EXA8 Network
| lan1 | lan3 | lan5 | lan7 |
|---|---|---|---|
| lan2 | lan4 | lan6 | lan8 |
The Cubro EXA8 ships with the flagship sub-millisecond passive Bridge, vertically linking all interface pairs by default.
NOTE: When running in BYPASS mode, no packets will be visible by the integrated platform!
Unless otherwise configured, the EXA8 will bridge passively its G1-8 interfaces vertically stacked
The interface for each pair are controlled by dedicated GPIO pins
ge5 ~ ge6 (pin 32);
ge7 ~ ge8 (pin 33);
ge1 ~ ge2 (pin 34);
ge3 ~ ge4 (pin 35);
gpioset <gpio_pin> set 0
gpioset <gpio_pin> set 1
gpioset 32 set 1
gpioset 33 set 1
gpioset 34 set 1
gpioset 35 set 1
When BYPASS mode is disabled, the integrated system can be configured to handle packets.
The following examples illustrates a simple bridge between interfaces lan1 and lan2 with sniffing capabilities
NOTE: Any interface and GPIO changes should be applied on each boot to avoid bypass default mode!
The following chapter requires the legacy kernel modules for the EXA8 available here
cp /lib/modules/4.14.11-svn258/kernel/drivers/net/dsa/mv88e6xxx/mv88e6xxx.ko /root/mv88e6xxx.ko.backup
cp mv88e6xxx.ko /lib/modules/4.14.11-svn258/kernel/drivers/net/dsa/mv88e6xxx/mv88e6xxx.ko
The interface enp5s0f1 is the logical interface to the CPU. It needs to be UP to enable any traffic flow on lan IFs
ifconfig enp5s0f1 up
ip link add veth0 type veth peer name veth1
brctl addbr br0
brctl addbr br1
brctl addif br0 lan1
brctl addif br0 veth0
brctl addif br1 lan2
brctl addif br1 veth1
ifconfig veth0 up
ifconfig veth1 up
ifconfig lan1 up
ifconfig lan2 up
ifconfig br0 up
ifconfig br1 up
- Thanks to @pstavirs for this contribution
Dump the current value of the uboot fdt_update variable:
$ sudo fw_printenv | grep fdt_update
fdt_update=fdt adds ${fdtcontroladdr};fdt resize 4096;fdt set /soc/pci/mrml-bridge0/mdio-nexus/mdio1 phandle <0x4e>;fdt set /soc/pci/mrml-bridge0/mdio-nexus/mdio1/switch0 status disabled;fdt mknode / switch;fdt set /switch compatible marvell,mvmdio-uio;fdt set /switch status okay;fdt set /switch mii-bus <0x4e>
The above output might be different. The important thing to note is the /soc/pci/mrml-bridge0/mdio-nexus/mdio1/switch0 which is the device that the kernel has a driver for is set to status disabled and /switch (Cubro proprietary driver) is set to status okay. We need to swap the status for those two devices:
To do that, update that variable -
$ sudo fw_setenv fdt_update fdt adds ${fdtcontroladdr};fdt resize 4096;fdt set /soc/pci/mrml-bridge0/mdio-nexus/mdio1 phandle <0x4e>;fdt set /soc/pci/mrml-bridge0/mdio-nexus/mdio1/switch0 status okay;fdt mknode / switch;fdt set /switch compatible marvell,mvmdio-uio;fdt set /switch status disabled;fdt set /switch mii-bus <0x4e>
Reboot the device and use sudo fw_printenv | grep fdt_update to verify that the change is reflected.
Use ifconfig -a to verify that the lan[1-8] devices are created. If they are down, use ifconfig lan[1-8] up to enable them. You will also need to enable the CPU port ifconfig enp5s0f1 up for the lan[1-8] devices to come up.