Skip to contentSkip to conversation

Brcm ASP 2.0 Ethernet controller

PATCH In review Justin Chen 6 participants 16 messages Apr 26, 2023 – May 3, 2023

Subsystems (8)
lore
Share & tools
lore
#0 lore raw

v2 cover Brcm ASP 2.0 Ethernet controller

v2
- Updates to yaml dt documentation
- Replace a couple functions with helper functions
- Minor formatting fixes
- Fix a few WoL issues

Add support for the Broadcom ASP 2.0 Ethernet controller which is first introduced with 72165.

Add support for 74165 10/100 integrated Ethernet PHY which also uses the ASP 2.0 Ethernet controller.

Florian Fainelli (2):
dt-bindings: net: Brcm ASP 2.0 Ethernet controller
net: phy: bcm7xxx: Add EPHY entry for 74165

Justin Chen (4):
dt-bindings: net: brcm,unimac-mdio: Add asp-v2.0
net: bcmasp: Add support for ASP2.0 Ethernet controller
net: phy: mdio-bcm-unimac: Add asp v2.0 support
MAINTAINERS: ASP 2.0 Ethernet driver maintainers

 .../devicetree/bindings/net/brcm,asp-v2.0.yaml     |  145 ++
 .../devicetree/bindings/net/brcm,unimac-mdio.yaml  |    2 +
 MAINTAINERS                                        |    9 +
 drivers/net/ethernet/broadcom/Kconfig              |   11 +
 drivers/net/ethernet/broadcom/Makefile             |    1 +
 drivers/net/ethernet/broadcom/asp2/Makefile        |    2 +
 drivers/net/ethernet/broadcom/asp2/bcmasp.c        | 1476 ++++++++++++++++++++
 drivers/net/ethernet/broadcom/asp2/bcmasp.h        |  636 +++++++++
 .../net/ethernet/broadcom/asp2/bcmasp_ethtool.c    |  585 ++++++++
 drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c   | 1435 +++++++++++++++++++
 .../net/ethernet/broadcom/asp2/bcmasp_intf_defs.h  |  238 ++++
 drivers/net/mdio/mdio-bcm-unimac.c                 |    2 +
 drivers/net/phy/bcm7xxx.c                          |    1 +
 include/linux/brcmphy.h                            |    1 +
 14 files changed, 4544 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
 create mode 100644 drivers/net/ethernet/broadcom/asp2/Makefile
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp.c
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp.h
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp_ethtool.c
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp_intf_defs.h
Signature · 1 line
2.7.4
#1 lore raw

PATCH v2 1/6 dt-bindings: net: brcm,unimac-mdio: Add asp-v2.0

The ASP 2.0 Ethernet controller uses a brcm unimac.

---
 Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml +2 -0
index 0be426ee1e44..6684810fcbf0 100644
@@ -22,6 +22,8 @@ properties:
       - brcm,genet-mdio-v3
       - brcm,genet-mdio-v4
       - brcm,genet-mdio-v5
+      - brcm,asp-v2.0-mdio
+      - brcm,asp-v2.1-mdio
       - brcm,unimac-mdio
 
   reg:
Signature · 1 line
2.7.4
maintainer
#2 lore raw

Re: [PATCH v2 net-next 1/6] dt-bindings: net: brcm,unimac-mdio: Add asp-v2.0

Quoted 13 lines · Justin Chen

The ASP 2.0 Ethernet controller uses a brcm unimac.

---
 Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml +0 -0
index 0be426ee1e44..6684810fcbf0 100644

Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml properties: · line 26

Quoted 5 lines
Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml +2 -0
@@ -22,6 +22,8 @@ properties:
       - brcm,genet-mdio-v3
       - brcm,genet-mdio-v4
       - brcm,genet-mdio-v5
+      - brcm,asp-v2.0-mdio
+      - brcm,asp-v2.1-mdio

How many SoCs does each of these correspond to? SoC specific compatibles are preferred to version numbers (because few vendors are disciplined at versioning and also not changing versions with every Soc).

Quoted 6 lines
Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml +0 -0
       - brcm,unimac-mdio
 
   reg:
Signature · 1 line
2.7.4
#3 lore raw

Re: [PATCH v2 net-next 1/6] dt-bindings: net: brcm,unimac-mdio: Add asp-v2.0

Quoted 9 lines · Rob Herring
Quoted 8 lines · Justin Chen

The ASP 2.0 Ethernet controller uses a brcm unimac.

---
  Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml | 2 ++
  1 file changed, 2 insertions(+)
Quoted 14 lines

Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml properties: · line 26

Quoted 10 lines
Documentation/devicetree/bindings/net/brcm,unimac-mdio.yaml +2 -0
index 0be426ee1e44..6684810fcbf0 100644
@@ -22,6 +22,8 @@ properties:
        - brcm,genet-mdio-v3
        - brcm,genet-mdio-v4
        - brcm,genet-mdio-v5
+      - brcm,asp-v2.0-mdio
+      - brcm,asp-v2.1-mdio

How many SoCs does each of these correspond to? SoC specific compatibles are preferred to version numbers (because few vendors are disciplined at versioning and also not changing versions with every Soc).

So far there is a 1:1 mapping between the number of versions and the number of SoCs, and the older SoC uses v2.0, while the newer one uses v2.1.

Signature · 1 line
Florian
#4 lore raw

PATCH v2 2/6 dt-bindings: net: Brcm ASP 2.0 Ethernet controller

In reply to Justin Chen

From: Florian Fainelli <[email protected]>

Add a binding document for the Broadcom ASP 2.0 Ethernet controller.

---
 .../devicetree/bindings/net/brcm,asp-v2.0.yaml     | 145 +++++++++++++++++++++
 1 file changed, 145 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +145 -0
new file mode 100644
index 000000000000..818d91692e6e
@@ -0,0 +1,145 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/brcm,asp-v2.0.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Broadcom ASP 2.0 Ethernet controller
+
+maintainers:
+  - Justin Chen <[email protected]>
+  - Florian Fainelli <[email protected]>
+
+description: Broadcom Ethernet controller first introduced with 72165
+
+properties:
+  '#address-cells':
+    const: 1
+  '#size-cells':
+    const: 1
+
+  compatible:
+    enum:
+      - brcm,asp-v2.0
+      - brcm,bcm72165-asp-v2.0
+      - brcm,asp-v2.1
+      - brcm,bcm74165-asp-v2.1
+
+  reg:
+    maxItems: 1
+
+  ranges: true
+
+  interrupts:
+    minItems: 1
+    items:
+      - description: RX/TX interrupt
+      - description: Port 0 Wake-on-LAN
+      - description: Port 1 Wake-on-LAN
+
+  clocks:
+    maxItems: 1
+
+  ethernet-ports:
+    type: object
+    properties:
+      '#address-cells':
+        const: 1
+      '#size-cells':
+        const: 0
+
+    patternProperties:
+      "^port@[0-9]+$":
+        type: object
+
+        $ref: ethernet-controller.yaml#
+
+        properties:
+          reg:
+            maxItems: 1
+            description: Port number
+
+          channel:
+            maxItems: 1
+            description: ASP channel number
+
+        required:
+          - reg
+          - channel
+
+    additionalProperties: false
+
+patternProperties:
+  "^mdio@[0-9a-f]+$":
+    type: object
+    $ref: "brcm,unimac-mdio.yaml"
+
+    description:
+      ASP internal UniMAC MDIO bus
+
+required:
+  - compatible
+  - reg
+  - interrupts
+  - clocks
+  - ranges
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/interrupt-controller/irq.h>
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+
+    ethernet@9c00000 {
+        compatible = "brcm,asp-v2.0";
+        reg = <0x9c00000 0x1fff14>;
+        interrupts = <GIC_SPI 51 IRQ_TYPE_LEVEL_HIGH>;
+        ranges;
+        clocks = <&scmi 14>;
+        #address-cells = <1>;
+        #size-cells = <1>;
+
+        mdio@c614 {
+            compatible = "brcm,asp-v2.0-mdio";
+            reg = <0xc614 0x8>;
+            reg-names = "mdio";
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            phy0: ethernet-phy@1 {
+                reg = <1>;
+            };
+       };
+
+        mdio@ce14 {
+            compatible = "brcm,asp-v2.0-mdio";
+            reg = <0xce14 0x8>;
+            reg-names = "mdio";
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            phy1: ethernet-phy@1 {
+                reg = <1>;
+            };
+        };
+
+        ethernet-ports {
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            port@0 {
+                reg = <0>;
+                channel = <8>;
+                phy-mode = "rgmii";
+                phy-handle = <&phy0>;
+            };
+
+            port@1 {
+                reg = <1>;
+                channel = <9>;
+                phy-mode = "rgmii";
+                phy-handle = <&phy1>;
+            };
+        };
+    };
Signature · 1 line
2.7.4
maintainer
#5 lore raw

Re: [PATCH v2 net-next 2/6] dt-bindings: net: Brcm ASP 2.0 Ethernet controller

Quoted 34 lines · Justin Chen

From: Florian Fainelli <[email protected]>

Add a binding document for the Broadcom ASP 2.0 Ethernet controller.

---
 .../devicetree/bindings/net/brcm,asp-v2.0.yaml     | 145 +++++++++++++++++++++
 1 file changed, 145 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +16 -0
new file mode 100644
index 000000000000..818d91692e6e
@@ -0,0 +1,145 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/brcm,asp-v2.0.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Broadcom ASP 2.0 Ethernet controller
+
+maintainers:
+  - Justin Chen <[email protected]>
+  - Florian Fainelli <[email protected]>
+
+description: Broadcom Ethernet controller first introduced with 72165
+
+properties:
+  '#address-cells':

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 26

Quoted 10 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +10 -0
+    const: 1
+  '#size-cells':
+    const: 1
+
+  compatible:
+    enum:
+      - brcm,asp-v2.0
+      - brcm,bcm72165-asp-v2.0
+      - brcm,asp-v2.1
+      - brcm,bcm74165-asp-v2.1

You have 1 SoC per version, so what's the point of versions? If you have more coming, then fine, but I'd expect it to be something like this:

compatible = "brcm,bcm74165-asp-v2.1", "brcm,asp-v2.1";

Also, the version in the SoC specific compatible is redundant. Just "brcm,bcm74165-asp" is enough.

v2.1 is not compatible with v2.0? What that means is would a client/OS that only understands what v2.0 is work with v2.1 h/w? If so, you should have fallback compatible.

Quoted 7 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +7 -0
+
+  reg:
+    maxItems: 1
+
+  ranges: true
+
+  interrupts:

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 43

Quoted 10 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +10 -0
+    minItems: 1
+    items:
+      - description: RX/TX interrupt
+      - description: Port 0 Wake-on-LAN
+      - description: Port 1 Wake-on-LAN
+
+  clocks:
+    maxItems: 1
+
+  ethernet-ports:

The ethernet-switch.yaml schema doesn't work for you?

Quoted 11 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +11 -0
+    type: object
+    properties:
+      '#address-cells':
+        const: 1
+      '#size-cells':
+        const: 0
+
+    patternProperties:
+      "^port@[0-9]+$":
+        type: object
+

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 64

Quoted 10 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +10 -0
+        $ref: ethernet-controller.yaml#
+
+        properties:
+          reg:
+            maxItems: 1
+            description: Port number
+
+          channel:
+            maxItems: 1
+            description: ASP channel number

Not a standard property, so it needs a type and vendor prefix. However, what's the difference between channel and port? Can the port numbers correspond to the channels?

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 75

Quoted 11 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +11 -0
+
+        required:
+          - reg
+          - channel
+
+    additionalProperties: false
+
+patternProperties:
+  "^mdio@[0-9a-f]+$":
+    type: object
+    $ref: "brcm,unimac-mdio.yaml"

Drop quotes.

Quoted 20 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +20 -0
+
+    description:
+      ASP internal UniMAC MDIO bus
+
+required:
+  - compatible
+  - reg
+  - interrupts
+  - clocks
+  - ranges
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/interrupt-controller/irq.h>
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+
+    ethernet@9c00000 {
+        compatible = "brcm,asp-v2.0";

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 105

Quoted 10 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +10 -0
+        reg = <0x9c00000 0x1fff14>;
+        interrupts = <GIC_SPI 51 IRQ_TYPE_LEVEL_HIGH>;
+        ranges;
+        clocks = <&scmi 14>;
+        #address-cells = <1>;
+        #size-cells = <1>;
+
+        mdio@c614 {
+            compatible = "brcm,asp-v2.0-mdio";
+            reg = <0xc614 0x8>;

You have 1:1 ranges, is that really what you want? That means 0xc614 is an absolute address.

Quoted 43 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +40 -0
+            reg-names = "mdio";
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            phy0: ethernet-phy@1 {
+                reg = <1>;
+            };
+       };
+
+        mdio@ce14 {
+            compatible = "brcm,asp-v2.0-mdio";
+            reg = <0xce14 0x8>;
+            reg-names = "mdio";
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            phy1: ethernet-phy@1 {
+                reg = <1>;
+            };
+        };
+
+        ethernet-ports {
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            port@0 {
+                reg = <0>;
+                channel = <8>;
+                phy-mode = "rgmii";
+                phy-handle = <&phy0>;
+            };
+
+            port@1 {
+                reg = <1>;
+                channel = <9>;
+                phy-mode = "rgmii";
+                phy-handle = <&phy1>;
+            };
+        };
+    };
Signature · 1 line
2.7.4
#6 lore raw

Re: [PATCH v2 net-next 2/6] dt-bindings: net: Brcm ASP 2.0 Ethernet controller

Quoted 50 lines · Rob Herring
Quoted 34 lines · Justin Chen

From: Florian Fainelli <[email protected]>

Add a binding document for the Broadcom ASP 2.0 Ethernet controller.

---
 .../devicetree/bindings/net/brcm,asp-v2.0.yaml     | 145 +++++++++++++++++++++
 1 file changed, 145 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +16 -0
new file mode 100644
index 000000000000..818d91692e6e
@@ -0,0 +1,145 @@
+# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
+%YAML 1.2
+---
+$id: http://devicetree.org/schemas/net/brcm,asp-v2.0.yaml#
+$schema: http://devicetree.org/meta-schemas/core.yaml#
+
+title: Broadcom ASP 2.0 Ethernet controller
+
+maintainers:
+  - Justin Chen <[email protected]>
+  - Florian Fainelli <[email protected]>
+
+description: Broadcom Ethernet controller first introduced with 72165
+
+properties:
+  '#address-cells':

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 26

Quoted 10 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +10 -0
+    const: 1
+  '#size-cells':
+    const: 1
+
+  compatible:
+    enum:
+      - brcm,asp-v2.0
+      - brcm,bcm72165-asp-v2.0
+      - brcm,asp-v2.1
+      - brcm,bcm74165-asp-v2.1

You have 1 SoC per version, so what's the point of versions? If you have more coming, then fine, but I'd expect it to be something like this:

Quoted 9 lines

compatible = "brcm,bcm74165-asp-v2.1", "brcm,asp-v2.1";

Also, the version in the SoC specific compatible is redundant. Just "brcm,bcm74165-asp" is enough.

v2.1 is not compatible with v2.0? What that means is would a client/OS that only understands what v2.0 is work with v2.1 h/w? If so, you should have fallback compatible.

v2.1 is not compatible with v2.0 unfortunately. So no, a client/OS that only understands v2.0 will not work with v2.1 h/w.

Quoted 11 lines
Quoted 11 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +11 -0
+
+  reg:
+    maxItems: 1
+
+  ranges: true
+
+  interrupts:
+    minItems: 1
+    items:
+      - description: RX/TX interrupt
+      - description: Port 0 Wake-on-LAN
Quoted 9 lines

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 43

Quoted 6 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +6 -0
+      - description: Port 1 Wake-on-LAN
+
+  clocks:
+    maxItems: 1
+
+  ethernet-ports:

The ethernet-switch.yaml schema doesn't work for you?

Technically it is not a switch. But it might work... If we use port to reference the unimac and reg to reference the ethernet channel. I rather not though, just cause it is not a switch, so calling it an ethernet-switch is confusing.

Quoted 17 lines
Quoted 17 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +17 -0
+    type: object
+    properties:
+      '#address-cells':
+        const: 1
+      '#size-cells':
+        const: 0
+
+    patternProperties:
+      "^port@[0-9]+$":
+        type: object
+
+        $ref: ethernet-controller.yaml#
+
+        properties:
+          reg:
+            maxItems: 1
+            description: Port number
Quoted 9 lines

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 64

Quoted 4 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +4 -0
+
+          channel:
+            maxItems: 1
+            description: ASP channel number

Not a standard property, so it needs a type and vendor prefix. However, what's the difference between channel and port? Can the port numbers correspond to the channels?

Port refers to the unimac. In our case we currently have a maximum of
2. Channel refers to the ethernet hardware channel proper, in which we have many. So yes, you can have a port correlate to any channel.

Quoted 39 lines

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 75

Quoted 11 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +11 -0
+
+        required:
+          - reg
+          - channel
+
+    additionalProperties: false
+
+patternProperties:
+  "^mdio@[0-9a-f]+$":
+    type: object
+    $ref: "brcm,unimac-mdio.yaml"

Drop quotes.

Quoted 25 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +25 -0
+
+    description:
+      ASP internal UniMAC MDIO bus
+
+required:
+  - compatible
+  - reg
+  - interrupts
+  - clocks
+  - ranges
+
+additionalProperties: false
+
+examples:
+  - |
+    #include <dt-bindings/interrupt-controller/irq.h>
+    #include <dt-bindings/interrupt-controller/arm-gic.h>
+
+    ethernet@9c00000 {
+        compatible = "brcm,asp-v2.0";
+        reg = <0x9c00000 0x1fff14>;
+        interrupts = <GIC_SPI 51 IRQ_TYPE_LEVEL_HIGH>;
+        ranges;
+        clocks = <&scmi 14>;
+        #address-cells = <1>;
Quoted 8 lines

Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml line 105

Quoted 5 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +5 -0
+        #size-cells = <1>;
+
+        mdio@c614 {
+            compatible = "brcm,asp-v2.0-mdio";
+            reg = <0xc614 0x8>;

You have 1:1 ranges, is that really what you want? That means 0xc614 is an absolute address.

Ack, will fix.

Thanks for the review,
Justin

Quoted 44 lines
Quoted 44 lines
Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml new file +40 -1
+            reg-names = "mdio";
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            phy0: ethernet-phy@1 {
+                reg = <1>;
+            };
+       };
+
+        mdio@ce14 {
+            compatible = "brcm,asp-v2.0-mdio";
+            reg = <0xce14 0x8>;
+            reg-names = "mdio";
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            phy1: ethernet-phy@1 {
+                reg = <1>;
+            };
+        };
+
+        ethernet-ports {
+            #address-cells = <1>;
+            #size-cells = <0>;
+
+            port@0 {
+                reg = <0>;
+                channel = <8>;
+                phy-mode = "rgmii";
+                phy-handle = <&phy0>;
+            };
+
+            port@1 {
+                reg = <1>;
+                channel = <9>;
+                phy-mode = "rgmii";
+                phy-handle = <&phy1>;
+            };
+        };
+    };
--

2.7.4

#7 lore raw

PATCH v2 3/6 net: bcmasp: Add support for ASP2.0 Ethernet controller

In reply to Justin Chen

Add support for the Broadcom ASP 2.0 Ethernet controller which is first introduced with 72165. This controller features two distinct Ethernet ports that can be independently operated.

This patch supports:

- Wake-on-LAN using magic packets
- basic ethtool operations (link, counters, message level)
- MAC destination address filtering (promiscuous, ALL_MULTI, etc.)

---
 drivers/net/ethernet/broadcom/Kconfig              |   11 +
 drivers/net/ethernet/broadcom/Makefile             |    1 +
 drivers/net/ethernet/broadcom/asp2/Makefile        |    2 +
 drivers/net/ethernet/broadcom/asp2/bcmasp.c        | 1476 ++++++++++++++++++++
 drivers/net/ethernet/broadcom/asp2/bcmasp.h        |  636 +++++++++
 .../net/ethernet/broadcom/asp2/bcmasp_ethtool.c    |  585 ++++++++
 drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c   | 1435 +++++++++++++++++++
 .../net/ethernet/broadcom/asp2/bcmasp_intf_defs.h  |  238 ++++
 8 files changed, 4384 insertions(+)
 create mode 100644 drivers/net/ethernet/broadcom/asp2/Makefile
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp.c
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp.h
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp_ethtool.c
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c
 create mode 100644 drivers/net/ethernet/broadcom/asp2/bcmasp_intf_defs.h
drivers/net/ethernet/broadcom/Kconfig +11 -0
index 948586bf1b5b..d4166141145d 100644
@@ -255,4 +255,15 @@ config BNXT_HWMON
 	  Say Y if you want to expose the thermal sensor data on NetXtreme-C/E
 	  devices, via the hwmon sysfs interface.
 
+config BCMASP
+	tristate "Broadcom ASP 2.0 Ethernet support"
+	default ARCH_BRCMSTB
+	depends on OF
+	select MII
+	select PHYLIB
+	select MDIO_BCM_UNIMAC
+	help
+	  This configuration enables the Broadcom ASP 2.0 Ethernet controller
+	  driver which is present in Broadcom STB SoCs such as 72165.
+
 endif # NET_VENDOR_BROADCOM
drivers/net/ethernet/broadcom/Makefile +1 -0
index 0ddfb5b5d53c..bac5cb6ad0cd 100644
@@ -17,3 +17,4 @@ obj-$(CONFIG_BGMAC_BCMA) += bgmac-bcma.o bgmac-bcma-mdio.o
 obj-$(CONFIG_BGMAC_PLATFORM) += bgmac-platform.o
 obj-$(CONFIG_SYSTEMPORT) += bcmsysport.o
 obj-$(CONFIG_BNXT) += bnxt/
+obj-$(CONFIG_BCMASP) += asp2/
drivers/net/ethernet/broadcom/asp2/Makefile new file +2 -0
new file mode 100644
index 000000000000..e07550315f83
@@ -0,0 +1,2 @@
+obj-$(CONFIG_BCMASP) += bcm-asp.o
+bcm-asp-objs := bcmasp.o bcmasp_intf.o bcmasp_ethtool.o
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +1476 -0

1,476 lines not on this page · raw message

drivers/net/ethernet/broadcom/asp2/bcmasp.h new file +636 -0

636 lines not on this page · raw message

drivers/net/ethernet/broadcom/asp2/bcmasp_ethtool.c new file +585 -0

585 lines not on this page · raw message

drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c new file +1435 -0

282 lines not on this page · raw message

drivers/net/ethernet/broadcom/asp2/bcmasp_intf_defs.h new file +238 -0

238 lines not on this page · raw message

The rest of this diff is not shown. View the raw message.

Signature · 1 line
2.7.4

Part of the diff is not shown here. View the raw message.

#8 lore raw

Re: [PATCH v2 net-next 3/6] net: bcmasp: Add support for ASP2.0 Ethernet controller

Quoted 12 lines · Justin Chen

Add support for the Broadcom ASP 2.0 Ethernet controller which is first introduced with 72165. This controller features two distinct Ethernet ports that can be independently operated.

This patch supports:

- Wake-on-LAN using magic packets
- basic ethtool operations (link, counters, message level)
- MAC destination address filtering (promiscuous, ALL_MULTI, etc.)

...

Quoted 1 line
drivers/net/ethernet/broadcom/asp2/bcmasp.c +0 -0

...

Quoted 22 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +22 -0
+static int bcmasp_netfilt_get_reg_offset(struct bcmasp_priv *priv,
+					 struct bcmasp_net_filter *nfilt,
+					 enum asp_netfilt_reg_type reg_type,
+					 u32 offset)
+{
+	u32 block_index, filter_sel;
+
+	if (offset < 32) {
+		block_index = ASP_RX_FILTER_NET_L2;
+		filter_sel = nfilt->hw_index;
+	} else if (offset < 64) {
+		block_index = ASP_RX_FILTER_NET_L2;
+		filter_sel = nfilt->hw_index + 1;
+	} else if (offset < 96) {
+		block_index = ASP_RX_FILTER_NET_L3_0;
+		filter_sel = nfilt->hw_index;
+	} else if (offset < 128) {
+		block_index = ASP_RX_FILTER_NET_L3_0;
+		filter_sel = nfilt->hw_index + 1;
+	} else if (offset < 160) {
+		block_index = ASP_RX_FILTER_NET_L3_1;
+		filter_sel = nfilt->hw_index;

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 204

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+	} else if (offset < 192) {
+		block_index = ASP_RX_FILTER_NET_L3_1;
+		filter_sel = nfilt->hw_index + 1;
+	} else if (offset < 224) {
+		block_index = ASP_RX_FILTER_NET_L4;
+		filter_sel = nfilt->hw_index;
+	} else if (offset < 256) {
+		block_index = ASP_RX_FILTER_NET_L4;
+		filter_sel = nfilt->hw_index + 1;
+	}

block_index and filter_sel are uninitialised if offset doesn't match any of the conditions above. Can that happen?

Quoted 12 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +12 -0
+
+	switch (reg_type) {
+	case ASP_NETFILT_MATCH:
+		return ASP_RX_FILTER_NET_PAT(filter_sel, block_index,
+					     (offset % 32));
+	case ASP_NETFILT_MASK:
+		return ASP_RX_FILTER_NET_MASK(filter_sel, block_index,
+					      (offset % 32));
+	default:
+		return -EINVAL;
+	}
+}

...

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 323

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+static void bcmasp_netfilt_tcpip4_wr(struct bcmasp_priv *priv,
+				     struct bcmasp_net_filter *nfilt,
+				     struct ethtool_tcpip4_spec *match,
+				     struct ethtool_tcpip4_spec *mask,
+				     u32 offset)
+{
+	__be16 val_16, mask_16;
+
+	val_16 = htons(ETH_P_IP);
+	mask_16 = 0xFFFF;

mask_17 is __be16, but 0xFFFF is host byte order.

Please make sure there are no new warnings when building with W=1 C=1.

...

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 649

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+/* If no network filter found, return open filter.
+ * If no more open filters return NULL
+ */
+struct bcmasp_net_filter *bcmasp_netfilt_get_init(struct bcmasp_intf *intf,
+						  int loc, bool wake_filter,
+						  bool init)
+{
+	struct bcmasp_priv *priv = intf->parent;
+	struct bcmasp_net_filter *nfilter = NULL;
+	int i, open_index = -1;

Please use reverse xmas tree - longest line to shortest - for local variable declarations in networking code.

You can check for this using https://github.com/ecree-solarflare/xmastree

...

Quoted 7 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +7 -0
+static int bcmasp_combine_set_filter(struct bcmasp_intf *intf,
+				     unsigned char *addr, unsigned char *mask,
+				     int i)
+{
+	u64 addr1, addr2, mask1, mask2, mask3;
+	struct bcmasp_priv *priv = intf->parent;
+

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 865

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+	/* Switch to u64 to help with the calculations */
+	addr1 = ether_addr_to_u64(priv->mda_filters[i].addr);
+	mask1 = ether_addr_to_u64(priv->mda_filters[i].mask);
+	addr2 = ether_addr_to_u64(addr);
+	mask2 = ether_addr_to_u64(mask);
+
+	/* Check if one filter resides within the other */
+	mask3 = mask1 & mask2;
+	if (mask3 == mask1 && ((addr1 & mask1) == (addr2 & mask1))) {
+		/* Filter 2 resides within fitler 1, so everthing is good */

nit: s/fitler/filter/

Please consider running ./scripts/checkpatch.pl --codespell

...

Quoted 29 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +29 -0
+static void bcmasp_update_mib_counters(struct bcmasp_intf *priv)
+{
+	int i, j = 0;
+
+	for (i = 0; i < BCMASP_STATS_LEN; i++) {
+		const struct bcmasp_stats *s;
+		u16 offset = 0;
+		u32 val = 0;
+		char *p;
+
+		s = &bcmasp_gstrings_stats[i];
+		switch (s->type) {
+		case BCMASP_STAT_NETDEV:
+		case BCMASP_STAT_SOFT:
+			continue;
+		case BCMASP_STAT_RUNT:
+			offset += BCMASP_STAT_OFFSET;
+			fallthrough;
+		case BCMASP_STAT_MIB_TX:
+			offset += BCMASP_STAT_OFFSET;
+			fallthrough;
+		case BCMASP_STAT_MIB_RX:
+			val = umac_rl(priv, UMC_MIB_START + j + offset);
+			offset = 0;	/* Reset Offset */
+			break;
+		case BCMASP_STAT_RX_EDPKT:
+			val = rx_edpkt_core_rl(priv->parent, s->reg_offset);
+			break;
+		case BCMASP_STAT_RX_CTRL:

drivers/net/ethernet/broadcom/asp2/bcmasp.c

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+			offset = bcmasp_stat_fixup_offset(priv, s);
+			if (offset != ASP_RX_CTRL_FB_FILT_OUT_FRAME_COUNT)
+				offset += sizeof(u32) * priv->port;
+			val = rx_ctrl_core_rl(priv->parent, offset);
+			break;
+		}
+
+		j += s->stat_sizeof;
+		p = (char *)priv + s->stat_offset;
+		*(u32 *)p = val;

Is p always 32bit aligned?

Quoted 22 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +22 -0
+	}
+}
+
+static void bcmasp_get_ethtool_stats(struct net_device *dev,
+				     struct ethtool_stats *stats,
+				     u64 *data)
+{
+	struct bcmasp_intf *priv = netdev_priv(dev);
+	int i, j = 0;
+
+	if (netif_running(dev))
+		bcmasp_update_mib_counters(priv);
+
+	dev->netdev_ops->ndo_get_stats(dev);
+
+	for (i = 0; i < BCMASP_STATS_LEN; i++) {
+		const struct bcmasp_stats *s;
+		char *p;
+
+		s = &bcmasp_gstrings_stats[i];
+		if (!bcmasp_stat_available(priv, s->type))
+			continue;

drivers/net/ethernet/broadcom/asp2/bcmasp.c

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+		if (s->type == BCMASP_STAT_NETDEV)
+			p = (char *)&dev->stats;
+		else
+			p = (char *)priv;
+		p += s->stat_offset;
+		if (sizeof(unsigned long) != sizeof(u32) &&
+		    s->stat_sizeof == sizeof(unsigned long))
+			data[j] = *(unsigned long *)p;
+		else
+			data[j] = *(u32 *)p;

Maybe memcpy would make this a little easier to read.

Quoted 3 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +3 -0
+		j++;
+	}
+}

...

Quoted 1 line
drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c +0 -0

...

Quoted 7 lines
drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c +7 -0
+static int bcmasp_init_rx(struct bcmasp_intf *intf)
+{
+	struct device *kdev = &intf->parent->pdev->dev;
+	struct net_device *ndev = intf->ndev;
+	void *p;
+	dma_addr_t dma;
+	struct page *buffer_pg;

drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c line 700

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c +10 -0
+	u32 reg;
+	int ret;
+
+	intf->rx_buf_order = get_order(RING_BUFFER_SIZE);
+	buffer_pg = alloc_pages(GFP_KERNEL, intf->rx_buf_order);
+
+	dma = dma_map_page(kdev, buffer_pg, 0, RING_BUFFER_SIZE,
+			   DMA_FROM_DEVICE);
+	if (dma_mapping_error(kdev, dma)) {
+		netdev_err(ndev, "Cannot allocate RX buffer\n");

I think the core will log an error on allocation failure, so the message above is not needed.

Quoted 3 lines
drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c +3 -0
+		__free_pages(buffer_pg, intf->rx_buf_order);
+		return -ENOMEM;
+	}

...

#9 lore raw

Re: [PATCH v2 net-next 3/6] net: bcmasp: Add support for ASP2.0 Ethernet controller

Quoted 48 lines · Simon Horman
Quoted 12 lines · Justin Chen

Add support for the Broadcom ASP 2.0 Ethernet controller which is first introduced with 72165. This controller features two distinct Ethernet ports that can be independently operated.

This patch supports:

- Wake-on-LAN using magic packets
- basic ethtool operations (link, counters, message level)
- MAC destination address filtering (promiscuous, ALL_MULTI, etc.)

...

Quoted 1 line
drivers/net/ethernet/broadcom/asp2/bcmasp.c +0 -0

...

Quoted 27 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +27 -0
+static int bcmasp_netfilt_get_reg_offset(struct bcmasp_priv *priv,
+                                      struct bcmasp_net_filter *nfilt,
+                                      enum asp_netfilt_reg_type reg_type,
+                                      u32 offset)
+{
+     u32 block_index, filter_sel;
+
+     if (offset < 32) {
+             block_index = ASP_RX_FILTER_NET_L2;
+             filter_sel = nfilt->hw_index;
+     } else if (offset < 64) {
+             block_index = ASP_RX_FILTER_NET_L2;
+             filter_sel = nfilt->hw_index + 1;
+     } else if (offset < 96) {
+             block_index = ASP_RX_FILTER_NET_L3_0;
+             filter_sel = nfilt->hw_index;
+     } else if (offset < 128) {
+             block_index = ASP_RX_FILTER_NET_L3_0;
+             filter_sel = nfilt->hw_index + 1;
+     } else if (offset < 160) {
+             block_index = ASP_RX_FILTER_NET_L3_1;
+             filter_sel = nfilt->hw_index;
+     } else if (offset < 192) {
+             block_index = ASP_RX_FILTER_NET_L3_1;
+             filter_sel = nfilt->hw_index + 1;
+     } else if (offset < 224) {
+             block_index = ASP_RX_FILTER_NET_L4;
Quoted 9 lines

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 204

Quoted 5 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +5 -0
+             filter_sel = nfilt->hw_index;
+     } else if (offset < 256) {
+             block_index = ASP_RX_FILTER_NET_L4;
+             filter_sel = nfilt->hw_index + 1;
+     }

block_index and filter_sel are uninitialised if offset doesn't match any of the conditions above. Can that happen?

Nope. This is a helper function for netfilter read and write reg, we check offset sizes in those functions.

Quoted 107 lines
Quoted 12 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +12 -0
+
+     switch (reg_type) {
+     case ASP_NETFILT_MATCH:
+             return ASP_RX_FILTER_NET_PAT(filter_sel, block_index,
+                                          (offset % 32));
+     case ASP_NETFILT_MASK:
+             return ASP_RX_FILTER_NET_MASK(filter_sel, block_index,
+                                           (offset % 32));
+     default:
+             return -EINVAL;
+     }
+}

...

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 323

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+static void bcmasp_netfilt_tcpip4_wr(struct bcmasp_priv *priv,
+                                  struct bcmasp_net_filter *nfilt,
+                                  struct ethtool_tcpip4_spec *match,
+                                  struct ethtool_tcpip4_spec *mask,
+                                  u32 offset)
+{
+     __be16 val_16, mask_16;
+
+     val_16 = htons(ETH_P_IP);
+     mask_16 = 0xFFFF;

mask_17 is __be16, but 0xFFFF is host byte order.

Please make sure there are no new warnings when building with W=1 C=1.

...

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 649

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+/* If no network filter found, return open filter.
+ * If no more open filters return NULL
+ */
+struct bcmasp_net_filter *bcmasp_netfilt_get_init(struct bcmasp_intf *intf,
+                                               int loc, bool wake_filter,
+                                               bool init)
+{
+     struct bcmasp_priv *priv = intf->parent;
+     struct bcmasp_net_filter *nfilter = NULL;
+     int i, open_index = -1;

Please use reverse xmas tree - longest line to shortest - for local variable declarations in networking code.

You can check for this using https://github.com/ecree-solarflare/xmastree

...

Quoted 7 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +7 -0
+static int bcmasp_combine_set_filter(struct bcmasp_intf *intf,
+                                  unsigned char *addr, unsigned char *mask,
+                                  int i)
+{
+     u64 addr1, addr2, mask1, mask2, mask3;
+     struct bcmasp_priv *priv = intf->parent;
+

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 865

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+     /* Switch to u64 to help with the calculations */
+     addr1 = ether_addr_to_u64(priv->mda_filters[i].addr);
+     mask1 = ether_addr_to_u64(priv->mda_filters[i].mask);
+     addr2 = ether_addr_to_u64(addr);
+     mask2 = ether_addr_to_u64(mask);
+
+     /* Check if one filter resides within the other */
+     mask3 = mask1 & mask2;
+     if (mask3 == mask1 && ((addr1 & mask1) == (addr2 & mask1))) {
+             /* Filter 2 resides within fitler 1, so everthing is good */

nit: s/fitler/filter/

Please consider running ./scripts/checkpatch.pl --codespell

...

Quoted 33 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +33 -0
+static void bcmasp_update_mib_counters(struct bcmasp_intf *priv)
+{
+     int i, j = 0;
+
+     for (i = 0; i < BCMASP_STATS_LEN; i++) {
+             const struct bcmasp_stats *s;
+             u16 offset = 0;
+             u32 val = 0;
+             char *p;
+
+             s = &bcmasp_gstrings_stats[i];
+             switch (s->type) {
+             case BCMASP_STAT_NETDEV:
+             case BCMASP_STAT_SOFT:
+                     continue;
+             case BCMASP_STAT_RUNT:
+                     offset += BCMASP_STAT_OFFSET;
+                     fallthrough;
+             case BCMASP_STAT_MIB_TX:
+                     offset += BCMASP_STAT_OFFSET;
+                     fallthrough;
+             case BCMASP_STAT_MIB_RX:
+                     val = umac_rl(priv, UMC_MIB_START + j + offset);
+                     offset = 0;     /* Reset Offset */
+                     break;
+             case BCMASP_STAT_RX_EDPKT:
+                     val = rx_edpkt_core_rl(priv->parent, s->reg_offset);
+                     break;
+             case BCMASP_STAT_RX_CTRL:
+                     offset = bcmasp_stat_fixup_offset(priv, s);
+                     if (offset != ASP_RX_CTRL_FB_FILT_OUT_FRAME_COUNT)
+                             offset += sizeof(u32) * priv->port;
+                     val = rx_ctrl_core_rl(priv->parent, offset);
Quoted 9 lines

drivers/net/ethernet/broadcom/asp2/bcmasp.c

Quoted 6 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +6 -0
+                     break;
+             }
+
+             j += s->stat_sizeof;
+             p = (char *)priv + s->stat_offset;
+             *(u32 *)p = val;

Is p always 32bit aligned?

Nope. I can make sure it is 32 bit aligned.

Acked, the other comments. Will submit v3 when net-next window is open. Thank you for the review.

Justin

Quoted 71 lines
Quoted 22 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +22 -0
+     }
+}
+
+static void bcmasp_get_ethtool_stats(struct net_device *dev,
+                                  struct ethtool_stats *stats,
+                                  u64 *data)
+{
+     struct bcmasp_intf *priv = netdev_priv(dev);
+     int i, j = 0;
+
+     if (netif_running(dev))
+             bcmasp_update_mib_counters(priv);
+
+     dev->netdev_ops->ndo_get_stats(dev);
+
+     for (i = 0; i < BCMASP_STATS_LEN; i++) {
+             const struct bcmasp_stats *s;
+             char *p;
+
+             s = &bcmasp_gstrings_stats[i];
+             if (!bcmasp_stat_available(priv, s->type))
+                     continue;

drivers/net/ethernet/broadcom/asp2/bcmasp.c

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +10 -0
+             if (s->type == BCMASP_STAT_NETDEV)
+                     p = (char *)&dev->stats;
+             else
+                     p = (char *)priv;
+             p += s->stat_offset;
+             if (sizeof(unsigned long) != sizeof(u32) &&
+                 s->stat_sizeof == sizeof(unsigned long))
+                     data[j] = *(unsigned long *)p;
+             else
+                     data[j] = *(u32 *)p;

Maybe memcpy would make this a little easier to read.

Quoted 3 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c +3 -0
+             j++;
+     }
+}

...

Quoted 1 line
drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c +0 -0

...

Quoted 7 lines
drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c +7 -0
+static int bcmasp_init_rx(struct bcmasp_intf *intf)
+{
+     struct device *kdev = &intf->parent->pdev->dev;
+     struct net_device *ndev = intf->ndev;
+     void *p;
+     dma_addr_t dma;
+     struct page *buffer_pg;

drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c line 700

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c +10 -0
+     u32 reg;
+     int ret;
+
+     intf->rx_buf_order = get_order(RING_BUFFER_SIZE);
+     buffer_pg = alloc_pages(GFP_KERNEL, intf->rx_buf_order);
+
+     dma = dma_map_page(kdev, buffer_pg, 0, RING_BUFFER_SIZE,
+                        DMA_FROM_DEVICE);
+     if (dma_mapping_error(kdev, dma)) {
+             netdev_err(ndev, "Cannot allocate RX buffer\n");

I think the core will log an error on allocation failure, so the message above is not needed.

Quoted 3 lines
drivers/net/ethernet/broadcom/asp2/bcmasp_intf.c +3 -0
+             __free_pages(buffer_pg, intf->rx_buf_order);
+             return -ENOMEM;
+     }

...

#10 lore raw

Re: [PATCH v2 net-next 3/6] net: bcmasp: Add support for ASP2.0 Ethernet controller

Quoted 3 lines · Justin Chen
Quoted 2 lines · Simon Horman

On Wed, Apr 26, 2023 at 11:54:29AM -0700, Justin Chen wrote:

...

Quoted 43 lines
Quoted 33 lines
Quoted 33 lines
drivers/net/ethernet/broadcom/asp2/bcmasp_ethtool.c new file +33 -0
+static void bcmasp_update_mib_counters(struct bcmasp_intf *priv)
+{
+     int i, j = 0;
+
+     for (i = 0; i < BCMASP_STATS_LEN; i++) {
+             const struct bcmasp_stats *s;
+             u16 offset = 0;
+             u32 val = 0;
+             char *p;
+
+             s = &bcmasp_gstrings_stats[i];
+             switch (s->type) {
+             case BCMASP_STAT_NETDEV:
+             case BCMASP_STAT_SOFT:
+                     continue;
+             case BCMASP_STAT_RUNT:
+                     offset += BCMASP_STAT_OFFSET;
+                     fallthrough;
+             case BCMASP_STAT_MIB_TX:
+                     offset += BCMASP_STAT_OFFSET;
+                     fallthrough;
+             case BCMASP_STAT_MIB_RX:
+                     val = umac_rl(priv, UMC_MIB_START + j + offset);
+                     offset = 0;     /* Reset Offset */
+                     break;
+             case BCMASP_STAT_RX_EDPKT:
+                     val = rx_edpkt_core_rl(priv->parent, s->reg_offset);
+                     break;
+             case BCMASP_STAT_RX_CTRL:
+                     offset = bcmasp_stat_fixup_offset(priv, s);
+                     if (offset != ASP_RX_CTRL_FB_FILT_OUT_FRAME_COUNT)
+                             offset += sizeof(u32) * priv->port;
+                     val = rx_ctrl_core_rl(priv->parent, offset);
Quoted 9 lines

drivers/net/ethernet/broadcom/asp2/bcmasp_ethtool.c line 286

Quoted 6 lines
drivers/net/ethernet/broadcom/asp2/bcmasp_ethtool.c new file +6 -0
+                     break;
+             }
+
+             j += s->stat_sizeof;
+             p = (char *)priv + s->stat_offset;
+             *(u32 *)p = val;

Is p always 32bit aligned?

Quoted 1 line

Nope. I can make sure it is 32 bit aligned.

I'm not sure if it helps, but you could also consider put_unaligned().

Quoted 2 lines

Acked, the other comments. Will submit v3 when net-next window is open. Thank you for the review.

Likewise, thanks.

#11 lore raw

Re: [PATCH v2 net-next 3/6] net: bcmasp: Add support for ASP2.0 Ethernet controller

In reply to Justin Chen

Le 26/04/2023 à 20:54, Justin Chen a écrit :

Quoted 13 lines

Add support for the Broadcom ASP 2.0 Ethernet controller which is first introduced with 72165. This controller features two distinct Ethernet ports that can be independently operated.

This patch supports:

- Wake-on-LAN using magic packets
- basic ethtool operations (link, counters, message level)
- MAC destination address filtering (promiscuous, ALL_MULTI, etc.)

---

[...]

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 839

Quoted 4 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +4 -0
+void bcmasp_disable_all_filters(struct bcmasp_intf *intf)
+{
+	struct bcmasp_priv *priv = intf->parent;
+	unsigned int i;

Hi,

Nit: Some loop index are unsigned int, but most are int. This could be done consistantly.

Quoted 8 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +8 -0
+
+	/* Disable all filters held by this port */
+	for (i = ASP_RX_FILT_MDA_RES_COUNT(intf); i < NUM_MDA_FILTERS; i++) {
+		if (priv->mda_filters[i].en &&
+		    priv->mda_filters[i].port == intf->port)
+			bcmasp_en_mda_filter(intf, 0, i);
+	}
+}

[...]

Quoted 25 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +25 -0
+static int bcmasp_probe(struct platform_device *pdev)
+{
+	struct device_node *ports_node, *intf_node;
+	const struct bcmasp_plat_data *pdata;
+	struct device *dev = &pdev->dev;
+	int ret, i, count = 0, port;
+	struct bcmasp_priv *priv;
+	struct bcmasp_intf *intf;
+
+	priv = devm_kzalloc(dev, sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
+
+	priv->irq = platform_get_irq(pdev, 0);
+	if (priv->irq <= 0) {
+		dev_err(dev, "invalid interrupt\n");
+		return -EINVAL;
+	}
+
+	priv->clk = devm_clk_get_optional_enabled(dev, "sw_asp");
+	if (IS_ERR(priv->clk)) {
+		dev_err(dev, "failed to request clock\n");
+		return PTR_ERR(priv->clk);
+	}
+

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 1246

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +10 -0
+	/* Base from parent node */
+	priv->base = devm_platform_ioremap_resource(pdev, 0);
+	if (IS_ERR(priv->base)) {
+		dev_err(dev, "failed to iomap\n");
+		return PTR_ERR(priv->base);
+	}
+
+	ret = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));
+	if (ret)
+		ret = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(32));

I don't think that this fallback is needed.
See [1].

More over, using dev_err_probe() would slighly simplify the probe function. (saves a few LoC, logs the error code in a human reading format)

[1]: https://lore.kernel.org/lkml/[email protected]/T/#m022abc0051ede3ba1feeb06cefd59e2a8a5c7864

Quoted 5 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +5 -0
+	if (ret) {
+		dev_err(&pdev->dev, "unable to set DMA mask: %d\n", ret);
+		return ret;
+	}
+

[...]

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 1391

Quoted 5 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +5 -0
+static int __maybe_unused bcmasp_suspend(struct device *d)
+{
+	struct bcmasp_priv *priv = dev_get_drvdata(d);
+	struct bcmasp_intf *intf;
+	unsigned int i;

Same

Review

Quoted 1 line
(quoted diff) +1 -0
+	int ret = 0;

no need to initialize, but it is mostmy a matter of taste.

Quoted 22 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +22 -0
+
+	for (i = 0; i < priv->intf_count; i++) {
+		intf = priv->intfs[i];
+		if (!intf)
+			continue;
+
+		ret = bcmasp_interface_suspend(intf);
+		if (ret)
+			break;
+	}
+
+	ret = clk_prepare_enable(priv->clk);
+	if (ret)
+		return ret;
+
+	/* Whether Wake-on-LAN is enabled or not, we can always disable
+	 * the shared TX clock
+	 */
+	bcmasp_core_clock_set(priv, 0, ASP_CTRL_CLOCK_CTRL_ASP_TX_DISABLE);
+
+	bcmasp_core_clock_select(priv, true);
+

drivers/net/ethernet/broadcom/asp2/bcmasp.c line 1424

Quoted 10 lines
drivers/net/ethernet/broadcom/asp2/bcmasp.c new file +10 -0
+	clk_disable_unprepare(priv->clk);
+
+	return ret;
+}
+
+static int __maybe_unused bcmasp_resume(struct device *d)
+{
+	struct bcmasp_priv *priv = dev_get_drvdata(d);
+	struct bcmasp_intf *intf;
+	unsigned int i;

same

Review

Quoted 1 line
(quoted diff) +1 -0
+	int ret = 0;

no need to initialize, but it is mostmy a matter of taste.

Just my 2c,
CJ

#12 lore raw

PATCH v2 4/6 net: phy: mdio-bcm-unimac: Add asp v2.0 support

In reply to Justin Chen

Add mdio compat string for ASP 2.0 ethernet driver.

---
 drivers/net/mdio/mdio-bcm-unimac.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/net/mdio/mdio-bcm-unimac.c +2 -0
index bfc9be23c973..6b26a0803696 100644
@@ -334,6 +334,8 @@ static SIMPLE_DEV_PM_OPS(unimac_mdio_pm_ops,
 			 unimac_mdio_suspend, unimac_mdio_resume);
 
 static const struct of_device_id unimac_mdio_ids[] = {
+	{ .compatible = "brcm,asp-v2.1-mdio", },
+	{ .compatible = "brcm,asp-v2.0-mdio", },
 	{ .compatible = "brcm,genet-mdio-v5", },
 	{ .compatible = "brcm,genet-mdio-v4", },
 	{ .compatible = "brcm,genet-mdio-v3", },
Signature · 1 line
2.7.4
#13 lore raw

PATCH v2 5/6 net: phy: bcm7xxx: Add EPHY entry for 74165

In reply to Justin Chen

From: Florian Fainelli <[email protected]>

74165 is a 16nm process SoC with a 10/100 integrated Ethernet PHY, utilize the recently defined 16nm EPHY macro to configure that PHY.

---
 drivers/net/phy/bcm7xxx.c | 1 +
 include/linux/brcmphy.h   | 1 +
 2 files changed, 2 insertions(+)
drivers/net/phy/bcm7xxx.c +1 -0
index 06be71ecd2f8..5c03c379cb5e 100644
@@ -913,6 +913,7 @@ static struct phy_driver bcm7xxx_driver[] = {
 	BCM7XXX_28NM_GPHY(PHY_ID_BCM7278, "Broadcom BCM7278"),
 	BCM7XXX_28NM_GPHY(PHY_ID_BCM7364, "Broadcom BCM7364"),
 	BCM7XXX_28NM_GPHY(PHY_ID_BCM7366, "Broadcom BCM7366"),
+	BCM7XXX_16NM_EPHY(PHY_ID_BCM74165, "Broadcom BCM74165"),
 	BCM7XXX_28NM_GPHY(PHY_ID_BCM74371, "Broadcom BCM74371"),
 	BCM7XXX_28NM_GPHY(PHY_ID_BCM7439, "Broadcom BCM7439"),
 	BCM7XXX_28NM_GPHY(PHY_ID_BCM7439_2, "Broadcom BCM7439 (2)"),
include/linux/brcmphy.h +1 -0
index 9e77165f3ef6..e11c2e9a5398 100644
@@ -44,6 +44,7 @@
 #define PHY_ID_BCM7366			0x600d8490
 #define PHY_ID_BCM7346			0x600d8650
 #define PHY_ID_BCM7362			0x600d84b0
+#define PHY_ID_BCM74165			0x359052c0
 #define PHY_ID_BCM7425			0x600d86b0
 #define PHY_ID_BCM7429			0x600d8730
 #define PHY_ID_BCM7435			0x600d8750
Signature · 1 line
2.7.4
#14 lore raw

PATCH v2 6/6 MAINTAINERS: ASP 2.0 Ethernet driver maintainers

In reply to Justin Chen

Add maintainers entry for ASP 2.0 Ethernet driver.

---
 MAINTAINERS | 9 +++++++++
 1 file changed, 9 insertions(+)
MAINTAINERS +9 -0
index 4fc57dfd5fd0..24cbe1c0fc06 100644
@@ -4189,6 +4189,15 @@ F:	drivers/net/mdio/mdio-bcm-unimac.c
 F:	include/linux/platform_data/bcmgenet.h
 F:	include/linux/platform_data/mdio-bcm-unimac.h
 
+BROADCOM ASP 2.0 ETHERNET DRIVER
+M:	Justin Chen <[email protected]>
+M:	Florian Fainelli <[email protected]>
+L:	[email protected]
+L:	[email protected]
+S:	Supported
+F:	Documentation/devicetree/bindings/net/brcm,asp-v2.0.yaml
+F:	drivers/net/ethernet/broadcom/asp2/
+
 BROADCOM IPROC ARM ARCHITECTURE
 M:	Ray Jui <[email protected]>
 M:	Scott Branden <[email protected]>
Signature · 1 line
2.7.4
maintainer
#15 lore raw
In reply to Justin Chen
Quoted 11 lines · Justin Chen

v2
- Updates to yaml dt documentation
- Replace a couple functions with helper functions
- Minor formatting fixes
- Fix a few WoL issues

Add support for the Broadcom ASP 2.0 Ethernet controller which is first introduced with 72165.

Add support for 74165 10/100 integrated Ethernet PHY which also uses the ASP 2.0 Ethernet controller.

Note that the changelog should be located after the cover letter itself. Changelog in the individual patches should be more helpful, too. In that case, it should be located after the tags and a '---' separator.

## Form letter - net-next-closed

The merge window for v6.3 has begun and therefore net-next is closed for new drivers, features, code refactoring and optimizations. We are currently accepting bug fixes only.

Please repost when net-next reopens after May 8th.

RFC patches sent for review only are obviously welcome at any time.

Signature · 1 line
pw-bot: defer