Hi All, I am very new to VPP. I basically just compiled it, assigned some interfaces and was ready to start testing performance improvements over kernel with an ixia box. All in all i want to use 8 x 10G interfaces in my setup and pass through them a total of ~36Gbps of traffic but i got stuck a bit because 3 of the interfaces are not working fine.
I did this test with the interfaces assigned to the linux kernel so i know for sure the interfaces are working fine. Originally the interfaces were all transiting through a L2 switch JUNIPER QFX in order to reach IXIA. However seeing that there are problems on 3 of the 8 interfaces i converted the config on QFX to Layer 3 in order to simplify things. Further i will refer to just one interface out of the 3 that are not working. Namely TenGigabitEthernet82/0/1 (0000:82:00.1) All in all what i have now is this: QFX xe-0/0/22:1 -----------------------TenGigabitEthernet82/0/1 VPP-box. 192.168.131.4/24 192.168.131.1/24 Looking on the device running VPP i cannot see any counter increasing when i am doing a ping test from QFX towards 192.168.131.1 however on QFX side i am seeing LLDP packets bing sent from the VPP here is a sample of an LLDP packets for MAC reference: 16:10:49.122580 In Juniper PCAP Flags [Ext, In], PCAP Extension(s) total length 16 Device Media Type Extension TLV #3, length 1, value: Ethernet (1) Logical Interface Encapsulation Extension TLV #6, length 1, value: Ethernet (14) Device Interface Index Extension TLV #1, length 2, value: 727 Logical Interface Index Extension TLV #4, length 4, value: 578 -----original packet----- 3c:fd:fe:13:b7:c2 > 01:80:c2:00:00:0e, ethertype LLDP (0x88cc), length 110: LLDP, length 96 Chassis ID TLV (1), length 7 Subtype MAC address (4): 3c:fd:fe:13:b7:c2 0x0000: 043c fdfe 13b7 c2 Port ID TLV (2), length 7 Subtype MAC address (3): 3c:fd:fe:13:b7:c2 0x0000: 033c fdfe 13b7 c2 Time to Live TLV (3), length 2: TTL 121s 0x0000: 0079 i can confirm this e above is the correct MAC address for that specific port on the VPP box by looking at dmsg log at the point of DPDK assignment: [ 7.043183] i40e 0000:82:00.1: fw 5.0.40043 api 1.5 nvm 5.05 0x8000289f 17.5.12 [ 7.272171] i40e 0000:82:00.1: MAC address: 3c:fd:fe:13:b7:c2 [ 7.301191] i40e 0000:82:00.1 eth13: NIC Link is Up, 10 Gbps Full Duplex, Flow Control: RX/TX [ 7.330768] i40e 0000:82:00.1: PCI-Express: Speed 8.0GT/s Width x8 [ 7.343107] i40e 0000:82:00.1: Features: PF-id[1] VFs: 32 VSIs: 34 QP: 56 RSS FD_ATR FD_SB NTUPLE DCB VxLAN Geneve PTP VEPA Here are some details about the interfaces: vpp# show pci Address Sock VID:PID Link Speed Driver Product Name Vital Product Data 0000:01:00.0 0 8086:1572 8.0 GT/s x8 i40e X710 10GbE Controller 0000:01:00.1 0 8086:1572 8.0 GT/s x8 i40e X710 10GbE Controller 0000:01:00.2 0 8086:1572 8.0 GT/s x8 igb_uio X710 10GbE Controller <<<<working fine 0000:01:00.3 0 8086:1572 8.0 GT/s x8 igb_uio X710 10GbE Controller <<<<working fine 0000:04:00.0 0 8086:1572 8.0 GT/s x8 i40e X710 10GbE Controller 0000:04:00.1 0 8086:1572 8.0 GT/s x8 igb_uio X710 10GbE Controller <<<<working fine 0000:04:00.2 0 8086:1572 8.0 GT/s x8 i40e X710 10GbE Controller 0000:04:00.3 0 8086:1572 8.0 GT/s x8 i40e X710 10GbE Controller 0000:05:00.0 0 8086:1572 8.0 GT/s x8 i40e X710 10GbE Controller 0000:05:00.1 0 8086:1572 8.0 GT/s x8 i40e X710 10GbE Controller 0000:05:00.2 0 8086:1572 8.0 GT/s x8 i40e X710 10GbE Controller 0000:05:00.3 0 8086:1572 8.0 GT/s x8 igb_uio X710 10GbE Controller <<<<working fine 0000:82:00.0 1 8086:1572 8.0 GT/s x8 igb_uio X710 10GbE Controller <<<<working fine 0000:82:00.1 1 8086:1572 8.0 GT/s x8 igb_uio X710 10GbE Controller <<<<NOT working 0000:82:00.2 1 8086:1572 8.0 GT/s x8 igb_uio X710 10GbE Controller <<<<NOT working 0000:82:00.3 1 8086:1572 8.0 GT/s x8 igb_uio X710 10GbE Controller <<<<NOT working vpp# show interface Name Idx State Counter Count TenGigabitEthernet1/0/2 1 up rx packets 4 rx bytes 316 tx packets 4 tx bytes 298 drops 4 ip4 2 TenGigabitEthernet1/0/3 2 up rx packets 4 rx bytes 316 tx packets 4 tx bytes 298 drops 4 ip4 2 TenGigabitEthernet4/0/1 3 up rx packets 3 rx bytes 218 tx packets 3 tx bytes 200 drops 3 ip4 1 TenGigabitEthernet5/0/3 4 up rx packets 3 rx bytes 218 tx packets 3 tx bytes 200 drops 3 ip4 1 TenGigabitEthernet82/0/0 5 up rx packets 509 rx bytes 49806 tx packets 509 tx bytes 49788 drops 4 ip4 507 TenGigabitEthernet82/0/1 6 up tx packets 10 tx bytes 420 drops 15 TenGigabitEthernet82/0/2 7 up tx packets 21 tx bytes 882 drops 15 TenGigabitEthernet82/0/3 8 up tx packets 10 tx bytes 420 drops 10 local0 0 down vpp# show interface address TenGigabitEthernet1/0/2 (up): L3 192.168.134.1/24 TenGigabitEthernet1/0/3 (up): L3 192.168.135.1/24 TenGigabitEthernet4/0/1 (up): L3 192.168.136.1/24 TenGigabitEthernet5/0/3 (up): L3 192.168.137.1/24 TenGigabitEthernet82/0/0 (up): L3 192.168.130.1/24 TenGigabitEthernet82/0/1 (up): L3 192.168.131.1/24 TenGigabitEthernet82/0/2 (up): L3 192.168.132.1/24 TenGigabitEthernet82/0/3 (up): L3 192.168.133.1/24 I have tried multiple things: 1. I have tried running a trace but whenever i am doing traffic to any of this interface that is not working i am getting "No packets in trace buffer". However when i am trying to do traffic towards one of the working interfaces, i am seeing packets in the trace. 2. I tried initiating traffic towards this non-working interfaces but i see that the counter is not increasing on 82/0/1. 3. Another "strange" thing i have noticed is that the drop counter is under the "show interface" on VPP does not increase when i am doing the test. Considering the above, my guess is that the packets are not even getting to be processed by the VPP and i do not know why yet. I am pretty sure about the physical connection so all i can think of is some VPP problem. Is there a limit of interfaces i can enable ? How can i debug this further considering i do not see anything on the trace for this particular interface? When any of the 3 non working interfaces are moved back to the linux kernel, they are working absolutely fine. Thanks in advance for your help.
