Thread (25 messages) 25 messages, 4 authors, 2021-09-10

Re: [PATCH net-next 1/5 v2] net: dsa: rtl8366rb: support bridge offloading

flat view

From: Florian Fainelli <f.fainelli@gmail.com>
Date: 2021-08-31 01:01:33


On 8/30/2021 2:48 PM, Linus Walleij wrote:
quoted hunk ↗ jump to hunk
From: DENG Qingfang <dqfext@gmail.com>

Use port isolation registers to configure bridge offloading.

Tested on the D-Link DIR-685, switching between ports and
sniffing ports to make sure no packets leak.

Cc: Vladimir Oltean <olteanv@gmail.com>
Cc: Alvin Šipraga <redacted>
Cc: Mauri Sandberg <redacted>
Signed-off-by: DENG Qingfang <dqfext@gmail.com>
Signed-off-by: Linus Walleij <redacted>
---
ChangeLog v1->v2:
- introduce RTL8366RB_PORT_ISO_PORTS() to shift the port
   mask into place so we are not confused by the enable
   bit.
- Use this with dsa_user_ports() to isolate the CPU port
   from itself.
---
  drivers/net/dsa/rtl8366rb.c | 87 +++++++++++++++++++++++++++++++++++++
  1 file changed, 87 insertions(+)
diff --git a/drivers/net/dsa/rtl8366rb.c b/drivers/net/dsa/rtl8366rb.c
index a89093bc6c6a..50ee7cd62484 100644
--- a/drivers/net/dsa/rtl8366rb.c
+++ b/drivers/net/dsa/rtl8366rb.c
@@ -300,6 +300,13 @@
  #define RTL8366RB_INTERRUPT_STATUS_REG	0x0442
  #define RTL8366RB_NUM_INTERRUPT		14 /* 0..13 */
  
+/* Port isolation registers */
+#define RTL8366RB_PORT_ISO_BASE		0x0F08
+#define RTL8366RB_PORT_ISO(pnum)	(RTL8366RB_PORT_ISO_BASE + (pnum))
+#define RTL8366RB_PORT_ISO_EN		BIT(0)
+#define RTL8366RB_PORT_ISO_PORTS_MASK	GENMASK(7, 1)
+#define RTL8366RB_PORT_ISO_PORTS(pmask)	(pmask << 1)
+
  /* bits 0..5 enable force when cleared */
  #define RTL8366RB_MAC_FORCE_CTRL_REG	0x0F11
  
@@ -835,6 +842,22 @@ static int rtl8366rb_setup(struct dsa_switch *ds)
  	if (ret)
  		return ret;
  
+	/* Isolate all user ports so only the CPU port can access them */
+	for (i = 0; i < RTL8366RB_PORT_NUM_CPU; i++) {
+		ret = regmap_write(smi->map, RTL8366RB_PORT_ISO(i),
+				   RTL8366RB_PORT_ISO_EN |
+				   RTL8366RB_PORT_ISO_PORTS(BIT(RTL8366RB_PORT_NUM_CPU)));
+		if (ret)
+			return ret;
+	}
+	/* CPU port can access all ports */
+	dev_info(smi->dev, "DSA user port mask: %08x\n", dsa_user_ports(ds));
As mentioned by Alvin, this should be a dev_dbg() or completely eliminated.
quoted hunk ↗ jump to hunk
+	ret = regmap_write(smi->map, RTL8366RB_PORT_ISO(RTL8366RB_PORT_NUM_CPU),
+			   RTL8366RB_PORT_ISO_PORTS(dsa_user_ports(ds))|
+			   RTL8366RB_PORT_ISO_EN);
+	if (ret)
+		return ret;
+
  	/* Set up the "green ethernet" feature */
  	ret = rtl8366rb_jam_table(rtl8366rb_green_jam,
  				  ARRAY_SIZE(rtl8366rb_green_jam), smi, false);
@@ -1127,6 +1150,68 @@ rtl8366rb_port_disable(struct dsa_switch *ds, int port)
  	rb8366rb_set_port_led(smi, port, false);
  }
  
+static int
+rtl8366rb_port_bridge_join(struct dsa_switch *ds, int port,
+			   struct net_device *bridge)
+{
+	struct realtek_smi *smi = ds->priv;
+	unsigned int port_bitmap = 0;
+	int ret, i;
+
+	/* Loop over all other ports than this one */
This sentence is a bit weird, how about: Loop over all ports but this one?
+	for (i = 0; i < RTL8366RB_PORT_NUM_CPU; i++) {
+		/* Handled last */
This comment is a bit misleading as it would suggest that the loop does act on 'port' when really it does that outside of the loop.

With those nitpicks fixed:

Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
-- 
Florian
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help