[2/2] net: ethernet: Add driver for Sunplus SP7021

Commit Message

呂芳騰 Nov. 3, 2021, 11:02 a.m. UTC

Add Ethernet driver for Sunplus SP7021.

Signed-off-by: Wells Lu <[e-mail]>
---
 MAINTAINERS                                  |   1 +
 drivers/net/ethernet/Kconfig                 |   1 +
 drivers/net/ethernet/Makefile                |   1 +
 drivers/net/ethernet/sunplus/Kconfig         |  20 +
 drivers/net/ethernet/sunplus/Makefile        |   6 +
 drivers/net/ethernet/sunplus/l2sw_define.h   | 221 ++++++++
 drivers/net/ethernet/sunplus/l2sw_desc.c     | 233 ++++++++
 drivers/net/ethernet/sunplus/l2sw_desc.h     |  21 +
 drivers/net/ethernet/sunplus/l2sw_driver.c   | 779 +++++++++++++++++++++++++++
 drivers/net/ethernet/sunplus/l2sw_driver.h   |  23 +
 drivers/net/ethernet/sunplus/l2sw_hal.c      | 422 +++++++++++++++
 drivers/net/ethernet/sunplus/l2sw_hal.h      |  47 ++
 drivers/net/ethernet/sunplus/l2sw_int.c      | 326 +++++++++++
 drivers/net/ethernet/sunplus/l2sw_int.h      |  16 +
 drivers/net/ethernet/sunplus/l2sw_mac.c      |  68 +++
 drivers/net/ethernet/sunplus/l2sw_mac.h      |  24 +
 drivers/net/ethernet/sunplus/l2sw_mdio.c     | 118 ++++
 drivers/net/ethernet/sunplus/l2sw_mdio.h     |  19 +
 drivers/net/ethernet/sunplus/l2sw_register.h |  99 ++++
 19 files changed, 2445 insertions(+)
 create mode 100644 drivers/net/ethernet/sunplus/Kconfig
 create mode 100644 drivers/net/ethernet/sunplus/Makefile
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_define.h
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_desc.c
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_desc.h
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_driver.c
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_driver.h
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_hal.c
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_hal.h
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_int.c
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_int.h
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_mac.c
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_mac.h
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_mdio.c
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_mdio.h
 create mode 100644 drivers/net/ethernet/sunplus/l2sw_register.h

Comments

Denis Kirjanov Nov. 3, 2021, 12:05 p.m. UTC | #1

11/3/21 2:02 PM, Wells Lu пишет:
> Add Ethernet driver for Sunplus SP7021.
> 
> Signed-off-by: Wells Lu <[e-mail]>
> ---
>   MAINTAINERS                                  |   1 +
>   drivers/net/ethernet/Kconfig                 |   1 +
>   drivers/net/ethernet/Makefile                |   1 +
>   drivers/net/ethernet/sunplus/Kconfig         |  20 +
>   drivers/net/ethernet/sunplus/Makefile        |   6 +
>   drivers/net/ethernet/sunplus/l2sw_define.h   | 221 ++++++++
>   drivers/net/ethernet/sunplus/l2sw_desc.c     | 233 ++++++++
>   drivers/net/ethernet/sunplus/l2sw_desc.h     |  21 +
>   drivers/net/ethernet/sunplus/l2sw_driver.c   | 779 +++++++++++++++++++++++++++
>   drivers/net/ethernet/sunplus/l2sw_driver.h   |  23 +
>   drivers/net/ethernet/sunplus/l2sw_hal.c      | 422 +++++++++++++++
>   drivers/net/ethernet/sunplus/l2sw_hal.h      |  47 ++
>   drivers/net/ethernet/sunplus/l2sw_int.c      | 326 +++++++++++
>   drivers/net/ethernet/sunplus/l2sw_int.h      |  16 +
>   drivers/net/ethernet/sunplus/l2sw_mac.c      |  68 +++
>   drivers/net/ethernet/sunplus/l2sw_mac.h      |  24 +
>   drivers/net/ethernet/sunplus/l2sw_mdio.c     | 118 ++++
>   drivers/net/ethernet/sunplus/l2sw_mdio.h     |  19 +
>   drivers/net/ethernet/sunplus/l2sw_register.h |  99 ++++
>   19 files changed, 2445 insertions(+)
>   create mode 100644 drivers/net/ethernet/sunplus/Kconfig
>   create mode 100644 drivers/net/ethernet/sunplus/Makefile
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_define.h
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_desc.c
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_desc.h
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_driver.c
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_driver.h
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_hal.c
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_hal.h
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_int.c
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_int.h
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_mac.c
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_mac.h
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_mdio.c
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_mdio.h
>   create mode 100644 drivers/net/ethernet/sunplus/l2sw_register.h
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index 4669c16..ca676ec 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -18006,6 +18006,7 @@ L:	[e-mail]
>   S:	Maintained
>   W:	https://sunplus-tibbo.atlassian.net/wiki/spaces/doc/overview
>   F:	Documentation/devicetree/bindings/net/sunplus,sp7021-l2sw.yaml
> +F:	drivers/net/ethernet/sunplus/
>   
>   SUPERH
>   M:	Yoshinori Sato <[e-mail]>
> diff --git a/drivers/net/ethernet/Kconfig b/drivers/net/ethernet/Kconfig
> index 412ae3e..0084852 100644
> --- a/drivers/net/ethernet/Kconfig
> +++ b/drivers/net/ethernet/Kconfig
> @@ -176,6 +176,7 @@ source "drivers/net/ethernet/smsc/Kconfig"
>   source "drivers/net/ethernet/socionext/Kconfig"
>   source "drivers/net/ethernet/stmicro/Kconfig"
>   source "drivers/net/ethernet/sun/Kconfig"
> +source "drivers/net/ethernet/sunplus/Kconfig"
>   source "drivers/net/ethernet/synopsys/Kconfig"
>   source "drivers/net/ethernet/tehuti/Kconfig"
>   source "drivers/net/ethernet/ti/Kconfig"
> diff --git a/drivers/net/ethernet/Makefile b/drivers/net/ethernet/Makefile
> index aaa5078..e4ce162 100644
> --- a/drivers/net/ethernet/Makefile
> +++ b/drivers/net/ethernet/Makefile
> @@ -87,6 +87,7 @@ obj-$(CONFIG_NET_VENDOR_SMSC) += smsc/
>   obj-$(CONFIG_NET_VENDOR_SOCIONEXT) += socionext/
>   obj-$(CONFIG_NET_VENDOR_STMICRO) += stmicro/
>   obj-$(CONFIG_NET_VENDOR_SUN) += sun/
> +obj-$(CONFIG_NET_VENDOR_SUNPLUS) += sunplus/
>   obj-$(CONFIG_NET_VENDOR_TEHUTI) += tehuti/
>   obj-$(CONFIG_NET_VENDOR_TI) += ti/
>   obj-$(CONFIG_NET_VENDOR_TOSHIBA) += toshiba/
> diff --git a/drivers/net/ethernet/sunplus/Kconfig b/drivers/net/ethernet/sunplus/Kconfig
> new file mode 100644
> index 0000000..a9e3a4c
> --- /dev/null
> +++ b/drivers/net/ethernet/sunplus/Kconfig
> @@ -0,0 +1,20 @@
> +# SPDX-License-Identifier: GPL-2.0
> +#
> +# Sunplus Ethernet device configuration
> +#
> +
> +config NET_VENDOR_SUNPLUS
> +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> +	depends on ETHERNET && SOC_SP7021
> +	select PHYLIB
> +	select PINCTRL_SPPCTL
> +	select COMMON_CLK_SP7021
> +	select RESET_SUNPLUS
> +	select NVMEM_SUNPLUS_OCOTP
> +	help
> +	  If you have Sunplus dual 10M/100M Ethernet (with L2 switch)
> +	  devices, say Y.
> +	  The network device supports dual 10M/100M Ethernet interfaces,
> +	  or one 10/100M Ethernet interface with two LAN ports.
> +	  To compile this driver as a module, choose M here.  The module
> +	  will be called sp_l2sw.
> diff --git a/drivers/net/ethernet/sunplus/Makefile b/drivers/net/ethernet/sunplus/Makefile
> new file mode 100644
> index 0000000..b401cec
> --- /dev/null
> +++ b/drivers/net/ethernet/sunplus/Makefile
> @@ -0,0 +1,6 @@
> +# SPDX-License-Identifier: GPL-2.0
> +#
> +# Makefile for the Sunplus network device drivers.
> +#
> +obj-$(CONFIG_NET_VENDOR_SUNPLUS) += sp_l2sw.o
> +sp_l2sw-objs := l2sw_driver.o l2sw_int.o l2sw_hal.o l2sw_desc.o l2sw_mac.o l2sw_mdio.o
> diff --git a/drivers/net/ethernet/sunplus/l2sw_define.h b/drivers/net/ethernet/sunplus/l2sw_define.h
> new file mode 100644
> index 0000000..c1049c5
> --- /dev/null
> +++ b/drivers/net/ethernet/sunplus/l2sw_define.h
> @@ -0,0 +1,221 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/* Copyright Sunplus Technology Co., Ltd.
> + *       All rights reserved.
> + */
> +
> +#ifndef __L2SW_DEFINE_H__
> +#define __L2SW_DEFINE_H__
> +
> +#include <linux/module.h>
> +#include <linux/init.h>
> +#include <linux/sched.h>
> +#include <linux/kernel.h>
> +#include <linux/slab.h>
> +#include <linux/errno.h>
> +#include <linux/types.h>
> +#include <linux/interrupt.h>
> +#include <linux/kdev_t.h>
> +#include <linux/in.h>
> +#include <linux/netdevice.h>
> +#include <linux/etherdevice.h>
> +#include <linux/ip.h>
> +#include <linux/tcp.h>
> +#include <linux/skbuff.h>
> +#include <linux/ethtool.h>
> +#include <linux/platform_device.h>
> +#include <linux/phy.h>
> +#include <linux/mii.h>
> +#include <linux/if_vlan.h>
> +#include <linux/io.h>
> +#include <linux/dma-mapping.h>
> +#include <linux/of_address.h>
> +#include <linux/of_mdio.h>
> +
> +#undef pr_fmt
> +#define pr_fmt(fmt)     "[L2SW]" fmt
> +
> +//define MAC interrupt status bit
> +#define MAC_INT_DAISY_MODE_CHG          BIT(31)
> +#define MAC_INT_IP_CHKSUM_ERR           BIT(23)
> +#define MAC_INT_WDOG_TIMER1_EXP         BIT(22)
> +#define MAC_INT_WDOG_TIMER0_EXP         BIT(21)
> +#define MAC_INT_INTRUDER_ALERT          BIT(20)
> +#define MAC_INT_PORT_ST_CHG             BIT(19)
> +#define MAC_INT_BC_STORM                BIT(18)
> +#define MAC_INT_MUST_DROP_LAN           BIT(17)
> +#define MAC_INT_GLOBAL_QUE_FULL         BIT(16)
> +#define MAC_INT_TX_SOC_PAUSE_ON         BIT(15)
> +#define MAC_INT_RX_SOC_QUE_FULL         BIT(14)
> +#define MAC_INT_TX_LAN1_QUE_FULL        BIT(9)
> +#define MAC_INT_TX_LAN0_QUE_FULL        BIT(8)
> +#define MAC_INT_RX_L_DESCF              BIT(7)
> +#define MAC_INT_RX_H_DESCF              BIT(6)
> +#define MAC_INT_RX_DONE_L               BIT(5)
> +#define MAC_INT_RX_DONE_H               BIT(4)
> +#define MAC_INT_TX_DONE_L               BIT(3)
> +#define MAC_INT_TX_DONE_H               BIT(2)
> +#define MAC_INT_TX_DES_ERR              BIT(1)
> +#define MAC_INT_RX_DES_ERR              BIT(0)
> +
> +#define MAC_INT_RX                      (MAC_INT_RX_DONE_H | MAC_INT_RX_DONE_L | \
> +					MAC_INT_RX_DES_ERR)
> +#define MAC_INT_TX                      (MAC_INT_TX_DONE_L | MAC_INT_TX_DONE_H | \
> +					MAC_INT_TX_DES_ERR)
> +#define MAC_INT_MASK_DEF                (MAC_INT_DAISY_MODE_CHG | MAC_INT_IP_CHKSUM_ERR | \
> +					MAC_INT_WDOG_TIMER1_EXP | MAC_INT_WDOG_TIMER0_EXP | \
> +					MAC_INT_INTRUDER_ALERT | MAC_INT_BC_STORM | \
> +					MAC_INT_MUST_DROP_LAN | MAC_INT_GLOBAL_QUE_FULL | \
> +					MAC_INT_TX_SOC_PAUSE_ON | MAC_INT_RX_SOC_QUE_FULL | \
> +					MAC_INT_TX_LAN1_QUE_FULL | MAC_INT_TX_LAN0_QUE_FULL | \
> +					MAC_INT_RX_L_DESCF | MAC_INT_RX_H_DESCF)
> +
> +/*define port ability*/
> +#define PORT_ABILITY_LINK_ST_P1         BIT(25)
> +#define PORT_ABILITY_LINK_ST_P0         BIT(24)
> +
> +/*define PHY command bit*/
> +#define PHY_WT_DATA_MASK                0xffff0000
> +#define PHY_RD_CMD                      0x00004000
> +#define PHY_WT_CMD                      0x00002000
> +#define PHY_REG_MASK                    0x00001f00
> +#define PHY_ADR_MASK                    0x0000001f
> +
> +/*define PHY status bit*/
> +#define PHY_RD_DATA_MASK                0xffff0000
> +#define PHY_RD_RDY                      BIT(1)
> +#define PHY_WT_DONE                     BIT(0)
> +
> +/*define other register bit*/
> +#define RX_MAX_LEN_MASK                 0x00011000
> +#define ROUTE_MODE_MASK                 0x00000060
> +#define POK_INT_THS_MASK                0x000E0000
> +#define VLAN_TH_MASK                    0x00000007
> +
> +/*define tx descriptor bit*/
> +#define OWN_BIT                         BIT(31)
> +#define FS_BIT                          BIT(25)
> +#define LS_BIT                          BIT(24)
> +#define LEN_MASK                        0x000007FF
> +#define PKTSP_MASK                      0x00007000
> +#define PKTSP_PORT1                     0x00001000
> +#define TO_VLAN_MASK                    0x0003F000
> +#define TO_VLAN_GROUP1                  0x00002000
> +
> +#define EOR_BIT                         BIT(31)
> +
> +/*define rx descriptor bit*/
> +#define ERR_CODE                        (0xf << 26)
> +#define RX_TCP_UDP_CHKSUM_BIT           BIT(23)
> +#define RX_IP_CHKSUM_BIT                BIT(18)
> +
> +#define OWC_BIT                         BIT(31)
> +#define TXOK_BIT                        BIT(26)
> +#define LNKF_BIT                        BIT(25)
> +#define BUR_BIT                         BIT(22)
> +#define TWDE_BIT                        BIT(20)
> +#define CC_MASK                         0x000f0000
> +#define TBE_MASK                        0x00070000
> +
> +// Address table search
> +#define MAC_ADDR_LOOKUP_IDLE            BIT(2)
> +#define MAC_SEARCH_NEXT_ADDR            BIT(1)
> +#define MAC_BEGIN_SEARCH_ADDR           BIT(0)
> +
> +// Address table search
> +#define MAC_HASK_LOOKUP_ADDR_MASK       (0x3ff << 22)
> +#define MAC_AT_TABLE_END                BIT(1)
> +#define MAC_AT_DATA_READY               BIT(0)
> +
> +#define MAC_PHY_ADDR                    0x01	/* define by hardware */
> +
> +/*config descriptor*/
> +#define TX_DESC_NUM                     16
> +#define MAC_GUARD_DESC_NUM              2
> +#define RX_QUEUE0_DESC_NUM              16
> +#define RX_QUEUE1_DESC_NUM              16
> +#define TX_DESC_QUEUE_NUM               1
> +#define RX_DESC_QUEUE_NUM               2
> +
> +#define MAC_TX_BUFF_SIZE                1536
> +#define MAC_RX_LEN_MAX                  2047
> +
> +#define DESC_ALIGN_BYTE                 32
> +#define RX_OFFSET                       0
> +#define TX_OFFSET                       0
> +
> +#define ETHERNET_MAC_ADDR_LEN           6
> +
> +struct mac_desc {
> +	u32 cmd1;
> +	u32 cmd2;
> +	u32 addr1;
> +	u32 addr2;
> +};
> +
> +struct skb_info {
> +	struct sk_buff *skb;
> +	u32 mapping;
> +	u32 len;
> +};
> +
> +struct l2sw_common {
> +	struct net_device *net_dev;
> +	struct platform_device *pdev;
> +	int dual_nic;
> +	int sa_learning;
> +
> +	void *desc_base;
> +	dma_addr_t desc_dma;
> +	s32 desc_size;
> +	struct clk *clk;
> +	struct reset_control *rstc;
> +	int irq;
> +
> +	struct mac_desc *rx_desc[RX_DESC_QUEUE_NUM];
> +	struct skb_info *rx_skb_info[RX_DESC_QUEUE_NUM];
> +	u32 rx_pos[RX_DESC_QUEUE_NUM];
> +	u32 rx_desc_num[RX_DESC_QUEUE_NUM];
> +	u32 rx_desc_buff_size;
> +
> +	struct mac_desc *tx_desc;
> +	struct skb_info tx_temp_skb_info[TX_DESC_NUM];
> +	u32 tx_done_pos;
> +	u32 tx_pos;
> +	u32 tx_desc_full;
> +
> +	struct mii_bus *mii_bus;
> +	struct phy_device *phy_dev;
> +
> +	struct napi_struct rx_napi;
> +	struct napi_struct tx_napi;
> +
> +	spinlock_t rx_lock;      // spinlock for accessing rx buffer
> +	spinlock_t tx_lock;      // spinlock for accessing tx buffer
> +	spinlock_t ioctl_lock;   // spinlock for ioctl operations
> +	struct mutex store_mode; // mutex for dynamic mode change

run checkpatch.pl and fix the errors

Philipp Zabel Nov. 3, 2021, 12:10 p.m. UTC | #2

On Wed, 2021-11-03 at 19:02 +0800, Wells Lu wrote:
[...]
> diff --git a/drivers/net/ethernet/sunplus/l2sw_driver.c b/drivers/net/ethernet/sunplus/l2sw_driver.c
> new file mode 100644
> index 0000000..3dfd0dd
> --- /dev/null
> +++ b/drivers/net/ethernet/sunplus/l2sw_driver.c
> @@ -0,0 +1,779 @@
[...]
> +static int l2sw_probe(struct platform_device *pdev)
> +{
> +	struct l2sw_common *comm;
> +	struct resource *r_mem;
> +	struct net_device *net_dev, *net_dev2;
> +	struct l2sw_mac *mac, *mac2;
> +	u32 mode;
> +	int ret = 0;
> +	int rc;
> +
> +	if (platform_get_drvdata(pdev))
> +		return -ENODEV;
> +
> +	// Allocate memory for l2sw 'common' area.
> +	comm = kmalloc(sizeof(*comm), GFP_KERNEL);

I'd use devm_kzalloc() here for initialization and to simplify the
cleanup path.

> +	if (!comm)
> +		return -ENOMEM;
> +	pr_debug(" comm = %p\n", comm);

What is this useful for?

> +	memset(comm, '\0', sizeof(struct l2sw_common));

Not needed with kzalloc, See above.

[...]
> +	// Get memory resoruce 0 from dts.
> +	r_mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> +	if (r_mem) {
> +		pr_debug(" res->name = \"%s\", r_mem->start = %pa\n", r_mem->name, &r_mem->start);
> +		if (l2sw_reg_base_set(devm_ioremap(&pdev->dev, r_mem->start,
> +						   (r_mem->end - r_mem->start + 1))) != 0) {
> +			pr_err(" ioremap failed!\n");
> +			ret = -ENOMEM;
> +			goto out_free_comm;
> +		}
> +	} else {
> +		pr_err(" No MEM resource 0 found!\n");
> +		ret = -ENXIO;
> +		goto out_free_comm;
> +	}
> +
> +	// Get memory resoruce 1 from dts.
> +	r_mem = platform_get_resource(pdev, IORESOURCE_MEM, 1);
> +	if (r_mem) {
> +		pr_debug(" res->name = \"%s\", r_mem->start = %pa\n", r_mem->name, &r_mem->start);
> +		if (moon5_reg_base_set(devm_ioremap(&pdev->dev, r_mem->start,
> +						    (r_mem->end - r_mem->start + 1))) != 0) {
> +			pr_err(" ioremap failed!\n");
> +			ret = -ENOMEM;
> +			goto out_free_comm;
> +		}
> +	} else {
> +		pr_err(" No MEM resource 1 found!\n");
> +		ret = -ENXIO;
> +		goto out_free_comm;
> +	}

Using devm_ioremap_resource() would simplify both a lot.

[...]
> +	comm->rstc = devm_reset_control_get(&pdev->dev, NULL);

Please use devm_reset_control_get_exclusive().

> +	if (IS_ERR(comm->rstc)) {
> +		dev_err(&pdev->dev, "Failed to retrieve reset controller!\n");
> +		ret = PTR_ERR(comm->rstc);
> +		goto out_free_comm;
> +	}
> +
> +	// Enable clock.
> +	clk_prepare_enable(comm->clk);
> +	udelay(1);
> +
> +	ret = reset_control_assert(comm->rstc);

No need to assign to ret if you ignore it anyway.

> +	udelay(1);
> +	ret = reset_control_deassert(comm->rstc);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed to deassert reset line (err = %d)!\n", ret);
> +		ret = -ENODEV;
> +		goto out_free_comm;
> +	}
> +	udelay(1);

regards
Philipp

Wells Lu 呂芳騰 Nov. 3, 2021, 2:08 p.m. UTC | #3

Hi,

Thanks a lot for review.

> From: Denis Kirjanov <[e-mail]>
> Sent: Wednesday, November 3, 2021 8:05 PM
> To: Wells Lu <[e-mail]>; [e-mail];
> [e-mail]; [e-mail]; [e-mail];
> [e-mail]; [e-mail];
> [e-mail]
> Cc: Wells Lu 呂芳騰 <[e-mail]>
> Subject: Re: [PATCH 2/2] net: ethernet: Add driver for Sunplus SP7021
> 
> 
> 
> 11/3/21 2:02 PM, Wells Lu пишет:
> > Add Ethernet driver for Sunplus SP7021.
> >
> > Signed-off-by: Wells Lu <[e-mail]>
> > ---
> >   MAINTAINERS                                  |   1 +
> >   drivers/net/ethernet/Kconfig                 |   1 +
> >   drivers/net/ethernet/Makefile                |   1 +
> >   drivers/net/ethernet/sunplus/Kconfig         |  20 +
> >   drivers/net/ethernet/sunplus/Makefile        |   6 +
> >   drivers/net/ethernet/sunplus/l2sw_define.h   | 221 ++++++++
> >   drivers/net/ethernet/sunplus/l2sw_desc.c     | 233 ++++++++
> >   drivers/net/ethernet/sunplus/l2sw_desc.h     |  21 +
> >   drivers/net/ethernet/sunplus/l2sw_driver.c   | 779
> +++++++++++++++++++++++++++
> >   drivers/net/ethernet/sunplus/l2sw_driver.h   |  23 +
> >   drivers/net/ethernet/sunplus/l2sw_hal.c      | 422 +++++++++++++++
> >   drivers/net/ethernet/sunplus/l2sw_hal.h      |  47 ++
> >   drivers/net/ethernet/sunplus/l2sw_int.c      | 326 +++++++++++
> >   drivers/net/ethernet/sunplus/l2sw_int.h      |  16 +
> >   drivers/net/ethernet/sunplus/l2sw_mac.c      |  68 +++
> >   drivers/net/ethernet/sunplus/l2sw_mac.h      |  24 +
> >   drivers/net/ethernet/sunplus/l2sw_mdio.c     | 118 ++++
> >   drivers/net/ethernet/sunplus/l2sw_mdio.h     |  19 +
> >   drivers/net/ethernet/sunplus/l2sw_register.h |  99 ++++
> >   19 files changed, 2445 insertions(+)
> >   create mode 100644 drivers/net/ethernet/sunplus/Kconfig
> >   create mode 100644 drivers/net/ethernet/sunplus/Makefile
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_define.h
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_desc.c
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_desc.h
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_driver.c
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_driver.h
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_hal.c
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_hal.h
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_int.c
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_int.h
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_mac.c
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_mac.h
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_mdio.c
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_mdio.h
> >   create mode 100644 drivers/net/ethernet/sunplus/l2sw_register.h
> >
> > diff --git a/MAINTAINERS b/MAINTAINERS index 4669c16..ca676ec 100644
> > --- a/MAINTAINERS
> > +++ b/MAINTAINERS
> > @@ -18006,6 +18006,7 @@ L:	[e-mail]
> >   S:	Maintained
> >   W:	https://sunplus-tibbo.atlassian.net/wiki/spaces/doc/overview
> >   F:	Documentation/devicetree/bindings/net/sunplus,sp7021-l2sw.yaml
> > +F:	drivers/net/ethernet/sunplus/
> >
> >   SUPERH
> >   M:	Yoshinori Sato <[e-mail]>
> > diff --git a/drivers/net/ethernet/Kconfig
> > b/drivers/net/ethernet/Kconfig index 412ae3e..0084852 100644
> > --- a/drivers/net/ethernet/Kconfig
> > +++ b/drivers/net/ethernet/Kconfig
> > @@ -176,6 +176,7 @@ source "drivers/net/ethernet/smsc/Kconfig"
> >   source "drivers/net/ethernet/socionext/Kconfig"
> >   source "drivers/net/ethernet/stmicro/Kconfig"
> >   source "drivers/net/ethernet/sun/Kconfig"
> > +source "drivers/net/ethernet/sunplus/Kconfig"
> >   source "drivers/net/ethernet/synopsys/Kconfig"
> >   source "drivers/net/ethernet/tehuti/Kconfig"
> >   source "drivers/net/ethernet/ti/Kconfig"
> > diff --git a/drivers/net/ethernet/Makefile
> > b/drivers/net/ethernet/Makefile index aaa5078..e4ce162 100644
> > --- a/drivers/net/ethernet/Makefile
> > +++ b/drivers/net/ethernet/Makefile
> > @@ -87,6 +87,7 @@ obj-$(CONFIG_NET_VENDOR_SMSC) += smsc/
> >   obj-$(CONFIG_NET_VENDOR_SOCIONEXT) += socionext/
> >   obj-$(CONFIG_NET_VENDOR_STMICRO) += stmicro/
> >   obj-$(CONFIG_NET_VENDOR_SUN) += sun/
> > +obj-$(CONFIG_NET_VENDOR_SUNPLUS) += sunplus/
> >   obj-$(CONFIG_NET_VENDOR_TEHUTI) += tehuti/
> >   obj-$(CONFIG_NET_VENDOR_TI) += ti/
> >   obj-$(CONFIG_NET_VENDOR_TOSHIBA) += toshiba/ diff --git
> > a/drivers/net/ethernet/sunplus/Kconfig
> > b/drivers/net/ethernet/sunplus/Kconfig
> > new file mode 100644
> > index 0000000..a9e3a4c
> > --- /dev/null
> > +++ b/drivers/net/ethernet/sunplus/Kconfig
> > @@ -0,0 +1,20 @@
> > +# SPDX-License-Identifier: GPL-2.0
> > +#
> > +# Sunplus Ethernet device configuration #
> > +
> > +config NET_VENDOR_SUNPLUS
> > +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> > +	depends on ETHERNET && SOC_SP7021
> > +	select PHYLIB
> > +	select PINCTRL_SPPCTL
> > +	select COMMON_CLK_SP7021
> > +	select RESET_SUNPLUS
> > +	select NVMEM_SUNPLUS_OCOTP
> > +	help
> > +	  If you have Sunplus dual 10M/100M Ethernet (with L2 switch)
> > +	  devices, say Y.
> > +	  The network device supports dual 10M/100M Ethernet interfaces,
> > +	  or one 10/100M Ethernet interface with two LAN ports.
> > +	  To compile this driver as a module, choose M here.  The module
> > +	  will be called sp_l2sw.
> > diff --git a/drivers/net/ethernet/sunplus/Makefile
> > b/drivers/net/ethernet/sunplus/Makefile
> > new file mode 100644
> > index 0000000..b401cec
> > --- /dev/null
> > +++ b/drivers/net/ethernet/sunplus/Makefile
> > @@ -0,0 +1,6 @@
> > +# SPDX-License-Identifier: GPL-2.0
> > +#
> > +# Makefile for the Sunplus network device drivers.
> > +#
> > +obj-$(CONFIG_NET_VENDOR_SUNPLUS) += sp_l2sw.o sp_l2sw-objs :=
> > +l2sw_driver.o l2sw_int.o l2sw_hal.o l2sw_desc.o l2sw_mac.o
> > +l2sw_mdio.o
> > diff --git a/drivers/net/ethernet/sunplus/l2sw_define.h
> > b/drivers/net/ethernet/sunplus/l2sw_define.h
> > new file mode 100644
> > index 0000000..c1049c5
> > --- /dev/null
> > +++ b/drivers/net/ethernet/sunplus/l2sw_define.h
> > @@ -0,0 +1,221 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +/* Copyright Sunplus Technology Co., Ltd.
> > + *       All rights reserved.
> > + */
> > +
> > +#ifndef __L2SW_DEFINE_H__
> > +#define __L2SW_DEFINE_H__
> > +
> > +#include <linux/module.h>
> > +#include <linux/init.h>
> > +#include <linux/sched.h>
> > +#include <linux/kernel.h>
> > +#include <linux/slab.h>
> > +#include <linux/errno.h>
> > +#include <linux/types.h>
> > +#include <linux/interrupt.h>
> > +#include <linux/kdev_t.h>
> > +#include <linux/in.h>
> > +#include <linux/netdevice.h>
> > +#include <linux/etherdevice.h>
> > +#include <linux/ip.h>
> > +#include <linux/tcp.h>
> > +#include <linux/skbuff.h>
> > +#include <linux/ethtool.h>
> > +#include <linux/platform_device.h>
> > +#include <linux/phy.h>
> > +#include <linux/mii.h>
> > +#include <linux/if_vlan.h>
> > +#include <linux/io.h>
> > +#include <linux/dma-mapping.h>
> > +#include <linux/of_address.h>
> > +#include <linux/of_mdio.h>
> > +
> > +#undef pr_fmt
> > +#define pr_fmt(fmt)     "[L2SW]" fmt
> > +
> > +//define MAC interrupt status bit
> > +#define MAC_INT_DAISY_MODE_CHG          BIT(31)
> > +#define MAC_INT_IP_CHKSUM_ERR           BIT(23)
> > +#define MAC_INT_WDOG_TIMER1_EXP         BIT(22)
> > +#define MAC_INT_WDOG_TIMER0_EXP         BIT(21)
> > +#define MAC_INT_INTRUDER_ALERT          BIT(20)
> > +#define MAC_INT_PORT_ST_CHG             BIT(19)
> > +#define MAC_INT_BC_STORM                BIT(18)
> > +#define MAC_INT_MUST_DROP_LAN           BIT(17)
> > +#define MAC_INT_GLOBAL_QUE_FULL         BIT(16)
> > +#define MAC_INT_TX_SOC_PAUSE_ON         BIT(15)
> > +#define MAC_INT_RX_SOC_QUE_FULL         BIT(14)
> > +#define MAC_INT_TX_LAN1_QUE_FULL        BIT(9)
> > +#define MAC_INT_TX_LAN0_QUE_FULL        BIT(8)
> > +#define MAC_INT_RX_L_DESCF              BIT(7)
> > +#define MAC_INT_RX_H_DESCF              BIT(6)
> > +#define MAC_INT_RX_DONE_L               BIT(5)
> > +#define MAC_INT_RX_DONE_H               BIT(4)
> > +#define MAC_INT_TX_DONE_L               BIT(3)
> > +#define MAC_INT_TX_DONE_H               BIT(2)
> > +#define MAC_INT_TX_DES_ERR              BIT(1)
> > +#define MAC_INT_RX_DES_ERR              BIT(0)
> > +
> > +#define MAC_INT_RX                      (MAC_INT_RX_DONE_H |
> MAC_INT_RX_DONE_L | \
> > +					MAC_INT_RX_DES_ERR)
> > +#define MAC_INT_TX                      (MAC_INT_TX_DONE_L |
> MAC_INT_TX_DONE_H | \
> > +					MAC_INT_TX_DES_ERR)
> > +#define MAC_INT_MASK_DEF
> (MAC_INT_DAISY_MODE_CHG | MAC_INT_IP_CHKSUM_ERR | \
> > +					MAC_INT_WDOG_TIMER1_EXP |
> MAC_INT_WDOG_TIMER0_EXP | \
> > +					MAC_INT_INTRUDER_ALERT | MAC_INT_BC_STORM
> | \
> > +					MAC_INT_MUST_DROP_LAN |
> MAC_INT_GLOBAL_QUE_FULL | \
> > +					MAC_INT_TX_SOC_PAUSE_ON |
> MAC_INT_RX_SOC_QUE_FULL | \
> > +					MAC_INT_TX_LAN1_QUE_FULL |
> MAC_INT_TX_LAN0_QUE_FULL | \
> > +					MAC_INT_RX_L_DESCF | MAC_INT_RX_H_DESCF)
> > +
> > +/*define port ability*/
> > +#define PORT_ABILITY_LINK_ST_P1         BIT(25)
> > +#define PORT_ABILITY_LINK_ST_P0         BIT(24)
> > +
> > +/*define PHY command bit*/
> > +#define PHY_WT_DATA_MASK                0xffff0000
> > +#define PHY_RD_CMD                      0x00004000
> > +#define PHY_WT_CMD                      0x00002000
> > +#define PHY_REG_MASK                    0x00001f00
> > +#define PHY_ADR_MASK                    0x0000001f
> > +
> > +/*define PHY status bit*/
> > +#define PHY_RD_DATA_MASK                0xffff0000
> > +#define PHY_RD_RDY                      BIT(1)
> > +#define PHY_WT_DONE                     BIT(0)
> > +
> > +/*define other register bit*/
> > +#define RX_MAX_LEN_MASK                 0x00011000
> > +#define ROUTE_MODE_MASK                 0x00000060
> > +#define POK_INT_THS_MASK                0x000E0000
> > +#define VLAN_TH_MASK                    0x00000007
> > +
> > +/*define tx descriptor bit*/
> > +#define OWN_BIT                         BIT(31)
> > +#define FS_BIT                          BIT(25)
> > +#define LS_BIT                          BIT(24)
> > +#define LEN_MASK                        0x000007FF
> > +#define PKTSP_MASK                      0x00007000
> > +#define PKTSP_PORT1                     0x00001000
> > +#define TO_VLAN_MASK                    0x0003F000
> > +#define TO_VLAN_GROUP1                  0x00002000
> > +
> > +#define EOR_BIT                         BIT(31)
> > +
> > +/*define rx descriptor bit*/
> > +#define ERR_CODE                        (0xf << 26)
> > +#define RX_TCP_UDP_CHKSUM_BIT           BIT(23)
> > +#define RX_IP_CHKSUM_BIT                BIT(18)
> > +
> > +#define OWC_BIT                         BIT(31)
> > +#define TXOK_BIT                        BIT(26)
> > +#define LNKF_BIT                        BIT(25)
> > +#define BUR_BIT                         BIT(22)
> > +#define TWDE_BIT                        BIT(20)
> > +#define CC_MASK                         0x000f0000
> > +#define TBE_MASK                        0x00070000
> > +
> > +// Address table search
> > +#define MAC_ADDR_LOOKUP_IDLE            BIT(2)
> > +#define MAC_SEARCH_NEXT_ADDR            BIT(1)
> > +#define MAC_BEGIN_SEARCH_ADDR           BIT(0)
> > +
> > +// Address table search
> > +#define MAC_HASK_LOOKUP_ADDR_MASK       (0x3ff << 22)
> > +#define MAC_AT_TABLE_END                BIT(1)
> > +#define MAC_AT_DATA_READY               BIT(0)
> > +
> > +#define MAC_PHY_ADDR                    0x01	/* define by
> hardware */
> > +
> > +/*config descriptor*/
> > +#define TX_DESC_NUM                     16
> > +#define MAC_GUARD_DESC_NUM              2
> > +#define RX_QUEUE0_DESC_NUM              16
> > +#define RX_QUEUE1_DESC_NUM              16
> > +#define TX_DESC_QUEUE_NUM               1
> > +#define RX_DESC_QUEUE_NUM               2
> > +
> > +#define MAC_TX_BUFF_SIZE                1536
> > +#define MAC_RX_LEN_MAX                  2047
> > +
> > +#define DESC_ALIGN_BYTE                 32
> > +#define RX_OFFSET                       0
> > +#define TX_OFFSET                       0
> > +
> > +#define ETHERNET_MAC_ADDR_LEN           6
> > +
> > +struct mac_desc {
> > +	u32 cmd1;
> > +	u32 cmd2;
> > +	u32 addr1;
> > +	u32 addr2;
> > +};
> > +
> > +struct skb_info {
> > +	struct sk_buff *skb;
> > +	u32 mapping;
> > +	u32 len;
> > +};
> > +
> > +struct l2sw_common {
> > +	struct net_device *net_dev;
> > +	struct platform_device *pdev;
> > +	int dual_nic;
> > +	int sa_learning;
> > +
> > +	void *desc_base;
> > +	dma_addr_t desc_dma;
> > +	s32 desc_size;
> > +	struct clk *clk;
> > +	struct reset_control *rstc;
> > +	int irq;
> > +
> > +	struct mac_desc *rx_desc[RX_DESC_QUEUE_NUM];
> > +	struct skb_info *rx_skb_info[RX_DESC_QUEUE_NUM];
> > +	u32 rx_pos[RX_DESC_QUEUE_NUM];
> > +	u32 rx_desc_num[RX_DESC_QUEUE_NUM];
> > +	u32 rx_desc_buff_size;
> > +
> > +	struct mac_desc *tx_desc;
> > +	struct skb_info tx_temp_skb_info[TX_DESC_NUM];
> > +	u32 tx_done_pos;
> > +	u32 tx_pos;
> > +	u32 tx_desc_full;
> > +
> > +	struct mii_bus *mii_bus;
> > +	struct phy_device *phy_dev;
> > +
> > +	struct napi_struct rx_napi;
> > +	struct napi_struct tx_napi;
> > +
> > +	spinlock_t rx_lock;      // spinlock for accessing rx buffer
> > +	spinlock_t tx_lock;      // spinlock for accessing tx buffer
> > +	spinlock_t ioctl_lock;   // spinlock for ioctl operations
> > +	struct mutex store_mode; // mutex for dynamic mode change
> 
> run checkpatch.pl and fix the errors

I ran checkpatch.pl with -f option for all *.c and *.h files, like:

   $ checkpatch.pl -f *.c
   $ checkpatch.pl -f *.h

I got 0 warnings and 0 errors for all files.

Also, I ran checkpath.pl for patch files, except cover-letter patch, as 
shown below:

   $ checkpatch.pl 0001*.patch
   $ checkpatch.pl 0002*.patch

I got 0 warnings and 0 errors for the 2 patch files. But the second patch 
file shows some 'CHECK' items, like 

    "Unnecessary parentheses around...",
    "Macro argument reuse 'N' may be better as'(N)'..."

I am sure the code has no problem.
Do I need to clean 'CHECK' items?


Best regards,

Wells Lu 呂芳騰 Nov. 3, 2021, 3:11 p.m. UTC | #4

Thanks a lot for your review.

> From: Philipp Zabel <[e-mail]>
> Sent: Wednesday, November 3, 2021 8:10 PM
> To: Wells Lu <[e-mail]>; [e-mail];
> [e-mail]; [e-mail]; [e-mail];
> [e-mail]; [e-mail]
> Cc: Wells Lu 呂芳騰 <[e-mail]>
> Subject: Re: [PATCH 2/2] net: ethernet: Add driver for Sunplus SP7021
> 
> On Wed, 2021-11-03 at 19:02 +0800, Wells Lu wrote:
> [...]
> > diff --git a/drivers/net/ethernet/sunplus/l2sw_driver.c
> > b/drivers/net/ethernet/sunplus/l2sw_driver.c
> > new file mode 100644
> > index 0000000..3dfd0dd
> > --- /dev/null
> > +++ b/drivers/net/ethernet/sunplus/l2sw_driver.c
> > @@ -0,0 +1,779 @@
> [...]
> > +static int l2sw_probe(struct platform_device *pdev) {
> > +	struct l2sw_common *comm;
> > +	struct resource *r_mem;
> > +	struct net_device *net_dev, *net_dev2;
> > +	struct l2sw_mac *mac, *mac2;
> > +	u32 mode;
> > +	int ret = 0;
> > +	int rc;
> > +
> > +	if (platform_get_drvdata(pdev))
> > +		return -ENODEV;
> > +
> > +	// Allocate memory for l2sw 'common' area.
> > +	comm = kmalloc(sizeof(*comm), GFP_KERNEL);
> 
> I'd use devm_kzalloc() here for initialization and to simplify the cleanup
> path.

I'll replace kmalloc() with devm_kzalloc() and remove the kfree(...) in
cleanup path (platform remove) in PATCH v2.

> > +	if (!comm)
> > +		return -ENOMEM;
> > +	pr_debug(" comm = %p\n", comm);
> 
> What is this useful for?

'pr_debug(" comm = %p\n", comm);' is for early-stage debug.
I'll remove the line in PATCH v2.

> > +	memset(comm, '\0', sizeof(struct l2sw_common));
> 
> Not needed with kzalloc, See above.

Yes, I'll remove the line 'memset(...);' as devm_kzalloc() will be used in PATCH v2.

> [...]
> > +	// Get memory resoruce 0 from dts.
> > +	r_mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > +	if (r_mem) {
> > +		pr_debug(" res->name = \"%s\", r_mem->start = %pa\n",
> r_mem->name, &r_mem->start);
> > +		if (l2sw_reg_base_set(devm_ioremap(&pdev->dev, r_mem->start,
> > +						   (r_mem->end - r_mem->start + 1))) != 0) {
> > +			pr_err(" ioremap failed!\n");
> > +			ret = -ENOMEM;
> > +			goto out_free_comm;
> > +		}
> > +	} else {
> > +		pr_err(" No MEM resource 0 found!\n");
> > +		ret = -ENXIO;
> > +		goto out_free_comm;
> > +	}
> > +
> > +	// Get memory resoruce 1 from dts.
> > +	r_mem = platform_get_resource(pdev, IORESOURCE_MEM, 1);
> > +	if (r_mem) {
> > +		pr_debug(" res->name = \"%s\", r_mem->start = %pa\n",
> r_mem->name, &r_mem->start);
> > +		if (moon5_reg_base_set(devm_ioremap(&pdev->dev,
> r_mem->start,
> > +						    (r_mem->end - r_mem->start + 1))) != 0) {
> > +			pr_err(" ioremap failed!\n");
> > +			ret = -ENOMEM;
> > +			goto out_free_comm;
> > +		}
> > +	} else {
> > +		pr_err(" No MEM resource 1 found!\n");
> > +		ret = -ENXIO;
> > +		goto out_free_comm;
> > +	}
> 
> Using devm_ioremap_resource() would simplify both a lot.

Yes, I'll replace devm_ioremap() with devm_ioremap_resource() in PATCH v2.

> [...]
> > +	comm->rstc = devm_reset_control_get(&pdev->dev, NULL);
> 
> Please use devm_reset_control_get_exclusive().

Yes, I'll replace devm_reset_control_get() with devm_reset_control_get_exclusive() in PATCH v2.

> > +	if (IS_ERR(comm->rstc)) {
> > +		dev_err(&pdev->dev, "Failed to retrieve reset controller!\n");
> > +		ret = PTR_ERR(comm->rstc);
> > +		goto out_free_comm;
> > +	}
> > +
> > +	// Enable clock.
> > +	clk_prepare_enable(comm->clk);
> > +	udelay(1);
> > +
> > +	ret = reset_control_assert(comm->rstc);
> 
> No need to assign to ret if you ignore it anyway.

Ok, I'll remove 'ret =' in PATCH v2

> > +	udelay(1);
> > +	ret = reset_control_deassert(comm->rstc);
> > +	if (ret) {
> > +		dev_err(&pdev->dev, "Failed to deassert reset line (err = %d)!\n",
> ret);
> > +		ret = -ENODEV;
> > +		goto out_free_comm;
> > +	}
> > +	udelay(1);
> 
> regards
> Philipp

Best regards,
Wells

Randy Dunlap Nov. 3, 2021, 3:52 p.m. UTC | #5

Hi--

On 11/3/21 4:02 AM, Wells Lu wrote:
> diff --git a/drivers/net/ethernet/sunplus/Kconfig b/drivers/net/ethernet/sunplus/Kconfig
> new file mode 100644
> index 0000000..a9e3a4c
> --- /dev/null
> +++ b/drivers/net/ethernet/sunplus/Kconfig
> @@ -0,0 +1,20 @@
> +# SPDX-License-Identifier: GPL-2.0
> +#
> +# Sunplus Ethernet device configuration
> +#
> +
> +config NET_VENDOR_SUNPLUS
> +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> +	depends on ETHERNET && SOC_SP7021
> +	select PHYLIB
> +	select PINCTRL_SPPCTL
> +	select COMMON_CLK_SP7021
> +	select RESET_SUNPLUS
> +	select NVMEM_SUNPLUS_OCOTP
> +	help
> +	  If you have Sunplus dual 10M/100M Ethernet (with L2 switch)
> +	  devices, say Y.
> +	  The network device supports dual 10M/100M Ethernet interfaces,
> +	  or one 10/100M Ethernet interface with two LAN ports.
> +	  To compile this driver as a module, choose M here.  The module
> +	  will be called sp_l2sw.

Please use NET_VENDOR_SUNPLUS in the same way that other
NET_VENDOR_wyxz kconfig symbols are used. It should just enable
or disable any specific device drivers under it.

Andrew Lunn Nov. 3, 2021, 4:51 p.m. UTC | #6

> +config NET_VENDOR_SUNPLUS
> +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"

The "with L2 Switch" is causing lots of warning bells to ring for me.

I don't see any references to switchdev or DSA in this driver. How is
the switch managed? There have been a few examples in the past of
similar two port switches being first supported in Dual MAC
mode. Later trying to actually use the switch in the Linux was always
ran into problems, and basically needed a new driver. So i want to
make sure you don't have this problem.

In the Linux world, Ethernet switches default to having there
ports/interfaces separated. This effectively gives you your dual MAC
mode by default.  You then create a Linux bridge, and add the
ports/interfaces to the bridge. switchdev is used to offload the
bridge, telling the hardware to enable the L2 switch between the
ports.

So you don't need the mode parameter in DT. switchdev tells you
this. Switchdev gives user space access to the address table etc.

> +obj-$(CONFIG_NET_VENDOR_SUNPLUS) += sp_l2sw.o
...
> +struct l2sw_common {

Please change your prefix. l2sw is a common prefix, there are other
silicon vendors using l2sw. I would suggest sp_l2sw or spl2sw.

> +static int ethernet_do_ioctl(struct net_device *net_dev, struct ifreq *ifr, int cmd)
> +{
> +	struct l2sw_mac *mac = netdev_priv(net_dev);
> +	struct l2sw_common *comm = mac->comm;
> +	struct mii_ioctl_data *data = if_mii(ifr);
> +	unsigned long flags;
> +
> +	pr_debug(" if = %s, cmd = %04x\n", ifr->ifr_ifrn.ifrn_name, cmd);
> +	pr_debug(" phy_id = %d, reg_num = %d, val_in = %04x\n", data->phy_id,
> +		 data->reg_num, data->val_in);

You should not be using any of the pr_ functions. You have a net_dev,
so netdev_dbg().

> +
> +	// Check parameters' range.
> +	if ((cmd == SIOCGMIIREG) || (cmd == SIOCSMIIREG)) {
> +		if (data->reg_num > 31) {
> +			pr_err(" reg_num (= %d) excesses range!\n", (int)data->reg_num);

Don't spam the kernel log for things like this.

> +			return -EINVAL;
> +		}
> +	}
> +
> +	switch (cmd) {
> +	case SIOCGMIIPHY:
> +		if (comm->dual_nic && (strcmp(ifr->ifr_ifrn.ifrn_name, "eth1") == 0))

You cannot rely on the name, systemd has probably renamed it. If you
have using phylib correctly, net_dev->phydev is what you want.


> +			return comm->phy2_addr;
> +		else
> +			return comm->phy1_addr;
> +
> +	case SIOCGMIIREG:
> +		spin_lock_irqsave(&comm->ioctl_lock, flags);
> +		data->val_out = mdio_read(data->phy_id, data->reg_num);
> +		spin_unlock_irqrestore(&comm->ioctl_lock, flags);
> +		pr_debug(" val_out = %04x\n", data->val_out);
> +		break;
> +
> +	case SIOCSMIIREG:
> +		spin_lock_irqsave(&comm->ioctl_lock, flags);
> +		mdio_write(data->phy_id, data->reg_num, data->val_in);
> +		spin_unlock_irqrestore(&comm->ioctl_lock, flags);
> +		break;
> +

You should be using phylink_mii_ioctl() or phy_mii_ioctl().

You locking is also suspect.

> +static int ethernet_change_mtu(struct net_device *ndev, int new_mtu)
> +{
> +	if (netif_running(ndev))
> +		return -EBUSY;
> +
> +	if (new_mtu < 68 || new_mtu > ETH_DATA_LEN)
> +		return -EINVAL;

The core will do this for you, if you set the values in the ndev
correct at probe time.

> +
> +	ndev->mtu = new_mtu;
> +
> +	return 0;
> +}
> +


> +static int mdio_access(u8 op_cd, u8 dev_reg_addr, u8 phy_addr, u32 wdata)
> +{
> +	u32 value, time = 0;
> +
> +	HWREG_W(phy_cntl_reg0, (wdata << 16) | (op_cd << 13) | (dev_reg_addr << 8) | phy_addr);
> +	wmb();			// make sure settings are effective.

That suggests you are using the wrong macros to access the registers.

> +	do {
> +		if (++time > MDIO_RW_TIMEOUT_RETRY_NUMBERS) {
> +			pr_err(" mdio failed to operate!\n");
> +			time = 0;
> +		}
> +
> +		value = HWREG_R(phy_cntl_reg1);
> +	} while ((value & 0x3) == 0);

include/linux/iopoll.h.


> +	if (time == 0)
> +		return -1;

-ETIMDEOUT. One of the advantages of iopoll.h is that reusing code
 avoids issues like this.

> +u32 mdio_read(u32 phy_id, u16 regnum)
> +{
> +	int ret;

Please check for C45 and return -EOPNOTSUPP.

> +
> +	ret = mdio_access(MDIO_READ_CMD, regnum, phy_id, 0);
> +	if (ret < 0)
> +		return -EIO;
> +
> +	return ret;
> +}
> +
> +u32 mdio_write(u32 phy_id, u32 regnum, u16 val)
> +{
> +	int ret;
> +

Please check for C45 and return -EOPNOTSUPP.

> +	ret = mdio_access(MDIO_WRITE_CMD, regnum, phy_id, val);
> +	if (ret < 0)
> +		return -EIO;
> +
> +	return 0;
> +}
> +
> +inline void tx_trigger(void)

No inline functions in C code. Let the compiler decide.

> +int phy_cfg(struct l2sw_mac *mac)
> +{
> +	// Bug workaround:
> +	// Flow-control of phy should be enabled. L2SW IP flow-control will refer
> +	// to the bit to decide to enable or disable flow-control.
> +	mdio_write(mac->comm->phy1_addr, 4, mdio_read(mac->comm->phy1_addr, 4) | (1 << 10));
> +	mdio_write(mac->comm->phy2_addr, 4, mdio_read(mac->comm->phy2_addr, 4) | (1 << 10));

This should be in the PHY driver. The MAC driver should never need to
touch PHY registers.

> +++ b/drivers/net/ethernet/sunplus/l2sw_mdio.c
> @@ -0,0 +1,118 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/* Copyright Sunplus Technology Co., Ltd.
> + *       All rights reserved.
> + */
> +
> +#include "l2sw_mdio.h"
> +
> +static int mii_read(struct mii_bus *bus, int phy_id, int regnum)
> +{
> +	return mdio_read(phy_id, regnum);
> +}
> +
> +static int mii_write(struct mii_bus *bus, int phy_id, int regnum, u16 val)
> +{
> +	return mdio_write(phy_id, regnum, val);
> +}
> +
> +u32 mdio_init(struct platform_device *pdev, struct net_device *net_dev)
> +{
> +	struct l2sw_mac *mac = netdev_priv(net_dev);
> +	struct mii_bus *mii_bus;
> +	struct device_node *mdio_node;
> +	u32 ret;
> +
> +	mii_bus = mdiobus_alloc();
> +	if (!mii_bus) {
> +		pr_err(" Failed to allocate mdio_bus memory!\n");
> +		return -ENOMEM;
> +	}
> +
> +	mii_bus->name = "sunplus_mii_bus";
> +	mii_bus->parent = &pdev->dev;
> +	mii_bus->priv = mac;
> +	mii_bus->read = mii_read;
> +	mii_bus->write = mii_write;
> +	snprintf(mii_bus->id, MII_BUS_ID_SIZE, "%s-mii", dev_name(&pdev->dev));
> +
> +	mdio_node = of_get_parent(mac->comm->phy1_node);
> +	ret = of_mdiobus_register(mii_bus, mdio_node);
> +	if (ret) {
> +		pr_err(" Failed to register mii bus (ret = %d)!\n", ret);
> +		mdiobus_free(mii_bus);
> +		return ret;
> +	}
> +
> +	mac->comm->mii_bus = mii_bus;
> +	return ret;
> +}
> +
> +void mdio_remove(struct net_device *net_dev)
> +{
> +	struct l2sw_mac *mac = netdev_priv(net_dev);
> +
> +	if (mac->comm->mii_bus) {
> +		mdiobus_unregister(mac->comm->mii_bus);
> +		mdiobus_free(mac->comm->mii_bus);
> +		mac->comm->mii_bus = NULL;
> +	}
> +}

You MDIO code is pretty scattered around. Please bring it all together
in one file.

> +static void mii_linkchange(struct net_device *netdev)
> +{
> +}

Nothing to do? Seems very odd. Don't you need to tell the MAC it
should do 10Mbps or 100Mbps? What about pause?

> +
> +int mac_phy_probe(struct net_device *netdev)
> +{
> +	struct l2sw_mac *mac = netdev_priv(netdev);
> +	struct phy_device *phydev;
> +	int i;
> +
> +	phydev = of_phy_connect(mac->net_dev, mac->comm->phy1_node, mii_linkchange,
> +				0, PHY_INTERFACE_MODE_RGMII_ID);

You should not hard code PHY_INTERFACE_MODE_RGMII_ID. Use the DT
property "phy-mode"

> +	if (!phydev) {
> +		pr_err(" \"%s\" has no phy found\n", netdev->name);
> +		return -1;

-ENODEV;

Never use -1, pick an error code.

> +	}
> +
> +	if (mac->comm->phy2_node) {
> +		of_phy_connect(mac->net_dev, mac->comm->phy2_node, mii_linkchange,
> +			       0, PHY_INTERFACE_MODE_RGMII_ID);
> +	}
> +
> +	linkmode_clear_bit(ETHTOOL_LINK_MODE_Pause_BIT, phydev->supported);
> +	linkmode_clear_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT, phydev->supported);

So the MAC does not support pause? I'm then confused about phy_cfg().

   Andrew

Wells Lu 呂芳騰 Nov. 3, 2021, 6:08 p.m. UTC | #7

> 
> Hi--
> 
> On 11/3/21 4:02 AM, Wells Lu wrote:
> > diff --git a/drivers/net/ethernet/sunplus/Kconfig
> > b/drivers/net/ethernet/sunplus/Kconfig
> > new file mode 100644
> > index 0000000..a9e3a4c
> > --- /dev/null
> > +++ b/drivers/net/ethernet/sunplus/Kconfig
> > @@ -0,0 +1,20 @@
> > +# SPDX-License-Identifier: GPL-2.0
> > +#
> > +# Sunplus Ethernet device configuration #
> > +
> > +config NET_VENDOR_SUNPLUS
> > +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> > +	depends on ETHERNET && SOC_SP7021
> > +	select PHYLIB
> > +	select PINCTRL_SPPCTL
> > +	select COMMON_CLK_SP7021
> > +	select RESET_SUNPLUS
> > +	select NVMEM_SUNPLUS_OCOTP
> > +	help
> > +	  If you have Sunplus dual 10M/100M Ethernet (with L2 switch)
> > +	  devices, say Y.
> > +	  The network device supports dual 10M/100M Ethernet interfaces,
> > +	  or one 10/100M Ethernet interface with two LAN ports.
> > +	  To compile this driver as a module, choose M here.  The module
> > +	  will be called sp_l2sw.
> 
> Please use NET_VENDOR_SUNPLUS in the same way that other
> NET_VENDOR_wyxz kconfig symbols are used. It should just enable or
> disable any specific device drivers under it.
> 
> 
> --
> ~Randy

I looked up Kconfig file of other vendors, but not sure what I should do.
Do I need to modify Kconfig file in the form as shown below?

# SPDX-License-Identifier: GPL-2.0
#
# Sunplus device configuration
#

config NET_VENDOR_SUNPLUS
	bool "Sunplus devices"
	default y
	depends on ARCH_SUNPLUS
	---help---
	  If you have a network (Ethernet) card belonging to this
	  class, say Y here.

	  Note that the answer to this question doesn't directly
	  affect the kernel: saying N will just cause the configurator
	  to skip all the questions about Sunplus cards. If you say Y,
	  you will be asked for your specific card in the following
	  questions.

if NET_VENDOR_SUNPLUS

config SP7021_EMAC
	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
	depends on ETHERNET && SOC_SP7021
	select PHYLIB
	select PINCTRL_SPPCTL
	select COMMON_CLK_SP7021
	select RESET_SUNPLUS
	select NVMEM_SUNPLUS_OCOTP
	help
	  If you have Sunplus dual 10M/100M Ethernet (with L2 switch)
	  devices, say Y.
	  The network device supports dual 10M/100M Ethernet interfaces,
	  or one 10/100M Ethernet interface with two LAN ports.
	  To compile this driver as a module, choose M here.  The module
	  will be called sp_l2sw.

endif # NET_VENDOR_SUNPLUS

Best regards,
Wells

Andrew Lunn Nov. 3, 2021, 7:30 p.m. UTC | #8

> config NET_VENDOR_SUNPLUS
> 	bool "Sunplus devices"
> 	default y
> 	depends on ARCH_SUNPLUS

Does it actually depend on ARCH_SUNPLUS? What do you make use of?

Ideally, you want it to also build with COMPILE_TEST, so that the
driver gets build by 0-day and all the other build bots.

> 	---help---
> 	  If you have a network (Ethernet) card belonging to this
> 	  class, say Y here.
> 
> 	  Note that the answer to this question doesn't directly
> 	  affect the kernel: saying N will just cause the configurator
> 	  to skip all the questions about Sunplus cards. If you say Y,
> 	  you will be asked for your specific card in the following
> 	  questions.
> 
> if NET_VENDOR_SUNPLUS
> 
> config SP7021_EMAC
> 	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> 	depends on ETHERNET && SOC_SP7021

Does it actually depend on SOC_SP7021 to build?

     Andrew

Randy Dunlap Nov. 3, 2021, 8:26 p.m. UTC | #9

On 11/3/21 11:08 AM, Wells Lu 呂芳騰 wrote:
>>
>> Hi--
>>
>> On 11/3/21 4:02 AM, Wells Lu wrote:
>>> diff --git a/drivers/net/ethernet/sunplus/Kconfig
>>> b/drivers/net/ethernet/sunplus/Kconfig
>>> new file mode 100644
>>> index 0000000..a9e3a4c
>>> --- /dev/null
>>> +++ b/drivers/net/ethernet/sunplus/Kconfig
>>> @@ -0,0 +1,20 @@
>>> +# SPDX-License-Identifier: GPL-2.0
>>> +#
>>> +# Sunplus Ethernet device configuration #
>>> +
>>> +config NET_VENDOR_SUNPLUS
>>> +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
>>> +	depends on ETHERNET && SOC_SP7021
>>> +	select PHYLIB
>>> +	select PINCTRL_SPPCTL
>>> +	select COMMON_CLK_SP7021
>>> +	select RESET_SUNPLUS
>>> +	select NVMEM_SUNPLUS_OCOTP
>>> +	help
>>> +	  If you have Sunplus dual 10M/100M Ethernet (with L2 switch)
>>> +	  devices, say Y.
>>> +	  The network device supports dual 10M/100M Ethernet interfaces,
>>> +	  or one 10/100M Ethernet interface with two LAN ports.
>>> +	  To compile this driver as a module, choose M here.  The module
>>> +	  will be called sp_l2sw.
>>
>> Please use NET_VENDOR_SUNPLUS in the same way that other
>> NET_VENDOR_wyxz kconfig symbols are used. It should just enable or
>> disable any specific device drivers under it.
>>
>>
>> --
>> ~Randy
> 
> I looked up Kconfig file of other vendors, but not sure what I should do.
> Do I need to modify Kconfig file in the form as shown below?

Hi,

Yes, this is the correct general idea, but also consider
Andrew's comments.

Thanks.

> # SPDX-License-Identifier: GPL-2.0
> #
> # Sunplus device configuration
> #
> 
> config NET_VENDOR_SUNPLUS
> 	bool "Sunplus devices"
> 	default y
> 	depends on ARCH_SUNPLUS
> 	---help---
> 	  If you have a network (Ethernet) card belonging to this
> 	  class, say Y here.
> 
> 	  Note that the answer to this question doesn't directly
> 	  affect the kernel: saying N will just cause the configurator
> 	  to skip all the questions about Sunplus cards. If you say Y,
> 	  you will be asked for your specific card in the following
> 	  questions.
> 
> if NET_VENDOR_SUNPLUS
> 
> config SP7021_EMAC
> 	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> 	depends on ETHERNET && SOC_SP7021
> 	select PHYLIB
> 	select PINCTRL_SPPCTL
> 	select COMMON_CLK_SP7021
> 	select RESET_SUNPLUS
> 	select NVMEM_SUNPLUS_OCOTP
> 	help
> 	  If you have Sunplus dual 10M/100M Ethernet (with L2 switch)
> 	  devices, say Y.
> 	  The network device supports dual 10M/100M Ethernet interfaces,
> 	  or one 10/100M Ethernet interface with two LAN ports.
> 	  To compile this driver as a module, choose M here.  The module
> 	  will be called sp_l2sw.
> 
> endif # NET_VENDOR_SUNPLUS
> 
> Best regards,
> Wells
>

Wells Lu 呂芳騰 Nov. 4, 2021, 5:31 a.m. UTC | #10

Hi,

Thanks a lot for review.

> 
> > config NET_VENDOR_SUNPLUS
> > 	bool "Sunplus devices"
> > 	default y
> > 	depends on ARCH_SUNPLUS
> 
> Does it actually depend on ARCH_SUNPLUS? What do you make use of?

ARCH_SUNPLUS will be defined for Sunplus family series SoC.
Ethernet devices of Sunplus are designed and used for Sunplus SoC.
So far, only two SoC of Sunplus have the network device.
I'd like to show up the selection only for Sunplus SoC.

> 
> Ideally, you want it to also build with COMPILE_TEST, so that the driver gets
> build by 0-day and all the other build bots.

I am not sure if this is mandatory or not.
Should I add COMPILE_TEST as below?

	depends on ARCH_SUNPLUS | COMPILE_TEST

> 
> > 	---help---
> > 	  If you have a network (Ethernet) card belonging to this
> > 	  class, say Y here.
> >
> > 	  Note that the answer to this question doesn't directly
> > 	  affect the kernel: saying N will just cause the configurator
> > 	  to skip all the questions about Sunplus cards. If you say Y,
> > 	  you will be asked for your specific card in the following
> > 	  questions.
> >
> > if NET_VENDOR_SUNPLUS
> >
> > config SP7021_EMAC
> > 	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> > 	depends on ETHERNET && SOC_SP7021
> 
> Does it actually depend on SOC_SP7021 to build?
> 
>      Andrew

Yes, the device is now only for Sunplus SP7021 SoC.
Devices in each SoC may have a bit difference because of adding new 
function or improving something.

Andrew Lunn Nov. 4, 2021, 12:59 p.m. UTC | #11

On Thu, Nov 04, 2021 at 05:31:57AM +0000, Wells Lu 呂芳騰 wrote:
> Hi,
> 
> Thanks a lot for review.
> 
> > 
> > > config NET_VENDOR_SUNPLUS
> > > 	bool "Sunplus devices"
> > > 	default y
> > > 	depends on ARCH_SUNPLUS
> > 
> > Does it actually depend on ARCH_SUNPLUS? What do you make use of?
> 
> ARCH_SUNPLUS will be defined for Sunplus family series SoC.
> Ethernet devices of Sunplus are designed and used for Sunplus SoC.
> So far, only two SoC of Sunplus have the network device.
> I'd like to show up the selection only for Sunplus SoC.

So it does not actually depend on ARCH_SUNPLUS. There are a few cases
where drivers have needed to call into arch specific code, which stops
them building for any other arch.

> > Ideally, you want it to also build with COMPILE_TEST, so that the driver gets
> > build by 0-day and all the other build bots.
> 
> I am not sure if this is mandatory or not.
> Should I add COMPILE_TEST as below?
> 
> 	depends on ARCH_SUNPLUS | COMPILE_TEST

Yes.

> Yes, the device is now only for Sunplus SP7021 SoC.
> Devices in each SoC may have a bit difference because of adding new 
> function or improving something.

If it will compile with COMPILE_TEST on x86, mips, etc, you should
allow it to compile with COMPILE_TEST. You get better compile testing
that way.

     Andrew

Randy Dunlap Nov. 4, 2021, 2:55 p.m. UTC | #12

On 11/4/21 5:59 AM, Andrew Lunn wrote:
> On Thu, Nov 04, 2021 at 05:31:57AM +0000, Wells Lu 呂芳騰 wrote:
>> Hi,
>>
>> Thanks a lot for review.
>>
>>>
>>>> config NET_VENDOR_SUNPLUS
>>>> 	bool "Sunplus devices"
>>>> 	default y
>>>> 	depends on ARCH_SUNPLUS
>>>
>>> Does it actually depend on ARCH_SUNPLUS? What do you make use of?
>>
>> ARCH_SUNPLUS will be defined for Sunplus family series SoC.
>> Ethernet devices of Sunplus are designed and used for Sunplus SoC.
>> So far, only two SoC of Sunplus have the network device.
>> I'd like to show up the selection only for Sunplus SoC.
> 
> So it does not actually depend on ARCH_SUNPLUS. There are a few cases
> where drivers have needed to call into arch specific code, which stops
> them building for any other arch.
> 
>>> Ideally, you want it to also build with COMPILE_TEST, so that the driver gets
>>> build by 0-day and all the other build bots.
>>
>> I am not sure if this is mandatory or not.
>> Should I add COMPILE_TEST as below?
>>
>> 	depends on ARCH_SUNPLUS | COMPILE_TEST
> 
> Yes.

Yes, but use "||" instead of one "|".

> 
>> Yes, the device is now only for Sunplus SP7021 SoC.
>> Devices in each SoC may have a bit difference because of adding new
>> function or improving something.
> 
> If it will compile with COMPILE_TEST on x86, mips, etc, you should
> allow it to compile with COMPILE_TEST. You get better compile testing
> that way.
> 
>       Andrew
>

Wells Lu 呂芳騰 Nov. 4, 2021, 5:46 p.m. UTC | #13

> On Thu, Nov 04, 2021 at 05:31:57AM +0000, Wells Lu 呂芳騰 wrote:
> > Hi,
> >
> > Thanks a lot for review.
> >
> > >
> > > > config NET_VENDOR_SUNPLUS
> > > > 	bool "Sunplus devices"
> > > > 	default y
> > > > 	depends on ARCH_SUNPLUS
> > >
> > > Does it actually depend on ARCH_SUNPLUS? What do you make use of?
> >
> > ARCH_SUNPLUS will be defined for Sunplus family series SoC.
> > Ethernet devices of Sunplus are designed and used for Sunplus SoC.
> > So far, only two SoC of Sunplus have the network device.
> > I'd like to show up the selection only for Sunplus SoC.
> 
> So it does not actually depend on ARCH_SUNPLUS. There are a few cases where
> drivers have needed to call into arch specific code, which stops them building
> for any other arch.
> 
> > > Ideally, you want it to also build with COMPILE_TEST, so that the
> > > driver gets build by 0-day and all the other build bots.
> >
> > I am not sure if this is mandatory or not.
> > Should I add COMPILE_TEST as below?
> >
> > 	depends on ARCH_SUNPLUS | COMPILE_TEST
> 
> Yes.
> 
> > Yes, the device is now only for Sunplus SP7021 SoC.
> > Devices in each SoC may have a bit difference because of adding new
> > function or improving something.
> 
> If it will compile with COMPILE_TEST on x86, mips, etc, you should allow it to
> compile with COMPILE_TEST. You get better compile testing that way.
> 
>      Andrew

No, we only develop arm-based SoC, never for x86 or mips.
We never compile the driver for x86 or mips machine.

Wells Lu 呂芳騰 Nov. 4, 2021, 5:51 p.m. UTC | #14

> On 11/4/21 5:59 AM, Andrew Lunn wrote:
> > On Thu, Nov 04, 2021 at 05:31:57AM +0000, Wells Lu 呂芳騰 wrote:
> >> Hi,
> >>
> >> Thanks a lot for review.
> >>
> >>>
> >>>> config NET_VENDOR_SUNPLUS
> >>>> 	bool "Sunplus devices"
> >>>> 	default y
> >>>> 	depends on ARCH_SUNPLUS
> >>>
> >>> Does it actually depend on ARCH_SUNPLUS? What do you make use of?
> >>
> >> ARCH_SUNPLUS will be defined for Sunplus family series SoC.
> >> Ethernet devices of Sunplus are designed and used for Sunplus SoC.
> >> So far, only two SoC of Sunplus have the network device.
> >> I'd like to show up the selection only for Sunplus SoC.
> >
> > So it does not actually depend on ARCH_SUNPLUS. There are a few cases
> > where drivers have needed to call into arch specific code, which stops
> > them building for any other arch.
> >
> >>> Ideally, you want it to also build with COMPILE_TEST, so that the
> >>> driver gets build by 0-day and all the other build bots.
> >>
> >> I am not sure if this is mandatory or not.
> >> Should I add COMPILE_TEST as below?
> >>
> >> 	depends on ARCH_SUNPLUS | COMPILE_TEST
> >
> > Yes.
> 
> Yes, but use "||" instead of one "|".
> 
> >
> >> Yes, the device is now only for Sunplus SP7021 SoC.
> >> Devices in each SoC may have a bit difference because of adding new
> >> function or improving something.
> >
> > If it will compile with COMPILE_TEST on x86, mips, etc, you should
> > allow it to compile with COMPILE_TEST. You get better compile testing
> > that way.
> >
> >       Andrew
> >
> 
> 
> --
> ~Randy

I will do it in PATCH v2.

Thanks,

Andrew Lunn Nov. 4, 2021, 6:21 p.m. UTC | #15

> No, we only develop arm-based SoC, never for x86 or mips.
> We never compile the driver for x86 or mips machine.

You don't, but the Linux community does build for those
architectures. Most people do tree wide refactoring work using
x86. Tree wide cleanups using x86, etc. Any changes like that could
touch your driver. The harder is it to build, the less build testing
it will get, and tree wide changes which break it are less likely to
get noticed.  So you really do want it to compile cleanly for all
architectures. If it does not, it normally actually means you are
doing something wrong, something you need to fix anyway. So please do
build it for x86 and make sure it builds cleanly.

	  Andrew

Wells Lu 呂芳騰 Nov. 4, 2021, 7:03 p.m. UTC | #16

> > No, we only develop arm-based SoC, never for x86 or mips.
> > We never compile the driver for x86 or mips machine.
> 
> You don't, but the Linux community does build for those architectures. Most
> people do tree wide refactoring work using x86. Tree wide cleanups using x86,
> etc. Any changes like that could touch your driver. The harder is it to build, the
> less build testing it will get, and tree wide changes which break it are less likely
> to get noticed.  So you really do want it to compile cleanly for all
> architectures. If it does not, it normally actually means you are doing
> something wrong, something you need to fix anyway. So please do build it for
> x86 and make sure it builds cleanly.
> 
> 	  Andrew

Ok, I understand.
I'll add COMPILE_TEST and compile driver for x86.

Thanks,

Wells Lu 呂芳騰 Nov. 5, 2021, 11:25 a.m. UTC | #17

Hi,

Thanks a lot for your review.

> > +config NET_VENDOR_SUNPLUS
> > +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> 
> The "with L2 Switch" is causing lots of warning bells to ring for me.
> 
> I don't see any references to switchdev or DSA in this driver. How is the
> switch managed? There have been a few examples in the past of similar two
> port switches being first supported in Dual MAC mode. Later trying to
> actually use the switch in the Linux was always ran into problems, and
> basically needed a new driver. So i want to make sure you don't have this
> problem.
> 
> In the Linux world, Ethernet switches default to having there
> ports/interfaces separated. This effectively gives you your dual MAC mode by
> default.  You then create a Linux bridge, and add the ports/interfaces to the
> bridge. switchdev is used to offload the bridge, telling the hardware to
> enable the L2 switch between the ports.
> 
> So you don't need the mode parameter in DT. switchdev tells you this.
> Switchdev gives user space access to the address table etc.

The L2 switch of Ethernet of SP7021 is not used to forward packets 
between two network interfaces.

Sunplus Dual Ethernet devices consists of one CPU port, two LAN 
ports, and a L2 switch. L2 switch is a circuitry which receives packets 
from CPU or LAN ports and then forwards them other ports. Rules of 
forwarding packets are set by driver.

Ethernet driver of SP7021 of Sunplus supports 3 operation modes:
  - Dual NIC mode
  - An NIC with two LAN ports mode (daisy-chain mode)
  - An NIC with two LAN ports mode 2

Dual NIC mode
Ethernet driver creates two net-device interfaces (eg: eth0 and eth1). 
Each has its dedicated LAN port. For example, LAN port 0 is for 
net-device interface eth0. LAN port 1 is for net-device interface 
eth1. Packets from LAN port 0 will be always forwarded to eth0 and 
vice versa by L2 switch. Similarly, packets from LAN port 1 will be 
always forwarded to eth1 and vice versa by L2 switch. Packets will 
never be forwarded between two LAN ports, or between eth0 and 
LAN port 1, or between eth1 and LAN port 0. The two network 
devices work independently.

An NIC with two LAN ports mode (daisy-chain mode)
Ethernet driver creates one net-device interface (eg: eth0), but the 
net-device interface has two LAN ports. In this mode, a packet from 
one LAN port will be either forwarded to net-device interface (eht0) 
if its destination address matches MAC address of net-device 
interface (eth0), or forwarded to other LAN port. A packet from 
net-device interface (eth0) will be forwarded to a LAN port if its 
destination address is learnt by L2 switch, or forwarded to both 
LAN ports if its destination has not been learnt yet.

An NIC with two LAN ports mode 2
This mode is similar to “An NIC with two LAN ports mode”. The 
difference is that a packet from net-device interface (eth0) will be 
always forwarded to both LAN ports. Learning function of L2 switch 
is turned off in this mode. This means L2 switch will never learn the 
source address of a packet. So, it always forward packets to both 
LAN ports. This mode works like you have 2-port Ethernet hub.

> 
> > +obj-$(CONFIG_NET_VENDOR_SUNPLUS) += sp_l2sw.o
> ...
> > +struct l2sw_common {
> 
> Please change your prefix. l2sw is a common prefix, there are other silicon
> vendors using l2sw. I would suggest sp_l2sw or spl2sw.

Ok, I'll modify two struct names in next patch as shown below:
l2sw_common --> sp_common
l2sw_mac --> sp_mac

Should I also modify prefix of file name?

> > +static int ethernet_do_ioctl(struct net_device *net_dev, struct ifreq
> > +*ifr, int cmd) {
> > +	struct l2sw_mac *mac = netdev_priv(net_dev);
> > +	struct l2sw_common *comm = mac->comm;
> > +	struct mii_ioctl_data *data = if_mii(ifr);
> > +	unsigned long flags;
> > +
> > +	pr_debug(" if = %s, cmd = %04x\n", ifr->ifr_ifrn.ifrn_name, cmd);
> > +	pr_debug(" phy_id = %d, reg_num = %d, val_in = %04x\n",
> data->phy_id,
> > +		 data->reg_num, data->val_in);
> 
> You should not be using any of the pr_ functions. You have a net_dev, so
> netdev_dbg().

Ok, I will replace pr_xxx() functions with netdev_xxx() functions 
where 'net_dev' is available in next patch.


> > +
> > +	// Check parameters' range.
> > +	if ((cmd == SIOCGMIIREG) || (cmd == SIOCSMIIREG)) {
> > +		if (data->reg_num > 31) {
> > +			pr_err(" reg_num (= %d) excesses range!\n",
> (int)data->reg_num);
> 
> Don't spam the kernel log for things like this.

Ok, I'll remove the pr_err() line in next patch.


> > +			return -EINVAL;
> > +		}
> > +	}
> > +
> > +	switch (cmd) {
> > +	case SIOCGMIIPHY:
> > +		if (comm->dual_nic && (strcmp(ifr->ifr_ifrn.ifrn_name, "eth1") ==
> > +0))
> 
> You cannot rely on the name, systemd has probably renamed it. If you have
> using phylib correctly, net_dev->phydev is what you want.

Ok, I'll use name of the second net device to do compare, 
instead of using fixed string "eth1", in next patch.


> 
> 
> > +			return comm->phy2_addr;
> > +		else
> > +			return comm->phy1_addr;
> > +
> > +	case SIOCGMIIREG:
> > +		spin_lock_irqsave(&comm->ioctl_lock, flags);
> > +		data->val_out = mdio_read(data->phy_id, data->reg_num);
> > +		spin_unlock_irqrestore(&comm->ioctl_lock, flags);
> > +		pr_debug(" val_out = %04x\n", data->val_out);
> > +		break;
> > +
> > +	case SIOCSMIIREG:
> > +		spin_lock_irqsave(&comm->ioctl_lock, flags);
> > +		mdio_write(data->phy_id, data->reg_num, data->val_in);
> > +		spin_unlock_irqrestore(&comm->ioctl_lock, flags);
> > +		break;
> > +
> 
> You should be using phylink_mii_ioctl() or phy_mii_ioctl().
> 
> You locking is also suspect.

Ok, I'll use phy_mii_ioctl() for SIOCSMIIREG and SIOCGMIIREG commands 
and remove spin_lock in next patch.


> > +static int ethernet_change_mtu(struct net_device *ndev, int new_mtu)
> > +{
> > +	if (netif_running(ndev))
> > +		return -EBUSY;
> > +
> > +	if (new_mtu < 68 || new_mtu > ETH_DATA_LEN)
> > +		return -EINVAL;
> 
> The core will do this for you, if you set the values in the ndev correct at
> probe time.

Ok, I'll remove the whole function ethernet_change_mtu () in next patch 
as the core will do everything for changing mtu.


> > +
> > +	ndev->mtu = new_mtu;
> > +
> > +	return 0;
> > +}
> > +
> 
> 
> > +static int mdio_access(u8 op_cd, u8 dev_reg_addr, u8 phy_addr, u32
> > +wdata) {
> > +	u32 value, time = 0;
> > +
> > +	HWREG_W(phy_cntl_reg0, (wdata << 16) | (op_cd << 13) |
> (dev_reg_addr << 8) | phy_addr);
> > +	wmb();			// make sure settings are effective.
> 
> That suggests you are using the wrong macros to access the registers.

Ok, I'll modify HWREG_W() and HWREG_R() and remove wmb().


> > +	do {
> > +		if (++time > MDIO_RW_TIMEOUT_RETRY_NUMBERS) {
> > +			pr_err(" mdio failed to operate!\n");
> > +			time = 0;
> > +		}
> > +
> > +		value = HWREG_R(phy_cntl_reg1);
> > +	} while ((value & 0x3) == 0);
> 
> include/linux/iopoll.h.
> 
> 
> > +	if (time == 0)
> > +		return -1;
> 
> -ETIMDEOUT. One of the advantages of iopoll.h is that reusing code  avoids
> issues like this.

Ok, I'll replace the do-while loop with read_poll_timeout().


> > +u32 mdio_read(u32 phy_id, u16 regnum) {
> > +	int ret;
> 
> Please check for C45 and return -EOPNOTSUPP.

Ok, I'll modify mdio_read() to return -EOPNOTSUPP 
when error (time-out) occurs.

> > +
> > +	ret = mdio_access(MDIO_READ_CMD, regnum, phy_id, 0);
> > +	if (ret < 0)
> > +		return -EIO;
> > +
> > +	return ret;
> > +}
> > +
> > +u32 mdio_write(u32 phy_id, u32 regnum, u16 val) {
> > +	int ret;
> > +
> 
> Please check for C45 and return -EOPNOTSUPP.

Ok, I'll modify mdio_read() to return -EOPNOTSUPP 
when error (time-out) occurs.


> > +
> > +inline void tx_trigger(void)
> 
> No inline functions in C code. Let the compiler decide.

Ok, I'll remove 'inline' for all functions in next patch.


> > +int phy_cfg(struct l2sw_mac *mac)
> > +{
> > +	// Bug workaround:
> > +	// Flow-control of phy should be enabled. L2SW IP flow-control will refer
> > +	// to the bit to decide to enable or disable flow-control.
> > +	mdio_write(mac->comm->phy1_addr, 4,
> mdio_read(mac->comm->phy1_addr, 4) | (1 << 10));
> > +	mdio_write(mac->comm->phy2_addr, 4,
> mdio_read(mac->comm->phy2_addr,
> > +4) | (1 << 10));
> 
> This should be in the PHY driver. The MAC driver should never need to touch
> PHY registers.

Sunplus Ethernet MAC integrates MDIO controller. 
So Ethernet driver has MDIO- and PHY-related code. 
To work-around a circuitry bug, we need to enable 
bit 10 of register 4 of PHY.
Where should we place the code?


> > +++ b/drivers/net/ethernet/sunplus/l2sw_mdio.c
> > @@ -0,0 +1,118 @@
> > +// SPDX-License-Identifier: GPL-2.0
> > +/* Copyright Sunplus Technology Co., Ltd.
> > + *       All rights reserved.
> > + */
> > +
> > +#include "l2sw_mdio.h"
> > +
> > +static int mii_read(struct mii_bus *bus, int phy_id, int regnum) {
> > +	return mdio_read(phy_id, regnum);
> > +}
> > +
> > +static int mii_write(struct mii_bus *bus, int phy_id, int regnum, u16
> > +val) {
> > +	return mdio_write(phy_id, regnum, val); }
> > +
> > +u32 mdio_init(struct platform_device *pdev, struct net_device
> > +*net_dev) {
> > +	struct l2sw_mac *mac = netdev_priv(net_dev);
> > +	struct mii_bus *mii_bus;
> > +	struct device_node *mdio_node;
> > +	u32 ret;
> > +
> > +	mii_bus = mdiobus_alloc();
> > +	if (!mii_bus) {
> > +		pr_err(" Failed to allocate mdio_bus memory!\n");
> > +		return -ENOMEM;
> > +	}
> > +
> > +	mii_bus->name = "sunplus_mii_bus";
> > +	mii_bus->parent = &pdev->dev;
> > +	mii_bus->priv = mac;
> > +	mii_bus->read = mii_read;
> > +	mii_bus->write = mii_write;
> > +	snprintf(mii_bus->id, MII_BUS_ID_SIZE, "%s-mii",
> > +dev_name(&pdev->dev));
> > +
> > +	mdio_node = of_get_parent(mac->comm->phy1_node);
> > +	ret = of_mdiobus_register(mii_bus, mdio_node);
> > +	if (ret) {
> > +		pr_err(" Failed to register mii bus (ret = %d)!\n", ret);
> > +		mdiobus_free(mii_bus);
> > +		return ret;
> > +	}
> > +
> > +	mac->comm->mii_bus = mii_bus;
> > +	return ret;
> > +}
> > +
> > +void mdio_remove(struct net_device *net_dev) {
> > +	struct l2sw_mac *mac = netdev_priv(net_dev);
> > +
> > +	if (mac->comm->mii_bus) {
> > +		mdiobus_unregister(mac->comm->mii_bus);
> > +		mdiobus_free(mac->comm->mii_bus);
> > +		mac->comm->mii_bus = NULL;
> > +	}
> > +}
> 
> You MDIO code is pretty scattered around. Please bring it all together in one
> file.

Ok, I'll move mdio_read() and mdio_write() to 'l2sw_mdio.c',
but keep mdio_access() in 'l2sw_hal.c' because it is a function 
which accesses hardware registers actually.


> > +static void mii_linkchange(struct net_device *netdev) { }
> 
> Nothing to do? Seems very odd. Don't you need to tell the MAC it should do
> 10Mbps or 100Mbps? What about pause?

No, hardware does it automatically.
Sunplus MAC integrates MDIO controller.
It reads PHY status and set MAC automatically.


> > +
> > +int mac_phy_probe(struct net_device *netdev) {
> > +	struct l2sw_mac *mac = netdev_priv(netdev);
> > +	struct phy_device *phydev;
> > +	int i;
> > +
> > +	phydev = of_phy_connect(mac->net_dev, mac->comm->phy1_node,
> mii_linkchange,
> > +				0, PHY_INTERFACE_MODE_RGMII_ID);
> 
> You should not hard code PHY_INTERFACE_MODE_RGMII_ID. Use the DT
> property "phy-mode"

Ok, I'll do it in next patch.
I create a new file 'l2sw_phy.c' and move phy-related functions
from 'l2sw_hal.c' to it.


> > +	if (!phydev) {
> > +		pr_err(" \"%s\" has no phy found\n", netdev->name);
> > +		return -1;
> 
> -ENODEV;
> 
> Never use -1, pick an error code.

Ok, I'll modify it in next patch.

> 
> > +	}
> > +
> > +	if (mac->comm->phy2_node) {
> > +		of_phy_connect(mac->net_dev, mac->comm->phy2_node,
> mii_linkchange,
> > +			       0, PHY_INTERFACE_MODE_RGMII_ID);
> > +	}
> > +
> > +	linkmode_clear_bit(ETHTOOL_LINK_MODE_Pause_BIT,
> phydev->supported);
> > +	linkmode_clear_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT,
> > +phydev->supported);
> 
> So the MAC does not support pause? I'm then confused about phy_cfg().

Yes, MAC supports pause. MAC (hardware) takes care of pause 
automatically.

Should I remove the two lines?


> 
>    Andrew

Andrew Lunn Nov. 5, 2021, 1:37 p.m. UTC | #18

> > > +config NET_VENDOR_SUNPLUS
> > > +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> > 
> > The "with L2 Switch" is causing lots of warning bells to ring for me.
> > 
> > I don't see any references to switchdev or DSA in this driver. How is the
> > switch managed? There have been a few examples in the past of similar two
> > port switches being first supported in Dual MAC mode. Later trying to
> > actually use the switch in the Linux was always ran into problems, and
> > basically needed a new driver. So i want to make sure you don't have this
> > problem.
> > 
> > In the Linux world, Ethernet switches default to having there
> > ports/interfaces separated. This effectively gives you your dual MAC mode by
> > default.  You then create a Linux bridge, and add the ports/interfaces to the
> > bridge. switchdev is used to offload the bridge, telling the hardware to
> > enable the L2 switch between the ports.
> > 
> > So you don't need the mode parameter in DT. switchdev tells you this.
> > Switchdev gives user space access to the address table etc.
> 
> The L2 switch of Ethernet of SP7021 is not used to forward packets 
> between two network interfaces.
> 
> Sunplus Dual Ethernet devices consists of one CPU port, two LAN 
> ports, and a L2 switch. L2 switch is a circuitry which receives packets 
> from CPU or LAN ports and then forwards them other ports. Rules of 
> forwarding packets are set by driver.
> 
> Ethernet driver of SP7021 of Sunplus supports 3 operation modes:
>   - Dual NIC mode
>   - An NIC with two LAN ports mode (daisy-chain mode)
>   - An NIC with two LAN ports mode 2
> 
> Dual NIC mode
> Ethernet driver creates two net-device interfaces (eg: eth0 and eth1). 
> Each has its dedicated LAN port. For example, LAN port 0 is for 
> net-device interface eth0. LAN port 1 is for net-device interface 
> eth1. Packets from LAN port 0 will be always forwarded to eth0 and 
> vice versa by L2 switch. Similarly, packets from LAN port 1 will be 
> always forwarded to eth1 and vice versa by L2 switch. Packets will 
> never be forwarded between two LAN ports, or between eth0 and 
> LAN port 1, or between eth1 and LAN port 0. The two network 
> devices work independently.
> 
> An NIC with two LAN ports mode (daisy-chain mode)
> Ethernet driver creates one net-device interface (eg: eth0), but the 
> net-device interface has two LAN ports. In this mode, a packet from 
> one LAN port will be either forwarded to net-device interface (eht0) 
> if its destination address matches MAC address of net-device 
> interface (eth0), or forwarded to other LAN port. A packet from 
> net-device interface (eth0) will be forwarded to a LAN port if its 
> destination address is learnt by L2 switch, or forwarded to both 
> LAN ports if its destination has not been learnt yet.
> 
> An NIC with two LAN ports mode 2
> This mode is similar to “An NIC with two LAN ports mode”. The 
> difference is that a packet from net-device interface (eth0) will be 
> always forwarded to both LAN ports. Learning function of L2 switch 
> is turned off in this mode. This means L2 switch will never learn the 
> source address of a packet. So, it always forward packets to both 
> LAN ports. This mode works like you have 2-port Ethernet hub.

So here you describe how the hardware can be used. Dual is two
interfaces. Daisy-chain is what you get by taking those two interfaces
and adding them to a bridge. The bridge then forwards frames between
the interfaces and the CPU as needed, based on learning. And your
third mode is the bridge always performs flooding.

A linux driver must follow the linux networking model. You cannot make
up your own model. In the linux world, you model the external
ports. The hardware always has two external ports, so you need to
always have two netdev interfaces. To bridge packets between those two
interfaces, you create a bridge and you add the interfaces to the
bridge. That is the model you need to follow. switchdev gives you the
API calls you need to implement this.

> > > +struct l2sw_common {
> > 
> > Please change your prefix. l2sw is a common prefix, there are other silicon
> > vendors using l2sw. I would suggest sp_l2sw or spl2sw.
> 
> Ok, I'll modify two struct names in next patch as shown below:
> l2sw_common --> sp_common
> l2sw_mac --> sp_mac
> 
> Should I also modify prefix of file name?

You need to modify the prefix everywhere you use it.  Function names,
variable names, all symbols. Search and replace throughout the whole
code.

> > > +			return -EINVAL;
> > > +		}
> > > +	}
> > > +
> > > +	switch (cmd) {
> > > +	case SIOCGMIIPHY:
> > > +		if (comm->dual_nic && (strcmp(ifr->ifr_ifrn.ifrn_name, "eth1") ==
> > > +0))
> > 
> > You cannot rely on the name, systemd has probably renamed it. If you have
> > using phylib correctly, net_dev->phydev is what you want.
> 
> Ok, I'll use name of the second net device to do compare, 
> instead of using fixed string "eth1", in next patch.

No. There are always two interfaces. You always have two netdev
structures. Each netdev structure has a phydev. So use netdev->phydev.

This is another advantage of the Linux model. In your daisy chain
mode, how do i control the two PHYs? How do i see one is up and one is
down? How do i configure one to 10Half and the other 100Full?

> > > +int phy_cfg(struct l2sw_mac *mac)
> > > +{
> > > +	// Bug workaround:
> > > +	// Flow-control of phy should be enabled. L2SW IP flow-control will refer
> > > +	// to the bit to decide to enable or disable flow-control.
> > > +	mdio_write(mac->comm->phy1_addr, 4,
> > mdio_read(mac->comm->phy1_addr, 4) | (1 << 10));
> > > +	mdio_write(mac->comm->phy2_addr, 4,
> > mdio_read(mac->comm->phy2_addr,
> > > +4) | (1 << 10));
> > 
> > This should be in the PHY driver. The MAC driver should never need to touch
> > PHY registers.
> 
> Sunplus Ethernet MAC integrates MDIO controller. 
> So Ethernet driver has MDIO- and PHY-related code. 
> To work-around a circuitry bug, we need to enable 
> bit 10 of register 4 of PHY.
> Where should we place the code?

The silicon is integrated, but it is still a collection of standard
blocks. Linux models those blocks independently. There is a subsystem
for the MAC, a subsystem for the MDIO bus master and a subsystem for
the PHY. You register a driver with each of these subsystems. PHY
drivers live in drivers/net/phy. Put a PHY driver in there, which
includes this workaround.

> > > +static void mii_linkchange(struct net_device *netdev) { }
> > 
> > Nothing to do? Seems very odd. Don't you need to tell the MAC it should do
> > 10Mbps or 100Mbps? What about pause?
> 
> No, hardware does it automatically.
> Sunplus MAC integrates MDIO controller.
> It reads PHY status and set MAC automatically.

The PHY is external? So you have no idea what PHY that is? It could be
a Marvell PHY, a microchip PHY, an Atheros PHY. Often PHYs have
pages. In order to read the temperature sensor you change the page,
read a register, and then hopefully change the page back again. If the
PHY supports Fibre as well as copper, it can put the fibre registers
in a second page. The PHY driver knows about this, it will flip the
pages as needed. The phylib core has a mutex, so that only one
operation happens at a time. So a page flip does not happen
unexpectedly.

Your MAC hardware does not take this mutex. It has no idea what page
is selected when it reads registers. Instead of getting the basic mode
register, it could get the LED control register...

The MAC should never directly access the PHY. Please disable this
hardware, and use the mii_linkchange callback to configure the MAC.

> > So the MAC does not support pause? I'm then confused about phy_cfg().
 
> Yes, MAC supports pause. MAC (hardware) takes care of pause 
> automatically.
> 
> Should I remove the two lines?

Yes.

And you need to configure the MAC based on the results of the
auto-neg.

	Andrew

Wells Lu 呂芳騰 Nov. 8, 2021, 9:37 a.m. UTC | #19

> > > > +config NET_VENDOR_SUNPLUS
> > > > +	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
> > >
> > > The "with L2 Switch" is causing lots of warning bells to ring for me.
> > >
> > > I don't see any references to switchdev or DSA in this driver. How
> > > is the switch managed? There have been a few examples in the past of
> > > similar two port switches being first supported in Dual MAC mode.
> > > Later trying to actually use the switch in the Linux was always ran
> > > into problems, and basically needed a new driver. So i want to make
> > > sure you don't have this problem.
> > >
> > > In the Linux world, Ethernet switches default to having there
> > > ports/interfaces separated. This effectively gives you your dual MAC
> > > mode by default.  You then create a Linux bridge, and add the
> > > ports/interfaces to the bridge. switchdev is used to offload the
> > > bridge, telling the hardware to enable the L2 switch between the ports.
> > >
> > > So you don't need the mode parameter in DT. switchdev tells you this.
> > > Switchdev gives user space access to the address table etc.
> >
> > The L2 switch of Ethernet of SP7021 is not used to forward packets
> > between two network interfaces.
> >
> > Sunplus Dual Ethernet devices consists of one CPU port, two LAN ports,
> > and a L2 switch. L2 switch is a circuitry which receives packets from
> > CPU or LAN ports and then forwards them other ports. Rules of
> > forwarding packets are set by driver.
> >
> > Ethernet driver of SP7021 of Sunplus supports 3 operation modes:
> >   - Dual NIC mode
> >   - An NIC with two LAN ports mode (daisy-chain mode)
> >   - An NIC with two LAN ports mode 2
> >
> > Dual NIC mode
> > Ethernet driver creates two net-device interfaces (eg: eth0 and eth1).
> > Each has its dedicated LAN port. For example, LAN port 0 is for
> > net-device interface eth0. LAN port 1 is for net-device interface
> > eth1. Packets from LAN port 0 will be always forwarded to eth0 and
> > vice versa by L2 switch. Similarly, packets from LAN port 1 will be
> > always forwarded to eth1 and vice versa by L2 switch. Packets will
> > never be forwarded between two LAN ports, or between eth0 and LAN port
> > 1, or between eth1 and LAN port 0. The two network devices work
> > independently.
> >
> > An NIC with two LAN ports mode (daisy-chain mode) Ethernet driver
> > creates one net-device interface (eg: eth0), but the net-device
> > interface has two LAN ports. In this mode, a packet from one LAN port
> > will be either forwarded to net-device interface (eht0) if its
> > destination address matches MAC address of net-device interface
> > (eth0), or forwarded to other LAN port. A packet from net-device
> > interface (eth0) will be forwarded to a LAN port if its destination
> > address is learnt by L2 switch, or forwarded to both LAN ports if its
> > destination has not been learnt yet.
> >
> > An NIC with two LAN ports mode 2
> > This mode is similar to “An NIC with two LAN ports mode”. The
> > difference is that a packet from net-device interface (eth0) will be
> > always forwarded to both LAN ports. Learning function of L2 switch is
> > turned off in this mode. This means L2 switch will never learn the
> > source address of a packet. So, it always forward packets to both LAN
> > ports. This mode works like you have 2-port Ethernet hub.
> 
> So here you describe how the hardware can be used. Dual is two interfaces.
> Daisy-chain is what you get by taking those two interfaces and adding them to
> a bridge. The bridge then forwards frames between the interfaces and the CPU
> as needed, based on learning. And your third mode is the bridge always
> performs flooding.
> 
> A linux driver must follow the linux networking model. You cannot make up
> your own model. In the linux world, you model the external ports. The
> hardware always has two external ports, so you need to always have two
> netdev interfaces. To bridge packets between those two interfaces, you create
> a bridge and you add the interfaces to the bridge. That is the model you need
> to follow. switchdev gives you the API calls you need to implement this.

Thank you very much for your explanation.

I realize that we need to follow the Linux networking model.
I'll remove all descriptions about L2 switch or daisy-chain mode.

I'd like to modify Sunplus Ethernet driver to fulfill Linux networking model.
Here is my proposal:

SP7021 Ethernet supports 3 operation modes:
 - Dual Ethernet mode
   In this mode, driver creates two net-device interfaces. Each connects
   to PHY. There are two LAN ports totally.
   I am sorry that EMAC of SP7021 cannot support L2 switch functions
   of Linux switch-device model because it only has partial function of 
   switch.

 - One Ethernet mode
   In this mode, driver creates one net-device interface. It connects to
   to a PHY (There is only one LAN port).
   The LAN port is then connected to a 3-port Ethernet hub.
   The 3-port Ethernet hub is a hardware circuitry. All operations 
   (packet forwarding) are done by hardware. No software 
   intervention is needed. Actually, even just power-on, no software 
   running, two LAN ports of SP7021 work well as 2-port hub.

 - One Ethernet mode 2
   This is mode is similar to previous mode, but a bit different settings
   to the hub.

Please kindly comment if my proposal is feasible or not


> > > > +struct l2sw_common {
> > >
> > > Please change your prefix. l2sw is a common prefix, there are other
> > > silicon vendors using l2sw. I would suggest sp_l2sw or spl2sw.
> >
> > Ok, I'll modify two struct names in next patch as shown below:
> > l2sw_common --> sp_common
> > l2sw_mac --> sp_mac
> >
> > Should I also modify prefix of file name?
> 
> You need to modify the prefix everywhere you use it.  Function names,
> variable names, all symbols. Search and replace throughout the whole code.

Yes, I'll do in next patch.


> > > > +			return -EINVAL;
> > > > +		}
> > > > +	}
> > > > +
> > > > +	switch (cmd) {
> > > > +	case SIOCGMIIPHY:
> > > > +		if (comm->dual_nic && (strcmp(ifr->ifr_ifrn.ifrn_name, "eth1")
> > > > +==
> > > > +0))
> > >
> > > You cannot rely on the name, systemd has probably renamed it. If you
> > > have using phylib correctly, net_dev->phydev is what you want.
> >
> > Ok, I'll use name of the second net device to do compare, instead of
> > using fixed string "eth1", in next patch.
> 
> No. There are always two interfaces. You always have two netdev structures.
> Each netdev structure has a phydev. So use netdev->phydev.

Yes, I'll modify driver to use 'netdev->phydev'.


> This is another advantage of the Linux model. In your daisy chain mode, how
> do i control the two PHYs? How do i see one is up and one is down? How do i
> configure one to 10Half and the other 100Full?

No software intervention is needed.
Hardware circuitry of EMAC of Sunplus SP7021 does it well.
EMAC will communicate with PHY chips (via MDIO bus) automatically.
Actually, just giving power to SP7021, the two LAN ports act as 2-port 
Ethernet hub, forwarding packets between ports.


> > > > +int phy_cfg(struct l2sw_mac *mac) {
> > > > +	// Bug workaround:
> > > > +	// Flow-control of phy should be enabled. L2SW IP flow-control will
> refer
> > > > +	// to the bit to decide to enable or disable flow-control.
> > > > +	mdio_write(mac->comm->phy1_addr, 4,
> > > mdio_read(mac->comm->phy1_addr, 4) | (1 << 10));
> > > > +	mdio_write(mac->comm->phy2_addr, 4,
> > > mdio_read(mac->comm->phy2_addr,
> > > > +4) | (1 << 10));
> > >
> > > This should be in the PHY driver. The MAC driver should never need
> > > to touch PHY registers.
> >
> > Sunplus Ethernet MAC integrates MDIO controller.
> > So Ethernet driver has MDIO- and PHY-related code.
> > To work-around a circuitry bug, we need to enable bit 10 of register 4
> > of PHY.
> > Where should we place the code?
> 
> The silicon is integrated, but it is still a collection of standard blocks. Linux
> models those blocks independently. There is a subsystem for the MAC, a
> subsystem for the MDIO bus master and a subsystem for the PHY. You register
> a driver with each of these subsystems. PHY drivers live in drivers/net/phy. Put
> a PHY driver in there, which includes this workaround.
> 
> > > > +static void mii_linkchange(struct net_device *netdev) { }
> > >
> > > Nothing to do? Seems very odd. Don't you need to tell the MAC it
> > > should do 10Mbps or 100Mbps? What about pause?
> >
> > No, hardware does it automatically.
> > Sunplus MAC integrates MDIO controller.
> > It reads PHY status and set MAC automatically.
> 
> The PHY is external? So you have no idea what PHY that is? It could be a
> Marvell PHY, a microchip PHY, an Atheros PHY. Often PHYs have pages. In order
> to read the temperature sensor you change the page, read a register, and then
> hopefully change the page back again. If the PHY supports Fibre as well as
> copper, it can put the fibre registers in a second page. The PHY driver knows
> about this, it will flip the pages as needed. The phylib core has a mutex, so that
> only one operation happens at a time. So a page flip does not happen
> unexpectedly.
>
> Your MAC hardware does not take this mutex. It has no idea what page is
> selected when it reads registers. Instead of getting the basic mode register, it
> could get the LED control register...
> 
> The MAC should never directly access the PHY. Please disable this hardware,
> and use the mii_linkchange callback to configure the MAC.

Yes, the PHYs are external. SP7021 are connected to two ICPlus IP101 PHY.
EMAC of SP7021 communicates with PHY via MDIO bus automatically after 
It is setup and enabled. Sorry that the function cannot be disabled.

The mentioned bug is not a bug of PHY, but a hardware bug of EMAC.
SP7021 EMAC will read bit 10 (pause bit) of register 4 of PHY and then set
pause mode of EMAC itself. At initial, bit 10 of register 4 of PHY is 0. This
results in pause mode of EMAC be turned off. Due to timing issue, setting 
bit 10 on PHY driver is not feasible, we need to set it on EMAC driver right 
just after MAC is enabled.


> > > So the MAC does not support pause? I'm then confused about phy_cfg().
> 
> > Yes, MAC supports pause. MAC (hardware) takes care of pause
> > automatically.
> >
> > Should I remove the two lines?
> 
> Yes.

Yes, I'll remove them in next patch.


> And you need to configure the MAC based on the results of the auto-neg.
> 
> 	Andrew

Andrew Lunn Nov. 8, 2021, 1:15 p.m. UTC | #20

> SP7021 Ethernet supports 3 operation modes:
>  - Dual Ethernet mode
>    In this mode, driver creates two net-device interfaces. Each connects
>    to PHY. There are two LAN ports totally.
>    I am sorry that EMAC of SP7021 cannot support L2 switch functions
>    of Linux switch-device model because it only has partial function of 
>    switch.

This is fine.

> 
>  - One Ethernet mode
>    In this mode, driver creates one net-device interface. It connects to
>    to a PHY (There is only one LAN port).
>    The LAN port is then connected to a 3-port Ethernet hub.
>    The 3-port Ethernet hub is a hardware circuitry. All operations 
>    (packet forwarding) are done by hardware. No software 
>    intervention is needed. Actually, even just power-on, no software 
>    running, two LAN ports of SP7021 work well as 2-port hub.

We need to dig into the details of this mode. I would initially say
no, until we really do know it is impossible to do it correctly.  Even
if it is impossible to do it correctly, i'm still temped to reject
this mode.

How does spanning tree work? Who sends and receives the BPDU?

Is there PTP support? How do you send and receive the PTP frames?

Is IGMP snooping supported?

All of these have one thing in common, you need to be able to egress
frames out a specific port of the switch, and you need to know what
port a received frames ingressed on. If you can do that, you can
probably do proper support in Linux.

Is the datasheet available?


   Andrew

Wells Lu 呂芳騰 Nov. 8, 2021, 2:26 p.m. UTC | #21

> > SP7021 Ethernet supports 3 operation modes:
> >  - Dual Ethernet mode
> >    In this mode, driver creates two net-device interfaces. Each connects
> >    to PHY. There are two LAN ports totally.
> >    I am sorry that EMAC of SP7021 cannot support L2 switch functions
> >    of Linux switch-device model because it only has partial function of
> >    switch.
> 
> This is fine.

Thanks a lot!


> >
> >  - One Ethernet mode
> >    In this mode, driver creates one net-device interface. It connects to
> >    to a PHY (There is only one LAN port).
> >    The LAN port is then connected to a 3-port Ethernet hub.
> >    The 3-port Ethernet hub is a hardware circuitry. All operations
> >    (packet forwarding) are done by hardware. No software
> >    intervention is needed. Actually, even just power-on, no software
> >    running, two LAN ports of SP7021 work well as 2-port hub.
> 
> We need to dig into the details of this mode. I would initially say no, until we
> really do know it is impossible to do it correctly.  Even if it is impossible to do
> it correctly, i'm still temped to reject this mode.
> 
> How does spanning tree work? Who sends and receives the BPDU?
> 
> Is there PTP support? How do you send and receive the PTP frames?
> 
> Is IGMP snooping supported?
> 
> All of these have one thing in common, you need to be able to egress frames
> out a specific port of the switch, and you need to know what port a received
> frames ingressed on. If you can do that, you can probably do proper support in
> Linux.

The "L2 switch" is a very simple witch. It has 3 ports: CPU, LAN port 0 and LAN 
port 1. A packet is always forwarded to other two ports if source-address (of MAC)
learning function is off, or forwarded to one of the two ports if source-address
learning function is on and source address is learnt (recorded by switch).

The switch will not recognize type of packets, regardless BPDU, PTP or any other
packets. If turning off source-address learning function, it works like an Ethernet
plus a 2-port hub.


> Is the datasheet available?

Yes, refer to on-line document of SP7021:
https://sunplus-tibbo.atlassian.net/wiki/spaces/doc/pages/462553090/15.+Ethernet+Switch


> 
>    Andrew

Andrew Lunn Nov. 8, 2021, 2:52 p.m. UTC | #22

> The switch will not recognize type of packets, regardless BPDU, PTP or any other
> packets. If turning off source-address learning function, it works like an Ethernet
> plus a 2-port hub.

So without STP, there is no way to stop an loop, and a broadcast storm
taking down your network?

Looking at the TX descriptor, there are two bits:

          [18]: force forward to port 0
          [19]: force forward to port 1

When the switch is enabled, can these two bits be used?

In the RX descriptor there is:

pkt_sp:
          000: from port0
          001: from port1
          110: soc0 loopback
          101: soc1 loopback

Are these bits used when the switch is enabled?

0.31 port control 1 (port cntl1) blocking state seems to have what you
need for STP.

    Andrew

Wells Lu 呂芳騰 Nov. 8, 2021, 4:47 p.m. UTC | #23

> > The switch will not recognize type of packets, regardless BPDU, PTP or
> > any other packets. If turning off source-address learning function, it
> > works like an Ethernet plus a 2-port hub.
> 
> So without STP, there is no way to stop an loop, and a broadcast storm taking
> down your network?

Do you mean connecting two PHY ports to the same LAN? We never 
connect two PHY ports to the same LAN (or hub). I never think of this 
loop problem. I thought only WAN has the loop problem.

The switch has some kinds of flow control, refer to 0.2 "Flow control threshold"
and 0.3 "CPU port flow control threshold". It will drop extra packets.
How an Ethernet hub take care of this situation?
Is that reasonable to connect two ports of an Ethernet hub together?


> Looking at the TX descriptor, there are two bits:
> 
>           [18]: force forward to port 0
>           [19]: force forward to port 1
> 
> When the switch is enabled, can these two bits be used?

Yes, for example, when bit 19 of TX descriptor is enabled, a packet from CPU 
port is forwarded to LAN port 0 forcibly.


> In the RX descriptor there is:
> 
> pkt_sp:
>           000: from port0
>           001: from port1
>           110: soc0 loopback
>           101: soc1 loopback
> 
> Are these bits used when the switch is enabled?

Yes, E- MAC driver uses these bits to tell where a packet comes from.
Note that soc1 port (CPU port) has been removed in this chip.


> 0.31 port control 1 (port cntl1) blocking state seems to have what you need for
> STP.

From document, if bit 17 or bit 16 of port_cntl1 register is set, only RMC 
packets will be forwarded to other LAN port. I am not sure whether 
enabling the bits helps the issue. Should I enable the bits?
Sorry, I don't know what is a RMC packet?
Could you please teach me?


>     Andrew

Thank you for review.

Andrew Lunn Nov. 8, 2021, 5:32 p.m. UTC | #24

On Mon, Nov 08, 2021 at 04:47:34PM +0000, Wells Lu 呂芳騰 wrote:
> > > The switch will not recognize type of packets, regardless BPDU, PTP or
> > > any other packets. If turning off source-address learning function, it
> > > works like an Ethernet plus a 2-port hub.
> > 
> > So without STP, there is no way to stop an loop, and a broadcast storm taking
> > down your network?
> 
> Do you mean connecting two PHY ports to the same LAN? We never 
> connect two PHY ports to the same LAN (or hub). I never think of this 
> loop problem. I thought only WAN has the loop problem.

Any Ethernet network can have a loop. Often loops a deliberate because
they give redundancy. STP will detect this loop, and somewhere in the
network one of the switches will block traffic to break the loop. But
if something in the network breaks, the port can be unblocked to allow
traffic to flow, redundancy. Well behaved switches should always
implement STP.

> How an Ethernet hub take care of this situation?

STP. Run tcpdump on your network. Depending on how your network is
configured, you might see BPDU from your building switches.

> Is that reasonable to connect two ports of an Ethernet hub together?

It is not just together. You cannot guarantee any Ethernet network is
a tree. You could connect the two ports to two different hubs, but
those hubs are connected together, and so you get a loop.

> > Looking at the TX descriptor, there are two bits:
> > 
> >           [18]: force forward to port 0
> >           [19]: force forward to port 1
> > 
> > When the switch is enabled, can these two bits be used?
> 
> Yes, for example, when bit 19 of TX descriptor is enabled, a packet from CPU 
> port is forwarded to LAN port 0 forcibly.
> 
> 
> > In the RX descriptor there is:
> > 
> > pkt_sp:
> >           000: from port0
> >           001: from port1
> >           110: soc0 loopback
> >           101: soc1 loopback
> > 
> > Are these bits used when the switch is enabled?
> 
> Yes, E- MAC driver uses these bits to tell where a packet comes from.
> Note that soc1 port (CPU port) has been removed in this chip.
 
Right. So you can have two netdev when in L2 switch mode.

You need to think about the Linux model some more. In linux,
networking hardware is there to accelerate what the Linux stack can do
in software. Take for example a simple SoC will have two Ethernet
interfaces. You can perform software bridging on those two interfaces:

ip link add name br0 type bridge
ip link set dev br0 up
ip link set dev eth0 master br0
ip link set dev eth1 master br0

The software bridge will decided which interface to send a packet
out. The software will perform learning etc.

You can use your dual MAC setup exactly like this. But you can also go
further. You can use the hardware to accelerate switching packets
between eth0 and eth1. But also Linux can still send packets out
specific ports using these bits. The software bridge and the hardware
bridge work together. This is the correct way to do this in Linux.

> Sorry, I don't know what is a RMC packet?

Sorry, i have no idea.

       Andrew

Wells Lu 呂芳騰 Nov. 9, 2021, 2:39 p.m. UTC | #25

> On Mon, Nov 08, 2021 at 04:47:34PM +0000, Wells Lu 呂芳騰 wrote:
> > > > The switch will not recognize type of packets, regardless BPDU,
> > > > PTP or any other packets. If turning off source-address learning
> > > > function, it works like an Ethernet plus a 2-port hub.
> > >
> > > So without STP, there is no way to stop an loop, and a broadcast
> > > storm taking down your network?
> >
> > Do you mean connecting two PHY ports to the same LAN? We never connect
> > two PHY ports to the same LAN (or hub). I never think of this loop
> > problem. I thought only WAN has the loop problem.
> 
> Any Ethernet network can have a loop. Often loops a deliberate because they
> give redundancy. STP will detect this loop, and somewhere in the network one
> of the switches will block traffic to break the loop. But if something in the
> network breaks, the port can be unblocked to allow traffic to flow, redundancy.
> Well behaved switches should always implement STP.

I don't know how to implement STP in L2 switch like SP7021.
How about one NIC + 2-port simple frame-flooding hub?
Someone told me that some low-cost Ethernet hub just does frame-flooding 
to other ports. Let users take care of use.
If this is acceptable, I'd like to have Ethernet of SP7021 have two operation 
modes:
 - Dual NIC mode
 - Single NIC with 2-port frame-flooding hub mode

If this is not acceptable, can I, instead, implement the two operation modes:
 - Dual NIC mode
 - Single NIC mode

RMII pins of PHY ports of SP7021 are multiplexable. I'd like to switch RMII 
pins of the second PHY for other use if single NIC mode is used.
In fact, some SP7021 boards have dual Ethernet and some have only one
Ethernet. We really need the two operation modes.


> > How an Ethernet hub take care of this situation?
> 
> STP. Run tcpdump on your network. Depending on how your network is
> configured, you might see BPDU from your building switches.

Thanks a lot. I understand.

> > Is that reasonable to connect two ports of an Ethernet hub together?
> 
> It is not just together. You cannot guarantee any Ethernet network is a tree. You
> could connect the two ports to two different hubs, but those hubs are
> connected together, and so you get a loop.

Thanks for explanation. I got it.

> > > Looking at the TX descriptor, there are two bits:
> > >
> > >           [18]: force forward to port 0
> > >           [19]: force forward to port 1
> > >
> > > When the switch is enabled, can these two bits be used?
> >
> > Yes, for example, when bit 19 of TX descriptor is enabled, a packet
> > from CPU port is forwarded to LAN port 0 forcibly.
> >
> >
> > > In the RX descriptor there is:
> > >
> > > pkt_sp:
> > >           000: from port0
> > >           001: from port1
> > >           110: soc0 loopback
> > >           101: soc1 loopback
> > >
> > > Are these bits used when the switch is enabled?
> >
> > Yes, E- MAC driver uses these bits to tell where a packet comes from.
> > Note that soc1 port (CPU port) has been removed in this chip.
> 
> Right. So you can have two netdev when in L2 switch mode.
> 
> You need to think about the Linux model some more. In linux, networking
> hardware is there to accelerate what the Linux stack can do in software. Take
> for example a simple SoC will have two Ethernet interfaces. You can perform
> software bridging on those two interfaces:
> 
> ip link add name br0 type bridge
> ip link set dev br0 up
> ip link set dev eth0 master br0
> ip link set dev eth1 master br0
> 
> The software bridge will decided which interface to send a packet out. The
> software will perform learning etc.
> 
> You can use your dual MAC setup exactly like this. But you can also go further.
> You can use the hardware to accelerate switching packets between eth0 and
> eth1. But also Linux can still send packets out specific ports using these bits.
> The software bridge and the hardware bridge work together. This is the correct
> way to do this in Linux.

Sorry, I am not capable to do that.
I need to study more about Linux switch device, L2 switch of SP7021, procotols,...
So what I can use now is pure software bridge for dual Ethernet case.

> > Sorry, I don't know what is a RMC packet?
>
> Sorry, i have no idea.

After looking up some data, I find RMC means reserved multi-cast.
RMC packets means packets with DA = 0x0180c2000000, 0x0180c2000002 ~ 0x0180c200000f,
except the PAUSE packet (DA = 0x0180c2000001)


>        Andrew

Thank you for your review.

Andrew Lunn Nov. 9, 2021, 3:32 p.m. UTC | #26

> I don't know how to implement STP in L2 switch like SP7021.

That is the nice thing about using Linux. It already knows how to
implement STP. The bridge will do it for you. You just need to add the
callbacks in the driver which are needed. Please take a look at other
switchdev drivers.

> If this is acceptable, I'd like to have Ethernet of SP7021 have two operation 
> modes:
>  - Dual NIC mode
>  - Single NIC with 2-port frame-flooding hub mode

No, sorry. Do it correctly, or do not do it. Please start with a clean
driver doing Dual NIC mode. You can add L2 support later, once you
have done the research to understand switchdev, etc.

> RMII pins of PHY ports of SP7021 are multiplexable. I'd like to switch RMII 
> pins of the second PHY for other use if single NIC mode is used.
> In fact, some SP7021 boards have dual Ethernet and some have only one
> Ethernet. We really need the two operation modes.

Only using a subset of ports in a switch is common. The common binding
for DSA switches is described in:

Documentation/devicetree/bindings/net/dsa/dsa.yaml and for example
Documentation/devicetree/bindings/net/dsa/hirschmann,hellcreek.yaml is
a memory mapped switch. Notice the reg numbers:

           ethernet-ports {
                #address-cells = <1>;
                #size-cells = <0>;

                port@0 {
                    reg = <0>;
                    label = "cpu";
                    ethernet = <&gmac0>;
                };

                port@2 {
                    reg = <2>;
                    label = "lan0";
                    phy-handle = <&phy1>;
                };

reg = <1> is missing in this example. Port 1 of the switch is not
used. You can do the same with a 2 port switch, when you don't want to
make use of a port. Just don't list it in DT.

> After looking up some data, I find RMC means reserved multi-cast.
> RMC packets means packets with DA = 0x0180c2000000, 0x0180c2000002 ~ 0x0180c200000f,
> except the PAUSE packet (DA = 0x0180c2000001)

Ah, good. BPDUs use 01:80:C2:00:00:00. So they will be passed when the
port is in blocking mode. PTP uses 01:80:C2:00:00:0E. So the hardware
designers appear to of designed a proper L2 switch with everything you
need for a managed switch. What is missing is software. The more i
learn about this hardware, the more i've convinced you need to write
proper Linux support for it, not your mode hacks.

    Andrew

Wells Lu 呂芳騰 Nov. 9, 2021, 5:05 p.m. UTC | #27

> > I don't know how to implement STP in L2 switch like SP7021.
> 
> That is the nice thing about using Linux. It already knows how to implement
> STP. The bridge will do it for you. You just need to add the callbacks in the
> driver which are needed. Please take a look at other switchdev drivers.
> 
> > If this is acceptable, I'd like to have Ethernet of SP7021 have two
> > operation
> > modes:
> >  - Dual NIC mode
> >  - Single NIC with 2-port frame-flooding hub mode
> 
> No, sorry. Do it correctly, or do not do it. Please start with a clean driver doing
> Dual NIC mode. You can add L2 support later, once you have done the research
> to understand switchdev, etc.

Sorry, I will go with Dual NIC mode. I'll do a whole cleanup on driver for this.
Please kindly review again.

I need time to study more about switchdev and propose a plan to high
management of company. However, Sunplus is not a networking company, 
but targets on Linux-based industrial control, autonomous mobile robot, ...


> > RMII pins of PHY ports of SP7021 are multiplexable. I'd like to switch
> > RMII pins of the second PHY for other use if single NIC mode is used.
> > In fact, some SP7021 boards have dual Ethernet and some have only one
> > Ethernet. We really need the two operation modes.
> 
> Only using a subset of ports in a switch is common. The common binding for
> DSA switches is described in:
> 
> Documentation/devicetree/bindings/net/dsa/dsa.yaml and for example
> Documentation/devicetree/bindings/net/dsa/hirschmann,hellcreek.yaml is a
> memory mapped switch. Notice the reg numbers:
> 
>            ethernet-ports {
>                 #address-cells = <1>;
>                 #size-cells = <0>;
> 
>                 port@0 {
>                     reg = <0>;
>                     label = "cpu";
>                     ethernet = <&gmac0>;
>                 };
> 
>                 port@2 {
>                     reg = <2>;
>                     label = "lan0";
>                     phy-handle = <&phy1>;
>                 };
> 
> reg = <1> is missing in this example. Port 1 of the switch is not used. You can
> do the same with a 2 port switch, when you don't want to make use of a port.
> Just don't list it in DT.

Thank you for routing me to the document.
Now I know there are switch device examples in folder dsa/.
We can refer to them when we want to make a switch.


> > After looking up some data, I find RMC means reserved multi-cast.
> > RMC packets means packets with DA = 0x0180c2000000, 0x0180c2000002 ~
> > 0x0180c200000f, except the PAUSE packet (DA = 0x0180c2000001)
> 
> Ah, good. BPDUs use 01:80:C2:00:00:00. So they will be passed when the port is
> in blocking mode. PTP uses 01:80:C2:00:00:0E. So the hardware designers
> appear to of designed a proper L2 switch with everything you need for a
> managed switch. What is missing is software. The more i learn about this
> hardware, the more i've convinced you need to write proper Linux support for
> it, not your mode hacks.
> 
>     Andrew

Thanks for confirming that the L2 switch is good enough for switch device.
Actually, the IP was licensed from other company long ago. We don’t know 
all details.

Pavel Skripkin Nov. 14, 2021, 7:19 p.m. UTC | #28

Hi, Wells!

On 11/3/21 14:02, Wells Lu wrote:

[code snip]

> +		if (comm->dual_nic) {
> +			struct net_device *net_dev2 = mac->next_netdev;
> +
> +			if (!netif_running(net_dev2)) {
> +				mac_hw_stop(mac);
> +
> +				mac2 = netdev_priv(net_dev2);
> +

(*)

> +				// unregister and free net device.
> +				unregister_netdev(net_dev2);
> +				free_netdev(net_dev2);
> +				mac->next_netdev = NULL;
> +				pr_info(" Unregistered and freed net device \"eth1\"!\n");
> +
> +				comm->dual_nic = 0;
> +				mac_switch_mode(mac);
> +				rx_mode_set(net_dev);
> +				mac_hw_addr_del(mac2);
> +

mac2 is net_dev2 private data (*), so it will become freed after 
free_netdev() call.

FWIW the latest `smatch` should warn about this type of bugs.

> +				// If eth0 is up, turn on lan 0 and 1 when
> +				// switching to daisy-chain mode.
> +				if (comm->enable & 0x1)
> +					comm->enable = 0x3;

[code snip]

> +static int l2sw_remove(struct platform_device *pdev)
> +{
> +	struct net_device *net_dev;
> +	struct net_device *net_dev2;
> +	struct l2sw_mac *mac;
> +
> +	net_dev = platform_get_drvdata(pdev);
> +	if (!net_dev)
> +		return 0;
> +	mac = netdev_priv(net_dev);
> +
> +	// Unregister and free 2nd net device.
> +	net_dev2 = mac->next_netdev;
> +	if (net_dev2) {
> +		unregister_netdev(net_dev2);
> +		free_netdev(net_dev2);
> +	}
> +

Is it save here to free mac->next_netdev before unregistering "parent" 
netdev? I haven't checked the whole code, just asking :)

> +	sysfs_remove_group(&pdev->dev.kobj, &l2sw_attribute_group);
> +
> +	mac->comm->enable = 0;
> +	soc0_stop(mac);
> +
> +	napi_disable(&mac->comm->rx_napi);
> +	netif_napi_del(&mac->comm->rx_napi);
> +	napi_disable(&mac->comm->tx_napi);
> +	netif_napi_del(&mac->comm->tx_napi);
> +
> +	mdio_remove(net_dev);
> +
> +	// Unregister and free 1st net device.
> +	unregister_netdev(net_dev);
> +	free_netdev(net_dev);
> +
> +	clk_disable(mac->comm->clk);
> +
> +	// Free 'common' area.
> +	kfree(mac->comm);

Same here with `mac`.

> +	return 0;
> +}


I haven't read the whole thread, i am sorry if these questions were 
already discussed.



With regards,
Pavel Skripkin

Wells Lu 呂芳騰 Nov. 17, 2021, 9:28 a.m. UTC | #29

Hi Pavel,

> Hi, Wells!
> 
> On 11/3/21 14:02, Wells Lu wrote:
> 
> [code snip]
> 
> > +		if (comm->dual_nic) {
> > +			struct net_device *net_dev2 = mac->next_netdev;
> > +
> > +			if (!netif_running(net_dev2)) {
> > +				mac_hw_stop(mac);
> > +
> > +				mac2 = netdev_priv(net_dev2);
> > +
> 
> (*)
> 
> > +				// unregister and free net device.
> > +				unregister_netdev(net_dev2);
> > +				free_netdev(net_dev2);
> > +				mac->next_netdev = NULL;
> > +				pr_info(" Unregistered and freed net device \"eth1\"!\n");
> > +
> > +				comm->dual_nic = 0;
> > +				mac_switch_mode(mac);
> > +				rx_mode_set(net_dev);
> > +				mac_hw_addr_del(mac2);
> > +
> 
> mac2 is net_dev2 private data (*), so it will become freed after
> free_netdev() call.
> 
> FWIW the latest `smatch` should warn about this type of bugs.

Yes, this is indeed a bug.
But the code paragraph has been removed thoroughly in [PATCH v2].


> > +				// If eth0 is up, turn on lan 0 and 1 when
> > +				// switching to daisy-chain mode.
> > +				if (comm->enable & 0x1)
> > +					comm->enable = 0x3;
> 
> [code snip]
> 
> > +static int l2sw_remove(struct platform_device *pdev) {
> > +	struct net_device *net_dev;
> > +	struct net_device *net_dev2;
> > +	struct l2sw_mac *mac;
> > +
> > +	net_dev = platform_get_drvdata(pdev);
> > +	if (!net_dev)
> > +		return 0;
> > +	mac = netdev_priv(net_dev);
> > +
> > +	// Unregister and free 2nd net device.
> > +	net_dev2 = mac->next_netdev;
> > +	if (net_dev2) {
> > +		unregister_netdev(net_dev2);
> > +		free_netdev(net_dev2);
> > +	}
> > +
> 
> Is it save here to free mac->next_netdev before unregistering "parent"
> netdev? I haven't checked the whole code, just asking :)

Yes, I think it is save.
netdev2 should be unregistered and freed before net_dev.
If net_dev is unregistered and freed in advance,
mac->next_netdev becomes danger because 'mac' has been freed.


> > +	sysfs_remove_group(&pdev->dev.kobj, &l2sw_attribute_group);
> > +
> > +	mac->comm->enable = 0;
> > +	soc0_stop(mac);
> > +
> > +	napi_disable(&mac->comm->rx_napi);
> > +	netif_napi_del(&mac->comm->rx_napi);
> > +	napi_disable(&mac->comm->tx_napi);
> > +	netif_napi_del(&mac->comm->tx_napi);
> > +
> > +	mdio_remove(net_dev);
> > +
> > +	// Unregister and free 1st net device.
> > +	unregister_netdev(net_dev);
> > +	free_netdev(net_dev);
> > +
> > +	clk_disable(mac->comm->clk);
> > +
> > +	// Free 'common' area.
> > +	kfree(mac->comm);
> 
> Same here with `mac`.

This is indeed a bug.
But the statement, Kfree(mac->comm);, has been removed in [PATCH v2].
In [PATCH v2], structure data 'mac->comm' is allocated by
devm_kzalloc(). No more need to free it here.


> > +	return 0;
> > +}
> 
> 
> I haven't read the whole thread, i am sorry if these questions were already discussed.
> 
> 
> 
> With regards,
> Pavel Skripkin


Thank you very much for your review!

Best regards,
Wells Lu

Patch

diff --git a/MAINTAINERS b/MAINTAINERS
index 4669c16..ca676ec 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -18006,6 +18006,7 @@  L:	[e-mail]
 S:	Maintained
 W:	https://sunplus-tibbo.atlassian.net/wiki/spaces/doc/overview
 F:	Documentation/devicetree/bindings/net/sunplus,sp7021-l2sw.yaml
+F:	drivers/net/ethernet/sunplus/
 
 SUPERH
 M:	Yoshinori Sato <[e-mail]>
diff --git a/drivers/net/ethernet/Kconfig b/drivers/net/ethernet/Kconfig
index 412ae3e..0084852 100644
--- a/drivers/net/ethernet/Kconfig
+++ b/drivers/net/ethernet/Kconfig
@@ -176,6 +176,7 @@  source "drivers/net/ethernet/smsc/Kconfig"
 source "drivers/net/ethernet/socionext/Kconfig"
 source "drivers/net/ethernet/stmicro/Kconfig"
 source "drivers/net/ethernet/sun/Kconfig"
+source "drivers/net/ethernet/sunplus/Kconfig"
 source "drivers/net/ethernet/synopsys/Kconfig"
 source "drivers/net/ethernet/tehuti/Kconfig"
 source "drivers/net/ethernet/ti/Kconfig"
diff --git a/drivers/net/ethernet/Makefile b/drivers/net/ethernet/Makefile
index aaa5078..e4ce162 100644
--- a/drivers/net/ethernet/Makefile
+++ b/drivers/net/ethernet/Makefile
@@ -87,6 +87,7 @@  obj-$(CONFIG_NET_VENDOR_SMSC) += smsc/
 obj-$(CONFIG_NET_VENDOR_SOCIONEXT) += socionext/
 obj-$(CONFIG_NET_VENDOR_STMICRO) += stmicro/
 obj-$(CONFIG_NET_VENDOR_SUN) += sun/
+obj-$(CONFIG_NET_VENDOR_SUNPLUS) += sunplus/
 obj-$(CONFIG_NET_VENDOR_TEHUTI) += tehuti/
 obj-$(CONFIG_NET_VENDOR_TI) += ti/
 obj-$(CONFIG_NET_VENDOR_TOSHIBA) += toshiba/
diff --git a/drivers/net/ethernet/sunplus/Kconfig b/drivers/net/ethernet/sunplus/Kconfig
new file mode 100644
index 0000000..a9e3a4c
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/Kconfig
@@ -0,0 +1,20 @@ 
+# SPDX-License-Identifier: GPL-2.0
+#
+# Sunplus Ethernet device configuration
+#
+
+config NET_VENDOR_SUNPLUS
+	tristate "Sunplus Dual 10M/100M Ethernet (with L2 switch) devices"
+	depends on ETHERNET && SOC_SP7021
+	select PHYLIB
+	select PINCTRL_SPPCTL
+	select COMMON_CLK_SP7021
+	select RESET_SUNPLUS
+	select NVMEM_SUNPLUS_OCOTP
+	help
+	  If you have Sunplus dual 10M/100M Ethernet (with L2 switch)
+	  devices, say Y.
+	  The network device supports dual 10M/100M Ethernet interfaces,
+	  or one 10/100M Ethernet interface with two LAN ports.
+	  To compile this driver as a module, choose M here.  The module
+	  will be called sp_l2sw.
diff --git a/drivers/net/ethernet/sunplus/Makefile b/drivers/net/ethernet/sunplus/Makefile
new file mode 100644
index 0000000..b401cec
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/Makefile
@@ -0,0 +1,6 @@ 
+# SPDX-License-Identifier: GPL-2.0
+#
+# Makefile for the Sunplus network device drivers.
+#
+obj-$(CONFIG_NET_VENDOR_SUNPLUS) += sp_l2sw.o
+sp_l2sw-objs := l2sw_driver.o l2sw_int.o l2sw_hal.o l2sw_desc.o l2sw_mac.o l2sw_mdio.o
diff --git a/drivers/net/ethernet/sunplus/l2sw_define.h b/drivers/net/ethernet/sunplus/l2sw_define.h
new file mode 100644
index 0000000..c1049c5
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_define.h
@@ -0,0 +1,221 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#ifndef __L2SW_DEFINE_H__
+#define __L2SW_DEFINE_H__
+
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/slab.h>
+#include <linux/errno.h>
+#include <linux/types.h>
+#include <linux/interrupt.h>
+#include <linux/kdev_t.h>
+#include <linux/in.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/ip.h>
+#include <linux/tcp.h>
+#include <linux/skbuff.h>
+#include <linux/ethtool.h>
+#include <linux/platform_device.h>
+#include <linux/phy.h>
+#include <linux/mii.h>
+#include <linux/if_vlan.h>
+#include <linux/io.h>
+#include <linux/dma-mapping.h>
+#include <linux/of_address.h>
+#include <linux/of_mdio.h>
+
+#undef pr_fmt
+#define pr_fmt(fmt)     "[L2SW]" fmt
+
+//define MAC interrupt status bit
+#define MAC_INT_DAISY_MODE_CHG          BIT(31)
+#define MAC_INT_IP_CHKSUM_ERR           BIT(23)
+#define MAC_INT_WDOG_TIMER1_EXP         BIT(22)
+#define MAC_INT_WDOG_TIMER0_EXP         BIT(21)
+#define MAC_INT_INTRUDER_ALERT          BIT(20)
+#define MAC_INT_PORT_ST_CHG             BIT(19)
+#define MAC_INT_BC_STORM                BIT(18)
+#define MAC_INT_MUST_DROP_LAN           BIT(17)
+#define MAC_INT_GLOBAL_QUE_FULL         BIT(16)
+#define MAC_INT_TX_SOC_PAUSE_ON         BIT(15)
+#define MAC_INT_RX_SOC_QUE_FULL         BIT(14)
+#define MAC_INT_TX_LAN1_QUE_FULL        BIT(9)
+#define MAC_INT_TX_LAN0_QUE_FULL        BIT(8)
+#define MAC_INT_RX_L_DESCF              BIT(7)
+#define MAC_INT_RX_H_DESCF              BIT(6)
+#define MAC_INT_RX_DONE_L               BIT(5)
+#define MAC_INT_RX_DONE_H               BIT(4)
+#define MAC_INT_TX_DONE_L               BIT(3)
+#define MAC_INT_TX_DONE_H               BIT(2)
+#define MAC_INT_TX_DES_ERR              BIT(1)
+#define MAC_INT_RX_DES_ERR              BIT(0)
+
+#define MAC_INT_RX                      (MAC_INT_RX_DONE_H | MAC_INT_RX_DONE_L | \
+					MAC_INT_RX_DES_ERR)
+#define MAC_INT_TX                      (MAC_INT_TX_DONE_L | MAC_INT_TX_DONE_H | \
+					MAC_INT_TX_DES_ERR)
+#define MAC_INT_MASK_DEF                (MAC_INT_DAISY_MODE_CHG | MAC_INT_IP_CHKSUM_ERR | \
+					MAC_INT_WDOG_TIMER1_EXP | MAC_INT_WDOG_TIMER0_EXP | \
+					MAC_INT_INTRUDER_ALERT | MAC_INT_BC_STORM | \
+					MAC_INT_MUST_DROP_LAN | MAC_INT_GLOBAL_QUE_FULL | \
+					MAC_INT_TX_SOC_PAUSE_ON | MAC_INT_RX_SOC_QUE_FULL | \
+					MAC_INT_TX_LAN1_QUE_FULL | MAC_INT_TX_LAN0_QUE_FULL | \
+					MAC_INT_RX_L_DESCF | MAC_INT_RX_H_DESCF)
+
+/*define port ability*/
+#define PORT_ABILITY_LINK_ST_P1         BIT(25)
+#define PORT_ABILITY_LINK_ST_P0         BIT(24)
+
+/*define PHY command bit*/
+#define PHY_WT_DATA_MASK                0xffff0000
+#define PHY_RD_CMD                      0x00004000
+#define PHY_WT_CMD                      0x00002000
+#define PHY_REG_MASK                    0x00001f00
+#define PHY_ADR_MASK                    0x0000001f
+
+/*define PHY status bit*/
+#define PHY_RD_DATA_MASK                0xffff0000
+#define PHY_RD_RDY                      BIT(1)
+#define PHY_WT_DONE                     BIT(0)
+
+/*define other register bit*/
+#define RX_MAX_LEN_MASK                 0x00011000
+#define ROUTE_MODE_MASK                 0x00000060
+#define POK_INT_THS_MASK                0x000E0000
+#define VLAN_TH_MASK                    0x00000007
+
+/*define tx descriptor bit*/
+#define OWN_BIT                         BIT(31)
+#define FS_BIT                          BIT(25)
+#define LS_BIT                          BIT(24)
+#define LEN_MASK                        0x000007FF
+#define PKTSP_MASK                      0x00007000
+#define PKTSP_PORT1                     0x00001000
+#define TO_VLAN_MASK                    0x0003F000
+#define TO_VLAN_GROUP1                  0x00002000
+
+#define EOR_BIT                         BIT(31)
+
+/*define rx descriptor bit*/
+#define ERR_CODE                        (0xf << 26)
+#define RX_TCP_UDP_CHKSUM_BIT           BIT(23)
+#define RX_IP_CHKSUM_BIT                BIT(18)
+
+#define OWC_BIT                         BIT(31)
+#define TXOK_BIT                        BIT(26)
+#define LNKF_BIT                        BIT(25)
+#define BUR_BIT                         BIT(22)
+#define TWDE_BIT                        BIT(20)
+#define CC_MASK                         0x000f0000
+#define TBE_MASK                        0x00070000
+
+// Address table search
+#define MAC_ADDR_LOOKUP_IDLE            BIT(2)
+#define MAC_SEARCH_NEXT_ADDR            BIT(1)
+#define MAC_BEGIN_SEARCH_ADDR           BIT(0)
+
+// Address table search
+#define MAC_HASK_LOOKUP_ADDR_MASK       (0x3ff << 22)
+#define MAC_AT_TABLE_END                BIT(1)
+#define MAC_AT_DATA_READY               BIT(0)
+
+#define MAC_PHY_ADDR                    0x01	/* define by hardware */
+
+/*config descriptor*/
+#define TX_DESC_NUM                     16
+#define MAC_GUARD_DESC_NUM              2
+#define RX_QUEUE0_DESC_NUM              16
+#define RX_QUEUE1_DESC_NUM              16
+#define TX_DESC_QUEUE_NUM               1
+#define RX_DESC_QUEUE_NUM               2
+
+#define MAC_TX_BUFF_SIZE                1536
+#define MAC_RX_LEN_MAX                  2047
+
+#define DESC_ALIGN_BYTE                 32
+#define RX_OFFSET                       0
+#define TX_OFFSET                       0
+
+#define ETHERNET_MAC_ADDR_LEN           6
+
+struct mac_desc {
+	u32 cmd1;
+	u32 cmd2;
+	u32 addr1;
+	u32 addr2;
+};
+
+struct skb_info {
+	struct sk_buff *skb;
+	u32 mapping;
+	u32 len;
+};
+
+struct l2sw_common {
+	struct net_device *net_dev;
+	struct platform_device *pdev;
+	int dual_nic;
+	int sa_learning;
+
+	void *desc_base;
+	dma_addr_t desc_dma;
+	s32 desc_size;
+	struct clk *clk;
+	struct reset_control *rstc;
+	int irq;
+
+	struct mac_desc *rx_desc[RX_DESC_QUEUE_NUM];
+	struct skb_info *rx_skb_info[RX_DESC_QUEUE_NUM];
+	u32 rx_pos[RX_DESC_QUEUE_NUM];
+	u32 rx_desc_num[RX_DESC_QUEUE_NUM];
+	u32 rx_desc_buff_size;
+
+	struct mac_desc *tx_desc;
+	struct skb_info tx_temp_skb_info[TX_DESC_NUM];
+	u32 tx_done_pos;
+	u32 tx_pos;
+	u32 tx_desc_full;
+
+	struct mii_bus *mii_bus;
+	struct phy_device *phy_dev;
+
+	struct napi_struct rx_napi;
+	struct napi_struct tx_napi;
+
+	spinlock_t rx_lock;      // spinlock for accessing rx buffer
+	spinlock_t tx_lock;      // spinlock for accessing tx buffer
+	spinlock_t ioctl_lock;   // spinlock for ioctl operations
+	struct mutex store_mode; // mutex for dynamic mode change
+
+	struct device_node *phy1_node;
+	struct device_node *phy2_node;
+	u8 phy1_addr;
+	u8 phy2_addr;
+
+	u8 enable;
+};
+
+struct l2sw_mac {
+	struct platform_device *pdev;
+	struct net_device *net_dev;
+	struct l2sw_common *comm;
+	struct net_device *next_netdev;
+
+	struct net_device_stats dev_stats;
+
+	u8 mac_addr[ETHERNET_MAC_ADDR_LEN];
+
+	u8 lan_port;
+	u8 to_vlan;
+	u8 cpu_port;
+	u8 vlan_id;
+};
+
+#endif
diff --git a/drivers/net/ethernet/sunplus/l2sw_desc.c b/drivers/net/ethernet/sunplus/l2sw_desc.c
new file mode 100644
index 0000000..345738f
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_desc.c
@@ -0,0 +1,233 @@ 
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#include "l2sw_desc.h"
+#include "l2sw_define.h"
+
+void rx_descs_flush(struct l2sw_common *comm)
+{
+	u32 i, j;
+	struct mac_desc *rx_desc;
+	struct skb_info *rx_skbinfo;
+
+	for (i = 0; i < RX_DESC_QUEUE_NUM; i++) {
+		rx_desc = comm->rx_desc[i];
+		rx_skbinfo = comm->rx_skb_info[i];
+		for (j = 0; j < comm->rx_desc_num[i]; j++) {
+			rx_desc[j].addr1 = rx_skbinfo[j].mapping;
+			rx_desc[j].cmd2 = (j == comm->rx_desc_num[i] - 1) ?
+					  EOR_BIT | comm->rx_desc_buff_size :
+					  comm->rx_desc_buff_size;
+			wmb();	// Set OWN_BIT after other fields are ready.
+			rx_desc[j].cmd1 = OWN_BIT;
+		}
+	}
+}
+
+void tx_descs_clean(struct l2sw_common *comm)
+{
+	u32 i;
+	s32 buflen;
+
+	if (!comm->tx_desc)
+		return;
+
+	for (i = 0; i < TX_DESC_NUM; i++) {
+		comm->tx_desc[i].cmd1 = 0;
+		wmb();		// Clear OWN_BIT and then set other fields.
+		comm->tx_desc[i].cmd2 = 0;
+		comm->tx_desc[i].addr1 = 0;
+		comm->tx_desc[i].addr2 = 0;
+
+		if (comm->tx_temp_skb_info[i].mapping) {
+			buflen = (comm->tx_temp_skb_info[i].skb) ?
+				 comm->tx_temp_skb_info[i].skb->len :
+				 MAC_TX_BUFF_SIZE;
+			dma_unmap_single(&comm->pdev->dev, comm->tx_temp_skb_info[i].mapping,
+					 buflen, DMA_TO_DEVICE);
+			comm->tx_temp_skb_info[i].mapping = 0;
+		}
+
+		if (comm->tx_temp_skb_info[i].skb) {
+			dev_kfree_skb(comm->tx_temp_skb_info[i].skb);
+			comm->tx_temp_skb_info[i].skb = NULL;
+		}
+	}
+}
+
+void rx_descs_clean(struct l2sw_common *comm)
+{
+	u32 i, j;
+	struct mac_desc *rx_desc;
+	struct skb_info *rx_skbinfo;
+
+	for (i = 0; i < RX_DESC_QUEUE_NUM; i++) {
+		if (!comm->rx_skb_info[i])
+			continue;
+
+		rx_desc = comm->rx_desc[i];
+		rx_skbinfo = comm->rx_skb_info[i];
+		for (j = 0; j < comm->rx_desc_num[i]; j++) {
+			rx_desc[j].cmd1 = 0;
+			wmb();	// Clear OWN_BIT and then set other fields.
+			rx_desc[j].cmd2 = 0;
+			rx_desc[j].addr1 = 0;
+
+			if (rx_skbinfo[j].skb) {
+				dma_unmap_single(&comm->pdev->dev, rx_skbinfo[j].mapping,
+						 comm->rx_desc_buff_size, DMA_FROM_DEVICE);
+				dev_kfree_skb(rx_skbinfo[j].skb);
+				rx_skbinfo[j].skb = NULL;
+				rx_skbinfo[j].mapping = 0;
+			}
+		}
+
+		kfree(rx_skbinfo);
+		comm->rx_skb_info[i] = NULL;
+	}
+}
+
+void descs_clean(struct l2sw_common *comm)
+{
+	rx_descs_clean(comm);
+	tx_descs_clean(comm);
+}
+
+void descs_free(struct l2sw_common *comm)
+{
+	u32 i;
+
+	descs_clean(comm);
+	comm->tx_desc = NULL;
+	for (i = 0; i < RX_DESC_QUEUE_NUM; i++)
+		comm->rx_desc[i] = NULL;
+
+	/*  Free descriptor area  */
+	if (comm->desc_base) {
+		dma_free_coherent(&comm->pdev->dev, comm->desc_size, comm->desc_base,
+				  comm->desc_dma);
+		comm->desc_base = NULL;
+		comm->desc_dma = 0;
+		comm->desc_size = 0;
+	}
+}
+
+u32 tx_descs_init(struct l2sw_common *comm)
+{
+	memset(comm->tx_desc, '\0', sizeof(struct mac_desc) * (TX_DESC_NUM + MAC_GUARD_DESC_NUM));
+	return 0;
+}
+
+u32 rx_descs_init(struct l2sw_common *comm)
+{
+	struct sk_buff *skb;
+	u32 i, j;
+	struct mac_desc *rx_desc;
+	struct skb_info *rx_skbinfo;
+
+	for (i = 0; i < RX_DESC_QUEUE_NUM; i++) {
+		comm->rx_skb_info[i] = kmalloc_array(comm->rx_desc_num[i],
+						     sizeof(struct skb_info), GFP_KERNEL);
+		if (!comm->rx_skb_info[i])
+			goto MEM_ALLOC_FAIL;
+
+		rx_skbinfo = comm->rx_skb_info[i];
+		rx_desc = comm->rx_desc[i];
+		for (j = 0; j < comm->rx_desc_num[i]; j++) {
+			skb = __dev_alloc_skb(comm->rx_desc_buff_size + RX_OFFSET,
+					      GFP_ATOMIC | GFP_DMA);
+			if (!skb)
+				goto MEM_ALLOC_FAIL;
+
+			skb->dev = comm->net_dev;
+			skb_reserve(skb, RX_OFFSET);	/* +data +tail */
+
+			rx_skbinfo[j].skb = skb;
+			rx_skbinfo[j].mapping = dma_map_single(&comm->pdev->dev, skb->data,
+							       comm->rx_desc_buff_size,
+							       DMA_FROM_DEVICE);
+			rx_desc[j].addr1 = rx_skbinfo[j].mapping;
+			rx_desc[j].addr2 = 0;
+			rx_desc[j].cmd2 = (j == comm->rx_desc_num[i] - 1) ?
+					  EOR_BIT | comm->rx_desc_buff_size :
+					  comm->rx_desc_buff_size;
+			wmb();	// Set OWN_BIT after other fields are effective.
+			rx_desc[j].cmd1 = OWN_BIT;
+		}
+	}
+
+	return 0;
+
+MEM_ALLOC_FAIL:
+	rx_descs_clean(comm);
+	return -ENOMEM;
+}
+
+u32 descs_alloc(struct l2sw_common *comm)
+{
+	u32 i;
+	s32 desc_size;
+
+	/* Alloc descriptor area  */
+	desc_size = (TX_DESC_NUM + MAC_GUARD_DESC_NUM) * sizeof(struct mac_desc);
+	for (i = 0; i < RX_DESC_QUEUE_NUM; i++)
+		desc_size += comm->rx_desc_num[i] * sizeof(struct mac_desc);
+
+	comm->desc_base = dma_alloc_coherent(&comm->pdev->dev, desc_size, &comm->desc_dma,
+					     GFP_KERNEL);
+	if (!comm->desc_base)
+		return -ENOMEM;
+
+	comm->desc_size = desc_size;
+
+	/* Setup Tx descriptor */
+	comm->tx_desc = (struct mac_desc *)comm->desc_base;
+
+	/* Setup Rx descriptor */
+	comm->rx_desc[0] = &comm->tx_desc[TX_DESC_NUM + MAC_GUARD_DESC_NUM];
+	for (i = 1; i < RX_DESC_QUEUE_NUM; i++)
+		comm->rx_desc[i] = comm->rx_desc[i - 1] + comm->rx_desc_num[i - 1];
+
+	return 0;
+}
+
+u32 descs_init(struct l2sw_common *comm)
+{
+	u32 i, rc;
+
+	// Initialize rx descriptor's data
+	comm->rx_desc_num[0] = RX_QUEUE0_DESC_NUM;
+#if RX_DESC_QUEUE_NUM > 1
+	comm->rx_desc_num[1] = RX_QUEUE1_DESC_NUM;
+#endif
+
+	for (i = 0; i < RX_DESC_QUEUE_NUM; i++) {
+		comm->rx_desc[i] = NULL;
+		comm->rx_skb_info[i] = NULL;
+		comm->rx_pos[i] = 0;
+	}
+	comm->rx_desc_buff_size = MAC_RX_LEN_MAX;
+
+	// Initialize tx descriptor's data
+	comm->tx_done_pos = 0;
+	comm->tx_desc = NULL;
+	comm->tx_pos = 0;
+	comm->tx_desc_full = 0;
+	for (i = 0; i < TX_DESC_NUM; i++)
+		comm->tx_temp_skb_info[i].skb = NULL;
+
+	// Allocate tx & rx descriptors.
+	rc = descs_alloc(comm);
+	if (rc) {
+		pr_err(" Failed to allocate tx & rx descriptors!\n");
+		return rc;
+	}
+
+	rc = tx_descs_init(comm);
+	if (rc)
+		return rc;
+
+	return rx_descs_init(comm);
+}
diff --git a/drivers/net/ethernet/sunplus/l2sw_desc.h b/drivers/net/ethernet/sunplus/l2sw_desc.h
new file mode 100644
index 0000000..d0647cb
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_desc.h
@@ -0,0 +1,21 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#ifndef __L2SW_DESC_H__
+#define __L2SW_DESC_H__
+
+#include "l2sw_define.h"
+
+void rx_descs_flush(struct l2sw_common *comm);
+void tx_descs_clean(struct l2sw_common *comm);
+void rx_descs_clean(struct l2sw_common *comm);
+void descs_clean(struct l2sw_common *comm);
+void descs_free(struct l2sw_common *comm);
+u32 tx_descs_init(struct l2sw_common *comm);
+u32 rx_descs_init(struct l2sw_common *comm);
+u32 descs_alloc(struct l2sw_common *comm);
+u32 descs_init(struct l2sw_common *comm);
+
+#endif
diff --git a/drivers/net/ethernet/sunplus/l2sw_driver.c b/drivers/net/ethernet/sunplus/l2sw_driver.c
new file mode 100644
index 0000000..3dfd0dd
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_driver.c
@@ -0,0 +1,779 @@ 
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#include <linux/clk.h>
+#include <linux/reset.h>
+#include <linux/nvmem-consumer.h>
+#include "l2sw_driver.h"
+
+static const char def_mac_addr[ETHERNET_MAC_ADDR_LEN] = {
+	0x88, 0x88, 0x88, 0x88, 0x88, 0x80
+};
+
+/*********************************************************************
+ *
+ * net_device_ops
+ *
+ **********************************************************************/
+static int ethernet_open(struct net_device *net_dev)
+{
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+
+	pr_debug(" Open port = %x\n", mac->lan_port);
+
+	mac->comm->enable |= mac->lan_port;
+
+	mac_hw_start(mac);
+	write_sw_int_mask0(read_sw_int_mask0() & ~(MAC_INT_TX | MAC_INT_RX));
+
+	netif_carrier_on(net_dev);
+	if (netif_carrier_ok(net_dev)) {
+		pr_debug(" Open netif_start_queue.\n");
+		netif_start_queue(net_dev);
+	}
+
+	return 0;
+}
+
+static int ethernet_stop(struct net_device *net_dev)
+{
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+
+	netif_stop_queue(net_dev);
+	netif_carrier_off(net_dev);
+
+	mac->comm->enable &= ~mac->lan_port;
+
+	mac_hw_stop(mac);
+
+	return 0;
+}
+
+/* Transmit a packet (called by the kernel) */
+static int ethernet_start_xmit(struct sk_buff *skb, struct net_device *net_dev)
+{
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+	struct l2sw_common *comm = mac->comm;
+	u32 tx_pos;
+	u32 cmd1;
+	u32 cmd2;
+	struct mac_desc *txdesc;
+	struct skb_info *skbinfo;
+	unsigned long flags;
+
+	if (unlikely(comm->tx_desc_full == 1)) {	/* no desc left, wait for tx interrupt */
+		pr_err(" TX descriptor queue full when xmit!\n");
+		return NETDEV_TX_BUSY;
+	}
+
+	/* if skb size shorter than 60, fill it with '\0' */
+	if (unlikely(skb->len < ETH_ZLEN)) {
+		if (skb_tailroom(skb) >= (ETH_ZLEN - skb->len)) {
+			memset(__skb_put(skb, ETH_ZLEN - skb->len), '\0', ETH_ZLEN - skb->len);
+		} else {
+			struct sk_buff *old_skb = skb;
+
+			skb = dev_alloc_skb(ETH_ZLEN + TX_OFFSET);
+			if (skb) {
+				memset(skb->data + old_skb->len, '\0', ETH_ZLEN - old_skb->len);
+				memcpy(skb->data, old_skb->data, old_skb->len);
+				skb_put(skb, ETH_ZLEN);	/* add data to an sk_buff */
+				dev_kfree_skb_irq(old_skb);
+			} else {
+				skb = old_skb;
+			}
+		}
+	}
+
+	spin_lock_irqsave(&mac->comm->tx_lock, flags);
+	tx_pos = comm->tx_pos;
+	txdesc = &comm->tx_desc[tx_pos];
+	skbinfo = &comm->tx_temp_skb_info[tx_pos];
+	skbinfo->len = skb->len;
+	skbinfo->skb = skb;
+	skbinfo->mapping = dma_map_single(&mac->pdev->dev, skb->data, skb->len, DMA_TO_DEVICE);
+	cmd1 = (OWN_BIT | FS_BIT | LS_BIT | (mac->to_vlan << 12) | (skb->len & LEN_MASK));
+	cmd2 = skb->len & LEN_MASK;
+
+	if (tx_pos == (TX_DESC_NUM - 1))
+		cmd2 |= EOR_BIT;
+
+	txdesc->addr1 = skbinfo->mapping;
+	txdesc->cmd2 = cmd2;
+	wmb();			// Set OWN_BIT after other fields of descriptor are effective.
+	txdesc->cmd1 = cmd1;
+
+	NEXT_TX(tx_pos);
+
+	if (unlikely(tx_pos == comm->tx_done_pos)) {
+		netif_stop_queue(net_dev);
+		comm->tx_desc_full = 1;
+		//pr_info(" TX Descriptor Queue Full!\n");
+	}
+	comm->tx_pos = tx_pos;
+	wmb();			// make sure settings are effective.
+
+	/* trigger gmac to transmit */
+	tx_trigger();
+
+	spin_unlock_irqrestore(&mac->comm->tx_lock, flags);
+	return NETDEV_TX_OK;
+}
+
+static void ethernet_set_rx_mode(struct net_device *net_dev)
+{
+	if (net_dev) {
+		struct l2sw_mac *mac = netdev_priv(net_dev);
+		struct l2sw_common *comm = mac->comm;
+		unsigned long flags;
+
+		spin_lock_irqsave(&comm->ioctl_lock, flags);
+		rx_mode_set(net_dev);
+		spin_unlock_irqrestore(&comm->ioctl_lock, flags);
+	}
+}
+
+static int ethernet_set_mac_address(struct net_device *net_dev, void *addr)
+{
+	struct sockaddr *hwaddr = (struct sockaddr *)addr;
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+
+	if (netif_running(net_dev)) {
+		pr_err(" Device %s is busy!\n", net_dev->name);
+		return -EBUSY;
+	}
+
+	memcpy(net_dev->dev_addr, hwaddr->sa_data, net_dev->addr_len);
+
+	/* Delete the old Ethernet MAC address */
+	pr_debug(" HW Addr = %pM\n", mac->mac_addr);
+	if (is_valid_ether_addr(mac->mac_addr))
+		mac_hw_addr_del(mac);
+
+	/* Set the Ethernet MAC address */
+	memcpy(mac->mac_addr, hwaddr->sa_data, net_dev->addr_len);
+	mac_hw_addr_set(mac);
+
+	return 0;
+}
+
+static int ethernet_do_ioctl(struct net_device *net_dev, struct ifreq *ifr, int cmd)
+{
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+	struct l2sw_common *comm = mac->comm;
+	struct mii_ioctl_data *data = if_mii(ifr);
+	unsigned long flags;
+
+	pr_debug(" if = %s, cmd = %04x\n", ifr->ifr_ifrn.ifrn_name, cmd);
+	pr_debug(" phy_id = %d, reg_num = %d, val_in = %04x\n", data->phy_id,
+		 data->reg_num, data->val_in);
+
+	// Check parameters' range.
+	if ((cmd == SIOCGMIIREG) || (cmd == SIOCSMIIREG)) {
+		if (data->reg_num > 31) {
+			pr_err(" reg_num (= %d) excesses range!\n", (int)data->reg_num);
+			return -EINVAL;
+		}
+	}
+
+	switch (cmd) {
+	case SIOCGMIIPHY:
+		if (comm->dual_nic && (strcmp(ifr->ifr_ifrn.ifrn_name, "eth1") == 0))
+			return comm->phy2_addr;
+		else
+			return comm->phy1_addr;
+
+	case SIOCGMIIREG:
+		spin_lock_irqsave(&comm->ioctl_lock, flags);
+		data->val_out = mdio_read(data->phy_id, data->reg_num);
+		spin_unlock_irqrestore(&comm->ioctl_lock, flags);
+		pr_debug(" val_out = %04x\n", data->val_out);
+		break;
+
+	case SIOCSMIIREG:
+		spin_lock_irqsave(&comm->ioctl_lock, flags);
+		mdio_write(data->phy_id, data->reg_num, data->val_in);
+		spin_unlock_irqrestore(&comm->ioctl_lock, flags);
+		break;
+
+	default:
+		pr_err(" ioctl #%d has not implemented yet!\n", cmd);
+		return -EOPNOTSUPP;
+	}
+	return 0;
+}
+
+static int ethernet_change_mtu(struct net_device *ndev, int new_mtu)
+{
+	if (netif_running(ndev))
+		return -EBUSY;
+
+	if (new_mtu < 68 || new_mtu > ETH_DATA_LEN)
+		return -EINVAL;
+
+	ndev->mtu = new_mtu;
+
+	return 0;
+}
+
+static void ethernet_tx_timeout(struct net_device *net_dev, unsigned int txqueue)
+{
+}
+
+static struct net_device_stats *ethernet_get_stats(struct net_device *net_dev)
+{
+	struct l2sw_mac *mac;
+
+	mac = netdev_priv(net_dev);
+	return &mac->dev_stats;
+}
+
+static const struct net_device_ops netdev_ops = {
+	.ndo_open = ethernet_open,
+	.ndo_stop = ethernet_stop,
+	.ndo_start_xmit = ethernet_start_xmit,
+	.ndo_set_rx_mode = ethernet_set_rx_mode,
+	.ndo_set_mac_address = ethernet_set_mac_address,
+	.ndo_do_ioctl = ethernet_do_ioctl,
+	.ndo_change_mtu = ethernet_change_mtu,
+	.ndo_tx_timeout = ethernet_tx_timeout,
+	.ndo_get_stats = ethernet_get_stats,
+};
+
+char *sp7021_otp_read_mac(struct device *_d, ssize_t *_l, char *_name)
+{
+	char *ret = NULL;
+	struct nvmem_cell *c = nvmem_cell_get(_d, _name);
+
+	if (IS_ERR_OR_NULL(c)) {
+		pr_err(" OTP %s read failure: %ld", _name, PTR_ERR(c));
+		return NULL;
+	}
+
+	ret = nvmem_cell_read(c, _l);
+	nvmem_cell_put(c);
+	pr_debug(" %zd bytes are read from OTP %s.", *_l, _name);
+
+	return ret;
+}
+
+static void check_mac_vendor_id_and_convert(char *mac_addr)
+{
+	// Byte order of MAC address of some samples are reversed.
+	// Check vendor id and convert byte order if it is wrong.
+	if ((mac_addr[5] == 0xFC) && (mac_addr[4] == 0x4B) && (mac_addr[3] == 0xBC) &&
+	    ((mac_addr[0] != 0xFC) || (mac_addr[1] != 0x4B) || (mac_addr[2] != 0xBC))) {
+		char tmp;
+
+		// Swap mac_addr[0] and mac_addr[5]
+		tmp = mac_addr[0];
+		mac_addr[0] = mac_addr[5];
+		mac_addr[5] = tmp;
+
+		// Swap mac_addr[1] and mac_addr[4]
+		tmp = mac_addr[1];
+		mac_addr[1] = mac_addr[4];
+		mac_addr[4] = tmp;
+
+		// Swap mac_addr[2] and mac_addr[3]
+		tmp = mac_addr[2];
+		mac_addr[2] = mac_addr[3];
+		mac_addr[3] = tmp;
+	}
+}
+
+/*********************************************************************
+ *
+ * platform_driver
+ *
+ **********************************************************************/
+static u32 init_netdev(struct platform_device *pdev, int eth_no, struct net_device **r_ndev)
+{
+	u32 ret = -ENODEV;
+	struct l2sw_mac *mac;
+	struct net_device *net_dev;
+	char *m_addr_name = (eth_no == 0) ? "mac_addr0" : "mac_addr1";
+	ssize_t otp_l = 0;
+	char *otp_v;
+
+	/* allocate the devices, and also allocate l2sw_mac, we can get it by netdev_priv() */
+	net_dev = alloc_etherdev(sizeof(struct l2sw_mac));
+	if (!net_dev) {
+		*r_ndev = NULL;
+		return -ENOMEM;
+	}
+	SET_NETDEV_DEV(net_dev, &pdev->dev);
+	net_dev->netdev_ops = &netdev_ops;
+
+	mac = netdev_priv(net_dev);
+	mac->net_dev = net_dev;
+	mac->pdev = pdev;
+	mac->next_netdev = NULL;
+
+	// Get property 'mac-addr0' or 'mac-addr1' from dts.
+	otp_v = sp7021_otp_read_mac(&pdev->dev, &otp_l, m_addr_name);
+	if ((otp_l < 6) || IS_ERR_OR_NULL(otp_v)) {
+		pr_info(" OTP mac %s (len = %zd) is invalid, using default!\n",
+			m_addr_name, otp_l);
+		otp_l = 0;
+	} else {
+		// Check if mac-address is valid or not. If not, copy from default.
+		memcpy(mac->mac_addr, otp_v, 6);
+
+		// Byte order of Some samples are reversed. Convert byte order here.
+		check_mac_vendor_id_and_convert(mac->mac_addr);
+
+		if (!is_valid_ether_addr(mac->mac_addr)) {
+			pr_info(" Invalid mac in OTP[%s] = %pM, use default!\n",
+				m_addr_name, mac->mac_addr);
+			otp_l = 0;
+		}
+	}
+	if (otp_l != 6) {
+		memcpy(mac->mac_addr, def_mac_addr, ETHERNET_MAC_ADDR_LEN);
+		mac->mac_addr[5] += eth_no;
+	}
+
+	pr_info(" HW Addr = %pM\n", mac->mac_addr);
+
+	memcpy(net_dev->dev_addr, mac->mac_addr, ETHERNET_MAC_ADDR_LEN);
+
+	ret = register_netdev(net_dev);
+	if (ret != 0) {
+		pr_err(" Failed to register net device \"%s\" (ret = %d)!\n", net_dev->name, ret);
+		free_netdev(net_dev);
+		*r_ndev = NULL;
+		return ret;
+	}
+	pr_info(" Registered net device \"%s\" successfully.\n", net_dev->name);
+
+	*r_ndev = net_dev;
+	return 0;
+}
+
+static ssize_t mode_show(struct device *dev, struct device_attribute *attr, char *buf)
+{
+	struct net_device *net_dev = dev_get_drvdata(dev);
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+
+	return sprintf(buf, "%d\n", (mac->comm->dual_nic) ? 1 : (mac->comm->sa_learning) ? 0 : 2);
+}
+
+static ssize_t mode_store(struct device *dev, struct device_attribute *attr,
+			  const char *buf, size_t count)
+{
+	struct net_device *net_dev = dev_get_drvdata(dev);
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+	struct l2sw_common *comm = mac->comm;
+	struct net_device *net_dev2;
+	struct l2sw_mac *mac2;
+
+	if (buf[0] == '1') {
+		// Switch to dual NIC mode.
+		mutex_lock(&comm->store_mode);
+		if (!comm->dual_nic) {
+			mac_hw_stop(mac);
+
+			comm->dual_nic = 1;
+			comm->sa_learning = 0;
+			mac_disable_port_sa_learning();
+			mac_addr_table_del_all();
+			mac_switch_mode(mac);
+			rx_mode_set(net_dev);
+
+			init_netdev(mac->pdev, 1, &net_dev2);	// Initialize the 2nd net device.
+			if (net_dev2) {
+				mac->next_netdev = net_dev2;	// Pointed by previous net device.
+				mac2 = netdev_priv(net_dev2);
+				mac2->comm = comm;
+				net_dev2->irq = comm->irq;
+
+				mac_switch_mode(mac);
+				rx_mode_set(net_dev2);
+				mac_hw_addr_set(mac2);
+			}
+
+			comm->enable &= 0x1;	// Keep lan 0, but always turn off lan 1.
+			mac_hw_start(mac);
+		}
+		mutex_unlock(&comm->store_mode);
+	} else if ((buf[0] == '0') || (buf[0] == '2')) {
+		// Switch to one NIC with daisy-chain mode.
+		mutex_lock(&comm->store_mode);
+
+		if (buf[0] == '2') {
+			if (comm->sa_learning == 1) {
+				comm->sa_learning = 0;
+				mac_disable_port_sa_learning();
+				mac_addr_table_del_all();
+			}
+		} else {
+			if (comm->sa_learning == 0) {
+				comm->sa_learning = 1;
+				mac_enable_port_sa_learning();
+			}
+		}
+
+		if (comm->dual_nic) {
+			struct net_device *net_dev2 = mac->next_netdev;
+
+			if (!netif_running(net_dev2)) {
+				mac_hw_stop(mac);
+
+				mac2 = netdev_priv(net_dev2);
+
+				// unregister and free net device.
+				unregister_netdev(net_dev2);
+				free_netdev(net_dev2);
+				mac->next_netdev = NULL;
+				pr_info(" Unregistered and freed net device \"eth1\"!\n");
+
+				comm->dual_nic = 0;
+				mac_switch_mode(mac);
+				rx_mode_set(net_dev);
+				mac_hw_addr_del(mac2);
+
+				// If eth0 is up, turn on lan 0 and 1 when
+				// switching to daisy-chain mode.
+				if (comm->enable & 0x1)
+					comm->enable = 0x3;
+				else
+					comm->enable = 0;
+
+				mac_hw_start(mac);
+			} else {
+				pr_err(" Error: Net device \"%s\" is running!\n", net_dev2->name);
+			}
+		}
+		mutex_unlock(&comm->store_mode);
+	} else {
+		pr_err(" Error: Unknown mode \"%c\"!\n", buf[0]);
+	}
+
+	return count;
+}
+
+static DEVICE_ATTR_RW(mode);
+static struct attribute *l2sw_sysfs_entries[] = {
+	&dev_attr_mode.attr,
+	NULL,
+};
+
+static struct attribute_group l2sw_attribute_group = {
+	.attrs = l2sw_sysfs_entries,
+};
+
+static int soc0_open(struct l2sw_mac *mac)
+{
+	struct l2sw_common *comm = mac->comm;
+	u32 rc;
+
+	mac_hw_stop(mac);
+
+	rc = descs_init(comm);
+	if (rc) {
+		pr_err(" Fail to initialize mac descriptors!\n");
+		goto INIT_DESC_FAIL;
+	}
+
+	/*start hardware port, open interrupt, start system tx queue */
+	mac_init(mac);
+	return 0;
+
+INIT_DESC_FAIL:
+	descs_free(comm);
+	return rc;
+}
+
+static int soc0_stop(struct l2sw_mac *mac)
+{
+	mac_hw_stop(mac);
+
+	descs_free(mac->comm);
+	return 0;
+}
+
+static int l2sw_probe(struct platform_device *pdev)
+{
+	struct l2sw_common *comm;
+	struct resource *r_mem;
+	struct net_device *net_dev, *net_dev2;
+	struct l2sw_mac *mac, *mac2;
+	u32 mode;
+	int ret = 0;
+	int rc;
+
+	if (platform_get_drvdata(pdev))
+		return -ENODEV;
+
+	// Allocate memory for l2sw 'common' area.
+	comm = kmalloc(sizeof(*comm), GFP_KERNEL);
+	if (!comm)
+		return -ENOMEM;
+	pr_debug(" comm = %p\n", comm);
+	memset(comm, '\0', sizeof(struct l2sw_common));
+	comm->pdev = pdev;
+
+	spin_lock_init(&comm->rx_lock);
+	spin_lock_init(&comm->tx_lock);
+	spin_lock_init(&comm->ioctl_lock);
+	mutex_init(&comm->store_mode);
+
+	// Get memory resoruce 0 from dts.
+	r_mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (r_mem) {
+		pr_debug(" res->name = \"%s\", r_mem->start = %pa\n", r_mem->name, &r_mem->start);
+		if (l2sw_reg_base_set(devm_ioremap(&pdev->dev, r_mem->start,
+						   (r_mem->end - r_mem->start + 1))) != 0) {
+			pr_err(" ioremap failed!\n");
+			ret = -ENOMEM;
+			goto out_free_comm;
+		}
+	} else {
+		pr_err(" No MEM resource 0 found!\n");
+		ret = -ENXIO;
+		goto out_free_comm;
+	}
+
+	// Get memory resoruce 1 from dts.
+	r_mem = platform_get_resource(pdev, IORESOURCE_MEM, 1);
+	if (r_mem) {
+		pr_debug(" res->name = \"%s\", r_mem->start = %pa\n", r_mem->name, &r_mem->start);
+		if (moon5_reg_base_set(devm_ioremap(&pdev->dev, r_mem->start,
+						    (r_mem->end - r_mem->start + 1))) != 0) {
+			pr_err(" ioremap failed!\n");
+			ret = -ENOMEM;
+			goto out_free_comm;
+		}
+	} else {
+		pr_err(" No MEM resource 1 found!\n");
+		ret = -ENXIO;
+		goto out_free_comm;
+	}
+
+	// Get irq resource from dts.
+	if (l2sw_get_irq(pdev, comm) != 0) {
+		ret = -ENXIO;
+		goto out_free_comm;
+	}
+
+	// Get L2-switch mode.
+	ret = of_property_read_u32(pdev->dev.of_node, "mode", &mode);
+	if (ret)
+		mode = 0;
+	pr_info(" L2 switch mode = %u\n", mode);
+	if (mode == 2) {
+		comm->dual_nic = 0;	// daisy-chain mode 2
+		comm->sa_learning = 0;
+	} else if (mode == 1) {
+		comm->dual_nic = 1;	// dual NIC mode
+		comm->sa_learning = 0;
+	} else {
+		comm->dual_nic = 0;	// daisy-chain mode
+		comm->sa_learning = 1;
+	}
+
+	// Get resource of clock controller
+	comm->clk = devm_clk_get(&pdev->dev, NULL);
+	if (IS_ERR(comm->clk)) {
+		dev_err(&pdev->dev, "Failed to retrieve clock controller!\n");
+		ret = PTR_ERR(comm->clk);
+		goto out_free_comm;
+	}
+
+	comm->rstc = devm_reset_control_get(&pdev->dev, NULL);
+	if (IS_ERR(comm->rstc)) {
+		dev_err(&pdev->dev, "Failed to retrieve reset controller!\n");
+		ret = PTR_ERR(comm->rstc);
+		goto out_free_comm;
+	}
+
+	// Enable clock.
+	clk_prepare_enable(comm->clk);
+	udelay(1);
+
+	ret = reset_control_assert(comm->rstc);
+	udelay(1);
+	ret = reset_control_deassert(comm->rstc);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to deassert reset line (err = %d)!\n", ret);
+		ret = -ENODEV;
+		goto out_free_comm;
+	}
+	udelay(1);
+
+	ret = init_netdev(pdev, 0, &net_dev);	// Initialize the 1st net device (eth0)
+	if (!net_dev)
+		goto out_free_comm;
+
+	platform_set_drvdata(pdev, net_dev);	// Pointed by drvdata net device.
+
+	net_dev->irq = comm->irq;
+
+	mac = netdev_priv(net_dev);
+	mac->comm = comm;
+	comm->net_dev = net_dev;
+	pr_debug(" net_dev = %p, mac = %p, comm = %p\n", net_dev, mac, mac->comm);
+
+	comm->phy1_node = of_parse_phandle(pdev->dev.of_node, "phy-handle1", 0);
+	comm->phy2_node = of_parse_phandle(pdev->dev.of_node, "phy-handle2", 0);
+
+	// Get address of phy of ethernet from dts.
+	if (of_property_read_u32(comm->phy1_node, "reg", &rc) == 0) {
+		comm->phy1_addr = rc;
+	} else {
+		comm->phy1_addr = 0;
+		pr_info(" Cannot get address of phy of ethernet 1! Set to 0 by default.\n");
+	}
+
+	if (of_property_read_u32(comm->phy2_node, "reg", &rc) == 0) {
+		comm->phy2_addr = rc;
+	} else {
+		comm->phy2_addr = 1;
+		pr_info(" Cannot get address of phy of ethernet 2! Set to 1 by default.\n");
+	}
+
+	l2sw_enable_port(mac);
+
+	if (comm->phy1_node) {
+		ret = mdio_init(pdev, net_dev);
+		if (ret) {
+			pr_err(" Failed to initialize mdio!\n");
+			goto out_unregister_dev;
+		}
+
+		ret = mac_phy_probe(net_dev);
+		if (ret) {
+			pr_err(" Failed to probe phy!\n");
+			goto out_freemdio;
+		}
+	} else {
+		pr_err(" Failed to get phy-handle!\n");
+	}
+
+	phy_cfg(mac);
+
+	// Register irq to system.
+	if (l2sw_request_irq(pdev, comm, net_dev) != 0) {
+		ret = -ENODEV;
+		goto out_freemdio;
+	}
+
+	netif_napi_add(net_dev, &comm->rx_napi, rx_poll, RX_NAPI_WEIGHT);
+	napi_enable(&comm->rx_napi);
+	netif_napi_add(net_dev, &comm->tx_napi, tx_poll, TX_NAPI_WEIGHT);
+	napi_enable(&comm->tx_napi);
+
+	soc0_open(mac);
+
+	// Set MAC address
+	mac_hw_addr_set(mac);
+
+	/* Add the device attributes */
+	rc = sysfs_create_group(&pdev->dev.kobj, &l2sw_attribute_group);
+	if (rc)
+		dev_err(&pdev->dev, "Error creating sysfs files!\n");
+
+	mac_addr_table_del_all();
+	if (comm->sa_learning)
+		mac_enable_port_sa_learning();
+	else
+		mac_disable_port_sa_learning();
+	rx_mode_set(net_dev);
+
+	if (comm->dual_nic) {
+		init_netdev(pdev, 1, &net_dev2);
+		if (!net_dev2)
+			goto fail_to_init_2nd_port;
+		mac->next_netdev = net_dev2;	// Pointed by previous net device.
+
+		net_dev2->irq = comm->irq;
+		mac2 = netdev_priv(net_dev2);
+		mac2->comm = comm;
+		pr_debug(" net_dev = %p, mac = %p, comm = %p\n", net_dev2, mac2, mac2->comm);
+
+		mac_switch_mode(mac);
+		rx_mode_set(net_dev2);
+		mac_hw_addr_set(mac2);	// Set MAC address for the second net device.
+	}
+
+fail_to_init_2nd_port:
+	return 0;
+
+out_freemdio:
+	if (comm->mii_bus)
+		mdio_remove(net_dev);
+
+out_unregister_dev:
+	unregister_netdev(net_dev);
+
+out_free_comm:
+	kfree(comm);
+	return ret;
+}
+
+static int l2sw_remove(struct platform_device *pdev)
+{
+	struct net_device *net_dev;
+	struct net_device *net_dev2;
+	struct l2sw_mac *mac;
+
+	net_dev = platform_get_drvdata(pdev);
+	if (!net_dev)
+		return 0;
+	mac = netdev_priv(net_dev);
+
+	// Unregister and free 2nd net device.
+	net_dev2 = mac->next_netdev;
+	if (net_dev2) {
+		unregister_netdev(net_dev2);
+		free_netdev(net_dev2);
+	}
+
+	sysfs_remove_group(&pdev->dev.kobj, &l2sw_attribute_group);
+
+	mac->comm->enable = 0;
+	soc0_stop(mac);
+
+	napi_disable(&mac->comm->rx_napi);
+	netif_napi_del(&mac->comm->rx_napi);
+	napi_disable(&mac->comm->tx_napi);
+	netif_napi_del(&mac->comm->tx_napi);
+
+	mdio_remove(net_dev);
+
+	// Unregister and free 1st net device.
+	unregister_netdev(net_dev);
+	free_netdev(net_dev);
+
+	clk_disable(mac->comm->clk);
+
+	// Free 'common' area.
+	kfree(mac->comm);
+	return 0;
+}
+
+static const struct of_device_id sp_l2sw_of_match[] = {
+	{.compatible = "sunplus,sp7021-l2sw"},
+	{ /* sentinel */ }
+};
+
+MODULE_DEVICE_TABLE(of, sp_l2sw_of_match);
+
+static struct platform_driver l2sw_driver = {
+	.probe = l2sw_probe,
+	.remove = l2sw_remove,
+	.driver = {
+		.name = "sp_l2sw",
+		.owner = THIS_MODULE,
+		.of_match_table = sp_l2sw_of_match,
+	},
+};
+
+module_platform_driver(l2sw_driver);
+
+MODULE_AUTHOR("Wells Lu <[e-mail]>");
+MODULE_DESCRIPTION("Sunplus 10M/100M Ethernet (with L2 switch) driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/net/ethernet/sunplus/l2sw_driver.h b/drivers/net/ethernet/sunplus/l2sw_driver.h
new file mode 100644
index 0000000..a076b33
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_driver.h
@@ -0,0 +1,23 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#ifndef __L2SW_DRIVER_H__
+#define __L2SW_DRIVER_H__
+
+#include "l2sw_define.h"
+#include "l2sw_register.h"
+#include "l2sw_hal.h"
+#include "l2sw_int.h"
+#include "l2sw_mdio.h"
+#include "l2sw_mac.h"
+#include "l2sw_desc.h"
+
+#define NEXT_TX(N)              ((N) = (((N) + 1) == TX_DESC_NUM) ? 0 : (N) + 1)
+#define NEXT_RX(QUEUE, N)       ((N) = (((N) + 1) == mac->comm->rx_desc_num[QUEUE]) ? 0 : (N) + 1)
+
+#define RX_NAPI_WEIGHT          16
+#define TX_NAPI_WEIGHT          16
+
+#endif
diff --git a/drivers/net/ethernet/sunplus/l2sw_hal.c b/drivers/net/ethernet/sunplus/l2sw_hal.c
new file mode 100644
index 0000000..2aac817
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_hal.c
@@ -0,0 +1,422 @@ 
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#include "l2sw_hal.h"
+
+static struct l2sw_reg *l2sw_reg_base;
+static struct moon5_reg *moon5_reg_base;
+
+int l2sw_reg_base_set(void __iomem *baseaddr)
+{
+	l2sw_reg_base = (struct l2sw_reg *)baseaddr;
+	pr_debug(" l2sw_reg_base = %p\n", l2sw_reg_base);
+
+	if (!l2sw_reg_base)
+		return -1;
+
+	return 0;
+}
+
+int moon5_reg_base_set(void __iomem *baseaddr)
+{
+	moon5_reg_base = (struct moon5_reg *)baseaddr;
+	pr_debug(" moon5_reg_base = %p\n", moon5_reg_base);
+
+	if (!moon5_reg_base)
+		return -1;
+
+	return 0;
+}
+
+void mac_hw_stop(struct l2sw_mac *mac)
+{
+	struct l2sw_common *comm = mac->comm;
+	u32 reg, disable;
+
+	if (comm->enable == 0) {
+		HWREG_W(sw_int_mask_0, 0xffffffff);
+		HWREG_W(sw_int_status_0, 0xffffffff & (~MAC_INT_PORT_ST_CHG));
+
+		reg = HWREG_R(cpu_cntl);
+		HWREG_W(cpu_cntl, (0x3 << 6) | reg);	// Disable cpu 0 and cpu 1.
+	}
+
+	if (comm->dual_nic) {
+		disable = ((~comm->enable) & 0x3) << 24;
+		reg = HWREG_R(port_cntl0);
+		HWREG_W(port_cntl0, disable | reg);	// Disable lan 0 and lan 1.
+		wmb();		// make sure settings are effective.
+	}
+}
+
+void mac_hw_reset(struct l2sw_mac *mac)
+{
+}
+
+void mac_hw_start(struct l2sw_mac *mac)
+{
+	struct l2sw_common *comm = mac->comm;
+	u32 reg;
+
+	//enable cpu port 0 (6) & port 0 crc padding (8)
+	reg = HWREG_R(cpu_cntl);
+	HWREG_W(cpu_cntl, (reg & (~(0x1 << 6))) | (0x1 << 8));
+	wmb();			// make sure settings are effective.
+
+	//enable lan 0 & lan 1
+	reg = HWREG_R(port_cntl0);
+	HWREG_W(port_cntl0, reg & (~(comm->enable << 24)));
+	wmb();			// make sure settings are effective.
+}
+
+void mac_hw_addr_set(struct l2sw_mac *mac)
+{
+	u32 reg;
+
+	HWREG_W(w_mac_15_0, mac->mac_addr[0] + (mac->mac_addr[1] << 8));
+	HWREG_W(w_mac_47_16, mac->mac_addr[2] + (mac->mac_addr[3] << 8) +
+		(mac->mac_addr[4] << 16) + (mac->mac_addr[5] << 24));
+	wmb();			// make sure settings are effective.
+
+	// Set aging=1
+	HWREG_W(wt_mac_ad0, (mac->cpu_port << 10) + (mac->vlan_id << 7) + (1 << 4) + 0x1);
+	wmb();			// make sure settings are effective.
+
+	do {
+		reg = HWREG_R(wt_mac_ad0);
+		ndelay(10);
+		pr_debug(" wt_mac_ad0 = %08x\n", reg);
+	} while ((reg & (0x1 << 1)) == 0x0);
+	pr_debug(" mac_ad0 = %08x, mac_ad = %08x%04x\n", HWREG_R(wt_mac_ad0),
+		 HWREG_R(w_mac_47_16), HWREG_R(w_mac_15_0) & 0xffff);
+}
+
+void mac_hw_addr_del(struct l2sw_mac *mac)
+{
+	u32 reg;
+
+	HWREG_W(w_mac_15_0, mac->mac_addr[0] + (mac->mac_addr[1] << 8));
+	HWREG_W(w_mac_47_16, mac->mac_addr[2] + (mac->mac_addr[3] << 8) +
+		(mac->mac_addr[4] << 16) + (mac->mac_addr[5] << 24));
+	wmb();			// make sure settings are effective.
+
+	HWREG_W(wt_mac_ad0, (0x1 << 12) + (mac->vlan_id << 7) + 0x1);
+	wmb();			// make sure settings are effective.
+	do {
+		reg = HWREG_R(wt_mac_ad0);
+		ndelay(10);
+		pr_debug(" wt_mac_ad0 = %08x\n", reg);
+	} while ((reg & (0x1 << 1)) == 0x0);
+	pr_debug(" mac_ad0 = %08x, mac_ad = %08x%04x\n", HWREG_R(wt_mac_ad0),
+		 HWREG_R(w_mac_47_16), HWREG_R(w_mac_15_0) & 0xffff);
+}
+
+void mac_addr_table_del_all(void)
+{
+	u32 reg;
+
+	// Wait for address table being idle.
+	do {
+		reg = HWREG_R(addr_tbl_srch);
+		ndelay(10);
+	} while (!(reg & MAC_ADDR_LOOKUP_IDLE));
+
+	// Search address table from start.
+	HWREG_W(addr_tbl_srch, HWREG_R(addr_tbl_srch) | MAC_BEGIN_SEARCH_ADDR);
+	mb();			// make sure settings are effective.
+	while (1) {
+		do {
+			reg = HWREG_R(addr_tbl_st);
+			ndelay(10);
+			pr_debug(" addr_tbl_st = %08x\n", reg);
+		} while (!(reg & (MAC_AT_TABLE_END | MAC_AT_DATA_READY)));
+
+		if (reg & MAC_AT_TABLE_END)
+			break;
+
+		pr_debug(" addr_tbl_st = %08x\n", reg);
+		pr_debug(" @AT #%u: port=%01x, cpu=%01x, vid=%u, aging=%u, proxy=%u, mc_ingress=%u\n",
+			 (reg >> 22) & 0x3ff, (reg >> 12) & 0x3, (reg >> 10) & 0x3,
+			 (reg >> 7) & 0x7, (reg >> 4) & 0x7, (reg >> 3) & 0x1,
+			 (reg >> 2) & 0x1);
+
+		// Delete all entries which are learnt from lan ports.
+		if ((reg >> 12) & 0x3) {
+			HWREG_W(w_mac_15_0, HWREG_R(mac_ad_ser0));
+			wmb();	// make sure settings are effective.
+			HWREG_W(w_mac_47_16, HWREG_R(mac_ad_ser1));
+			wmb();	// make sure settings are effective.
+
+			HWREG_W(wt_mac_ad0, (0x1 << 12) + (reg & (0x7 << 7)) + 0x1);
+			wmb();	// make sure settings are effective.
+			do {
+				reg = HWREG_R(wt_mac_ad0);
+				ndelay(10);
+				pr_debug(" wt_mac_ad0 = %08x\n", reg);
+			} while ((reg & (0x1 << 1)) == 0x0);
+			pr_debug(" mac_ad0 = %08x, mac_ad = %08x%04x\n", HWREG_R(wt_mac_ad0),
+				 HWREG_R(w_mac_47_16), HWREG_R(w_mac_15_0) & 0xffff);
+		}
+		// Search next.
+		wmb();		// make sure settings are effective.
+
+		HWREG_W(addr_tbl_srch, HWREG_R(addr_tbl_srch) | MAC_SEARCH_NEXT_ADDR);
+	}
+}
+
+void mac_hw_init(struct l2sw_mac *mac)
+{
+	struct l2sw_common *comm = mac->comm;
+	u32 reg;
+
+	reg = HWREG_R(cpu_cntl);
+	HWREG_W(cpu_cntl, (0x3 << 6) | reg);	// Disable cpu0 and cpu 1.
+	wmb();			// make sure settings are effective.
+
+	/* descriptor base address */
+	HWREG_W(tx_lbase_addr_0, mac->comm->desc_dma);
+	HWREG_W(tx_hbase_addr_0, mac->comm->desc_dma + sizeof(struct mac_desc) * TX_DESC_NUM);
+	HWREG_W(rx_hbase_addr_0, mac->comm->desc_dma +
+		sizeof(struct mac_desc) * (TX_DESC_NUM + MAC_GUARD_DESC_NUM));
+	HWREG_W(rx_lbase_addr_0, mac->comm->desc_dma +
+		sizeof(struct mac_desc) * (TX_DESC_NUM + MAC_GUARD_DESC_NUM + RX_QUEUE0_DESC_NUM));
+	wmb();			// make sure settings are effective.
+
+	// Threshold values
+	HWREG_W(fl_cntl_th, 0x4a3a2d1d);	// Fc_rls_th=0x4a,   Fc_set_th=0x3a,
+						// Drop_rls_th=0x2d, Drop_set_th=0x1d
+	HWREG_W(cpu_fl_cntl_th, 0x4a3a1212);	// Cpu_rls_th=0x4a,  Cpu_set_th=0x3a,
+						// Cpu_th=0x12,      Port_th=0x12
+	HWREG_W(pri_fl_cntl, 0xf6680000);	// mtcc_lmt=0xf,     Pri_th_l=6,
+						// Pri_th_h=6,       weigh_8x_en=1
+
+	// High-active LED
+	reg = HWREG_R(led_port0);
+	HWREG_W(led_port0, reg | (1 << 28));
+
+	/* phy address */
+	reg = HWREG_R(mac_force_mode);
+	reg = (reg & (~(0x1f << 16))) | ((mac->comm->phy1_addr & 0x1f) << 16);
+	reg = (reg & (~(0x1f << 24))) | ((mac->comm->phy2_addr & 0x1f) << 24);
+	HWREG_W(mac_force_mode, reg);
+
+	//disable cpu port0 aging (12)
+	//disable cpu port0 learning (14)
+	//enable UC and MC packets
+	reg = HWREG_R(cpu_cntl);
+	HWREG_W(cpu_cntl, (reg & (~((0x1 << 14) | (0x3c << 0)))) | (0x1 << 12));
+
+	mac_switch_mode(mac);
+
+	if (!comm->dual_nic) {
+		//enable lan 0 & lan 1
+		reg = HWREG_R(port_cntl0);
+		HWREG_W(port_cntl0, reg & (~(0x3 << 24)));
+		wmb();		// make sure settings are effective.
+	}
+
+	wmb();			// make sure settings are effective.
+	HWREG_W(sw_int_mask_0, MAC_INT_MASK_DEF);
+}
+
+void mac_switch_mode(struct l2sw_mac *mac)
+{
+	struct l2sw_common *comm = mac->comm;
+	u32 reg;
+
+	if (comm->dual_nic) {
+		mac->cpu_port = 0x1;	// soc0
+		mac->lan_port = 0x1;	// forward to port 0
+		mac->to_vlan = 0x1;	// vlan group: 0
+		mac->vlan_id = 0x0;	// vlan group: 0
+
+		if (mac->next_netdev) {
+			struct l2sw_mac *mac2 = netdev_priv(mac->next_netdev);
+
+			mac2->cpu_port = 0x1;	// soc0
+			mac2->lan_port = 0x2;	// forward to port 1
+			mac2->to_vlan = 0x2;	// vlan group: 1
+			mac2->vlan_id = 0x1;	// vlan group: 1
+		}
+		//port 0: VLAN group 0
+		//port 1: VLAN group 1
+		HWREG_W(pvid_config0, (1 << 4) + 0);
+
+		//VLAN group 0: cpu0+port0
+		//VLAN group 1: cpu0+port1
+		HWREG_W(vlan_memset_config0, (0xa << 8) + 0x9);
+
+		//RMC forward: to cpu
+		//LED: 60mS
+		//BC storm prev: 31 BC
+		reg = HWREG_R(sw_glb_cntl);
+		HWREG_W(sw_glb_cntl, (reg & (~((0x3 << 25) | (0x3 << 23) | (0x3 << 4)))) |
+			(0x1 << 25) | (0x1 << 23) | (0x1 << 4));
+	} else {
+		mac->cpu_port = 0x1;	// soc0
+		mac->lan_port = 0x3;	// forward to port 0 and 1
+		mac->to_vlan = 0x1;	// vlan group: 0
+		mac->vlan_id = 0x0;	// vlan group: 0
+		comm->enable = 0x3;	// enable lan 0 and 1
+
+		//port 0: VLAN group 0
+		//port 1: VLAN group 0
+		HWREG_W(pvid_config0, (0 << 4) + 0);
+
+		//VLAN group 0: cpu0+port1+port0
+		HWREG_W(vlan_memset_config0, (0x0 << 8) + 0xb);
+
+		//RMC forward: broadcast
+		//LED: 60mS
+		//BC storm prev: 31 BC
+		reg = HWREG_R(sw_glb_cntl);
+		HWREG_W(sw_glb_cntl, (reg & (~((0x3 << 25) | (0x3 << 23) | (0x3 << 4)))) |
+			(0x0 << 25) | (0x1 << 23) | (0x1 << 4));
+	}
+}
+
+void mac_disable_port_sa_learning(void)
+{
+	u32 reg;
+
+	// Disable lan port SA learning.
+	reg = HWREG_R(port_cntl1);
+	HWREG_W(port_cntl1, reg | (0x3 << 8));
+}
+
+void mac_enable_port_sa_learning(void)
+{
+	u32 reg;
+
+	// Disable lan port SA learning.
+	reg = HWREG_R(port_cntl1);
+	HWREG_W(port_cntl1, reg & (~(0x3 << 8)));
+}
+
+void rx_mode_set(struct net_device *net_dev)
+{
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+	u32 mask, reg, rx_mode;
+
+	pr_debug(" net_dev->flags = %08x\n", net_dev->flags);
+
+	mask = (mac->lan_port << 2) | (mac->lan_port << 0);
+	reg = HWREG_R(cpu_cntl);
+
+	if (net_dev->flags & IFF_PROMISC) {	/* Set promiscuous mode */
+		// Allow MC and unknown UC packets
+		rx_mode = (mac->lan_port << 2) | (mac->lan_port << 0);
+	} else if ((!netdev_mc_empty(net_dev) && (net_dev->flags & IFF_MULTICAST)) ||
+		   (net_dev->flags & IFF_ALLMULTI)) {
+		// Allow MC packets
+		rx_mode = (mac->lan_port << 2);
+	} else {
+		// Disable MC and unknown UC packets
+		rx_mode = 0;
+	}
+
+	HWREG_W(cpu_cntl, (reg & (~mask)) | ((~rx_mode) & mask));
+	pr_debug(" cpu_cntl = %08x\n", HWREG_R(cpu_cntl));
+}
+
+static int mdio_access(u8 op_cd, u8 dev_reg_addr, u8 phy_addr, u32 wdata)
+{
+	u32 value, time = 0;
+
+	HWREG_W(phy_cntl_reg0, (wdata << 16) | (op_cd << 13) | (dev_reg_addr << 8) | phy_addr);
+	wmb();			// make sure settings are effective.
+	do {
+		if (++time > MDIO_RW_TIMEOUT_RETRY_NUMBERS) {
+			pr_err(" mdio failed to operate!\n");
+			time = 0;
+		}
+
+		value = HWREG_R(phy_cntl_reg1);
+	} while ((value & 0x3) == 0);
+
+	if (time == 0)
+		return -1;
+
+	return value >> 16;
+}
+
+u32 mdio_read(u32 phy_id, u16 regnum)
+{
+	int ret;
+
+	ret = mdio_access(MDIO_READ_CMD, regnum, phy_id, 0);
+	if (ret < 0)
+		return -EIO;
+
+	return ret;
+}
+
+u32 mdio_write(u32 phy_id, u32 regnum, u16 val)
+{
+	int ret;
+
+	ret = mdio_access(MDIO_WRITE_CMD, regnum, phy_id, val);
+	if (ret < 0)
+		return -EIO;
+
+	return 0;
+}
+
+inline void tx_trigger(void)
+{
+	HWREG_W(cpu_tx_trig, (0x1 << 1));
+}
+
+inline u32 read_sw_int_mask0(void)
+{
+	return HWREG_R(sw_int_mask_0);
+}
+
+inline void write_sw_int_mask0(u32 value)
+{
+	HWREG_W(sw_int_mask_0, value);
+}
+
+inline void write_sw_int_status0(u32 value)
+{
+	HWREG_W(sw_int_status_0, value);
+}
+
+inline u32 read_sw_int_status0(void)
+{
+	return HWREG_R(sw_int_status_0);
+}
+
+inline u32 read_port_ability(void)
+{
+	return HWREG_R(port_ability);
+}
+
+void l2sw_enable_port(struct l2sw_mac *mac)
+{
+	u32 reg;
+
+	//set clock
+	reg = MOON5REG_R(mo4_l2sw_clksw_ctl);
+	MOON5REG_W(mo4_l2sw_clksw_ctl, reg | (0xf << 16) | 0xf);
+
+	//phy address
+	reg = HWREG_R(mac_force_mode);
+	reg = (reg & (~(0x1f << 16))) | ((mac->comm->phy1_addr & 0x1f) << 16);
+	reg = (reg & (~(0x1f << 24))) | ((mac->comm->phy2_addr & 0x1f) << 24);
+	HWREG_W(mac_force_mode, reg);
+	wmb();			// make sure settings are effective.
+}
+
+int phy_cfg(struct l2sw_mac *mac)
+{
+	// Bug workaround:
+	// Flow-control of phy should be enabled. L2SW IP flow-control will refer
+	// to the bit to decide to enable or disable flow-control.
+	mdio_write(mac->comm->phy1_addr, 4, mdio_read(mac->comm->phy1_addr, 4) | (1 << 10));
+	mdio_write(mac->comm->phy2_addr, 4, mdio_read(mac->comm->phy2_addr, 4) | (1 << 10));
+
+	return 0;
+}
diff --git a/drivers/net/ethernet/sunplus/l2sw_hal.h b/drivers/net/ethernet/sunplus/l2sw_hal.h
new file mode 100644
index 0000000..7f04659
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_hal.h
@@ -0,0 +1,47 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#ifndef __L2SW_HAL_H__
+#define __L2SW_HAL_H__
+
+#include "l2sw_register.h"
+#include "l2sw_define.h"
+#include "l2sw_desc.h"
+
+#define HWREG_W(M, N)           (l2sw_reg_base->M = N)
+#define HWREG_R(M)              (l2sw_reg_base->M)
+#define MOON5REG_W(M, N)        (moon5_reg_base->M = N)
+#define MOON5REG_R(M)           (moon5_reg_base->M)
+
+#define MDIO_RW_TIMEOUT_RETRY_NUMBERS 500
+#define MDIO_READ_CMD                 0x02
+#define MDIO_WRITE_CMD                0x01
+
+int  l2sw_reg_base_set(void __iomem *baseaddr);
+int  moon5_reg_base_set(void __iomem *baseaddr);
+void mac_hw_stop(struct l2sw_mac *mac);
+void mac_hw_reset(struct l2sw_mac *mac);
+void mac_hw_start(struct l2sw_mac *mac);
+void mac_hw_addr_set(struct l2sw_mac *mac);
+void mac_hw_addr_del(struct l2sw_mac *mac);
+void mac_addr_table_del_all(void);
+void mac_hw_init(struct l2sw_mac *mac);
+void mac_switch_mode(struct l2sw_mac *mac);
+void mac_disable_port_sa_learning(void);
+void mac_enable_port_sa_learning(void);
+void rx_mode_set(struct net_device *net_dev);
+u32  mdio_read(u32 phy_id, u16 regnum);
+u32  mdio_write(u32 phy_id, u32 regnum, u16 val);
+void tx_trigger(void);
+u32  read_sw_int_mask0(void);
+void write_sw_int_mask0(u32 value);
+void write_sw_int_status0(u32 value);
+u32  read_sw_int_status0(void);
+u32  read_port_ability(void);
+int  phy_cfg(struct l2sw_mac *mac);
+void l2sw_enable_port(struct l2sw_mac *mac);
+void regs_print(void);
+
+#endif
diff --git a/drivers/net/ethernet/sunplus/l2sw_int.c b/drivers/net/ethernet/sunplus/l2sw_int.c
new file mode 100644
index 0000000..865d2d1
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_int.c
@@ -0,0 +1,326 @@ 
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#include "l2sw_define.h"
+#include "l2sw_int.h"
+#include "l2sw_driver.h"
+#include "l2sw_hal.h"
+
+static inline void port_status_change(struct l2sw_mac *mac)
+{
+	u32 reg;
+	struct net_device *net_dev = (struct net_device *)mac->net_dev;
+
+	reg = read_port_ability();
+	if (mac->comm->dual_nic) {
+		if (!netif_carrier_ok(net_dev) && (reg & PORT_ABILITY_LINK_ST_P0)) {
+			netif_carrier_on(net_dev);
+			netif_start_queue(net_dev);
+		} else if (netif_carrier_ok(net_dev) && !(reg & PORT_ABILITY_LINK_ST_P0)) {
+			netif_carrier_off(net_dev);
+			netif_stop_queue(net_dev);
+		}
+
+		if (mac->next_netdev) {
+			struct net_device *ndev2 = mac->next_netdev;
+
+			if (!netif_carrier_ok(ndev2) && (reg & PORT_ABILITY_LINK_ST_P1)) {
+				netif_carrier_on(ndev2);
+				netif_start_queue(ndev2);
+			} else if (netif_carrier_ok(ndev2) && !(reg & PORT_ABILITY_LINK_ST_P1)) {
+				netif_carrier_off(ndev2);
+				netif_stop_queue(ndev2);
+			}
+		}
+	} else {
+		if (!netif_carrier_ok(net_dev) &&
+		    (reg & (PORT_ABILITY_LINK_ST_P1 | PORT_ABILITY_LINK_ST_P0))) {
+			netif_carrier_on(net_dev);
+			netif_start_queue(net_dev);
+		} else if (netif_carrier_ok(net_dev) &&
+		    !(reg & (PORT_ABILITY_LINK_ST_P1 | PORT_ABILITY_LINK_ST_P0))) {
+			netif_carrier_off(net_dev);
+			netif_stop_queue(net_dev);
+		}
+	}
+}
+
+static inline void rx_skb(struct l2sw_mac *mac, struct sk_buff *skb)
+{
+	mac->dev_stats.rx_packets++;
+	mac->dev_stats.rx_bytes += skb->len;
+	netif_receive_skb(skb);
+}
+
+int rx_poll(struct napi_struct *napi, int budget)
+{
+	struct l2sw_common *comm = container_of(napi, struct l2sw_common, rx_napi);
+	struct l2sw_mac *mac = netdev_priv(comm->net_dev);
+	struct sk_buff *skb, *new_skb;
+	struct skb_info *sinfo;
+	struct mac_desc *desc;
+	struct mac_desc *h_desc;
+	u32 rx_pos, pkg_len;
+	u32 cmd;
+	u32 num, rx_count;
+	s32 queue;
+	int ndev2_pkt;
+	struct net_device_stats *dev_stats;
+
+	spin_lock(&comm->rx_lock);
+
+	// Process high-priority queue and then low-priority queue.
+	for (queue = 0; queue < RX_DESC_QUEUE_NUM; queue++) {
+		rx_pos = comm->rx_pos[queue];
+		rx_count = comm->rx_desc_num[queue];
+
+		for (num = 0; num < rx_count; num++) {
+			sinfo = comm->rx_skb_info[queue] + rx_pos;
+			desc = comm->rx_desc[queue] + rx_pos;
+			cmd = desc->cmd1;
+
+			if (cmd & OWN_BIT)
+				break;
+
+			if (comm->dual_nic && ((cmd & PKTSP_MASK) == PKTSP_PORT1)) {
+				struct l2sw_mac *mac2;
+
+				ndev2_pkt = 1;
+				mac2 = (mac->next_netdev) ? netdev_priv(mac->next_netdev) : NULL;
+				dev_stats = (mac2) ? &mac2->dev_stats : &mac->dev_stats;
+			} else {
+				ndev2_pkt = 0;
+				dev_stats = &mac->dev_stats;
+			}
+
+			pkg_len = cmd & LEN_MASK;
+			if (unlikely((cmd & ERR_CODE) || (pkg_len < 64))) {
+				dev_stats->rx_length_errors++;
+				dev_stats->rx_dropped++;
+				goto NEXT;
+			}
+
+			if (unlikely(cmd & RX_IP_CHKSUM_BIT)) {
+				dev_stats->rx_crc_errors++;
+				dev_stats->rx_dropped++;
+				goto NEXT;
+			}
+
+			/* allocate an skbuff for receiving, and it's an inline function */
+			new_skb = __dev_alloc_skb(comm->rx_desc_buff_size + RX_OFFSET,
+						  GFP_ATOMIC | GFP_DMA);
+			if (unlikely(!new_skb)) {
+				dev_stats->rx_dropped++;
+				goto NEXT;
+			}
+			new_skb->dev = mac->net_dev;
+
+			dma_unmap_single(&mac->pdev->dev, sinfo->mapping, comm->rx_desc_buff_size,
+					 DMA_FROM_DEVICE);
+
+			skb = sinfo->skb;
+			skb->ip_summed = CHECKSUM_NONE;
+
+			/*skb_put will judge if tail exceeds end, but __skb_put won't */
+			__skb_put(skb, (pkg_len - 4 > comm->rx_desc_buff_size) ?
+				       comm->rx_desc_buff_size : pkg_len - 4);
+
+			sinfo->mapping = dma_map_single(&mac->pdev->dev, new_skb->data,
+							comm->rx_desc_buff_size,
+							DMA_FROM_DEVICE);
+			sinfo->skb = new_skb;
+
+			if (ndev2_pkt) {
+				struct net_device *netdev2 = mac->next_netdev;
+
+				if (netdev2) {
+					skb->protocol = eth_type_trans(skb, netdev2);
+					rx_skb(netdev_priv(netdev2), skb);
+				}
+			} else {
+				skb->protocol = eth_type_trans(skb, mac->net_dev);
+				rx_skb(mac, skb);
+			}
+
+			desc->addr1 = sinfo->mapping;
+
+NEXT:
+			desc->cmd2 = (rx_pos == comm->rx_desc_num[queue] - 1) ?
+				     EOR_BIT | MAC_RX_LEN_MAX : MAC_RX_LEN_MAX;
+			wmb();	// Set OWN_BIT after other fields of descriptor are effective.
+			desc->cmd1 = OWN_BIT | (comm->rx_desc_buff_size & LEN_MASK);
+
+			NEXT_RX(queue, rx_pos);
+
+			// If there are packets in high-priority queue,
+			// stop processing low-priority queue.
+			if ((queue == 1) && ((h_desc->cmd1 & OWN_BIT) == 0))
+				break;
+		}
+
+		comm->rx_pos[queue] = rx_pos;
+
+		// Save pointer to last rx descriptor of high-priority queue.
+		if (queue == 0)
+			h_desc = comm->rx_desc[queue] + rx_pos;
+	}
+
+	spin_unlock(&comm->rx_lock);
+
+	wmb();			// make sure settings are effective.
+	write_sw_int_mask0(read_sw_int_mask0() & ~MAC_INT_RX);
+
+	napi_complete(napi);
+	return 0;
+}
+
+int tx_poll(struct napi_struct *napi, int budget)
+{
+	struct l2sw_common *comm = container_of(napi, struct l2sw_common, tx_napi);
+	struct l2sw_mac *mac = netdev_priv(comm->net_dev);
+	u32 tx_done_pos;
+	u32 cmd;
+	struct skb_info *skbinfo;
+	struct l2sw_mac *smac;
+
+	spin_lock(&comm->tx_lock);
+
+	tx_done_pos = comm->tx_done_pos;
+	while ((tx_done_pos != comm->tx_pos) || (comm->tx_desc_full == 1)) {
+		cmd = comm->tx_desc[tx_done_pos].cmd1;
+		if (cmd & OWN_BIT)
+			break;
+
+		skbinfo = &comm->tx_temp_skb_info[tx_done_pos];
+		if (unlikely(!skbinfo->skb))
+			pr_err(" skb is null!\n");
+
+		smac = mac;
+		if (mac->next_netdev && ((cmd & TO_VLAN_MASK) == TO_VLAN_GROUP1))
+			smac = netdev_priv(mac->next_netdev);
+
+		if (unlikely(cmd & (ERR_CODE))) {
+			smac->dev_stats.tx_errors++;
+		} else {
+			smac->dev_stats.tx_packets++;
+			smac->dev_stats.tx_bytes += skbinfo->len;
+		}
+
+		dma_unmap_single(&mac->pdev->dev, skbinfo->mapping, skbinfo->len, DMA_TO_DEVICE);
+		skbinfo->mapping = 0;
+		dev_kfree_skb_irq(skbinfo->skb);
+		skbinfo->skb = NULL;
+
+		NEXT_TX(tx_done_pos);
+		if (comm->tx_desc_full == 1)
+			comm->tx_desc_full = 0;
+	}
+
+	comm->tx_done_pos = tx_done_pos;
+	if (!comm->tx_desc_full) {
+		if (netif_queue_stopped(mac->net_dev))
+			netif_wake_queue(mac->net_dev);
+
+		if (mac->next_netdev) {
+			if (netif_queue_stopped(mac->next_netdev))
+				netif_wake_queue(mac->next_netdev);
+		}
+	}
+
+	spin_unlock(&comm->tx_lock);
+
+	wmb();			// make sure settings are effective.
+	write_sw_int_mask0(read_sw_int_mask0() & ~MAC_INT_TX);
+
+	napi_complete(napi);
+	return 0;
+}
+
+irqreturn_t ethernet_interrupt(int irq, void *dev_id)
+{
+	struct net_device *net_dev;
+	struct l2sw_mac *mac;
+	struct l2sw_common *comm;
+	u32 status;
+
+	net_dev = (struct net_device *)dev_id;
+	if (unlikely(!net_dev)) {
+		pr_err(" net_dev is null!\n");
+		return -1;
+	}
+
+	mac = netdev_priv(net_dev);
+	comm = mac->comm;
+
+	status = read_sw_int_status0();
+	if (unlikely(status == 0)) {
+		pr_err(" Interrput status is null!\n");
+		goto OUT;
+	}
+	write_sw_int_status0(status);
+
+	if (status & MAC_INT_RX) {
+		// Disable RX interrupts.
+		write_sw_int_mask0(read_sw_int_mask0() | MAC_INT_RX);
+
+		if (unlikely(status & MAC_INT_RX_DES_ERR)) {
+			pr_err(" Illegal RX Descriptor!\n");
+			mac->dev_stats.rx_fifo_errors++;
+		}
+		if (napi_schedule_prep(&comm->rx_napi))
+			__napi_schedule(&comm->rx_napi);
+	}
+
+	if (status & MAC_INT_TX) {
+		// Disable TX interrupts.
+		write_sw_int_mask0(read_sw_int_mask0() | MAC_INT_TX);
+
+		if (unlikely(status & MAC_INT_TX_DES_ERR)) {
+			pr_err(" Illegal TX Descriptor Error\n");
+			mac->dev_stats.tx_fifo_errors++;
+			mac_soft_reset(mac);
+			wmb();			// make sure settings are effective.
+			write_sw_int_mask0(read_sw_int_mask0() & ~MAC_INT_TX);
+		} else {
+			if (napi_schedule_prep(&comm->tx_napi))
+				__napi_schedule(&comm->tx_napi);
+		}
+	}
+
+	if (status & MAC_INT_PORT_ST_CHG)	/* link status changed */
+		port_status_change(mac);
+
+OUT:
+	return IRQ_HANDLED;
+}
+
+int l2sw_get_irq(struct platform_device *pdev, struct l2sw_common *comm)
+{
+	struct resource *res;
+
+	res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
+	if (!res) {
+		pr_err(" No IRQ resource found!\n");
+		return -1;
+	}
+
+	pr_debug(" res->name = \"%s\", res->start = %pa\n", res->name, &res->start);
+	comm->irq = res->start;
+	return 0;
+}
+
+int l2sw_request_irq(struct platform_device *pdev, struct l2sw_common *comm,
+		     struct net_device *net_dev)
+{
+	int rc;
+
+	rc = devm_request_irq(&pdev->dev, comm->irq, ethernet_interrupt, 0,
+			      net_dev->name, net_dev);
+	if (rc != 0) {
+		pr_err(" Failed to request irq #%d for \"%s\" (rc = %d)!\n",
+		       net_dev->irq, net_dev->name, rc);
+	}
+	return rc;
+}
diff --git a/drivers/net/ethernet/sunplus/l2sw_int.h b/drivers/net/ethernet/sunplus/l2sw_int.h
new file mode 100644
index 0000000..b459481
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_int.h
@@ -0,0 +1,16 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#ifndef __L2SW_INT_H__
+#define __L2SW_INT_H__
+
+int rx_poll(struct napi_struct *napi, int budget);
+int tx_poll(struct napi_struct *napi, int budget);
+irqreturn_t ethernet_interrupt(int irq, void *dev_id);
+int l2sw_get_irq(struct platform_device *pdev, struct l2sw_common *comm);
+int l2sw_request_irq(struct platform_device *pdev, struct l2sw_common *comm,
+		     struct net_device *net_dev);
+
+#endif
diff --git a/drivers/net/ethernet/sunplus/l2sw_mac.c b/drivers/net/ethernet/sunplus/l2sw_mac.c
new file mode 100644
index 0000000..4b5b1d7
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_mac.c
@@ -0,0 +1,68 @@ 
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#include "l2sw_mac.h"
+
+bool mac_init(struct l2sw_mac *mac)
+{
+	u32 i;
+
+	for (i = 0; i < RX_DESC_QUEUE_NUM; i++)
+		mac->comm->rx_pos[i] = 0;
+	mb();			// make sure settings are effective.
+
+	mac_hw_init(mac);
+	mb();			// make sure settings are effective.
+
+	return 1;
+}
+
+void mac_soft_reset(struct l2sw_mac *mac)
+{
+	u32 i;
+	struct net_device *net_dev2;
+
+	if (netif_carrier_ok(mac->net_dev)) {
+		netif_carrier_off(mac->net_dev);
+		netif_stop_queue(mac->net_dev);
+	}
+
+	net_dev2 = mac->next_netdev;
+	if (net_dev2) {
+		if (netif_carrier_ok(net_dev2)) {
+			netif_carrier_off(net_dev2);
+			netif_stop_queue(net_dev2);
+		}
+	}
+
+	mac_hw_reset(mac);
+	mac_hw_stop(mac);
+	mb();			// make sure settings are effective.
+
+	rx_descs_flush(mac->comm);
+	mac->comm->tx_pos = 0;
+	mac->comm->tx_done_pos = 0;
+	mac->comm->tx_desc_full = 0;
+
+	for (i = 0; i < RX_DESC_QUEUE_NUM; i++)
+		mac->comm->rx_pos[i] = 0;
+	mb();			// make sure settings are effective.
+
+	mac_hw_init(mac);
+	mac_hw_start(mac);
+	mb();			// make sure settings are effective.
+
+	if (!netif_carrier_ok(mac->net_dev)) {
+		netif_carrier_on(mac->net_dev);
+		netif_start_queue(mac->net_dev);
+	}
+
+	if (net_dev2) {
+		if (!netif_carrier_ok(net_dev2)) {
+			netif_carrier_on(net_dev2);
+			netif_start_queue(net_dev2);
+		}
+	}
+}
diff --git a/drivers/net/ethernet/sunplus/l2sw_mac.h b/drivers/net/ethernet/sunplus/l2sw_mac.h
new file mode 100644
index 0000000..a053130
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_mac.h
@@ -0,0 +1,24 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#ifndef __L2SW_MAC_H__
+#define __L2SW_MAC_H__
+
+#include "l2sw_define.h"
+#include "l2sw_hal.h"
+
+bool mac_init(struct l2sw_mac *mac);
+
+void mac_soft_reset(struct l2sw_mac *mac);
+
+//calculate the empty tx descriptor number
+#define TX_DESC_AVAIL(mac) \
+	(((mac)->tx_pos != (mac)->tx_done_pos) ? \
+	(((mac)->tx_done_pos < (mac)->tx_pos) ? \
+	(TX_DESC_NUM - ((mac)->tx_pos - (mac)->tx_done_pos)) : \
+	((mac)->tx_done_pos - (mac)->tx_pos)) : \
+	((mac)->tx_desc_full ? 0 : TX_DESC_NUM))
+
+#endif
diff --git a/drivers/net/ethernet/sunplus/l2sw_mdio.c b/drivers/net/ethernet/sunplus/l2sw_mdio.c
new file mode 100644
index 0000000..9008ab9
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_mdio.c
@@ -0,0 +1,118 @@ 
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#include "l2sw_mdio.h"
+
+static int mii_read(struct mii_bus *bus, int phy_id, int regnum)
+{
+	return mdio_read(phy_id, regnum);
+}
+
+static int mii_write(struct mii_bus *bus, int phy_id, int regnum, u16 val)
+{
+	return mdio_write(phy_id, regnum, val);
+}
+
+u32 mdio_init(struct platform_device *pdev, struct net_device *net_dev)
+{
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+	struct mii_bus *mii_bus;
+	struct device_node *mdio_node;
+	u32 ret;
+
+	mii_bus = mdiobus_alloc();
+	if (!mii_bus) {
+		pr_err(" Failed to allocate mdio_bus memory!\n");
+		return -ENOMEM;
+	}
+
+	mii_bus->name = "sunplus_mii_bus";
+	mii_bus->parent = &pdev->dev;
+	mii_bus->priv = mac;
+	mii_bus->read = mii_read;
+	mii_bus->write = mii_write;
+	snprintf(mii_bus->id, MII_BUS_ID_SIZE, "%s-mii", dev_name(&pdev->dev));
+
+	mdio_node = of_get_parent(mac->comm->phy1_node);
+	ret = of_mdiobus_register(mii_bus, mdio_node);
+	if (ret) {
+		pr_err(" Failed to register mii bus (ret = %d)!\n", ret);
+		mdiobus_free(mii_bus);
+		return ret;
+	}
+
+	mac->comm->mii_bus = mii_bus;
+	return ret;
+}
+
+void mdio_remove(struct net_device *net_dev)
+{
+	struct l2sw_mac *mac = netdev_priv(net_dev);
+
+	if (mac->comm->mii_bus) {
+		mdiobus_unregister(mac->comm->mii_bus);
+		mdiobus_free(mac->comm->mii_bus);
+		mac->comm->mii_bus = NULL;
+	}
+}
+
+static void mii_linkchange(struct net_device *netdev)
+{
+}
+
+int mac_phy_probe(struct net_device *netdev)
+{
+	struct l2sw_mac *mac = netdev_priv(netdev);
+	struct phy_device *phydev;
+	int i;
+
+	phydev = of_phy_connect(mac->net_dev, mac->comm->phy1_node, mii_linkchange,
+				0, PHY_INTERFACE_MODE_RGMII_ID);
+	if (!phydev) {
+		pr_err(" \"%s\" has no phy found\n", netdev->name);
+		return -1;
+	}
+
+	if (mac->comm->phy2_node) {
+		of_phy_connect(mac->net_dev, mac->comm->phy2_node, mii_linkchange,
+			       0, PHY_INTERFACE_MODE_RGMII_ID);
+	}
+
+	linkmode_clear_bit(ETHTOOL_LINK_MODE_Pause_BIT, phydev->supported);
+	linkmode_clear_bit(ETHTOOL_LINK_MODE_Asym_Pause_BIT, phydev->supported);
+
+	for (i = 0; i < sizeof(phydev->supported) / sizeof(long); i++)
+		phydev->advertising[i] = phydev->supported[i];
+
+	phydev->irq = PHY_MAC_INTERRUPT;
+	mac->comm->phy_dev = phydev;
+
+	return 0;
+}
+
+void mac_phy_start(struct net_device *netdev)
+{
+	struct l2sw_mac *mac = netdev_priv(netdev);
+
+	phy_start(mac->comm->phy_dev);
+}
+
+void mac_phy_stop(struct net_device *netdev)
+{
+	struct l2sw_mac *mac = netdev_priv(netdev);
+
+	if (mac->comm->phy_dev)
+		phy_stop(mac->comm->phy_dev);
+}
+
+void mac_phy_remove(struct net_device *netdev)
+{
+	struct l2sw_mac *mac = netdev_priv(netdev);
+
+	if (mac->comm->phy_dev) {
+		phy_disconnect(mac->comm->phy_dev);
+		mac->comm->phy_dev = NULL;
+	}
+}
diff --git a/drivers/net/ethernet/sunplus/l2sw_mdio.h b/drivers/net/ethernet/sunplus/l2sw_mdio.h
new file mode 100644
index 0000000..698fc0b
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_mdio.h
@@ -0,0 +1,19 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#ifndef __L2SW_MDIO_H__
+#define __L2SW_MDIO_H__
+
+#include "l2sw_define.h"
+#include "l2sw_hal.h"
+
+u32  mdio_init(struct platform_device *pdev, struct net_device *net_dev);
+void mdio_remove(struct net_device *net_dev);
+int  mac_phy_probe(struct net_device *netdev);
+void mac_phy_start(struct net_device *netdev);
+void mac_phy_stop(struct net_device *netdev);
+void mac_phy_remove(struct net_device *netdev);
+
+#endif
diff --git a/drivers/net/ethernet/sunplus/l2sw_register.h b/drivers/net/ethernet/sunplus/l2sw_register.h
new file mode 100644
index 0000000..a54a296
--- /dev/null
+++ b/drivers/net/ethernet/sunplus/l2sw_register.h
@@ -0,0 +1,99 @@ 
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright Sunplus Technology Co., Ltd.
+ *       All rights reserved.
+ */
+
+#ifndef __L2SW_REGISTER_H__
+#define __L2SW_REGISTER_H__
+
+#include "l2sw_define.h"
+
+/* TYPE: RegisterFile_L2SW */
+struct l2sw_reg {
+	u32 sw_int_status_0;
+	u32 sw_int_mask_0;
+	u32 fl_cntl_th;
+	u32 cpu_fl_cntl_th;
+	u32 pri_fl_cntl;
+	u32 vlan_pri_th;
+	u32 en_tos_bus;
+	u32 tos_map0;
+	u32 tos_map1;
+	u32 tos_map2;
+	u32 tos_map3;
+	u32 tos_map4;
+	u32 tos_map5;
+	u32 tos_map6;
+	u32 tos_map7;
+	u32 global_que_status;
+	u32 addr_tbl_srch;
+	u32 addr_tbl_st;
+	u32 mac_ad_ser0;
+	u32 mac_ad_ser1;
+	u32 wt_mac_ad0;
+	u32 w_mac_15_0;
+	u32 w_mac_47_16;
+	u32 pvid_config0;
+	u32 pvid_config1;
+	u32 vlan_memset_config0;
+	u32 vlan_memset_config1;
+	u32 port_ability;
+	u32 port_st;
+	u32 cpu_cntl;
+	u32 port_cntl0;
+	u32 port_cntl1;
+	u32 port_cntl2;
+	u32 sw_glb_cntl;
+	u32 l2sw_rsv1;
+	u32 led_port0;
+	u32 led_port1;
+	u32 led_port2;
+	u32 led_port3;
+	u32 led_port4;
+	u32 watch_dog_trig_rst;
+	u32 watch_dog_stop_cpu;
+	u32 phy_cntl_reg0;
+	u32 phy_cntl_reg1;
+	u32 mac_force_mode;
+	u32 vlan_group_config0;
+	u32 vlan_group_config1;
+	u32 flow_ctrl_th3;
+	u32 queue_status_0;
+	u32 debug_cntl;
+	u32 l2sw_rsv2;
+	u32 mem_test_info;
+	u32 sw_int_status_1;
+	u32 sw_int_mask_1;
+	u32 l2sw_rsv3[76];
+	u32 cpu_tx_trig;
+	u32 tx_hbase_addr_0;
+	u32 tx_lbase_addr_0;
+	u32 rx_hbase_addr_0;
+	u32 rx_lbase_addr_0;
+	u32 tx_hw_addr_0;
+	u32 tx_lw_addr_0;
+	u32 rx_hw_addr_0;
+	u32 rx_lw_addr_0;
+	u32 cpu_port_cntl_reg_0;
+	u32 tx_hbase_addr_1;
+	u32 tx_lbase_addr_1;
+	u32 rx_hbase_addr_1;
+	u32 rx_lbase_addr_1;
+	u32 tx_hw_addr_1;
+	u32 tx_lw_addr_1;
+	u32 rx_hw_addr_1;
+	u32 rx_lw_addr_1;
+	u32 cpu_port_cntl_reg_1;
+};
+
+/* TYPE: RegisterFile_MOON5 */
+struct moon5_reg {
+	u32 mo5_thermal_ctl_0;
+	u32 mo5_thermal_ctl_1;
+	u32 mo4_thermal_ctl_2;
+	u32 mo4_thermal_ctl_3;
+	u32 mo4_tmds_l2sw_ctl;
+	u32 mo4_l2sw_clksw_ctl;
+};
+
+#endif