Re: [PATCH net-next v4 1/2] net: dsa: microchip: ksz8: Make flow control, speed, and duplex on CPU port configurable

2 messages, 2 authors, 2023-05-20 · open the first message on its own page

Re: [PATCH net-next v4 1/2] net: dsa: microchip: ksz8: Make flow control, speed, and duplex on CPU port configurable

From: Vladimir Oltean <olteanv@gmail.com>
Date: 2023-05-19 23:28:11

On Fri, May 19, 2023 at 02:46:59PM +0200, Oleksij Rempel wrote:
quoted hunk
diff --git a/drivers/net/dsa/microchip/ksz8795.c b/drivers/net/dsa/microchip/ksz8795.c
index f56fca1b1a22..9eedccbf5b7c 100644
--- a/drivers/net/dsa/microchip/ksz8795.c
+++ b/drivers/net/dsa/microchip/ksz8795.c
@@ -1371,6 +1371,57 @@ void ksz8_config_cpu_port(struct dsa_switch *ds)
+/**
+ * ksz8_cpu_port_link_up - Configures the CPU port of the switch.
+ * @dev: The KSZ device instance.
+ * @speed: The desired link speed.
+ * @duplex: The desired duplex mode.
+ * @tx_pause: If true, enables transmit pause.
+ * @rx_pause: If true, enables receive pause.
+ *
+ * Description:
+ * The function configures flow control and speed settings for the CPU
+ * port of the switch based on the desired settings, current duplex mode, and
+ * speed.
+ */
+static void ksz8_cpu_port_link_up(struct ksz_device *dev, int speed, int duplex,
+				  bool tx_pause, bool rx_pause)
+{
+	u8 ctrl = 0;
+
+	/* SW_FLOW_CTRL, SW_HALF_DUPLEX, and SW_10_MBIT bits are bootstrappable
+	 * at least on KSZ8873. They can have different values depending on your
+	 * board setup.
+	 */
+	if (duplex) {
+		if (tx_pause || rx_pause)
+			ctrl |= SW_FLOW_CTRL;
+	} else {
+		ctrl |= SW_HALF_DUPLEX;
+	}
+
+	/* This hardware only supports SPEED_10 and SPEED_100. For SPEED_10
+	 * we need to set the SW_10_MBIT bit. Otherwise, we can leave it 0.
+	 */
+	if (speed == SPEED_10)
+		ctrl |= SW_10_MBIT;
+
+	ksz_rmw8(dev, REG_SW_CTRL_4, SW_HALF_DUPLEX | SW_FLOW_CTRL |
+		 SW_10_MBIT, ctrl);
REG_SW_CTRL_4 ... S_REPLACE_VID_CTRL ... dev->info->regs[P_XMII_CTRL_1] ...
at some point we will need one more consolidation effort here, since we
have at least 3 ways of reaching the same register.
quoted hunk
+}
+
+void ksz8_phylink_mac_link_up(struct ksz_device *dev, int port,
+			      unsigned int mode, phy_interface_t interface,
+			      struct phy_device *phydev, int speed, int duplex,
+			      bool tx_pause, bool rx_pause)
+{
+	/* If the port is the CPU port, apply special handling. Only the CPU
+	 * port is configured via global registers.
+	 */
+	if (dev->cpu_port == port)
+		ksz8_cpu_port_link_up(dev, speed, duplex, tx_pause, rx_pause);
+}
+
 static int ksz8_handle_global_errata(struct dsa_switch *ds)
 {
 	struct ksz_device *dev = ds->priv;
@@ -1419,8 +1470,6 @@ int ksz8_setup(struct dsa_switch *ds)
 	 */
 	ds->vlan_filtering_is_global = true;
 
-	ksz_cfg(dev, S_REPLACE_VID_CTRL, SW_FLOW_CTRL, true);
-
 	/* Enable automatic fast aging when link changed detected. */
 	ksz_cfg(dev, S_LINK_AGING_CTRL, SW_LINK_AUTO_AGING, true);
 
diff --git a/drivers/net/dsa/microchip/ksz_common.c b/drivers/net/dsa/microchip/ksz_common.c
index a4428be5f483..6e19ad70c671 100644
--- a/drivers/net/dsa/microchip/ksz_common.c
+++ b/drivers/net/dsa/microchip/ksz_common.c
@@ -210,6 +210,7 @@ static const struct ksz_dev_ops ksz8_dev_ops = {
 	.mirror_add = ksz8_port_mirror_add,
 	.mirror_del = ksz8_port_mirror_del,
 	.get_caps = ksz8_get_caps,
+	.phylink_mac_link_up = ksz8_phylink_mac_link_up,
Another future consolidation to consider: since all ksz_dev_ops now
provide .phylink_mac_link_up(), the "if" condition here is no longer
necessary:

static void ksz_phylink_mac_link_up(struct dsa_switch *ds, int port,
				    unsigned int mode,
				    phy_interface_t interface,
				    struct phy_device *phydev, int speed,
				    int duplex, bool tx_pause, bool rx_pause)
{
	struct ksz_device *dev = ds->priv;

	if (dev->dev_ops->phylink_mac_link_up)
		dev->dev_ops->phylink_mac_link_up(dev, port, mode, interface,
						  phydev, speed, duplex,
						  tx_pause, rx_pause);
}

which reminds me of the fact that I also had a patch to remove
dev->dev_ops->phylink_mac_config():
https://patchwork.kernel.org/project/netdevbpf/patch/20230316161250.3286055-5-vladimir.oltean@nxp.com/

I give up with that patch set now, since there's zero reviewer interest.
If you want and you think it's useful, you might want to adapt it for
KSZ8873.
 	.config_cpu_port = ksz8_config_cpu_port,
 	.enable_stp_addr = ksz8_enable_stp_addr,
 	.reset = ksz8_reset_switch,
-- 
2.39.2
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>

Re: [PATCH net-next v4 1/2] net: dsa: microchip: ksz8: Make flow control, speed, and duplex on CPU port configurable

From: Oleksij Rempel <o.rempel@pengutronix.de>
Date: 2023-05-20 04:56:54

On Sat, May 20, 2023 at 02:28:02AM +0300, Vladimir Oltean wrote:
quoted
+	/* This hardware only supports SPEED_10 and SPEED_100. For SPEED_10
+	 * we need to set the SW_10_MBIT bit. Otherwise, we can leave it 0.
+	 */
+	if (speed == SPEED_10)
+		ctrl |= SW_10_MBIT;
+
+	ksz_rmw8(dev, REG_SW_CTRL_4, SW_HALF_DUPLEX | SW_FLOW_CTRL |
+		 SW_10_MBIT, ctrl);
REG_SW_CTRL_4 ... S_REPLACE_VID_CTRL ... dev->info->regs[P_XMII_CTRL_1] ...
at some point we will need one more consolidation effort here, since we
have at least 3 ways of reaching the same register.
Agree, the register access is a bit messy now. Your idea about the
regfield API sounds good. We should try it.

Should i convert this patch to use dev->info->regs?
quoted
 	.mirror_del = ksz8_port_mirror_del,
 	.get_caps = ksz8_get_caps,
+	.phylink_mac_link_up = ksz8_phylink_mac_link_up,
Another future consolidation to consider: since all ksz_dev_ops now
provide .phylink_mac_link_up(), the "if" condition here is no longer
necessary:

static void ksz_phylink_mac_link_up(struct dsa_switch *ds, int port,
				    unsigned int mode,
				    phy_interface_t interface,
				    struct phy_device *phydev, int speed,
				    int duplex, bool tx_pause, bool rx_pause)
{
	struct ksz_device *dev = ds->priv;

	if (dev->dev_ops->phylink_mac_link_up)
		dev->dev_ops->phylink_mac_link_up(dev, port, mode, interface,
						  phydev, speed, duplex,
						  tx_pause, rx_pause);
}

which reminds me of the fact that I also had a patch to remove
dev->dev_ops->phylink_mac_config():
https://patchwork.kernel.org/project/netdevbpf/patch/20230316161250.3286055-5-vladimir.oltean@nxp.com/

I give up with that patch set now, since there's zero reviewer interest.
If you want and you think it's useful, you might want to adapt it for
KSZ8873.
Sounds good. I'll take it in to my mainlining queue for KSZ8873.
quoted
 	.config_cpu_port = ksz8_config_cpu_port,
 	.enable_stp_addr = ksz8_enable_stp_addr,
 	.reset = ksz8_reset_switch,
-- 
2.39.2
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Regards,
Oleksij
-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help