3.6.3.4. PRU_ICSSG Two Port Ethernet Switch

PRU_ICSSG can operate as VLAN aware Switch mode with two external physical ports and one internal host port. By default, interfaces come up in Dual independent EMAC mode and can be changed to operate in Switch mode at runtime. Note that changing from Dual EMAC to Switch mode needs loading of different firmwares to various PRU cores and thus have to follow specific sequence as shown in below sections:

3.6.3.4.1. Enabling Switch mode

Example assuming ETH1 and ETH2 as ICSSG0 interfaces:

ip link add name br0 type bridge
ip link set dev eth1 master br0
ip link set dev eth2 master br0
ip link set dev br0 up
bridge vlan add dev br0 vid 1 pvid untagged self

3.6.3.4.2. Going back to Dual EMAC mode

ip link set dev br0 down
ip link set dev eth1 nomaster
ip link set dev eth2 nomaster
ip link del name br0 type bridge

3.6.3.4.3. Turn On/Off Spanning Tree Protocol (STP)

ip link set dev br0 type bridge stp_state 1/0

3.6.3.4.4. VLAN configuration

bridge vlan add dev br0 vid 1 pvid untagged self <---- add cpu port to VLAN 1

This step is mandatory for bridge/default_pvid.

3.6.3.4.5. Adding extra VLANs

  1. untagged:

    bridge vlan add dev eth1 vid 100 pvid untagged master
bridge vlan add dev eth1 vid 100 pvid untagged master
bridge vlan add dev br0 vid 100 pvid untagged self <---- Add cpu port to VLAN100

  2. tagged:

    bridge vlan add dev eth1 vid 100 master
bridge vlan add dev eth1 vid 100 master
bridge vlan add dev br0 vid 100 pvid tagged self <---- Add cpu port to VLAN100

3.6.3.4.6. Forwarding Data Bases (FDBs)

Forwarding entries for MAC addresses are automatically added on the appropriate switch port upon detection as default operation as an unmanaged bridge. For managed bridge operation manually add FDB entries as required.

Manually adding FDBs:

bridge fdb add aa:bb:cc:dd:ee:ff dev eth1 master vlan 100
bridge fdb add aa:bb:cc:dd:ee:fe dev eth1 master <---- Add on all VLANs

3.6.3.4.7. Multicast Data Bases (MDBs)

Multicast entries are automatically added on the appropriate switch port upon detection as default operation as an unmanaged bridge. For managed bridge operation manually add MDB entries as required.

Manually adding MDBs:

bridge mdb add dev br0 port eth1 grp 239.1.1.1 permanent vid 100
bridge mdb add dev br0 port eth1 grp 239.1.1.1 permanent <---- Add on all VLANs

3.6.3.4.8. Multicast flooding

CPU port mcast_flooding is always on

Turning flooding on/off on switch ports:

bridge link set dev eth1 mcast_flood on/off

3.6.3.4.9. Limitations of switch mode offloading support

  • VLAN offloading is limited to 0-256 VIDs.

  • MDB/FDB static entries are limited to 511 entries and different FDBs can hash to same bucket and thus may not be completely offloaded

3.6.3.4.10. Cut Through Forwarding

Caution

The Cut Through configuration interface could be changed significantly in the future depending on Linux Kernel mainline development.

ICSSG Switch firmware now supports cut through forwarding. Cut Through feature allows forwarding packet from one external port to another without being stored in Port FIFOs thus reducing overall latency for packet forwarding.

Limitations

  1. Feature is not available in dual EMAC mode.

  2. Cut Through forwarding is only supported between external ports and no support for Cut Through to host port currently.

  3. Cut Through is not supported with 10/100 half-duplex.

  4. Cut Through is not supported with any form of flow control.

  5. If Intersperced Express Traffic (IET) is enabled, then Cut Through can only be enabled on an express priority queue and not on preemptible queues.

The devlink command to enable cut-through takes a u16 value. All 16 bits of it represent queues. BIT 0 to BIT 7 are slice0 queues where as BIT 8 to BIT 15 are slice1 queues. To enable cut-through forwarding in both directions, cut-through will need to be enabled on both slice0 and slice1.

In general, to enable cut through for slice0 qM, qN queues and slice1 qX, qY queues, pass the value \(2^M + 2^N + 2^{8+X} + 2^{8+Y}\) to the devlink command.

Assuming eth1 and eth2 are the active ports of ICSSG1 on AM64xx-EVM, to enable cut through run below commands,

ip link set eth1 down
ip link set eth2 down

# To enable cut thru for slice0 q0
devlink dev param set platform/icssg1-eth name cut_thru value 1 cmode runtime

# To enable cut thru for slice0 q1
devlink dev param set platform/icssg1-eth name cut_thru value 2 cmode runtime

# To enable cut thru for slice0 q0, q1
devlink dev param set platform/icssg1-eth name cut_thru value 3 cmode runtime

# To enable cut thru for slice0 q0, q1 and slice1 q0, q1
devlink dev param set platform/icssg1-eth name cut_thru value 771 cmode runtime

ip link set eth1 up
ip link set eth2 up

To show the current cut through status, run the below command,

devlink dev param show platform/icssg1-eth name cut_thru

platform/icssg1-eth:
 name cut_thru type driver-specific
   values:
     cmode runtime value 771

Here the value is shown as 771 (Binary 0000 0011 0000 0011). This means that cut through is enabled for q0, q1 of slice0 and q0, q1 of slice1. As bit 0,1,8 and 9 are set in the binary representation of the value 711.

To disable cut through on all queues

devlink dev param set platform/icssg1-eth name cut_thru value 0 cmode runtime

3.6.3.4.11. SRAM Requirement

  • The ICSSG Ethernet driver can be used in switch mode. Two icssg ports are used per switch.

  • SRAM Required for each switch (per two ports) is 256 KB.

  • For each ICSSG switch, the SRAM required needs to be contiguous.

  • PRUETH only uses the required amount of SRAM from the SRAM/MSMC pool. If PRUETH doesn’t get the required amount of SRAM, the prueth_probe() API will return with -ENOMEM error.