[PATCH v1 0/7] add Ethernet support for S700
This series adds Ethernet Support Actions Semi OWL S700 SoC.
First patch(1/7) series adds Ethernet clock for S900 and S700, S900 part is just compiled tested.
Patches(from 2/7 and 3/7) adds support for RTL 8201F PHY module and introduce configuration option "RTL8201F_PHY_S700_RMII_TIMINGS" to fulfil specific timing requirements for S700.
Patches(from 4/21 to 7/21) are there to enable Ethenet support in S700, MAC is based on Designware IP.
These patches re-uses the existing driver(drivers/net/designware.c) and programs SoC specific bits to enable ethernet. SoC specific glue code is kept in dwmac_s700.c file, did it this way as found it more cleaner but having done that I am not really sure, if it's bit of a overkill to have it place in glue code or we can keep this glue code somewhere in machine file ?
Series is tested on Cubieboard7-lite by tftping all the images needed to boot Linux.
Amit Singh Tomar (7): clk: actions: Add Ethernet clocks net: phy: realtek: Add support for RTL8201F PHY module. net: phy: realtek: Introduce PHY_RTL8201F_S700_RMII_TIMINGS to adjust rx/tx timings net: designware: s700: Add glue code for S700 mac arm: dts: s700: add node for ethernet controller owl: Kconfig: Enable DM eth for OWL platform configs: Enable mac and phy configs
arch/arm/Kconfig | 1 + arch/arm/dts/s700-u-boot.dtsi | 13 ++++++ arch/arm/include/asm/arch-owl/regs_s700.h | 8 ++++ arch/arm/include/asm/arch-owl/regs_s900.h | 4 ++ configs/cubieboard7_defconfig | 4 ++ drivers/clk/owl/clk_owl.c | 9 +++++ drivers/clk/owl/clk_owl.h | 2 - drivers/net/Kconfig | 7 ++++ drivers/net/Makefile | 1 + drivers/net/dwmac_s700.c | 66 +++++++++++++++++++++++++++++++ drivers/net/phy/Kconfig | 9 +++++ drivers/net/phy/realtek.c | 55 ++++++++++++++++++++++++++ 12 files changed, 177 insertions(+), 2 deletions(-) create mode 100644 drivers/net/dwmac_s700.c
This commit adds clocks needed for ethernet operations for Actions OWL family of SoCs (S700 and S900).
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com --- arch/arm/include/asm/arch-owl/regs_s700.h | 2 ++ arch/arm/include/asm/arch-owl/regs_s900.h | 4 ++++ drivers/clk/owl/clk_owl.c | 9 +++++++++ drivers/clk/owl/clk_owl.h | 2 -- 4 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/arch/arm/include/asm/arch-owl/regs_s700.h b/arch/arm/include/asm/arch-owl/regs_s700.h index 2f21c15cca8c..90459ae95eeb 100644 --- a/arch/arm/include/asm/arch-owl/regs_s700.h +++ b/arch/arm/include/asm/arch-owl/regs_s700.h @@ -53,4 +53,6 @@ #define CMU_CVBSPLL 0x00B8 #define CMU_SSTSCLK 0x00C0
+#define CMU_DEVCLKEN1_ETH BIT(23) + #endif diff --git a/arch/arm/include/asm/arch-owl/regs_s900.h b/arch/arm/include/asm/arch-owl/regs_s900.h index 9e9106ddaab9..084bc9b8c3fb 100644 --- a/arch/arm/include/asm/arch-owl/regs_s900.h +++ b/arch/arm/include/asm/arch-owl/regs_s900.h @@ -61,4 +61,8 @@ #define CMU_TVOUTPLLDEBUG0 (0x00EC) #define CMU_TVOUTPLLDEBUG1 (0x00FC)
+#define CMU_DEVCLKEN1_ETH BIT(22) +#define CLK_ETHERNET CLK_ETH_MAC +#define CMU_ETHERNETPLL CMU_ASSISTPLL + #endif diff --git a/drivers/clk/owl/clk_owl.c b/drivers/clk/owl/clk_owl.c index 5607b2b7b516..b80862e17f51 100644 --- a/drivers/clk/owl/clk_owl.c +++ b/drivers/clk/owl/clk_owl.c @@ -85,6 +85,11 @@ int owl_clk_enable(struct clk *clk) /* Enable UART3 interface clock */ setbits_le32(priv->base + CMU_DEVCLKEN1, CMU_DEVCLKEN1_UART3); break; + case CLK_RMII_REF: + case CLK_ETHERNET: + setbits_le32(priv->base + CMU_DEVCLKEN1, CMU_DEVCLKEN1_ETH); + setbits_le32(priv->base + CMU_ETHERNETPLL, 5); + break; default: return -EINVAL; } @@ -110,6 +115,10 @@ int owl_clk_disable(struct clk *clk) /* Disable UART3 interface clock */ clrbits_le32(priv->base + CMU_DEVCLKEN1, CMU_DEVCLKEN1_UART3); break; + case CLK_RMII_REF: + case CLK_ETHERNET: + clrbits_le32(priv->base + CMU_DEVCLKEN1, CMU_DEVCLKEN1_ETH); + break; default: return -EINVAL; } diff --git a/drivers/clk/owl/clk_owl.h b/drivers/clk/owl/clk_owl.h index b8d33624c5f5..a2065a726d46 100644 --- a/drivers/clk/owl/clk_owl.h +++ b/drivers/clk/owl/clk_owl.h @@ -59,6 +59,4 @@ struct owl_clk_priv { #define CMU_DEVCLKEN1_UART5 BIT(21) #define CMU_DEVCLKEN1_UART3 BIT(11)
-#define CMU_DEVCLKEN1_ETH_S700 BIT(23) - #endif
On Sat, May 09, 2020 at 07:55:09PM +0530, Amit Singh Tomar wrote:
This commit adds clocks needed for ethernet operations for Actions OWL family of SoCs (S700 and S900).
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
Applied to u-boot/master, thanks!
This patch adds support for Realtek PHY RTL8201F 10/100Mbs (with variants: RTL8201FN and RTL8201FL) PHYceiver. It is present on Actions Semi Cubieboard7 board.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com --- drivers/net/phy/realtek.c | 20 ++++++++++++++++++++ 1 file changed, 20 insertions(+)
diff --git a/drivers/net/phy/realtek.c b/drivers/net/phy/realtek.c index 8f1d75963259..b9c373bfe3cc 100644 --- a/drivers/net/phy/realtek.c +++ b/drivers/net/phy/realtek.c @@ -157,6 +157,14 @@ static int rtl8211x_config(struct phy_device *phydev) return 0; }
+/* RealTek RTL8201F */ +static int rtl8201f_config(struct phy_device *phydev) +{ + genphy_config_aneg(phydev); + + return 0; +} + static int rtl8211f_config(struct phy_device *phydev) { u16 reg; @@ -386,12 +394,24 @@ static struct phy_driver RTL8211F_driver = { .writeext = &rtl8211f_phy_extwrite, };
+/* Support for RTL8201F PHY */ +static struct phy_driver RTL8201F_driver = { + .name = "RealTek RTL8201F 10/100Mbps Ethernet", + .uid = 0x1cc816, + .mask = 0xffffff, + .features = PHY_BASIC_FEATURES, + .config = &rtl8201f_config, + .startup = &rtl8211e_startup, + .shutdown = &genphy_shutdown, +}; + int phy_realtek_init(void) { phy_register(&RTL8211B_driver); phy_register(&RTL8211E_driver); phy_register(&RTL8211F_driver); phy_register(&RTL8211DN_driver); + phy_register(&RTL8201F_driver);
return 0; }
On Sat, May 09, 2020 at 07:55:10PM +0530, Amit Singh Tomar wrote:
This patch adds support for Realtek PHY RTL8201F 10/100Mbs (with variants: RTL8201FN and RTL8201FL) PHYceiver. It is present on Actions Semi Cubieboard7 board.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
Applied to u-boot/master, thanks!
RTL8201F PHY module found on Actions Semi Cubieboard7 seems to have specific Rx/Tx interface timings requirement for proper PHY operations. These timing values are not documented anywhere and picked from vendor code.
This commits lets proper packets to be transmitted over the network.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com --- drivers/net/phy/Kconfig | 9 +++++++++ drivers/net/phy/realtek.c | 35 +++++++++++++++++++++++++++++++++++ 2 files changed, 44 insertions(+)
diff --git a/drivers/net/phy/Kconfig b/drivers/net/phy/Kconfig index d1f049e62ab7..123c17c04458 100644 --- a/drivers/net/phy/Kconfig +++ b/drivers/net/phy/Kconfig @@ -235,6 +235,15 @@ config RTL8211F_PHY_FORCE_EEE_RXC_ON Default n, which means that the PHY state is not changed. To work around the issues, change this setting to y.
+config RTL8201F_PHY_S700_RMII_TIMINGS + bool "Ethernet PHY RTL8201F: adjust RMII Tx Interface timings" + depends on PHY_REALTEK + help + This provides an option to configure specific timing requirements (needed + for proper PHY operations) for the PHY module present on ACTION SEMI S700 + based cubieboard7. Exact timing requiremnets seems to be SoC specific + (and it's undocumented) that comes from vendor code itself. + config PHY_SMSC bool "Microchip(SMSC) Ethernet PHYs support"
diff --git a/drivers/net/phy/realtek.c b/drivers/net/phy/realtek.c index b9c373bfe3cc..a549a106876f 100644 --- a/drivers/net/phy/realtek.c +++ b/drivers/net/phy/realtek.c @@ -13,6 +13,7 @@ #define PHY_RTL8211x_FORCE_MASTER BIT(1) #define PHY_RTL8211E_PINE64_GIGABIT_FIX BIT(2) #define PHY_RTL8211F_FORCE_EEE_RXC_ON BIT(3) +#define PHY_RTL8201F_S700_RMII_TIMINGS BIT(4)
#define PHY_AUTONEGOTIATE_TIMEOUT 5000
@@ -58,6 +59,15 @@ #define MIIM_RTL8211F_TX_DELAY 0x100 #define MIIM_RTL8211F_LCR 0x10
+#define RTL8201F_RMSR 0x10 + +#define RMSR_RX_TIMING_SHIFT BIT(2) +#define RMSR_RX_TIMING_MASK GENMASK(7, 4) +#define RMSR_RX_TIMING_VAL 0x4 +#define RMSR_TX_TIMING_SHIFT BIT(3) +#define RMSR_TX_TIMING_MASK GENMASK(11, 8) +#define RMSR_TX_TIMING_VAL 0x5 + static int rtl8211f_phy_extread(struct phy_device *phydev, int addr, int devaddr, int regnum) { @@ -112,6 +122,15 @@ static int rtl8211f_probe(struct phy_device *phydev) return 0; }
+static int rtl8210f_probe(struct phy_device *phydev) +{ +#ifdef CONFIG_RTL8201F_PHY_S700_RMII_TIMINGS + phydev->flags |= PHY_RTL8201F_S700_RMII_TIMINGS; +#endif + + return 0; +} + /* RealTek RTL8211x */ static int rtl8211x_config(struct phy_device *phydev) { @@ -160,6 +179,21 @@ static int rtl8211x_config(struct phy_device *phydev) /* RealTek RTL8201F */ static int rtl8201f_config(struct phy_device *phydev) { + unsigned int reg; + + if (phydev->flags & PHY_RTL8201F_S700_RMII_TIMINGS) { + phy_write(phydev, MDIO_DEVAD_NONE, MIIM_RTL8211F_PAGE_SELECT, + 7); + reg = phy_read(phydev, MDIO_DEVAD_NONE, RTL8201F_RMSR); + reg &= ~(RMSR_RX_TIMING_MASK | RMSR_TX_TIMING_MASK); + /* Set the needed Rx/Tx Timings for proper PHY operation */ + reg |= (RMSR_RX_TIMING_VAL << RMSR_RX_TIMING_SHIFT) + | (RMSR_TX_TIMING_VAL << RMSR_TX_TIMING_SHIFT); + phy_write(phydev, MDIO_DEVAD_NONE, RTL8201F_RMSR, reg); + phy_write(phydev, MDIO_DEVAD_NONE, MIIM_RTL8211F_PAGE_SELECT, + 0); + } + genphy_config_aneg(phydev);
return 0; @@ -400,6 +434,7 @@ static struct phy_driver RTL8201F_driver = { .uid = 0x1cc816, .mask = 0xffffff, .features = PHY_BASIC_FEATURES, + .probe = &rtl8210f_probe, .config = &rtl8201f_config, .startup = &rtl8211e_startup, .shutdown = &genphy_shutdown,
On Sat, May 09, 2020 at 07:55:11PM +0530, Amit Singh Tomar wrote:
RTL8201F PHY module found on Actions Semi Cubieboard7 seems to have specific Rx/Tx interface timings requirement for proper PHY operations. These timing values are not documented anywhere and picked from vendor code.
This commits lets proper packets to be transmitted over the network.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
Applied to u-boot/master, thanks!
This patchs adds glue logic to enable designware mac present on Action Semi based S700 SoC, Configures SoC specific bits.
Undocumented bit that programs the PHY interface select register comes from vendor source.
It has been tested on Cubieboard7-lite based on S700 SoC.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com --- arch/arm/include/asm/arch-owl/regs_s700.h | 6 +++ drivers/net/Kconfig | 7 ++++ drivers/net/Makefile | 1 + drivers/net/dwmac_s700.c | 66 +++++++++++++++++++++++++++++++ 4 files changed, 80 insertions(+) create mode 100644 drivers/net/dwmac_s700.c
diff --git a/arch/arm/include/asm/arch-owl/regs_s700.h b/arch/arm/include/asm/arch-owl/regs_s700.h index 90459ae95eeb..0f79faec69c1 100644 --- a/arch/arm/include/asm/arch-owl/regs_s700.h +++ b/arch/arm/include/asm/arch-owl/regs_s700.h @@ -55,4 +55,10 @@
#define CMU_DEVCLKEN1_ETH BIT(23)
+#define GPIO_MFP_PWM (0xE01B0000) +#define MFP_CTL0 (GPIO_MFP_PWM + 0x40) +#define MFP_CTL1 (GPIO_MFP_PWM + 0x44) +#define MFP_CTL2 (GPIO_MFP_PWM + 0x48) +#define MFP_CTL3 (GPIO_MFP_PWM + 0x4C) + #endif diff --git a/drivers/net/Kconfig b/drivers/net/Kconfig index a2587a29e165..1a04e333efdf 100644 --- a/drivers/net/Kconfig +++ b/drivers/net/Kconfig @@ -225,6 +225,13 @@ config ETH_DESIGNWARE_SOCFPGA Altera system manager to correctly interface with the PHY. This code handles those SoC specifics.
+config ETH_DESIGNWARE_S700 + bool "Actins S700 glue driver for Synopsys Designware Ethernet MAC" + depends on DM_ETH && ETH_DESIGNWARE + help + This provides glue layer to use Synopsys Designware Ethernet MAC + present on Actions S700 SoC. + config ETHOC bool "OpenCores 10/100 Mbps Ethernet MAC" help diff --git a/drivers/net/Makefile b/drivers/net/Makefile index 6d9b8772b1a5..6f01d0169240 100644 --- a/drivers/net/Makefile +++ b/drivers/net/Makefile @@ -18,6 +18,7 @@ obj-$(CONFIG_CS8900) += cs8900.o obj-$(CONFIG_TULIP) += dc2114x.o obj-$(CONFIG_ETH_DESIGNWARE) += designware.o obj-$(CONFIG_ETH_DESIGNWARE_SOCFPGA) += dwmac_socfpga.o +obj-$(CONFIG_ETH_DESIGNWARE_S700) += dwmac_s700.o obj-$(CONFIG_DRIVER_DM9000) += dm9000x.o obj-$(CONFIG_DNET) += dnet.o obj-$(CONFIG_E1000) += e1000.o diff --git a/drivers/net/dwmac_s700.c b/drivers/net/dwmac_s700.c new file mode 100644 index 000000000000..a5d544e91e06 --- /dev/null +++ b/drivers/net/dwmac_s700.c @@ -0,0 +1,66 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * Copyright (C) 2020 Amit Singh Tomar amittomer25@gmail.com + * + * Actions DWMAC specific glue layer + */ + +#include <common.h> +#include <asm/io.h> +#include <dm.h> +#include <clk.h> +#include <phy.h> +#include <regmap.h> +#include <reset.h> +#include <syscon.h> +#include "designware.h" +#include <asm/arch-owl/regs_s700.h> + +/* pin control for MAC */ +#define RMII_TXD01_MFP_CTL0 (0x0 << 16) +#define RMII_RXD01_MFP_CTL0 (0x0 << 8) +#define RMII_TXEN_TXER_MFP_CTL0 (0x0 << 13) +#define RMII_REF_CLK_MFP_CTL0 (0x0 << 6) +#define CLKO_25M_EN_MFP_CTL3 BIT(30) + +DECLARE_GLOBAL_DATA_PTR; + +static void dwmac_board_setup(void) +{ + clrbits_le32(MFP_CTL0, (RMII_TXD01_MFP_CTL0 | RMII_RXD01_MFP_CTL0 | + RMII_TXEN_TXER_MFP_CTL0 | RMII_REF_CLK_MFP_CTL0)); + + setbits_le32(MFP_CTL3, CLKO_25M_EN_MFP_CTL3); +} + +static int dwmac_s700_probe(struct udevice *dev) +{ + dwmac_board_setup(); + + /* This is undocumented, phy interface select register */ + writel(0x4, 0xe024c0a0); + + return designware_eth_probe(dev); +} + +static int dwmac_s700_ofdata_to_platdata(struct udevice *dev) +{ + return designware_eth_ofdata_to_platdata(dev); +} + +static const struct udevice_id dwmac_s700_ids[] = { + {.compatible = "actions,s700-ethernet"}, + { } +}; + +U_BOOT_DRIVER(dwmac_s700) = { + .name = "dwmac_s700", + .id = UCLASS_ETH, + .of_match = dwmac_s700_ids, + .ofdata_to_platdata = dwmac_s700_ofdata_to_platdata, + .probe = dwmac_s700_probe, + .ops = &designware_eth_ops, + .priv_auto_alloc_size = sizeof(struct dw_eth_dev), + .platdata_auto_alloc_size = sizeof(struct eth_pdata), + .flags = DM_FLAG_ALLOC_PRIV_DMA, +};
On Sat, May 09, 2020 at 07:55:12PM +0530, Amit Singh Tomar wrote:
This patchs adds glue logic to enable designware mac present on Action Semi based S700 SoC, Configures SoC specific bits.
Undocumented bit that programs the PHY interface select register comes from vendor source.
It has been tested on Cubieboard7-lite based on S700 SoC.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
Applied to u-boot/master, thanks!
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com --- arch/arm/dts/s700-u-boot.dtsi | 13 +++++++++++++ 1 file changed, 13 insertions(+)
diff --git a/arch/arm/dts/s700-u-boot.dtsi b/arch/arm/dts/s700-u-boot.dtsi index a527cccc75f2..1b2768272c62 100644 --- a/arch/arm/dts/s700-u-boot.dtsi +++ b/arch/arm/dts/s700-u-boot.dtsi @@ -6,6 +6,19 @@ /{ soc { u-boot,dm-pre-reloc; + + gmac: ethernet@e0220000 { + compatible = "actions,s700-ethernet"; + reg = <0 0xe0220000 0 0x2000>; + interrupts = <GIC_SPI 55 IRQ_TYPE_LEVEL_HIGH>; + interrupt-names = "macirq"; + local-mac-address = [ 00 18 fe 66 66 66 ]; + clocks = <&cmu CLK_ETHERNET>, <&cmu CLK_RMII_REF>; + clock-names = "ethernet", "rmii_ref"; + phy-mode = "rmii"; + status = "okay"; + }; + }; };
On 09/05/2020 15:25, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
But that should not be the S700 SoC .dtsi, instead the cubieboard .dts file, since you specify the PHY mode in here (which is board specific).
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
arch/arm/dts/s700-u-boot.dtsi | 13 +++++++++++++ 1 file changed, 13 insertions(+)
diff --git a/arch/arm/dts/s700-u-boot.dtsi b/arch/arm/dts/s700-u-boot.dtsi index a527cccc75f2..1b2768272c62 100644 --- a/arch/arm/dts/s700-u-boot.dtsi +++ b/arch/arm/dts/s700-u-boot.dtsi @@ -6,6 +6,19 @@ /{ soc { u-boot,dm-pre-reloc;
gmac: ethernet@e0220000 {compatible = "actions,s700-ethernet";reg = <0 0xe0220000 0 0x2000>;interrupts = <GIC_SPI 55 IRQ_TYPE_LEVEL_HIGH>;interrupt-names = "macirq";local-mac-address = [ 00 18 fe 66 66 66 ];
Is there another solution to that? Maybe put that in the environment instead? Or generate this randomly or ideally by hashing some serial number?
Cheers, Andre.
clocks = <&cmu CLK_ETHERNET>, <&cmu CLK_RMII_REF>;clock-names = "ethernet", "rmii_ref";phy-mode = "rmii";status = "okay";};- };
};
On Tue, May 12, 2020 at 03:18:33PM +0100, André Przywara wrote:
On 09/05/2020 15:25, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
But that should not be the S700 SoC .dtsi, instead the cubieboard .dts file, since you specify the PHY mode in here (which is board specific).
The general way to move forward here is that bindings should be getting proposed to Linux and accepted there, and U-Boot can take a WIP of them, that gets updated later on to match, but we shouldn't get it here first.
Hi,
The general way to move forward here is that bindings should be getting proposed to Linux and accepted there, and U-Boot can take a WIP of them, that gets updated later on to match, but we shouldn't get it here first.
I do have a plan to propose this binding to Linux but this is kind of ready(I mean, the other bits like PHY and driver) for U-boot. So thought of posting it to the list here.
Thanks -Amit
On Tue, May 12, 2020 at 08:12:37PM +0530, Amit Tomer wrote:
Hi,
The general way to move forward here is that bindings should be getting proposed to Linux and accepted there, and U-Boot can take a WIP of them, that gets updated later on to match, but we shouldn't get it here first.
I do have a plan to propose this binding to Linux but this is kind of ready(I mean, the other bits like PHY and driver) for U-boot. So thought of posting it to the list here.
For some initial review, fine, but we need to have the Linux bindings being reviewed at the same time, and before things come in to U-Boot. Thanks!
On 12/05/2020 15:25, Tom Rini wrote:
On Tue, May 12, 2020 at 03:18:33PM +0100, André Przywara wrote:
On 09/05/2020 15:25, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
But that should not be the S700 SoC .dtsi, instead the cubieboard .dts file, since you specify the PHY mode in here (which is board specific).
The general way to move forward here is that bindings should be getting proposed to Linux and accepted there, and U-Boot can take a WIP of them, that gets updated later on to match, but we shouldn't get it here first.
Yes, this is of course a stop-gap measure, but a useful one: Being able to TFTP a kernel helps with developing the kernel port, especially since U-Boot doesn't support MMC (yet). And there are easier things than pushing a DT binding into the kernel tree without having a Linux driver ready ...
Cheers, Andre
On Tue, May 12, 2020 at 03:53:55PM +0100, André Przywara wrote:
On 12/05/2020 15:25, Tom Rini wrote:
On Tue, May 12, 2020 at 03:18:33PM +0100, André Przywara wrote:
On 09/05/2020 15:25, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
But that should not be the S700 SoC .dtsi, instead the cubieboard .dts file, since you specify the PHY mode in here (which is board specific).
The general way to move forward here is that bindings should be getting proposed to Linux and accepted there, and U-Boot can take a WIP of them, that gets updated later on to match, but we shouldn't get it here first.
Yes, this is of course a stop-gap measure, but a useful one: Being able to TFTP a kernel helps with developing the kernel port, especially since U-Boot doesn't support MMC (yet). And there are easier things than pushing a DT binding into the kernel tree without having a Linux driver ready ...
So, we're at -rc2 for v2020.07. The DDR calculation stuff I can see getting pulled in. Is the ethernet driver for this SoC so far from done that it's not ready for linux-next? Things don't have to be in mainline proper, but the expectation is that it's making reasonable progress there and been reviewed so that binding changes between what we take in U-Boot at first and a final re-sync once it is in mainline are minimal.
So, we're at -rc2 for v2020.07. The DDR calculation stuff I can see getting pulled in. Is the ethernet driver for this SoC so far from done that it's not ready for linux-next? Things don't have to be in mainline proper, but the expectation is that it's making reasonable progress there and been reviewed so that binding changes between what we take in U-Boot at first and a final re-sync once it is in mainline are minimal.
To be honest, we haven't put any of the Ethernet bits for Linux yet. We have put few patches for review to support MMC for S700 in Linux, and once those gets merged we would start working on Ethernet.
Thanks -Amit
On 12/05/2020 16:09, Tom Rini wrote:
Hi,
On Tue, May 12, 2020 at 03:53:55PM +0100, André Przywara wrote:
On 12/05/2020 15:25, Tom Rini wrote:
On Tue, May 12, 2020 at 03:18:33PM +0100, André Przywara wrote:
On 09/05/2020 15:25, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
But that should not be the S700 SoC .dtsi, instead the cubieboard .dts file, since you specify the PHY mode in here (which is board specific).
The general way to move forward here is that bindings should be getting proposed to Linux and accepted there, and U-Boot can take a WIP of them, that gets updated later on to match, but we shouldn't get it here first.
Yes, this is of course a stop-gap measure, but a useful one: Being able to TFTP a kernel helps with developing the kernel port, especially since U-Boot doesn't support MMC (yet). And there are easier things than pushing a DT binding into the kernel tree without having a Linux driver ready ...
So, we're at -rc2 for v2020.07. The DDR calculation stuff I can see getting pulled in. Is the ethernet driver for this SoC so far from done that it's not ready for linux-next? Things don't have to be in mainline proper, but the expectation is that it's making reasonable progress there and been reviewed so that binding changes between what we take in U-Boot at first and a final re-sync once it is in mainline are minimal.
I don't think there is any particular rush in getting this actually merged. After all it seems like the users of this board can be counted on one hand. I think it's nice to have this on the list, so interested parties can pull it in if there is a need. But we can surely wait how the binding evolves in the kernel tree, though this might take some time.
Cheers, Andre.
On Tue, May 12, 2020 at 05:39:46PM +0100, André Przywara wrote:
On 12/05/2020 16:09, Tom Rini wrote:
Hi,
On Tue, May 12, 2020 at 03:53:55PM +0100, André Przywara wrote:
On 12/05/2020 15:25, Tom Rini wrote:
On Tue, May 12, 2020 at 03:18:33PM +0100, André Przywara wrote:
On 09/05/2020 15:25, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
But that should not be the S700 SoC .dtsi, instead the cubieboard .dts file, since you specify the PHY mode in here (which is board specific).
The general way to move forward here is that bindings should be getting proposed to Linux and accepted there, and U-Boot can take a WIP of them, that gets updated later on to match, but we shouldn't get it here first.
Yes, this is of course a stop-gap measure, but a useful one: Being able to TFTP a kernel helps with developing the kernel port, especially since U-Boot doesn't support MMC (yet). And there are easier things than pushing a DT binding into the kernel tree without having a Linux driver ready ...
So, we're at -rc2 for v2020.07. The DDR calculation stuff I can see getting pulled in. Is the ethernet driver for this SoC so far from done that it's not ready for linux-next? Things don't have to be in mainline proper, but the expectation is that it's making reasonable progress there and been reviewed so that binding changes between what we take in U-Boot at first and a final re-sync once it is in mainline are minimal.
I don't think there is any particular rush in getting this actually merged. After all it seems like the users of this board can be counted on one hand. I think it's nice to have this on the list, so interested parties can pull it in if there is a need. But we can surely wait how the binding evolves in the kernel tree, though this might take some time.
OK, I think that'll be good enough for a while then. I'll mark this as RFC in patchwork so I don't forget it's not quite ready for merging anyhow. Thanks!
Hi,
On Tue, May 12, 2020 at 7:49 PM André Przywara andre.przywara@arm.com wrote:
On 09/05/2020 15:25, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
But that should not be the S700 SoC .dtsi, instead the cubieboard .dts file, since you specify the PHY mode in here (which is board specific).
But MAC is present on SoC , so I thought of adding it s700 specific u-boot.dtsi as it is already hosting few other things .
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
arch/arm/dts/s700-u-boot.dtsi | 13 +++++++++++++ 1 file changed, 13 insertions(+)
diff --git a/arch/arm/dts/s700-u-boot.dtsi b/arch/arm/dts/s700-u-boot.dtsi index a527cccc75f2..1b2768272c62 100644 --- a/arch/arm/dts/s700-u-boot.dtsi +++ b/arch/arm/dts/s700-u-boot.dtsi @@ -6,6 +6,19 @@ /{ soc { u-boot,dm-pre-reloc;
gmac: ethernet@e0220000 {compatible = "actions,s700-ethernet";reg = <0 0xe0220000 0 0x2000>;interrupts = <GIC_SPI 55 IRQ_TYPE_LEVEL_HIGH>;interrupt-names = "macirq";local-mac-address = [ 00 18 fe 66 66 66 ];Is there another solution to that? Maybe put that in the environment instead? Or generate this randomly or ideally by hashing some serial number?
Yeah, I tried having "CONFIG_NET_RANDOM_ETHADDR" other day which seems to work(would double check it).
But is there a reason not to use "local-mac-address", since driver does not support parsing this property ?
Thanks -Amit.
On 12/05/2020 15:37, Amit Tomer wrote:
Hi,
On Tue, May 12, 2020 at 7:49 PM André Przywara andre.przywara@arm.com wrote:
On 09/05/2020 15:25, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
But that should not be the S700 SoC .dtsi, instead the cubieboard .dts file, since you specify the PHY mode in here (which is board specific).
But MAC is present on SoC , so I thought of adding it s700 specific u-boot.dtsi as it is already hosting few other things .
The MAC is indeed, but not the PHY and the MAC address. And since this -u-boot.dtsi trick only works once in the chain, you have to move this to the board version of the file. Doesn't really make any difference looking at the vast number of boards using the S700 ;-)
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
arch/arm/dts/s700-u-boot.dtsi | 13 +++++++++++++ 1 file changed, 13 insertions(+)
diff --git a/arch/arm/dts/s700-u-boot.dtsi b/arch/arm/dts/s700-u-boot.dtsi index a527cccc75f2..1b2768272c62 100644 --- a/arch/arm/dts/s700-u-boot.dtsi +++ b/arch/arm/dts/s700-u-boot.dtsi @@ -6,6 +6,19 @@ /{ soc { u-boot,dm-pre-reloc;
gmac: ethernet@e0220000 {compatible = "actions,s700-ethernet";reg = <0 0xe0220000 0 0x2000>;interrupts = <GIC_SPI 55 IRQ_TYPE_LEVEL_HIGH>;interrupt-names = "macirq";local-mac-address = [ 00 18 fe 66 66 66 ];Is there another solution to that? Maybe put that in the environment instead? Or generate this randomly or ideally by hashing some serial number?
Yeah, I tried having "CONFIG_NET_RANDOM_ETHADDR" other day which seems to work(would double check it).
But is there a reason not to use "local-mac-address", since driver does not support parsing this property ?
I don't think local-mac-address is meant to be in a .dts file, actually. It's more useful to communicate some MAC address between firmware and bootloaders/kernels. You can use this in your development setup, but this should not be submitted. Otherwise every board on the planet would use the same MAC address.
Cheers, Andre
On Sat, May 09, 2020 at 07:55:13PM +0530, Amit Singh Tomar wrote:
This patch adds node for ethernet controller found on Action Semi OWL S700 SoC.
Since, there is no upstream Linux binding exist for S700 ethernet controller, Changes are put in u-boot specific dtsi file.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
Applied to u-boot/master, thanks!
This patch selects CONFIG_DM_ETH (ethernet driver is base on DM model) for Action semi owl SoC.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com --- arch/arm/Kconfig | 1 + 1 file changed, 1 insertion(+)
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 0d463088a2e4..80e07f92febc 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -904,6 +904,7 @@ config ARCH_MX5 config ARCH_OWL bool "Actions Semi OWL SoCs" select DM + select DM_ETH select DM_SERIAL select OWL_SERIAL select CLK
On Sat, May 09, 2020 at 07:55:14PM +0530, Amit Singh Tomar wrote:
This patch selects CONFIG_DM_ETH (ethernet driver is base on DM model) for Action semi owl SoC.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
Applied to u-boot/master, thanks!
This patch adds MAC and PHY related configs (needed for proper ethernet operations) for Action Semi S700 SoC.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com --- configs/cubieboard7_defconfig | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/configs/cubieboard7_defconfig b/configs/cubieboard7_defconfig index 637dc9e9fb79..c82afc340313 100644 --- a/configs/cubieboard7_defconfig +++ b/configs/cubieboard7_defconfig @@ -10,3 +10,7 @@ CONFIG_BOOTARGS="console=ttyOWL3,115200n8" # CONFIG_DISPLAY_BOARDINFO is not set CONFIG_SYS_PROMPT="U-Boot => " CONFIG_DEFAULT_DEVICE_TREE="s700-cubieboard7" +CONFIG_ETH_DESIGNWARE_S700=y +CONFIG_ETH_DESIGNWARE=y +CONFIG_PHY_REALTEK=y +CONFIG_RTL8201F_PHY_S700_RMII_TIMINGS=y
On Sat, May 09, 2020 at 07:55:15PM +0530, Amit Singh Tomar wrote:
This patch adds MAC and PHY related configs (needed for proper ethernet operations) for Action Semi S700 SoC.
Signed-off-by: Amit Singh Tomar amittomer25@gmail.com
Applied to u-boot/master, thanks!
participants (4)
-
Amit Singh Tomar -
Amit Tomer
-
André Przywara -
Tom Rini