Hi there,
I am trying to setup and run VPP HPS in my local vagrant vm. There I have 2 x
1Gbs (82540EM EC) configured for VPP. I configured HPS placement per:
/etc/vpp/startup.conf
cpu {
...
main-core 0
corelist-workers 1, 2
corelist-hqos-threads 3
...
}
dpdk {
...
socket-mem 1024
dev 0000:00:08.0 {
num-rx-queues 1
}
dev 0000:00:09.0 {
num-tx-queues 1
hqos
}
}
vpp# show dpdk version
DPDK Version: DPDK 16.11.0
DPDK EAL init args: -c f -n 4 --huge-dir /run/vpp/hugepages --file-prefix
vpp -w 0000:00:08.0 -w 0000:00:09.0 --master-lcore 0 --socket-mem 1024
I assigned the following static arp-s and a route for my simple test:
vpp# sh int
Name Idx State Counter Count
GigabitEthernet0/8/0 1 down
GigabitEthernet0/9/0 2 down
local0 0 down
vpp# set int ip address GigabitEthernet0/8/0 10.10.0.1/24
vpp# set interface state GigabitEthernet0/8/0 up
vpp# set int ip address GigabitEthernet0/9/0 10.10.0.10/24
vpp# set interface state GigabitEthernet0/9/0 up
vpp# set ip arp static GigabitEthernet0/9/0 10.10.10.1 08:00:27:25:8A:0F
vpp# set ip arp static GigabitEthernet0/9/0 10.10.10.2 08:00:27:25:8A:0F
vpp# set ip arp static GigabitEthernet0/9/0 10.10.10.3 08:00:27:25:8A:0F
vpp# set ip arp static GigabitEthernet0/9/0 10.10.10.4 08:00:27:25:8A:0F
vpp# ip route add 10.10.10.255/24 via GigabitEthernet0/9/0
vpp# show dpdk interface placement
Thread 2 (vpp_wk_0 at lcore 1):
GigabitEthernet0/8/0 queue 0
Thread 3 (vpp_wk_1 at lcore 2):
GigabitEthernet0/9/0 queue 0
I sent the traffic from my traffic generator (Dst IPv4 range 10.10.10.1-4, and
random DSCP). From the trace log I cannot see the packet hitting the HPS:
vpp# clear trace
vpp# trace add dpdk-input 10
vpp# show trace
------------------- Start of thread 0 vpp_main -------------------
No packets in trace buffer
------------------- Start of thread 1 vpp_hqos-threads_0 -------------------
No packets in trace buffer
------------------- Start of thread 2 vpp_wk_0 -------------------
Packet 1
00:09:53:685732: dpdk-input
GigabitEthernet0/8/0 rx queue 0
buffer 0x4e03: current data 14, length 110, free-list 0, totlen-nifb 0, trace
0x0
PKT MBUF: port 0, nb_segs 1, pkt_len 124
buf_len 2176, data_len 124, ol_flags 0x0, data_off 128, phys_addr 0xc2d33fc0
packet_type 0x0
IP4: 08:00:27:25:8a:0f -> 08:00:27:f1:2a:09
UDP: 10.10.0.254 -> 10.10.10.3
tos 0xca, ttl 64, length 110, checksum 0x5aa1
fragment id 0x0000
UDP: 3333 -> 4444
length 90, checksum 0xc1c4
00:09:53:685823: ip4-input
UDP: 10.10.0.254 -> 10.10.10.3
tos 0xca, ttl 64, length 110, checksum 0x5aa1
fragment id 0x0000
UDP: 3333 -> 4444
length 90, checksum 0xc1c4
00:09:53:685845: ip4-lookup
fib 0 dpo-idx 4 flow hash: 0x00000000
UDP: 10.10.0.254 -> 10.10.10.3
tos 0xca, ttl 64, length 110, checksum 0x5aa1
fragment id 0x0000
UDP: 3333 -> 4444
length 90, checksum 0xc1c4
00:09:53:685853: ip4-rewrite
tx_sw_if_index 2 dpo-idx 4 : ipv4 via 10.10.10.3 GigabitEthernet0/9/0: IP4:
08:00:27:68:0c:e4 -> 08:00:27:25:8a:0f flow hash: 0x00000000
IP4: 08:00:27:68:0c:e4 -> 08:00:27:25:8a:0f
UDP: 10.10.0.254 -> 10.10.10.3
tos 0xca, ttl 63, length 110, checksum 0x5ba1
fragment id 0x0000
UDP: 3333 -> 4444
length 90, checksum 0xc1c4
00:09:53:685858: GigabitEthernet0/9/0-output
GigabitEthernet0/9/0
IP4: 08:00:27:68:0c:e4 -> 08:00:27:25:8a:0f
UDP: 10.10.0.254 -> 10.10.10.3
tos 0xca, ttl 63, length 110, checksum 0x5ba1
fragment id 0x0000
UDP: 3333 -> 4444
length 90, checksum 0xc1c4
00:09:53:685871: GigabitEthernet0/9/0-tx
GigabitEthernet0/9/0 tx queue 2
buffer 0x4e03: current data 0, length 124, free-list 0, totlen-nifb 0, trace
0x0
IP4: 08:00:27:68:0c:e4 -> 08:00:27:25:8a:0f
UDP: 10.10.0.254 -> 10.10.10.3
tos 0xca, ttl 63, length 110, checksum 0x5ba1
fragment id 0x0000
UDP: 3333 -> 4444
length 90, checksum 0xc1c4
...
The VPP documentation shows the following HPS cli commands:
set dpdk interface hqos pipe <if-name> subport <n> pipe <n> profile <n>
set dpdk interface hqos pktfield <if-name> id <n> offset <n> mask <n>
set dpdk interface hqos placement <if-name> thread <n>
set dpdk interface hqos subport <if-name> subport <n> [rate <n>] [bktsize
<n>] [tc0 <n>] [tc1 <n>] [tc2 <n>] [tc3 <n>] [period <n>]
set dpdk interface hqos tctbl <if-name> entry <n> tc <n> queue <n>
set dpdk interface placement <if-name> [queue <n>] thread <n>
show dpdk hqos queue <if-name> subport <subport> pipe <pipe> tc <tc> tc_q
<tc_q>
show dpdk interface hqos placement
show dpdk interface hqos <if-name>
show dpdk interface placement
This is the HPS placement shown in my setup:
vpp# show dpdk interface hqos placement
Thread 1 (vpp_hqos-threads_0 at lcore 3):
GigabitEthernet0/9/0 queue 0
vpp# show dpdk interface placement
Thread 2 (vpp_wk_0 at lcore 1):
GigabitEthernet0/8/0 queue 0
Thread 3 (vpp_wk_1 at lcore 2):
GigabitEthernet0/9/0 queue 0
vpp# show dpdk interface hqos placement
Thread 1 (vpp_hqos-threads_0 at lcore 3):
GigabitEthernet0/9/0 queue 0
How someone can configure VPS HPS port parameters form the VPP CLI? I only see
the default port attributes set as per
https://docs.fd.io/vpp/17.01/d5/dd9/hqos_8c_source.html
..
95 /* port */
96 .port = {
97 .name = NULL, /* Set at init */
98 .socket = 0, /* Set at init */
99 .rate = 1250000000, /* Assuming 10GbE port */
100 .mtu = 14 + 1500, /* Assuming Ethernet/IPv4 pkt (Ethernet
FCS not included) */
101 .frame_overhead = RTE_SCHED_FRAME_OVERHEAD_DEFAULT,
102 .n_subports_per_port = 1,
103 .n_pipes_per_subport = 4096,
104 .qsize = {64, 64, 64, 64},
105 .pipe_profiles = NULL, /* Set at config */
106 .n_pipe_profiles = 1,
I cannot change the default setting:
vpp# set dpdk interface hqos pipe GigabitEthernet0/9/0 subport 0 pipe 0 profile 0
vpp# set dpdk interface hqos subport GigabitEthernet0/9/0 subport 0 rate
125000000 bktsize 10000 tc0 30517 tc1 30517 tc2 30517 tc3 30517 period 80
vpp# show dpdk interface hqos GigabitEthernet0/9/0
Thread:
Input SWQ size = 4096 packets
Enqueue burst size = 256 packets
Dequeue burst size = 220 packets
Packet field 0: slab position = 0, slab bitmask = 0x0000000000000000
Packet field 1: slab position = 40, slab bitmask = 0x0000000fff000000
Packet field 2: slab position = 8, slab bitmask = 0x00000000000000fc
Packet field 2 translation table:
[ 0 .. 15]: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
[16 .. 31]: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
[32 .. 47]: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
[48 .. 63]: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Port:
Rate = 1250000000 bytes/second
MTU = 1514 bytes
Frame overhead = 24 bytes
Number of subports = 1
Number of pipes per subport = 4096
Packet queue size: TC0 = 64, TC1 = 64, TC2 = 64, TC3 = 64 packets
Number of pipe profiles = 1
Pipe profile 0:
Rate = 305175 bytes/second
Token bucket size = 1000000 bytes
Traffic class rate: TC0 = 305175, TC1 = 305175, TC2 = 305175, TC3 = 305175
bytes/second
TC period = 40 milliseconds
TC0 WRR weights: Q0 = 1, Q1 = 1, Q2 = 1, Q3 = 1
TC1 WRR weights: Q0 = 1, Q1 = 1, Q2 = 1, Q3 = 1
TC2 WRR weights: Q0 = 1, Q1 = 1, Q2 = 1, Q3 = 1
TC3 WRR weights: Q0 = 1, Q1 = 1, Q2 = 1, Q3 = 1
I appreciate any help with my HPS setup.
Thank you,
Zhelyko Oluic
::DISCLAIMER::
----------------------------------------------------------------------------------------------------------------------------------------------------
The contents of this e-mail and any attachment(s) are confidential and intended
for the named recipient(s) only.
E-mail transmission is not guaranteed to be secure or error-free as information
could be intercepted, corrupted,
lost, destroyed, arrive late or incomplete, or may contain viruses in
transmission. The e mail and its contents
(with or without referred errors) shall therefore not attach any liability on
the originator or HCL or its affiliates.
Views or opinions, if any, presented in this email are solely those of the
author and may not necessarily reflect the
views or opinions of HCL or its affiliates. Any form of reproduction,
dissemination, copying, disclosure, modification,
distribution and / or publication of this message without the prior written
consent of authorized representative of
HCL is strictly prohibited. If you have received this email in error please
delete it and notify the sender immediately.
Before opening any email and/or attachments, please check them for viruses and
other defects.
----------------------------------------------------------------------------------------------------------------------------------------------------
_______________________________________________
vpp-dev mailing list
vpp-dev@lists.fd.io
https://lists.fd.io/mailman/listinfo/vpp-dev
