84. 82599 ixgbe_get_vf_queue Include Extra Information Tests¶

84.1. Description¶

VF can get following information in ixgbe driver:

Get the TC’s configured by PF for a given VF.
Get the User priority to TC mapping information for a given VF.

84.2. Prerequisites¶

Hardware: Ixgbe connect tester to pf with cable.
software: dpdk: http://dpdk.org/git/dpdk scapy: http://www.secdev.org/projects/scapy/
bind the pf to dpdk driver:
```
./usertools/dpdk-devbind.py -b igb_uio 05:00.0
```
the mac address of 05:00.0 is 00:00:00:00:01:00

create 1 vf from pf:

echo 1 >/sys/bus/pci/devices/0000:05:00.0/max_vfs

Detach VF from the host, bind them to pci-stub driver:

modprobe pci-stub

using lspci -nn|grep -i ethernet got VF device id “8086 10ed”, then:

echo "8086 10ed" > /sys/bus/pci/drivers/pci-stub/new_id
echo "0000:05:10.0" > /sys/bus/pci/drivers/ixgbevf/unbind
echo "0000:05:10.0" > /sys/bus/pci/drivers/pci-stub/bind

Lauch the VM with the VF PCI passthrough:

taskset -c 2-5 qemu-system-x86_64 \
-enable-kvm -m 8192 -smp cores=4,sockets=1 -cpu host -name dpdk1-vm1 \
-drive file=/home/VM/centOS7_1.img \
-device pci-assign,host=05:10.0 \
-netdev tap,id=ipvm1,ifname=tap3,script=/etc/qemu-ifup -device rtl8139,netdev=ipvm1,id=net0,mac=00:00:00:00:00:01 \
-localtime -vnc :2 -daemonize

login VM, get VF’s mac adress is 2e:ae:7f:16:6f:e7

84.3. Test case 1: DPDK PF, kernel VF, enable DCB mode with TC=4¶

start the testpmd on PF:

./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -c 1ffff -n 4 -- -i --rxq=4 --txq=4 --nb-cores=16
testpmd> port stop 0
testpmd> port config 0 dcb vt on 4 pfc off
testpmd> port start 0

check if VF port is linked. if vf port is down, up the port:

ifconfig eth0 up

then you can see VF information in PF side:

PMD: VF 0: enabling multicast promiscuous
PMD: VF 0: disabling multicast promiscuous

check VF’s queue number:
```
ethtool -S eth0
```
there is 1 tx queue and 4 rx queues which equals TC number.

send packet from tester to VF:

pkt1 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=0, vlan=0)/IP()/Raw('x'*20)
pkt2 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=1, vlan=0)/IP()/Raw('x'*20)
pkt3 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=2, vlan=0)/IP()/Raw('x'*20)
pkt4 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=3, vlan=0)/IP()/Raw('x'*20)
pkt5 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=4, vlan=0)/IP()/Raw('x'*20)
pkt6 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=5, vlan=0)/IP()/Raw('x'*20)
pkt7 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=6, vlan=0)/IP()/Raw('x'*20)
pkt8 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=7, vlan=0)/IP()/Raw('x'*20)
pkt9 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/Dot1Q(prio=0, vlan=1)/IP()/Raw('x'*20)
pkt10 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/IP()/Raw('x'*20)

check the packets with different User Priority mapping to TC:
```
ethtool -S eth0
```
check the NIC statistics to check the packets increasing of different rx queue. pkt1 to queue 0, pkt2 to queue 1, pkt3 to queue 2, pkt4 to queue 3, pkt5-pkt8 to queue 0, VF can’t get pkt9, pkt10 to queue 0.

84.4. Test case 2: DPDK PF, kernel VF, disable DCB mode¶

start the testpmd on PF:

./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -c 1ffff -n 4 -- -i --rxq=2 --txq=2 --nb-cores=16

check if VF port is linked. if vf port is down, up the port:
```
ifconfig eth0 up
```
if vf port is still down, reload the ixgbevf driver:
```
rmmod ixgbevf
modprobe ixgbevf
```
then you can see VF information in PF side:
```
PMD: VF 0: enabling multicast promiscuous
PMD: VF 0: disabling multicast promiscuous
```
set vlan insert to vf:
```
set vf vlan insert 0 0 1
```
check VF’s queue number:
```
ethtool -S eth0
```
there is 2 tx queues and 2 rx queues as default number.

send packet from tester to VF:

pkt1 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/IP()/Raw('x'*20)
pkt2 = Ether(dst="2e:ae:7f:16:6f:e7", src="00:02:00:00:00:01")/IP(src="192.168.0.1", dst="192.168.0.3")/UDP(sport=23,dport=24)/Raw('x'*20)

check the NIC statistics to verify the different packets mapping to different queues according RSS rule:
```
ethtool -S eth0
```
send 100 pkt1 to VF, all the packets received by queue 0, then, send 100 pkt2 to VF, all the packets received by queue 1.