[PATCH v4] bmips: bcm6368-enetsw: Bump max MTU

Tue Oct 3 10:53:35 PDT 2023

The max frame size "jumbo" for this ethernet switch is 2048 bytes,
excluding the VLAN headers, which is

2048 - VLAN_ETH_HLEN - VLAN_HLEN = 2048 - 18 - 4 = 2026

not

ETH_DATA_LEN + ENETSW_TAG_SIZE = 1500 + 4 + 6 = 1510

as it is right now. The available overhead is needed when using the DSA
switch with a cascaded Marvell DSA switch, which is something that exist
in real products, in this case the Inteno XG6846.

We need to adjust the RX buffer size to fit the new max frame size.

Before this patch (on the lan1 DSA port in this case):
dsa_slave_change_mtu: master->max_mtu = 9724, dev->max_mtu = 10218, DSA overhead = 8
dsa_slave_change_mtu: master = extsw, dev = lan1
dsa_slave_change_mtu: master->max_mtu = 1510, dev->max_mtu = 9724, DSA overhead = 6
dsa_slave_change_mtu: master = eth0, dev = extsw
dsa_slave_change_mtu new_master_mtu 1514 > mtu_limit 1510
mdio_mux-0.1:00: nonfatal error -34 setting MTU to 1500 on port 0

My added debug prints before the nonfatal error: the first switch from the top
is the Marvell switch, the second in the bcm6368-enetsw with its 1510 limit.

After this patch the error is gone.

OpenWrt adds a VLAN to each port so we get VLAN tags on all frames.

Testing the new 2026 MTU:

    eth0             ext1              enp7s0
 .--------.     .-----------.         .------.
 | enetsw | <-> | mv88e8152 | <-----> | host |
 `--------´     `-----------´         `------´

On the router:

Notice that the max MTU for the ext1 switch (mv88e8152)needs to be 18
bytes less than the eth0 (enetsw). That's the VLAN overhead since OpenWrt
sets up one VLAN per external port.

On the host we just set the max MTU for test:

Test pininging from host:
PING 192.168.1.220 (192.168.1.220) 1984(2012) bytes of data.
1992 bytes from 192.168.1.220: icmp_seq=1 ttl=64 time=0.687 ms
1992 bytes from 192.168.1.220: icmp_seq=2 ttl=64 time=0.670 ms

Test pinging from router:
PING 192.168.1.2 (192.168.1.2): 1984 data bytes
1992 bytes from 192.168.1.2: seq=0 ttl=64 time=1.021 ms
1992 bytes from 192.168.1.2: seq=1 ttl=64 time=0.909 ms

The max IP packet without headers is 1984, the outgoing ICMP packet is
1992 bytes, then 32 bytes are added up to 2048 as described below.

Ping packet on router side as viewed with tcpdump:

00:15:13.516325 AF Unknown (1429722180), length 2044:
        0x0000:  3d93 bcae c56b a83d 8874 03fd 0004 8100  =....k.=.t......
        0x0010:  0000 dada 0000 c020 0fff 0800 4500 07dc  ............E...
        0x0020:  0000 4000 4001 aef2 c0a8 0102 c0a8 01dc  .. at .@...........
        0x0030:  0800 43b6 0092 0001 9c4c 1c65 0000 0000  ..C......L.e....
        0x0040:  2963 0400 0000 0000 1011 1213 1415 1617  )c..............
        0x0050:  1819 1a1b 1c1d 1e1f 2021 2223 2425 2627  .........!"#$%&'

3d93 = First four bytes are the last two bytes of the destination
ethernet address I don't know why the first four are missing,
but it sure explains why the paket is 2044 bytes and not 2048
which is the actual max MTU.

bcae c56b a83b = source ethernet address

8874 03fd 0004 = Broadcom enetsw DSA tag

8100 0000 = VLAN 802.1Q header

dada 0000 c020 0fff 0800 = EDSA tag for the Marvell (outer) switch,
 0800 is the ethertype obviously included in the tag

Next follows the contents of the ping packet as it appears if
we dump it on the DSA interface such as tcpdump -i lan1
etc, there we get the stripped out packet.

At the end 4 bytes of FCS.

Cc: Álvaro Fernández Rojas <noltari at gmail.com>
Cc: Jonas Gorski <jonas.gorski at gmail.com>
Signed-off-by: Linus Walleij <linus.walleij at linaro.org>
---
ChangeLog v3->v4:
- Adjust the RX buffer size and we can use the max "jumbo"
  frame size 2048.
ChangeLog v2->v3:
- Make a more believable case for the max MTU with tcpdump
  example.
ChangeLog v1->v2:
- Do some better research after help on IRC, did some ping tests.
---
 .../files/drivers/net/ethernet/broadcom/bcm6368-enetsw.c  | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/target/linux/bmips/files/drivers/net/ethernet/broadcom/bcm6368-enetsw.c b/target/linux/bmips/files/drivers/net/ethernet/broadcom/bcm6368-enetsw.c
index 321e95dbbb3d..879089a4a3c4 100644
--- a/target/linux/bmips/files/drivers/net/ethernet/broadcom/bcm6368-enetsw.c
+++ b/target/linux/bmips/files/drivers/net/ethernet/broadcom/bcm6368-enetsw.c
@@ -22,7 +22,9 @@
 #include <linux/reset.h>
 #include <linux/version.h>
 
-/* MTU */
+/* The max frame size is 2048 including VLAN headers but excluding the 6 byte DSA tag */
+#define ENETSW_MAX_FRAME		2048
+#define ENETSW_MAX_MTU			(ENETSW_MAX_FRAME - VLAN_ETH_HLEN - VLAN_HLEN)
 #define ENETSW_TAG_SIZE			(6 + VLAN_HLEN)
 #define ENETSW_MTU_OVERHEAD		(VLAN_ETH_HLEN + VLAN_HLEN + \
 					 ENETSW_TAG_SIZE)
@@ -1006,7 +1008,7 @@ static int bcm6368_enetsw_probe(struct platform_device *pdev)
 		dev_info(dev, "random mac\n");
 	}
 
-	priv->rx_buf_size = ALIGN(ndev->mtu + ENETSW_MTU_OVERHEAD,
+	priv->rx_buf_size = ALIGN(ENETSW_MAX_FRAME,
 				  ENETSW_DMA_MAXBURST * 4);
 
 	priv->rx_frag_size = ENETSW_FRAG_SIZE(priv->rx_buf_size);
@@ -1067,7 +1069,7 @@ static int bcm6368_enetsw_probe(struct platform_device *pdev)
 	ndev->netdev_ops = &bcm6368_enetsw_ops;
 	ndev->min_mtu = ETH_ZLEN;
 	ndev->mtu = ETH_DATA_LEN + ENETSW_TAG_SIZE;
-	ndev->max_mtu = ETH_DATA_LEN + ENETSW_TAG_SIZE;
+	ndev->max_mtu = ENETSW_MAX_MTU;
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(6,1,0)
 	netif_napi_add(ndev, &priv->napi, bcm6368_enetsw_poll);
 #else
-- 
2.34.1


