[U-boot,0/4] Add ethernet support for MT7622
This patch series adds and enable ethernet support for MT7622, including 1. add sgmii mode support in mediatek eth driver 2. add mt7622 support in mediatek eth driver 3. add ethernet and sgmii dts node for mt7622 4. enable mt7622 ethernet support in defconfig
MarkLee (4): eth: mtk-eth: add sgmii mode support in mediatek eth driver eth: mtk-eth: add mt7622 support in mediatek eth driver arm: dts: mediatek: add ethernet and sgmii dts node for mt7622 configs: mediatek: enable mt7622 ethernet support
arch/arm/dts/mt7622-rfb.dts | 13 ++++++++++ arch/arm/dts/mt7622.dtsi | 45 +++++++++++++++++++++++++++++++++ configs/mt7622_rfb_defconfig | 4 +++ drivers/net/mtk_eth.c | 49 +++++++++++++++++++++++++++++++++++- drivers/net/mtk_eth.h | 15 +++++++++++ 5 files changed, 125 insertions(+), 1 deletion(-)
This patch add sgmii init part for the mediatek SoC that support sgmii mode. It is a must for mt7622.
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com --- drivers/net/mtk_eth.c | 44 +++++++++++++++++++++++++++++++++++++++++++ drivers/net/mtk_eth.h | 15 +++++++++++++++ 2 files changed, 59 insertions(+)
diff --git a/drivers/net/mtk_eth.c b/drivers/net/mtk_eth.c index c22e590387..8354e82244 100644 --- a/drivers/net/mtk_eth.c +++ b/drivers/net/mtk_eth.c @@ -151,6 +151,7 @@ struct mtk_eth_priv { void __iomem *fe_base; void __iomem *gmac_base; void __iomem *ethsys_base; + void __iomem *sgmii_base;
struct mii_dev *mdio_bus; int (*mii_read)(struct mtk_eth_priv *priv, u8 phy, u8 reg); @@ -750,6 +751,24 @@ static int mtk_phy_probe(struct udevice *dev) return 0; }
+static void mtk_sgmii_init(struct mtk_eth_priv *priv) +{ + /* Set SGMII GEN2 speed(2.5G) */ + clrsetbits_le32(priv->sgmii_base + SGMSYS_GEN2_SPEED, + SGMSYS_SPEED_2500, SGMSYS_SPEED_2500); + + /* Disable SGMII AN */ + clrsetbits_le32(priv->sgmii_base + SGMSYS_PCS_CONTROL_1, + SGMII_AN_ENABLE, 0); + + /* SGMII force mode setting */ + writel(SGMII_FORCE_MODE, priv->sgmii_base + SGMSYS_SGMII_MODE); + + /* Release PHYA power down state */ + clrsetbits_le32(priv->sgmii_base + SGMSYS_QPHY_PWR_STATE_CTRL, + SGMII_PHYA_PWD, 0); +} + static void mtk_mac_init(struct mtk_eth_priv *priv) { int i, ge_mode = 0; @@ -758,8 +777,13 @@ static void mtk_mac_init(struct mtk_eth_priv *priv) switch (priv->phy_interface) { case PHY_INTERFACE_MODE_RGMII_RXID: case PHY_INTERFACE_MODE_RGMII: + ge_mode = GE_MODE_RGMII; + break; case PHY_INTERFACE_MODE_SGMII: ge_mode = GE_MODE_RGMII; + mtk_ethsys_rmw(priv, ETHSYS_SYSCFG0_REG, SYSCFG0_SGMII_SEL_M, + SYSCFG0_SGMII_SEL(priv->gmac_id)); + mtk_sgmii_init(priv); break; case PHY_INTERFACE_MODE_MII: case PHY_INTERFACE_MODE_GMII: @@ -1104,6 +1128,26 @@ static int mtk_eth_ofdata_to_platdata(struct udevice *dev) } }
+ if (priv->phy_interface == PHY_INTERFACE_MODE_SGMII) { + /* get corresponding sgmii phandle */ + ret = dev_read_phandle_with_args(dev, "mediatek,sgmiisys", + NULL, 0, 0, &args); + if (ret) + return ret; + + regmap = syscon_node_to_regmap(args.node); + + if (IS_ERR(regmap)) + return PTR_ERR(regmap); + + priv->sgmii_base = regmap_get_range(regmap, 0); + + if (!priv->sgmii_base) { + dev_err(dev, "Unable to find sgmii\n"); + return -ENODEV; + } + } + /* check for switch first, otherwise phy will be used */ priv->sw = SW_NONE; priv->switch_init = NULL; diff --git a/drivers/net/mtk_eth.h b/drivers/net/mtk_eth.h index fe89a03739..9bb037d440 100644 --- a/drivers/net/mtk_eth.h +++ b/drivers/net/mtk_eth.h @@ -20,6 +20,8 @@ #define ETHSYS_SYSCFG0_REG 0x14 #define SYSCFG0_GE_MODE_S(n) (12 + ((n) * 2)) #define SYSCFG0_GE_MODE_M 0x3 +#define SYSCFG0_SGMII_SEL_M (0x3 << 8) +#define SYSCFG0_SGMII_SEL(gmac) ((!(gmac)) ? BIT(9) : BIT(8))
#define ETHSYS_CLKCFG0_REG 0x2c #define ETHSYS_TRGMII_CLK_SEL362_5 BIT(11) @@ -30,6 +32,19 @@ #define GE_MODE_MII_PHY 2 #define GE_MODE_RMII 3
+/* SGMII subsystem config registers */ +#define SGMSYS_PCS_CONTROL_1 0x0 +#define SGMII_AN_ENABLE BIT(12) + +#define SGMSYS_SGMII_MODE 0x20 +#define SGMII_FORCE_MODE 0x31120019 + +#define SGMSYS_QPHY_PWR_STATE_CTRL 0xe8 +#define SGMII_PHYA_PWD BIT(4) + +#define SGMSYS_GEN2_SPEED 0x2028 +#define SGMSYS_SPEED_2500 BIT(2) + /* Frame Engine Registers */
/* PDMA */
On Tue, Jan 21, 2020 at 07:31:57PM +0800, MarkLee wrote:
This patch add sgmii init part for the mediatek SoC that support sgmii mode. It is a must for mt7622.
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com
Applied to u-boot/master, thanks!
This patch add mt7622 support in mediatek eth driver
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com --- drivers/net/mtk_eth.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/net/mtk_eth.c b/drivers/net/mtk_eth.c index 7721ffb46c..85bbc2a3c5 100644 --- a/drivers/net/mtk_eth.c +++ b/drivers/net/mtk_eth.c @@ -136,7 +136,8 @@ enum mtk_switch {
enum mtk_soc { SOC_MT7623, - SOC_MT7629 + SOC_MT7629, + SOC_MT7622 };
struct mtk_eth_priv { @@ -1196,6 +1197,7 @@ static int mtk_eth_ofdata_to_platdata(struct udevice *dev) static const struct udevice_id mtk_eth_ids[] = { { .compatible = "mediatek,mt7629-eth", .data = SOC_MT7629 }, { .compatible = "mediatek,mt7623-eth", .data = SOC_MT7623 }, + { .compatible = "mediatek,mt7622-eth", .data = SOC_MT7622 }, {} };
On Tue, Jan 21, 2020 at 07:31:58PM +0800, MarkLee wrote:
This patch add mt7622 support in mediatek eth driver
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com
Applied to u-boot/master, thanks!
This patch add eth and sgmii dts node for mt7622 to support ethernet
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com --- arch/arm/dts/mt7622-rfb.dts | 13 +++++++++++ arch/arm/dts/mt7622.dtsi | 45 +++++++++++++++++++++++++++++++++++++ 2 files changed, 58 insertions(+)
diff --git a/arch/arm/dts/mt7622-rfb.dts b/arch/arm/dts/mt7622-rfb.dts index ec30f5c6eb..f05c3fe14d 100644 --- a/arch/arm/dts/mt7622-rfb.dts +++ b/arch/arm/dts/mt7622-rfb.dts @@ -178,3 +178,16 @@ pinctrl-0 = <&watchdog_pins>; status = "okay"; }; + +ð { + status = "okay"; + mediatek,gmac-id = <0>; + phy-mode = "sgmii"; + mediatek,switch = "mt7531"; + reset-gpios = <&gpio 54 GPIO_ACTIVE_HIGH>; + + fixed-link { + speed = <1000>; + full-duplex; + }; +}; diff --git a/arch/arm/dts/mt7622.dtsi b/arch/arm/dts/mt7622.dtsi index 7dcca5c6af..1e8ec9b48b 100644 --- a/arch/arm/dts/mt7622.dtsi +++ b/arch/arm/dts/mt7622.dtsi @@ -7,6 +7,9 @@ #include <dt-bindings/interrupt-controller/irq.h> #include <dt-bindings/interrupt-controller/arm-gic.h> #include <dt-bindings/clock/mt7622-clk.h> +#include <dt-bindings/power/mt7629-power.h> +#include <dt-bindings/reset/mt7629-reset.h> +#include <dt-bindings/gpio/gpio.h>
/ { compatible = "mediatek,mt7622"; @@ -182,4 +185,46 @@ clock-names = "source", "hclk"; status = "disabled"; }; + + ethsys: syscon@1b000000 { + compatible = "mediatek,mt7622-ethsys", "syscon"; + reg = <0x1b000000 0x1000>; + #clock-cells = <1>; + #reset-cells = <1>; + }; + + eth: ethernet@1b100000 { + compatible = "mediatek,mt7622-eth", "syscon"; + reg = <0x1b100000 0x20000>; + clocks = <&topckgen CLK_TOP_ETH_SEL>, + <ðsys CLK_ETH_ESW_EN>, + <ðsys CLK_ETH_GP0_EN>, + <ðsys CLK_ETH_GP1_EN>, + <ðsys CLK_ETH_GP2_EN>, + <&sgmiisys CLK_SGMII_TX250M_EN>, + <&sgmiisys CLK_SGMII_RX250M_EN>, + <&sgmiisys CLK_SGMII_CDR_REF>, + <&sgmiisys CLK_SGMII_CDR_FB>, + <&topckgen CLK_TOP_SGMIIPLL>, + <&apmixedsys CLK_APMIXED_ETH2PLL>; + clock-names = "ethif", "esw", "gp0", "gp1", "gp2", + "sgmii_tx250m", "sgmii_rx250m", + "sgmii_cdr_ref", "sgmii_cdr_fb", "sgmii_ck", + "eth2pll"; + power-domains = <&scpsys MT7629_POWER_DOMAIN_ETHSYS>; + resets = <ðsys ETHSYS_FE_RST>; + reset-names = "fe"; + mediatek,ethsys = <ðsys>; + mediatek,sgmiisys = <&sgmiisys>; + #address-cells = <1>; + #size-cells = <0>; + status = "disabled"; + }; + + sgmiisys: sgmiisys@1b128000 { + compatible = "mediatek,mt7622-sgmiisys", "syscon"; + reg = <0x1b128000 0x3000>; + #clock-cells = <1>; + }; + };
On Tue, Jan 21, 2020 at 07:31:59PM +0800, MarkLee wrote:
This patch add eth and sgmii dts node for mt7622 to support ethernet
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com
Applied to u-boot/master, thanks!
This patch enable mt7622 ethernet support in its defconfig
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com --- configs/mt7622_rfb_defconfig | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/configs/mt7622_rfb_defconfig b/configs/mt7622_rfb_defconfig index e1917e70e7..806087a3d6 100644 --- a/configs/mt7622_rfb_defconfig +++ b/configs/mt7622_rfb_defconfig @@ -34,6 +34,10 @@ CONFIG_SPI_FLASH_SPANSION=y CONFIG_SPI_FLASH_STMICRO=y CONFIG_SPI_FLASH_WINBOND=y CONFIG_DM_ETH=y +CONFIG_PHY_FIXED=y +CONFIG_MEDIATEK_ETH=y +CONFIG_NET_RANDOM_ETHADDR=y +CONFIG_CMD_PING=y CONFIG_PINCTRL=y CONFIG_PINCONF=y CONFIG_PINCTRL_MT7622=y
On Tue, Jan 21, 2020 at 07:32:00PM +0800, MarkLee wrote:
This patch enable mt7622 ethernet support in its defconfig
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com
configs/mt7622_rfb_defconfig | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/configs/mt7622_rfb_defconfig b/configs/mt7622_rfb_defconfig index e1917e70e7..806087a3d6 100644 --- a/configs/mt7622_rfb_defconfig +++ b/configs/mt7622_rfb_defconfig @@ -34,6 +34,10 @@ CONFIG_SPI_FLASH_SPANSION=y CONFIG_SPI_FLASH_STMICRO=y CONFIG_SPI_FLASH_WINBOND=y CONFIG_DM_ETH=y +CONFIG_PHY_FIXED=y +CONFIG_MEDIATEK_ETH=y +CONFIG_NET_RANDOM_ETHADDR=y +CONFIG_CMD_PING=y CONFIG_PINCTRL=y CONFIG_PINCONF=y CONFIG_PINCTRL_MT7622=y
This leads to warnings in the ethernet driver: drivers/net/mtk_eth.c: In function 'mtk_eth_fifo_init': drivers/net/mtk_eth.c:856:21: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] flush_dcache_range((u32)pkt_base, (u32)(pkt_base + TOTAL_PKT_BUF_SIZE)); ^ drivers/net/mtk_eth.c:856:36: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] ^ drivers/net/mtk_eth.c: In function 'mtk_eth_send': drivers/net/mtk_eth.c:968:21: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] flush_dcache_range((u32)pkt_base, (u32)pkt_base + drivers/net/mtk_eth.c:968:36: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] drivers/net/mtk_eth.c: In function 'mtk_eth_recv': drivers/net/mtk_eth.c:994:26: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] invalidate_dcache_range((u32)pkt_base, (u32)pkt_base + ^ drivers/net/mtk_eth.c:994:41: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] ^ drivers/net/mtk_eth.c: In function 'mtk_eth_probe': drivers/net/mtk_eth.c:1026:18: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast] priv->fe_base = (void *)iobase; ^ drivers/net/mtk_eth.c:1029:20: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast] priv->gmac_base = (void *)(iobase + GMAC_BASE);
Hi tom,
thanks for testing
imho the first 3 can be fixed by this:
diff --git a/drivers/net/mtk_eth.c b/drivers/net/mtk_eth.c index 6cffc3f32a..d13020a624 100644 --- a/drivers/net/mtk_eth.c +++ b/drivers/net/mtk_eth.c @@ -853,7 +853,7 @@ static void mtk_eth_fifo_init(struct mtk_eth_priv *priv) memset(priv->rx_ring_noc, 0, NUM_RX_DESC * sizeof(struct pdma_rxdesc)); memset(priv->pkt_pool, 0, TOTAL_PKT_BUF_SIZE);
- flush_dcache_range((u32)pkt_base, (u32)(pkt_base + TOTAL_PKT_BUF_SIZE)); + flush_dcache_range((int)pkt_base, (int)(pkt_base + TOTAL_PKT_BUF_SIZE));
priv->rx_dma_owner_idx0 = 0; priv->tx_cpu_owner_idx0 = 0; @@ -965,7 +965,7 @@ static int mtk_eth_send(struct udevice *dev, void *packet, int length)
pkt_base = (void *)phys_to_virt(priv->tx_ring_noc[idx].txd_info1.SDP0); memcpy(pkt_base, packet, length); - flush_dcache_range((u32)pkt_base, (u32)pkt_base + + flush_dcache_range((int)pkt_base, (int)pkt_base + roundup(length, ARCH_DMA_MINALIGN));
priv->tx_ring_noc[idx].txd_info2.SDL0 = length; @@ -991,8 +991,8 @@ static int mtk_eth_recv(struct udevice *dev, int flags, uchar **packetp)
length = priv->rx_ring_noc[idx].rxd_info2.PLEN0; pkt_base = (void *)phys_to_virt(priv->rx_ring_noc[idx].rxd_info1.PDP0); - invalidate_dcache_range((u32)pkt_base, (u32)pkt_base + - roundup(length, ARCH_DMA_MINALIGN)); + invalidate_dcache_range((int)pkt_base, (unsigned int)(pkt_base + + roundup(length, ARCH_DMA_MINALIGN)));
if (packetp) *packetp = pkt_base;
did not get the last 2 (after fixing the first 3)...
i can send full patch if it's ok
regards Frank
Gesendet: Montag, 27. Januar 2020 um 19:49 Uhr Von: "Tom Rini" trini@konsulko.com An: MarkLee Mark-MC.Lee@mediatek.com Cc: "Albert Aribaud" albert.u.boot@aribaud.net, "Ryder Lee" ryder.lee@mediatek.com, "Steven Liu" steven.liu@mediatek.com, "Joe Hershberger" joe.hershberger@ni.com, u-boot@lists.denx.de, GSS_MTK_Uboot_upstream GSS_MTK_Uboot_upstream@mediatek.com Betreff: Re: [U-boot,4/4] configs: mediatek: enable mt7622 ethernet support
On Tue, Jan 21, 2020 at 07:32:00PM +0800, MarkLee wrote:
This patch enable mt7622 ethernet support in its defconfig
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com
configs/mt7622_rfb_defconfig | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/configs/mt7622_rfb_defconfig b/configs/mt7622_rfb_defconfig index e1917e70e7..806087a3d6 100644 --- a/configs/mt7622_rfb_defconfig +++ b/configs/mt7622_rfb_defconfig @@ -34,6 +34,10 @@ CONFIG_SPI_FLASH_SPANSION=y CONFIG_SPI_FLASH_STMICRO=y CONFIG_SPI_FLASH_WINBOND=y CONFIG_DM_ETH=y +CONFIG_PHY_FIXED=y +CONFIG_MEDIATEK_ETH=y +CONFIG_NET_RANDOM_ETHADDR=y +CONFIG_CMD_PING=y CONFIG_PINCTRL=y CONFIG_PINCONF=y CONFIG_PINCTRL_MT7622=y
This leads to warnings in the ethernet driver: drivers/net/mtk_eth.c: In function 'mtk_eth_fifo_init': drivers/net/mtk_eth.c:856:21: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] flush_dcache_range((u32)pkt_base, (u32)(pkt_base + TOTAL_PKT_BUF_SIZE)); ^ drivers/net/mtk_eth.c:856:36: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] ^ drivers/net/mtk_eth.c: In function 'mtk_eth_send': drivers/net/mtk_eth.c:968:21: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] flush_dcache_range((u32)pkt_base, (u32)pkt_base + drivers/net/mtk_eth.c:968:36: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] drivers/net/mtk_eth.c: In function 'mtk_eth_recv': drivers/net/mtk_eth.c:994:26: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] invalidate_dcache_range((u32)pkt_base, (u32)pkt_base + ^ drivers/net/mtk_eth.c:994:41: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast] ^ drivers/net/mtk_eth.c: In function 'mtk_eth_probe': drivers/net/mtk_eth.c:1026:18: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast] priv->fe_base = (void *)iobase; ^ drivers/net/mtk_eth.c:1029:20: error: cast to pointer from integer of different size [-Werror=int-to-pointer-cast] priv->gmac_base = (void *)(iobase + GMAC_BASE);
-- Tom
On Tue, Jan 21, 2020 at 07:32:00PM +0800, MarkLee wrote:
This patch enable mt7622 ethernet support in its defconfig
Signed-off-by: MarkLee Mark-MC.Lee@mediatek.com
Applied to u-boot/master, thanks!
Hi Mark,
does it depend on another patch(set)?
i got unsupported switch, because in eth-driver [1] there is a check only for mt7530 (else print this error) and in dts [2], mediatek,switch is set to mt7531 (which is right, but trigger the error).
i found no patch adding check for mt7531 (or change this check) in patchwork
[1] https://github.com/frank-w/u-boot/blob/aed9e2af65651d036d742b30f8b4ce2d5f4d2... [2] https://github.com/frank-w/u-boot/blob/29464fe9546c83ad34fdcc395f6e45a9a8538...
regards Frank
On Tue, 2020-01-21 at 13:53 +0100, Frank Wunderlich wrote:
Hi Mark,
does it depend on another patch(set)?
i got unsupported switch, because in eth-driver [1] there is a check only for mt7530 (else print this error) and in dts [2], mediatek,switch is set to mt7531 (which is right, but trigger the error).
i found no patch adding check for mt7531 (or change this check) in patchwork
[1] https://github.com/frank-w/u-boot/blob/aed9e2af65651d036d742b30f8b4ce2d5f4d2... [2] https://github.com/frank-w/u-boot/blob/29464fe9546c83ad34fdcc395f6e45a9a8538...
regards Frank
Hi Frank : This patches set just focus on adding mt7622 into mtk eth MAC driver with sgmii support, and yes for some mt7622 product(bpi-r64..etc) that has a internal switch(mt7531) still need patches to support mt7531 to make networking function work fully.
We do have plan to send a separate patches for mt7531 switch once this mt7622 eth patch set accepted by uBoot. Thanks to you put a big spotlight on mt7622 development, we will do it asap.
Mark
Hi
Gesendet: Mittwoch, 22. Januar 2020 um 09:43 Uhr Von: "mtk15127" Mark-MC.Lee@mediatek.com
This patches set just focus on adding mt7622 into mtk eth MAC driver with sgmii support, and yes for some mt7622 product(bpi-r64..etc) that has a internal switch(mt7531) still need patches to support mt7531 to make networking function work fully.
i wondered about changes in ref-board dts. does it contain also mt7531 as switch? imho we need an extra dts for bpi-r64 because ram is different (1G instead of 256M) and for future changes which are only for r64.
We do have plan to send a separate patches for mt7531 switch once this mt7622 eth patch set accepted by uBoot.
@tom is this Patchset ok or does it need any changes? any way to merge it in actual merge-window, so mark can work on switch-part?
@Mark, is there any pre-release of switch-driver that i can try before you post to mailinglist?
regards Frank
participants (4)
-
Frank Wunderlich -
MarkLee -
mtk15127
-
Tom Rini