further mt7988 devicetree work
Subsystems (7)
PATCH v8 2/16 dt-bindings: net: mediatek,net: allow up to 8 IRQs
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Increase the maximum IRQ count to 8 (4 FE + 4 RSS/LRO).
Frame-engine-IRQs (max 4):
MT7621, MT7628: 1 IRQ
MT7622, MT7623: 3 IRQs (only two used by the driver for now) MT7981, MT7986, MT7988: 4 IRQs (only two used by the driver for now)
Mediatek Filogic SoCs (mt798x) have 4 additional IRQs for RSS and/or LRO.
---
v8: separate irq-count change from interrupt-names patch
--- Documentation/devicetree/bindings/net/mediatek,net.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Documentation/devicetree/bindings/net/mediatek,net.yaml +1 -1
@@ -40,7 +40,7 @@ properties: interrupts: minItems: 1 - maxItems: 4 + maxItems: 8 power-domains: maxItems: 1
Signature · 1 line
2.43.0
PATCH v8 4/16 dt-bindings: net: mediatek,net: add sram property
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Meditak Filogic SoCs (MT798x) have dedicated MMIO-SRAM for dma operations.
MT7981 and MT7986 currently use static offset to ethernet MAC register which will be changed in separate patch once this way is accepted.
Add "sram" property to map ethernet controller to dedicated mmio-sram node.
---
v8:
- splitted out mac subnode pattern
- dropped reg naming change
- rephrased description
- drop change of reg-name
v6:
- split out the interrupt-names into separate patch
- update irq(name) min count to 4
- add sram-property
- drop second reg entry and minitems as there is only 1 item left again
v5:
- fix v4 logmessage and change description a bit describing how i get the irq count.
- update binding for 8 irqs with different names (rx,tx => fe0..fe3) including the 2 reserved irqs which can be used later
- change rx-ringX to pdmaX to be closer to hardware documentation
v4:
- increase max interrupts to 6 because of adding RSS/LRO interrupts (4) and dropping 2 reserved irqs (0+3) around rx+tx
- dropped Robs RB due to this change
- allow interrupt names
- add interrupt-names without reserved IRQs on mt7988
this requires mtk driver patch:
https://patchwork.kernel.org/project/netdevbpf/patch/[email protected]/
v2:
- change reg to list of items
--- Documentation/devicetree/bindings/net/mediatek,net.yaml | 4 ++++ 1 file changed, 4 insertions(+)
Documentation/devicetree/bindings/net/mediatek,net.yaml +4 -0
@@ -66,6 +66,10 @@ properties: - const: gmac - const: ppe + sram: + $ref: /schemas/types.yaml#/definitions/phandle + description: phandle to mmio SRAM + mediatek,ethsys: $ref: /schemas/types.yaml#/definitions/phandle description:
Signature · 1 line
2.43.0
PATCH v8 14/16 arm64: dts: mediatek: mt7988a-bpi-r4: add aliases for ethernet
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Add aliases for gmacs to allow bootloader setting mac-adresses.
--- arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi | 6 ++++++ 1 file changed, 6 insertions(+)
arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi +6 -0
@@ -8,6 +8,12 @@ #include "mt7988a.dtsi" / { + aliases { + ethernet0 = &gmac0; + ethernet1 = &gmac1; + ethernet2 = &gmac2; + }; + chosen { stdout-path = "serial0:115200n8"; };
Signature · 1 line
2.43.0
PATCH v8 15/16 arm64: dts: mediatek: mt7988a-bpi-r4: add sfp cages and link to gmac
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Add SFP cages to Bananapi-R4 board. The 2.5g phy variant only contains the wan-SFP, so add this to common dtsi and the lan-sfp only to the dual-SFP variant.
---
v4:
- update 2g5-board (reorder and drop phy-connection-type)
- order sfp properties
v3:
- enable mac with 2.5g phy on r4 phy variant because driver is now mainline
--- .../mediatek/mt7988a-bananapi-bpi-r4-2g5.dts | 11 +++++++++++ .../dts/mediatek/mt7988a-bananapi-bpi-r4.dts | 19 +++++++++++++++++++ .../dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi | 19 +++++++++++++++++++ 3 files changed, 49 insertions(+)
arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4-2g5.dts +11 -0
@@ -9,3 +9,14 @@ / { model = "Banana Pi BPI-R4 (1x SFP+, 1x 2.5GbE)"; chassis-type = "embedded"; }; + +&gmac1 { + phy = <&int_2p5g_phy>; + phy-mode = "internal"; + status = "okay"; +}; + +&int_2p5g_phy { + pinctrl-0 = <&i2p5gbe_led0_pins>; + pinctrl-names = "i2p5gbe-led"; +};
arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dts +19 -0
@@ -8,6 +8,25 @@ / { compatible = "bananapi,bpi-r4", "mediatek,mt7988a"; model = "Banana Pi BPI-R4 (2x SFP+)"; chassis-type = "embedded"; + + /* SFP2 cage (LAN) */ + sfp2: sfp2 { + compatible = "sff,sfp"; + i2c-bus = <&i2c_sfp2>; + maximum-power-milliwatt = <3000>; + + los-gpios = <&pio 2 GPIO_ACTIVE_HIGH>; + mod-def0-gpios = <&pio 83 GPIO_ACTIVE_LOW>; + rate-select0-gpios = <&pio 3 GPIO_ACTIVE_LOW>; + tx-disable-gpios = <&pio 0 GPIO_ACTIVE_HIGH>; + tx-fault-gpios = <&pio 1 GPIO_ACTIVE_HIGH>; + }; +}; + +&gmac1 { + managed = "in-band-status"; + phy-mode = "usxgmii"; + sfp = <&sfp2>; }; &pca9545 {
arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi +19 -0
@@ -63,6 +63,19 @@ reg_3p3v: regulator-3p3v { regulator-boot-on; regulator-always-on; }; + + /* SFP1 cage (WAN) */ + sfp1: sfp1 { + compatible = "sff,sfp"; + i2c-bus = <&i2c_sfp1>; + maximum-power-milliwatt = <3000>; + + los-gpios = <&pio 54 GPIO_ACTIVE_HIGH>; + mod-def0-gpios = <&pio 82 GPIO_ACTIVE_LOW>; + rate-select0-gpios = <&pio 21 GPIO_ACTIVE_LOW>; + tx-disable-gpios = <&pio 70 GPIO_ACTIVE_HIGH>; + tx-fault-gpios = <&pio 69 GPIO_ACTIVE_HIGH>; + }; }; &cci {
@@ -133,6 +146,12 @@ map-cpu-active-low { }; }; +&gmac2 { + managed = "in-band-status"; + phy-mode = "usxgmii"; + sfp = <&sfp1>; +}; + &i2c0 { pinctrl-names = "default"; pinctrl-0 = <&i2c0_pins>;
Signature · 1 line
2.43.0
PATCH v8 16/16 arm64: dts: mediatek: mt7988a-bpi-r4: configure switch phys and leds
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Assign pinctrl to switch phys and leds.
---
v4:
- reorder switch phy(-led) properties
v2:
- add labels and led-function and include after dropping from soc dtsi
--- .../dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi | 61 +++++++++++++++++++ 1 file changed, 61 insertions(+)
arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi +61 -0
@@ -4,6 +4,7 @@ #include <dt-bindings/gpio/gpio.h> #include <dt-bindings/regulator/richtek,rt5190a-regulator.h> +#include <dt-bindings/leds/common.h> #include "mt7988a.dtsi"
@@ -152,6 +153,66 @@ &gmac2 { sfp = <&sfp1>; }; +&gsw_phy0 { + pinctrl-0 = <&gbe0_led0_pins>; + pinctrl-names = "gbe-led"; +}; + +&gsw_phy0_led0 { + function = LED_FUNCTION_WAN; + color = <LED_COLOR_ID_GREEN>; + status = "okay"; +}; + +&gsw_port0 { + label = "wan"; +}; + +&gsw_phy1 { + pinctrl-0 = <&gbe1_led0_pins>; + pinctrl-names = "gbe-led"; +}; + +&gsw_phy1_led0 { + function = LED_FUNCTION_LAN; + color = <LED_COLOR_ID_GREEN>; + status = "okay"; +}; + +&gsw_port1 { + label = "lan1"; +}; + +&gsw_phy2 { + pinctrl-0 = <&gbe2_led0_pins>; + pinctrl-names = "gbe-led"; +}; + +&gsw_phy2_led0 { + function = LED_FUNCTION_LAN; + color = <LED_COLOR_ID_GREEN>; + status = "okay"; +}; + +&gsw_port2 { + label = "lan2"; +}; + +&gsw_phy3 { + pinctrl-0 = <&gbe3_led0_pins>; + pinctrl-names = "gbe-led"; +}; + +&gsw_phy3_led0 { + function = LED_FUNCTION_LAN; + color = <LED_COLOR_ID_GREEN>; + status = "okay"; +}; + +&gsw_port3 { + label = "lan3"; +}; + &i2c0 { pinctrl-names = "default"; pinctrl-0 = <&i2c0_pins>;
Signature · 1 line
2.43.0
PATCH v8 11/16 arm64: dts: mediatek: mt7988a-bpi-r4: add proc-supply for cci
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
CCI requires proc-supply. Add it on board level.
--- arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi | 4 ++++ 1 file changed, 4 insertions(+)
arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi +4 -0
@@ -40,6 +40,10 @@ reg_3p3v: regulator-3p3v { }; }; +&cci { + proc-supply = <&rt5190_buck3>; +}; + &cpu0 { proc-supply = <&rt5190_buck3>; };
Signature · 1 line
2.43.0
PATCH v8 13/16 arm64: dts: mediatek: mt7988a-bpi-r4: add gpio leds
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Bananapi R4 has a green and a blue led which can be switched by gpio. Green led is for running state so default on.
Green led also shares pin with eeprom writeprotect where led off allows writing to eeprom.
--- .../dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+)
arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi +19 -0
@@ -21,6 +21,25 @@ fan: pwm-fan { status = "okay"; }; + gpio-leds { + compatible = "gpio-leds"; + + led_green: led-green { + function = LED_FUNCTION_STATUS; + color = <LED_COLOR_ID_GREEN>; + gpios = <&pio 79 GPIO_ACTIVE_HIGH>; + default-state = "on"; + }; + + led_blue: led-blue { + function = LED_FUNCTION_WPS; + color = <LED_COLOR_ID_BLUE>; + gpios = <&pio 63 GPIO_ACTIVE_HIGH>; + default-state = "off"; + }; + }; + + reg_1p8v: regulator-1p8v { compatible = "regulator-fixed"; regulator-name = "fixed-1.8V";
Signature · 1 line
2.43.0
PATCH v8 8/16 arm64: dts: mediatek: mt7988: add cci node
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Add cci devicetree node for cpu frequency scaling.
---
v3:
- add mt7988-cci compatible as suggested by angelo
--- arch/arm64/boot/dts/mediatek/mt7988a.dtsi | 33 +++++++++++++++++++++++ 1 file changed, 33 insertions(+)
arch/arm64/boot/dts/mediatek/mt7988a.dtsi +33 -0
@@ -12,6 +12,35 @@ / { #address-cells = <2>; #size-cells = <2>; + cci: cci { + compatible = "mediatek,mt7988-cci", "mediatek,mt8183-cci"; + clocks = <&mcusys CLK_MCU_BUS_DIV_SEL>, + <&topckgen CLK_TOP_XTAL>; + clock-names = "cci", "intermediate"; + operating-points-v2 = <&cci_opp>; + }; + + cci_opp: opp-table-cci { + compatible = "operating-points-v2"; + opp-shared; + opp-480000000 { + opp-hz = /bits/ 64 <480000000>; + opp-microvolt = <850000>; + }; + opp-660000000 { + opp-hz = /bits/ 64 <660000000>; + opp-microvolt = <850000>; + }; + opp-900000000 { + opp-hz = /bits/ 64 <900000000>; + opp-microvolt = <850000>; + }; + opp-1080000000 { + opp-hz = /bits/ 64 <1080000000>; + opp-microvolt = <900000>; + }; + }; + cpus { #address-cells = <1>; #size-cells = <0>;
@@ -25,6 +54,7 @@ cpu0: cpu@0 { <&topckgen CLK_TOP_XTAL>; clock-names = "cpu", "intermediate"; operating-points-v2 = <&cluster0_opp>; + mediatek,cci = <&cci>; }; cpu1: cpu@1 {
@@ -36,6 +66,7 @@ cpu1: cpu@1 { <&topckgen CLK_TOP_XTAL>; clock-names = "cpu", "intermediate"; operating-points-v2 = <&cluster0_opp>; + mediatek,cci = <&cci>; }; cpu2: cpu@2 {
@@ -47,6 +78,7 @@ cpu2: cpu@2 { <&topckgen CLK_TOP_XTAL>; clock-names = "cpu", "intermediate"; operating-points-v2 = <&cluster0_opp>; + mediatek,cci = <&cci>; }; cpu3: cpu@3 {
@@ -58,6 +90,7 @@ cpu3: cpu@3 { <&topckgen CLK_TOP_XTAL>; clock-names = "cpu", "intermediate"; operating-points-v2 = <&cluster0_opp>; + mediatek,cci = <&cci>; }; cluster0_opp: opp-table-0 {
Signature · 1 line
2.43.0
PATCH v8 12/16 arm64: dts: mediatek: mt7988a-bpi-r4: drop unused pins
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Pins were moved from SoC dtsi to Board level dtsi without cleaning up to needed ones. Drop the unused pins now.
--- .../dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi | 89 ------------------- 1 file changed, 89 deletions(-)
arch/arm64/boot/dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi +0 -89
@@ -223,18 +223,6 @@ &pcie3 { }; &pio { - mdio0_pins: mdio0-pins { - mux { - function = "eth"; - groups = "mdc_mdio0"; - }; - - conf { - pins = "SMI_0_MDC", "SMI_0_MDIO"; - drive-strength = <8>; - }; - }; - i2c0_pins: i2c0-g0-pins { mux { function = "i2c";
@@ -249,20 +237,6 @@ mux { }; }; - i2c1_sfp_pins: i2c1-sfp-g0-pins { - mux { - function = "i2c"; - groups = "i2c1_sfp"; - }; - }; - - i2c2_0_pins: i2c2-g0-pins { - mux { - function = "i2c"; - groups = "i2c2_0"; - }; - }; - i2c2_1_pins: i2c2-g1-pins { mux { function = "i2c";
@@ -298,34 +272,6 @@ mux { }; }; - gbe0_led1_pins: gbe0-led1-pins { - mux { - function = "led"; - groups = "gbe0_led1"; - }; - }; - - gbe1_led1_pins: gbe1-led1-pins { - mux { - function = "led"; - groups = "gbe1_led1"; - }; - }; - - gbe2_led1_pins: gbe2-led1-pins { - mux { - function = "led"; - groups = "gbe2_led1"; - }; - }; - - gbe3_led1_pins: gbe3-led1-pins { - mux { - function = "led"; - groups = "gbe3_led1"; - }; - }; - i2p5gbe_led0_pins: 2p5gbe-led0-pins { mux { function = "led";
@@ -333,13 +279,6 @@ mux { }; }; - i2p5gbe_led1_pins: 2p5gbe-led1-pins { - mux { - function = "led"; - groups = "2p5gbe_led1"; - }; - }; - mmc0_pins_emmc_45: mmc0-emmc-45-pins { mux { function = "flash";
@@ -361,40 +300,12 @@ mux { }; }; - snfi_pins: snfi-pins { - mux { - function = "flash"; - groups = "snfi"; - }; - }; - - spi0_pins: spi0-pins { - mux { - function = "spi"; - groups = "spi0"; - }; - }; - spi0_flash_pins: spi0-flash-pins { mux { function = "spi"; groups = "spi0", "spi0_wp_hold"; }; }; - - spi2_pins: spi2-pins { - mux { - function = "spi"; - groups = "spi2"; - }; - }; - - spi2_flash_pins: spi2-flash-pins { - mux { - function = "spi"; - groups = "spi2", "spi2_wp_hold"; - }; - }; }; &pwm {
Signature · 1 line
2.43.0
PATCH v8 6/16 dt-bindings: net: dsa: mediatek,mt7530: add internal mdio bus
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Mt7988 buildin switch has own mdio bus where ge-phys are connected. Add related property for this.
---
v2:
- change from patternproperty to property
- add unevaluatedProperties and mediatek,pio subproperty
--- .../devicetree/bindings/net/dsa/mediatek,mt7530.yaml | 10 ++++++++++ 1 file changed, 10 insertions(+)
Documentation/devicetree/bindings/net/dsa/mediatek,mt7530.yaml +10 -0
@@ -136,6 +136,16 @@ properties: See Documentation/devicetree/bindings/regulator/mt6323-regulator.txt for details for the regulator setup on these boards. + mdio: + $ref: /schemas/net/mdio.yaml# + unevaluatedProperties: false + + properties: + mediatek,pio: + $ref: /schemas/types.yaml#/definitions/phandle + description: + Phandle pointing to the mediatek pinctrl node. + mediatek,mcm: type: boolean description:
Signature · 1 line
2.43.0
PATCH v8 10/16 arm64: dts: mediatek: mt7988: add switch node
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Add mt7988 builtin mt753x switch nodes.
---
v4:
- drop phy-mode for gsw-phy
- reorder phy-mode after phy-handle
- drop interrupt parent from switch
v2:
- drop labels and led-function too (have to be in board)
--- arch/arm64/boot/dts/mediatek/mt7988a.dtsi | 148 ++++++++++++++++++++++ 1 file changed, 148 insertions(+)
arch/arm64/boot/dts/mediatek/mt7988a.dtsi +148 -0
@@ -742,6 +742,154 @@ ethsys: clock-controller@15000000 { #reset-cells = <1>; }; + switch: switch@15020000 { + compatible = "mediatek,mt7988-switch"; + reg = <0 0x15020000 0 0x8000>; + interrupt-controller; + #interrupt-cells = <1>; + interrupts = <GIC_SPI 209 IRQ_TYPE_LEVEL_HIGH>; + resets = <ðwarp MT7988_ETHWARP_RST_SWITCH>; + + ports { + #address-cells = <1>; + #size-cells = <0>; + + gsw_port0: port@0 { + reg = <0>; + phy-handle = <&gsw_phy0>; + phy-mode = "internal"; + }; + + gsw_port1: port@1 { + reg = <1>; + phy-handle = <&gsw_phy1>; + phy-mode = "internal"; + }; + + gsw_port2: port@2 { + reg = <2>; + phy-handle = <&gsw_phy2>; + phy-mode = "internal"; + }; + + gsw_port3: port@3 { + reg = <3>; + phy-handle = <&gsw_phy3>; + phy-mode = "internal"; + }; + + port@6 { + reg = <6>; + ethernet = <&gmac0>; + phy-mode = "internal"; + + fixed-link { + speed = <10000>; + full-duplex; + pause; + }; + }; + }; + + mdio { + #address-cells = <1>; + #size-cells = <0>; + mediatek,pio = <&pio>; + + gsw_phy0: ethernet-phy@0 { + compatible = "ethernet-phy-ieee802.3-c22"; + reg = <0>; + interrupts = <0>; + nvmem-cells = <&phy_calibration_p0>; + nvmem-cell-names = "phy-cal-data"; + + leds { + #address-cells = <1>; + #size-cells = <0>; + + gsw_phy0_led0: led@0 { + reg = <0>; + status = "disabled"; + }; + + gsw_phy0_led1: led@1 { + reg = <1>; + status = "disabled"; + }; + }; + }; + + gsw_phy1: ethernet-phy@1 { + compatible = "ethernet-phy-ieee802.3-c22"; + reg = <1>; + interrupts = <1>; + nvmem-cells = <&phy_calibration_p1>; + nvmem-cell-names = "phy-cal-data"; + + leds { + #address-cells = <1>; + #size-cells = <0>; + + gsw_phy1_led0: led@0 { + reg = <0>; + status = "disabled"; + }; + + gsw_phy1_led1: led@1 { + reg = <1>; + status = "disabled"; + }; + }; + }; + + gsw_phy2: ethernet-phy@2 { + compatible = "ethernet-phy-ieee802.3-c22"; + reg = <2>; + interrupts = <2>; + nvmem-cells = <&phy_calibration_p2>; + nvmem-cell-names = "phy-cal-data"; + + leds { + #address-cells = <1>; + #size-cells = <0>; + + gsw_phy2_led0: led@0 { + reg = <0>; + status = "disabled"; + }; + + gsw_phy2_led1: led@1 { + reg = <1>; + status = "disabled"; + }; + }; + }; + + gsw_phy3: ethernet-phy@3 { + compatible = "ethernet-phy-ieee802.3-c22"; + reg = <3>; + interrupts = <3>; + nvmem-cells = <&phy_calibration_p3>; + nvmem-cell-names = "phy-cal-data"; + + leds { + #address-cells = <1>; + #size-cells = <0>; + + gsw_phy3_led0: led@0 { + reg = <0>; + status = "disabled"; + }; + + gsw_phy3_led1: led@1 { + reg = <1>; + status = "disabled"; + }; + }; + }; + }; + }; + ethwarp: clock-controller@15031000 { compatible = "mediatek,mt7988-ethwarp"; reg = <0 0x15031000 0 0x1000>;
Signature · 1 line
2.43.0
v8 cover further mt7988 devicetree work
From: Frank Wunderlich <[email protected]>
This series continues mt7988 devicetree work
- Extend cpu frequency scaling with CCI
- GPIO leds
- Basic network-support (ethernet controller + builtin switch + SFP Cages)
depencies (i hope this list is complete and latest patches/series linked):
support interrupt-names is optional again as i re-added the reserved IRQs (they are not unusable as i thought and can allow features in future)
https://patchwork.kernel.org/project/netdevbpf/patch/[email protected]/
needs change in mtk ethernet driver for the sram to be read from separate node:
https://patchwork.kernel.org/project/netdevbpf/patch/c2b9242229d06af4e468204bcf42daa1535c3a72.1751461762.git.daniel@makrotopia.org/
for SFP-Function (macs currently disabled):
PCS clearance which is a 1.5 year discussion currently ongoing
Daniel asked netdev for a way 2 go:
https://lore.kernel.org/netdev/[email protected]/
e.g. something like this (one of):
* https://patchwork.kernel.org/project/netdevbpf/patch/[email protected]/ (v6)
* https://patchwork.kernel.org/project/netdevbpf/patch/[email protected]/ (v4)
* https://patchwork.kernel.org/project/netdevbpf/patch/ba4e359584a6b3bc4b3470822c42186d5b0856f9.1721910728.git.daniel@makrotopia.org/
full usxgmii driver:
https://patchwork.kernel.org/project/netdevbpf/patch/07845ec900ba41ff992875dce12c622277592c32.1702352117.git.daniel@makrotopia.org/
first PCS-discussion is here:
https://patchwork.kernel.org/project/netdevbpf/patch/8aa905080bdb6760875d62cb3b2b41258837f80e.1702352117.git.daniel@makrotopia.org/
some more here:
https://lore.kernel.org/netdev/[email protected]/
and then dts nodes for sgmiisys+usxgmii+2g5 firmware
when above depencies are solved the mac1/2 can be enabled and 2.5G phy/SFP slots will work.
changes:
v8:
- splitted binding into irq-count and irq-names and updated description
- fixed typo in mt7621 section "interrupt-namess"
- splitted binding changes into fixes (mac) and additions (sram)
- dropped change of reg (simple count to description)
- change ethernet register size to 0x40000
v7:
- fixed rebasing error while splitting dt-binding patch
v6:
binding:
- split out the interrupt-names into separate patch
- update irq(name) min count to 4
- move interrupt-names up
- add sram-property
- drop second reg entry and minitems as there is only 1 item left
dts:
- fix whitespace-errors for pdma irqs (spaces vs. tabs)
- move sram from eth reg to own sram node (needs CONFIG_SRAM)
v5:
- add reserved irqs and change names
- update binding for 8 irqs with different names (rx,tx => fe1+fe2, rx-ringX => pdmaX) (dropped Robs RB due to this change again, sorry)
v4:
net-binding:
- allow interrupt names and increase max interrupts to 6 because of RSS/LRO interrupts (dropped Robs RB due to this change)
dts-patches:
- add interrupts for RSS/LRO and interrupt-names for ethernet node
- eth-reg and clock whitespace-fix
- comment for fixed-link on gmac0
- drop phy-mode properties as suggested by andrew
- drop phy-connection-type on 2g5 board
- reorder some properties
- update 2g5 phy node
- unit-name dec instead of hex to match reg property
- move compatible before reg
- drop phy-mode
v3:
- dropped patches already applied (SPI+thermal)
- added soc specific cci compatible (new binding patch + changed dts)
- enable 2g5 phy because driver is now merged
- add patch for cleaning up unnecessary pins
- add patch for gpio-leds
- add patch for adding ethernet aliases
v2:
- change reg to list of items in eth binding
- changed mt7530 binding:
- unevaluatedProperties=false
- mediatek,pio subproperty
- from patternProperty to property
- board specific properties like led function and labels moved to bpi-r4 dtsi
Frank Wunderlich (16):
dt-bindings: net: mediatek,net: update mac subnode pattern for mt7988
dt-bindings: net: mediatek,net: allow up to 8 IRQs
dt-bindings: net: mediatek,net: allow irq names
dt-bindings: net: mediatek,net: add sram property
dt-bindings: net: dsa: mediatek,mt7530: add dsa-port definition for
mt7988
dt-bindings: net: dsa: mediatek,mt7530: add internal mdio bus
dt-bindings: interconnect: add mt7988-cci compatible
arm64: dts: mediatek: mt7988: add cci node
arm64: dts: mediatek: mt7988: add basic ethernet-nodes
arm64: dts: mediatek: mt7988: add switch node
arm64: dts: mediatek: mt7988a-bpi-r4: add proc-supply for cci
arm64: dts: mediatek: mt7988a-bpi-r4: drop unused pins
arm64: dts: mediatek: mt7988a-bpi-r4: add gpio leds
arm64: dts: mediatek: mt7988a-bpi-r4: add aliases for ethernet
arm64: dts: mediatek: mt7988a-bpi-r4: add sfp cages and link to gmac
arm64: dts: mediatek: mt7988a-bpi-r4: configure switch phys and leds
.../bindings/interconnect/mediatek,cci.yaml | 11 +- .../bindings/net/dsa/mediatek,mt7530.yaml | 24 +- .../devicetree/bindings/net/mediatek,net.yaml | 44 ++- .../mediatek/mt7988a-bananapi-bpi-r4-2g5.dts | 11 + .../dts/mediatek/mt7988a-bananapi-bpi-r4.dts | 19 ++ .../dts/mediatek/mt7988a-bananapi-bpi-r4.dtsi | 198 ++++++----- arch/arm64/boot/dts/mediatek/mt7988a.dtsi | 318 +++++++++++++++++- 7 files changed, 527 insertions(+), 98 deletions(-)
Signature · 1 line
2.43.0
PATCH v8 3/16 dt-bindings: net: mediatek,net: allow irq names
From: Frank Wunderlich <[email protected]>
In preparation for MT7988 and RSS/LRO allow the interrupt-names property.
In this way driver can request the interrupts by name which is much more readable in the driver code and SoC's dtsi than relying on a specific order.
Frame-engine-IRQs (fe0..3):
MT7621, MT7628: 1 IRQ
MT7622, MT7623: 3 IRQs (only two used by the driver for now) MT7981, MT7986: 4 IRQs (only two used by the driver for now)
RSS/LRO IRQs (pdma0..3) only on Filogic (MT798x) with count of 4.
Set boundaries for all compatibles same as irq count.
---
v8:
- fixed typo in mt7621 section "interrupt-namess"
- separated interrupt count from interrupt-names
- rephrased description a bit to explain the "why"
v7: fixed wrong rebase
v6: new patch splitted from the mt7988 changes
--- .../devicetree/bindings/net/mediatek,net.yaml | 36 +++++++++++++++++++ 1 file changed, 36 insertions(+)
Documentation/devicetree/bindings/net/mediatek,net.yaml +36 -0
@@ -42,6 +42,18 @@ properties: minItems: 1 maxItems: 8 + interrupt-names: + minItems: 1 + items: + - const: fe0 + - const: fe1 + - const: fe2 + - const: fe3 + - const: pdma0 + - const: pdma1 + - const: pdma2 + - const: pdma3 + power-domains: maxItems: 1
@@ -135,6 +147,10 @@ allOf: minItems: 3 maxItems: 3 + interrupt-names: + minItems: 3 + maxItems: 3 + clocks: minItems: 4 maxItems: 4
@@ -166,6 +182,9 @@ allOf: interrupts: maxItems: 1 + interrupt-names: + maxItems: 1 + clocks: minItems: 2 maxItems: 2
@@ -192,6 +211,10 @@ allOf: minItems: 3 maxItems: 3 + interrupt-names: + minItems: 3 + maxItems: 3 + clocks: minItems: 11 maxItems: 11
@@ -232,6 +255,10 @@ allOf: minItems: 3 maxItems: 3 + interrupt-names: + minItems: 3 + maxItems: 3 + clocks: minItems: 17 maxItems: 17
@@ -274,6 +301,9 @@ allOf: interrupts: minItems: 4 + interrupt-names: + minItems: 4 + clocks: minItems: 15 maxItems: 15
@@ -312,6 +342,9 @@ allOf: interrupts: minItems: 4 + interrupt-names: + minItems: 4 + clocks: minItems: 15 maxItems: 15
@@ -350,6 +383,9 @@ allOf: interrupts: minItems: 4 + interrupt-names: + minItems: 4 + clocks: minItems: 24 maxItems: 24
Signature · 1 line
2.43.0
PATCH v8 1/16 dt-bindings: net: mediatek,net: update mac subnode pattern for mt7988
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
MT7888 have 3 Macs and so its nodes have names from mac0 - mac2. Update pattern to fix this.
--- Documentation/devicetree/bindings/net/mediatek,net.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Documentation/devicetree/bindings/net/mediatek,net.yaml +1 -1
@@ -382,7 +382,7 @@ allOf: - const: xgp3 patternProperties: - "^mac@[0-1]$": + "^mac@[0-2]$": type: object unevaluatedProperties: false allOf:
Signature · 1 line
2.43.0
PATCH v8 7/16 dt-bindings: interconnect: add mt7988-cci compatible
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Add compatible for Mediatek MT7988 SoC with mediatek,mt8183-cci fallback which is taken by driver.
---
v2:
- no RFC
- drop "items" as sugested by conor
--- .../bindings/interconnect/mediatek,cci.yaml | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-)
Documentation/devicetree/bindings/interconnect/mediatek,cci.yaml +8 -3
@@ -17,9 +17,14 @@ description: | properties: compatible: - enum: - - mediatek,mt8183-cci - - mediatek,mt8186-cci + oneOf: + - enum: + - mediatek,mt8183-cci + - mediatek,mt8186-cci + - items: + - enum: + - mediatek,mt7988-cci + - const: mediatek,mt8183-cci clocks: items:
Signature · 1 line
2.43.0
PATCH v8 5/16 dt-bindings: net: dsa: mediatek,mt7530: add dsa-port definition for mt7988
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Add own dsa-port binding for SoC with internal switch where only phy-mode 'internal' is valid.
--- .../bindings/net/dsa/mediatek,mt7530.yaml | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-)
Documentation/devicetree/bindings/net/dsa/mediatek,mt7530.yaml +13 -1
@@ -190,6 +190,18 @@ required: - reg $defs: + builtin-dsa-port: + patternProperties: + "^(ethernet-)?ports$": + patternProperties: + "^(ethernet-)?port@[0-6]$": + if: + required: [ ethernet ] + then: + properties: + phy-mode: + const: internal + mt7530-dsa-port: patternProperties: "^(ethernet-)?ports$":
@@ -297,7 +309,7 @@ allOf: - airoha,en7581-switch - airoha,an7583-switch then: - $ref: "#/$defs/mt7530-dsa-port" + $ref: "#/$defs/builtin-dsa-port" properties: gpio-controller: false mediatek,mcm: false
Signature · 1 line
2.43.0
PATCH v8 9/16 arm64: dts: mediatek: mt7988: add basic ethernet-nodes
In reply to Frank WunderlichFrom: Frank Wunderlich <[email protected]>
Add basic ethernet related nodes.
Mac1+2 needs pcs (sgmii+usxgmii) to work correctly which will be linked later when driver is merged.
---
v8:
- change ethernet register size to 0x40000
range from 0x15140000 ~ 0x1517ffff is not usable on mt7988 => 0xDEADBEEF
v6:
- fix whitespace-errors for pdma irqs (spaces vs. tabs)
- move sram from eth reg to own sram node (needs CONFIG_SRAM)
v5:
- add reserved irqs and change names to fe0..fe3
- change rx-ringX to pdmaX to be closer to documentation
v4:
- comment for fixed-link on gmac0
- update 2g5 phy node
- unit-name dec instead of hex to match reg property
- move compatible before reg
- drop phy-mode
- add interrupts for RSS
- add interrupt-names and drop reserved irqs for ethernet
- some reordering
- eth-reg and clock whitespace-fix based on angelos review
--- arch/arm64/boot/dts/mediatek/mt7988a.dtsi | 137 +++++++++++++++++++++- 1 file changed, 134 insertions(+), 3 deletions(-)
arch/arm64/boot/dts/mediatek/mt7988a.dtsi +134 -3
@@ -680,7 +680,28 @@ xphyu3port0: usb-phy@11e13000 { }; }; - clock-controller@11f40000 { + xfi_tphy0: phy@11f20000 { + compatible = "mediatek,mt7988-xfi-tphy"; + reg = <0 0x11f20000 0 0x10000>; + clocks = <&xfi_pll CLK_XFIPLL_PLL_EN>, + <&topckgen CLK_TOP_XFI_PHY_0_XTAL_SEL>; + clock-names = "xfipll", "topxtal"; + resets = <&watchdog 14>; + mediatek,usxgmii-performance-errata; + #phy-cells = <0>; + }; + + xfi_tphy1: phy@11f30000 { + compatible = "mediatek,mt7988-xfi-tphy"; + reg = <0 0x11f30000 0 0x10000>; + clocks = <&xfi_pll CLK_XFIPLL_PLL_EN>, + <&topckgen CLK_TOP_XFI_PHY_1_XTAL_SEL>; + clock-names = "xfipll", "topxtal"; + resets = <&watchdog 15>; + #phy-cells = <0>; + }; + + xfi_pll: clock-controller@11f40000 { compatible = "mediatek,mt7988-xfi-pll"; reg = <0 0x11f40000 0 0x1000>; resets = <&watchdog 16>;
@@ -714,19 +735,129 @@ phy_calibration_p3: calib@97c { }; }; - clock-controller@15000000 { + ethsys: clock-controller@15000000 { compatible = "mediatek,mt7988-ethsys", "syscon"; reg = <0 0x15000000 0 0x1000>; #clock-cells = <1>; #reset-cells = <1>; }; - clock-controller@15031000 { + ethwarp: clock-controller@15031000 { compatible = "mediatek,mt7988-ethwarp"; reg = <0 0x15031000 0 0x1000>; #clock-cells = <1>; #reset-cells = <1>; }; + + eth: ethernet@15100000 { + compatible = "mediatek,mt7988-eth"; + reg = <0 0x15100000 0 0x40000>; + interrupts = <GIC_SPI 196 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 197 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 198 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 199 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 189 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 190 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 191 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 192 IRQ_TYPE_LEVEL_HIGH>; + interrupt-names = "fe0", "fe1", "fe2", "fe3", "pdma0", + "pdma1", "pdma2", "pdma3"; + clocks = <ðsys CLK_ETHDMA_CRYPT0_EN>, + <ðsys CLK_ETHDMA_FE_EN>, + <ðsys CLK_ETHDMA_GP2_EN>, + <ðsys CLK_ETHDMA_GP1_EN>, + <ðsys CLK_ETHDMA_GP3_EN>, + <ðwarp CLK_ETHWARP_WOCPU2_EN>, + <ðwarp CLK_ETHWARP_WOCPU1_EN>, + <ðwarp CLK_ETHWARP_WOCPU0_EN>, + <ðsys CLK_ETHDMA_ESW_EN>, + <&topckgen CLK_TOP_ETH_GMII_SEL>, + <&topckgen CLK_TOP_ETH_REFCK_50M_SEL>, + <&topckgen CLK_TOP_ETH_SYS_200M_SEL>, + <&topckgen CLK_TOP_ETH_SYS_SEL>, + <&topckgen CLK_TOP_ETH_XGMII_SEL>, + <&topckgen CLK_TOP_ETH_MII_SEL>, + <&topckgen CLK_TOP_NETSYS_SEL>, + <&topckgen CLK_TOP_NETSYS_500M_SEL>, + <&topckgen CLK_TOP_NETSYS_PAO_2X_SEL>, + <&topckgen CLK_TOP_NETSYS_SYNC_250M_SEL>, + <&topckgen CLK_TOP_NETSYS_PPEFB_250M_SEL>, + <&topckgen CLK_TOP_NETSYS_WARP_SEL>, + <ðsys CLK_ETHDMA_XGP1_EN>, + <ðsys CLK_ETHDMA_XGP2_EN>, + <ðsys CLK_ETHDMA_XGP3_EN>; + clock-names = "crypto", "fe", "gp2", "gp1", "gp3", + "ethwarp_wocpu2", "ethwarp_wocpu1", + "ethwarp_wocpu0", "esw", "top_eth_gmii_sel", + "top_eth_refck_50m_sel", "top_eth_sys_200m_sel", + "top_eth_sys_sel", "top_eth_xgmii_sel", + "top_eth_mii_sel", "top_netsys_sel", + "top_netsys_500m_sel", "top_netsys_pao_2x_sel", + "top_netsys_sync_250m_sel", + "top_netsys_ppefb_250m_sel", + "top_netsys_warp_sel","xgp1", "xgp2", "xgp3"; + assigned-clocks = <&topckgen CLK_TOP_NETSYS_2X_SEL>, + <&topckgen CLK_TOP_NETSYS_GSW_SEL>, + <&topckgen CLK_TOP_USXGMII_SBUS_0_SEL>, + <&topckgen CLK_TOP_USXGMII_SBUS_1_SEL>, + <&topckgen CLK_TOP_SGM_0_SEL>, + <&topckgen CLK_TOP_SGM_1_SEL>; + assigned-clock-parents = <&apmixedsys CLK_APMIXED_NET2PLL>, + <&topckgen CLK_TOP_NET1PLL_D4>, + <&topckgen CLK_TOP_NET1PLL_D8_D4>, + <&topckgen CLK_TOP_NET1PLL_D8_D4>, + <&apmixedsys CLK_APMIXED_SGMPLL>, + <&apmixedsys CLK_APMIXED_SGMPLL>; + sram = <ð_sram>; + #address-cells = <1>; + #size-cells = <0>; + mediatek,ethsys = <ðsys>; + mediatek,infracfg = <&topmisc>; + + gmac0: mac@0 { + compatible = "mediatek,eth-mac"; + reg = <0>; + phy-mode = "internal"; + + /* Connected to internal switch */ + fixed-link { + speed = <10000>; + full-duplex; + pause; + }; + }; + + gmac1: mac@1 { + compatible = "mediatek,eth-mac"; + reg = <1>; + status = "disabled"; + }; + + gmac2: mac@2 { + compatible = "mediatek,eth-mac"; + reg = <2>; + status = "disabled"; + }; + + mdio_bus: mdio-bus { + #address-cells = <1>; + #size-cells = <0>; + + /* internal 2.5G PHY */ + int_2p5g_phy: ethernet-phy@15 { + compatible = "ethernet-phy-ieee802.3-c45"; + reg = <15>; + }; + }; + }; + + eth_sram: sram@15400000 { + compatible = "mmio-sram"; + reg = <0 0x15400000 0 0x200000>; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0 0x15400000 0 0x200000>; + }; }; thermal-zones {
Signature · 1 line
2.43.0
Re: [PATCH v8 01/16] dt-bindings: net: mediatek,net: update mac subnode pattern for mt7988
In reply to Frank WunderlichQuoted 10 lines · Frank Wunderlich
From: Frank Wunderlich <[email protected]>
MT7888 have 3 Macs and so its nodes have names from mac0 - mac2. Update pattern to fix this.
Fixes c94a9aabec36 ("dt-bindings: net: mediatek,net: add mt7988-eth binding")Signed-off-by Frank Wunderlich <[email protected]>--- Documentation/devicetree/bindings/net/mediatek,net.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Best regards,
Krzysztof
Re: [PATCH v8 02/16] dt-bindings: net: mediatek,net: allow up to 8 IRQs
In reply to Frank WunderlichQuoted 3 lines · Frank Wunderlich
From: Frank Wunderlich <[email protected]>
Increase the maximum IRQ count to 8 (4 FE + 4 RSS/LRO).
Because? Hardware was updated? It was missing before?
Quoted 5 lines
Frame-engine-IRQs (max 4):
MT7621, MT7628: 1 IRQ
MT7622, MT7623: 3 IRQs (only two used by the driver for now) MT7981, MT7986, MT7988: 4 IRQs (only two used by the driver for now)
You updated commit msg - looks fine - but same problem as before in your code. Now MT7981 has 4-8 interrupts, even though you say here it has only 4.
Quoted 3 lines
Mediatek Filogic SoCs (mt798x) have 4 additional IRQs for RSS and/or LRO.
Although I don't know how to treat this. Just say how many interrupts are there (MT7981, MT7986, MT7988: 4 FE and 4 RSS), not 4 but later actually 4+4.
I also do not understand why 7 interrupts is now valid... Are these not connected physically?
Best regards,
Krzysztof
Re: [PATCH v8 04/16] dt-bindings: net: mediatek,net: add sram property
In reply to Frank WunderlichQuoted 3 lines · Frank Wunderlich
From: Frank Wunderlich <[email protected]>
Meditak Filogic SoCs (MT798x) have dedicated MMIO-SRAM for dma operations.
You need "sram: false" for other variants.
Best regards,
Krzysztof
Re: [PATCH v8 02/16] dt-bindings: net: mediatek,net: allow up to 8 IRQs
In reply to Krzysztof KozlowskiAm 7. Juli 2025 08:31:11 MESZ schrieb Krzysztof Kozlowski <[email protected]>:
Quoted 6 lines · Frank Wunderlich
Quoted 3 lines
From: Frank Wunderlich <[email protected]>
Increase the maximum IRQ count to 8 (4 FE + 4 RSS/LRO).
Because? Hardware was updated? It was missing before?
There is no RSS support in driver yet,so IRQs were not added to existing DTS yet.
Quoted 9 lines
Quoted 5 lines
Frame-engine-IRQs (max 4):
MT7621, MT7628: 1 IRQ
MT7622, MT7623: 3 IRQs (only two used by the driver for now) MT7981, MT7986, MT7988: 4 IRQs (only two used by the driver for now)You updated commit msg - looks fine - but same problem as before in your code. Now MT7981 has 4-8 interrupts, even though you say here it has only 4.
Ethernet works with 4,but can be 8 for MT798x.
I cannot increase the MinItems here as it will
throw error because currently only 4 are defined in DTS.same for MT7986.
Quoted 7 lines
Quoted 3 lines
Mediatek Filogic SoCs (mt798x) have 4 additional IRQs for RSS and/or LRO.
Although I don't know how to treat this. Just say how many interrupts are there (MT7981, MT7986, MT7988: 4 FE and 4 RSS), not 4 but later actually 4+4.
First block is for Frame Engine IRQs and second for RSS/LRO. Only mention total count across all SoCs is imho more confusing.
Quoted 2 lines
I also do not understand why 7 interrupts is now valid... Are these not connected physically?
7 does not make sense but i know no way to allow 8 with min 4 without between (5-7).
Quoted 2 lines
Best regards,
Krzysztof
Hi
Thanks for taking time for review again.
First block are the frame engine IRQs which are max 4 and on all SoCs. The RSS IRQs are only valid on Filogic (MT798x),so there a total of 8, but on MT7981 and MT7986 not yet added as i prepare the RSS/LRO driver in background. We just want to add the IRQs for MT7988 now.
regards Frank
Re: (subset) [PATCH v8 00/16] further mt7988 devicetree work
In reply to Frank WunderlichQuoted 9 lines · Frank Wunderlich
From: Frank Wunderlich <[email protected]>
This series continues mt7988 devicetree work
- Extend cpu frequency scaling with CCI
- GPIO leds
- Basic network-support (ethernet controller + builtin switch + SFP Cages)[...]
Applied to v6.16-next/dts64, thanks!
[07/16] dt-bindings: interconnect: add mt7988-cci compatible
commit: bd9e0f5d90959d2d07986084fbd58042b62aa549
[08/16] arm64: dts: mediatek: mt7988: add cci node
commit: 0cbdb6d04689f8c05074e348c8e0a42b229ef9a3
[11/16] arm64: dts: mediatek: mt7988a-bpi-r4: add proc-supply for cci
commit: b5a4ad957114b59a74b3e3f598ae0785dd86cd32
[12/16] arm64: dts: mediatek: mt7988a-bpi-r4: drop unused pins
commit: bc51660cd5fd2d0ee9a65b59e0c65e2d1b65975a
[13/16] arm64: dts: mediatek: mt7988a-bpi-r4: add gpio leds
commit: 5a40efb8c9d26e51db8acc61e920c3eda9407c02Cheers,
Angelo
AngeloGioacchino Del Regno
maintainerRe: [PATCH v8 02/16] dt-bindings: net: mediatek,net: allow up to 8 IRQs
In reply to Frank WunderlichIl 07/07/25 09:30, Frank Wunderlich ha scritto:
Quoted 10 lines
Am 7. Juli 2025 08:31:11 MESZ schrieb Krzysztof Kozlowski <[email protected]>:
Quoted 6 lines · Frank Wunderlich
Quoted 3 lines
From: Frank Wunderlich <[email protected]>
Increase the maximum IRQ count to 8 (4 FE + 4 RSS/LRO).
Because? Hardware was updated? It was missing before?
There is no RSS support in driver yet,so IRQs were not added to existing DTS yet.
That's the problem. It's the hardware that you should've described, not the driver.
In short, you should've allowed the interrupts from the get-go, and you wouldn't be in this situation now :-)
Quoted 34 lines
Quoted 9 lines
Quoted 5 lines
Frame-engine-IRQs (max 4):
MT7621, MT7628: 1 IRQ
MT7622, MT7623: 3 IRQs (only two used by the driver for now) MT7981, MT7986, MT7988: 4 IRQs (only two used by the driver for now)You updated commit msg - looks fine - but same problem as before in your code. Now MT7981 has 4-8 interrupts, even though you say here it has only 4.
Ethernet works with 4,but can be 8 for MT798x.
I cannot increase the MinItems here as it will
throw error because currently only 4 are defined in DTS.same for MT7986.Quoted 7 lines
Quoted 3 lines
Mediatek Filogic SoCs (mt798x) have 4 additional IRQs for RSS and/or LRO.
Although I don't know how to treat this. Just say how many interrupts are there (MT7981, MT7986, MT7988: 4 FE and 4 RSS), not 4 but later actually 4+4.
First block is for Frame Engine IRQs and second for RSS/LRO. Only mention total count across all SoCs is imho more confusing.
Quoted 2 lines
I also do not understand why 7 interrupts is now valid... Are these not connected physically?
7 does not make sense but i know no way to allow 8 with min 4 without between (5-7).
Quoted 2 lines
Best regards,
KrzysztofHi
Quoted 7 lines
Thanks for taking time for review again.
First block are the frame engine IRQs which are max 4 and on all SoCs. The RSS IRQs are only valid on Filogic (MT798x),so there a total of 8, but on MT7981 and MT7986 not yet added as i prepare the RSS/LRO driver in background. We just want to add the IRQs for MT7988 now.
regards Frank
Again, it's not the driver but the hardware that you're describing.
As long as you are fixing the description of the hardware, even for all three, I am personally even fine with breaking the ABI, because the hardware description has been wrong for all that time.
Just don't send those as Fixes commits, but next time you upstream something you must keep in mind that in bindings/dts you're describing hardware - the driver is something that should not drive any decision in what you write in bindings.
We're humans, so stuff like this happens - I'm not saying that you shall not make mistakes - but again please, for the next time, please please please keep in mind what I just said :-)
Now the options are two:
- Break the ABI; or
- Allow 4 or 8 interrupts (not 5, not 6, not 7)
and that - not just on MT7988 but also on 81 and 86 in one go.
Not sure if the second one is feasible, and I'm considering the first option only because of that; if the second option can be done, act like I never ever considered the first.
Cheers,
Angelo
Re: [PATCH v8 03/16] dt-bindings: net: mediatek,net: allow irq names
In reply to Frank WunderlichIl 06/07/25 15:21, Frank Wunderlich ha scritto:
Quoted 13 lines
From: Frank Wunderlich <[email protected]>
In preparation for MT7988 and RSS/LRO allow the interrupt-names property.
In this way driver can request the interrupts by name which is much more readable in the driver code and SoC's dtsi than relying on a specific order.Frame-engine-IRQs (fe0..3):
MT7621, MT7628: 1 IRQ
MT7622, MT7623: 3 IRQs (only two used by the driver for now) MT7981, MT7986: 4 IRQs (only two used by the driver for now)
Quoted 5 lines
RSS/LRO IRQs (pdma0..3) only on Filogic (MT798x) with count of 4.
Set boundaries for all compatibles same as irq count.
Signed-off-by Frank Wunderlich <[email protected]>
I'm fine with that, as long as you don't break the driver's ability to keep getting the interrupts without any interrupt-names, at least for the currently supported interrupts.
This commit ain't touching any driver, anyway, and the change is valid IMO, so:
Quoted 107 lines
---
v8:
- fixed typo in mt7621 section "interrupt-namess"
- separated interrupt count from interrupt-names
- rephrased description a bit to explain the "why"
v7: fixed wrong rebase
v6: new patch splitted from the mt7988 changes--- .../devicetree/bindings/net/mediatek,net.yaml | 36 +++++++++++++++++++ 1 file changed, 36 insertions(+)Documentation/devicetree/bindings/net/mediatek,net.yaml +36
-0@@ -42,6 +42,18 @@ properties: minItems: 1 maxItems: 8 + interrupt-names: + minItems: 1 + items: + - const: fe0 + - const: fe1 + - const: fe2 + - const: fe3 + - const: pdma0 + - const: pdma1 + - const: pdma2 + - const: pdma3 + power-domains: maxItems: 1@@ -135,6 +147,10 @@ allOf: minItems: 3 maxItems: 3 + interrupt-names: + minItems: 3 + maxItems: 3 + clocks: minItems: 4 maxItems: 4@@ -166,6 +182,9 @@ allOf: interrupts: maxItems: 1 + interrupt-names: + maxItems: 1 + clocks: minItems: 2 maxItems: 2@@ -192,6 +211,10 @@ allOf: minItems: 3 maxItems: 3 + interrupt-names: + minItems: 3 + maxItems: 3 + clocks: minItems: 11 maxItems: 11@@ -232,6 +255,10 @@ allOf: minItems: 3 maxItems: 3 + interrupt-names: + minItems: 3 + maxItems: 3 + clocks: minItems: 17 maxItems: 17@@ -274,6 +301,9 @@ allOf: interrupts: minItems: 4 + interrupt-names: + minItems: 4 + clocks: minItems: 15 maxItems: 15@@ -312,6 +342,9 @@ allOf: interrupts: minItems: 4 + interrupt-names: + minItems: 4 + clocks: minItems: 15 maxItems: 15@@ -350,6 +383,9 @@ allOf: interrupts: minItems: 4 + interrupt-names: + minItems: 4 + clocks: minItems: 24 maxItems: 24
Re: [PATCH v8 04/16] dt-bindings: net: mediatek,net: add sram property
In reply to Frank WunderlichIl 06/07/25 15:21, Frank Wunderlich ha scritto:
Quoted 10 lines
From: Frank Wunderlich <[email protected]>
Meditak Filogic SoCs (MT798x) have dedicated MMIO-SRAM for dma operations.
MT7981 and MT7986 currently use static offset to ethernet MAC register which will be changed in separate patch once this way is accepted.
Add "sram" property to map ethernet controller to dedicated mmio-sram node.
Signed-off-by Frank Wunderlich <[email protected]>
Honestly, I was more comfortable adding the sram to the reg list instead, but I know that you got a negative review for that. Oh well.
Please disallow the sram property on SoCs that don't have the SRAM (sram: false), after which:
Re: [PATCH v8 02/16] dt-bindings: net: mediatek,net: allow up to 8 IRQs
In reply to AngeloGioacchino Del RegnoHi Angelo,
Am 7. Juli 2025 12:06:02 MESZ schrieb AngeloGioacchino Del Regno <[email protected]>:
Quoted 9 lines
Il 07/07/25 09:30, Frank Wunderlich ha scritto:
Quoted 8 lines
Am 7. Juli 2025 08:31:11 MESZ schrieb Krzysztof Kozlowski <[email protected]>:
Quoted 6 lines · Frank Wunderlich
Quoted 3 lines
From: Frank Wunderlich <[email protected]>
Increase the maximum IRQ count to 8 (4 FE + 4 RSS/LRO).
Because? Hardware was updated? It was missing before?
Quoted 7 lines
Quoted 2 lines
There is no RSS support in driver yet,so IRQs were not added to existing DTS yet.
That's the problem. It's the hardware that you should've described, not the driver.
In short, you should've allowed the interrupts from the get-go, and you wouldn't be in this situation now :-)
I have not upstreamed MT7981 or MT7986. I also do not want to say anybody else did this wrong.
I'm happy that MT7986 is working in mainline. It was basicly not taken into account that these IRQs may be needed in future.
The technical documents are often not complete and we get some information step-by-step while testing. Or it was not seen when documents are too large :) many reasons why it was "forgotten to add".
We use what we get from sdk and docs and try to make it compatible with mainline....no optimal process,but it is like it is.
We are all humans :)
Quoted 1 line
Again, it's not the driver but the hardware that you're describing.
Frame engine works with 4,but we wanted to do better than mt798[16] and add all known IRQs for MT7988 and update the older SoCs in next step.
Quoted 3 lines
As long as you are fixing the description of the hardware, even for all three, I am personally even fine with breaking the ABI, because the hardware description has been wrong for all that time.
As some patches now are applied i can add the missing IRQs for MT798[16] in next round and maybe the sram too to not having the binding errors long time.
Quoted 8 lines
Just don't send those as Fixes commits, but next time you upstream something you must keep in mind that in bindings/dts you're describing hardware - the driver is something that should not drive any decision in what you write in bindings.
We're humans, so stuff like this happens - I'm not saying that you shall not make mistakes - but again please, for the next time, please please please keep in mind what I just said :-)
Quoted 5 lines
Now the options are two:
- Break the ABI; or
- Allow 4 or 8 interrupts (not 5, not 6, not 7)and that - not just on MT7988 but also on 81 and 86 in one go.
I would add dts patches for MT798[16] in next version adding the RSS/LRO IRQs, interrupt-names and the sram node and set the interrupt minItems to 8 for these.
Quoted 3 lines
Not sure if the second one is feasible, and I'm considering the first option only because of that; if the second option can be done, act like I never ever considered the first.
Maybe the cleanest way is like i described above? Afaik second is not possible.
My intention is that RSS/LRO is ONLY working with interrupt-names so driver can handle also old dtb (skip function if irqs are not found via name).
I hope this is a good way to go...
Quoted 2 lines
Cheers,
Angelo
regards Frank
Re: [PATCH v8 02/16] dt-bindings: net: mediatek,net: allow up to 8 IRQs
Quoted 20 lines · Frank Wunderlich
Hi Angelo,
Am 7. Juli 2025 12:06:02 MESZ schrieb AngeloGioacchino Del Regno <[email protected]>:
Quoted 9 lines
Il 07/07/25 09:30, Frank Wunderlich ha scritto:
Quoted 8 lines
Am 7. Juli 2025 08:31:11 MESZ schrieb Krzysztof Kozlowski <[email protected]>:
Quoted 6 lines · Frank Wunderlich
Quoted 3 lines
From: Frank Wunderlich <[email protected]>
Increase the maximum IRQ count to 8 (4 FE + 4 RSS/LRO).
Because? Hardware was updated? It was missing before?
Quoted 7 lines
Quoted 2 lines
There is no RSS support in driver yet,so IRQs were not added to existing DTS yet.
That's the problem. It's the hardware that you should've described, not the driver.
In short, you should've allowed the interrupts from the get-go, and you wouldn't be in this situation now :-)
Quoted 6 lines
I have not upstreamed MT7981 or MT7986. I also do not want to say anybody else did this wrong.
I'm happy that MT7986 is working in mainline. It was basicly not taken into account that these IRQs may be needed in future.The technical documents are often not complete and we get some information step-by-step while testing. Or it was not seen when documents are too large :) many reasons why it was "forgotten to add".
We use what we get from sdk and docs and try to make it compatible with mainline....no optimal process,but it is like it is.
Then explain in the commit msg that hardware description was incomplete and was missing this and that.
This is the valid reason for doing the change.
Best regards,
Krzysztof