PRU_ICSSG Ethernet Introduction

Driver is developed complying to Linux Driver model and implements netdev and NAPI APIs along with other basic driver functions. Driver supports multiple instances of ICSSG each of which has two slices. Each slice consists of a pair of Programmable Real-Time Units (PRUs), Auxillary Programmable Real-time Units (RTUs) and TX Programmable Real-Time Units (TX_PRUS). So each ICSSG instance supports two Ethernet interfaces i.e. one per slice.

One of the key difference in the driver implementation compared to PRU-ICSS driver is the use of UDMA driver interface and ring accelerator driver available on Keystone III SoC to send/receive frames between ARM Host processor and PRU_ICSSG. This allows for efficient moving of data and is more efficient than the shared memory transport used in PRU-ICSS. Supported platforms

SoC Number of external ports
AM65X 3 x ICSSG, 6 Ports max
AM64X 2 x ICSSG, 4 Ports max Features supported

  • 1G/100M/10M Full-Duplex Ethernet ports
  • 100M/10M Half-Duplex Ethernet ports
  • RGMII mode with TX delay (configured in DTS). RX delay is not supported and has to be provided by PHY.
  • MII mode
  • Multiple TX queues (upto 4), single RX queue.
  • PTP Ordinary clock
  • PPS Out
  • Promiscuous mode

Features not supported

  • VLAN HW filtering
  • All-multi mode is always enabled
  • Multi-cast HW filtering
  • Different MII modes for Real-Time Ethernet ports (MII_G_RT1 and MII_G_RT2). For example, MII_G_RT1=MII and MII_G_RT1=RGMII


Both MIIx ports have to be enabled in DT even if one of them is not used (no Ethernet PHY wired) for proper PRU_ICSSG Ethernet driver work. Use fixed-link for unused port as workaround. Driver Configuration

The TI Processor SDK has ICSSG driver enabled by default on supported platforms. In case of custom builds, please ensure following configs are enabled.


Module Build

Module build for the ICSSG driver is supported. To do this, use option ‘m’ for above configs, where applicable. User guide Bringing Up interface

The network interface can be configured automatically depending on root file system or configured manually. Manual configuration:

ip addr add dev eth1
ip link set dev eth1 up

< or >

ifconfig eth1 <ip> netmask <mask> up Get information (ethtool) Get driver information

The interface can be identified by using ethtool -i|--driver DEVNAME command. It also provides some information about supported features.

~# ethtool -i eth1
driver: icssg-prueth
bus-info: pruss2_eth
supports-statistics: yes
supports-test: no
supports-eeprom-access: no
supports-register-dump: no
supports-priv-flags: no Display time stamping capabilities

The interface time stamping capabilities can be retrieved by using ethtool -T|--show-time-stamping DEVNAME command.

ethtool -T eth2
Time stamping parameters for eth2:
     hardware-transmit     (SOF_TIMESTAMPING_TX_HARDWARE)
     software-transmit     (SOF_TIMESTAMPING_TX_SOFTWARE)
     hardware-receive      (SOF_TIMESTAMPING_RX_HARDWARE)
     software-receive      (SOF_TIMESTAMPING_RX_SOFTWARE)
     software-system-clock (SOF_TIMESTAMPING_SOFTWARE)
     hardware-raw-clock    (SOF_TIMESTAMPING_RAW_HARDWARE)
PTP Hardware Clock: 2
Hardware Transmit Timestamp Modes:
     off                   (HWTSTAMP_TX_OFF)
     on                    (HWTSTAMP_TX_ON)
Hardware Receive Filter Modes:
     none                  (HWTSTAMP_FILTER_NONE)
     all                   (HWTSTAMP_FILTER_ALL) Show permanent hardware address

The interface permanent hardware address can be retrieved by using ethtool -P|--show-permaddr DEVNAME command.

~# ethtool -P eth1
Permanent address: 70:ff:76:1d:5c:64 Query Channels information

The interface DMA Channels information can be retrieved by using ethtool-l|--show-channels DEVNAME command.

# ethtool -l eth1
Channel parameters for eth1:
Pre-set maximums:
RX:             1
TX:             4
Other:          0
Combined:       0
Current hardware settings:
RX:             1
TX:             1
Other:          0
Combined:       0 Show adapter statistics

The interface statistics can be retrieved by using ethtool -S|--statistics DEVNAME command. It displays statistic for the ethernet port.

# ethtool -S eth1
NIC statistics:
     rx_good_frames: 53
     rx_broadcast_frames: 1
     rx_multicast_frames: 53
     rx_crc_error_frames: 0
     rx_mii_error_frames: 0
     rx_odd_nibble_frames: 0
     rx_frame_max_size: 2000
     rx_max_size_error_frames: 0
     rx_frame_min_size: 64
     rx_min_size_error_frames: 11
     rx_overrun_frames: 0
     rx_class0_hits: 64
     rx_class1_hits: 0
     rx_class2_hits: 0
     rx_class3_hits: 0
     rx_class4_hits: 0
     rx_class5_hits: 0
     rx_class6_hits: 0
     rx_class7_hits: 0
     rx_class8_hits: 0
     rx_class9_hits: 0
     rx_class10_hits: 0
     rx_class11_hits: 0
     rx_class12_hits: 0
     rx_class13_hits: 0
     rx_class14_hits: 0
     rx_class15_hits: 0
     rx_smd_frags: 0
     rx_bucket1_size: 64
     rx_bucket2_size: 128
     rx_bucket3_size: 256
     rx_bucket4_size: 512
     rx_64B_frames: 2
     rx_bucket1_frames: 13
     rx_bucket2_frames: 30
     rx_bucket3_frames: 20
     rx_bucket4_frames: 1
     rx_bucket5_frames: 0
     rx_total_bytes: 7864
     rx_tx_total_bytes: 24165
     tx_good_frames: 98
     tx_broadcast_frames: 0
     tx_multicast_frames: 98
     tx_odd_nibble_frames: 0
     tx_underflow_errors: 0
     tx_frame_max_size: 2000
     tx_max_size_error_frames: 0
     tx_frame_min_size: 64
     tx_min_size_error_frames: 0
     tx_bucket1_size: 64
     tx_bucket2_size: 128
     tx_bucket3_size: 256
     tx_bucket4_size: 512
     tx_64B_frames: 0
     tx_bucket1_frames: 0
     tx_bucket2_frames: 68
     tx_bucket3_frames: 21
     tx_bucket4_frames: 9
     tx_bucket5_frames: 0
     tx_total_bytes: 12479 Show EEE settings

The interface EEE settings can be retrieved by using ethtool --show-eee DEVNAME command.

ethtool --show-eee eth1
EEE Settings for eth1:
   EEE status: disabled
   Tx LPI: disabled
   Supported EEE link modes:  100baseT/Full
   Advertised EEE link modes:  Not reported
   Link partner advertised EEE link modes:  100baseT/Full
                   1000baseT/Full VLAN Config

VLAN can be added/deleted using ip or vconfig utility.


ip link add link eth1 name eth1.5 type vlan id 5

< or >

vconfig add eth1 5

VLAN del

ip link del eth1.5

< or >

vconfig rem eth1 5

VLAN IP assigning

IP address can be assigned to the VLAN interface either via udhcpc when a VLAN aware dhcp server is present or via static ip assigning using ip or ifconfig.

Once VLAN is added, it will create a new entry in Ethernet interfaces like eth1.5, below is an example how it check the vlan interface

ip addr add dev eth1.5

< or >

ifconfig eth1.5

~# ifconfig eth1.5
eth1.5    Link encap:Ethernet  HWaddr 70:FF:76:1D:5C:64
          inet addr:  Bcast:  Mask:
          inet6 addr: fe80::72ff:76ff:fe1d:5c64/64 Scope:Link
          RX packets:0 errors:0 dropped:0 overruns:0 frame:0
          TX packets:45 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:0 (0.0 B)  TX bytes:7590 (7.4 KiB)

VLAN Packet Send/Receive

To Send or receive packets with the VLAN tag, bind the socket to the proper Ethernet interface shown above and can send/receive via that socket-fd. Multicast Add/Delete

Multicast MAC address can be added/deleted using ip maddr commands or Linux socket ioctl SIOCADDMULTI/SIOCDELMULTI.

Show muliticast address

~# ip maddr show eth1
3:      eth1
        link  33:33:00:00:00:01 users 2
        link  01:00:5e:00:00:01 users 2
        link  01:00:5e:00:00:fb users 2
        link  33:33:ff:1d:5c:64 users 2
        link  01:00:5e:00:00:fc users 2
        link  33:33:00:01:00:03 users 2
        link  33:33:00:00:00:fb users 2
        link  01:80:c2:00:00:21 users 2
        inet6 ff02::fb
        inet6 ff02::1:3
        inet6 ff02::1:ff1d:5c64
        inet6 ff02::1
        inet6 ff01::1

Add muliticast address

~# ip maddr add 01:00:5e:00:00:05 dev eth1
~# ip maddr show dev eth1
3:      eth1
        link  33:33:00:00:00:01 users 2
        link  01:00:5e:00:00:01 users 2
        link  01:00:5e:00:00:fb users 2
        link  33:33:ff:1d:5c:64 users 2
        link  01:00:5e:00:00:fc users 2
        link  33:33:00:01:00:03 users 2
        link  33:33:00:00:00:fb users 2
        link  01:80:c2:00:00:21 users 2
        link  01:00:5e:00:00:05 static
        inet6 ff02::fb
        inet6 ff02::1:3
        inet6 ff02::1:ff1d:5c64
        inet6 ff02::1
        inet6 ff01::1

Delete muliticast address

# ip maddr del 01:00:5e:00:00:05 dev eth1 Promiscous Mode

By default promiscous mode is disabled. It can be enabled by using the below command.

Please note running a tool like tcpdump will itself enable promiscous mode.

ip link set eth0 promisc on Configure interface (ethtool) Change generic options

The interface generic options can be configured by using ethtool -s|--change DEVNAME command. The main purpose of this command is to configure physical link settings (PHY) like speed, duplex, auto-negotiation.

Below commands will be redirected to the phy driver:

# ethtool -s <dev>
[ speed %d ]
[ duplex half|full ]
[ autoneg on|off ]
[ wol p|u|m|b|a|g|s|d... ]
[ sopass %x:%x:%x:%x:%x:%x ]


ICSSG Ethernet driver does not perform any kind of WOL specific actions or configurations.

Below is an example of forcing link speed to 100M and duplexity to full:

# ethtool -s eth1 duplex full speed 100
[   74.768324] icssg-prueth pruss2_eth eth1: Link is Down
[   78.592924] icssg-prueth pruss2_eth eth1: Link is Up - 100Mbps/Full - flow control off Restart N-WAY (PHY) negotiation

The interface PHY auto-negotiation can be restarted by using ethtool -r|--negotiate DEVNAME command.

# ethtool -r eth1
[  273.151655] icssg-prueth pruss2_eth eth1: Link is Down
[  276.225423] icssg-prueth pruss2_eth eth1: Link is Up - 1Gbps/Full - flow control off Set Channels parameters

The interface DMA channels parameters can be set by using ethtool -L\|--set-channels DEVNAME command. It allows to control number of TX channels driver is allowed to work with at DMA level. The maximum number of TX channels is 4. Supported options [ tx N ]:

# ethtool -L eth1 tx 4 PTP Ordinary Clock

The PRU Ethernet & IEP drivers implement the Linux PTP hardware clock subsystem APIs, the PRU-ICSS PTP clock can therefore be adjusted by using those standard APIs. See PTP hardware clock infrastructure for Linux for more details.

The IEP0 is used by PRU Ethernet driver and Firmware PTP hardware clock and shared between PRU Ethernet ports. The IEP1 is used for Firmware purposes.

The PTP Ordinary Clock (OC) implementation is provided by the linuxptp application.

ptp4l -f oc.cfg

oc.cfg is a ptp4l configuration file.

Example oc.cfg for OC,

tx_timestamp_timeout 10
logMinPdelayReqInterval -3
logSyncInterval -3
twoStepFlag 1
summary_interval 0
delay_mechanism P2P
network_transport L2

where eth1 is the intended PRU-ICSSG Ethernet port over which the OC functionality is provided.

See The Linux PTP Project for more details about linuxptp in general and ptp4l(8) - Linux man page about ptp4l configurations in particular.

Here is a sample screen display of ptp4l for PRU-ICSS Ethernet port as PTP/OC in slave mode:

# ptp4l -f oc.cfg -s -m
ptp4l[1255.613]: selected /dev/ptp2 as PTP clock
ptp4l[1255.664]: port 1: INITIALIZING to LISTENING on INITIALIZE
ptp4l[1255.665]: port 0: INITIALIZING to LISTENING on INITIALIZE
ptp4l[1255.665]: port 1: link up
ptp4l[1263.081]: selected best master clock 70ff76.fffe.1d5c64
ptp4l[1269.343]: selected best master clock 70ff76.fffe.1d5c64
ptp4l[1271.367]: port 1: new foreign master d494a1.fffe.8c36e9-1
ptp4l[1275.368]: selected best master clock d494a1.fffe.8c36e9
ptp4l[1275.368]: port 1: LISTENING to UNCALIBRATED on RS_SLAVE
ptp4l[1276.381]: rms 789386424832367360 max 1578772849664738816 freq -60377 +/- 22862 delay   229 +/-   6
ptp4l[1277.385]: rms  473 max  729 freq -67059 +/- 642 delay   251 +/-   4
ptp4l[1278.389]: rms  792 max  830 freq -65620 +/- 211 delay   253 +/-   0
ptp4l[1279.393]: rms  504 max  667 freq -65335 +/-  17 delay   255 +/-   1
ptp4l[1280.397]: rms  166 max  271 freq -65484 +/-  59 delay   251 +/-   2
ptp4l[1281.401]: rms   26 max   42 freq -65649 +/-  34 delay   249 +/-   1
ptp4l[1282.405]: rms   43 max   50 freq -65727 +/-  10 delay   253 +/-   3
ptp4l[1283.409]: rms   26 max   39 freq -65739 +/-   6 delay   256 +/-   1
ptp4l[1284.412]: rms    5 max    7 freq -65725 +/-   3 delay   253 +/-   1
ptp4l[1285.416]: rms    5 max    7 freq -65717 +/-   6 delay   252 +/-   1
ptp4l[1286.420]: rms   11 max   14 freq -65698 +/-   6 delay   252 +/-   1
ptp4l[1287.424]: rms    8 max   12 freq -65693 +/-   5 delay   254 +/-   1
ptp4l[1288.427]: rms    7 max   12 freq -65687 +/-   4 delay   251 +/-   2
ptp4l[1289.430]: rms    4 max    8 freq -65686 +/-   3 delay   249 +/-   1
ptp4l[1290.434]: rms    5 max    8 freq -65693 +/-   7 delay   249 +/-   1
ptp4l[1291.438]: rms    4 max    9 freq -65696 +/-   5 delay   251 +/-   1
ptp4l[1292.441]: rms    7 max    9 freq -65682 +/-   5 delay   253 +/-   0
ptp4l[1293.445]: rms   11 max   14 freq -65667 +/-   4 delay   252 +/-   0
ptp4l[1294.448]: rms    8 max   14 freq -65662 +/-   5 delay   254 +/-   1
ptp4l[1295.452]: rms    6 max    8 freq -65659 +/-   5 delay   254 +/-   2
ptp4l[1296.456]: rms    3 max    7 freq -65657 +/-   2 delay   251 +/-   0
ptp4l[1297.459]: rms    4 max    5 freq -65661 +/-   6 delay   256 +/-   2
... PPS Pulse Per Second support

PPS hardware pin is available only on the IDK application card i.e. ICSSG0 port 0 and ICSSG1 port 1. They are available at LEDs LD2 and LD5 respectively.

PPS can be tested using testptp.c tool.

To find out the PTP device number i.e. PTP Hardware Clock, use ethtool -T DEVNAME


For PPS to work, the firmware needs to be running so the ICSSG network interface must be brought up.

To turn on PPS,

# ip link set dev eth1 up
# ./testptp -d /dev/ptp2 -P 1
pps for system time request okay

You should be able to see either LD2 or LD5 blink at 1 second interval.

To turn off PPS,

# ./testptp -d /dev/ptp2 -P 0
pps for system time request okay Tips Ethernet PHYs/MDIO bindings

The PRU_ICSSG Ethernet driver follows standard Linux DT bindings for MDIO bus, Ethernet controlers and PHYs which can be found at:

The existing TI Ethernet PHYs DT bindings: 100M/10M Half-Duplex

The 10/100 half duplex (HD) support depends on HW capability to route PHY output pin (COL) to ICSSG GPI1O pin (PRGx_PRU0/1_GPI10) as input. To indicate that HW supports HD the DT “ti,half-duplex-capable” property shell be added to the corresponding ICSSG “ethernet-mii0” port node.

For example:

icssg0_eth: icssg0-eth {

   icssg0_emac1: ethernet-mii1 {


&main_pmx0 {

   icssg0_rgmii_pins_default: icssg0-rgmii-pins-default {
         pinctrl-single,pins = <
                AM65X_IOPAD(0x026c, PIN_INPUT, 1) /* (AA28) PRG0_PRU1_GPO10.PRG0_PRU1_GPI10 - col */
}; MII Support


On AM64x-evm the DP83869HM are strapped to RGMII configuration by default. To use MII interface the k3-am642-evm-icssg1-dualemac-mii.dtbo overlay file has to be applied using the following command in uboot.

setenv bootcmd 'run findfdt; run envboot; run init_${boot}; run get_kern_${boot}; run get_fdt_${boot}; setenv name_overlays k3-am642-evm-icssg1-dualemac-mii.dtbo; run get_overlay_${boot}; run run_kern'