* [PATCH net-next v3 01/12] net: libwx: add mailbox api for wangxun vf drivers
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 02/12] net: libwx: add base vf api for " Mengyuan Lou
` (11 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Implements the mailbox interfaces for Wangxun vf drivers which
will be used in txgbevf and ngbevf.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
drivers/net/ethernet/wangxun/libwx/wx_mbx.c | 243 +++++++++++++++++++
drivers/net/ethernet/wangxun/libwx/wx_mbx.h | 22 ++
drivers/net/ethernet/wangxun/libwx/wx_type.h | 3 +
3 files changed, 268 insertions(+)
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_mbx.c b/drivers/net/ethernet/wangxun/libwx/wx_mbx.c
index 73af5f11c3bd..2aa03eadf064 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_mbx.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_mbx.c
@@ -174,3 +174,246 @@ int wx_check_for_rst_pf(struct wx *wx, u16 vf)
return 0;
}
+
+static u32 wx_read_v2p_mailbox(struct wx *wx)
+{
+ u32 mailbox = rd32(wx, WX_VXMAILBOX);
+
+ mailbox |= wx->mbx.mailbox;
+ wx->mbx.mailbox |= mailbox & WX_VXMAILBOX_R2C_BITS;
+
+ return mailbox;
+}
+
+static u32 wx_mailbox_get_lock_vf(struct wx *wx)
+{
+ wr32(wx, WX_VXMAILBOX, WX_VXMAILBOX_VFU);
+ return wx_read_v2p_mailbox(wx);
+}
+
+/**
+ * wx_obtain_mbx_lock_vf - obtain mailbox lock
+ * @wx: pointer to the HW structure
+ *
+ * Return: return 0 on success and -EBUSY on failure
+ **/
+static int wx_obtain_mbx_lock_vf(struct wx *wx)
+{
+ int count = 5, ret;
+ u32 mailbox;
+
+ ret = readx_poll_timeout_atomic(wx_mailbox_get_lock_vf, wx, mailbox,
+ (mailbox & WX_VXMAILBOX_VFU),
+ 1, count);
+ if (ret)
+ wx_err(wx, "Failed to obtain mailbox lock for VF.\n");
+
+ return ret;
+}
+
+static int wx_check_for_bit_vf(struct wx *wx, u32 mask)
+{
+ u32 mailbox = wx_read_v2p_mailbox(wx);
+
+ wx->mbx.mailbox &= ~mask;
+
+ return (mailbox & mask ? 0 : -EBUSY);
+}
+
+/**
+ * wx_check_for_ack_vf - checks to see if the PF has ACK'd
+ * @wx: pointer to the HW structure
+ *
+ * Return: return 0 if the PF has set the status bit or else -EBUSY
+ **/
+static int wx_check_for_ack_vf(struct wx *wx)
+{
+ /* read clear the pf ack bit */
+ return wx_check_for_bit_vf(wx, WX_VXMAILBOX_PFACK);
+}
+
+/**
+ * wx_check_for_msg_vf - checks to see if the PF has sent mail
+ * @wx: pointer to the HW structure
+ *
+ * Return: return 0 if the PF has got req bit or else -EBUSY
+ **/
+int wx_check_for_msg_vf(struct wx *wx)
+{
+ /* read clear the pf sts bit */
+ return wx_check_for_bit_vf(wx, WX_VXMAILBOX_PFSTS);
+}
+
+/**
+ * wx_check_for_rst_vf - checks to see if the PF has reset
+ * @wx: pointer to the HW structure
+ *
+ * Return: return 0 if the PF has set the reset done and -EBUSY on failure
+ **/
+int wx_check_for_rst_vf(struct wx *wx)
+{
+ /* read clear the pf reset done bit */
+ return wx_check_for_bit_vf(wx,
+ WX_VXMAILBOX_RSTD |
+ WX_VXMAILBOX_RSTI);
+}
+
+/**
+ * wx_poll_for_msg - Wait for message notification
+ * @wx: pointer to the HW structure
+ *
+ * Return: return 0 if the VF has successfully received a message notification
+ **/
+static int wx_poll_for_msg(struct wx *wx)
+{
+ struct wx_mbx_info *mbx = &wx->mbx;
+ u32 val;
+
+ return readx_poll_timeout_atomic(wx_check_for_msg_vf, wx, val,
+ (val == 0), mbx->udelay, mbx->timeout);
+}
+
+/**
+ * wx_poll_for_ack - Wait for message acknowledgment
+ * @wx: pointer to the HW structure
+ *
+ * Return: return 0 if the VF has successfully received a message ack
+ **/
+static int wx_poll_for_ack(struct wx *wx)
+{
+ struct wx_mbx_info *mbx = &wx->mbx;
+ u32 val;
+
+ return readx_poll_timeout_atomic(wx_check_for_ack_vf, wx, val,
+ (val == 0), mbx->udelay, mbx->timeout);
+}
+
+/**
+ * wx_read_posted_mbx - Wait for message notification and receive message
+ * @wx: pointer to the HW structure
+ * @msg: The message buffer
+ * @size: Length of buffer
+ *
+ * Return: returns 0 if it successfully received a message notification and
+ * copied it into the receive buffer.
+ **/
+int wx_read_posted_mbx(struct wx *wx, u32 *msg, u16 size)
+{
+ int ret;
+
+ ret = wx_poll_for_msg(wx);
+ /* if ack received read message, otherwise we timed out */
+ if (ret)
+ return ret;
+
+ return wx_read_mbx_vf(wx, msg, size);
+}
+
+/**
+ * wx_write_posted_mbx - Write a message to the mailbox, wait for ack
+ * @wx: pointer to the HW structure
+ * @msg: The message buffer
+ * @size: Length of buffer
+ *
+ * Return: returns 0 if it successfully copied message into the buffer and
+ * received an ack to that message within delay * timeout period
+ **/
+int wx_write_posted_mbx(struct wx *wx, u32 *msg, u16 size)
+{
+ int ret;
+
+ /* send msg */
+ ret = wx_write_mbx_vf(wx, msg, size);
+ /* if msg sent wait until we receive an ack */
+ if (ret)
+ return ret;
+
+ return wx_poll_for_ack(wx);
+}
+
+/**
+ * wx_write_mbx_vf - Write a message to the mailbox
+ * @wx: pointer to the HW structure
+ * @msg: The message buffer
+ * @size: Length of buffer
+ *
+ * Return: returns 0 if it successfully copied message into the buffer
+ **/
+int wx_write_mbx_vf(struct wx *wx, u32 *msg, u16 size)
+{
+ struct wx_mbx_info *mbx = &wx->mbx;
+ int ret, i;
+
+ /* mbx->size is up to 15 */
+ if (size > mbx->size) {
+ wx_err(wx, "Invalid mailbox message size %d", size);
+ return -EINVAL;
+ }
+
+ /* lock the mailbox to prevent pf/vf race condition */
+ ret = wx_obtain_mbx_lock_vf(wx);
+ if (ret)
+ return ret;
+
+ /* flush msg and acks as we are overwriting the message buffer */
+ wx_check_for_msg_vf(wx);
+ wx_check_for_ack_vf(wx);
+
+ /* copy the caller specified message to the mailbox memory buffer */
+ for (i = 0; i < size; i++)
+ wr32a(wx, WX_VXMBMEM, i, msg[i]);
+
+ /* Drop VFU and interrupt the PF to tell it a message has been sent */
+ wr32(wx, WX_VXMAILBOX, WX_VXMAILBOX_REQ);
+
+ return 0;
+}
+
+/**
+ * wx_read_mbx_vf - Reads a message from the inbox intended for vf
+ * @wx: pointer to the HW structure
+ * @msg: The message buffer
+ * @size: Length of buffer
+ *
+ * Return: returns 0 if it successfully copied message into the buffer
+ **/
+int wx_read_mbx_vf(struct wx *wx, u32 *msg, u16 size)
+{
+ struct wx_mbx_info *mbx = &wx->mbx;
+ int ret, i;
+
+ /* limit read to size of mailbox and mbx->size is up to 15 */
+ if (size > mbx->size)
+ size = mbx->size;
+
+ /* lock the mailbox to prevent pf/vf race condition */
+ ret = wx_obtain_mbx_lock_vf(wx);
+ if (ret)
+ return ret;
+
+ /* copy the message from the mailbox memory buffer */
+ for (i = 0; i < size; i++)
+ msg[i] = rd32a(wx, WX_VXMBMEM, i);
+
+ /* Acknowledge receipt and release mailbox, then we're done */
+ wr32(wx, WX_VXMAILBOX, WX_VXMAILBOX_ACK);
+
+ return 0;
+}
+
+int wx_init_mbx_params_vf(struct wx *wx)
+{
+ wx->vfinfo = kzalloc(sizeof(struct vf_data_storage),
+ GFP_KERNEL);
+ if (!wx->vfinfo)
+ return -ENOMEM;
+
+ /* Initialize mailbox parameters */
+ wx->mbx.size = WX_VXMAILBOX_SIZE;
+ wx->mbx.mailbox = WX_VXMAILBOX;
+ wx->mbx.udelay = 10;
+ wx->mbx.timeout = 1000;
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_init_mbx_params_vf);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_mbx.h b/drivers/net/ethernet/wangxun/libwx/wx_mbx.h
index 05aae138dbc3..82df9218490a 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_mbx.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_mbx.h
@@ -11,6 +11,20 @@
#define WX_PXMAILBOX_ACK BIT(1) /* Ack message recv'd from VF */
#define WX_PXMAILBOX_PFU BIT(3) /* PF owns the mailbox buffer */
+/* VF Registers */
+#define WX_VXMAILBOX 0x600
+#define WX_VXMAILBOX_REQ BIT(0) /* Request for PF Ready bit */
+#define WX_VXMAILBOX_ACK BIT(1) /* Ack PF message received */
+#define WX_VXMAILBOX_VFU BIT(2) /* VF owns the mailbox buffer */
+#define WX_VXMAILBOX_PFU BIT(3) /* PF owns the mailbox buffer */
+#define WX_VXMAILBOX_PFSTS BIT(4) /* PF wrote a message in the MB */
+#define WX_VXMAILBOX_PFACK BIT(5) /* PF ack the previous VF msg */
+#define WX_VXMAILBOX_RSTI BIT(6) /* PF has reset indication */
+#define WX_VXMAILBOX_RSTD BIT(7) /* PF has indicated reset done */
+#define WX_VXMAILBOX_R2C_BITS (WX_VXMAILBOX_RSTD | \
+ WX_VXMAILBOX_PFSTS | WX_VXMAILBOX_PFACK)
+
+#define WX_VXMBMEM 0x00C00 /* 16*4B */
#define WX_PXMBMEM(i) (0x5000 + (64 * (i))) /* i=[0,63] */
#define WX_VFLRE(i) (0x4A0 + (4 * (i))) /* i=[0,1] */
@@ -74,4 +88,12 @@ int wx_check_for_rst_pf(struct wx *wx, u16 mbx_id);
int wx_check_for_msg_pf(struct wx *wx, u16 mbx_id);
int wx_check_for_ack_pf(struct wx *wx, u16 mbx_id);
+int wx_read_posted_mbx(struct wx *wx, u32 *msg, u16 size);
+int wx_write_posted_mbx(struct wx *wx, u32 *msg, u16 size);
+int wx_check_for_rst_vf(struct wx *wx);
+int wx_check_for_msg_vf(struct wx *wx);
+int wx_read_mbx_vf(struct wx *wx, u32 *msg, u16 size);
+int wx_write_mbx_vf(struct wx *wx, u32 *msg, u16 size);
+int wx_init_mbx_params_vf(struct wx *wx);
+
#endif /* _WX_MBX_H_ */
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h
index c363379126c0..3d4785865bb2 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_type.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h
@@ -825,6 +825,9 @@ struct wx_bus_info {
struct wx_mbx_info {
u16 size;
+ u32 mailbox;
+ u32 udelay;
+ u32 timeout;
};
struct wx_thermal_sensor_data {
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 02/12] net: libwx: add base vf api for vf drivers
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 01/12] net: libwx: add mailbox api for wangxun vf drivers Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 03/12] net: libwx: add wangxun vf common api Mengyuan Lou
` (10 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Implement mbox_write_and_read_ack functions which are
used to set basic functions like set_mac, get_link.etc
for vf.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
drivers/net/ethernet/wangxun/libwx/Makefile | 1 +
drivers/net/ethernet/wangxun/libwx/wx_hw.c | 2 +-
drivers/net/ethernet/wangxun/libwx/wx_hw.h | 1 +
drivers/net/ethernet/wangxun/libwx/wx_type.h | 1 +
drivers/net/ethernet/wangxun/libwx/wx_vf.c | 473 +++++++++++++++++++
drivers/net/ethernet/wangxun/libwx/wx_vf.h | 63 +++
6 files changed, 540 insertions(+), 1 deletion(-)
create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf.c
create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf.h
diff --git a/drivers/net/ethernet/wangxun/libwx/Makefile b/drivers/net/ethernet/wangxun/libwx/Makefile
index 9b78b604a94e..ddf0bb921676 100644
--- a/drivers/net/ethernet/wangxun/libwx/Makefile
+++ b/drivers/net/ethernet/wangxun/libwx/Makefile
@@ -5,3 +5,4 @@
obj-$(CONFIG_LIBWX) += libwx.o
libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_ptp.o wx_mbx.o wx_sriov.o
+libwx-objs += wx_vf.o
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
index 0f4be72116b8..82dd76f0326e 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
@@ -1107,7 +1107,7 @@ static int wx_write_uc_addr_list(struct net_device *netdev, int pool)
* by the MO field of the MCSTCTRL. The MO field is set during initialization
* to mc_filter_type.
**/
-static u32 wx_mta_vector(struct wx *wx, u8 *mc_addr)
+u32 wx_mta_vector(struct wx *wx, u8 *mc_addr)
{
u32 vector = 0;
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.h b/drivers/net/ethernet/wangxun/libwx/wx_hw.h
index 26a56cba60b9..718015611da6 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_hw.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.h
@@ -29,6 +29,7 @@ void wx_mac_set_default_filter(struct wx *wx, u8 *addr);
int wx_add_mac_filter(struct wx *wx, u8 *addr, u16 pool);
int wx_del_mac_filter(struct wx *wx, u8 *addr, u16 pool);
void wx_flush_sw_mac_table(struct wx *wx);
+u32 wx_mta_vector(struct wx *wx, u8 *mc_addr);
int wx_set_mac(struct net_device *netdev, void *p);
void wx_disable_rx(struct wx *wx);
int wx_set_vf_spoofchk(struct net_device *netdev, int vf, bool setting);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h
index 3d4785865bb2..d14e46ac244a 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_type.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h
@@ -1214,6 +1214,7 @@ struct wx {
void *priv;
u8 __iomem *hw_addr;
+ u8 __iomem *b4_addr; /* vf only */
struct pci_dev *pdev;
struct net_device *netdev;
struct wx_bus_info bus;
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf.c b/drivers/net/ethernet/wangxun/libwx/wx_vf.c
new file mode 100644
index 000000000000..165b83e9098f
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf.c
@@ -0,0 +1,473 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#include <linux/etherdevice.h>
+#include <linux/pci.h>
+
+#include "wx_type.h"
+#include "wx_hw.h"
+#include "wx_mbx.h"
+#include "wx_vf.h"
+
+static void wx_virt_clr_reg(struct wx *wx)
+{
+ u32 vfsrrctl, i;
+
+ /* VRSRRCTL default values (BSIZEPACKET = 2048, BSIZEHEADER = 256) */
+ vfsrrctl = WX_VXRXDCTL_HDRSZ(wx_hdr_sz(WX_RX_HDR_SIZE));
+ vfsrrctl |= WX_VXRXDCTL_BUFSZ(wx_buf_sz(WX_RX_BUF_SIZE));
+
+ /* clear all rxd ctl */
+ for (i = 0; i < WX_VF_MAX_RING_NUMS; i++)
+ wr32m(wx, WX_VXRXDCTL(i),
+ WX_VXRXDCTL_HDRSZ_MASK | WX_VXRXDCTL_BUFSZ_MASK,
+ vfsrrctl);
+
+ rd32(wx, WX_VXSTATUS);
+}
+
+/**
+ * wx_init_hw_vf - virtual function hardware initialization
+ * @wx: pointer to hardware structure
+ *
+ * Initialize the mac address
+ **/
+void wx_init_hw_vf(struct wx *wx)
+{
+ wx_get_mac_addr_vf(wx, wx->mac.addr);
+}
+EXPORT_SYMBOL(wx_init_hw_vf);
+
+static int wx_mbx_write_and_read_reply(struct wx *wx, u32 *req_buf,
+ u32 *resp_buf, u16 size)
+{
+ int ret;
+
+ ret = wx_write_posted_mbx(wx, req_buf, size);
+ if (ret)
+ return ret;
+
+ return wx_read_posted_mbx(wx, resp_buf, size);
+}
+
+/**
+ * wx_reset_hw_vf - Performs hardware reset
+ * @wx: pointer to hardware structure
+ *
+ * Resets the hardware by resetting the transmit and receive units, masks and
+ * clears all interrupts.
+ *
+ * Return: returns 0 on success, negative error code on failure
+ **/
+int wx_reset_hw_vf(struct wx *wx)
+{
+ struct wx_mbx_info *mbx = &wx->mbx;
+ u32 msgbuf[4] = {WX_VF_RESET};
+ u8 *addr = (u8 *)(&msgbuf[1]);
+ u32 b4_buf[16] = {0};
+ u32 timeout = 200;
+ int ret;
+ u32 i;
+
+ /* Call wx stop to disable tx/rx and clear interrupts */
+ wx_stop_adapter_vf(wx);
+
+ /* reset the api version */
+ wx->vfinfo->vf_api = wx_mbox_api_null;
+
+ /* backup msix vectors */
+ if (wx->b4_addr) {
+ for (i = 0; i < 16; i++)
+ b4_buf[i] = readl(wx->b4_addr + i * 4);
+ }
+
+ wr32m(wx, WX_VXCTRL, WX_VXCTRL_RST, WX_VXCTRL_RST);
+ rd32(wx, WX_VXSTATUS);
+
+ /* we cannot reset while the RSTI / RSTD bits are asserted */
+ while (!wx_check_for_rst_vf(wx) && timeout) {
+ timeout--;
+ udelay(5);
+ }
+
+ /* restore msix vectors */
+ if (wx->b4_addr) {
+ for (i = 0; i < 16; i++)
+ writel(b4_buf[i], wx->b4_addr + i * 4);
+ }
+
+ /* amlite: bme */
+ if (wx->mac.type == wx_mac_aml || wx->mac.type == wx_mac_aml40)
+ wr32(wx, WX_VX_PF_BME, WX_VF_BME_ENABLE);
+
+ if (!timeout)
+ return -EBUSY;
+
+ /* Reset VF registers to initial values */
+ wx_virt_clr_reg(wx);
+
+ /* mailbox timeout can now become active */
+ mbx->timeout = 2000;
+
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+
+ if (msgbuf[0] != (WX_VF_RESET | WX_VT_MSGTYPE_ACK) &&
+ msgbuf[0] != (WX_VF_RESET | WX_VT_MSGTYPE_NACK))
+ return -EINVAL;
+
+ if (msgbuf[0] == (WX_VF_RESET | WX_VT_MSGTYPE_ACK))
+ ether_addr_copy(wx->mac.perm_addr, addr);
+
+ wx->mac.mc_filter_type = msgbuf[3];
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_reset_hw_vf);
+
+/**
+ * wx_stop_adapter_vf - Generic stop Tx/Rx units
+ * @wx: pointer to hardware structure
+ *
+ * Clears interrupts, disables transmit and receive units.
+ **/
+void wx_stop_adapter_vf(struct wx *wx)
+{
+ u32 reg_val;
+ u16 i;
+
+ /* Clear interrupt mask to stop from interrupts being generated */
+ wr32(wx, WX_VXIMS, WX_VF_IRQ_CLEAR_MASK);
+
+ /* Clear any pending interrupts, flush previous writes */
+ wr32(wx, WX_VXICR, U32_MAX);
+
+ /* Disable the transmit unit. Each queue must be disabled. */
+ for (i = 0; i < wx->mac.max_tx_queues; i++)
+ wr32(wx, WX_VXTXDCTL(i), WX_VXTXDCTL_FLUSH);
+
+ /* Disable the receive unit by stopping each queue */
+ for (i = 0; i < wx->mac.max_rx_queues; i++) {
+ reg_val = rd32(wx, WX_VXRXDCTL(i));
+ reg_val &= ~WX_VXRXDCTL_ENABLE;
+ wr32(wx, WX_VXRXDCTL(i), reg_val);
+ }
+ /* Clear packet split and pool config */
+ wr32(wx, WX_VXMRQC, 0);
+
+ /* flush all queues disables */
+ rd32(wx, WX_VXSTATUS);
+}
+EXPORT_SYMBOL(wx_stop_adapter_vf);
+
+/**
+ * wx_set_rar_vf - set device MAC address
+ * @wx: pointer to hardware structure
+ * @index: Receive address register to write
+ * @addr: Address to put into receive address register
+ * @enable_addr: set flag that address is active
+ *
+ * Return: returns 0 on success, negative error code on failure
+ **/
+int wx_set_rar_vf(struct wx *wx, u32 index, u8 *addr, u32 enable_addr)
+{
+ u32 msgbuf[3] = {WX_VF_SET_MAC_ADDR};
+ u8 *msg_addr = (u8 *)(&msgbuf[1]);
+ int ret;
+
+ memcpy(msg_addr, addr, ETH_ALEN);
+
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+ msgbuf[0] &= ~WX_VT_MSGTYPE_CTS;
+
+ /* if nacked the address was rejected, use "perm_addr" */
+ if (msgbuf[0] == (WX_VF_SET_MAC_ADDR | WX_VT_MSGTYPE_NACK)) {
+ wx_get_mac_addr_vf(wx, wx->mac.addr);
+ return -EINVAL;
+ }
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_set_rar_vf);
+
+/**
+ * wx_update_mc_addr_list_vf - Update Multicast addresses
+ * @wx: pointer to the HW structure
+ * @netdev: pointer to the net device structure
+ *
+ * Updates the Multicast Table Array.
+ *
+ * Return: returns 0 on success, negative error code on failure
+ **/
+int wx_update_mc_addr_list_vf(struct wx *wx, struct net_device *netdev)
+{
+ u32 msgbuf[WX_VXMAILBOX_SIZE] = {WX_VF_SET_MULTICAST};
+ u16 *vector_l = (u16 *)&msgbuf[1];
+ struct netdev_hw_addr *ha;
+ u32 cnt, i;
+
+ cnt = netdev_mc_count(netdev);
+ if (cnt > 28)
+ cnt = 28;
+ msgbuf[0] |= cnt << WX_VT_MSGINFO_SHIFT;
+
+ i = 0;
+ netdev_for_each_mc_addr(ha, netdev) {
+ if (i == cnt)
+ break;
+ if (is_link_local_ether_addr(ha->addr))
+ continue;
+
+ vector_l[i++] = wx_mta_vector(wx, ha->addr);
+ }
+
+ return wx_write_posted_mbx(wx, msgbuf, ARRAY_SIZE(msgbuf));
+}
+EXPORT_SYMBOL(wx_update_mc_addr_list_vf);
+
+/**
+ * wx_update_xcast_mode_vf - Update Multicast mode
+ * @wx: pointer to the HW structure
+ * @xcast_mode: new multicast mode
+ *
+ * Updates the Multicast Mode of VF.
+ *
+ * Return: returns 0 on success, negative error code on failure
+ **/
+int wx_update_xcast_mode_vf(struct wx *wx, int xcast_mode)
+{
+ u32 msgbuf[2] = {WX_VF_UPDATE_XCAST_MODE, xcast_mode};
+ int ret = 0;
+
+ if (wx->vfinfo->vf_api < wx_mbox_api_13)
+ return -EINVAL;
+
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+
+ msgbuf[0] &= ~WX_VT_MSGTYPE_CTS;
+ if (msgbuf[0] == (WX_VF_UPDATE_XCAST_MODE | WX_VT_MSGTYPE_NACK))
+ return -EINVAL;
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_update_xcast_mode_vf);
+
+/**
+ * wx_get_link_state_vf - Get VF link state from PF
+ * @wx: pointer to the HW structure
+ * @link_state: link state storage
+ *
+ * Return: return state of the operation error or success.
+ **/
+int wx_get_link_state_vf(struct wx *wx, u16 *link_state)
+{
+ u32 msgbuf[2] = {WX_VF_GET_LINK_STATE};
+ int ret;
+
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+
+ if (msgbuf[0] & WX_VT_MSGTYPE_NACK)
+ return -EINVAL;
+
+ *link_state = msgbuf[1];
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_get_link_state_vf);
+
+/**
+ * wx_set_vfta_vf - Set/Unset vlan filter table address
+ * @wx: pointer to the HW structure
+ * @vlan: 12 bit VLAN ID
+ * @vind: unused by VF drivers
+ * @vlan_on: if true then set bit, else clear bit
+ * @vlvf_bypass: boolean flag indicating updating default pool is okay
+ *
+ * Turn on/off specified VLAN in the VLAN filter table.
+ *
+ * Return: returns 0 on success, negative error code on failure
+ **/
+int wx_set_vfta_vf(struct wx *wx, u32 vlan, u32 vind, bool vlan_on,
+ bool vlvf_bypass)
+{
+ u32 msgbuf[2] = {WX_VF_SET_VLAN, vlan};
+ bool vlan_offload = false;
+ int ret;
+
+ /* Setting the 8 bit field MSG INFO to TRUE indicates "add" */
+ msgbuf[0] |= vlan_on << WX_VT_MSGINFO_SHIFT;
+ /* if vf vlan offload is disabled, allow to create vlan under pf port vlan */
+ msgbuf[0] |= BIT(vlan_offload);
+
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+
+ if (msgbuf[0] & WX_VT_MSGTYPE_ACK)
+ return 0;
+
+ return msgbuf[0] & WX_VT_MSGTYPE_NACK;
+}
+EXPORT_SYMBOL(wx_set_vfta_vf);
+
+void wx_get_mac_addr_vf(struct wx *wx, u8 *mac_addr)
+{
+ ether_addr_copy(mac_addr, wx->mac.perm_addr);
+}
+EXPORT_SYMBOL(wx_get_mac_addr_vf);
+
+int wx_get_fw_version_vf(struct wx *wx)
+{
+ u32 msgbuf[2] = {WX_VF_GET_FW_VERSION};
+ int ret;
+
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+
+ if (msgbuf[0] & WX_VT_MSGTYPE_NACK)
+ return -EINVAL;
+ snprintf(wx->eeprom_id, 32, "0x%08x", msgbuf[1]);
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_get_fw_version_vf);
+
+int wx_set_uc_addr_vf(struct wx *wx, u32 index, u8 *addr)
+{
+ u32 msgbuf[3] = {WX_VF_SET_MACVLAN};
+ u8 *msg_addr = (u8 *)(&msgbuf[1]);
+ int ret;
+
+ /* If index is one then this is the start of a new list and needs
+ * indication to the PF so it can do it's own list management.
+ * If it is zero then that tells the PF to just clear all of
+ * this VF's macvlans and there is no new list.
+ */
+ msgbuf[0] |= index << WX_VT_MSGINFO_SHIFT;
+ if (addr)
+ memcpy(msg_addr, addr, 6);
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+
+ msgbuf[0] &= ~WX_VT_MSGTYPE_CTS;
+
+ if (msgbuf[0] == (WX_VF_SET_MACVLAN | WX_VT_MSGTYPE_NACK))
+ return -EINVAL;
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_set_uc_addr_vf);
+
+/**
+ * wx_rlpml_set_vf - Set the maximum receive packet length
+ * @wx: pointer to the HW structure
+ * @max_size: value to assign to max frame size
+ *
+ * Return: returns 0 on success, negative error code on failure
+ **/
+int wx_rlpml_set_vf(struct wx *wx, u16 max_size)
+{
+ u32 msgbuf[2] = {WX_VF_SET_LPE, max_size};
+ int ret;
+
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+ if ((msgbuf[0] & WX_VF_SET_LPE) &&
+ (msgbuf[0] & WX_VT_MSGTYPE_NACK))
+ return -EINVAL;
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_rlpml_set_vf);
+
+/**
+ * wx_negotiate_api_version - Negotiate supported API version
+ * @wx: pointer to the HW structure
+ * @api: integer containing requested API version
+ *
+ * Return: returns 0 on success, negative error code on failure
+ **/
+int wx_negotiate_api_version(struct wx *wx, int api)
+{
+ u32 msgbuf[2] = {WX_VF_API_NEGOTIATE, api};
+ int ret;
+
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+
+ msgbuf[0] &= ~WX_VT_MSGTYPE_CTS;
+
+ /* Store value and return 0 on success */
+ if (msgbuf[0] == (WX_VF_API_NEGOTIATE | WX_VT_MSGTYPE_NACK))
+ return -EINVAL;
+ wx->vfinfo->vf_api = api;
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_negotiate_api_version);
+
+int wx_get_queues_vf(struct wx *wx, u32 *num_tcs, u32 *default_tc)
+{
+ u32 msgbuf[5] = {WX_VF_GET_QUEUES};
+ int ret;
+
+ /* do nothing if API doesn't support wx_get_queues */
+ if (wx->vfinfo->vf_api < wx_mbox_api_13)
+ return -EINVAL;
+
+ /* Fetch queue configuration from the PF */
+ ret = wx_mbx_write_and_read_reply(wx, msgbuf, msgbuf,
+ ARRAY_SIZE(msgbuf));
+ if (ret)
+ return ret;
+ msgbuf[0] &= ~WX_VT_MSGTYPE_CTS;
+
+ /* if we didn't get an ACK there must have been
+ * some sort of mailbox error so we should treat it
+ * as such
+ */
+ if (msgbuf[0] != (WX_VF_GET_QUEUES | WX_VT_MSGTYPE_ACK))
+ return -EINVAL;
+ /* record and validate values from message */
+ wx->mac.max_tx_queues = msgbuf[WX_VF_TX_QUEUES];
+ if (wx->mac.max_tx_queues == 0 ||
+ wx->mac.max_tx_queues > WX_VF_MAX_TX_QUEUES)
+ wx->mac.max_tx_queues = WX_VF_MAX_TX_QUEUES;
+
+ wx->mac.max_rx_queues = msgbuf[WX_VF_RX_QUEUES];
+ if (wx->mac.max_rx_queues == 0 ||
+ wx->mac.max_rx_queues > WX_VF_MAX_RX_QUEUES)
+ wx->mac.max_rx_queues = WX_VF_MAX_RX_QUEUES;
+
+ *num_tcs = msgbuf[WX_VF_TRANS_VLAN];
+ /* in case of unknown state assume we cannot tag frames */
+ if (*num_tcs > wx->mac.max_rx_queues)
+ *num_tcs = 1;
+ *default_tc = msgbuf[WX_VF_DEF_QUEUE];
+ /* default to queue 0 on out-of-bounds queue number */
+ if (*default_tc >= wx->mac.max_tx_queues)
+ *default_tc = 0;
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_get_queues_vf);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf.h b/drivers/net/ethernet/wangxun/libwx/wx_vf.h
new file mode 100644
index 000000000000..c523ef3e8502
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf.h
@@ -0,0 +1,63 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#ifndef _WX_VF_H_
+#define _WX_VF_H_
+
+#define WX_VF_MAX_RING_NUMS 8
+#define WX_VX_PF_BME 0x4B8
+#define WX_VF_BME_ENABLE BIT(0)
+#define WX_VXSTATUS 0x4
+#define WX_VXCTRL 0x8
+#define WX_VXCTRL_RST BIT(0)
+
+#define WX_VXMRQC 0x78
+#define WX_VXICR 0x100
+#define WX_VXIMS 0x108
+#define WX_VF_IRQ_CLEAR_MASK 7
+#define WX_VF_MAX_TX_QUEUES 4
+#define WX_VF_MAX_RX_QUEUES 4
+#define WX_VXTXDCTL(r) (0x3010 + (0x40 * (r)))
+#define WX_VXRXDCTL(r) (0x1010 + (0x40 * (r)))
+#define WX_VXRXDCTL_ENABLE BIT(0)
+#define WX_VXTXDCTL_FLUSH BIT(26)
+
+#define WX_VXRXDCTL_RSCMAX(f) FIELD_PREP(GENMASK(24, 23), f)
+#define WX_VXRXDCTL_BUFLEN(f) FIELD_PREP(GENMASK(6, 1), f)
+#define WX_VXRXDCTL_BUFSZ(f) FIELD_PREP(GENMASK(11, 8), f)
+#define WX_VXRXDCTL_HDRSZ(f) FIELD_PREP(GENMASK(15, 12), f)
+
+#define WX_VXRXDCTL_RSCMAX_MASK GENMASK(24, 23)
+#define WX_VXRXDCTL_BUFLEN_MASK GENMASK(6, 1)
+#define WX_VXRXDCTL_BUFSZ_MASK GENMASK(11, 8)
+#define WX_VXRXDCTL_HDRSZ_MASK GENMASK(15, 12)
+
+#define wx_conf_size(v, mwidth, uwidth) ({ \
+ typeof(v) _v = (v); \
+ (_v == 2 << (mwidth) ? 0 : _v >> (uwidth)); \
+})
+#define wx_buf_len(v) wx_conf_size(v, 13, 7)
+#define wx_hdr_sz(v) wx_conf_size(v, 10, 6)
+#define wx_buf_sz(v) wx_conf_size(v, 14, 10)
+#define wx_pkt_thresh(v) wx_conf_size(v, 4, 0)
+
+#define WX_RX_HDR_SIZE 256
+#define WX_RX_BUF_SIZE 2048
+
+void wx_init_hw_vf(struct wx *wx);
+int wx_reset_hw_vf(struct wx *wx);
+void wx_get_mac_addr_vf(struct wx *wx, u8 *mac_addr);
+void wx_stop_adapter_vf(struct wx *wx);
+int wx_get_fw_version_vf(struct wx *wx);
+int wx_set_rar_vf(struct wx *wx, u32 index, u8 *addr, u32 enable_addr);
+int wx_update_mc_addr_list_vf(struct wx *wx, struct net_device *netdev);
+int wx_set_uc_addr_vf(struct wx *wx, u32 index, u8 *addr);
+int wx_rlpml_set_vf(struct wx *wx, u16 max_size);
+int wx_negotiate_api_version(struct wx *wx, int api);
+int wx_get_queues_vf(struct wx *wx, u32 *num_tcs, u32 *default_tc);
+int wx_update_xcast_mode_vf(struct wx *wx, int xcast_mode);
+int wx_get_link_state_vf(struct wx *wx, u16 *link_state);
+int wx_set_vfta_vf(struct wx *wx, u32 vlan, u32 vind, bool vlan_on,
+ bool vlvf_bypass);
+
+#endif /* _WX_VF_H_ */
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 03/12] net: libwx: add wangxun vf common api
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 01/12] net: libwx: add mailbox api for wangxun vf drivers Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 02/12] net: libwx: add base vf api for " Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-09 13:51 ` Vadim Fedorenko
2025-07-04 9:49 ` [PATCH net-next v3 04/12] net: wangxun: add txgbevf build Mengyuan Lou
` (9 subsequent siblings)
12 siblings, 1 reply; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Add common wx_configure_vf and wx_set_mac_vf for
ngbevf and txgbevf.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
drivers/net/ethernet/wangxun/libwx/Makefile | 2 +-
drivers/net/ethernet/wangxun/libwx/wx_hw.c | 3 +-
drivers/net/ethernet/wangxun/libwx/wx_hw.h | 1 +
drivers/net/ethernet/wangxun/libwx/wx_type.h | 4 +
drivers/net/ethernet/wangxun/libwx/wx_vf.h | 50 ++++
.../net/ethernet/wangxun/libwx/wx_vf_common.c | 196 ++++++++++++
.../net/ethernet/wangxun/libwx/wx_vf_common.h | 14 +
.../net/ethernet/wangxun/libwx/wx_vf_lib.c | 280 ++++++++++++++++++
.../net/ethernet/wangxun/libwx/wx_vf_lib.h | 14 +
9 files changed, 562 insertions(+), 2 deletions(-)
create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c
create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf_lib.h
diff --git a/drivers/net/ethernet/wangxun/libwx/Makefile b/drivers/net/ethernet/wangxun/libwx/Makefile
index ddf0bb921676..a71b0ad77de3 100644
--- a/drivers/net/ethernet/wangxun/libwx/Makefile
+++ b/drivers/net/ethernet/wangxun/libwx/Makefile
@@ -5,4 +5,4 @@
obj-$(CONFIG_LIBWX) += libwx.o
libwx-objs := wx_hw.o wx_lib.o wx_ethtool.o wx_ptp.o wx_mbx.o wx_sriov.o
-libwx-objs += wx_vf.o
+libwx-objs += wx_vf.o wx_vf_lib.o wx_vf_common.o
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
index 82dd76f0326e..27bb33788701 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
@@ -1827,7 +1827,7 @@ void wx_disable_rx_queue(struct wx *wx, struct wx_ring *ring)
}
EXPORT_SYMBOL(wx_disable_rx_queue);
-static void wx_enable_rx_queue(struct wx *wx, struct wx_ring *ring)
+void wx_enable_rx_queue(struct wx *wx, struct wx_ring *ring)
{
u8 reg_idx = ring->reg_idx;
u32 rxdctl;
@@ -1843,6 +1843,7 @@ static void wx_enable_rx_queue(struct wx *wx, struct wx_ring *ring)
reg_idx);
}
}
+EXPORT_SYMBOL(wx_enable_rx_queue);
static void wx_configure_srrctl(struct wx *wx,
struct wx_ring *rx_ring)
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.h b/drivers/net/ethernet/wangxun/libwx/wx_hw.h
index 718015611da6..2393a743b564 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_hw.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.h
@@ -38,6 +38,7 @@ void wx_enable_sec_rx_path(struct wx *wx);
void wx_set_rx_mode(struct net_device *netdev);
int wx_change_mtu(struct net_device *netdev, int new_mtu);
void wx_disable_rx_queue(struct wx *wx, struct wx_ring *ring);
+void wx_enable_rx_queue(struct wx *wx, struct wx_ring *ring);
void wx_configure_rx(struct wx *wx);
void wx_configure(struct wx *wx);
void wx_start_hw(struct wx *wx);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h
index d14e46ac244a..9e5b0d1fcb21 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_type.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h
@@ -828,6 +828,8 @@ struct wx_mbx_info {
u32 mailbox;
u32 udelay;
u32 timeout;
+ /* lock mbx access */
+ spinlock_t mbx_lock;
};
struct wx_thermal_sensor_data {
@@ -1289,6 +1291,8 @@ struct wx {
u32 *isb_mem;
u32 isb_tag[WX_ISB_MAX];
bool misc_irq_domain;
+ u32 eims_other;
+ u32 eims_enable_mask;
#define WX_MAX_RETA_ENTRIES 128
#define WX_RSS_INDIR_TBL_MAX 64
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf.h b/drivers/net/ethernet/wangxun/libwx/wx_vf.h
index c523ef3e8502..e863a74c291d 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf.h
@@ -14,6 +14,7 @@
#define WX_VXMRQC 0x78
#define WX_VXICR 0x100
#define WX_VXIMS 0x108
+#define WX_VXIMC 0x10C
#define WX_VF_IRQ_CLEAR_MASK 7
#define WX_VF_MAX_TX_QUEUES 4
#define WX_VF_MAX_RX_QUEUES 4
@@ -22,6 +23,12 @@
#define WX_VXRXDCTL_ENABLE BIT(0)
#define WX_VXTXDCTL_FLUSH BIT(26)
+#define WX_VXITR(i) (0x200 + (4 * (i))) /* i=[0,1] */
+#define WX_VXITR_MASK GENMASK(8, 0)
+#define WX_VXITR_CNT_WDIS BIT(31)
+#define WX_VXIVAR_MISC 0x260
+#define WX_VXIVAR(i) (0x240 + (4 * (i))) /* i=[0,3] */
+
#define WX_VXRXDCTL_RSCMAX(f) FIELD_PREP(GENMASK(24, 23), f)
#define WX_VXRXDCTL_BUFLEN(f) FIELD_PREP(GENMASK(6, 1), f)
#define WX_VXRXDCTL_BUFSZ(f) FIELD_PREP(GENMASK(11, 8), f)
@@ -44,6 +51,49 @@
#define WX_RX_HDR_SIZE 256
#define WX_RX_BUF_SIZE 2048
+#define WX_RXBUFFER_2048 (2048)
+#define WX_RXBUFFER_3072 3072
+
+/* Receive Path */
+#define WX_VXRDBAL(r) (0x1000 + (0x40 * (r)))
+#define WX_VXRDBAH(r) (0x1004 + (0x40 * (r)))
+#define WX_VXRDT(r) (0x1008 + (0x40 * (r)))
+#define WX_VXRDH(r) (0x100C + (0x40 * (r)))
+
+#define WX_VXRXDCTL_RSCEN BIT(29)
+#define WX_VXRXDCTL_DROP BIT(30)
+#define WX_VXRXDCTL_VLAN BIT(31)
+
+#define WX_VXTDBAL(r) (0x3000 + (0x40 * (r)))
+#define WX_VXTDBAH(r) (0x3004 + (0x40 * (r)))
+#define WX_VXTDT(r) (0x3008 + (0x40 * (r)))
+#define WX_VXTDH(r) (0x300C + (0x40 * (r)))
+
+#define WX_VXTXDCTL_ENABLE BIT(0)
+#define WX_VXTXDCTL_BUFLEN(f) FIELD_PREP(GENMASK(6, 1), f)
+#define WX_VXTXDCTL_PTHRESH(f) FIELD_PREP(GENMASK(11, 8), f)
+#define WX_VXTXDCTL_WTHRESH(f) FIELD_PREP(GENMASK(22, 16), f)
+
+#define WX_VXMRQC_PSR(f) FIELD_PREP(GENMASK(5, 1), f)
+#define WX_VXMRQC_PSR_MASK GENMASK(5, 1)
+#define WX_VXMRQC_PSR_L4HDR BIT(0)
+#define WX_VXMRQC_PSR_L3HDR BIT(1)
+#define WX_VXMRQC_PSR_L2HDR BIT(2)
+#define WX_VXMRQC_PSR_TUNHDR BIT(3)
+#define WX_VXMRQC_PSR_TUNMAC BIT(4)
+
+#define WX_VXRSSRK(i) (0x80 + ((i) * 4)) /* i=[0,9] */
+#define WX_VXRETA(i) (0xC0 + ((i) * 4)) /* i=[0,15] */
+
+#define WX_VXMRQC_RSS(f) FIELD_PREP(GENMASK(31, 16), f)
+#define WX_VXMRQC_RSS_MASK GENMASK(31, 16)
+#define WX_VXMRQC_RSS_ALG_IPV4_TCP BIT(0)
+#define WX_VXMRQC_RSS_ALG_IPV4 BIT(1)
+#define WX_VXMRQC_RSS_ALG_IPV6 BIT(4)
+#define WX_VXMRQC_RSS_ALG_IPV6_TCP BIT(5)
+#define WX_VXMRQC_RSS_EN BIT(8)
+#define WX_VXMRQC_RSS_HASH(f) FIELD_PREP(GENMASK(15, 13), f)
+
void wx_init_hw_vf(struct wx *wx);
int wx_reset_hw_vf(struct wx *wx);
void wx_get_mac_addr_vf(struct wx *wx, u8 *mac_addr);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
new file mode 100644
index 000000000000..aac420bf578b
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
@@ -0,0 +1,196 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#include <linux/etherdevice.h>
+#include <linux/pci.h>
+
+#include "wx_type.h"
+#include "wx_mbx.h"
+#include "wx_lib.h"
+#include "wx_vf.h"
+#include "wx_vf_lib.h"
+#include "wx_vf_common.h"
+
+static irqreturn_t wx_msix_misc_vf(int __always_unused irq, void *data)
+{
+ struct wx *wx = data;
+
+ /* Clear the interrupt */
+ if (netif_running(wx->netdev))
+ wr32(wx, WX_VXIMC, wx->eims_other);
+
+ return IRQ_HANDLED;
+}
+
+int wx_request_msix_irqs_vf(struct wx *wx)
+{
+ struct net_device *netdev = wx->netdev;
+ int vector, err;
+
+ for (vector = 0; vector < wx->num_q_vectors; vector++) {
+ struct wx_q_vector *q_vector = wx->q_vector[vector];
+ struct msix_entry *entry = &wx->msix_q_entries[vector];
+
+ if (q_vector->tx.ring && q_vector->rx.ring)
+ snprintf(q_vector->name, sizeof(q_vector->name) - 1,
+ "%s-TxRx-%d", netdev->name, entry->entry);
+ else
+ /* skip this unused q_vector */
+ continue;
+
+ err = request_irq(entry->vector, wx_msix_clean_rings, 0,
+ q_vector->name, q_vector);
+ if (err) {
+ wx_err(wx, "request_irq failed for MSIX interrupt %s Error: %d\n",
+ q_vector->name, err);
+ goto free_queue_irqs;
+ }
+ }
+
+ err = request_threaded_irq(wx->msix_entry->vector, NULL,
+ wx_msix_misc_vf, IRQF_ONESHOT,
+ netdev->name, wx);
+ if (err) {
+ wx_err(wx, "request_irq for msix_other failed: %d\n", err);
+ goto free_queue_irqs;
+ }
+
+ return 0;
+
+free_queue_irqs:
+ while (vector) {
+ vector--;
+ free_irq(wx->msix_q_entries[vector].vector,
+ wx->q_vector[vector]);
+ }
+ wx_reset_interrupt_capability(wx);
+ return err;
+}
+EXPORT_SYMBOL(wx_request_msix_irqs_vf);
+
+void wx_negotiate_api_vf(struct wx *wx)
+{
+ int api[] = {
+ wx_mbox_api_13,
+ wx_mbox_api_null};
+ int err = 0, idx = 0;
+
+ spin_lock_bh(&wx->mbx.mbx_lock);
+ while (api[idx] != wx_mbox_api_null) {
+ err = wx_negotiate_api_version(wx, api[idx]);
+ if (!err)
+ break;
+ idx++;
+ }
+ spin_unlock_bh(&wx->mbx.mbx_lock);
+}
+EXPORT_SYMBOL(wx_negotiate_api_vf);
+
+void wx_reset_vf(struct wx *wx)
+{
+ struct net_device *netdev = wx->netdev;
+ int ret = 0;
+
+ ret = wx_reset_hw_vf(wx);
+ if (!ret)
+ wx_init_hw_vf(wx);
+ wx_negotiate_api_vf(wx);
+ if (is_valid_ether_addr(wx->mac.addr)) {
+ eth_hw_addr_set(netdev, wx->mac.addr);
+ ether_addr_copy(netdev->perm_addr, wx->mac.addr);
+ }
+}
+EXPORT_SYMBOL(wx_reset_vf);
+
+void wx_set_rx_mode_vf(struct net_device *netdev)
+{
+ struct wx *wx = netdev_priv(netdev);
+ unsigned int flags = netdev->flags;
+ int xcast_mode;
+
+ xcast_mode = (flags & IFF_ALLMULTI) ? WXVF_XCAST_MODE_ALLMULTI :
+ (flags & (IFF_BROADCAST | IFF_MULTICAST)) ?
+ WXVF_XCAST_MODE_MULTI : WXVF_XCAST_MODE_NONE;
+ /* request the most inclusive mode we need */
+ if (flags & IFF_PROMISC)
+ xcast_mode = WXVF_XCAST_MODE_PROMISC;
+ else if (flags & IFF_ALLMULTI)
+ xcast_mode = WXVF_XCAST_MODE_ALLMULTI;
+ else if (flags & (IFF_BROADCAST | IFF_MULTICAST))
+ xcast_mode = WXVF_XCAST_MODE_MULTI;
+ else
+ xcast_mode = WXVF_XCAST_MODE_NONE;
+
+ spin_lock_bh(&wx->mbx.mbx_lock);
+ wx_update_xcast_mode_vf(wx, xcast_mode);
+ wx_update_mc_addr_list_vf(wx, netdev);
+ wx_write_uc_addr_list_vf(netdev);
+ spin_unlock_bh(&wx->mbx.mbx_lock);
+}
+EXPORT_SYMBOL(wx_set_rx_mode_vf);
+
+/**
+ * wx_configure_rx_vf - Configure Receive Unit after Reset
+ * @wx: board private structure
+ *
+ * Configure the Rx unit of the MAC after a reset.
+ **/
+static void wx_configure_rx_vf(struct wx *wx)
+{
+ struct net_device *netdev = wx->netdev;
+ int i, ret;
+
+ wx_setup_psrtype_vf(wx);
+ wx_setup_vfmrqc_vf(wx);
+
+ spin_lock_bh(&wx->mbx.mbx_lock);
+ ret = wx_rlpml_set_vf(wx,
+ netdev->mtu + ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN);
+ spin_unlock_bh(&wx->mbx.mbx_lock);
+ if (ret)
+ wx_dbg(wx, "Failed to set MTU at %d\n", netdev->mtu);
+
+ /* Setup the HW Rx Head and Tail Descriptor Pointers and
+ * the Base and Length of the Rx Descriptor Ring
+ */
+ for (i = 0; i < wx->num_rx_queues; i++) {
+ struct wx_ring *rx_ring = wx->rx_ring[i];
+#ifdef HAVE_SWIOTLB_SKIP_CPU_SYNC
+ wx_set_rx_buffer_len_vf(wx, rx_ring);
+#endif
+ wx_configure_rx_ring_vf(wx, rx_ring);
+ }
+}
+
+void wx_configure_vf(struct wx *wx)
+{
+ wx_set_rx_mode_vf(wx->netdev);
+ wx_configure_tx_vf(wx);
+ wx_configure_rx_vf(wx);
+}
+EXPORT_SYMBOL(wx_configure_vf);
+
+int wx_set_mac_vf(struct net_device *netdev, void *p)
+{
+ struct wx *wx = netdev_priv(netdev);
+ struct sockaddr *addr = p;
+ int ret;
+
+ ret = eth_prepare_mac_addr_change(netdev, addr);
+ if (ret)
+ return ret;
+
+ spin_lock_bh(&wx->mbx.mbx_lock);
+ ret = wx_set_rar_vf(wx, 1, (u8 *)addr->sa_data, 1);
+ spin_unlock_bh(&wx->mbx.mbx_lock);
+
+ if (ret)
+ return -EPERM;
+
+ memcpy(wx->mac.addr, addr->sa_data, netdev->addr_len);
+ memcpy(wx->mac.perm_addr, addr->sa_data, netdev->addr_len);
+ eth_hw_addr_set(netdev, addr->sa_data);
+
+ return 0;
+}
+EXPORT_SYMBOL(wx_set_mac_vf);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
new file mode 100644
index 000000000000..9bee9de86cb2
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#ifndef _WX_VF_COMMON_H_
+#define _WX_VF_COMMON_H_
+
+int wx_request_msix_irqs_vf(struct wx *wx);
+void wx_negotiate_api_vf(struct wx *wx);
+void wx_reset_vf(struct wx *wx);
+void wx_set_rx_mode_vf(struct net_device *netdev);
+void wx_configure_vf(struct wx *wx);
+int wx_set_mac_vf(struct net_device *netdev, void *p);
+
+#endif /* _WX_VF_COMMON_H_ */
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c
new file mode 100644
index 000000000000..5d48df7a849f
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c
@@ -0,0 +1,280 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#include <linux/etherdevice.h>
+#include <linux/pci.h>
+
+#include "wx_type.h"
+#include "wx_hw.h"
+#include "wx_lib.h"
+#include "wx_vf.h"
+#include "wx_vf_lib.h"
+
+static void wx_write_eitr_vf(struct wx_q_vector *q_vector)
+{
+ struct wx *wx = q_vector->wx;
+ int v_idx = q_vector->v_idx;
+ u32 itr_reg;
+
+ itr_reg = q_vector->itr & WX_VXITR_MASK;
+
+ /* set the WDIS bit to not clear the timer bits and cause an
+ * immediate assertion of the interrupt
+ */
+ itr_reg |= WX_VXITR_CNT_WDIS;
+
+ wr32(wx, WX_VXITR(v_idx), itr_reg);
+}
+
+static void wx_set_ivar_vf(struct wx *wx, s8 direction, u8 queue,
+ u8 msix_vector)
+{
+ u32 ivar, index;
+
+ if (direction == -1) {
+ /* other causes */
+ msix_vector |= WX_PX_IVAR_ALLOC_VAL;
+ ivar = rd32(wx, WX_VXIVAR_MISC);
+ ivar &= ~0xFF;
+ ivar |= msix_vector;
+ wr32(wx, WX_VXIVAR_MISC, ivar);
+ } else {
+ /* tx or rx causes */
+ msix_vector |= WX_PX_IVAR_ALLOC_VAL;
+ index = ((16 * (queue & 1)) + (8 * direction));
+ ivar = rd32(wx, WX_VXIVAR(queue >> 1));
+ ivar &= ~(0xFF << index);
+ ivar |= (msix_vector << index);
+ wr32(wx, WX_VXIVAR(queue >> 1), ivar);
+ }
+}
+
+void wx_configure_msix_vf(struct wx *wx)
+{
+ int v_idx;
+
+ wx->eims_enable_mask = 0;
+ for (v_idx = 0; v_idx < wx->num_q_vectors; v_idx++) {
+ struct wx_q_vector *q_vector = wx->q_vector[v_idx];
+ struct wx_ring *ring;
+
+ wx_for_each_ring(ring, q_vector->rx)
+ wx_set_ivar_vf(wx, 0, ring->reg_idx, v_idx);
+
+ wx_for_each_ring(ring, q_vector->tx)
+ wx_set_ivar_vf(wx, 1, ring->reg_idx, v_idx);
+
+ /* add q_vector eims value to global eims_enable_mask */
+ wx->eims_enable_mask |= BIT(v_idx);
+ wx_write_eitr_vf(q_vector);
+ }
+
+ wx_set_ivar_vf(wx, -1, 1, v_idx);
+
+ /* setup eims_other and add value to global eims_enable_mask */
+ wx->eims_other = BIT(v_idx);
+ wx->eims_enable_mask |= wx->eims_other;
+}
+
+int wx_write_uc_addr_list_vf(struct net_device *netdev)
+{
+ struct wx *wx = netdev_priv(netdev);
+ int count = 0;
+
+ if (!netdev_uc_empty(netdev)) {
+ struct netdev_hw_addr *ha;
+
+ netdev_for_each_uc_addr(ha, netdev)
+ wx_set_uc_addr_vf(wx, ++count, ha->addr);
+ } else {
+ /*
+ * If the list is empty then send message to PF driver to
+ * clear all macvlans on this VF.
+ */
+ wx_set_uc_addr_vf(wx, 0, NULL);
+ }
+
+ return count;
+}
+
+/**
+ * wx_configure_tx_ring_vf - Configure Tx ring after Reset
+ * @wx: board private structure
+ * @ring: structure containing ring specific data
+ *
+ * Configure the Tx descriptor ring after a reset.
+ **/
+static void wx_configure_tx_ring_vf(struct wx *wx, struct wx_ring *ring)
+{
+ u8 reg_idx = ring->reg_idx;
+ u64 tdba = ring->dma;
+ u32 txdctl = 0;
+ int ret;
+
+ /* disable queue to avoid issues while updating state */
+ wr32(wx, WX_VXTXDCTL(reg_idx), WX_VXTXDCTL_FLUSH);
+ wr32(wx, WX_VXTDBAL(reg_idx), tdba & DMA_BIT_MASK(32));
+ wr32(wx, WX_VXTDBAH(reg_idx), tdba >> 32);
+
+ /* enable relaxed ordering */
+ pcie_capability_clear_and_set_word(wx->pdev, PCI_EXP_DEVCTL,
+ 0, PCI_EXP_DEVCTL_RELAX_EN);
+
+ /* reset head and tail pointers */
+ wr32(wx, WX_VXTDH(reg_idx), 0);
+ wr32(wx, WX_VXTDT(reg_idx), 0);
+ ring->tail = wx->hw_addr + WX_VXTDT(reg_idx);
+
+ /* reset ntu and ntc to place SW in sync with hardwdare */
+ ring->next_to_clean = 0;
+ ring->next_to_use = 0;
+
+ txdctl |= WX_VXTXDCTL_BUFLEN(wx_buf_len(ring->count));
+ txdctl |= WX_VXTXDCTL_ENABLE;
+
+ /* reinitialize tx_buffer_info */
+ memset(ring->tx_buffer_info, 0,
+ sizeof(struct wx_tx_buffer) * ring->count);
+
+ wr32(wx, WX_VXTXDCTL(reg_idx), txdctl);
+ /* poll to verify queue is enabled */
+ ret = read_poll_timeout(rd32, txdctl, txdctl & WX_VXTXDCTL_ENABLE,
+ 1000, 10000, true, wx, WX_VXTXDCTL(reg_idx));
+ if (ret == -ETIMEDOUT)
+ wx_err(wx, "Could not enable Tx Queue %d\n", reg_idx);
+}
+
+/**
+ * wx_configure_tx_vf - Configure Transmit Unit after Reset
+ * @wx: board private structure
+ *
+ * Configure the Tx unit of the MAC after a reset.
+ **/
+void wx_configure_tx_vf(struct wx *wx)
+{
+ u32 i;
+
+ /* Setup the HW Tx Head and Tail descriptor pointers */
+ for (i = 0; i < wx->num_tx_queues; i++)
+ wx_configure_tx_ring_vf(wx, wx->tx_ring[i]);
+}
+
+static void wx_configure_srrctl_vf(struct wx *wx, struct wx_ring *ring,
+ int index)
+{
+ u32 srrctl;
+
+ srrctl = rd32m(wx, WX_VXRXDCTL(index),
+ (u32)~(WX_VXRXDCTL_HDRSZ_MASK | WX_VXRXDCTL_BUFSZ_MASK));
+ srrctl |= WX_VXRXDCTL_DROP;
+ srrctl |= WX_VXRXDCTL_HDRSZ(wx_hdr_sz(WX_RX_HDR_SIZE));
+ srrctl |= WX_VXRXDCTL_BUFSZ(wx_buf_sz(WX_RX_BUF_SIZE));
+
+ wr32(wx, WX_VXRXDCTL(index), srrctl);
+}
+
+void wx_setup_psrtype_vf(struct wx *wx)
+{
+ /* PSRTYPE must be initialized */
+ u32 psrtype = WX_VXMRQC_PSR_L2HDR |
+ WX_VXMRQC_PSR_L3HDR |
+ WX_VXMRQC_PSR_L4HDR |
+ WX_VXMRQC_PSR_TUNHDR |
+ WX_VXMRQC_PSR_TUNMAC;
+
+ wr32m(wx, WX_VXMRQC, WX_VXMRQC_PSR_MASK, WX_VXMRQC_PSR(psrtype));
+}
+
+void wx_setup_vfmrqc_vf(struct wx *wx)
+{
+ u16 rss_i = wx->num_rx_queues;
+ u32 vfmrqc = 0, vfreta = 0;
+ u8 i, j;
+
+ /* Fill out hash function seeds */
+ netdev_rss_key_fill(wx->rss_key, sizeof(wx->rss_key));
+ for (i = 0; i < WX_RSS_KEY_SIZE / 4; i++)
+ wr32(wx, WX_VXRSSRK(i), wx->rss_key[i]);
+
+ for (i = 0, j = 0; i < WX_MAX_RETA_ENTRIES; i++, j++) {
+ if (j == rss_i)
+ j = 0;
+
+ wx->rss_indir_tbl[i] = j;
+
+ vfreta |= j << (i & 0x3) * 8;
+ if ((i & 3) == 3) {
+ wr32(wx, WX_VXRETA(i >> 2), vfreta);
+ vfreta = 0;
+ }
+ }
+
+ /* Perform hash on these packet types */
+ vfmrqc |= WX_VXMRQC_RSS_ALG_IPV4 |
+ WX_VXMRQC_RSS_ALG_IPV4_TCP |
+ WX_VXMRQC_RSS_ALG_IPV6 |
+ WX_VXMRQC_RSS_ALG_IPV6_TCP;
+
+ vfmrqc |= WX_VXMRQC_RSS_EN;
+
+ if (wx->num_rx_queues > 3)
+ vfmrqc |= WX_VXMRQC_RSS_HASH(2);
+ else if (wx->num_rx_queues > 1)
+ vfmrqc |= WX_VXMRQC_RSS_HASH(1);
+ wr32m(wx, WX_VXMRQC, WX_VXMRQC_RSS_MASK, WX_VXMRQC_RSS(vfmrqc));
+}
+
+void wx_configure_rx_ring_vf(struct wx *wx, struct wx_ring *ring)
+{
+ u8 reg_idx = ring->reg_idx;
+ union wx_rx_desc *rx_desc;
+ u64 rdba = ring->dma;
+ u32 rxdctl;
+
+ /* disable queue to avoid issues while updating state */
+ rxdctl = rd32(wx, WX_VXRXDCTL(reg_idx));
+ wx_disable_rx_queue(wx, ring);
+
+ wr32(wx, WX_VXRDBAL(reg_idx), rdba & DMA_BIT_MASK(32));
+ wr32(wx, WX_VXRDBAH(reg_idx), rdba >> 32);
+
+ /* enable relaxed ordering */
+ pcie_capability_clear_and_set_word(wx->pdev, PCI_EXP_DEVCTL,
+ 0, PCI_EXP_DEVCTL_RELAX_EN);
+
+ /* reset head and tail pointers */
+ wr32(wx, WX_VXRDH(reg_idx), 0);
+ wr32(wx, WX_VXRDT(reg_idx), 0);
+ ring->tail = wx->hw_addr + WX_VXRDT(reg_idx);
+
+ /* initialize rx_buffer_info */
+ memset(ring->rx_buffer_info, 0,
+ sizeof(struct wx_rx_buffer) * ring->count);
+
+ /* initialize Rx descriptor 0 */
+ rx_desc = WX_RX_DESC(ring, 0);
+ rx_desc->wb.upper.length = 0;
+
+ /* reset ntu and ntc to place SW in sync with hardwdare */
+ ring->next_to_clean = 0;
+ ring->next_to_use = 0;
+ ring->next_to_alloc = 0;
+
+ wx_configure_srrctl_vf(wx, ring, reg_idx);
+
+ /* allow any size packet since we can handle overflow */
+ rxdctl &= ~WX_VXRXDCTL_BUFLEN_MASK;
+ rxdctl |= WX_VXRXDCTL_BUFLEN(wx_buf_len(ring->count));
+ rxdctl |= WX_VXRXDCTL_ENABLE | WX_VXRXDCTL_VLAN;
+
+ /* enable RSC */
+ rxdctl &= ~WX_VXRXDCTL_RSCMAX_MASK;
+ rxdctl |= WX_VXRXDCTL_RSCMAX(0);
+ rxdctl |= WX_VXRXDCTL_RSCEN;
+
+ wr32(wx, WX_VXRXDCTL(reg_idx), rxdctl);
+
+ /* pf/vf reuse */
+ wx_enable_rx_queue(wx, ring);
+ wx_alloc_rx_buffers(ring, wx_desc_unused(ring));
+}
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.h b/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.h
new file mode 100644
index 000000000000..43ea126b79eb
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_lib.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#ifndef _WX_VF_LIB_H_
+#define _WX_VF_LIB_H_
+
+void wx_configure_msix_vf(struct wx *wx);
+int wx_write_uc_addr_list_vf(struct net_device *netdev);
+void wx_setup_psrtype_vf(struct wx *wx);
+void wx_setup_vfmrqc_vf(struct wx *wx);
+void wx_configure_tx_vf(struct wx *wx);
+void wx_configure_rx_ring_vf(struct wx *wx, struct wx_ring *ring);
+
+#endif /* _WX_VF_LIB_H_ */
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* Re: [PATCH net-next v3 03/12] net: libwx: add wangxun vf common api
2025-07-04 9:49 ` [PATCH net-next v3 03/12] net: libwx: add wangxun vf common api Mengyuan Lou
@ 2025-07-09 13:51 ` Vadim Fedorenko
0 siblings, 0 replies; 15+ messages in thread
From: Vadim Fedorenko @ 2025-07-09 13:51 UTC (permalink / raw)
To: Mengyuan Lou, netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu
On 04/07/2025 10:49, Mengyuan Lou wrote:
> Add common wx_configure_vf and wx_set_mac_vf for
> ngbevf and txgbevf.
>
> Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
> ---
> drivers/net/ethernet/wangxun/libwx/Makefile | 2 +-
> drivers/net/ethernet/wangxun/libwx/wx_hw.c | 3 +-
> drivers/net/ethernet/wangxun/libwx/wx_hw.h | 1 +
> drivers/net/ethernet/wangxun/libwx/wx_type.h | 4 +
> drivers/net/ethernet/wangxun/libwx/wx_vf.h | 50 ++++
> .../net/ethernet/wangxun/libwx/wx_vf_common.c | 196 ++++++++++++
> .../net/ethernet/wangxun/libwx/wx_vf_common.h | 14 +
> .../net/ethernet/wangxun/libwx/wx_vf_lib.c | 280 ++++++++++++++++++
> .../net/ethernet/wangxun/libwx/wx_vf_lib.h | 14 +
> 9 files changed, 562 insertions(+), 2 deletions(-)
> create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
> create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
> create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf_lib.c
> create mode 100644 drivers/net/ethernet/wangxun/libwx/wx_vf_lib.h
>
[...]
> diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf.h b/drivers/net/ethernet/wangxun/libwx/wx_vf.h
> index c523ef3e8502..e863a74c291d 100644
> --- a/drivers/net/ethernet/wangxun/libwx/wx_vf.h
> +++ b/drivers/net/ethernet/wangxun/libwx/wx_vf.h
> @@ -14,6 +14,7 @@
> #define WX_VXMRQC 0x78
> #define WX_VXICR 0x100
> #define WX_VXIMS 0x108
> +#define WX_VXIMC 0x10C
> #define WX_VF_IRQ_CLEAR_MASK 7
> #define WX_VF_MAX_TX_QUEUES 4
> #define WX_VF_MAX_RX_QUEUES 4
> @@ -22,6 +23,12 @@
> #define WX_VXRXDCTL_ENABLE BIT(0)
> #define WX_VXTXDCTL_FLUSH BIT(26)
>
> +#define WX_VXITR(i) (0x200 + (4 * (i))) /* i=[0,1] */
> +#define WX_VXITR_MASK GENMASK(8, 0)
> +#define WX_VXITR_CNT_WDIS BIT(31)
> +#define WX_VXIVAR_MISC 0x260
> +#define WX_VXIVAR(i) (0x240 + (4 * (i))) /* i=[0,3] */
> +
> #define WX_VXRXDCTL_RSCMAX(f) FIELD_PREP(GENMASK(24, 23), f)
> #define WX_VXRXDCTL_BUFLEN(f) FIELD_PREP(GENMASK(6, 1), f)
> #define WX_VXRXDCTL_BUFSZ(f) FIELD_PREP(GENMASK(11, 8), f)
> @@ -44,6 +51,49 @@
> #define WX_RX_HDR_SIZE 256
> #define WX_RX_BUF_SIZE 2048
>
> +#define WX_RXBUFFER_2048 (2048)
extra parentheses are not needed
> +#define WX_RXBUFFER_3072 3072
> +
> +/* Receive Path */
> +#define WX_VXRDBAL(r) (0x1000 + (0x40 * (r)))
> +#define WX_VXRDBAH(r) (0x1004 + (0x40 * (r)))
> +#define WX_VXRDT(r) (0x1008 + (0x40 * (r)))
> +#define WX_VXRDH(r) (0x100C + (0x40 * (r)))
[...]
^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH net-next v3 04/12] net: wangxun: add txgbevf build
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (2 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 03/12] net: libwx: add wangxun vf common api Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 05/12] net: txgbevf: add sw init pci info and reset hardware Mengyuan Lou
` (8 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Add doc build infrastructure for txgbevf driver.
Implement the basic PCI driver loading and unloading interface.
Initialize the id_table which support 10/25/40G virtual
functions for Wangxun.
Ioremap the space of bar0 and bar4 which will be used.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
.../device_drivers/ethernet/index.rst | 1 +
.../ethernet/wangxun/txgbevf.rst | 16 ++
drivers/net/ethernet/wangxun/Kconfig | 18 ++
drivers/net/ethernet/wangxun/Makefile | 1 +
.../net/ethernet/wangxun/libwx/wx_vf_common.c | 38 +++++
.../net/ethernet/wangxun/libwx/wx_vf_common.h | 4 +
drivers/net/ethernet/wangxun/txgbevf/Makefile | 9 +
.../ethernet/wangxun/txgbevf/txgbevf_main.c | 154 ++++++++++++++++++
.../ethernet/wangxun/txgbevf/txgbevf_type.h | 20 +++
9 files changed, 261 insertions(+)
create mode 100644 Documentation/networking/device_drivers/ethernet/wangxun/txgbevf.rst
create mode 100644 drivers/net/ethernet/wangxun/txgbevf/Makefile
create mode 100644 drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
create mode 100644 drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
diff --git a/Documentation/networking/device_drivers/ethernet/index.rst b/Documentation/networking/device_drivers/ethernet/index.rst
index 139b4c75a191..e93453410772 100644
--- a/Documentation/networking/device_drivers/ethernet/index.rst
+++ b/Documentation/networking/device_drivers/ethernet/index.rst
@@ -58,6 +58,7 @@ Contents:
ti/tlan
ti/icssg_prueth
wangxun/txgbe
+ wangxun/txgbevf
wangxun/ngbe
.. only:: subproject and html
diff --git a/Documentation/networking/device_drivers/ethernet/wangxun/txgbevf.rst b/Documentation/networking/device_drivers/ethernet/wangxun/txgbevf.rst
new file mode 100644
index 000000000000..b2f759b7b518
--- /dev/null
+++ b/Documentation/networking/device_drivers/ethernet/wangxun/txgbevf.rst
@@ -0,0 +1,16 @@
+.. SPDX-License-Identifier: GPL-2.0+
+
+===========================================================================
+Linux Base Virtual Function Driver for Wangxun(R) 10/25/40 Gigabit Ethernet
+===========================================================================
+
+WangXun 10/25/40 Gigabit Virtual Function Linux driver.
+Copyright(c) 2015 - 2025 Beijing WangXun Technology Co., Ltd.
+
+Support
+=======
+For general information, go to the website at:
+https://www.net-swift.com
+
+If you got any problem, contact Wangxun support team via nic-support@net-swift.com
+and Cc: netdev.
diff --git a/drivers/net/ethernet/wangxun/Kconfig b/drivers/net/ethernet/wangxun/Kconfig
index e5fc942c28cc..a6ec73e4f300 100644
--- a/drivers/net/ethernet/wangxun/Kconfig
+++ b/drivers/net/ethernet/wangxun/Kconfig
@@ -64,4 +64,22 @@ config TXGBE
To compile this driver as a module, choose M here. The module
will be called txgbe.
+config TXGBEVF
+ tristate "Wangxun(R) 10/25/40G Virtual Function Ethernet support"
+ depends on PCI
+ depends on PCI_MSI
+ select LIBWX
+ select PHYLINK
+ help
+ This driver supports virtual functions for SP1000A, WX1820AL,
+ WX5XXX, WX5XXXAL.
+
+ This driver was formerly named txgbevf.
+
+ More specific information on configuring the driver is in
+ <file:Documentation/networking/device_drivers/ethernet/wangxun/txgbevf.rst>.
+
+ To compile this driver as a module, choose M here. MSI-X interrupt
+ support is required for this driver to work correctly.
+
endif # NET_VENDOR_WANGXUN
diff --git a/drivers/net/ethernet/wangxun/Makefile b/drivers/net/ethernet/wangxun/Makefile
index ca19311dbe38..71371d47a6ee 100644
--- a/drivers/net/ethernet/wangxun/Makefile
+++ b/drivers/net/ethernet/wangxun/Makefile
@@ -5,4 +5,5 @@
obj-$(CONFIG_LIBWX) += libwx/
obj-$(CONFIG_TXGBE) += txgbe/
+obj-$(CONFIG_TXGBEVF) += txgbevf/
obj-$(CONFIG_NGBE) += ngbe/
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
index aac420bf578b..4a3c7d61e5fd 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
@@ -11,6 +11,44 @@
#include "wx_vf_lib.h"
#include "wx_vf_common.h"
+int wxvf_suspend(struct device *dev_d)
+{
+ struct pci_dev *pdev = to_pci_dev(dev_d);
+ struct wx *wx = pci_get_drvdata(pdev);
+
+ netif_device_detach(wx->netdev);
+ pci_disable_device(pdev);
+
+ return 0;
+}
+EXPORT_SYMBOL(wxvf_suspend);
+
+void wxvf_shutdown(struct pci_dev *pdev)
+{
+ wxvf_suspend(&pdev->dev);
+}
+EXPORT_SYMBOL(wxvf_shutdown);
+
+int wxvf_resume(struct device *dev_d)
+{
+ struct pci_dev *pdev = to_pci_dev(dev_d);
+ struct wx *wx = pci_get_drvdata(pdev);
+
+ pci_set_master(pdev);
+ netif_device_attach(wx->netdev);
+
+ return 0;
+}
+EXPORT_SYMBOL(wxvf_resume);
+
+void wxvf_remove(struct pci_dev *pdev)
+{
+ pci_release_selected_regions(pdev,
+ pci_select_bars(pdev, IORESOURCE_MEM));
+ pci_disable_device(pdev);
+}
+EXPORT_SYMBOL(wxvf_remove);
+
static irqreturn_t wx_msix_misc_vf(int __always_unused irq, void *data)
{
struct wx *wx = data;
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
index 9bee9de86cb2..f3b31f33407b 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
@@ -4,6 +4,10 @@
#ifndef _WX_VF_COMMON_H_
#define _WX_VF_COMMON_H_
+int wxvf_suspend(struct device *dev_d);
+void wxvf_shutdown(struct pci_dev *pdev);
+int wxvf_resume(struct device *dev_d);
+void wxvf_remove(struct pci_dev *pdev);
int wx_request_msix_irqs_vf(struct wx *wx);
void wx_negotiate_api_vf(struct wx *wx);
void wx_reset_vf(struct wx *wx);
diff --git a/drivers/net/ethernet/wangxun/txgbevf/Makefile b/drivers/net/ethernet/wangxun/txgbevf/Makefile
new file mode 100644
index 000000000000..4c7e6de04424
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/txgbevf/Makefile
@@ -0,0 +1,9 @@
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd.
+#
+# Makefile for the Wangxun(R) 10/25/40GbE virtual functions driver
+#
+
+obj-$(CONFIG_TXGBE) += txgbevf.o
+
+txgbevf-objs := txgbevf_main.o
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
new file mode 100644
index 000000000000..9e8ddec36913
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
@@ -0,0 +1,154 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/netdevice.h>
+#include <linux/string.h>
+#include <linux/etherdevice.h>
+
+#include "../libwx/wx_type.h"
+#include "../libwx/wx_vf_common.h"
+#include "txgbevf_type.h"
+
+/* txgbevf_pci_tbl - PCI Device ID Table
+ *
+ * Wildcard entries (PCI_ANY_ID) should come last
+ * Last entry must be all 0s
+ *
+ * { Vendor ID, Device ID, SubVendor ID, SubDevice ID,
+ * Class, Class Mask, private data (not used) }
+ */
+static const struct pci_device_id txgbevf_pci_tbl[] = {
+ { PCI_VDEVICE(WANGXUN, TXGBEVF_DEV_ID_SP1000), 0},
+ { PCI_VDEVICE(WANGXUN, TXGBEVF_DEV_ID_WX1820), 0},
+ { PCI_VDEVICE(WANGXUN, TXGBEVF_DEV_ID_AML500F), 0},
+ { PCI_VDEVICE(WANGXUN, TXGBEVF_DEV_ID_AML510F), 0},
+ { PCI_VDEVICE(WANGXUN, TXGBEVF_DEV_ID_AML5024), 0},
+ { PCI_VDEVICE(WANGXUN, TXGBEVF_DEV_ID_AML5124), 0},
+ { PCI_VDEVICE(WANGXUN, TXGBEVF_DEV_ID_AML503F), 0},
+ { PCI_VDEVICE(WANGXUN, TXGBEVF_DEV_ID_AML513F), 0},
+ /* required last entry */
+ { .device = 0 }
+};
+
+/**
+ * txgbevf_probe - Device Initialization Routine
+ * @pdev: PCI device information struct
+ * @ent: entry in txgbevf_pci_tbl
+ *
+ * Return: return 0 on success, negative on failure
+ *
+ * txgbevf_probe initializes an adapter identified by a pci_dev structure.
+ * The OS initialization, configuring of the adapter private structure,
+ * and a hardware reset occur.
+ **/
+static int txgbevf_probe(struct pci_dev *pdev,
+ const struct pci_device_id __always_unused *ent)
+{
+ struct net_device *netdev;
+ struct wx *wx = NULL;
+ int err;
+
+ err = pci_enable_device_mem(pdev);
+ if (err)
+ return err;
+
+ err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
+ if (err) {
+ dev_err(&pdev->dev,
+ "No usable DMA configuration, aborting\n");
+ goto err_pci_disable_dev;
+ }
+
+ err = pci_request_selected_regions(pdev,
+ pci_select_bars(pdev, IORESOURCE_MEM),
+ dev_driver_string(&pdev->dev));
+ if (err) {
+ dev_err(&pdev->dev,
+ "pci_request_selected_regions failed 0x%x\n", err);
+ goto err_pci_disable_dev;
+ }
+
+ pci_set_master(pdev);
+
+ netdev = devm_alloc_etherdev_mqs(&pdev->dev,
+ sizeof(struct wx),
+ TXGBEVF_MAX_TX_QUEUES,
+ TXGBEVF_MAX_RX_QUEUES);
+ if (!netdev) {
+ err = -ENOMEM;
+ goto err_pci_release_regions;
+ }
+
+ SET_NETDEV_DEV(netdev, &pdev->dev);
+
+ wx = netdev_priv(netdev);
+ wx->netdev = netdev;
+ wx->pdev = pdev;
+
+ wx->msg_enable = netif_msg_init(-1, NETIF_MSG_DRV |
+ NETIF_MSG_PROBE | NETIF_MSG_LINK);
+ wx->hw_addr = devm_ioremap(&pdev->dev,
+ pci_resource_start(pdev, 0),
+ pci_resource_len(pdev, 0));
+ if (!wx->hw_addr) {
+ err = -EIO;
+ goto err_pci_release_regions;
+ }
+
+ wx->b4_addr = devm_ioremap(&pdev->dev,
+ pci_resource_start(pdev, 4),
+ pci_resource_len(pdev, 4));
+ if (!wx->b4_addr) {
+ err = -EIO;
+ goto err_pci_release_regions;
+ }
+
+ netdev->features |= NETIF_F_HIGHDMA;
+
+ pci_set_drvdata(pdev, wx);
+
+ return 0;
+
+err_pci_release_regions:
+ pci_release_selected_regions(pdev,
+ pci_select_bars(pdev, IORESOURCE_MEM));
+err_pci_disable_dev:
+ pci_disable_device(pdev);
+ return err;
+}
+
+/**
+ * txgbevf_remove - Device Removal Routine
+ * @pdev: PCI device information struct
+ *
+ * txgbevf_remove is called by the PCI subsystem to alert the driver
+ * that it should release a PCI device. The could be caused by a
+ * Hot-Plug event, or because the driver is going to be removed from
+ * memory.
+ **/
+static void txgbevf_remove(struct pci_dev *pdev)
+{
+ wxvf_remove(pdev);
+}
+
+static DEFINE_SIMPLE_DEV_PM_OPS(txgbevf_pm_ops, wxvf_suspend, wxvf_resume);
+
+static struct pci_driver txgbevf_driver = {
+ .name = KBUILD_MODNAME,
+ .id_table = txgbevf_pci_tbl,
+ .probe = txgbevf_probe,
+ .remove = txgbevf_remove,
+ .shutdown = wxvf_shutdown,
+ /* Power Management Hooks */
+ .driver.pm = pm_sleep_ptr(&txgbevf_pm_ops)
+};
+
+module_pci_driver(txgbevf_driver);
+
+MODULE_DEVICE_TABLE(pci, txgbevf_pci_tbl);
+MODULE_AUTHOR("Beijing WangXun Technology Co., Ltd, <software@trustnetic.com>");
+MODULE_DESCRIPTION("WangXun(R) 10/25/40 Gigabit Virtual Function Network Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
new file mode 100644
index 000000000000..2ba9d0cb63d5
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
@@ -0,0 +1,20 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#ifndef _TXGBEVF_TYPE_H_
+#define _TXGBEVF_TYPE_H_
+
+/* Device IDs */
+#define TXGBEVF_DEV_ID_SP1000 0x1000
+#define TXGBEVF_DEV_ID_WX1820 0x2000
+#define TXGBEVF_DEV_ID_AML500F 0x500F
+#define TXGBEVF_DEV_ID_AML510F 0x510F
+#define TXGBEVF_DEV_ID_AML5024 0x5024
+#define TXGBEVF_DEV_ID_AML5124 0x5124
+#define TXGBEVF_DEV_ID_AML503F 0x503f
+#define TXGBEVF_DEV_ID_AML513F 0x513f
+
+#define TXGBEVF_MAX_RX_QUEUES 4
+#define TXGBEVF_MAX_TX_QUEUES 4
+
+#endif /* _TXGBEVF_TYPE_H_ */
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 05/12] net: txgbevf: add sw init pci info and reset hardware
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (3 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 04/12] net: wangxun: add txgbevf build Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 06/12] net: txgbevf: init interrupts and request irqs Mengyuan Lou
` (7 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Add sw init and reset hw for txgbevf virtual functions
which initialize basic parameters, and then register netdev.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
drivers/net/ethernet/wangxun/libwx/wx_hw.c | 3 +-
.../net/ethernet/wangxun/libwx/wx_vf_common.c | 32 +++++
.../net/ethernet/wangxun/libwx/wx_vf_common.h | 2 +
.../ethernet/wangxun/txgbevf/txgbevf_main.c | 111 ++++++++++++++++++
.../ethernet/wangxun/txgbevf/txgbevf_type.h | 4 +
5 files changed, 151 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
index 27bb33788701..6e830436a19b 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
@@ -2369,7 +2369,8 @@ int wx_sw_init(struct wx *wx)
wx->bus.device = PCI_SLOT(pdev->devfn);
wx->bus.func = PCI_FUNC(pdev->devfn);
- if (wx->oem_svid == PCI_VENDOR_ID_WANGXUN) {
+ if (wx->oem_svid == PCI_VENDOR_ID_WANGXUN ||
+ pdev->is_virtfn) {
wx->subsystem_vendor_id = pdev->subsystem_vendor;
wx->subsystem_device_id = pdev->subsystem_device;
} else {
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
index 4a3c7d61e5fd..ed5daeec598a 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
@@ -43,6 +43,14 @@ EXPORT_SYMBOL(wxvf_resume);
void wxvf_remove(struct pci_dev *pdev)
{
+ struct wx *wx = pci_get_drvdata(pdev);
+ struct net_device *netdev;
+
+ netdev = wx->netdev;
+ unregister_netdev(netdev);
+ kfree(wx->vfinfo);
+ kfree(wx->rss_key);
+ kfree(wx->mac_table);
pci_release_selected_regions(pdev,
pci_select_bars(pdev, IORESOURCE_MEM));
pci_disable_device(pdev);
@@ -232,3 +240,27 @@ int wx_set_mac_vf(struct net_device *netdev, void *p)
return 0;
}
EXPORT_SYMBOL(wx_set_mac_vf);
+
+int wxvf_open(struct net_device *netdev)
+{
+ return 0;
+}
+EXPORT_SYMBOL(wxvf_open);
+
+static void wxvf_down(struct wx *wx)
+{
+ struct net_device *netdev = wx->netdev;
+
+ netif_tx_disable(netdev);
+ wx_reset_vf(wx);
+}
+
+int wxvf_close(struct net_device *netdev)
+{
+ struct wx *wx = netdev_priv(netdev);
+
+ wxvf_down(wx);
+
+ return 0;
+}
+EXPORT_SYMBOL(wxvf_close);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
index f3b31f33407b..272743a3c878 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
@@ -14,5 +14,7 @@ void wx_reset_vf(struct wx *wx);
void wx_set_rx_mode_vf(struct net_device *netdev);
void wx_configure_vf(struct wx *wx);
int wx_set_mac_vf(struct net_device *netdev, void *p);
+int wxvf_open(struct net_device *netdev);
+int wxvf_close(struct net_device *netdev);
#endif /* _WX_VF_COMMON_H_ */
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
index 9e8ddec36913..9918d5b2ee57 100644
--- a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
@@ -9,6 +9,9 @@
#include <linux/etherdevice.h>
#include "../libwx/wx_type.h"
+#include "../libwx/wx_hw.h"
+#include "../libwx/wx_mbx.h"
+#include "../libwx/wx_vf.h"
#include "../libwx/wx_vf_common.h"
#include "txgbevf_type.h"
@@ -33,6 +36,96 @@ static const struct pci_device_id txgbevf_pci_tbl[] = {
{ .device = 0 }
};
+static const struct net_device_ops txgbevf_netdev_ops = {
+ .ndo_open = wxvf_open,
+ .ndo_stop = wxvf_close,
+ .ndo_validate_addr = eth_validate_addr,
+ .ndo_set_mac_address = wx_set_mac_vf,
+};
+
+static void txgbevf_init_type_code(struct wx *wx)
+{
+ switch (wx->device_id) {
+ case TXGBEVF_DEV_ID_SP1000:
+ case TXGBEVF_DEV_ID_WX1820:
+ wx->mac.type = wx_mac_sp;
+ break;
+ case TXGBEVF_DEV_ID_AML500F:
+ case TXGBEVF_DEV_ID_AML510F:
+ case TXGBEVF_DEV_ID_AML5024:
+ case TXGBEVF_DEV_ID_AML5124:
+ case TXGBEVF_DEV_ID_AML503F:
+ case TXGBEVF_DEV_ID_AML513F:
+ wx->mac.type = wx_mac_aml;
+ break;
+ default:
+ wx->mac.type = wx_mac_unknown;
+ break;
+ }
+}
+
+static int txgbevf_sw_init(struct wx *wx)
+{
+ struct net_device *netdev = wx->netdev;
+ struct pci_dev *pdev = wx->pdev;
+ int err;
+
+ /* Initialize pcie info and common capability flags */
+ err = wx_sw_init(wx);
+ if (err < 0)
+ goto err_wx_sw_init;
+
+ /* Initialize the mailbox */
+ err = wx_init_mbx_params_vf(wx);
+ if (err)
+ goto err_init_mbx_params;
+
+ /* Initialize the device type */
+ txgbevf_init_type_code(wx);
+ /* lock to protect mailbox accesses */
+ spin_lock_init(&wx->mbx.mbx_lock);
+
+ err = wx_reset_hw_vf(wx);
+ if (err) {
+ wx_err(wx, "PF still in reset state. Is the PF interface up?\n");
+ goto err_reset_hw;
+ }
+ wx_init_hw_vf(wx);
+ wx_negotiate_api_vf(wx);
+ if (is_zero_ether_addr(wx->mac.addr))
+ dev_info(&pdev->dev,
+ "MAC address not assigned by administrator.\n");
+ eth_hw_addr_set(netdev, wx->mac.addr);
+
+ if (!is_valid_ether_addr(netdev->dev_addr)) {
+ dev_info(&pdev->dev, "Assigning random MAC address\n");
+ eth_hw_addr_random(netdev);
+ ether_addr_copy(wx->mac.addr, netdev->dev_addr);
+ ether_addr_copy(wx->mac.perm_addr, netdev->dev_addr);
+ }
+
+ wx->mac.max_tx_queues = TXGBEVF_MAX_TX_QUEUES;
+ wx->mac.max_rx_queues = TXGBEVF_MAX_RX_QUEUES;
+ /* Enable dynamic interrupt throttling rates */
+ wx->rx_itr_setting = 1;
+ wx->tx_itr_setting = 1;
+ /* set default ring sizes */
+ wx->tx_ring_count = TXGBEVF_DEFAULT_TXD;
+ wx->rx_ring_count = TXGBEVF_DEFAULT_RXD;
+ /* set default work limits */
+ wx->tx_work_limit = TXGBEVF_DEFAULT_TX_WORK;
+ wx->rx_work_limit = TXGBEVF_DEFAULT_RX_WORK;
+
+ return 0;
+err_reset_hw:
+ kfree(wx->vfinfo);
+err_init_mbx_params:
+ kfree(wx->rss_key);
+ kfree(wx->mac_table);
+err_wx_sw_init:
+ return err;
+}
+
/**
* txgbevf_probe - Device Initialization Routine
* @pdev: PCI device information struct
@@ -106,12 +199,30 @@ static int txgbevf_probe(struct pci_dev *pdev,
goto err_pci_release_regions;
}
+ netdev->netdev_ops = &txgbevf_netdev_ops;
+
+ /* setup the private structure */
+ err = txgbevf_sw_init(wx);
+ if (err)
+ goto err_pci_release_regions;
+
netdev->features |= NETIF_F_HIGHDMA;
+ eth_hw_addr_set(netdev, wx->mac.perm_addr);
+ ether_addr_copy(netdev->perm_addr, wx->mac.addr);
+
+ err = register_netdev(netdev);
+ if (err)
+ goto err_register;
+
pci_set_drvdata(pdev, wx);
return 0;
+err_register:
+ kfree(wx->vfinfo);
+ kfree(wx->rss_key);
+ kfree(wx->mac_table);
err_pci_release_regions:
pci_release_selected_regions(pdev,
pci_select_bars(pdev, IORESOURCE_MEM));
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
index 2ba9d0cb63d5..8f4f08ce06c0 100644
--- a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
@@ -16,5 +16,9 @@
#define TXGBEVF_MAX_RX_QUEUES 4
#define TXGBEVF_MAX_TX_QUEUES 4
+#define TXGBEVF_DEFAULT_TXD 128
+#define TXGBEVF_DEFAULT_RXD 128
+#define TXGBEVF_DEFAULT_TX_WORK 256
+#define TXGBEVF_DEFAULT_RX_WORK 256
#endif /* _TXGBEVF_TYPE_H_ */
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 06/12] net: txgbevf: init interrupts and request irqs
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (4 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 05/12] net: txgbevf: add sw init pci info and reset hardware Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 07/12] net: txgbevf: Support Rx and Tx process path Mengyuan Lou
` (6 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Add irq alloc flow functions for vf.
Alloc pcie msix irqs for drivers and request_irq for tx/rx rings
and misc other events.
If the application is successful, config vertors for interrupts.
Enable interrupts mask in wxvf_irq_enable.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
drivers/net/ethernet/wangxun/libwx/wx_hw.c | 6 +++
drivers/net/ethernet/wangxun/libwx/wx_lib.c | 9 +++-
drivers/net/ethernet/wangxun/libwx/wx_type.h | 1 +
.../net/ethernet/wangxun/libwx/wx_vf_common.c | 40 +++++++++++++++--
.../ethernet/wangxun/txgbevf/txgbevf_main.c | 44 +++++++++++++++++++
.../ethernet/wangxun/txgbevf/txgbevf_type.h | 2 +
6 files changed, 97 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_hw.c b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
index 6e830436a19b..58e9d6a38802 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_hw.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_hw.c
@@ -11,6 +11,7 @@
#include "wx_type.h"
#include "wx_lib.h"
#include "wx_sriov.h"
+#include "wx_vf.h"
#include "wx_hw.h"
static int wx_phy_read_reg_mdi(struct mii_bus *bus, int phy_addr, int devnum, int regnum)
@@ -124,6 +125,11 @@ void wx_intr_enable(struct wx *wx, u64 qmask)
{
u32 mask;
+ if (wx->pdev->is_virtfn) {
+ wr32(wx, WX_VXIMC, qmask);
+ return;
+ }
+
mask = (qmask & U32_MAX);
if (mask)
wr32(wx, WX_PX_IMC(0), mask);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_lib.c b/drivers/net/ethernet/wangxun/libwx/wx_lib.c
index 55e252789db3..0e76be1c8154 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_lib.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_lib.c
@@ -1819,7 +1819,7 @@ static int wx_set_interrupt_capability(struct wx *wx)
/* We will try to get MSI-X interrupts first */
ret = wx_acquire_msix_vectors(wx);
- if (ret == 0 || (ret == -ENOMEM))
+ if (ret == 0 || (ret == -ENOMEM) || pdev->is_virtfn)
return ret;
/* Disable VMDq support */
@@ -2170,7 +2170,12 @@ int wx_init_interrupt_scheme(struct wx *wx)
int ret;
/* Number of supported queues */
- wx_set_num_queues(wx);
+ if (wx->pdev->is_virtfn) {
+ if (wx->set_num_queues)
+ wx->set_num_queues(wx);
+ } else {
+ wx_set_num_queues(wx);
+ }
/* Set interrupt mode */
ret = wx_set_interrupt_capability(wx);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h
index 9e5b0d1fcb21..58e9988388a7 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_type.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h
@@ -1324,6 +1324,7 @@ struct wx {
int (*setup_tc)(struct net_device *netdev, u8 tc);
void (*do_reset)(struct net_device *netdev);
int (*ptp_setup_sdp)(struct wx *wx);
+ void (*set_num_queues)(struct wx *wx);
bool pps_enabled;
u64 pps_width;
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
index ed5daeec598a..7442b195425f 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
@@ -17,6 +17,7 @@ int wxvf_suspend(struct device *dev_d)
struct wx *wx = pci_get_drvdata(pdev);
netif_device_detach(wx->netdev);
+ wx_clear_interrupt_scheme(wx);
pci_disable_device(pdev);
return 0;
@@ -35,6 +36,7 @@ int wxvf_resume(struct device *dev_d)
struct wx *wx = pci_get_drvdata(pdev);
pci_set_master(pdev);
+ wx_init_interrupt_scheme(wx);
netif_device_attach(wx->netdev);
return 0;
@@ -51,6 +53,7 @@ void wxvf_remove(struct pci_dev *pdev)
kfree(wx->vfinfo);
kfree(wx->rss_key);
kfree(wx->mac_table);
+ wx_clear_interrupt_scheme(wx);
pci_release_selected_regions(pdev,
pci_select_bars(pdev, IORESOURCE_MEM));
pci_disable_device(pdev);
@@ -93,9 +96,8 @@ int wx_request_msix_irqs_vf(struct wx *wx)
}
}
- err = request_threaded_irq(wx->msix_entry->vector, NULL,
- wx_msix_misc_vf, IRQF_ONESHOT,
- netdev->name, wx);
+ err = request_threaded_irq(wx->msix_entry->vector, wx_msix_misc_vf,
+ NULL, IRQF_ONESHOT, netdev->name, wx);
if (err) {
wx_err(wx, "request_irq for msix_other failed: %d\n", err);
goto free_queue_irqs;
@@ -241,9 +243,35 @@ int wx_set_mac_vf(struct net_device *netdev, void *p)
}
EXPORT_SYMBOL(wx_set_mac_vf);
+static void wxvf_irq_enable(struct wx *wx)
+{
+ wr32(wx, WX_VXIMC, wx->eims_enable_mask);
+}
+
+static void wxvf_up_complete(struct wx *wx)
+{
+ wx_configure_msix_vf(wx);
+
+ /* clear any pending interrupts, may auto mask */
+ wr32(wx, WX_VXICR, U32_MAX);
+ wxvf_irq_enable(wx);
+}
+
int wxvf_open(struct net_device *netdev)
{
+ struct wx *wx = netdev_priv(netdev);
+ int err;
+
+ err = wx_request_msix_irqs_vf(wx);
+ if (err)
+ goto err_reset;
+
+ wxvf_up_complete(wx);
+
return 0;
+err_reset:
+ wx_reset_vf(wx);
+ return err;
}
EXPORT_SYMBOL(wxvf_open);
@@ -251,8 +279,13 @@ static void wxvf_down(struct wx *wx)
{
struct net_device *netdev = wx->netdev;
+ netif_tx_stop_all_queues(netdev);
netif_tx_disable(netdev);
+ wx_napi_disable_all(wx);
wx_reset_vf(wx);
+
+ wx_clean_all_tx_rings(wx);
+ wx_clean_all_rx_rings(wx);
}
int wxvf_close(struct net_device *netdev)
@@ -260,6 +293,7 @@ int wxvf_close(struct net_device *netdev)
struct wx *wx = netdev_priv(netdev);
wxvf_down(wx);
+ wx_free_irq(wx);
return 0;
}
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
index 9918d5b2ee57..a61e4a0781cf 100644
--- a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
@@ -10,6 +10,7 @@
#include "../libwx/wx_type.h"
#include "../libwx/wx_hw.h"
+#include "../libwx/wx_lib.h"
#include "../libwx/wx_mbx.h"
#include "../libwx/wx_vf.h"
#include "../libwx/wx_vf_common.h"
@@ -43,6 +44,39 @@ static const struct net_device_ops txgbevf_netdev_ops = {
.ndo_set_mac_address = wx_set_mac_vf,
};
+static void txgbevf_set_num_queues(struct wx *wx)
+{
+ u32 def_q = 0, num_tcs = 0;
+ u16 rss, queue;
+ int ret = 0;
+
+ /* Start with base case */
+ wx->num_rx_queues = 1;
+ wx->num_tx_queues = 1;
+
+ spin_lock_bh(&wx->mbx.mbx_lock);
+ /* fetch queue configuration from the PF */
+ ret = wx_get_queues_vf(wx, &num_tcs, &def_q);
+ spin_unlock_bh(&wx->mbx.mbx_lock);
+
+ if (ret)
+ return;
+
+ /* we need as many queues as traffic classes */
+ if (num_tcs > 1) {
+ wx->num_rx_queues = num_tcs;
+ } else {
+ rss = min_t(u16, num_online_cpus(), TXGBEVF_MAX_RSS_NUM);
+ queue = min_t(u16, wx->mac.max_rx_queues, wx->mac.max_tx_queues);
+ rss = min_t(u16, queue, rss);
+
+ if (wx->vfinfo->vf_api >= wx_mbox_api_13) {
+ wx->num_rx_queues = rss;
+ wx->num_tx_queues = rss;
+ }
+ }
+}
+
static void txgbevf_init_type_code(struct wx *wx)
{
switch (wx->device_id) {
@@ -80,6 +114,8 @@ static int txgbevf_sw_init(struct wx *wx)
if (err)
goto err_init_mbx_params;
+ /* max q_vectors */
+ wx->mac.max_msix_vectors = TXGBEVF_MAX_MSIX_VECTORS;
/* Initialize the device type */
txgbevf_init_type_code(wx);
/* lock to protect mailbox accesses */
@@ -116,6 +152,8 @@ static int txgbevf_sw_init(struct wx *wx)
wx->tx_work_limit = TXGBEVF_DEFAULT_TX_WORK;
wx->rx_work_limit = TXGBEVF_DEFAULT_RX_WORK;
+ wx->set_num_queues = txgbevf_set_num_queues;
+
return 0;
err_reset_hw:
kfree(wx->vfinfo);
@@ -211,6 +249,10 @@ static int txgbevf_probe(struct pci_dev *pdev,
eth_hw_addr_set(netdev, wx->mac.perm_addr);
ether_addr_copy(netdev->perm_addr, wx->mac.addr);
+ err = wx_init_interrupt_scheme(wx);
+ if (err)
+ goto err_free_sw_init;
+
err = register_netdev(netdev);
if (err)
goto err_register;
@@ -220,6 +262,8 @@ static int txgbevf_probe(struct pci_dev *pdev,
return 0;
err_register:
+ wx_clear_interrupt_scheme(wx);
+err_free_sw_init:
kfree(wx->vfinfo);
kfree(wx->rss_key);
kfree(wx->mac_table);
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
index 8f4f08ce06c0..1364d2b58bb0 100644
--- a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_type.h
@@ -14,6 +14,8 @@
#define TXGBEVF_DEV_ID_AML503F 0x503f
#define TXGBEVF_DEV_ID_AML513F 0x513f
+#define TXGBEVF_MAX_MSIX_VECTORS 2
+#define TXGBEVF_MAX_RSS_NUM 4
#define TXGBEVF_MAX_RX_QUEUES 4
#define TXGBEVF_MAX_TX_QUEUES 4
#define TXGBEVF_DEFAULT_TXD 128
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 07/12] net: txgbevf: Support Rx and Tx process path
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (5 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 06/12] net: txgbevf: init interrupts and request irqs Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 08/12] net: txgbevf: add link update flow Mengyuan Lou
` (5 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Improve the configuration of Rx and Tx ring.
Setup and alloc resources.
Configure Rx and Tx unit on hardware.
Add .ndo_start_xmit support and start all queues.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
.../net/ethernet/wangxun/libwx/wx_vf_common.c | 25 ++++++++++++++++++-
.../ethernet/wangxun/txgbevf/txgbevf_main.c | 2 ++
2 files changed, 26 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
index 7442b195425f..dc3ed0808e15 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
@@ -251,10 +251,14 @@ static void wxvf_irq_enable(struct wx *wx)
static void wxvf_up_complete(struct wx *wx)
{
wx_configure_msix_vf(wx);
+ smp_mb__before_atomic();
+ wx_napi_enable_all(wx);
/* clear any pending interrupts, may auto mask */
wr32(wx, WX_VXICR, U32_MAX);
wxvf_irq_enable(wx);
+ /* enable transmits */
+ netif_tx_start_all_queues(wx->netdev);
}
int wxvf_open(struct net_device *netdev)
@@ -262,13 +266,31 @@ int wxvf_open(struct net_device *netdev)
struct wx *wx = netdev_priv(netdev);
int err;
- err = wx_request_msix_irqs_vf(wx);
+ err = wx_setup_resources(wx);
if (err)
goto err_reset;
+ wx_configure_vf(wx);
+
+ err = wx_request_msix_irqs_vf(wx);
+ if (err)
+ goto err_free_resources;
+
+ /* Notify the stack of the actual queue counts. */
+ err = netif_set_real_num_tx_queues(netdev, wx->num_tx_queues);
+ if (err)
+ goto err_free_irq;
+
+ err = netif_set_real_num_rx_queues(netdev, wx->num_rx_queues);
+ if (err)
+ goto err_free_irq;
wxvf_up_complete(wx);
return 0;
+err_free_irq:
+ wx_free_irq(wx);
+err_free_resources:
+ wx_free_resources(wx);
err_reset:
wx_reset_vf(wx);
return err;
@@ -294,6 +316,7 @@ int wxvf_close(struct net_device *netdev)
wxvf_down(wx);
wx_free_irq(wx);
+ wx_free_resources(wx);
return 0;
}
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
index a61e4a0781cf..57e67804b8b7 100644
--- a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
@@ -40,6 +40,7 @@ static const struct pci_device_id txgbevf_pci_tbl[] = {
static const struct net_device_ops txgbevf_netdev_ops = {
.ndo_open = wxvf_open,
.ndo_stop = wxvf_close,
+ .ndo_start_xmit = wx_xmit_frame,
.ndo_validate_addr = eth_validate_addr,
.ndo_set_mac_address = wx_set_mac_vf,
};
@@ -258,6 +259,7 @@ static int txgbevf_probe(struct pci_dev *pdev,
goto err_register;
pci_set_drvdata(pdev, wx);
+ netif_tx_stop_all_queues(netdev);
return 0;
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 08/12] net: txgbevf: add link update flow
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (6 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 07/12] net: txgbevf: Support Rx and Tx process path Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 09/12] net: wangxun: add ngbevf build Mengyuan Lou
` (4 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Add link update flow to wangxun 10/25/40G virtual functions.
Get link status from pf in mbox, and if it is failed then
check the vx_status, because vx_status switching is too slow.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
drivers/net/ethernet/wangxun/libwx/wx_type.h | 2 +
drivers/net/ethernet/wangxun/libwx/wx_vf.c | 126 ++++++++++++++++++
drivers/net/ethernet/wangxun/libwx/wx_vf.h | 14 ++
.../net/ethernet/wangxun/libwx/wx_vf_common.c | 91 +++++++++++++
.../net/ethernet/wangxun/libwx/wx_vf_common.h | 2 +
.../ethernet/wangxun/txgbevf/txgbevf_main.c | 3 +
6 files changed, 238 insertions(+)
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_type.h b/drivers/net/ethernet/wangxun/libwx/wx_type.h
index 58e9988388a7..42b0e65fe983 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_type.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_type.h
@@ -1206,6 +1206,8 @@ enum wx_pf_flags {
WX_FLAG_PTP_PPS_ENABLED,
WX_FLAG_NEED_LINK_CONFIG,
WX_FLAG_NEED_SFP_RESET,
+ WX_FLAG_NEED_UPDATE_LINK,
+ WX_FLAG_NEED_DO_RESET,
WX_PF_FLAGS_NBITS /* must be last */
};
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf.c b/drivers/net/ethernet/wangxun/libwx/wx_vf.c
index 165b83e9098f..7567216a005f 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf.c
@@ -471,3 +471,129 @@ int wx_get_queues_vf(struct wx *wx, u32 *num_tcs, u32 *default_tc)
return 0;
}
EXPORT_SYMBOL(wx_get_queues_vf);
+
+static int wx_get_link_status_from_pf(struct wx *wx, u32 *msgbuf)
+{
+ u32 links_reg = msgbuf[1];
+
+ if (msgbuf[1] & WX_PF_NOFITY_VF_NET_NOT_RUNNING)
+ wx->notify_down = true;
+ else
+ wx->notify_down = false;
+
+ if (wx->notify_down) {
+ wx->link = false;
+ wx->speed = SPEED_UNKNOWN;
+ return 0;
+ }
+
+ wx->link = WX_PFLINK_STATUS(links_reg);
+ wx->speed = WX_PFLINK_SPEED(links_reg);
+
+ return 0;
+}
+
+static int wx_pf_ping_vf(struct wx *wx, u32 *msgbuf)
+{
+ if (!(msgbuf[0] & WX_VT_MSGTYPE_CTS))
+ /* msg is not CTS, we need to do reset */
+ return -EINVAL;
+
+ return 0;
+}
+
+static struct wx_link_reg_fields wx_speed_lookup_vf[] = {
+ {wx_mac_unknown},
+ {wx_mac_sp, SPEED_10000, SPEED_1000, SPEED_100, SPEED_UNKNOWN, SPEED_UNKNOWN},
+ {wx_mac_em, SPEED_1000, SPEED_100, SPEED_10, SPEED_UNKNOWN, SPEED_UNKNOWN},
+ {wx_mac_aml, SPEED_40000, SPEED_25000, SPEED_10000, SPEED_1000, SPEED_UNKNOWN},
+ {wx_mac_aml40, SPEED_40000, SPEED_25000, SPEED_10000, SPEED_1000, SPEED_UNKNOWN},
+};
+
+static void wx_check_physical_link(struct wx *wx)
+{
+ u32 val, link_val;
+ int ret;
+
+ /* get link status from hw status reg
+ * for SFP+ modules and DA cables, it can take up to 500usecs
+ * before the link status is correct
+ */
+ if (wx->mac.type == wx_mac_em)
+ ret = read_poll_timeout_atomic(rd32, val, val & GENMASK(4, 1),
+ 100, 500, false, wx, WX_VXSTATUS);
+ else
+ ret = read_poll_timeout_atomic(rd32, val, val & BIT(0), 100,
+ 500, false, wx, WX_VXSTATUS);
+ if (ret) {
+ wx->speed = SPEED_UNKNOWN;
+ wx->link = false;
+ return;
+ }
+
+ wx->link = true;
+ link_val = WX_VXSTATUS_SPEED(val);
+
+ if (link_val & BIT(0))
+ wx->speed = wx_speed_lookup_vf[wx->mac.type].bit0_f;
+ else if (link_val & BIT(1))
+ wx->speed = wx_speed_lookup_vf[wx->mac.type].bit1_f;
+ else if (link_val & BIT(2))
+ wx->speed = wx_speed_lookup_vf[wx->mac.type].bit2_f;
+ else if (link_val & BIT(3))
+ wx->speed = wx_speed_lookup_vf[wx->mac.type].bit3_f;
+ else
+ wx->speed = SPEED_UNKNOWN;
+}
+
+int wx_check_mac_link_vf(struct wx *wx)
+{
+ struct wx_mbx_info *mbx = &wx->mbx;
+ u32 msgbuf[2] = {0};
+ int ret = 0;
+
+ if (!mbx->timeout)
+ goto out;
+
+ wx_check_for_rst_vf(wx);
+ if (!wx_check_for_msg_vf(wx))
+ ret = wx_read_mbx_vf(wx, msgbuf, 2);
+ if (ret)
+ goto out;
+
+ switch (msgbuf[0] & GENMASK(8, 0)) {
+ case WX_PF_NOFITY_VF_LINK_STATUS | WX_PF_CONTROL_MSG:
+ ret = wx_get_link_status_from_pf(wx, msgbuf);
+ goto out;
+ case WX_PF_CONTROL_MSG:
+ ret = wx_pf_ping_vf(wx, msgbuf);
+ goto out;
+ case 0:
+ if (msgbuf[0] & WX_VT_MSGTYPE_NACK) {
+ /* msg is NACK, we must have lost CTS status */
+ ret = -EBUSY;
+ goto out;
+ }
+ /* no message, check link status */
+ wx_check_physical_link(wx);
+ goto out;
+ default:
+ break;
+ }
+
+ if (!(msgbuf[0] & WX_VT_MSGTYPE_CTS)) {
+ /* msg is not CTS and is NACK we must have lost CTS status */
+ if (msgbuf[0] & WX_VT_MSGTYPE_NACK)
+ ret = -EBUSY;
+ goto out;
+ }
+
+ /* the pf is talking, if we timed out in the past we reinit */
+ if (!mbx->timeout) {
+ ret = -EBUSY;
+ goto out;
+ }
+
+out:
+ return ret;
+}
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf.h b/drivers/net/ethernet/wangxun/libwx/wx_vf.h
index e863a74c291d..fec1126703e3 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf.h
@@ -94,6 +94,19 @@
#define WX_VXMRQC_RSS_EN BIT(8)
#define WX_VXMRQC_RSS_HASH(f) FIELD_PREP(GENMASK(15, 13), f)
+#define WX_PFLINK_STATUS(g) FIELD_GET(BIT(0), g)
+#define WX_PFLINK_SPEED(g) FIELD_GET(GENMASK(31, 1), g)
+#define WX_VXSTATUS_SPEED(g) FIELD_GET(GENMASK(4, 1), g)
+
+struct wx_link_reg_fields {
+ u32 mac_type;
+ u32 bit0_f;
+ u32 bit1_f;
+ u32 bit2_f;
+ u32 bit3_f;
+ u32 bit4_f;
+};
+
void wx_init_hw_vf(struct wx *wx);
int wx_reset_hw_vf(struct wx *wx);
void wx_get_mac_addr_vf(struct wx *wx, u8 *mac_addr);
@@ -109,5 +122,6 @@ int wx_update_xcast_mode_vf(struct wx *wx, int xcast_mode);
int wx_get_link_state_vf(struct wx *wx, u16 *link_state);
int wx_set_vfta_vf(struct wx *wx, u32 vlan, u32 vind, bool vlan_on,
bool vlvf_bypass);
+int wx_check_mac_link_vf(struct wx *wx);
#endif /* _WX_VF_H_ */
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
index dc3ed0808e15..ade2bfe563aa 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.c
@@ -48,6 +48,7 @@ void wxvf_remove(struct pci_dev *pdev)
struct wx *wx = pci_get_drvdata(pdev);
struct net_device *netdev;
+ cancel_work_sync(&wx->service_task);
netdev = wx->netdev;
unregister_netdev(netdev);
kfree(wx->vfinfo);
@@ -64,6 +65,7 @@ static irqreturn_t wx_msix_misc_vf(int __always_unused irq, void *data)
{
struct wx *wx = data;
+ set_bit(WX_FLAG_NEED_UPDATE_LINK, wx->flags);
/* Clear the interrupt */
if (netif_running(wx->netdev))
wr32(wx, WX_VXIMC, wx->eims_other);
@@ -243,6 +245,24 @@ int wx_set_mac_vf(struct net_device *netdev, void *p)
}
EXPORT_SYMBOL(wx_set_mac_vf);
+void wxvf_watchdog_update_link(struct wx *wx)
+{
+ int err;
+
+ if (!test_bit(WX_FLAG_NEED_UPDATE_LINK, wx->flags))
+ return;
+
+ spin_lock_bh(&wx->mbx.mbx_lock);
+ err = wx_check_mac_link_vf(wx);
+ spin_unlock_bh(&wx->mbx.mbx_lock);
+ if (err) {
+ wx->link = false;
+ set_bit(WX_FLAG_NEED_DO_RESET, wx->flags);
+ }
+ clear_bit(WX_FLAG_NEED_UPDATE_LINK, wx->flags);
+}
+EXPORT_SYMBOL(wxvf_watchdog_update_link);
+
static void wxvf_irq_enable(struct wx *wx)
{
wr32(wx, WX_VXIMC, wx->eims_enable_mask);
@@ -250,6 +270,11 @@ static void wxvf_irq_enable(struct wx *wx)
static void wxvf_up_complete(struct wx *wx)
{
+ /* Always set the carrier off */
+ netif_carrier_off(wx->netdev);
+ mod_timer(&wx->service_timer, jiffies + HZ);
+ set_bit(WX_FLAG_NEED_UPDATE_LINK, wx->flags);
+
wx_configure_msix_vf(wx);
smp_mb__before_atomic();
wx_napi_enable_all(wx);
@@ -301,8 +326,10 @@ static void wxvf_down(struct wx *wx)
{
struct net_device *netdev = wx->netdev;
+ timer_delete_sync(&wx->service_timer);
netif_tx_stop_all_queues(netdev);
netif_tx_disable(netdev);
+ netif_carrier_off(netdev);
wx_napi_disable_all(wx);
wx_reset_vf(wx);
@@ -310,6 +337,34 @@ static void wxvf_down(struct wx *wx)
wx_clean_all_rx_rings(wx);
}
+static void wxvf_reinit_locked(struct wx *wx)
+{
+ while (test_and_set_bit(WX_STATE_RESETTING, wx->state))
+ usleep_range(1000, 2000);
+ wxvf_down(wx);
+ wx_free_irq(wx);
+ wx_configure_vf(wx);
+ wx_request_msix_irqs_vf(wx);
+ wxvf_up_complete(wx);
+ clear_bit(WX_STATE_RESETTING, wx->state);
+}
+
+static void wxvf_reset_subtask(struct wx *wx)
+{
+ if (!test_bit(WX_FLAG_NEED_DO_RESET, wx->flags))
+ return;
+ clear_bit(WX_FLAG_NEED_DO_RESET, wx->flags);
+
+ rtnl_lock();
+ if (test_bit(WX_STATE_RESETTING, wx->state) ||
+ !(netif_running(wx->netdev))) {
+ rtnl_unlock();
+ return;
+ }
+ wxvf_reinit_locked(wx);
+ rtnl_unlock();
+}
+
int wxvf_close(struct net_device *netdev)
{
struct wx *wx = netdev_priv(netdev);
@@ -321,3 +376,39 @@ int wxvf_close(struct net_device *netdev)
return 0;
}
EXPORT_SYMBOL(wxvf_close);
+
+static void wxvf_link_config_subtask(struct wx *wx)
+{
+ struct net_device *netdev = wx->netdev;
+
+ wxvf_watchdog_update_link(wx);
+ if (wx->link) {
+ if (netif_carrier_ok(netdev))
+ return;
+ netif_carrier_on(netdev);
+ netdev_info(netdev, "Link is Up - %s\n",
+ phy_speed_to_str(wx->speed));
+ } else {
+ if (!netif_carrier_ok(netdev))
+ return;
+ netif_carrier_off(netdev);
+ netdev_info(netdev, "Link is Down\n");
+ }
+}
+
+static void wxvf_service_task(struct work_struct *work)
+{
+ struct wx *wx = container_of(work, struct wx, service_task);
+
+ wxvf_link_config_subtask(wx);
+ wxvf_reset_subtask(wx);
+ wx_service_event_complete(wx);
+}
+
+void wxvf_init_service(struct wx *wx)
+{
+ timer_setup(&wx->service_timer, wx_service_timer, 0);
+ INIT_WORK(&wx->service_task, wxvf_service_task);
+ clear_bit(WX_STATE_SERVICE_SCHED, wx->state);
+}
+EXPORT_SYMBOL(wxvf_init_service);
diff --git a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
index 272743a3c878..cbbb1b178cb2 100644
--- a/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
+++ b/drivers/net/ethernet/wangxun/libwx/wx_vf_common.h
@@ -14,7 +14,9 @@ void wx_reset_vf(struct wx *wx);
void wx_set_rx_mode_vf(struct net_device *netdev);
void wx_configure_vf(struct wx *wx);
int wx_set_mac_vf(struct net_device *netdev, void *p);
+void wxvf_watchdog_update_link(struct wx *wx);
int wxvf_open(struct net_device *netdev);
int wxvf_close(struct net_device *netdev);
+void wxvf_init_service(struct wx *wx);
#endif /* _WX_VF_COMMON_H_ */
diff --git a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
index 57e67804b8b7..ebfce3cf753e 100644
--- a/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
+++ b/drivers/net/ethernet/wangxun/txgbevf/txgbevf_main.c
@@ -250,6 +250,7 @@ static int txgbevf_probe(struct pci_dev *pdev,
eth_hw_addr_set(netdev, wx->mac.perm_addr);
ether_addr_copy(netdev->perm_addr, wx->mac.addr);
+ wxvf_init_service(wx);
err = wx_init_interrupt_scheme(wx);
if (err)
goto err_free_sw_init;
@@ -266,6 +267,8 @@ static int txgbevf_probe(struct pci_dev *pdev,
err_register:
wx_clear_interrupt_scheme(wx);
err_free_sw_init:
+ timer_delete_sync(&wx->service_timer);
+ cancel_work_sync(&wx->service_task);
kfree(wx->vfinfo);
kfree(wx->rss_key);
kfree(wx->mac_table);
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 09/12] net: wangxun: add ngbevf build
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (7 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 08/12] net: txgbevf: add link update flow Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 10/12] net: ngbevf: add sw init pci info and reset hardware Mengyuan Lou
` (3 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Add doc build infrastructure for ngbevf driver.
Implement the basic PCI driver loading and unloading interface.
Initialize the id_table which support 1G virtual
functions for Wangxun.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
.../device_drivers/ethernet/index.rst | 1 +
.../ethernet/wangxun/ngbevf.rst | 16 ++
drivers/net/ethernet/wangxun/Kconfig | 15 ++
drivers/net/ethernet/wangxun/Makefile | 1 +
drivers/net/ethernet/wangxun/ngbevf/Makefile | 9 ++
.../net/ethernet/wangxun/ngbevf/ngbevf_main.c | 149 ++++++++++++++++++
.../net/ethernet/wangxun/ngbevf/ngbevf_type.h | 24 +++
7 files changed, 215 insertions(+)
create mode 100644 Documentation/networking/device_drivers/ethernet/wangxun/ngbevf.rst
create mode 100644 drivers/net/ethernet/wangxun/ngbevf/Makefile
create mode 100644 drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
create mode 100644 drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
diff --git a/Documentation/networking/device_drivers/ethernet/index.rst b/Documentation/networking/device_drivers/ethernet/index.rst
index e93453410772..40ac552641a3 100644
--- a/Documentation/networking/device_drivers/ethernet/index.rst
+++ b/Documentation/networking/device_drivers/ethernet/index.rst
@@ -60,6 +60,7 @@ Contents:
wangxun/txgbe
wangxun/txgbevf
wangxun/ngbe
+ wangxun/ngbevf
.. only:: subproject and html
diff --git a/Documentation/networking/device_drivers/ethernet/wangxun/ngbevf.rst b/Documentation/networking/device_drivers/ethernet/wangxun/ngbevf.rst
new file mode 100644
index 000000000000..a39e3d5a1038
--- /dev/null
+++ b/Documentation/networking/device_drivers/ethernet/wangxun/ngbevf.rst
@@ -0,0 +1,16 @@
+.. SPDX-License-Identifier: GPL-2.0+
+
+==================================================================
+Linux Base Virtual Function Driver for Wangxun(R) Gigabit Ethernet
+==================================================================
+
+WangXun Gigabit Virtual Function Linux driver.
+Copyright(c) 2015 - 2025 Beijing WangXun Technology Co., Ltd.
+
+Support
+=======
+For general information, go to the website at:
+https://www.net-swift.com
+
+If you got any problem, contact Wangxun support team via nic-support@net-swift.com
+and Cc: netdev.
diff --git a/drivers/net/ethernet/wangxun/Kconfig b/drivers/net/ethernet/wangxun/Kconfig
index a6ec73e4f300..c548f4e80565 100644
--- a/drivers/net/ethernet/wangxun/Kconfig
+++ b/drivers/net/ethernet/wangxun/Kconfig
@@ -82,4 +82,19 @@ config TXGBEVF
To compile this driver as a module, choose M here. MSI-X interrupt
support is required for this driver to work correctly.
+config NGBEVF
+ tristate "Wangxun(R) GbE Virtual Function Ethernet support"
+ depends on PCI_MSI
+ select LIBWX
+ help
+ This driver supports virtual functions for WX1860, WX1860AL.
+
+ This driver was formerly named ngbevf.
+
+ More specific information on configuring the driver is in
+ <file:Documentation/networking/device_drivers/ethernet/wangxun/ngbevf.rst>.
+
+ To compile this driver as a module, choose M here. MSI-X interrupt
+ support is required for this driver to work correctly.
+
endif # NET_VENDOR_WANGXUN
diff --git a/drivers/net/ethernet/wangxun/Makefile b/drivers/net/ethernet/wangxun/Makefile
index 71371d47a6ee..0a71a710b717 100644
--- a/drivers/net/ethernet/wangxun/Makefile
+++ b/drivers/net/ethernet/wangxun/Makefile
@@ -7,3 +7,4 @@ obj-$(CONFIG_LIBWX) += libwx/
obj-$(CONFIG_TXGBE) += txgbe/
obj-$(CONFIG_TXGBEVF) += txgbevf/
obj-$(CONFIG_NGBE) += ngbe/
+obj-$(CONFIG_NGBEVF) += ngbevf/
diff --git a/drivers/net/ethernet/wangxun/ngbevf/Makefile b/drivers/net/ethernet/wangxun/ngbevf/Makefile
new file mode 100644
index 000000000000..11a4f15e2ce3
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/ngbevf/Makefile
@@ -0,0 +1,9 @@
+# SPDX-License-Identifier: GPL-2.0
+# Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd.
+#
+# Makefile for the Wangxun(R) 1GbE virtual functions driver
+#
+
+obj-$(CONFIG_NGBE) += ngbevf.o
+
+ngbevf-objs := ngbevf_main.o
diff --git a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
new file mode 100644
index 000000000000..77025e7deeeb
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
@@ -0,0 +1,149 @@
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#include <linux/types.h>
+#include <linux/module.h>
+#include <linux/pci.h>
+#include <linux/netdevice.h>
+#include <linux/string.h>
+#include <linux/etherdevice.h>
+
+#include "../libwx/wx_type.h"
+#include "../libwx/wx_vf_common.h"
+#include "ngbevf_type.h"
+
+/* ngbevf_pci_tbl - PCI Device ID Table
+ *
+ * Wildcard entries (PCI_ANY_ID) should come last
+ * Last entry must be all 0s
+ *
+ * { Vendor ID, Device ID, SubVendor ID, SubDevice ID,
+ * Class, Class Mask, private data (not used) }
+ */
+static const struct pci_device_id ngbevf_pci_tbl[] = {
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860AL_W), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860A2), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860A2S), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860A4), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860A4S), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860AL2), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860AL2S), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860AL4), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860AL4S), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860NCSI), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860A1), 0},
+ { PCI_VDEVICE(WANGXUN, NGBEVF_DEV_ID_EM_WX1860AL1), 0},
+ /* required last entry */
+ { .device = 0 }
+};
+
+/**
+ * ngbevf_probe - Device Initialization Routine
+ * @pdev: PCI device information struct
+ * @ent: entry in ngbevf_pci_tbl
+ *
+ * Return: return 0 on success, negative on failure
+ *
+ * ngbevf_probe initializes an adapter identified by a pci_dev structure.
+ * The OS initialization, configuring of the adapter private structure,
+ * and a hardware reset occur.
+ **/
+static int ngbevf_probe(struct pci_dev *pdev,
+ const struct pci_device_id __always_unused *ent)
+{
+ struct net_device *netdev;
+ struct wx *wx = NULL;
+ int err;
+
+ err = pci_enable_device_mem(pdev);
+ if (err)
+ return err;
+
+ err = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
+ if (err) {
+ dev_err(&pdev->dev,
+ "No usable DMA configuration, aborting\n");
+ goto err_pci_disable_dev;
+ }
+
+ err = pci_request_selected_regions(pdev,
+ pci_select_bars(pdev, IORESOURCE_MEM),
+ dev_driver_string(&pdev->dev));
+ if (err) {
+ dev_err(&pdev->dev,
+ "pci_request_selected_regions failed 0x%x\n", err);
+ goto err_pci_disable_dev;
+ }
+
+ pci_set_master(pdev);
+
+ netdev = devm_alloc_etherdev_mqs(&pdev->dev,
+ sizeof(struct wx),
+ NGBEVF_MAX_TX_QUEUES,
+ NGBEVF_MAX_RX_QUEUES);
+ if (!netdev) {
+ err = -ENOMEM;
+ goto err_pci_release_regions;
+ }
+
+ SET_NETDEV_DEV(netdev, &pdev->dev);
+
+ wx = netdev_priv(netdev);
+ wx->netdev = netdev;
+ wx->pdev = pdev;
+
+ wx->msg_enable = netif_msg_init(-1, NETIF_MSG_DRV |
+ NETIF_MSG_PROBE | NETIF_MSG_LINK);
+ wx->hw_addr = devm_ioremap(&pdev->dev,
+ pci_resource_start(pdev, 0),
+ pci_resource_len(pdev, 0));
+ if (!wx->hw_addr) {
+ err = -EIO;
+ goto err_pci_release_regions;
+ }
+
+ netdev->features |= NETIF_F_HIGHDMA;
+ pci_set_drvdata(pdev, wx);
+
+ return 0;
+
+err_pci_release_regions:
+ pci_release_selected_regions(pdev,
+ pci_select_bars(pdev, IORESOURCE_MEM));
+err_pci_disable_dev:
+ pci_disable_device(pdev);
+ return err;
+}
+
+/**
+ * ngbevf_remove - Device Removal Routine
+ * @pdev: PCI device information struct
+ *
+ * ngbevf_remove is called by the PCI subsystem to alert the driver
+ * that it should release a PCI device. The could be caused by a
+ * Hot-Plug event, or because the driver is going to be removed from
+ * memory.
+ **/
+static void ngbevf_remove(struct pci_dev *pdev)
+{
+ wxvf_remove(pdev);
+}
+
+static DEFINE_SIMPLE_DEV_PM_OPS(ngbevf_pm_ops, wxvf_suspend, wxvf_resume);
+
+static struct pci_driver ngbevf_driver = {
+ .name = KBUILD_MODNAME,
+ .id_table = ngbevf_pci_tbl,
+ .probe = ngbevf_probe,
+ .remove = ngbevf_remove,
+ .shutdown = wxvf_shutdown,
+ /* Power Management Hooks */
+ .driver.pm = pm_sleep_ptr(&ngbevf_pm_ops)
+};
+
+module_pci_driver(ngbevf_driver);
+
+MODULE_DEVICE_TABLE(pci, ngbevf_pci_tbl);
+MODULE_AUTHOR("Beijing WangXun Technology Co., Ltd, <software@trustnetic.com>");
+MODULE_DESCRIPTION("WangXun(R) Gigabit PCI Express Network Driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
new file mode 100644
index 000000000000..c71a244ec6b9
--- /dev/null
+++ b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
@@ -0,0 +1,24 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/* Copyright (c) 2015 - 2025 Beijing WangXun Technology Co., Ltd. */
+
+#ifndef _NGBEVF_TYPE_H_
+#define _NGBEVF_TYPE_H_
+
+/* Device IDs */
+#define NGBEVF_DEV_ID_EM_WX1860AL_W 0x0110
+#define NGBEVF_DEV_ID_EM_WX1860A2 0x0111
+#define NGBEVF_DEV_ID_EM_WX1860A2S 0x0112
+#define NGBEVF_DEV_ID_EM_WX1860A4 0x0113
+#define NGBEVF_DEV_ID_EM_WX1860A4S 0x0114
+#define NGBEVF_DEV_ID_EM_WX1860AL2 0x0115
+#define NGBEVF_DEV_ID_EM_WX1860AL2S 0x0116
+#define NGBEVF_DEV_ID_EM_WX1860AL4 0x0117
+#define NGBEVF_DEV_ID_EM_WX1860AL4S 0x0118
+#define NGBEVF_DEV_ID_EM_WX1860NCSI 0x0119
+#define NGBEVF_DEV_ID_EM_WX1860A1 0x011a
+#define NGBEVF_DEV_ID_EM_WX1860AL1 0x011b
+
+#define NGBEVF_MAX_RX_QUEUES 1
+#define NGBEVF_MAX_TX_QUEUES 1
+
+#endif /* _NGBEVF_TYPE_H_ */
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 10/12] net: ngbevf: add sw init pci info and reset hardware
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (8 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 09/12] net: wangxun: add ngbevf build Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 11/12] net: ngbevf: init interrupts and request irqs Mengyuan Lou
` (2 subsequent siblings)
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Do sw init and reset hw for ngbevf virtual functions, then
register netdev.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
.../net/ethernet/wangxun/ngbevf/ngbevf_main.c | 91 +++++++++++++++++++
.../net/ethernet/wangxun/ngbevf/ngbevf_type.h | 4 +
2 files changed, 95 insertions(+)
diff --git a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
index 77025e7deeeb..4eea682f024b 100644
--- a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
+++ b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
@@ -9,6 +9,9 @@
#include <linux/etherdevice.h>
#include "../libwx/wx_type.h"
+#include "../libwx/wx_hw.h"
+#include "../libwx/wx_mbx.h"
+#include "../libwx/wx_vf.h"
#include "../libwx/wx_vf_common.h"
#include "ngbevf_type.h"
@@ -37,6 +40,75 @@ static const struct pci_device_id ngbevf_pci_tbl[] = {
{ .device = 0 }
};
+static const struct net_device_ops ngbevf_netdev_ops = {
+ .ndo_open = wxvf_open,
+ .ndo_stop = wxvf_close,
+ .ndo_validate_addr = eth_validate_addr,
+ .ndo_set_mac_address = wx_set_mac_vf,
+};
+
+static int ngbevf_sw_init(struct wx *wx)
+{
+ struct net_device *netdev = wx->netdev;
+ struct pci_dev *pdev = wx->pdev;
+ int err;
+
+ /* Initialize pcie info and common capability flags */
+ err = wx_sw_init(wx);
+ if (err < 0)
+ goto err_wx_sw_init;
+
+ /* Initialize the mailbox */
+ err = wx_init_mbx_params_vf(wx);
+ if (err)
+ goto err_init_mbx_params;
+
+ /* Initialize the device type */
+ wx->mac.type = wx_mac_em;
+ /* lock to protect mailbox accesses */
+ spin_lock_init(&wx->mbx.mbx_lock);
+
+ err = wx_reset_hw_vf(wx);
+ if (err) {
+ wx_err(wx, "PF still in reset state. Is the PF interface up?\n");
+ goto err_reset_hw;
+ }
+ wx_init_hw_vf(wx);
+ wx_negotiate_api_vf(wx);
+ if (is_zero_ether_addr(wx->mac.addr))
+ dev_info(&pdev->dev,
+ "MAC address not assigned by administrator.\n");
+ eth_hw_addr_set(netdev, wx->mac.addr);
+
+ if (!is_valid_ether_addr(netdev->dev_addr)) {
+ dev_info(&pdev->dev, "Assigning random MAC address\n");
+ eth_hw_addr_random(netdev);
+ ether_addr_copy(wx->mac.addr, netdev->dev_addr);
+ ether_addr_copy(wx->mac.perm_addr, netdev->dev_addr);
+ }
+
+ wx->mac.max_tx_queues = NGBEVF_MAX_TX_QUEUES;
+ wx->mac.max_rx_queues = NGBEVF_MAX_RX_QUEUES;
+ /* Enable dynamic interrupt throttling rates */
+ wx->rx_itr_setting = 1;
+ wx->tx_itr_setting = 1;
+ /* set default ring sizes */
+ wx->tx_ring_count = NGBEVF_DEFAULT_TXD;
+ wx->rx_ring_count = NGBEVF_DEFAULT_RXD;
+ /* set default work limits */
+ wx->tx_work_limit = NGBEVF_DEFAULT_TX_WORK;
+ wx->rx_work_limit = NGBEVF_DEFAULT_RX_WORK;
+
+ return 0;
+err_reset_hw:
+ kfree(wx->vfinfo);
+err_init_mbx_params:
+ kfree(wx->rss_key);
+ kfree(wx->mac_table);
+err_wx_sw_init:
+ return err;
+}
+
/**
* ngbevf_probe - Device Initialization Routine
* @pdev: PCI device information struct
@@ -102,11 +174,30 @@ static int ngbevf_probe(struct pci_dev *pdev,
goto err_pci_release_regions;
}
+ netdev->netdev_ops = &ngbevf_netdev_ops;
+
+ /* setup the private structure */
+ err = ngbevf_sw_init(wx);
+ if (err)
+ goto err_pci_release_regions;
+
netdev->features |= NETIF_F_HIGHDMA;
+
+ eth_hw_addr_set(netdev, wx->mac.perm_addr);
+ ether_addr_copy(netdev->perm_addr, wx->mac.addr);
+
+ err = register_netdev(netdev);
+ if (err)
+ goto err_register;
+
pci_set_drvdata(pdev, wx);
return 0;
+err_register:
+ kfree(wx->vfinfo);
+ kfree(wx->rss_key);
+ kfree(wx->mac_table);
err_pci_release_regions:
pci_release_selected_regions(pdev,
pci_select_bars(pdev, IORESOURCE_MEM));
diff --git a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
index c71a244ec6b9..dc29349304f1 100644
--- a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
+++ b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
@@ -20,5 +20,9 @@
#define NGBEVF_MAX_RX_QUEUES 1
#define NGBEVF_MAX_TX_QUEUES 1
+#define NGBEVF_DEFAULT_TXD 128
+#define NGBEVF_DEFAULT_RXD 128
+#define NGBEVF_DEFAULT_TX_WORK 256
+#define NGBEVF_DEFAULT_RX_WORK 256
#endif /* _NGBEVF_TYPE_H_ */
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 11/12] net: ngbevf: init interrupts and request irqs
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (9 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 10/12] net: ngbevf: add sw init pci info and reset hardware Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-04 9:49 ` [PATCH net-next v3 12/12] net: ngbevf: add link update flow Mengyuan Lou
2025-07-10 1:50 ` [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions patchwork-bot+netdevbpf
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Add specific parameters for irq alloc, then use
wx_init_interrupt_scheme to initialize interrupt
allocation in probe.
Add .ndo_start_xmit support and start all queues.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
.../net/ethernet/wangxun/ngbevf/ngbevf_main.c | 18 ++++++++++++++++++
.../net/ethernet/wangxun/ngbevf/ngbevf_type.h | 1 +
2 files changed, 19 insertions(+)
diff --git a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
index 4eea682f024b..a629b645d3a1 100644
--- a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
+++ b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
@@ -10,6 +10,7 @@
#include "../libwx/wx_type.h"
#include "../libwx/wx_hw.h"
+#include "../libwx/wx_lib.h"
#include "../libwx/wx_mbx.h"
#include "../libwx/wx_vf.h"
#include "../libwx/wx_vf_common.h"
@@ -43,10 +44,18 @@ static const struct pci_device_id ngbevf_pci_tbl[] = {
static const struct net_device_ops ngbevf_netdev_ops = {
.ndo_open = wxvf_open,
.ndo_stop = wxvf_close,
+ .ndo_start_xmit = wx_xmit_frame,
.ndo_validate_addr = eth_validate_addr,
.ndo_set_mac_address = wx_set_mac_vf,
};
+static void ngbevf_set_num_queues(struct wx *wx)
+{
+ /* Start with base case */
+ wx->num_rx_queues = 1;
+ wx->num_tx_queues = 1;
+}
+
static int ngbevf_sw_init(struct wx *wx)
{
struct net_device *netdev = wx->netdev;
@@ -65,6 +74,7 @@ static int ngbevf_sw_init(struct wx *wx)
/* Initialize the device type */
wx->mac.type = wx_mac_em;
+ wx->mac.max_msix_vectors = NGBEVF_MAX_MSIX_VECTORS;
/* lock to protect mailbox accesses */
spin_lock_init(&wx->mbx.mbx_lock);
@@ -98,6 +108,7 @@ static int ngbevf_sw_init(struct wx *wx)
/* set default work limits */
wx->tx_work_limit = NGBEVF_DEFAULT_TX_WORK;
wx->rx_work_limit = NGBEVF_DEFAULT_RX_WORK;
+ wx->set_num_queues = ngbevf_set_num_queues;
return 0;
err_reset_hw:
@@ -186,15 +197,22 @@ static int ngbevf_probe(struct pci_dev *pdev,
eth_hw_addr_set(netdev, wx->mac.perm_addr);
ether_addr_copy(netdev->perm_addr, wx->mac.addr);
+ err = wx_init_interrupt_scheme(wx);
+ if (err)
+ goto err_free_sw_init;
+
err = register_netdev(netdev);
if (err)
goto err_register;
pci_set_drvdata(pdev, wx);
+ netif_tx_stop_all_queues(netdev);
return 0;
err_register:
+ wx_clear_interrupt_scheme(wx);
+err_free_sw_init:
kfree(wx->vfinfo);
kfree(wx->rss_key);
kfree(wx->mac_table);
diff --git a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
index dc29349304f1..67e761089e99 100644
--- a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
+++ b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_type.h
@@ -18,6 +18,7 @@
#define NGBEVF_DEV_ID_EM_WX1860A1 0x011a
#define NGBEVF_DEV_ID_EM_WX1860AL1 0x011b
+#define NGBEVF_MAX_MSIX_VECTORS 1
#define NGBEVF_MAX_RX_QUEUES 1
#define NGBEVF_MAX_TX_QUEUES 1
#define NGBEVF_DEFAULT_TXD 128
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* [PATCH net-next v3 12/12] net: ngbevf: add link update flow
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (10 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 11/12] net: ngbevf: init interrupts and request irqs Mengyuan Lou
@ 2025-07-04 9:49 ` Mengyuan Lou
2025-07-10 1:50 ` [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions patchwork-bot+netdevbpf
12 siblings, 0 replies; 15+ messages in thread
From: Mengyuan Lou @ 2025-07-04 9:49 UTC (permalink / raw)
To: netdev
Cc: michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu, Mengyuan Lou
Add link update flow to wangxun 1G virtual functions.
Get link status from pf in mbox, and if it is failed then
check the vx_status, because vx_status switching is too slow.
Signed-off-by: Mengyuan Lou <mengyuanlou@net-swift.com>
---
drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
index a629b645d3a1..c1246ab5239c 100644
--- a/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
+++ b/drivers/net/ethernet/wangxun/ngbevf/ngbevf_main.c
@@ -197,6 +197,7 @@ static int ngbevf_probe(struct pci_dev *pdev,
eth_hw_addr_set(netdev, wx->mac.perm_addr);
ether_addr_copy(netdev->perm_addr, wx->mac.addr);
+ wxvf_init_service(wx);
err = wx_init_interrupt_scheme(wx);
if (err)
goto err_free_sw_init;
@@ -213,6 +214,8 @@ static int ngbevf_probe(struct pci_dev *pdev,
err_register:
wx_clear_interrupt_scheme(wx);
err_free_sw_init:
+ timer_delete_sync(&wx->service_timer);
+ cancel_work_sync(&wx->service_task);
kfree(wx->vfinfo);
kfree(wx->rss_key);
kfree(wx->mac_table);
--
2.30.1
^ permalink raw reply related [flat|nested] 15+ messages in thread* Re: [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions
2025-07-04 9:49 [PATCH net-next v3 00/12] Add vf drivers for wangxun virtual functions Mengyuan Lou
` (11 preceding siblings ...)
2025-07-04 9:49 ` [PATCH net-next v3 12/12] net: ngbevf: add link update flow Mengyuan Lou
@ 2025-07-10 1:50 ` patchwork-bot+netdevbpf
12 siblings, 0 replies; 15+ messages in thread
From: patchwork-bot+netdevbpf @ 2025-07-10 1:50 UTC (permalink / raw)
To: Mengyuan Lou
Cc: netdev, michal.swiatkowski, kuba, pabeni, horms, andrew+netdev,
duanqiangwen, linglingzhang, jiawenwu
Hello:
This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:
On Fri, 4 Jul 2025 17:49:11 +0800 you wrote:
> Introduces basic support for Wangxun’s virtual function (VF) network
> drivers, specifically txgbevf and ngbevf. These drivers provide SR-IOV
> VF functionality for Wangxun 10/25/40G network devices.
> The first three patches add common APIs for Wangxun VF drivers, including
> mailbox communication and shared initialization logic.These abstractions
> are placed in libwx to reduce duplication across VF drivers.
> Patches 4–8 introduce the txgbevf driver, including:
> PCI device initialization, Hardware reset, Interrupt setup, Rx/Tx datapath
> implementation and link status changeing flow.
> Patches 9–12 implement the ngbevf driver, mirroring the functionality
> added in txgbevf.
>
> [...]
Here is the summary with links:
- [net-next,v3,01/12] net: libwx: add mailbox api for wangxun vf drivers
https://git.kernel.org/netdev/net-next/c/8259946e6703
- [net-next,v3,02/12] net: libwx: add base vf api for vf drivers
https://git.kernel.org/netdev/net-next/c/ba3b8490bc2e
- [net-next,v3,03/12] net: libwx: add wangxun vf common api
https://git.kernel.org/netdev/net-next/c/eb4898fde1de
- [net-next,v3,04/12] net: wangxun: add txgbevf build
https://git.kernel.org/netdev/net-next/c/377d180bd71c
- [net-next,v3,05/12] net: txgbevf: add sw init pci info and reset hardware
https://git.kernel.org/netdev/net-next/c/4ee8afb44aee
- [net-next,v3,06/12] net: txgbevf: init interrupts and request irqs
https://git.kernel.org/netdev/net-next/c/fd0a2e03bf60
- [net-next,v3,07/12] net: txgbevf: Support Rx and Tx process path
https://git.kernel.org/netdev/net-next/c/ce12ba254655
- [net-next,v3,08/12] net: txgbevf: add link update flow
https://git.kernel.org/netdev/net-next/c/bf68010acc4b
- [net-next,v3,09/12] net: wangxun: add ngbevf build
https://git.kernel.org/netdev/net-next/c/a0008a3658a3
- [net-next,v3,10/12] net: ngbevf: add sw init pci info and reset hardware
https://git.kernel.org/netdev/net-next/c/85494c9bf5b0
- [net-next,v3,11/12] net: ngbevf: init interrupts and request irqs
https://git.kernel.org/netdev/net-next/c/0f71e3a6e59d
- [net-next,v3,12/12] net: ngbevf: add link update flow
https://git.kernel.org/netdev/net-next/c/cfeedf6a420d
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html
^ permalink raw reply [flat|nested] 15+ messages in thread