178. VF One-shot Rx Interrupt Tests

One-shot Rx interrupt feature will split rx interrupt handling from other interrupts like LSC interrupt. It implemented one handling mechanism to eliminate non-deterministic DPDK polling thread wakeup latency.

VFIO’ multiple interrupt vectors support mechanism to enable multiple event fds serving per Rx queue interrupt handling. UIO has limited interrupt support, specifically it only support a single interrupt vector, which is not suitable for enabling multi queues Rx/Tx interrupt.

178.1. Prerequisites

The suit support NIC: Intel® Ethernet 700 Series, Intel® Ethernet 800 Series and Intel® 82599 Gigabit Ethernet Controller.

Each of the 10Gb Ethernet* ports of the DUT is directly connected in full-duplex to a different port of the peer traffic generator.

Assume PF port PCI addresses is 0000:04:00.0, their Interfaces name is p786p0. Assume generated VF PCI address will be 0000:04:10.0.

Iommu pass through feature has been enabled in kernel:

intel_iommu=on iommu=pt

Modify the DPDK-l3fwd-power source code and recompile the l3fwd-power:

meson configure -Dexamples=l3fwd-power x86_64-native-linuxapp-gcc
ninja -C x86_64-native-linuxapp-gcc

Support igb_uio and vfio driver, if used vfio, kernel need 3.6+ and enable vt-d in bios. When used vfio, requested to insmod two drivers vfio and vfio-pci.

178.2. Test Case1: Check Interrupt for VF with vfio driver on ixgbe and i40e

1. Generate NIC VF, then bind it to vfio drvier:

   echo 1 > /sys/bus/pci/devices/0000\:04\:00.0/sriov_numvfs
   
   modprobe vfio-pci
   ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:04:10.0(vf_pci)
   

Notice: If your PF is kernel driver, make sure PF link is up when your start testpmd on VF.

2. Start l3fwd-power with VF:

   ./x86_64-native-linuxapp-gcc/examples/dpdk-l3fwd-power -l 1-3 -n 4 -- -P -p 0x01  --config '(0,0,2)'
   

3. Send packet with packet generator to the pf NIC, check that thread core2 waked up:

   sendp([Ether(dst='vf_mac')/IP()/UDP()/Raw(load='XXXXXXXXXXXXXXXXXX')], iface="tester_intf")
   
   L3FWD_POWER: lcore 2 is waked up from rx interrupt on port 0 queue 0
   

4. Check if threads on core 2 have returned to sleep mode:

   L3FWD_POWER: lcore 2 sleeps until interrupt triggers
   

178.3. Test Case2: VF interrupt pmd in VM with vfio-pci

1. Generate NIC VF, then bind it to vfio drvier:

   echo 1 > /sys/bus/pci/devices/0000\:04\:00.0/sriov_numvfs
   
   modprobe vfio-pci
   ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:04:10.0(vf_pci)
   

2. passthrough VF 0 to VM0 and start VM0:

   taskset -c 4,5,6,7 qemu-system-x86_64 \
   -name vm0 -enable-kvm -pidfile /tmp/.vm0.pid -daemonize -monitor unix:/tmp/vm0_monitor.sock,server,nowait \
   -device e1000,netdev=nttsip1  -netdev user,id=nttsip1,hostfwd=tcp:10.240.176.207:6000-:22 \
   -device vfio-pci,host=0000:04:02.0,id=pt_0 -cpu host -smp 4 -m 10240 \
   -chardev socket,path=/tmp/vm0_qga0.sock,server,nowait,id=vm0_qga0 -device virtio-serial \
   -device virtserialport,chardev=vm0_qga0,name=org.qemu.guest_agent.0 -vnc :11 \
   -drive file=/home/image/ubuntu16-0.img,format=qcow2,if=virtio,index=0,media=disk
   

3. Compile the l3fwd-power:

   meson configure -Dexamples=l3fwd-power x86_64-native-linuxapp-gcc
   ninja -C x86_64-native-linuxapp-gcc
   

4. Bind VF 0 to the vfio-pci driver:

   modprobe -r vfio_iommu_type1
   modprobe -r vfio
   modprobe vfio enable_unsafe_noiommu_mode=1
   modprobe vfio-pci
   
   ./usertools/dpdk-devbind.py -b vfio-pci 0000:00:04.0
   

5. start l3fwd-power in VM:

   ./x86_64-native-linuxapp-gcc/examples/dpdk-l3fwd-power -l 1-3 -n 4 -- -P -p 0x01  --config '(0,0,2)'
   

6. Send packet with packet generator to the VM, check that thread core2 waked up:

   sendp([Ether(dst='vf_mac')/IP()/UDP()/Raw(load='XXXXXXXXXXXXXXXXXX')], iface="tester_intf")
   
   L3FWD_POWER: lcore 2 is waked up from rx interrupt on port 0 queue 0
   

7. Check if threads on core 2 have returned to sleep mode:

   L3FWD_POWER: lcore 2 sleeps until interrupt triggers
   

178.4. Test Case3: vf multi-queue interrupt with vfio-pci on i40e

1. Generate NIC VF, then bind it to vfio drvier:

   echo 1 > /sys/bus/pci/devices/0000\:04\:00.0/sriov_numvfs
   modprobe vfio-pci
   ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:04:10.0(vf_pci)
   

Notice: If your PF is kernel driver, make sure PF link is up when your start testpmd on VF.

2. Start l3fwd-power with VF:

   ./x86_64-native-linuxapp-gcc/examples/dpdk-l3fwd-power -c 3f -n 4 -m 2048 -- -P -p 0x1 --config="(0,0,1),(0,1,2),(0,2,3),(0,3,4)"
   

3. Send UDP packets with random ip and dest mac = vf mac addr:

   for x in range(0,10):
    sendp(Ether(src="00:00:00:00:01:00",dst="vf_mac")/IP(src='2.1.1.' + str(x),dst='2.1.1.5')/UDP()/"Hello!0",iface="tester_intf")
   

4. Check if threads on all cores have waked up:

   L3FWD_POWER: lcore 1 is waked up from rx interrupt on port 0 queue 0
   L3FWD_POWER: lcore 2 is waked up from rx interrupt on port 0 queue 1
   L3FWD_POWER: lcore 3 is waked up from rx interrupt on port 0 queue 2
   L3FWD_POWER: lcore 4 is waked up from rx interrupt on port 0 queue 3
   

178.5. Test Case4: VF multi-queue interrupt in VM with vfio-pci on i40e

1. Generate NIC VF, then bind it to vfio drvier:

   echo 1 > /sys/bus/pci/devices/0000\:88:00.1/sriov_numvfs
   modprobe vfio-pci
   ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:88:0a.0(vf_pci)
   

Notice: If your PF is kernel driver, make sure PF link is up when your start testpmd on VF.

2. Passthrough VF 0 to VM0 and start VM0:

   taskset -c 4,5,6,7,8 qemu-system-x86_64 \
   -name vm0 -enable-kvm -pidfile /tmp/.vm0.pid -daemonize -monitor unix:/tmp/vm0_monitor.sock,server,nowait \
   -device e1000,netdev=nttsip1  -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6000-:22 \
   -device vfio-pci,host=0000:88:0a.0,id=pt_0 -cpu host -smp 5 -m 10240 \
   -chardev socket,path=/tmp/vm0_qga0.sock,server,nowait,id=vm0_qga0 -device virtio-serial \
   -device virtserialport,chardev=vm0_qga0,name=org.qemu.guest_agent.0 -vnc :11 \
   -drive file=/home/osimg/noiommu-ubt16.img,format=qcow2,if=virtio,index=0,media=disk
   

Notice: VM needs Kernel version > 4.8.0, mostly linux distribution don’t support vfio-noiommu mode by default, so testing this case need rebuild kernel to enable vfio-noiommu.

3. Bind VF 0 to the vfio-pci driver:

   modprobe -r vfio_iommu_type1
   modprobe -r vfio
   modprobe vfio enable_unsafe_noiommu_mode=1
   modprobe vfio-pci
   ./usertools/dpdk-devbind.py -b vfio-pci 0000:00:04.0
   

4.Start l3fwd-power in VM:

./x86_64-native-linuxapp-gcc/examples/dpdk-l3fwd-power -l 0-3 -n 4 -m 2048 -- -P -p 0x1 --config="(0,0,0),(0,1,1),(0,2,2),(0,3,3)"

5. Send UDP packets with random ip and dest mac = vf mac addr:

   for x in range(0,10):
    sendp(Ether(src="00:00:00:00:01:00",dst="vf_mac")/IP(src='2.1.1.' + str(x),dst='2.1.1.5')/UDP()/"Hello!0",iface="tester_intf")
   

6. Check if threads on core 0 to core 3 can be waked up in VM:

   L3FWD_POWER: lcore 0 is waked up from rx interrupt on port 0 queue 0
   L3FWD_POWER: lcore 1 is waked up from rx interrupt on port 0 queue 1
   L3FWD_POWER: lcore 2 is waked up from rx interrupt on port 0 queue 2
   L3FWD_POWER: lcore 3 is waked up from rx interrupt on port 0 queue 3