.. SPDX-License-Identifier: BSD-3-Clause Copyright(c) 2011-2019 Intel Corporation ======= QoS API ======= Description =========== The generic API for the Quality of Service (QoS) Traffic Management of Ethernet devices, which includes the following main features: hierarchical scheduling, traffic shaping, congestion management, packet marking. The TM setting commands are as following: 1. Add the port traffic management private shaper profile:: testpmd> add port tm node shaper profile (port_id) (shaper_profile_id) \ (cmit_tb_rate) (cmit_tb_size) (peak_tb_rate) (peak_tb_size) \ (packet_length_adjust) 2. Add nonleaf node to port traffic management hiearchy:: testpmd> add port tm nonleaf node (port_id) (node_id) (parent_node_id) \ (priority) (weight) (level_id) (shaper_profile_id) \ (n_sp_priorities) (stats_mask) (n_shared_shapers) \ [(shared_shaper_0) (shared_shaper_1) ...] \ 3. Add leaf node to port traffic management hiearchy:: testpmd> add port tm leaf node (port_id) (node_id) (parent_node_id) \ (priority) (weight) (level_id) (shaper_profile_id) \ (cman_mode) (wred_profile_id) (stats_mask) (n_shared_shapers) \ [(shared_shaper_id) (shared_shaper_id) ...] \ 4. Commit the traffic management hierarchy on the port:: testpmd> port tm hierarchy commit (port_id) (clean_on_fail) Prerequisites ============= For i40e, need enable rss For ixgbe, need disable rss. The DUT must have two 10G Ethernet ports connected to two ports on tester:: dut_port_0 <---> tester_port_0 dut_port_1 <---> tester_port_1 Assume two DUT 10G Ethernet ports' pci device id is as the following:: dut_port_0 : "0000:05:00.0" dut_port_1 : "0000:05:00.1" Bind two ports to dpdk driver:: ./usertools/dpdk-devbind.py -b igb_uio 05:00.0 05:00.1 Test Cases for I40e: ==================== Test Case: dcb 4 tc queue mapping ================================= 1. Start testpmd and set DCB:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 23-27 -n 4 --master-lcore=23 -- -i --nb-cores=4 --rxq=4 --txq=4 --rss-ip testpmd> port stop all testpmd> port config 0 dcb vt off 4 pfc off testpmd> port config 1 dcb vt off 4 pfc off testpmd> port start all testpmd> set fwd rxonly testpmd> set verbose 1 testpmd> start 2. Send packets with set vlan user priority 0, 1, 2, 3, verify:: pkt1 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=0)/IP()/Raw('x'*20) pkt2 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=1)/IP()/Raw('x'*20) pkt3 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=2)/IP()/Raw('x'*20) pkt4 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=3)/IP()/Raw('x'*20) packet with vlan user priority 0 should be received by queue 0 packet with vlan user priority 1 should be received by queue 1 packet with vlan user priority 2 should be received by queue 2 packet with vlan user priority 3 should be received by queue 3 Test Case: dcb 8 tc queue mapping ================================= 1. Start testpmd and set DCB:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 23-31 -n 4 --master-lcore=23 -- -i --nb-cores=8 --rxq=8 --txq=8 --rss-ip testpmd> port stop all testpmd> port config 0 dcb vt off 8 pfc off testpmd> port config 1 dcb vt off 8 pfc off testpmd> port start all testpmd> set fwd rxonly testpmd> set verbose 1 testpmd> start 2. Send packet with set vlan user priority 0, 1, 2, 3, 4, 5, 6, 7, verify:: pkt1 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=0)/IP()/Raw('x'*20) pkt2 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=1)/IP()/Raw('x'*20) pkt3 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=2)/IP()/Raw('x'*20) pkt4 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=3)/IP()/Raw('x'*20) pkt5 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=4)/IP()/Raw('x'*20) pkt6 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=5)/IP()/Raw('x'*20) pkt7 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=6)/IP()/Raw('x'*20) pkt8 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=7)/IP()/Raw('x'*20) packet with vlan user priority 0 should be received by queue 0 packet with vlan user priority 1 should be received by queue 1 packet with vlan user priority 2 should be received by queue 2 packet with vlan user priority 3 should be received by queue 3 packet with vlan user priority 4 should be received by queue 4 packet with vlan user priority 5 should be received by queue 5 packet with vlan user priority 6 should be received by queue 6 packet with vlan user priority 7 should be received by queue 7 Test Case: shaping 1 port 4 tc ============================== 1. Start testpmd and set DCB:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 23-27 -n 4 --master-lcore=23 -- -i --nb-cores=4 --rxq=4 --txq=4 --rss-ip testpmd> port stop all testpmd> port config 0 dcb vt off 4 pfc off testpmd> port config 1 dcb vt off 4 pfc off 2. Add root non leaf node:: testpmd> add port tm nonleaf node 1 1000000 -1 0 1 0 -1 1 0 0 3. Add private shaper 1, 2, 3, 4 for tc node 0-4:: testpmd> add port tm node shaper profile 1 1 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 2 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 3 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 4 0 0 25000000 0 0 4. Add tc node:: testpmd> add port tm nonleaf node 1 900000 1000000 0 1 1 1 1 0 0 testpmd> add port tm nonleaf node 1 900001 1000000 0 1 1 2 1 0 0 testpmd> add port tm nonleaf node 1 900002 1000000 0 1 1 3 1 0 0 testpmd> add port tm nonleaf node 1 900003 1000000 0 1 1 4 1 0 0 5. Add queue leaf node:: testpmd> add port tm leaf node 1 0 900000 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 1 900001 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 2 900002 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 3 900003 0 1 2 -1 0 0xffffffff 0 0 6. Commit the traffic management hierarchy on the port:: testpmd> port tm hierarchy commit 1 no testpmd> port start all testpmd> start 7. Send four flows with user priority 0, 1, 2, 3, verify shaping each flow to 200Mbps(25MBps). Test Case: shaping 1 port 8 tc =============================== 1. Start testpmd and set DCB:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 23-31 -n 4 --master-lcore=23 -- -i --nb-cores=8 --rxq=8 --txq=8 --rss-ip testpmd> port stop all testpmd> port config 0 dcb vt off 8 pfc off testpmd> port config 1 dcb vt off 8 pfc off 2. Add root non leaf node:: testpmd> add port tm nonleaf node 1 1000000 -1 0 1 0 -1 1 0 0 3. Add private shaper 1, 2, 3, 4, 5, 6, 7, 8, for tc node 0-7:: testpmd> add port tm node shaper profile 1 1 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 2 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 3 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 4 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 5 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 6 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 7 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 8 0 0 25000000 0 0 4. Add tc node:: testpmd> add port tm nonleaf node 1 900000 1000000 0 1 1 1 1 0 0 testpmd> add port tm nonleaf node 1 900001 1000000 0 1 1 2 1 0 0 testpmd> add port tm nonleaf node 1 900002 1000000 0 1 1 3 1 0 0 testpmd> add port tm nonleaf node 1 900003 1000000 0 1 1 4 1 0 0 testpmd> add port tm nonleaf node 1 900004 1000000 0 1 1 5 1 0 0 testpmd> add port tm nonleaf node 1 900005 1000000 0 1 1 6 1 0 0 testpmd> add port tm nonleaf node 1 900006 1000000 0 1 1 7 1 0 0 testpmd> add port tm nonleaf node 1 900007 1000000 0 1 1 8 1 0 0 5. Add queue leaf node:: testpmd> add port tm leaf node 1 0 900000 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 1 900001 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 2 900002 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 3 900003 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 4 900004 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 5 900005 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 6 900006 0 1 2 -1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 7 900007 0 1 2 -1 0 0xffffffff 0 0 6. Commit the traffic management hierarchy on the port:: testpmd> port tm hierarchy commit 1 no testpmd> port start all testpmd> start 7. Send four flows with user priority 0, 1, 2, 3, 4, 5, 6, 7, verify shaping each flow to 200Mbps(25MBps). Test Case: shaping for port =========================== 1. Start testpmd:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 23-27 -n 4 --master-lcore=23 -- -i --nb-cores=4 --rxq=4 --txq=4 --rss-ip testpmd> port stop 1 1. Add private shaper 0:: testpmd> add port tm node shaper profile 1 0 0 0 25000000 0 0 2. Add port (root nonleaf) node:: testpmd> add port tm nonleaf node 1 1000000 -1 0 1 0 0 1 0 0 3. Commit the traffic management hierarchy on the port:: testpmd> port tm hierarchy commit 1 no testpmd> port start 1 testpmd> start 4. Start transmitting, verify shaping the traffic to 200Mbps(25MBps). Test Cases for ixgbe: ===================== Test Case: dcb 4 tc queue mapping ================================= 1. Start testpmd and set DCB:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 3-7 -n 4 --master-lcore=3 -- -i --nb-cores=4 --rxq=4 --txq=4 --disable-rss testpmd> vlan set filter off 0 testpmd> vlan set filter off 1 testpmd> port stop all testpmd> port config 0 dcb vt off 4 pfc off testpmd> port config 1 dcb vt off 4 pfc off testpmd> port start all testpmd> set fwd rxonly testpmd> set verbose 1 testpmd> start 2. Send packets with set vlan user priority 0, 1, 2, 3, verify:: pkt1 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=0)/IP()/Raw('x'*20) pkt2 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=1)/IP()/Raw('x'*20) pkt3 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=2)/IP()/Raw('x'*20) pkt4 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=3)/IP()/Raw('x'*20) packet with vlan user priority 0 should be received by queue 0 packet with vlan user priority 1 should be received by queue 32 packet with vlan user priority 2 should be received by queue 64 packet with vlan user priority 3 should be received by queue 96 Test Case: dcb 8 tc queue mapping ================================= 1. Start testpmd and set DCB:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 3-11 -n 4 --master-lcore=3 -- -i --nb-cores=8 --rxq=8 --txq=8 --disable-rss testpmd> vlan set filter off 0 testpmd> vlan set filter off 1 testpmd> port stop all testpmd> port config 0 dcb vt off 8 pfc off testpmd> port config 1 dcb vt off 8 pfc off testpmd> port start all testpmd> set fwd rxonly testpmd> set verbose 1 testpmd> start 2. Send packet with set vlan user priority 0, 1, 2, 3, 4, 5, 6, 7, verify:: pkt1 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=0)/IP()/Raw('x'*20) pkt2 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=1)/IP()/Raw('x'*20) pkt3 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=2)/IP()/Raw('x'*20) pkt4 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=3)/IP()/Raw('x'*20) pkt5 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=4)/IP()/Raw('x'*20) pkt6 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=5)/IP()/Raw('x'*20) pkt7 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=6)/IP()/Raw('x'*20) pkt8 = Ether(dst="00:00:00:00:01:00", src="00:02:00:00:00:01")/Dot1Q(prio=7)/IP()/Raw('x'*20) packet with vlan user priority 0 should be received by queue 0 packet with vlan user priority 1 should be received by queue 16 packet with vlan user priority 2 should be received by queue 32 packet with vlan user priority 3 should be received by queue 48 packet with vlan user priority 4 should be received by queue 64 packet with vlan user priority 5 should be received by queue 80 packet with vlan user priority 6 should be received by queue 96 packet with vlan user priority 7 should be received by queue 112 Test Case: shaping for queue with 4 tc ====================================== 1. Start testpmd and set DCB:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 3-7 -n 4 --master-lcore=3 -- -i --nb-cores=4 --rxq=4 --txq=4 --disable-rss testpmd> vlan set filter off 0 testpmd> vlan set filter off 1 testpmd> port stop all testpmd> port config 0 dcb vt off 4 pfc off testpmd> port config 1 dcb vt off 4 pfc off testpmd> port start all 2. Add root non leaf node:: testpmd> add port tm nonleaf node 1 1000000 -1 0 1 0 -1 1 0 0 3. Add tc node:: testpmd> add port tm nonleaf node 1 900000 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900001 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900002 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900003 1000000 0 1 1 -1 1 0 0 4. Add private shaper 1, 2, 3, 4 for tc node 0-4:: testpmd> add port tm node shaper profile 1 0 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 1 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 2 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 3 0 0 25000000 0 0 5. Add queue leaf node:: testpmd> add port tm leaf node 1 0 900000 0 1 2 0 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 1 900001 0 1 2 1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 2 900002 0 1 2 2 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 3 900003 0 1 2 3 0 0xffffffff 0 0 6. Commit the traffic management hierarchy on the port:: testpmd> port tm hierarchy commit 1 no testpmd> port start all testpmd> start 7. Send four flows with user priority 0, 1, 2, 3, verify shaping each flow to 200Mbps(25MBps). Test Case: shaping for queue with 8 tc ====================================== 1. Start testpmd and set DCB:: ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 3-11 -n 4 --master-lcore=3 -- -i --nb-cores=8 --rxq=8 --txq=8 --disable-rss testpmd> vlan set filter off 0 testpmd> vlan set filter off 1 testpmd> port stop all testpmd> port config 0 dcb vt off 8 pfc off testpmd> port config 1 dcb vt off 8 pfc off testpmd> port start all 2. Add root non leaf node:: testpmd> add port tm nonleaf node 1 1000000 -1 0 1 0 -1 1 0 0 3. Add tc node:: testpmd> add port tm nonleaf node 1 900000 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900001 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900002 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900003 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900004 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900005 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900006 1000000 0 1 1 -1 1 0 0 testpmd> add port tm nonleaf node 1 900007 1000000 0 1 1 -1 1 0 0 4. Add private shaper 1, 2, 3, 4, 5, 6, 7, 8, for tc node 0-7:: testpmd> add port tm node shaper profile 1 0 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 1 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 2 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 3 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 4 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 5 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 6 0 0 25000000 0 0 testpmd> add port tm node shaper profile 1 7 0 0 25000000 0 0 5. Add queue leaf node:: testpmd> add port tm leaf node 1 0 900000 0 1 2 0 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 1 900001 0 1 2 1 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 2 900002 0 1 2 2 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 3 900003 0 1 2 3 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 4 900004 0 1 2 4 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 5 900005 0 1 2 5 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 6 900006 0 1 2 6 0 0xffffffff 0 0 testpmd> add port tm leaf node 1 7 900007 0 1 2 7 0 0xffffffff 0 0 6. Commit the traffic management hierarchy on the port:: testpmd> port tm hierarchy commit 1 no testpmd> port start all testpmd> start 7. Send four flows with user priority 0, 1, 2, 3, 4, 5, 6, 7, 8, verify shaping each flow to 200Mbps(25MBps).