Skip to contentSkip to conversation

net: dsa: bcm_sf2: Additional DT changes

PATCH In review Florian Fainelli 2 participants 4 messages Sep 22, 2020 – Sep 24, 2020

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

cover net: dsa: bcm_sf2: Additional DT changes

Hi David,

This patch series includes some additional changes to the bcm_sf2 in order to support the Device Tree firmwares provided on such platforms.

Florian Fainelli (2):
net: dsa: bcm_sf2: Disallow port 5 to be a DSA CPU port
net: dsa: bcm_sf2: Include address 0 for MDIO diversion

 drivers/net/dsa/bcm_sf2.c | 13 ++++++++++++-
 1 file changed, 12 insertions(+), 1 deletion(-)
Signature · 1 line
2.25.1
#1 lore raw

PATCH 1/2 net: dsa: bcm_sf2: Disallow port 5 to be a DSA CPU port

While the switch driver is written such that port 5 or 8 could be CPU ports, the use case on Broadcom STB chips is to use port 8 exclusively. The platform firmware does make port 5 comply to a proper DSA CPU port binding by specifiying an "ethernet" phandle. This is undesirable for now until we have an user-space configuration mechanism (such as devlink) which could support dynamically changing the port flavor at run time.

---
 drivers/net/dsa/bcm_sf2.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/net/dsa/bcm_sf2.c +11 -0
index 723820603107..2bd52b03de38 100644
@@ -457,6 +457,7 @@ static void bcm_sf2_identify_ports(struct bcm_sf2_priv *priv,
 {
 	struct device_node *port;
 	unsigned int port_num;
+	struct property *prop;
 	phy_interface_t mode;
 	int err;
 
@@ -483,6 +484,16 @@ static void bcm_sf2_identify_ports(struct bcm_sf2_priv *priv,
 
 		if (of_property_read_bool(port, "brcm,use-bcm-hdr"))
 			priv->brcm_tag_mask |= 1 << port_num;
+
+		/* Ensure that port 5 is not picked up as a DSA CPU port
+		 * flavour but a regular port instead. We should be using
+		 * devlink to be able to set the port flavour.
+		 */
+		if (port_num == 5 && priv->type == BCM7278_DEVICE_ID) {
+			prop = of_find_property(port, "ethernet", NULL);
+			if (prop)
+				of_remove_property(port, prop);
+		}
 	}
 }
 
Signature · 1 line
2.25.1
#2 lore raw

PATCH 2/2 net: dsa: bcm_sf2: Include address 0 for MDIO diversion

In reply to Florian Fainelli

We need to include MDIO address 0, which is how our Device Tree blobs indicate where to find the external BCM53125 switches.

---
 drivers/net/dsa/bcm_sf2.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/dsa/bcm_sf2.c +1 -1
index 2bd52b03de38..0b5b2b33b3b6 100644
@@ -538,7 +538,7 @@ static int bcm_sf2_mdio_register(struct dsa_switch *ds)
 	 * driver.
 	 */
 	if (of_machine_is_compatible("brcm,bcm7445d0"))
-		priv->indir_phy_mask |= (1 << BRCM_PSEUDO_PHY_ADDR);
+		priv->indir_phy_mask |= (1 << BRCM_PSEUDO_PHY_ADDR) | (1 << 0);
 	else
 		priv->indir_phy_mask = 0;
 
Signature · 1 line
2.25.1
maintainer
#3 lore raw
In reply to Florian Fainelli

From: Florian Fainelli <[email protected]>
Date: Tue, 22 Sep 2020 13:03:54 -0700

Quoted 2 lines

This patch series includes some additional changes to the bcm_sf2 in order to support the Device Tree firmwares provided on such platforms.

Series applied, thanks Florian.