(dev, err,
+				     "Could not get VF driver data\n");
+
+	err = enetc4_pf_struct_init(si);
+	if (err)
+		return err;
+
+	pf = enetc_si_priv(si);
+	err = enetc4_pf_init(pf);
+	if (err)
+		return err;
+
+	pinctrl_pm_select_default_state(dev);
+	enetc_get_si_caps(si);
+
+	return enetc4_pf_netdev_create(si);
+}
+
+static void enetc4_pf_remove(struct pci_dev *pdev)
+{
+	struct enetc_si *si = pci_get_drvdata(pdev);
+
+	enetc4_pf_netdev_destroy(si);
+}
+
+static const struct pci_device_id enetc4_pf_id_table[] = {
+	{ PCI_DEVICE(PCI_VENDOR_ID_NXP2, PCI_DEVICE_ID_NXP2_ENETC_PF) },
+	{ 0, } /* End of table. */
+};
+MODULE_DEVICE_TABLE(pci, enetc4_pf_id_table);
+
+static struct pci_driver enetc4_pf_driver = {
+	.name = KBUILD_MODNAME,
+	.id_table = enetc4_pf_id_table,
+	.probe = enetc4_pf_probe,
+	.remove = enetc4_pf_remove,
+};
+module_pci_driver(enetc4_pf_driver);
+
+MODULE_DESCRIPTION("ENETC4 PF Driver");
+MODULE_LICENSE("Dual BSD/GPL");
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c b/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
index 2563eb8ac7b6..4c223eb90c19 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_ethtool.c
@@ -775,9 +775,10 @@ static int enetc_get_coalesce(struct net_device *ndev,
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
 	struct enetc_int_vector *v = priv->int_vector[0];
+	u64 clk_freq = priv->sysclk_freq;
 
-	ic->tx_coalesce_usecs = enetc_cycles_to_usecs(priv->tx_ictt);
-	ic->rx_coalesce_usecs = enetc_cycles_to_usecs(v->rx_ictt);
+	ic->tx_coalesce_usecs = enetc_cycles_to_usecs(priv->tx_ictt, clk_freq);
+	ic->rx_coalesce_usecs = enetc_cycles_to_usecs(v->rx_ictt, clk_freq);
 
 	ic->tx_max_coalesced_frames = ENETC_TXIC_PKTTHR;
 	ic->rx_max_coalesced_frames = ENETC_RXIC_PKTTHR;
@@ -793,12 +794,13 @@ static int enetc_set_coalesce(struct net_device *ndev,
 			      struct netlink_ext_ack *extack)
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
+	u64 clk_freq = priv->sysclk_freq;
 	u32 rx_ictt, tx_ictt;
 	int i, ic_mode;
 	bool changed;
 
-	tx_ictt = enetc_usecs_to_cycles(ic->tx_coalesce_usecs);
-	rx_ictt = enetc_usecs_to_cycles(ic->rx_coalesce_usecs);
+	tx_ictt = enetc_usecs_to_cycles(ic->tx_coalesce_usecs, clk_freq);
+	rx_ictt = enetc_usecs_to_cycles(ic->rx_coalesce_usecs, clk_freq);
 
 	if (ic->rx_max_coalesced_frames != ENETC_RXIC_PKTTHR)
 		return -EOPNOTSUPP;
@@ -1178,7 +1180,7 @@ void enetc_mm_link_state_update(struct enetc_ndev_priv *priv, bool link)
 }
 EXPORT_SYMBOL_GPL(enetc_mm_link_state_update);
 
-static const struct ethtool_ops enetc_pf_ethtool_ops = {
+const struct ethtool_ops enetc_pf_ethtool_ops = {
 	.supported_coalesce_params = ETHTOOL_COALESCE_USECS |
 				     ETHTOOL_COALESCE_MAX_FRAMES |
 				     ETHTOOL_COALESCE_USE_ADAPTIVE_RX,
@@ -1213,7 +1215,7 @@ static const struct ethtool_ops enetc_pf_ethtool_ops = {
 	.get_mm_stats = enetc_get_mm_stats,
 };
 
-static const struct ethtool_ops enetc_vf_ethtool_ops = {
+const struct ethtool_ops enetc_vf_ethtool_ops = {
 	.supported_coalesce_params = ETHTOOL_COALESCE_USECS |
 				     ETHTOOL_COALESCE_MAX_FRAMES |
 				     ETHTOOL_COALESCE_USE_ADAPTIVE_RX,
@@ -1234,13 +1236,26 @@ static const struct ethtool_ops enetc_vf_ethtool_ops = {
 	.get_ts_info = enetc_get_ts_info,
 };
 
+const struct ethtool_ops enetc4_pf_ethtool_ops = {
+	.supported_coalesce_params = ETHTOOL_COALESCE_USECS |
+				     ETHTOOL_COALESCE_MAX_FRAMES |
+				     ETHTOOL_COALESCE_USE_ADAPTIVE_RX,
+	.get_ringparam = enetc_get_ringparam,
+	.get_coalesce = enetc_get_coalesce,
+	.set_coalesce = enetc_set_coalesce,
+	.get_link_ksettings = enetc_get_link_ksettings,
+	.set_link_ksettings = enetc_set_link_ksettings,
+	.get_link = ethtool_op_get_link,
+	.get_wol = enetc_get_wol,
+	.set_wol = enetc_set_wol,
+	.get_pauseparam = enetc_get_pauseparam,
+	.set_pauseparam = enetc_set_pauseparam,
+};
+
 void enetc_set_ethtool_ops(struct net_device *ndev)
 {
 	struct enetc_ndev_priv *priv = netdev_priv(ndev);
 
-	if (enetc_si_is_pf(priv->si))
-		ndev->ethtool_ops = &enetc_pf_ethtool_ops;
-	else
-		ndev->ethtool_ops = &enetc_vf_ethtool_ops;
+	ndev->ethtool_ops = priv->si->drvdata->eth_ops;
 }
 EXPORT_SYMBOL_GPL(enetc_set_ethtool_ops);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_hw.h b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
index 6a7b9b75d660..7c3285584f8a 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_hw.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc_hw.h
@@ -6,6 +6,8 @@
 #define ENETC_MM_VERIFY_SLEEP_US	USEC_PER_MSEC
 #define ENETC_MM_VERIFY_RETRIES		3
 
+#define ENETC_NUM_TC			8
+
 /* ENETC device IDs */
 #define ENETC_DEV_ID_PF		0xe100
 #define ENETC_DEV_ID_VF		0xef00
@@ -25,6 +27,7 @@
 #define ENETC_SIPCAPR0_RSS	BIT(8)
 #define ENETC_SIPCAPR0_QBV	BIT(4)
 #define ENETC_SIPCAPR0_QBU	BIT(3)
+#define ENETC_SIPCAPR0_RFS	BIT(2)
 #define ENETC_SIPCAPR1	0x24
 #define ENETC_SITGTGR	0x30
 #define ENETC_SIRBGCR	0x38
@@ -368,6 +371,10 @@ enum enetc_bdr_type {TX, RX};
 /** Global regs, offset: 2_0000h */
 #define ENETC_GLOBAL_BASE	0x20000
 #define ENETC_G_EIPBRR0		0x0bf8
+#define EIPBRR0_REVISION	GENMASK(15, 0)
+#define ENETC_REV_1_0		0x0100
+#define ENETC_REV_4_1		0X0401
+
 #define ENETC_G_EIPBRR1		0x0bfc
 #define ENETC_G_EPFBLPR(n)	(0xd00 + 4 * (n))
 #define ENETC_G_EPFBLPR1_XGMII	0x80000000
@@ -971,15 +978,17 @@ struct enetc_cbd {
 	u8 status_flags;
 };
 
-#define ENETC_CLK  400000000ULL
-static inline u32 enetc_cycles_to_usecs(u32 cycles)
+#define ENETC_CLK_400M		400000000ULL
+#define ENETC_CLK_333M		333000000ULL
+
+static inline u32 enetc_cycles_to_usecs(u32 cycles, u64 clk_freq)
 {
-	return (u32)div_u64(cycles * 1000000ULL, ENETC_CLK);
+	return (u32)div_u64(cycles * 1000000ULL, clk_freq);
 }
 
-static inline u32 enetc_usecs_to_cycles(u32 usecs)
+static inline u32 enetc_usecs_to_cycles(u32 usecs, u64 clk_freq)
 {
-	return (u32)div_u64(usecs * ENETC_CLK, 1000000ULL);
+	return (u32)div_u64(usecs * clk_freq, 1000000ULL);
 }
 
 /* Port traffic class frame preemption register */
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.c b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
index 7522316ddfea..527b317ddc11 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.c
@@ -946,6 +946,13 @@ static struct enetc_si *enetc_psi_create(struct pci_dev *pdev)
 		goto out_pci_remove;
 	}
 
+	si->revision = enetc_get_ip_revision(&si->hw);
+	err = enetc_get_driver_data(si);
+	if (err) {
+		dev_err(&pdev->dev, "Could not get PF driver data\n");
+		goto out_pci_remove;
+	}
+
 	err = enetc_setup_cbdr(&pdev->dev, &si->hw, ENETC_CBDR_DEFAULT_SIZE,
 			       &si->cbd_ring);
 	if (err)
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.h b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
index 39db9d5c2e50..19bf206438cd 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf.h
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
@@ -28,6 +28,14 @@ struct enetc_vf_state {
 	enum enetc_vf_flags flags;
 };
 
+struct enetc_port_caps {
+	u32 half_duplex:1;
+	int num_vsi;
+	int num_msix;
+	int num_rx_bdr;
+	int num_tx_bdr;
+};
+
 struct enetc_pf;
 
 struct enetc_pf_ops {
@@ -61,6 +69,7 @@ struct enetc_pf {
 	phy_interface_t if_mode;
 	struct phylink_config phylink_config;
 
+	struct enetc_port_caps caps;
 	const struct enetc_pf_ops *ops;
 };
 
@@ -81,6 +90,11 @@ int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node *node,
 			 const struct phylink_mac_ops *ops);
 void enetc_phylink_destroy(struct enetc_ndev_priv *priv);
 
+static inline u16 enetc_get_ip_revision(struct enetc_hw *hw)
+{
+	return enetc_global_rd(hw, ENETC_G_EIPBRR0) & EIPBRR0_REVISION;
+}
+
 static inline void enetc_pf_ops_register(struct enetc_pf *pf,
 					 const struct enetc_pf_ops *ops)
 {
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
index 94690ed92e3f..6d6059eac536 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
@@ -105,6 +105,7 @@ void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
 	si->ndev = ndev;
 
 	priv->msg_enable = (NETIF_MSG_WOL << 1) - 1;
+	priv->sysclk_freq = si->drvdata->sysclk_freq;
 	ndev->netdev_ops = ndev_ops;
 	enetc_set_ethtool_ops(ndev);
 	ndev->watchdog_timeo = 5 * HZ;
@@ -121,10 +122,17 @@ void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
 	ndev->vlan_features = NETIF_F_SG | NETIF_F_HW_CSUM |
 			      NETIF_F_TSO | NETIF_F_TSO6;
 
+	ndev->priv_flags |= IFF_UNICAST_FLT;
+
+	/* TODO: currently, i.MX95 ENETC driver does not support advanced features */
+	if (!is_enetc_rev1(si)) {
+		ndev->hw_features &= ~(NETIF_F_HW_VLAN_CTAG_FILTER | NETIF_F_LOOPBACK);
+		goto end;
+	}
+
 	if (si->num_rss)
 		ndev->hw_features |= NETIF_F_RXHASH;
 
-	ndev->priv_flags |= IFF_UNICAST_FLT;
 	ndev->xdp_features = NETDEV_XDP_ACT_BASIC | NETDEV_XDP_ACT_REDIRECT |
 			     NETDEV_XDP_ACT_NDO_XMIT | NETDEV_XDP_ACT_RX_SG |
 			     NETDEV_XDP_ACT_NDO_XMIT_SG;
@@ -136,6 +144,7 @@ void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
 		ndev->hw_features |= NETIF_F_HW_TC;
 	}
 
+end:
 	/* pick up primary MAC address from SI */
 	enetc_load_primary_mac_addr(&si->hw, ndev);
 }
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_qos.c b/drivers/net/ethernet/freescale/enetc/enetc_qos.c
index b65da49dd926..ccf86651455c 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_qos.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_qos.c
@@ -336,7 +336,7 @@ int enetc_setup_tc_cbs(struct net_device *ndev, void *type_data)
 	 *
 	 * (enetClockFrequency / portTransmitRate) * 100
 	 */
-	hi_credit_reg = (u32)div_u64((ENETC_CLK * 100ULL) * hi_credit_bit,
+	hi_credit_reg = (u32)div_u64((priv->sysclk_freq * 100ULL) * hi_credit_bit,
 				     port_transmit_rate * 1000000ULL);
 
 	enetc_port_wr(hw, ENETC_PTCCBSR1(tc), hi_credit_reg);
diff --git a/drivers/net/ethernet/freescale/enetc/enetc_vf.c b/drivers/net/ethernet/freescale/enetc/enetc_vf.c
index dfcaac302e24..31e630638090 100644
--- a/drivers/net/ethernet/freescale/enetc/enetc_vf.c
+++ b/drivers/net/ethernet/freescale/enetc/enetc_vf.c
@@ -128,6 +128,7 @@ static void enetc_vf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
 	si->ndev = ndev;
 
 	priv->msg_enable = (NETIF_MSG_IFUP << 1) - 1;
+	priv->sysclk_freq = si->drvdata->sysclk_freq;
 	ndev->netdev_ops = ndev_ops;
 	enetc_set_ethtool_ops(ndev);
 	ndev->watchdog_timeo = 5 * HZ;
@@ -164,6 +165,11 @@ static int enetc_vf_probe(struct pci_dev *pdev,
 		return dev_err_probe(&pdev->dev, err, "PCI probing failed\n");
 
 	si = pci_get_drvdata(pdev);
+	si->revision = ENETC_REV_1_0;
+	err = enetc_get_driver_data(si);
+	if (err)
+		return dev_err_probe(&pdev->dev, err,
+				     "Could not get VF driver data\n");
 
 	enetc_get_si_caps(si);
 
-- 
2.34.1

[PATCH v5 net-next 13/13] MAINTAINERS: update ENETC driver files and maintainers

[Wei Fang]

Add related YAML documentation and header files. Also, add maintainers
from the i.MX side as ENETC starts to be used on i.MX platforms.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
---
v5: no changes
---
 MAINTAINERS | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/MAINTAINERS b/MAINTAINERS
index 560a65b85297..cf442fcb9b49 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -9018,9 +9018,16 @@ F:	drivers/dma/fsl-edma*.*
 FREESCALE ENETC ETHERNET DRIVERS
 M:	Claudiu Manoil <claudiu.manoil@nxp.com>
 M:	Vladimir Oltean <vladimir.oltean@nxp.com>
+M:	Wei Fang <wei.fang@nxp.com>
+M:	Clark Wang <xiaoning.wang@nxp.com>
+L:	imx@lists.linux.dev
 L:	netdev@vger.kernel.org
 S:	Maintained
+F:	Documentation/devicetree/bindings/net/fsl,enetc*.yaml
+F:	Documentation/devicetree/bindings/net/nxp,netc-blk-ctrl.yaml
 F:	drivers/net/ethernet/freescale/enetc/
+F:	include/linux/fsl/enetc_mdio.h
+F:	include/linux/fsl/netc_global.h
 
 FREESCALE eTSEC ETHERNET DRIVER (GIANFAR)
 M:	Claudiu Manoil <claudiu.manoil@nxp.com>
-- 
2.34.1

Re: [PATCH v5 net-next 04/13] net: enetc: add initial netc-blk-ctrl driver support

[Vladimir Oltean]

On Thu, Oct 24, 2024 at 02:53:19PM +0800, Wei Fang wrote:
> The netc-blk-ctrl driver is used to configure Integrated Endpoint
> Register Block (IERB) and Privileged Register Block (PRB) of NETC.
> For i.MX platforms, it is also used to configure the NETCMIX block.
> 
> The IERB contains registers that are used for pre-boot initialization,
> debug, and non-customer configuration. The PRB controls global reset
> and global error handling for NETC. The NETCMIX block is mainly used
> to set MII protocol and PCS protocol of the links, it also contains
> settings for some other functions.
> 
> Note the IERB configuration registers can only be written after being
> unlocked by PRB, otherwise, all write operations are inhibited. A warm
> reset is performed when the IERB is unlocked, and it results in an FLR
> to all NETC devices. Therefore, all NETC device drivers must be probed
> or initialized after the warm reset is finished.
> 
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> Reviewed-by: Frank Li <Frank.Li@nxp.com>
> ---

Can U-Boot deal with the IERB/PRB configuration?

For LS1028A, the platform which initiated the IERB driver "trend", the
situation was a bit more complicated, as we realized the reset-time
defaults aren't what we need very late in the product life cycle, when
customer boards already had bootloaders and we didn't want to complicate
their process to have to redeploy in order to get access to such a basic
feature as flow control. Though if we knew it from day one, we would
have put the IERB fixups in U-Boot.

What is written in the IERB for MII/PCS protocols by default? I suppose
there's some other mechanism to preinitialize it with good values?

Re: [PATCH v5 net-next 05/13] net: enetc: extract common ENETC PF parts for LS1028A and i.MX95 platforms

[Vladimir Oltean]

On Thu, Oct 24, 2024 at 02:53:20PM +0800, Wei Fang wrote:
> diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.h b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> index c26bd66e4597..92a26b09cf57 100644
> --- a/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> +++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> @@ -58,3 +58,16 @@ struct enetc_pf {
>  int enetc_msg_psi_init(struct enetc_pf *pf);
>  void enetc_msg_psi_free(struct enetc_pf *pf);
>  void enetc_msg_handle_rxmsg(struct enetc_pf *pf, int mbox_id, u16 *status);
> +
> +void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr);
> +void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
> +				   const u8 *addr);
> +int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr);
> +int enetc_setup_mac_addresses(struct device_node *np, struct enetc_pf *pf);
> +void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
> +			   const struct net_device_ops *ndev_ops);
> +int enetc_mdiobus_create(struct enetc_pf *pf, struct device_node *node);
> +void enetc_mdiobus_destroy(struct enetc_pf *pf);
> +int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node *node,
> +			 const struct phylink_mac_ops *ops);
> +void enetc_phylink_destroy(struct enetc_ndev_priv *priv);

Could you put the prototypes of functions exported by enetc_pf_common.c
into a header named enetc_pf_common.h? It should be self-contained, i.e.
a dummy C file with just #include "enetc_pf_common.h" in it should compile
fine.

I know the enetc driver isn't there yet when it comes to thoroughly
respecting that, but for code we touch now, we should try to follow it.

Re: [PATCH v5 net-next 06/13] net: enetc: build enetc_pf_common.c as a separate module

[Vladimir Oltean]

On Thu, Oct 24, 2024 at 02:53:21PM +0800, Wei Fang wrote:
> diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.h b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> index 92a26b09cf57..39db9d5c2e50 100644
> --- a/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> +++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> @@ -28,6 +28,16 @@ struct enetc_vf_state {
>  	enum enetc_vf_flags flags;
>  };
>  
> +struct enetc_pf;
> +
> +struct enetc_pf_ops {
> +	void (*set_si_primary_mac)(struct enetc_hw *hw, int si, const u8 *addr);
> +	void (*get_si_primary_mac)(struct enetc_hw *hw, int si, u8 *addr);
> +	struct phylink_pcs *(*create_pcs)(struct enetc_pf *pf, struct mii_bus *bus);
> +	void (*destroy_pcs)(struct phylink_pcs *pcs);
> +	int (*enable_psfp)(struct enetc_ndev_priv *priv);
> +};
> +
>  struct enetc_pf {
>  	struct enetc_si *si;
>  	int num_vfs; /* number of active VFs, after sriov_init */
> @@ -50,6 +60,8 @@ struct enetc_pf {
>  
>  	phy_interface_t if_mode;
>  	struct phylink_config phylink_config;
> +
> +	const struct enetc_pf_ops *ops;
>  };
>  
>  #define phylink_to_enetc_pf(config) \
> @@ -59,9 +71,6 @@ int enetc_msg_psi_init(struct enetc_pf *pf);
>  void enetc_msg_psi_free(struct enetc_pf *pf);
>  void enetc_msg_handle_rxmsg(struct enetc_pf *pf, int mbox_id, u16 *status);
>  
> -void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr);
> -void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
> -				   const u8 *addr);
>  int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr);
>  int enetc_setup_mac_addresses(struct device_node *np, struct enetc_pf *pf);
>  void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
> @@ -71,3 +80,9 @@ void enetc_mdiobus_destroy(struct enetc_pf *pf);
>  int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node *node,
>  			 const struct phylink_mac_ops *ops);
>  void enetc_phylink_destroy(struct enetc_ndev_priv *priv);
> +
> +static inline void enetc_pf_ops_register(struct enetc_pf *pf,
> +					 const struct enetc_pf_ops *ops)
> +{
> +	pf->ops = ops;
> +}

I think this is more confusing than helpful? "register" suggests there
should also be an "unregister" coming. Either "set" or just open-code
the assignment?

> diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
> index bce81a4f6f88..94690ed92e3f 100644
> --- a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
> +++ b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
> @@ -8,19 +8,37 @@
>  
>  #include "enetc_pf.h"
>  
> +static int enetc_set_si_hw_addr(struct enetc_pf *pf, int si, u8 *mac_addr)
> +{
> +	struct enetc_hw *hw = &pf->si->hw;
> +
> +	if (pf->ops->set_si_primary_mac)
> +		pf->ops->set_si_primary_mac(hw, si, mac_addr);
> +	else
> +		return -EOPNOTSUPP;
> +
> +	return 0;

Don't artificially create errors when there are really no errors to handle.
Both enetc_pf_ops and enetc4_pf_ops provide .set_si_primary_mac(), so it
is unnecessary to handle the case where it isn't present. Those functions
return void, and void can be propagated to enetc_set_si_hw_addr() as well.

> +}
> +
>  int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr)
>  {
>  	struct enetc_ndev_priv *priv = netdev_priv(ndev);
> +	struct enetc_pf *pf = enetc_si_priv(priv->si);
>  	struct sockaddr *saddr = addr;
> +	int err;
>  
>  	if (!is_valid_ether_addr(saddr->sa_data))
>  		return -EADDRNOTAVAIL;
>  
> +	err = enetc_set_si_hw_addr(pf, 0, saddr->sa_data);
> +	if (err)
> +		return err;
> +
>  	eth_hw_addr_set(ndev, saddr->sa_data);
> -	enetc_pf_set_primary_mac_addr(&priv->si->hw, 0, saddr->sa_data);

This isn't one for one replacement, it also moves the function call
relative to eth_hw_addr_set() without making any mention about that
movement being safe. And even if safe, it is logically a separate change
which deserves its own merit analysis in another patch (if there's no
merit, leave it where it is).

I believe the merit was: enetc_set_si_hw_addr() can return error, thus
we simplify the control flow if we call it prior to eth_hw_addr_set() -
nothing to unroll. But as explained above, enetc_set_si_hw_addr() cannot
actually fail, so there is no real merit.

>  
>  	return 0;
>  }
> +EXPORT_SYMBOL_GPL(enetc_pf_set_mac_addr);
>  
>  static int enetc_setup_mac_address(struct device_node *np, struct enetc_pf *pf,
>  				   int si)
> @@ -38,8 +56,8 @@ static int enetc_setup_mac_address(struct device_node *np, struct enetc_pf *pf,
>  	}
>  
>  	/* (2) bootloader supplied MAC address */
> -	if (is_zero_ether_addr(mac_addr))
> -		enetc_pf_get_primary_mac_addr(hw, si, mac_addr);
> +	if (is_zero_ether_addr(mac_addr) && pf->ops->get_si_primary_mac)
> +		pf->ops->get_si_primary_mac(hw, si, mac_addr);

Again, if there's no reason for the method to be optional (both PF
drivers support it), don't make it optional.

A bit inconsistent that pf->ops->set_si_primary_mac() goes through a
wrapper function but this doesn't.

>  
>  	/* (3) choose a random one */
>  	if (is_zero_ether_addr(mac_addr)) {
> @@ -48,7 +66,9 @@ static int enetc_setup_mac_address(struct device_node *np, struct enetc_pf *pf,
>  			 si, mac_addr);
>  	}
>  
> -	enetc_pf_set_primary_mac_addr(hw, si, mac_addr);
> +	err = enetc_set_si_hw_addr(pf, si, mac_addr);
> +	if (err)
> +		return err;

This should go back to normal (no "err = ...; if (err) return err") once
the function is made void.

>  
>  	return 0;
>  }
> @@ -107,7 +129,8 @@ void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
>  			     NETDEV_XDP_ACT_NDO_XMIT | NETDEV_XDP_ACT_RX_SG |
>  			     NETDEV_XDP_ACT_NDO_XMIT_SG;
>  
> -	if (si->hw_features & ENETC_SI_F_PSFP && !enetc_psfp_enable(priv)) {
> +	if (si->hw_features & ENETC_SI_F_PSFP && pf->ops->enable_psfp &&
> +	    !pf->ops->enable_psfp(priv)) {

This one looks extremely weird in the existing code, but I suppose I'm
too late to the party to request you to clean up any of the PSFP code,
so I'll make a note to do it myself after your work. I haven't spotted
any actual bug, just weird coding patterns.

No change request here. I see the netc4_pf doesn't implement enable_psfp(),
so making it optional here is fine.

>  		priv->active_offloads |= ENETC_F_QCI;
>  		ndev->features |= NETIF_F_HW_TC;
>  		ndev->hw_features |= NETIF_F_HW_TC;
> @@ -116,6 +139,7 @@ void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
>  	/* pick up primary MAC address from SI */
>  	enetc_load_primary_mac_addr(&si->hw, ndev);
>  }
> +EXPORT_SYMBOL_GPL(enetc_pf_netdev_setup);
>  
>  static int enetc_mdio_probe(struct enetc_pf *pf, struct device_node *np)
>  {
> @@ -162,6 +186,9 @@ static int enetc_imdio_create(struct enetc_pf *pf)
>  	struct mii_bus *bus;
>  	int err;
>  
> +	if (!pf->ops->create_pcs)
> +		return -EOPNOTSUPP;
> +

I don't understand how this works. You don't implement create_pcs() for
netc4_pf until the very end of the series. Probing will fail for SerDes
interfaces (enetc_port_has_pcs() returns true) and that's fine?

A message maybe, stating what's the deal? Just that users figure out
quickly that it's an expected behavior, and not spend hours debugging
until they find out it's not their fault?

>  	bus = mdiobus_alloc_size(sizeof(*mdio_priv));
>  	if (!bus)
>  		return -ENOMEM;

Re: [PATCH v5 net-next 07/13] net: enetc: remove ERR050089 workaround for i.MX95

[Vladimir Oltean]

On Thu, Oct 24, 2024 at 02:53:22PM +0800, Wei Fang wrote:
> @@ -62,6 +65,12 @@ static int enetc_pci_mdio_probe(struct pci_dev *pdev,
>  		goto err_pci_mem_reg;
>  	}
>  
> +	if (pdev->vendor == PCI_VENDOR_ID_FREESCALE &&
> +	    pdev->device == ENETC_MDIO_DEV_ID) {
> +		static_branch_inc(&enetc_has_err050089);
> +		dev_info(&pdev->dev, "Enabled ERR050089 workaround\n");
> +	}
> +
>  	err = of_mdiobus_register(bus, dev->of_node);
>  	if (err)
>  		goto err_mdiobus_reg;

If of_mdiobus_register() fails, we should disable the static key.

Perhaps the snippets to enable and disable the static key should be in
helper functions to more easily insert the error teardown logic.

Re: [PATCH v5 net-next 09/13] net: enetc: add i.MX95 EMDIO support

[Vladimir Oltean]

On Thu, Oct 24, 2024 at 02:53:24PM +0800, Wei Fang wrote:
> The verdor ID and device ID of i.MX95 EMDIO are different from LS1028A
> EMDIO, so add new vendor ID and device ID to pci_device_id table to
> support i.MX95 EMDIO. And the i.MX95 EMDIO has two pins that need to be
> controlled, namely MDC and MDIO.
> 
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> Reviewed-by: Frank Li <Frank.Li@nxp.com>
> ---
> v5: no changes
> ---
>  drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c | 4 ++++
>  1 file changed, 4 insertions(+)
> 
> diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c b/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
> index 2445e35a764a..9968a1e9b5ef 100644
> --- a/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
> +++ b/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
> @@ -2,6 +2,7 @@
>  /* Copyright 2019 NXP */
>  #include <linux/fsl/enetc_mdio.h>
>  #include <linux/of_mdio.h>
> +#include <linux/pinctrl/consumer.h>
>  #include "enetc_pf.h"
>  
>  #define ENETC_MDIO_DEV_ID	0xee01
> @@ -71,6 +72,8 @@ static int enetc_pci_mdio_probe(struct pci_dev *pdev,
>  		dev_info(&pdev->dev, "Enabled ERR050089 workaround\n");
>  	}
>  
> +	pinctrl_pm_select_default_state(dev);
> +

Not an expert on pinctrl by any means.. but is this needed?
Documentation/driver-api/pin-control.rst says:

| When a device driver is about to probe the device core will automatically
| attempt to issue ``pinctrl_get_select_default()`` on these devices.
| This way driver writers do not need to add any of the boilerplate code
| of the type found below.

The documentation is obsolete, because pinctrl_get_select_default()
doesn't seem to be the current mechanism through which that happens. But
there is a pinctrl_bind_pins() function in really_probe() which looks
like it does that job. So.. is this needed?

Also, pinctrl_pm_select_default_state() seems to be the suspend/resume
(hence _pm_ name) variant of pinctrl_select_default_state(), but this is
called from probe().

>  	err = of_mdiobus_register(bus, dev->of_node);
>  	if (err)
>  		goto err_mdiobus_reg;
> @@ -113,6 +116,7 @@ static void enetc_pci_mdio_remove(struct pci_dev *pdev)
>  
>  static const struct pci_device_id enetc_pci_mdio_id_table[] = {
>  	{ PCI_DEVICE(PCI_VENDOR_ID_FREESCALE, ENETC_MDIO_DEV_ID) },
> +	{ PCI_DEVICE(PCI_VENDOR_ID_NXP2, PCI_DEVICE_ID_NXP2_NETC_EMDIO) },
>  	{ 0, } /* End of table. */
>  };
>  MODULE_DEVICE_TABLE(pci, enetc_pci_mdio_id_table);
> -- 
> 2.34.1
>

Re: [PATCH v5 net-next 09/13] net: enetc: add i.MX95 EMDIO support

[Vladimir Oltean]

On Thu, Oct 24, 2024 at 02:53:24PM +0800, Wei Fang wrote:
> The verdor ID and device ID of i.MX95 EMDIO are different from LS1028A

Also: s/verdor/vendor/

Re: [PATCH v5 net-next 08/13] PCI: Add NXP NETC vendor ID and device IDs

[Bjorn Helgaas]

On Thu, Oct 24, 2024 at 02:53:23PM +0800, Wei Fang wrote:
> NXP NETC is a multi-function RCiEP and it contains multiple functions,
> such as EMDIO, PTP Timer, ENETC PF and VF. Therefore, add these device
> IDs to pci_ids.h.
> 
> Below are the device IDs and corresponding drivers.
> PCI_DEVICE_ID_NXP2_ENETC_PF: nxp-enetc4
> PCI_DEVICE_ID_NXP2_NETC_EMDIO: fsl-enetc-mdio
> PCI_DEVICE_ID_NXP2_NETC_TIMER: ptp_netc
> PCI_DEVICE_ID_NXP2_ENETC_VF: fsl-enetc-vf
> 
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> Acked-by: Bjorn Helgaas <bhelgaas@google.com>

Please drop my ack.  I don't think these meet the spirit of the
guidance in pci_ids.h, which is there to minimize churn in that file
and make backports easier:

 *      Do not add new entries to this file unless the definitions
 *      are shared between multiple drivers.

PCI_DEVICE_ID_NXP2_NETC_TIMER and PCI_DEVICE_ID_NXP2_ENETC_VF aren't
used at all by this series, so they shouldn't be added to pci_ids.h.

PCI_DEVICE_ID_NXP2_NETC_EMDIO is used only by
drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c, so it should be
defined there, not in pci_ids.h.

PCI_DEVICE_ID_NXP2_ENETC_PF is used by enetc.c and enetc4_pf.c, but
it looks like those are basically part of the same driver, and it
could be defined in enetc4_hw.h or similar.

> ---
> v5: no changes
> ---
>  include/linux/pci_ids.h | 7 +++++++
>  1 file changed, 7 insertions(+)
> 
> diff --git a/include/linux/pci_ids.h b/include/linux/pci_ids.h
> index 4cf6aaed5f35..acd7ae774913 100644
> --- a/include/linux/pci_ids.h
> +++ b/include/linux/pci_ids.h
> @@ -1556,6 +1556,13 @@
>  #define PCI_DEVICE_ID_PHILIPS_SAA7146	0x7146
>  #define PCI_DEVICE_ID_PHILIPS_SAA9730	0x9730
>  
> +/* NXP has two vendor IDs, the other one is 0x1957 */
> +#define PCI_VENDOR_ID_NXP2		PCI_VENDOR_ID_PHILIPS
> +#define PCI_DEVICE_ID_NXP2_ENETC_PF	0xe101
> +#define PCI_DEVICE_ID_NXP2_NETC_EMDIO	0xee00
> +#define PCI_DEVICE_ID_NXP2_NETC_TIMER	0xee02
> +#define PCI_DEVICE_ID_NXP2_ENETC_VF	0xef00
> +
>  #define PCI_VENDOR_ID_EICON		0x1133
>  #define PCI_DEVICE_ID_EICON_DIVA20	0xe002
>  #define PCI_DEVICE_ID_EICON_DIVA20_U	0xe004
> -- 
> 2.34.1
> 

Re: [PATCH v5 net-next 10/13] net: enetc: extract enetc_int_vector_init/destroy() from enetc_alloc_msix()

[Vladimir Oltean]

On Thu, Oct 24, 2024 at 02:53:25PM +0800, Wei Fang wrote:
> From: Clark Wang <xiaoning.wang@nxp.com>
> 
> Extract enetc_int_vector_init() and enetc_int_vector_destroy() from
> enetc_alloc_msix() so that the code is more concise and readable. In
> addition, slightly different from before, the cleanup helper function
> is used to manage dynamically allocated memory resources.
> 
> Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> Reviewed-by: Frank Li <Frank.Li@nxp.com>
> Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
> ---
> v5: no changes
> ---
>  drivers/net/ethernet/freescale/enetc/enetc.c | 172 ++++++++++---------
>  1 file changed, 87 insertions(+), 85 deletions(-)
> 
> diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
> index 032d8eadd003..bd725561b8a2 100644
> --- a/drivers/net/ethernet/freescale/enetc/enetc.c
> +++ b/drivers/net/ethernet/freescale/enetc/enetc.c
> @@ -2965,6 +2965,87 @@ int enetc_ioctl(struct net_device *ndev, struct ifreq *rq, int cmd)
>  }
>  EXPORT_SYMBOL_GPL(enetc_ioctl);
>  
> +static int enetc_int_vector_init(struct enetc_ndev_priv *priv, int i,
> +				 int v_tx_rings)
> +{
> +	struct enetc_int_vector *v __free(kfree);

Please limit yourself to refactoring the code as-is into a separate function.
This function does not benefit in any way from the use of __free() and
no_free_ptr(). The established norm is that the normal teardown path
should be identical to the error unwind path, minus the last step.
Combining normal function calls with devres/scope-based cleanup/whatever
other "look, you don't get to care about error handling!!!" APIs there may be
makes that much more difficult to reason about. If the function is really
simple and you really don't get to care about error handling by using __free(),
then maybe its usage is tolerable, but that is hardly the general case.
The more intricate the error handling becomes, the less useful __free() is,
and the more it starts getting in the way of a human correctness reviewer.

FWIW, Documentation/process/maintainer-netdev.rst, section "Using
device-managed and cleanup.h constructs", appears to mostly state the
same position as me here.

Obviously, here, the established cleanup norm isn't followed anyway, but
a patch which brings it in line would be appreciated. I think that a
multi-patch approach, where the code is first moved and just moved, and
then successively transformed in obviously correct and easy to review
steps, would be best.

Since you're quite close currently to the 15-patch limit, I would try to
create a patch set just for preparing the enetc drivers, and introduce
the i.mx95 support in a separate set which has mostly "+" lines in its diff.
That way, there is also some time to not rush the ongoing IERB/PRB
dt-binding discussion, while you can progress on pure driver refactoring.

> +	struct enetc_bdr *bdr;
> +	int j, err;
> +
> +	v = kzalloc(struct_size(v, tx_ring, v_tx_rings), GFP_KERNEL);
> +	if (!v)
> +		return -ENOMEM;
> +
> +	bdr = &v->rx_ring;
> +	bdr->index = i;
> +	bdr->ndev = priv->ndev;
> +	bdr->dev = priv->dev;
> +	bdr->bd_count = priv->rx_bd_count;
> +	bdr->buffer_offset = ENETC_RXB_PAD;
> +	priv->rx_ring[i] = bdr;
> +
> +	err = xdp_rxq_info_reg(&bdr->xdp.rxq, priv->ndev, i, 0);
> +	if (err)
> +		return err;
> +
> +	err = xdp_rxq_info_reg_mem_model(&bdr->xdp.rxq,
> +					 MEM_TYPE_PAGE_SHARED, NULL);

MEM_TYPE_PAGE_SHARED fits on the previous line.

> +	if (err) {
> +		xdp_rxq_info_unreg(&bdr->xdp.rxq);
> +		return err;
> +	}
> +
> +	/* init defaults for adaptive IC */
> +	if (priv->ic_mode & ENETC_IC_RX_ADAPTIVE) {
> +		v->rx_ictt = 0x1;
> +		v->rx_dim_en = true;
> +	}
> +
> +	INIT_WORK(&v->rx_dim.work, enetc_rx_dim_work);
> +	netif_napi_add(priv->ndev, &v->napi, enetc_poll);
> +	v->count_tx_rings = v_tx_rings;
> +
> +	for (j = 0; j < v_tx_rings; j++) {
> +		int idx;
> +
> +		/* default tx ring mapping policy */
> +		idx = priv->bdr_int_num * j + i;
> +		__set_bit(idx, &v->tx_rings_map);
> +		bdr = &v->tx_ring[j];
> +		bdr->index = idx;
> +		bdr->ndev = priv->ndev;
> +		bdr->dev = priv->dev;
> +		bdr->bd_count = priv->tx_bd_count;
> +		priv->tx_ring[idx] = bdr;
> +	}
> +
> +	priv->int_vector[i] = no_free_ptr(v);
> +
> +	return 0;
> +}

RE: [PATCH v5 net-next 04/13] net: enetc: add initial netc-blk-ctrl driver support

[Wei Fang]

> On Thu, Oct 24, 2024 at 02:53:19PM +0800, Wei Fang wrote:
> > The netc-blk-ctrl driver is used to configure Integrated Endpoint
> > Register Block (IERB) and Privileged Register Block (PRB) of NETC.
> > For i.MX platforms, it is also used to configure the NETCMIX block.
> >
> > The IERB contains registers that are used for pre-boot initialization,
> > debug, and non-customer configuration. The PRB controls global reset
> > and global error handling for NETC. The NETCMIX block is mainly used
> > to set MII protocol and PCS protocol of the links, it also contains
> > settings for some other functions.
> >
> > Note the IERB configuration registers can only be written after being
> > unlocked by PRB, otherwise, all write operations are inhibited. A warm
> > reset is performed when the IERB is unlocked, and it results in an FLR
> > to all NETC devices. Therefore, all NETC device drivers must be probed
> > or initialized after the warm reset is finished.
> >
> > Signed-off-by: Wei Fang <wei.fang@nxp.com>
> > Reviewed-by: Frank Li <Frank.Li@nxp.com>
> > ---
> 
> Can U-Boot deal with the IERB/PRB configuration?
> 
> For LS1028A, the platform which initiated the IERB driver "trend", the situation
> was a bit more complicated, as we realized the reset-time defaults aren't what
> we need very late in the product life cycle, when customer boards already had
> bootloaders and we didn't want to complicate their process to have to redeploy
> in order to get access to such a basic feature as flow control. Though if we knew
> it from day one, we would have put the IERB fixups in U-Boot.

The situation of i.MX95 is different from LS1028A, i.MX95 needs to support system
suspend/resume feature. If the i.MX95 enters suspend mode, the NETC may
power off (depends on user case), so IERB and PRB will be reset, in this case, we need
to reconfigure the IERB & PRB, including NETCMIX.

> 
> What is written in the IERB for MII/PCS protocols by default? I suppose there's
> some other mechanism to preinitialize it with good values?

The MII/PCS protocols are set in NETCMIX not IERB, but the IERB will get these
info from NETCMIX, I mean the hardware, not the software. The default values
are all 0.

RE: [PATCH v5 net-next 05/13] net: enetc: extract common ENETC PF parts for LS1028A and i.MX95 platforms

[Wei Fang]

> On Thu, Oct 24, 2024 at 02:53:20PM +0800, Wei Fang wrote:
> > diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> > index c26bd66e4597..92a26b09cf57 100644
> > --- a/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> > +++ b/drivers/net/ethernet/freescale/enetc/enetc_pf.h
> > @@ -58,3 +58,16 @@ struct enetc_pf {
> >  int enetc_msg_psi_init(struct enetc_pf *pf);
> >  void enetc_msg_psi_free(struct enetc_pf *pf);
> >  void enetc_msg_handle_rxmsg(struct enetc_pf *pf, int mbox_id, u16
> *status);
> > +
> > +void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8
> *addr);
> > +void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
> > +				   const u8 *addr);
> > +int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr);
> > +int enetc_setup_mac_addresses(struct device_node *np, struct enetc_pf
> *pf);
> > +void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
> > +			   const struct net_device_ops *ndev_ops);
> > +int enetc_mdiobus_create(struct enetc_pf *pf, struct device_node *node);
> > +void enetc_mdiobus_destroy(struct enetc_pf *pf);
> > +int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node
> *node,
> > +			 const struct phylink_mac_ops *ops);
> > +void enetc_phylink_destroy(struct enetc_ndev_priv *priv);
> 
> Could you put the prototypes of functions exported by enetc_pf_common.c
> into a header named enetc_pf_common.h? It should be self-contained, i.e.
> a dummy C file with just #include "enetc_pf_common.h" in it should compile
> fine.
> 

Sure, I'll add a new header file.

> I know the enetc driver isn't there yet when it comes to thoroughly
> respecting that, but for code we touch now, we should try to follow it.

RE: [PATCH v5 net-next 06/13] net: enetc: build enetc_pf_common.c as a separate module

[Wei Fang]

> > +struct enetc_pf;
> > +
> > +struct enetc_pf_ops {
> > +	void (*set_si_primary_mac)(struct enetc_hw *hw, int si, const u8 *addr);
> > +	void (*get_si_primary_mac)(struct enetc_hw *hw, int si, u8 *addr);
> > +	struct phylink_pcs *(*create_pcs)(struct enetc_pf *pf, struct mii_bus *bus);
> > +	void (*destroy_pcs)(struct phylink_pcs *pcs);
> > +	int (*enable_psfp)(struct enetc_ndev_priv *priv);
> > +};
> > +
> >  struct enetc_pf {
> >  	struct enetc_si *si;
> >  	int num_vfs; /* number of active VFs, after sriov_init */
> > @@ -50,6 +60,8 @@ struct enetc_pf {
> >
> >  	phy_interface_t if_mode;
> >  	struct phylink_config phylink_config;
> > +
> > +	const struct enetc_pf_ops *ops;
> >  };
> >
> >  #define phylink_to_enetc_pf(config) \
> > @@ -59,9 +71,6 @@ int enetc_msg_psi_init(struct enetc_pf *pf);
> >  void enetc_msg_psi_free(struct enetc_pf *pf);
> >  void enetc_msg_handle_rxmsg(struct enetc_pf *pf, int mbox_id, u16
> *status);
> >
> > -void enetc_pf_get_primary_mac_addr(struct enetc_hw *hw, int si, u8 *addr);
> > -void enetc_pf_set_primary_mac_addr(struct enetc_hw *hw, int si,
> > -				   const u8 *addr);
> >  int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr);
> >  int enetc_setup_mac_addresses(struct device_node *np, struct enetc_pf
> *pf);
> >  void enetc_pf_netdev_setup(struct enetc_si *si, struct net_device *ndev,
> > @@ -71,3 +80,9 @@ void enetc_mdiobus_destroy(struct enetc_pf *pf);
> >  int enetc_phylink_create(struct enetc_ndev_priv *priv, struct device_node
> *node,
> >  			 const struct phylink_mac_ops *ops);
> >  void enetc_phylink_destroy(struct enetc_ndev_priv *priv);
> > +
> > +static inline void enetc_pf_ops_register(struct enetc_pf *pf,
> > +					 const struct enetc_pf_ops *ops)
> > +{
> > +	pf->ops = ops;
> > +}
> 
> I think this is more confusing than helpful? "register" suggests there
> should also be an "unregister" coming. Either "set" or just open-code
> the assignment?

Okay, I can remove this helper function, just use open-code.

> 
> > diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
> b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
> > index bce81a4f6f88..94690ed92e3f 100644
> > --- a/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
> > +++ b/drivers/net/ethernet/freescale/enetc/enetc_pf_common.c
> > @@ -8,19 +8,37 @@
> >
> >  #include "enetc_pf.h"
> >
> > +static int enetc_set_si_hw_addr(struct enetc_pf *pf, int si, u8 *mac_addr)
> > +{
> > +	struct enetc_hw *hw = &pf->si->hw;
> > +
> > +	if (pf->ops->set_si_primary_mac)
> > +		pf->ops->set_si_primary_mac(hw, si, mac_addr);
> > +	else
> > +		return -EOPNOTSUPP;
> > +
> > +	return 0;
> 
> Don't artificially create errors when there are really no errors to handle.
> Both enetc_pf_ops and enetc4_pf_ops provide .set_si_primary_mac(), so it
> is unnecessary to handle the case where it isn't present. Those functions
> return void, and void can be propagated to enetc_set_si_hw_addr() as well.
> 

I thought checking the pointer is safer, so you mean that pointers that are
definitely present in the current driver do not need to be checked?

> > +}
> > +
> >  int enetc_pf_set_mac_addr(struct net_device *ndev, void *addr)
> >  {
> >  	struct enetc_ndev_priv *priv = netdev_priv(ndev);
> > +	struct enetc_pf *pf = enetc_si_priv(priv->si);
> >  	struct sockaddr *saddr = addr;
> > +	int err;
> >
> >  	if (!is_valid_ether_addr(saddr->sa_data))
> >  		return -EADDRNOTAVAIL;
> >
> > +	err = enetc_set_si_hw_addr(pf, 0, saddr->sa_data);
> > +	if (err)
> > +		return err;
> > +
> >  	eth_hw_addr_set(ndev, saddr->sa_data);
> > -	enetc_pf_set_primary_mac_addr(&priv->si->hw, 0, saddr->sa_data);
> 
> This isn't one for one replacement, it also moves the function call
> relative to eth_hw_addr_set() without making any mention about that
> movement being safe. And even if safe, it is logically a separate change
> which deserves its own merit analysis in another patch (if there's no
> merit, leave it where it is).
> 
> I believe the merit was: enetc_set_si_hw_addr() can return error, thus
> we simplify the control flow if we call it prior to eth_hw_addr_set() -
> nothing to unroll. But as explained above, enetc_set_si_hw_addr() cannot
> actually fail, so there is no real merit.
> 
> >
> >  	return 0;
> >  }
> > +EXPORT_SYMBOL_GPL(enetc_pf_set_mac_addr);
> >
> >  static int enetc_setup_mac_address(struct device_node *np, struct
> enetc_pf *pf,
> >  				   int si)
> > @@ -38,8 +56,8 @@ static int enetc_setup_mac_address(struct device_node
> *np, struct enetc_pf *pf,
> >  	}
> >
> >  	/* (2) bootloader supplied MAC address */
> > -	if (is_zero_ether_addr(mac_addr))
> > -		enetc_pf_get_primary_mac_addr(hw, si, mac_addr);
> > +	if (is_zero_ether_addr(mac_addr) && pf->ops->get_si_primary_mac)
> > +		pf->ops->get_si_primary_mac(hw, si, mac_addr);
> 
> Again, if there's no reason for the method to be optional (both PF
> drivers support it), don't make it optional.
> 
> A bit inconsistent that pf->ops->set_si_primary_mac() goes through a
> wrapper function but this doesn't.
> 

If we really do not need to check these callback pointers, then I think I can
remove the wrapper.

> >
> >  	/* (3) choose a random one */
> >  	if (is_zero_ether_addr(mac_addr)) {
> > @@ -48,7 +66,9 @@ static int enetc_setup_mac_address(struct device_node
> *np, struct enetc_pf *pf,
> >  			 si, mac_addr);
> >  	}
> >
> > -	enetc_pf_set_primary_mac_addr(hw, si, mac_addr);
> > +	err = enetc_set_si_hw_addr(pf, si, mac_addr);
> > +	if (err)
> > +		return err;
> 
> This should go back to normal (no "err = ...; if (err) return err") once
> the function is made void.
> 
> >
> >  	return 0;
> >  }
> > @@ -107,7 +129,8 @@ void enetc_pf_netdev_setup(struct enetc_si *si,
> struct net_device *ndev,
> >  			     NETDEV_XDP_ACT_NDO_XMIT |
> NETDEV_XDP_ACT_RX_SG |
> >  			     NETDEV_XDP_ACT_NDO_XMIT_SG;
> >
> > -	if (si->hw_features & ENETC_SI_F_PSFP && !enetc_psfp_enable(priv)) {
> > +	if (si->hw_features & ENETC_SI_F_PSFP && pf->ops->enable_psfp &&
> > +	    !pf->ops->enable_psfp(priv)) {
> 
> This one looks extremely weird in the existing code, but I suppose I'm
> too late to the party to request you to clean up any of the PSFP code,
> so I'll make a note to do it myself after your work. I haven't spotted
> any actual bug, just weird coding patterns.
> 
> No change request here. I see the netc4_pf doesn't implement enable_psfp(),
> so making it optional here is fine.

Yes, PSFP is not supported in this patch set, I will remove it in future.

> 
> >  		priv->active_offloads |= ENETC_F_QCI;
> >  		ndev->features |= NETIF_F_HW_TC;
> >  		ndev->hw_features |= NETIF_F_HW_TC;
> > @@ -116,6 +139,7 @@ void enetc_pf_netdev_setup(struct enetc_si *si,
> struct net_device *ndev,
> >  	/* pick up primary MAC address from SI */
> >  	enetc_load_primary_mac_addr(&si->hw, ndev);
> >  }
> > +EXPORT_SYMBOL_GPL(enetc_pf_netdev_setup);
> >
> >  static int enetc_mdio_probe(struct enetc_pf *pf, struct device_node *np)
> >  {
> > @@ -162,6 +186,9 @@ static int enetc_imdio_create(struct enetc_pf *pf)
> >  	struct mii_bus *bus;
> >  	int err;
> >
> > +	if (!pf->ops->create_pcs)
> > +		return -EOPNOTSUPP;
> > +
> 
> I don't understand how this works. You don't implement create_pcs() for
> netc4_pf until the very end of the series. Probing will fail for SerDes
> interfaces (enetc_port_has_pcs() returns true) and that's fine?
> 

Currently, we have not add the PCS support, so the 10G ENETC is not supported
yet. And we also disable the 10G ENETC in DTS. Only the 1G ENETCs (without PCS)
are supported for i.MX95.

> A message maybe, stating what's the deal? Just that users figure out
> quickly that it's an expected behavior, and not spend hours debugging
> until they find out it's not their fault?

I will explain in the commit message that i.MX95 10G ENETC is not currently
supported.

> 
> >  	bus = mdiobus_alloc_size(sizeof(*mdio_priv));
> >  	if (!bus)
> >  		return -ENOMEM;

RE: [PATCH v5 net-next 07/13] net: enetc: remove ERR050089 workaround for i.MX95

[Wei Fang]

> On Thu, Oct 24, 2024 at 02:53:22PM +0800, Wei Fang wrote:
> > @@ -62,6 +65,12 @@ static int enetc_pci_mdio_probe(struct pci_dev *pdev,
> >  		goto err_pci_mem_reg;
> >  	}
> >
> > +	if (pdev->vendor == PCI_VENDOR_ID_FREESCALE &&
> > +	    pdev->device == ENETC_MDIO_DEV_ID) {
> > +		static_branch_inc(&enetc_has_err050089);
> > +		dev_info(&pdev->dev, "Enabled ERR050089 workaround\n");
> > +	}
> > +
> >  	err = of_mdiobus_register(bus, dev->of_node);
> >  	if (err)
> >  		goto err_mdiobus_reg;
> 
> If of_mdiobus_register() fails, we should disable the static key.

> 
> Perhaps the snippets to enable and disable the static key should be in
> helper functions to more easily insert the error teardown logic.

Okay, I will refine it

RE: [PATCH v5 net-next 09/13] net: enetc: add i.MX95 EMDIO support

[Wei Fang]

> >
> > Signed-off-by: Wei Fang <wei.fang@nxp.com>
> > Reviewed-by: Frank Li <Frank.Li@nxp.com>
> > ---
> > v5: no changes
> > ---
> >  drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c | 4 ++++
> >  1 file changed, 4 insertions(+)
> >
> > diff --git a/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
> b/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
> > index 2445e35a764a..9968a1e9b5ef 100644
> > --- a/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
> > +++ b/drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c
> > @@ -2,6 +2,7 @@
> >  /* Copyright 2019 NXP */
> >  #include <linux/fsl/enetc_mdio.h>
> >  #include <linux/of_mdio.h>
> > +#include <linux/pinctrl/consumer.h>
> >  #include "enetc_pf.h"
> >
> >  #define ENETC_MDIO_DEV_ID	0xee01
> > @@ -71,6 +72,8 @@ static int enetc_pci_mdio_probe(struct pci_dev *pdev,
> >  		dev_info(&pdev->dev, "Enabled ERR050089 workaround\n");
> >  	}
> >
> > +	pinctrl_pm_select_default_state(dev);
> > +
> 
> Not an expert on pinctrl by any means.. but is this needed?
> Documentation/driver-api/pin-control.rst says:
> 
> | When a device driver is about to probe the device core will automatically
> | attempt to issue ``pinctrl_get_select_default()`` on these devices.
> | This way driver writers do not need to add any of the boilerplate code
> | of the type found below.
> 
> The documentation is obsolete, because pinctrl_get_select_default()
> doesn't seem to be the current mechanism through which that happens. But
> there is a pinctrl_bind_pins() function in really_probe() which looks
> like it does that job. So.. is this needed?

I think it is no need, I will remove it, thanks.

RE: [PATCH v5 net-next 08/13] PCI: Add NXP NETC vendor ID and device IDs

[Wei Fang]

> On Thu, Oct 24, 2024 at 02:53:23PM +0800, Wei Fang wrote:
> > NXP NETC is a multi-function RCiEP and it contains multiple functions,
> > such as EMDIO, PTP Timer, ENETC PF and VF. Therefore, add these device
> > IDs to pci_ids.h.
> >
> > Below are the device IDs and corresponding drivers.
> > PCI_DEVICE_ID_NXP2_ENETC_PF: nxp-enetc4
> > PCI_DEVICE_ID_NXP2_NETC_EMDIO: fsl-enetc-mdio
> > PCI_DEVICE_ID_NXP2_NETC_TIMER: ptp_netc
> > PCI_DEVICE_ID_NXP2_ENETC_VF: fsl-enetc-vf
> >
> > Signed-off-by: Wei Fang <wei.fang@nxp.com>
> > Acked-by: Bjorn Helgaas <bhelgaas@google.com>
> 
> Please drop my ack.  I don't think these meet the spirit of the guidance in
> pci_ids.h, which is there to minimize churn in that file and make backports
> easier:
> 
>  *      Do not add new entries to this file unless the definitions
>  *      are shared between multiple drivers.
> 
> PCI_DEVICE_ID_NXP2_NETC_TIMER and PCI_DEVICE_ID_NXP2_ENETC_VF
> aren't used at all by this series, so they shouldn't be added to pci_ids.h.
> 
> PCI_DEVICE_ID_NXP2_NETC_EMDIO is used only by
> drivers/net/ethernet/freescale/enetc/enetc_pci_mdio.c, so it should be defined
> there, not in pci_ids.h.
> 
> PCI_DEVICE_ID_NXP2_ENETC_PF is used by enetc.c and enetc4_pf.c, but it looks
> like those are basically part of the same driver, and it could be defined in
> enetc4_hw.h or similar.
> 

I was not aware of this, I will remove this patch, thanks.

RE: [PATCH v5 net-next 10/13] net: enetc: extract enetc_int_vector_init/destroy() from enetc_alloc_msix()

[Wei Fang]

> On Thu, Oct 24, 2024 at 02:53:25PM +0800, Wei Fang wrote:
> > From: Clark Wang <xiaoning.wang@nxp.com>
> >
> > Extract enetc_int_vector_init() and enetc_int_vector_destroy() from
> > enetc_alloc_msix() so that the code is more concise and readable. In
> > addition, slightly different from before, the cleanup helper function
> > is used to manage dynamically allocated memory resources.
> >
> > Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
> > Signed-off-by: Wei Fang <wei.fang@nxp.com>
> > Reviewed-by: Frank Li <Frank.Li@nxp.com>
> > Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
> > ---
> > v5: no changes
> > ---
> >  drivers/net/ethernet/freescale/enetc/enetc.c | 172 ++++++++++---------
> >  1 file changed, 87 insertions(+), 85 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c
> b/drivers/net/ethernet/freescale/enetc/enetc.c
> > index 032d8eadd003..bd725561b8a2 100644
> > --- a/drivers/net/ethernet/freescale/enetc/enetc.c
> > +++ b/drivers/net/ethernet/freescale/enetc/enetc.c
> > @@ -2965,6 +2965,87 @@ int enetc_ioctl(struct net_device *ndev, struct
> ifreq *rq, int cmd)
> >  }
> >  EXPORT_SYMBOL_GPL(enetc_ioctl);
> >
> > +static int enetc_int_vector_init(struct enetc_ndev_priv *priv, int i,
> > +				 int v_tx_rings)
> > +{
> > +	struct enetc_int_vector *v __free(kfree);
> 
> Please limit yourself to refactoring the code as-is into a separate function.

Okay

> This function does not benefit in any way from the use of __free() and
> no_free_ptr(). The established norm is that the normal teardown path
> should be identical to the error unwind path, minus the last step.
> Combining normal function calls with devres/scope-based cleanup/whatever
> other "look, you don't get to care about error handling!!!" APIs there may be
> makes that much more difficult to reason about. If the function is really
> simple and you really don't get to care about error handling by using __free(),
> then maybe its usage is tolerable, but that is hardly the general case.
> The more intricate the error handling becomes, the less useful __free() is,
> and the more it starts getting in the way of a human correctness reviewer.
> 
> FWIW, Documentation/process/maintainer-netdev.rst, section "Using
> device-managed and cleanup.h constructs", appears to mostly state the
> same position as me here.
> 
> Obviously, here, the established cleanup norm isn't followed anyway, but
> a patch which brings it in line would be appreciated. I think that a
> multi-patch approach, where the code is first moved and just moved, and
> then successively transformed in obviously correct and easy to review
> steps, would be best.
> 
> Since you're quite close currently to the 15-patch limit, I would try to
> create a patch set just for preparing the enetc drivers, and introduce
> the i.mx95 support in a separate set which has mostly "+" lines in its diff.
> That way, there is also some time to not rush the ongoing IERB/PRB
> dt-binding discussion, while you can progress on pure driver refactoring.
> 

Thanks for your suggestion.

Re: [PATCH v5 net-next 11/13] net: enetc: optimize the allocation of tx_bdr

[Vladimir Oltean]

On Thu, Oct 24, 2024 at 02:53:26PM +0800, Wei Fang wrote:
> From: Clark Wang <xiaoning.wang@nxp.com>
> 
> There is a situation where num_tx_rings cannot be divided by bdr_int_num.
> For example, num_tx_rings is 8 and bdr_int_num is 3. According to the
> previous logic, this results in two tx_bdr corresponding memories not
> being allocated, so when sending packets to tx ring 6 or 7, wild pointers
> will be accessed. Of course, this issue doesn't exist on LS1028A, because
> its num_tx_rings is 8, and bdr_int_num is either 1 or 2. However, there
> is a risk for the upcoming i.MX95. Therefore, it is necessary to ensure
> that each tx_bdr can be allocated to the corresponding memory.
> 
> Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
> Reviewed-by: Frank Li <Frank.Li@nxp.com>
> ---
> v5: no changes
> ---
>  drivers/net/ethernet/freescale/enetc/enetc.c | 7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c b/drivers/net/ethernet/freescale/enetc/enetc.c
> index bd725561b8a2..bccbeb1f355c 100644
> --- a/drivers/net/ethernet/freescale/enetc/enetc.c
> +++ b/drivers/net/ethernet/freescale/enetc/enetc.c
> @@ -3049,10 +3049,10 @@ static void enetc_int_vector_destroy(struct enetc_ndev_priv *priv, int i)
>  int enetc_alloc_msix(struct enetc_ndev_priv *priv)
>  {
>  	struct pci_dev *pdev = priv->si->pdev;
> +	int v_tx_rings, v_remainder;
>  	int num_stack_tx_queues;
>  	int first_xdp_tx_ring;
>  	int i, n, err, nvec;
> -	int v_tx_rings;
>  
>  	nvec = ENETC_BDR_INT_BASE_IDX + priv->bdr_int_num;
>  	/* allocate MSIX for both messaging and Rx/Tx interrupts */
> @@ -3066,9 +3066,12 @@ int enetc_alloc_msix(struct enetc_ndev_priv *priv)
>  
>  	/* # of tx rings per int vector */
>  	v_tx_rings = priv->num_tx_rings / priv->bdr_int_num;
> +	v_remainder = priv->num_tx_rings % priv->bdr_int_num;
>  
>  	for (i = 0; i < priv->bdr_int_num; i++) {
> -		err = enetc_int_vector_init(priv, i, v_tx_rings);
> +		int num_tx_rings = i < v_remainder ? v_tx_rings + 1 : v_tx_rings;

It took me a moment to understand the mechanism through which this
works, even though I read the intention in the commit message.

Do you think this additional comment would help?

		/* Distribute the remaining TX rings to the first
		 * v_tx_rings interrupt vectors
		 */

> +
> +		err = enetc_int_vector_init(priv, i, num_tx_rings);
>  		if (err)
>  			goto fail;
>  	}
> -- 
> 2.34.1
>

Re: [PATCH v5 net-next 04/13] net: enetc: add initial netc-blk-ctrl driver support

[Vladimir Oltean]

On Fri, Oct 25, 2024 at 04:44:50AM +0300, Wei Fang wrote:
> > On Thu, Oct 24, 2024 at 02:53:19PM +0800, Wei Fang wrote:
> > Can U-Boot deal with the IERB/PRB configuration?
> > 
> > For LS1028A, the platform which initiated the IERB driver "trend", the situation
> > was a bit more complicated, as we realized the reset-time defaults aren't what
> > we need very late in the product life cycle, when customer boards already had
> > bootloaders and we didn't want to complicate their process to have to redeploy
> > in order to get access to such a basic feature as flow control. Though if we knew
> > it from day one, we would have put the IERB fixups in U-Boot.
> 
> The situation of i.MX95 is different from LS1028A, i.MX95 needs to support system
> suspend/resume feature. If the i.MX95 enters suspend mode, the NETC may
> power off (depends on user case), so IERB and PRB will be reset, in this case, we need
> to reconfigure the IERB & PRB, including NETCMIX.
> 
> > What is written in the IERB for MII/PCS protocols by default? I suppose there's
> > some other mechanism to preinitialize it with good values?
> 
> The MII/PCS protocols are set in NETCMIX not IERB, but the IERB will get these
> info from NETCMIX, I mean the hardware, not the software. The default values
> are all 0.

I am shocked that the NETCMIX/IERB blocks does not have a separate power
domain from the ENETC, to avoid powering them off, which loses the settings.
Please provide this explanation in the opening comments of this driver, it
is its entire "raison d'être".

Re: [PATCH v5 net-next 06/13] net: enetc: build enetc_pf_common.c as a separate module

[Vladimir Oltean]

On Fri, Oct 25, 2024 at 06:00:42AM +0300, Wei Fang wrote:
> > Don't artificially create errors when there are really no errors to handle.
> > Both enetc_pf_ops and enetc4_pf_ops provide .set_si_primary_mac(), so it
> > is unnecessary to handle the case where it isn't present. Those functions
> > return void, and void can be propagated to enetc_set_si_hw_addr() as well.
> 
> I thought checking the pointer is safer, so you mean that pointers that are
> definitely present in the current driver do not need to be checked?

Yes, there is no point to check for a condition which is impossible
to trigger in the current code. The callee and the caller are tightly
coupled (in the same driver), it's not a rigid API, so if the function
pointers should be made optional for some future hardware IP, the error
handling will be added when necessary. Ideally any change passes through
review, and any inconsistency should be spotted when added.

> > A bit inconsistent that pf->ops->set_si_primary_mac() goes through a
> > wrapper function but this doesn't.
> > 
> 
> If we really do not need to check these callback pointers, then I think I can
> remove the wrapper.

Fine without wrapping throughout, my comment was first and foremost
about consistency.

> > This one looks extremely weird in the existing code, but I suppose I'm
> > too late to the party to request you to clean up any of the PSFP code,
> > so I'll make a note to do it myself after your work. I haven't spotted
> > any actual bug, just weird coding patterns.
> > 
> > No change request here. I see the netc4_pf doesn't implement enable_psfp(),
> > so making it optional here is fine.
> 
> Yes, PSFP is not supported in this patch set, I will remove it in future.

If by "I will remove it in future" you mean "once NETC4 gains PSFP
support, I will make enable_psfp() non-optional", then ok, great.

> Currently, we have not add the PCS support, so the 10G ENETC is not supported
> yet. And we also disable the 10G ENETC in DTS. Only the 1G ENETCs (without PCS)
> are supported for i.MX95.

Also think about the case where the current version of the kernel
will boot on a newer version of the device tree, which does not have
'status = "disabled";' for those ports. It should do something reasonable.
In any case, "they're now disabled in the device tree" is not an argument.

My anecdotal and vague understanding of the Arm SystemReady (IR I think)
requirements is that the device tree is provided by the platform,
separately from the kernel/rootfs. It relies on the device tree ABI
being stable, backwards-compatible and forwards-compatible.

> > A message maybe, stating what's the deal? Just that users figure out
> > quickly that it's an expected behavior, and not spend hours debugging
> > until they find out it's not their fault?
> 
> I will explain in the commit message that i.MX95 10G ENETC is not currently
> supported.

By "a message, maybe" I actually meant dev_err("Message here\n"); rather
than silent/imprecise failure.

Re: [PATCH v5 net-next 02/13] dt-bindings: net: add i.MX95 ENETC support

[Rob Herring]

On Thu, Oct 24, 2024 at 02:53:17PM +0800, Wei Fang wrote:
> The ENETC of i.MX95 has been upgraded to revision 4.1, and the vendor
> ID and device ID have also changed, so add the new compatible strings
> for i.MX95 ENETC. In addition, i.MX95 supports configuration of RGMII
> or RMII reference clock.
> 
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> ---
> v2: Remove "nxp,imx95-enetc" compatible string.
> v3:
> 1. Add restriction to "clcoks" and "clock-names" properties and rename
> the clock, also remove the items from these two properties.
> 2. Remove unnecessary items for "pci1131,e101" compatible string.
> v4: Move clocks and clock-names to top level.
> v5: Add items to clocks and clock-names
> ---
>  .../devicetree/bindings/net/fsl,enetc.yaml    | 34 +++++++++++++++++--
>  1 file changed, 31 insertions(+), 3 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/net/fsl,enetc.yaml b/Documentation/devicetree/bindings/net/fsl,enetc.yaml
> index e152c93998fe..72d2d5d285cd 100644
> --- a/Documentation/devicetree/bindings/net/fsl,enetc.yaml
> +++ b/Documentation/devicetree/bindings/net/fsl,enetc.yaml
> @@ -20,14 +20,23 @@ maintainers:
>  
>  properties:
>    compatible:
> -    items:
> +    oneOf:
> +      - items:
> +          - enum:
> +              - pci1957,e100
> +          - const: fsl,enetc
>        - enum:
> -          - pci1957,e100
> -      - const: fsl,enetc
> +          - pci1131,e101
>  
>    reg:
>      maxItems: 1
>  
> +  clocks:
> +    maxItems: 1
> +
> +  clock-names:
> +    maxItems: 1
> +
>    mdio:
>      $ref: mdio.yaml
>      unevaluatedProperties: false
> @@ -40,6 +49,25 @@ required:
>  allOf:
>    - $ref: /schemas/pci/pci-device.yaml
>    - $ref: ethernet-controller.yaml
> +  - if:
> +      properties:
> +        compatible:
> +          contains:
> +            enum:
> +              - pci1131,e101
> +    then:
> +      properties:
> +        clocks:
> +          items:
> +            - description: MAC transmit/receiver reference clock

This goes in the top-level or can just be dropped.

> +
> +        clock-names:
> +          items:
> +            - const: ref

This goes in the top-level.


> +    else:

Then negate the 'if' schema (not: contains: ...) and you just need the 
part below:

> +      properties:
> +        clocks: false
> +        clock-names: false
>  
>  unevaluatedProperties: false
>  
> -- 
> 2.34.1
> 

Re: [PATCH v5 net-next 03/13] dt-bindings: net: add bindings for NETC blocks control

[Rob Herring (Arm)]

On Thu, 24 Oct 2024 14:53:18 +0800, Wei Fang wrote:
> Add bindings for NXP NETC blocks control. Usually, NETC has 2 blocks of
> 64KB registers, integrated endpoint register block (IERB) and privileged
> register block (PRB). IERB is used for pre-boot initialization for all
> NETC devices, such as ENETC, Timer, EMDIO and so on. And PRB controls
> global reset and global error handling for NETC. Moreover, for the i.MX
> platform, there is also a NETCMIX block for link configuration, such as
> MII protocol, PCS protocol, etc.
> 
> Signed-off-by: Wei Fang <wei.fang@nxp.com>
> ---
> v2 changes:
> 1. Rephrase the commit message.
> 2. Change unevaluatedProperties to additionalProperties.
> 3. Remove the useless lables from examples.
> v3 changes:
> 1. Remove the items from clocks and clock-names, add maxItems to clocks
> and rename the clock.
> v4 changes:
> 1. Reorder the required properties.
> 2. Add assigned-clocks, assigned-clock-parents and assigned-clock-rates.
> v5 changes:
> 1. Remove assigned-clocks, assigned-clock-parents and assigned-clock-rates
> 2. Remove minItems from reg and reg-names
> 3. Rename the node in the examples to be more general
> ---
>  .../bindings/net/nxp,netc-blk-ctrl.yaml       | 104 ++++++++++++++++++
>  1 file changed, 104 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/net/nxp,netc-blk-ctrl.yaml
> 

Reviewed-by: Rob Herring (Arm) <robh@kernel.org>

Re: [PATCH v5 net-next 05/13] net: enetc: extract common ENETC PF parts for LS1028A and i.MX95 platforms

[kernel test robot]

Hi Wei,

kernel test robot noticed the following build errors:

[auto build test ERROR on net-next/main]

url:    https://github.com/intel-lab-lkp/linux/commits/Wei-Fang/dt-bindings-net-add-compatible-string-for-i-MX95-EMDIO/20241024-151502
base:   net-next/main
patch link:    https://lore.kernel.org/r/20241024065328.521518-6-wei.fang%40nxp.com
patch subject: [PATCH v5 net-next 05/13] net: enetc: extract common ENETC PF parts for LS1028A and i.MX95 platforms
config: x86_64-buildonly-randconfig-001-20241026 (https://download.01.org/0day-ci/archive/20241026/202410260911.fvWnX8cx-lkp@intel.com/config)
compiler: clang version 19.1.2 (https://github.com/llvm/llvm-project 7ba7d8e2f7b6445b60679da826210cdde29eaf8b)
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20241026/202410260911.fvWnX8cx-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202410260911.fvWnX8cx-lkp@intel.com/

All errors (new ones prefixed by >>):

   In file included from drivers/net/ethernet/freescale/enetc/enetc_pf.c:7:
   In file included from include/linux/of_net.h:9:
   In file included from include/linux/phy.h:16:
   In file included from include/linux/ethtool.h:18:
   In file included from include/linux/if_ether.h:19:
   In file included from include/linux/skbuff.h:17:
   In file included from include/linux/bvec.h:10:
   In file included from include/linux/highmem.h:8:
   In file included from include/linux/cacheflush.h:5:
   In file included from arch/x86/include/asm/cacheflush.h:5:
   In file included from include/linux/mm.h:2213:
   include/linux/vmstat.h:518:36: warning: arithmetic between different enumeration types ('enum node_stat_item' and 'enum lru_list') [-Wenum-enum-conversion]
     518 |         return node_stat_name(NR_LRU_BASE + lru) + 3; // skip "nr_"
         |                               ~~~~~~~~~~~ ^ ~~~
>> drivers/net/ethernet/freescale/enetc/enetc_pf.c:906:14: error: call to undeclared function 'of_find_compatible_node'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
     906 |         ierb_node = of_find_compatible_node(NULL, NULL,
         |                     ^
>> drivers/net/ethernet/freescale/enetc/enetc_pf.c:906:12: error: incompatible integer to pointer conversion assigning to 'struct device_node *' from 'int' [-Wint-conversion]
     906 |         ierb_node = of_find_compatible_node(NULL, NULL,
         |                   ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
     907 |                                             "fsl,ls1028a-enetc-ierb");
         |                                             ~~~~~~~~~~~~~~~~~~~~~~~~~
>> drivers/net/ethernet/freescale/enetc/enetc_pf.c:908:21: error: call to undeclared function 'of_device_is_available'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
     908 |         if (!ierb_node || !of_device_is_available(ierb_node))
         |                            ^
>> drivers/net/ethernet/freescale/enetc/enetc_pf.c:912:2: error: call to undeclared function 'of_node_put'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
     912 |         of_node_put(ierb_node);
         |         ^
   drivers/net/ethernet/freescale/enetc/enetc_pf.c:1115:14: error: call to undeclared function 'of_device_is_available'; ISO C99 and later do not support implicit function declarations [-Wimplicit-function-declaration]
    1115 |         if (node && of_device_is_available(node))
         |                     ^
   1 warning and 5 errors generated.


vim +/of_find_compatible_node +906 drivers/net/ethernet/freescale/enetc/enetc_pf.c

07bf34a50e3279 Vladimir Oltean 2021-02-04  900  
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  901  static int enetc_pf_register_with_ierb(struct pci_dev *pdev)
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  902  {
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  903  	struct platform_device *ierb_pdev;
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  904  	struct device_node *ierb_node;
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  905  
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17 @906  	ierb_node = of_find_compatible_node(NULL, NULL,
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  907  					    "fsl,ls1028a-enetc-ierb");
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17 @908  	if (!ierb_node || !of_device_is_available(ierb_node))
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  909  		return -ENODEV;
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  910  
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  911  	ierb_pdev = of_find_device_by_node(ierb_node);
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17 @912  	of_node_put(ierb_node);
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  913  
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  914  	if (!ierb_pdev)
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  915  		return -EPROBE_DEFER;
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  916  
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  917  	return enetc_ierb_register_pf(ierb_pdev, pdev);
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  918  }
e7d48e5fbf30f8 Vladimir Oltean 2021-04-17  919  

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki

RE: [PATCH v5 net-next 11/13] net: enetc: optimize the allocation of tx_bdr

[Wei Fang]

> On Thu, Oct 24, 2024 at 02:53:26PM +0800, Wei Fang wrote:
> > From: Clark Wang <xiaoning.wang@nxp.com>
> >
> > There is a situation where num_tx_rings cannot be divided by bdr_int_num.
> > For example, num_tx_rings is 8 and bdr_int_num is 3. According to the
> > previous logic, this results in two tx_bdr corresponding memories not
> > being allocated, so when sending packets to tx ring 6 or 7, wild pointers
> > will be accessed. Of course, this issue doesn't exist on LS1028A, because
> > its num_tx_rings is 8, and bdr_int_num is either 1 or 2. However, there
> > is a risk for the upcoming i.MX95. Therefore, it is necessary to ensure
> > that each tx_bdr can be allocated to the corresponding memory.
> >
> > Signed-off-by: Clark Wang <xiaoning.wang@nxp.com>
> > Signed-off-by: Wei Fang <wei.fang@nxp.com>
> > Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
> > Reviewed-by: Frank Li <Frank.Li@nxp.com>
> > ---
> > v5: no changes
> > ---
> >  drivers/net/ethernet/freescale/enetc/enetc.c | 7 +++++--
> >  1 file changed, 5 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/net/ethernet/freescale/enetc/enetc.c
> b/drivers/net/ethernet/freescale/enetc/enetc.c
> > index bd725561b8a2..bccbeb1f355c 100644
> > --- a/drivers/net/ethernet/freescale/enetc/enetc.c
> > +++ b/drivers/net/ethernet/freescale/enetc/enetc.c
> > @@ -3049,10 +3049,10 @@ static void enetc_int_vector_destroy(struct
> enetc_ndev_priv *priv, int i)
> >  int enetc_alloc_msix(struct enetc_ndev_priv *priv)
> >  {
> >  	struct pci_dev *pdev = priv->si->pdev;
> > +	int v_tx_rings, v_remainder;
> >  	int num_stack_tx_queues;
> >  	int first_xdp_tx_ring;
> >  	int i, n, err, nvec;
> > -	int v_tx_rings;
> >
> >  	nvec = ENETC_BDR_INT_BASE_IDX + priv->bdr_int_num;
> >  	/* allocate MSIX for both messaging and Rx/Tx interrupts */
> > @@ -3066,9 +3066,12 @@ int enetc_alloc_msix(struct enetc_ndev_priv
> *priv)
> >
> >  	/* # of tx rings per int vector */
> >  	v_tx_rings = priv->num_tx_rings / priv->bdr_int_num;
> > +	v_remainder = priv->num_tx_rings % priv->bdr_int_num;
> >
> >  	for (i = 0; i < priv->bdr_int_num; i++) {
> > -		err = enetc_int_vector_init(priv, i, v_tx_rings);
> > +		int num_tx_rings = i < v_remainder ? v_tx_rings + 1 : v_tx_rings;
> 
> It took me a moment to understand the mechanism through which this
> works, even though I read the intention in the commit message.
> 
> Do you think this additional comment would help?

Yeah, it does help users understand quickly. I will add this comment.
> 
> 		/* Distribute the remaining TX rings to the first
> 		 * v_tx_rings interrupt vectors
> 		 */
> 
> > +
> > +		err = enetc_int_vector_init(priv, i, num_tx_rings);
> >  		if (err)
> >  			goto fail;
> >  	}
> > --
> > 2.34.1
> >

RE: [PATCH v5 net-next 04/13] net: enetc: add initial netc-blk-ctrl driver support

[Wei Fang]

> -----Original Message-----
> From: Vladimir Oltean <vladimir.oltean@nxp.com>
> Sent: 2024年10月25日 20:43
> To: Wei Fang <wei.fang@nxp.com>
> Cc: davem@davemloft.net; edumazet@google.com; kuba@kernel.org;
> pabeni@redhat.com; robh@kernel.org; krzk+dt@kernel.org;
> conor+dt@kernel.org; Claudiu Manoil <claudiu.manoil@nxp.com>; Clark Wang
> <xiaoning.wang@nxp.com>; Frank Li <frank.li@nxp.com>;
> christophe.leroy@csgroup.eu; linux@armlinux.org.uk; bhelgaas@google.com;
> horms@kernel.org; imx@lists.linux.dev; netdev@vger.kernel.org;
> devicetree@vger.kernel.org; linux-kernel@vger.kernel.org;
> linux-pci@vger.kernel.org; alexander.stein@ew.tq-group.com
> Subject: Re: [PATCH v5 net-next 04/13] net: enetc: add initial netc-blk-ctrl
> driver support
> 
> On Fri, Oct 25, 2024 at 04:44:50AM +0300, Wei Fang wrote:
> > > On Thu, Oct 24, 2024 at 02:53:19PM +0800, Wei Fang wrote:
> > > Can U-Boot deal with the IERB/PRB configuration?
> > >
> > > For LS1028A, the platform which initiated the IERB driver "trend", the
> situation
> > > was a bit more complicated, as we realized the reset-time defaults aren't
> what
> > > we need very late in the product life cycle, when customer boards already
> had
> > > bootloaders and we didn't want to complicate their process to have to
> redeploy
> > > in order to get access to such a basic feature as flow control. Though if we
> knew
> > > it from day one, we would have put the IERB fixups in U-Boot.
> >
> > The situation of i.MX95 is different from LS1028A, i.MX95 needs to support
> system
> > suspend/resume feature. If the i.MX95 enters suspend mode, the NETC may
> > power off (depends on user case), so IERB and PRB will be reset, in this case,
> we need
> > to reconfigure the IERB & PRB, including NETCMIX.
> >
> > > What is written in the IERB for MII/PCS protocols by default? I suppose
> there's
> > > some other mechanism to preinitialize it with good values?
> >
> > The MII/PCS protocols are set in NETCMIX not IERB, but the IERB will get
> these
> > info from NETCMIX, I mean the hardware, not the software. The default
> values
> > are all 0.
> 
> I am shocked that the NETCMIX/IERB blocks does not have a separate power
> domain from the ENETC, to avoid powering them off, which loses the settings.
> Please provide this explanation in the opening comments of this driver, it
> is its entire "raison d'être".

Hmm, it's a good idea, I can add this annotation at the beginning of the driver.
But this in not the entire "raison d'être", because we also hope to be able to
dynamically configure based on DTS, which is more flexible than under uboot.
For example, Timer binding for ENETC and switch if there are multiple Timer
instances, port selection if two ENETCs or one ENETC and one switch port share
the same physical port.

RE: [PATCH v5 net-next 06/13] net: enetc: build enetc_pf_common.c as a separate module

[Wei Fang]

> On Fri, Oct 25, 2024 at 06:00:42AM +0300, Wei Fang wrote:
> > > Don't artificially create errors when there are really no errors to handle.
> > > Both enetc_pf_ops and enetc4_pf_ops provide .set_si_primary_mac(), so it
> > > is unnecessary to handle the case where it isn't present. Those functions
> > > return void, and void can be propagated to enetc_set_si_hw_addr() as well.
> >
> > I thought checking the pointer is safer, so you mean that pointers that are
> > definitely present in the current driver do not need to be checked?
> 
> Yes, there is no point to check for a condition which is impossible
> to trigger in the current code. The callee and the caller are tightly
> coupled (in the same driver), it's not a rigid API, so if the function
> pointers should be made optional for some future hardware IP, the error
> handling will be added when necessary. Ideally any change passes through
> review, and any inconsistency should be spotted when added.
> 
> > > A bit inconsistent that pf->ops->set_si_primary_mac() goes through a
> > > wrapper function but this doesn't.
> > >
> >
> > If we really do not need to check these callback pointers, then I think I can
> > remove the wrapper.
> 
> Fine without wrapping throughout, my comment was first and foremost
> about consistency.
> 
> > > This one looks extremely weird in the existing code, but I suppose I'm
> > > too late to the party to request you to clean up any of the PSFP code,
> > > so I'll make a note to do it myself after your work. I haven't spotted
> > > any actual bug, just weird coding patterns.
> > >
> > > No change request here. I see the netc4_pf doesn't implement
> enable_psfp(),
> > > so making it optional here is fine.
> >
> > Yes, PSFP is not supported in this patch set, I will remove it in future.
> 
> If by "I will remove it in future" you mean "once NETC4 gains PSFP
> support, I will make enable_psfp() non-optional", then ok, great.
> 
> > Currently, we have not add the PCS support, so the 10G ENETC is not
> supported
> > yet. And we also disable the 10G ENETC in DTS. Only the 1G ENETCs (without
> PCS)
> > are supported for i.MX95.
> 
> Also think about the case where the current version of the kernel
> will boot on a newer version of the device tree, which does not have
> 'status = "disabled";' for those ports. It should do something reasonable.
> In any case, "they're now disabled in the device tree" is not an argument.
> 

For this case where the kernel is lower but the device tree is newer, I think
there is no problem. It just fails in probe() and does not affect other functions.
The old kernel does not support PCS, it is reasonable to return a '- EOPNOTSUPP'
error.

> My anecdotal and vague understanding of the Arm SystemReady (IR I think)
> requirements is that the device tree is provided by the platform,
> separately from the kernel/rootfs. It relies on the device tree ABI
> being stable, backwards-compatible and forwards-compatible.
> 
> > > A message maybe, stating what's the deal? Just that users figure out
> > > quickly that it's an expected behavior, and not spend hours debugging
> > > until they find out it's not their fault?
> >
> > I will explain in the commit message that i.MX95 10G ENETC is not currently
> > supported.
> 
> By "a message, maybe" I actually meant dev_err("Message here\n"); rather
> than silent/imprecise failure.

Okay, I'll add an explicit error message.

RE: [PATCH v5 net-next 02/13] dt-bindings: net: add i.MX95 ENETC support

[Wei Fang]

> 
> On Thu, Oct 24, 2024 at 02:53:17PM +0800, Wei Fang wrote:
> > The ENETC of i.MX95 has been upgraded to revision 4.1, and the vendor
> > ID and device ID have also changed, so add the new compatible strings
> > for i.MX95 ENETC. In addition, i.MX95 supports configuration of RGMII
> > or RMII reference clock.
> >
> > Signed-off-by: Wei Fang <wei.fang@nxp.com>
> > ---
> > v2: Remove "nxp,imx95-enetc" compatible string.
> > v3:
> > 1. Add restriction to "clcoks" and "clock-names" properties and rename
> > the clock, also remove the items from these two properties.
> > 2. Remove unnecessary items for "pci1131,e101" compatible string.
> > v4: Move clocks and clock-names to top level.
> > v5: Add items to clocks and clock-names
> > ---
> >  .../devicetree/bindings/net/fsl,enetc.yaml    | 34 +++++++++++++++++--
> >  1 file changed, 31 insertions(+), 3 deletions(-)
> >
> > diff --git a/Documentation/devicetree/bindings/net/fsl,enetc.yaml
> > b/Documentation/devicetree/bindings/net/fsl,enetc.yaml
> > index e152c93998fe..72d2d5d285cd 100644
> > --- a/Documentation/devicetree/bindings/net/fsl,enetc.yaml
> > +++ b/Documentation/devicetree/bindings/net/fsl,enetc.yaml
> > @@ -20,14 +20,23 @@ maintainers:
> >
> >  properties:
> >    compatible:
> > -    items:
> > +    oneOf:
> > +      - items:
> > +          - enum:
> > +              - pci1957,e100
> > +          - const: fsl,enetc
> >        - enum:
> > -          - pci1957,e100
> > -      - const: fsl,enetc
> > +          - pci1131,e101
> >
> >    reg:
> >      maxItems: 1
> >
> > +  clocks:
> > +    maxItems: 1
> > +
> > +  clock-names:
> > +    maxItems: 1
> > +
> >    mdio:
> >      $ref: mdio.yaml
> >      unevaluatedProperties: false
> > @@ -40,6 +49,25 @@ required:
> >  allOf:
> >    - $ref: /schemas/pci/pci-device.yaml
> >    - $ref: ethernet-controller.yaml
> > +  - if:
> > +      properties:
> > +        compatible:
> > +          contains:
> > +            enum:
> > +              - pci1131,e101
> > +    then:
> > +      properties:
> > +        clocks:
> > +          items:
> > +            - description: MAC transmit/receiver reference clock
> 
> This goes in the top-level or can just be dropped.
> 
> > +
> > +        clock-names:
> > +          items:
> > +            - const: ref
> 
> This goes in the top-level.
> 
> 
> > +    else:
> 
> Then negate the 'if' schema (not: contains: ...) and you just need the part
> below:

Okay, I'll improve it, many thanks.

> 
> > +      properties:
> > +        clocks: false
> > +        clock-names: false
> >
> >  unevaluatedProperties: false
> >
> > --
> > 2.34.1
> >