From patchwork Wed Aug 17 08:07:55 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Ling, WeiX" X-Patchwork-Id: 115199 Return-Path: X-Original-To: patchwork@inbox.dpdk.org Delivered-To: patchwork@inbox.dpdk.org Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id E798FA0032; Wed, 17 Aug 2022 10:12:15 +0200 (CEST) Received: from [217.70.189.124] (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id DFFD040E2D; Wed, 17 Aug 2022 10:12:15 +0200 (CEST) Received: from mga09.intel.com (mga09.intel.com [134.134.136.24]) by mails.dpdk.org (Postfix) with ESMTP id 670C04068E for ; Wed, 17 Aug 2022 10:12:13 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1660723933; x=1692259933; h=from:to:cc:subject:date:message-id:mime-version: content-transfer-encoding; bh=7NVK8+ZWIvSVYBXjlsBWb/qFCwFFrlzvtJAsoQPqqV0=; b=Sy+Gsf3vABrZ1LVmgqyjhWAcSsbPpRyfZsXw1p3CHSOQbORHtX21GICM QTIY1dfHZQyo0gYgBCeghS+TulIxCieID+oPmiAHjndxlLVbWk1YzTNza ryzz6j9UYpR2E98UrQyXHFA2MCgivWt7umRPXTxn/jNz78c/XVvFqZKIR XCGeY0T+ZCzNqbhoyq2Y9J14503u2v0ULOFjtdM1m7Pd58sQj2hTb/Z3h t+L4X8qrFO02ueVXSTVNSv1JZKtXGjAN3Vgo9tg+IIc/yWk7N3jc2FpCF Yfaks8CC+FnnwutBbj0DuF/8hxAeeQOGIWTcLS3zRh1XlwQPkt7SXEoSJ Q==; X-IronPort-AV: E=McAfee;i="6400,9594,10441"; a="293228304" X-IronPort-AV: E=Sophos;i="5.93,242,1654585200"; d="scan'208,223";a="293228304" Received: from orsmga007.jf.intel.com ([10.7.209.58]) by orsmga102.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 17 Aug 2022 01:12:12 -0700 X-IronPort-AV: E=Sophos;i="5.93,242,1654585200"; d="scan'208,223";a="603811386" Received: from unknown (HELO localhost.localdomain) ([10.239.252.222]) by orsmga007-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 17 Aug 2022 01:12:10 -0700 From: Wei Ling To: dts@dpdk.org Cc: Wei Ling Subject: [dts][PATCH V4 1/3] test_plans/basic_4k_pages_cbdma_test_plan: modify testplan to test virtio dequeue Date: Wed, 17 Aug 2022 04:07:55 -0400 Message-Id: <20220817080755.4018061-1-weix.ling@intel.com> X-Mailer: git-send-email 2.25.1 MIME-Version: 1.0 X-BeenThere: dts@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: test suite reviews and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dts-bounces@dpdk.org From DPDK-22.07, virtio support async dequeue for split and packed ring path, so modify basic_4k_pages_cbdma testplan to test the split and packed ring async dequeue feature. Signed-off-by: Wei Ling --- test_plans/basic_4k_pages_cbdma_test_plan.rst | 492 ++++++++++++------ 1 file changed, 334 insertions(+), 158 deletions(-) diff --git a/test_plans/basic_4k_pages_cbdma_test_plan.rst b/test_plans/basic_4k_pages_cbdma_test_plan.rst index 5bbd19b4..66ce000f 100644 --- a/test_plans/basic_4k_pages_cbdma_test_plan.rst +++ b/test_plans/basic_4k_pages_cbdma_test_plan.rst @@ -5,27 +5,26 @@ Basic test with CBDMA in 4K-pages test plan =========================================== -DPDK 19.02 add support for using virtio-user without hugepages. The --no-huge mode was augmented to use memfd-backed memory +DPDK 19.02 add support for using virtio-user without hugepages. The --no-huge mode was augmented to use memfd-backed memory (on systems that support memfd), to allow using virtio-user-based NICs without hugepages. Vhost asynchronous data path leverages DMA devices to offload memory copies from the CPU and it is implemented in an asynchronous way. -In addition, vhost supports M:N mapping between vrings and DMA virtual channels. Specifically, one vring can use multiple different DMA -channels and one DMA channel can be shared by multiple vrings at the same time. Vhost enqueue operation with CBDMA channels is supported -in both split and packed ring. +In addition, vhost supports M:N mapping between vrings and DMA virtual channels. Specifically, one vring can use multiple different DMA +channels and one DMA channel can be shared by multiple vrings at the same time. From DPDK22.07, Vhost enqueue and dequeue operation with +CBDMA channels is supported in both split and packed ring. This document provides the test plan for testing some basic functions with CBDMA device in 4K-pages memory environment. 1. Test split and packed ring virtio path in the PVP topology environmet. 2. Check Vhost tx offload function by verifing the TSO/cksum in the TCP/IP stack with vm2vm split ring and packed ring vhost-user/virtio-net mergeable path. -3.Check the payload of large packet (larger than 1MB) is valid after forwarding packets with vm2vm split ring and packed ring +3.Check the payload of large packet (larger than 1MB) is valid after forwarding packets with vm2vm split ring and packed ring vhost-user/virtio-net mergeable path. -4. Vhost-user using 1G hugepges and virtio-user using 4k-pages. -.. note: +Note: - 1. When CBDMA channels are bound to vfio driver, VA mode is the default and recommended. - For PA mode, page by page mapping may exceed IOMMU's max capability, better to use 1G guest hugepage. - 2. DPDK local patch that about vhost pmd is needed when testing Vhost asynchronous data path with testpmd. And case 4-5 have not yet been automated. +1. When CBDMA channels are bound to vfio driver, VA mode is the default and recommended. +For PA mode, page by page mapping may exceed IOMMU's max capability, better to use 1G guest hugepage. +2. DPDK local patch that about vhost pmd is needed when testing Vhost asynchronous data path with testpmd. And case 4-5 have not yet been automated. For more about dpdk-testpmd sample, please refer to the DPDK docments: https://doc.dpdk.org/guides/testpmd_app_ug/run_app.html @@ -45,11 +44,11 @@ General set up 2. Compile DPDK:: - # CC=gcc meson --werror -Denable_kmods=True -Dlibdir=lib -Dexamples=all --default-library= + # CC=gcc meson --werror -Denable_kmods=True -Dlibdir=lib -Dexamples=all --default-library=static # ninja -C -j 110 - For example: - CC=gcc meson --werror -Denable_kmods=True -Dlibdir=lib -Dexamples=all --default-library=x86_64-native-linuxapp-gcc - ninja -C x86_64-native-linuxapp-gcc -j 110 + For example: + CC=gcc meson --werror -Denable_kmods=True -Dlibdir=lib -Dexamples=all --default-library=static x86_64-native-linuxapp-gcc + ninja -C x86_64-native-linuxapp-gcc -j 110 3. Get the PCI device ID and DMA device ID of DUT, for example, 0000:18:00.0 is PCI device ID, 0000:00:04.0, 0000:00:04.1 is DMA device ID:: @@ -66,10 +65,10 @@ General set up 4. Prepare tmpfs with 4K-pages:: - mkdir /mnt/tmpfs_nohuge0 - mkdir /mnt/tmpfs_nohuge1 - mount tmpfs /mnt/tmpfs_nohuge0 -t tmpfs -o size=4G - mount tmpfs /mnt/tmpfs_nohuge1 -t tmpfs -o size=4G + mkdir /mnt/tmpfs_nohuge0 + mkdir /mnt/tmpfs_nohuge1 + mount tmpfs /mnt/tmpfs_nohuge0 -t tmpfs -o size=4G + mount tmpfs /mnt/tmpfs_nohuge1 -t tmpfs -o size=4G Test case ========= @@ -82,312 +81,489 @@ Common steps # ./usertools/dpdk-devbind.py -b vfio-pci For example, bind 1 NIC port and 1 CBDMA channels: - # ./usertools/dpdk-devbind.py -b vfio-pci 0000:af:00.0,0000:80:04.0 + # ./usertools/dpdk-devbind.py -b vfio-pci 0000:af:00.0 0000:80:04.0 -Test Case 1: Basic test vhost/virtio-user split ring with 4K-pages and cbdma enable ------------------------------------------------------------------------------------ -This case uses testpmd Traffic Generator(For example, Trex) to test split ring when vhost uses the asynchronous enqueue operations with CBDMA channels -in 4k-pages environment. And the mapping between vrings and dsa virtual channels is 1:1. +Test Case 1: Basic test vhost-user/virtio-user split ring vhost async operation using 4K-pages and cbdma enable +--------------------------------------------------------------------------------------------------------------- +This case tests basic functions of split ring virtio path when uses the asynchronous operations with CBDMA channels in 4K-pages memory environment and PVP vhost-user/virtio-user topology. -1. Bind 1 NIC port and 1 CBDMA channel to vfio-pci, as common steps 1. +1. Bind one port to vfio-pci, launch vhost:: -2. Launch vhost by below command:: + ./usertools/dpdk-devbind.py --bind=vfio-pci xx:xx.x + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 3-4 -n 4 -m 1024 --no-huge --file-prefix=vhost -a 0000:18:00.0 -a 0000:00:04.0 \ + --vdev 'net_vhost0,iface=/tmp/vhost-net,queues=1,dmas=[txq0;rxq0]' -- -i --no-numa --socket-num=0 --lcore-dma=[lcore4@0000:00:04.0] + testpmd>start - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 31-32 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:af:00.0 -a 0000:80:04.0 \ - --vdev 'net_vhost0,iface=/tmp/vhost-net,queues=1,dmas=[txq0]' -- -i --no-numa --socket-num=1 --lcore-dma=[lcore32@0000:80:04.0] - testpmd> start +2. Launch virtio-user with 4K-pages:: -3. Launch virtio-user with 4K-pages:: + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 5-6 -n 4 --no-huge -m 1024 --file-prefix=virtio-user \ + --vdev=net_virtio_user0,mac=00:01:02:03:04:05,path=/tmp/vhost-net,queues=1 -- -i + testpmd>start - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 33-34 -n 4 --no-huge -m 1024 --file-prefix=virtio-user --no-pci \ - --vdev=net_virtio_user0,mac=00:11:22:33:44:10,path=/tmp/vhost-net,queues=1 -- -i - testpmd> set fwd mac - testpmd> start +3. Send packet with packet generator with different packet size,includes [64, 128, 256, 512, 1024, 1518], check the throughput with below command:: -4. Send packet with packet generator with different packet size,includes [64, 128, 256, 512, 1024, 1518], check the throughput with below command:: + testpmd>show port stats all - testpmd> show port stats all +Test Case 2: Basic test vhost-user/virtio-user packed ring vhost async operation using 4K-pages and cbdma enable +---------------------------------------------------------------------------------------------------------------- +This case tests basic functions of packed ring virtio path when uses the asynchronous operations with CBDMA channels in 4K-pages memory environment and PVP vhost-user/virtio-user topology. -Test Case 2: Basic test vhost/virtio-user packed ring with 4K-pages and cbdma enable ------------------------------------------------------------------------------------- -This case uses testpmd Traffic Generator(For example, Trex) to test packed ring when vhost uses the asynchronous enqueue operations with CBDMA channels -in 4k-pages environment. And the mapping between vrings and dsa virtual channels is 1:1. +1. Bind one port to vfio-pci, launch vhost:: -1. Bind 1 NIC port and 1 CBDMA channel to vfio-pci, as common steps 1. + modprobe vfio-pci + ./usertools/dpdk-devbind.py --bind=vfio-pci xx:xx.x + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 3-4 -n 4 -m 1024 --no-huge --file-prefix=vhost -a 0000:18:00.0 -a 0000:00:04.0 \ + --vdev 'net_vhost0,iface=/tmp/vhost-net,queues=1,dmas=[txq0;rxq0]' -- -i --no-numa --socket-num=0 --lcore-dma=[lcore4@0000:00:04.0] + testpmd>start -2. Launch vhost by below command:: +2. Launch virtio-user with 4K-pages:: - ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:af:00.0 0000:80:04.0 - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 31-32 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:af:00.0 -a 0000:80:04.0 \ - --vdev 'net_vhost0,iface=/tmp/vhost-net,queues=1,dmas=[txq0]' -- -i --no-numa --socket-num=1 --lcore-dma=[lcore32@0000:80:04.0] - testpmd> start + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 5-6 -n 4 --no-huge -m 1024 --file-prefix=virtio-user \ + --vdev=net_virtio_user0,mac=00:01:02:03:04:05,path=/tmp/vhost-net,packed_vq=1,queues=1 -- -i + testpmd>start -3. Launch virtio-user with 4K-pages:: +3. Send packet with packet generator with different packet size,includes [64, 128, 256, 512, 1024, 1518], check the throughput with below command:: - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 33-34 -n 4 --no-huge -m 1024 --file-prefix=virtio-user --no-pci \ - --vdev=net_virtio_user0,mac=00:11:22:33:44:10,path=/tmp/vhost-net,packed_vq=1,queues=1 -- -i - testpmd> set fwd mac - testpmd> start + testpmd>show port stats all -4. Send packet with packet generator with different packet size,includes [64, 128, 256, 512, 1024, 1518], check the throughput with below command:: +Test Case 3: VM2VM vhost-user/virtio-net split ring vhost async operation test with tcp traffic using 4K-pages and cbdma enable +------------------------------------------------------------------------------------------------------------------------------- +This case test the function of Vhost TSO in the topology of vhost-user/virtio-net split ring mergeable path by verifing the TSO/cksum in the TCP/IP stack +when vhost uses the asynchronous operations with CBDMA channels in 4K-pages memory environment. - testpmd> show port stats all +1. Bind 2 CBDMA channels to vfio-pci, then launch vhost by below command:: -Test Case 3: VM2VM split ring vhost-user/virtio-net 4K-pages and CBDMA enable test with tcp traffic ---------------------------------------------------------------------------------------------------- -This case test the function of Vhost tx offload in the topology of vhost-user/virtio-net split ring mergeable path -by verifing the TSO/cksum in the TCP/IP stack when vhost uses the asynchronous enqueue operations with CBDMA channels -in 4k-pages environment. + rm -rf vhost-net* + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 2-4 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:00:04.0 -a 0000:00:04.1 \ + --vdev 'net_vhost0,iface=vhost-net0,queues=1,tso=1,dmas=[txq0;rxq0],dma_ring_size=2048' \ + --vdev 'net_vhost1,iface=vhost-net1,queues=1,tso=1,dmas=[txq0;rxq0],dma_ring_size=2048' \ + --iova=va -- -i --nb-cores=2 --txd=1024 --rxd=1024 --lcore-dma=[lcore3@0000:00:04.0,lcore4@0000:00:04.1] + testpmd>start -1. Bind 2 CBDMA channels to vfio-pci, as common steps 1. +2. Launch VM1 and VM2:: -2. Launch vhost by below command:: - - ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:80:04.0 0000:80:04.1 - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 30-32 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:80:04.0 -a 0000:80:04.1 \ - --vdev 'net_vhost0,iface=vhost-net0,queues=1,dmas=[txq0],dma_ring_size=2048' \ - --vdev 'net_vhost1,iface=vhost-net1,queues=1,dmas=[txq0],dma_ring_size=2048' \ - --iova=va -- -i --nb-cores=2 --txd=1024 --rxd=1024 --lcore-dma=[lcore31@0000:80:04.0,lcore32@0000:80:04.1] - testpmd> start - -3. Launch VM1 and VM2:: - - taskset -c 20,21,22,23,24,25,26,27 /home/QEMU/qemu-6.2.0/bin/qemu-system-x86_64 -name vm1 -enable-kvm -cpu host -smp 1 -m 4096 \ + taskset -c 32 qemu-system-x86_64 -name vm1 -enable-kvm -cpu host -smp 1 -m 4096 \ -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge0,share=on \ - -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004.img \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/osimg/ubuntu20-04.img \ -chardev socket,path=/tmp/vm1_qga0.sock,server,nowait,id=vm1_qga0 -device virtio-serial \ -device virtserialport,chardev=vm1_qga0,name=org.qemu.guest_agent.2 -daemonize \ -monitor unix:/tmp/vm1_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ - -netdev user,id=nttsip1,hostfwd=tcp:10.239.251.220:6000-:22 \ + -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6002-:22 \ -chardev socket,id=char0,path=./vhost-net0 \ -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce \ -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:01,disable-modern=false,mrg_rxbuf=on,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on -vnc :10 - taskset -c 48,49,50,51,52,53,54,55 /home/QEMU/qemu-6.2.0/bin/qemu-system-x86_64 -name vm2 -enable-kvm -cpu host -smp 1 -m 4096 \ + taskset -c 33 qemu-system-x86_64 -name vm2 -enable-kvm -cpu host -smp 1 -m 4096 \ -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge1,share=on \ - -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004_2.img \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/osimg/ubuntu20-04-2.img \ -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ - -netdev user,id=nttsip1,hostfwd=tcp:10.239.251.220:6001-:22 \ + -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6003-:22 \ -chardev socket,id=char0,path=./vhost-net1 \ -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce \ -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:02,disable-modern=false,mrg_rxbuf=on,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on -vnc :12 -4. On VM1, set virtio device IP and run arp protocal:: +3. On VM1, set virtio device IP and run arp protocal:: ifconfig ens5 1.1.1.2 arp -s 1.1.1.8 52:54:00:00:00:02 -5. On VM2, set virtio device IP and run arp protocal:: +4. On VM2, set virtio device IP and run arp protocal:: ifconfig ens5 1.1.1.8 arp -s 1.1.1.2 52:54:00:00:00:01 -6. Check the iperf performance between two VMs by below commands:: +5. Check the iperf performance between two VMs by below commands:: Under VM1, run: `iperf -s -i 1` Under VM2, run: `iperf -c 1.1.1.2 -i 1 -t 60` -7. Check 2VMs can receive and send big packets to each other:: +6. Check 2VMs can receive and send big packets to each other:: - testpmd> show port xstats all + testpmd>show port xstats all Port 0 should have tx packets above 1522 Port 1 should have rx packets above 1522 -Test Case 4: vm2vm vhost/virtio-net split ring multi queues with 4K-pages and cbdma enable -------------------------------------------------------------------------------------------- -This case uses iperf and scp to test the payload of large packet (larger than 1MB) is valid after packets forwarding in -vm2vm vhost-user/virtio-net split ring mergeable path when vhost uses the asynchronous enqueue operations with CBDMA channel. -The dynamic change of multi-queues number also test. +Test Case 4: VM2VM vhost-user/virtio-net packed ring vhost async operation test with tcp traffic using 4K-pages and cbdma enable +-------------------------------------------------------------------------------------------------------------------------------- +This case test the function of Vhost TSO in the topology of vhost-user/virtio-net packed ring mergeable path by verifing the TSO/cksum in the TCP/IP stack +when vhost uses the asynchronous operations with CBDMA channels in 4K-pages memory environment. -1. Bind 16 CBDMA channels to vfio-pci, as common steps 1. +1. Bind 2 CBDMA channels to vfio-pci, then launch vhost by below command:: -2. Launch vhost by below command:: + rm -rf vhost-net* + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 2-4 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:00:04.0 -a 0000:00:04.1 \ + --vdev 'net_vhost0,iface=vhost-net0,queues=1,tso=1,dmas=[txq0;rxq0],dma_ring_size=2048' \ + --vdev 'net_vhost1,iface=vhost-net1,queues=1,tso=1,dmas=[txq0;rxq0],dma_ring_size=2048' \ + --iova=va -- -i --nb-cores=2 --txd=1024 --rxd=1024 --lcore-dma=[lcore3@0000:00:04.0,lcore4@0000:00:04.1] + testpmd>start - ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:80:04.0 0000:80:04.1 0000:80:04.2 0000:80:04.3 0000:80:04.4 0000:80:04.5 0000:80:04.6 0000:80:04.7 \ - 0000:00:04.0 0000:00:04.1 0000:00:04.2 0000:00:04.3 0000:00:04.4 0000:00:04.5 0000:00:04.6 0000:00:04.7 +2. Launch VM1 and VM2:: - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 29-33 -n 4 --no-huge -m 1024 --file-prefix=vhost \ - -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:00:04.2 -a 0000:00:04.3 -a 0000:00:04.4 -a 0000:00:04.5 -a 0000:00:04.6 -a 0000:00:04.7 \ + taskset -c 32 qemu-system-x86_64 -name vm1 -enable-kvm -cpu host -smp 1 -m 4096 \ + -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge0,share=on \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/osimg/ubuntu20-04.img \ + -chardev socket,path=/tmp/vm1_qga0.sock,server,nowait,id=vm1_qga0 -device virtio-serial \ + -device virtserialport,chardev=vm1_qga0,name=org.qemu.guest_agent.2 -daemonize \ + -monitor unix:/tmp/vm1_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ + -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6002-:22 \ + -chardev socket,id=char0,path=./vhost-net0 \ + -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce \ + -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:01,disable-modern=false,mrg_rxbuf=on,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,packed=on -vnc :10 + + taskset -c 33 qemu-system-x86_64 -name vm2 -enable-kvm -cpu host -smp 1 -m 4096 \ + -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge1,share=on \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/osimg/ubuntu20-04-2.img \ + -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ + -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ + -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ + -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6003-:22 \ + -chardev socket,id=char0,path=./vhost-net1 \ + -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce \ + -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:02,disable-modern=false,mrg_rxbuf=on,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,packed=on -vnc :12 + +3. On VM1, set virtio device IP and run arp protocal:: + + ifconfig ens5 1.1.1.2 + arp -s 1.1.1.8 52:54:00:00:00:02 + +4. On VM2, set virtio device IP and run arp protocal:: + + ifconfig ens5 1.1.1.8 + arp -s 1.1.1.2 52:54:00:00:00:01 + +5. Check the iperf performance between two VMs by below commands:: + + Under VM1, run: `iperf -s -i 1` + Under VM2, run: `iperf -c 1.1.1.2 -i 1 -t 60` + +6. Check 2VMs can receive and send big packets to each other:: + + testpmd>show port xstats all + Port 0 should have tx packets above 1522 + Port 1 should have rx packets above 1522 + +Test Case 5: vm2vm vhost/virtio-net split ring multi queues using 4K-pages and cbdma enable +------------------------------------------------------------------------------------------- +This case uses iperf and scp to test the payload of large packet (larger than 1MB) is valid after packets forwarding in vm2vm vhost-user/virtio-net +split ring mergeable path when vhost uses the asynchronous operations with CBDMA channels in 4K-pages memory environment. The dynamic change of multi-queues number is also tested. + +1. Bind one port to vfio-pci, launch vhost:: + + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 1-5 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:00:04.0 -a 0000:00:04.1 -a 0000:00:04.2 -a 0000:00:04.3 -a 0000:00:04.4 -a 0000:00:04.5 -a 0000:00:04.6 -a 0000:00:04.7 \ -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:80:04.2 -a 0000:80:04.3 -a 0000:80:04.4 -a 0000:80:04.5 -a 0000:80:04.6 -a 0000:80:04.7 \ --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;txq6;txq7]' \ --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;txq6;txq7]' \ --iova=va -- -i --nb-cores=4 --txd=1024 --rxd=1024 --rxq=8 --txq=8 \ - --lcore-dma=[lcore30@0000:80:04.0,lcore30@0000:80:04.1,lcore30@0000:00:04.2,lcore30@0000:00:04.3,lcore30@0000:00:04.4,lcore30@0000:00:04.5,lcore31@0000:00:04.6,lcore31@0000:00:04.7,lcore32@0000:80:04.0,lcore32@0000:80:04.1,lcore32@0000:80:04.2,lcore32@0000:80:04.3,lcore32@0000:80:04.4,lcore32@0000:80:04.5,lcore32@0000:80:04.6,lcore33@0000:80:04.7] - - testpmd> start - + --lcore-dma=[lcore2@0000:00:04.0,lcore2@0000:00:04.1,lcore2@0000:00:04.2,lcore2@0000:00:04.3,lcore2@0000:00:04.4,lcore2@0000:00:04.5,lcore3@0000:00:04.6,lcore3@0000:00:04.7,lcore4@0000:80:04.0,lcore4@0000:80:04.1,lcore4@0000:80:04.2,lcore4@0000:80:04.3,lcore4@0000:80:04.4,lcore4@0000:80:04.5,lcore4@0000:80:04.6,lcore5@0000:80:04.7] + testpmd>start -3. Launch VM qemu:: +2. Launch VM qemu:: - taskset -c 20,21,22,23,24,25,26,27 /home/QEMU/qemu-6.2.0/bin/qemu-system-x86_64 -name vm1 -enable-kvm -cpu host -smp 8 -m 4096 \ - -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/huge,share=on \ - -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004.img \ - -chardev socket,path=/tmp/vm1_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ - -device virtserialport,chardev=vm1_qga0,name=org.qemu.guest_agent.2 -daemonize \ - -monitor unix:/tmp/vm1_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ - -netdev user,id=nttsip1,hostfwd=tcp:10.239.251.220:6000-:22 \ + taskset -c 32 qemu-system-x86_64 -name vm1 -enable-kvm -cpu host -smp 8 -m 4096 \ + -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge0,share=on \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/osimg/ubuntu20-04.img \ + -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ + -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ + -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ + -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6002-:22 \ -chardev socket,id=char0,path=./vhost-net0,server \ -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce,queues=8 \ -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:01,disable-modern=false,mrg_rxbuf=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on -vnc :10 - taskset -c 48,49,50,51,52,53,54,55 /home/QEMU/qemu-6.2.0/bin/qemu-system-x86_64 -name vm2 -enable-kvm -cpu host -smp 8 -m 4096 \ - -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_yinan1,share=on \ - -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004_2.img \ + taskset -c 40 qemu-system-x86_64 -name vm2 -enable-kvm -cpu host -smp 8 -m 4096 \ + -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge1,share=on \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/osimg/ubuntu20-04-2.img \ -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ - -netdev user,id=nttsip1,hostfwd=tcp:10.239.251.220:6001-:22 \ + -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6003-:22 \ -chardev socket,id=char0,path=./vhost-net1,server \ -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce,queues=8 \ -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:02,disable-modern=false,mrg_rxbuf=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on -vnc :12 -4. On VM1, set virtio device IP and run arp protocal:: +3. On VM1, set virtio device IP and run arp protocal:: ethtool -L ens5 combined 8 ifconfig ens5 1.1.1.2 arp -s 1.1.1.8 52:54:00:00:00:02 -5. On VM2, set virtio device IP and run arp protocal:: +4. On VM2, set virtio device IP and run arp protocal:: ethtool -L ens5 combined 8 ifconfig ens5 1.1.1.8 arp -s 1.1.1.2 52:54:00:00:00:01 -6. Scp 1MB file form VM1 to VM2:: +5. Scp 1MB file form VM1 to VM2:: Under VM1, run: `scp [xxx] root@1.1.1.8:/` [xxx] is the file name -7. Check the iperf performance between two VMs by below commands:: +6. Check the iperf performance between two VMs by below commands:: Under VM1, run: `iperf -s -i 1` Under VM2, run: `iperf -c 1.1.1.2 -i 1 -t 60` -8. Quit and relaunch vhost w/ diff CBDMA channels:: +7. Quit and relaunch vhost w/ diff CBDMA channels:: - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 29-33 -n 4 --no-huge -m 1024 --file-prefix=vhost \ - -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:00:04.2 -a 0000:00:04.3 -a 0000:00:04.4 -a 0000:00:04.5 -a 0000:00:04.6 -a 0000:00:04.7 \ + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 1-5 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:00:04.0 -a 0000:00:04.1 -a 0000:00:04.2 -a 0000:00:04.3 -a 0000:00:04.4 -a 0000:00:04.5 -a 0000:00:04.6 -a 0000:00:04.7 \ -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:80:04.2 -a 0000:80:04.3 -a 0000:80:04.4 -a 0000:80:04.5 -a 0000:80:04.6 -a 0000:80:04.7 \ - --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;txq6]' \ - --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=8,dmas=[txq1;txq2;txq3;txq4;txq5;txq6;txq7]' \ + --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;rxq2;rxq3;rxq4;rxq5;rxq6;rxq7]' \ + --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;rxq2;rxq3;rxq4;rxq5;rxq6;rxq7]' \ --iova=va -- -i --nb-cores=4 --txd=1024 --rxd=1024 --rxq=8 --txq=8 \ - --lcore-dma=[lcore30@0000:80:04.0,lcore30@0000:80:04.1,lcore30@0000:00:04.2,lcore30@0000:00:04.3,lcore31@0000:80:04.0,lcore31@0000:00:04.2,lcore31@0000:00:04.4,lcore31@0000:00:04.5,lcore31@0000:00:04.6,lcore31@0000:00:04.7,lcore32@0000:80:04.1,lcore32@0000:00:04.3,lcore32@0000:80:04.0,lcore32@0000:80:04.1,lcore32@0000:80:04.2,lcore32@0000:80:04.3,lcore32@0000:80:04.4,lcore32@0000:80:04.5,lcore32@0000:80:04.6,lcore33@0000:80:04.7] - testpmd> start + --lcore-dma=[lcore2@0000:00:04.0,lcore2@0000:00:04.1,lcore2@0000:00:04.2,lcore2@0000:00:04.3,lcore3@0000:00:04.0,lcore3@0000:00:04.2,lcore3@0000:00:04.4,lcore3@0000:00:04.5,lcore3@0000:00:04.6,lcore3@0000:00:04.7,lcore4@0000:00:04.1,lcore4@0000:00:04.3,lcore4@0000:80:04.0,lcore4@0000:80:04.1,lcore4@0000:80:04.2,lcore4@0000:80:04.3,lcore4@0000:80:04.4,lcore4@0000:80:04.5,lcore4@0000:80:04.6,lcore5@0000:80:04.7] + testpmd>start -9. Rerun step 6-7. +8. Rerun step 5-6. -10. Quit and relaunch vhost w/o CBDMA channels:: +9. Quit and relaunch vhost w/o CBDMA channels:: - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 29-33 -n 4 --no-huge -m 1024 --file-prefix=vhost \ + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 1-5 -n 4 --file-prefix=vhost \ --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=4' --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=4' \ -- -i --nb-cores=4 --txd=1024 --rxd=1024 --rxq=4 --txq=4 - testpmd> start + testpmd>start -11. On VM1, set virtio device:: +10. On VM1, set virtio device:: ethtool -L ens5 combined 4 -12. On VM2, set virtio device:: +11. On VM2, set virtio device:: ethtool -L ens5 combined 4 -13. Scp 1MB file form VM1 to VM2:: +12. Scp 1MB file form VM1 to VM2:: Under VM1, run: `scp [xxx] root@1.1.1.8:/` [xxx] is the file name -14. Check the iperf performance and compare with CBDMA enable performance, ensure CMDMA enable performance is higher:: +13. Check the iperf performance and compare with CBDMA enable performance, ensure CMDMA enable performance is higher:: Under VM1, run: `iperf -s -i 1` Under VM2, run: `iperf -c 1.1.1.2 -i 1 -t 60` -15. Quit and relaunch vhost with 1 queues:: +14. Quit and relaunch vhost with 1 queues:: - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 29-33 -n 4 --no-huge -m 1024 --file-prefix=vhost \ + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 1-5 -n 4 --no-huge -m 1024 --file-prefix=vhost \ --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=4' --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=4' \ -- -i --nb-cores=4 --txd=1024 --rxd=1024 --rxq=1 --txq=1 - testpmd> start + testpmd>start -16. On VM1, set virtio device:: +15. On VM1, set virtio device:: ethtool -L ens5 combined 1 -17. On VM2, set virtio device:: +16. On VM2, set virtio device:: ethtool -L ens5 combined 1 -18. Scp 1MB file form VM1 to VM2M, check packets can be forwarding success by scp:: +17. Scp 1MB file form VM1 to VM2M, check packets can be forwarding success by scp:: Under VM1, run: `scp [xxx] root@1.1.1.8:/` [xxx] is the file name -19. Check the iperf performance, ensure queue0 can work from vhost side:: +18. Check the iperf performance, ensure queue0 can work from vhost side:: Under VM1, run: `iperf -s -i 1` Under VM2, run: `iperf -c 1.1.1.2 -i 1 -t 60` -Test Case 5: vm2vm vhost/virtio-net split packed ring multi queues with 1G/4k-pages and cbdma enable ----------------------------------------------------------------------------------------------------- -This case uses iperf and scp to test the payload of large packet (larger than 1MB) is valid after packets forwarding in -vm2vm vhost-user/virtio-net multi-queues mergeable path when vhost uses the asynchronous enqueue operations with CBDMA -channels. And one virtio-net is split ring, the other is packed ring. The vhost run in 1G hugepages and the virtio-user run in 4k-pages environment. +Test Case 6: vm2vm vhost/virtio-net packed ring multi queues using 4K-pages and cbdma enable +-------------------------------------------------------------------------------------------- +This case uses iperf and scp to test the payload of large packet (larger than 1MB) is valid after packets forwarding in vm2vm vhost-user/virtio-net +packed ring mergeable path when vhost uses the asynchronous operations with CBDMA channels in 4K-pages memory environment. + +1. Bind one port to vfio-pci, launch vhost:: + + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 1-5 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:00:04.0 -a 0000:00:04.1 -a 0000:00:04.2 -a 0000:00:04.3 -a 0000:00:04.4 -a 0000:00:04.5 -a 0000:00:04.6 -a 0000:00:04.7 \ + -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:80:04.2 -a 0000:80:04.3 -a 0000:80:04.4 -a 0000:80:04.5 -a 0000:80:04.6 -a 0000:80:04.7 \ + --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;txq6;txq7]' \ + --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;txq6;txq7]' \ + --iova=va -- -i --nb-cores=4 --txd=1024 --rxd=1024 --rxq=8 --txq=8 \ + --lcore-dma=[lcore2@0000:00:04.0,lcore2@0000:00:04.1,lcore2@0000:00:04.2,lcore2@0000:00:04.3,lcore2@0000:00:04.4,lcore2@0000:00:04.5,lcore3@0000:00:04.6,lcore3@0000:00:04.7,lcore4@0000:80:04.0,lcore4@0000:80:04.1,lcore4@0000:80:04.2,lcore4@0000:80:04.3,lcore4@0000:80:04.4,lcore4@0000:80:04.5,lcore4@0000:80:04.6,lcore5@0000:80:04.7] + testpmd>start + +2. Launch VM qemu:: + + taskset -c 32 qemu-system-x86_64 -name vm1 -enable-kvm -cpu host -smp 8 -m 4096 \ + -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge0,share=on \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/osimg/ubuntu20-04.img \ + -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ + -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ + -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ + -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6002-:22 \ + -chardev socket,id=char0,path=./vhost-net0,server \ + -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce,queues=8 \ + -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:01,disable-modern=false,mrg_rxbuf=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on,packed=on -vnc :10 + + taskset -c 40 qemu-system-x86_64 -name vm2 -enable-kvm -cpu host -smp 8 -m 4096 \ + -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge1,share=on \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/osimg/ubuntu20-04-2.img \ + -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ + -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ + -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ + -netdev user,id=nttsip1,hostfwd=tcp:127.0.0.1:6003-:22 \ + -chardev socket,id=char0,path=./vhost-net1,server \ + -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce,queues=8 \ + -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:02,disable-modern=false,mrg_rxbuf=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on,packed=on -vnc :12 + +3. On VM1, set virtio device IP and run arp protocal:: + + ethtool -L ens5 combined 8 + ifconfig ens5 1.1.1.2 + arp -s 1.1.1.8 52:54:00:00:00:02 + +4. On VM2, set virtio device IP and run arp protocal:: -1. Bind 16 CBDMA channel to vfio-pci, as common steps 1. + ethtool -L ens5 combined 8 + ifconfig ens5 1.1.1.8 + arp -s 1.1.1.2 52:54:00:00:00:01 -2. Launch vhost by below command:: +5. Scp 1MB file form VM1 to VM2:: + + Under VM1, run: `scp [xxx] root@1.1.1.8:/` [xxx] is the file name + +6. Check the iperf performance between two VMs by below commands:: + + Under VM1, run: `iperf -s -i 1` + Under VM2, run: `iperf -c 1.1.1.2 -i 1 -t 60` + +Test Case 7: vm2vm vhost/virtio-net split ring multi queues using 1G/4k-pages and cbdma enable +---------------------------------------------------------------------------------------------- +This case uses iperf and scp to test the payload of large packet (larger than 1MB) is valid after packets forwarding in vm2vm vhost-user/virtio-net +split ring mergeable path when vhost uses the asynchronous operations with CBDMA channels,the back-end is in 1G-pages memory environment and the front-end is in 4k-pages memory environment. + +1. Bind 16 CBDMA channel to vfio-pci, launch vhost:: ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:80:04.0 0000:80:04.1 0000:80:04.2 0000:80:04.3 0000:80:04.4 0000:80:04.5 0000:80:04.6 0000:80:04.7 \ 0000:00:04.0 0000:00:04.1 0000:00:04.2 0000:00:04.3 0000:00:04.4 0000:00:04.5 0000:00:04.6 0000:00:04.7 - # ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 29-33 -n 4 -m 1024 --file-prefix=vhost \ + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 29-33 -n 4 -m 1024 --file-prefix=vhost \ -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:00:04.2 -a 0000:00:04.3 -a 0000:00:04.4 -a 0000:00:04.5 -a 0000:00:04.6 -a 0000:00:04.7 \ -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:80:04.2 -a 0000:80:04.3 -a 0000:80:04.4 -a 0000:80:04.5 -a 0000:80:04.6 -a 0000:80:04.7 \ - --vdev 'net_vhost0,iface=vhost-net0,queues=8,dmas=[txq0;txq1;txq2;txq3;]' \ - --vdev 'net_vhost1,iface=vhost-net1,queues=8,dmas=[txq0;txq1;txq2;txq3;]' \ + --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;rxq2;rxq3;rxq4;rxq5;rxq6;rxq7]' \ + --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;rxq2;rxq3;rxq4;rxq5;rxq6;rxq7]' \ --iova=va -- -i --nb-cores=4 --txd=1024 --rxd=1024 --rxq=8 --txq=8 \ --lcore-dma=[lcore30@0000:80:04.0,lcore30@0000:80:04.1,lcore30@0000:00:04.2,lcore30@0000:00:04.3,lcore31@0000:00:04.4,lcore31@0000:00:04.5,lcore31@0000:00:04.6,lcore31@0000:00:04.7,lcore32@0000:80:04.0,lcore32@0000:80:04.1,lcore32@0000:80:04.2,lcore32@0000:80:04.3,lcore33@0000:80:04.4,lcore33@0000:80:04.5,lcore33@0000:80:04.6,lcore33@0000:80:04.7] - testpmd> start + testpmd>start -3. Launch VM qemu:: +2. Launch VM qemu:: taskset -c 20,21,22,23,24,25,26,27 /home/QEMU/qemu-6.2.0/bin/qemu-system-x86_64 -name vm1 -enable-kvm -cpu host -smp 8 -m 4096 \ -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge0,share=on \ - -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004.img \ - -chardev socket,path=/tmp/vm1_qga0.sock,server,nowait,id=vm1_qga0 -device virtio-serial \ - -device virtserialport,chardev=vm1_qga0,name=org.qemu.guest_agent.2 -daemonize \ - -monitor unix:/tmp/vm1_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004.img \ + -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ + -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ + -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ -netdev user,id=nttsip1,hostfwd=tcp:10.239.251.220:6000-:22 \ - -chardev socket,id=char0,path=./vhost-net0 \ + -chardev socket,id=char0,path=./vhost-net0,server \ -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce,queues=8 \ -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:01,disable-modern=false,mrg_rxbuf=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on -vnc :10 taskset -c 48,49,50,51,52,53,54,55 /home/QEMU/qemu-6.2.0/bin/qemu-system-x86_64 -name vm2 -enable-kvm -cpu host -smp 8 -m 4096 \ -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge1,share=on \ - -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004_2.img \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004_2.img \ -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ -netdev user,id=nttsip1,hostfwd=tcp:10.239.251.220:6001-:22 \ - -chardev socket,id=char0,path=./vhost-net1 \ + -chardev socket,id=char0,path=./vhost-net1,server \ -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce,queues=8 \ - -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:02,disable-modern=false,mrg_rxbuf=on,packed=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on -vnc :12 + -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:02,disable-modern=false,mrg_rxbuf=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on -vnc :12 -4. On VM1, set virtio device IP and run arp protocal:: +3. On VM1, set virtio device IP and run arp protocal:: ethtool -L ens5 combined 8 ifconfig ens5 1.1.1.2 arp -s 1.1.1.8 52:54:00:00:00:02 -5. On VM2, set virtio device IP and run arp protocal:: +4. On VM2, set virtio device IP and run arp protocal:: ethtool -L ens5 combined 8 ifconfig ens5 1.1.1.8 arp -s 1.1.1.2 52:54:00:00:00:01 -6. Scp 1MB file form VM1 to VM2:: +5. Scp 1MB file form VM1 to VM2:: - Under VM1, run: `scp [xxx] root@1.1.1.8:/` [xxx] is the file name + Under VM1, run: `scp [xxx] root@1.1.1.8:/` [xxx] is the file name -7. Check the iperf performance between two VMs by below commands:: +6. Check the iperf performance between two VMs by below commands:: Under VM1, run: `iperf -s -i 1` Under VM2, run: `iperf -c 1.1.1.2 -i 1 -t 60` + +7. Quit and relaunch vhost w/ diff CBDMA channels:: + + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 1-5 -n 4 --no-huge -m 1024 --file-prefix=vhost -a 0000:00:04.0 -a 0000:00:04.1 -a 0000:00:04.2 -a 0000:00:04.3 -a 0000:00:04.4 -a 0000:00:04.5 -a 0000:00:04.6 -a 0000:00:04.7 \ + -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:80:04.2 -a 0000:80:04.3 -a 0000:80:04.4 -a 0000:80:04.5 -a 0000:80:04.6 -a 0000:80:04.7 \ + --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;txq6;txq7;rxq0;rxq1;rxq2;rxq3;rxq4;rxq5;rxq6;rxq7]' \ + --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;txq6;txq7;rxq0;rxq1;rxq2;rxq3;rxq4;rxq5;rxq6;rxq7]' \ + --iova=va -- -i --nb-cores=4 --txd=1024 --rxd=1024 --rxq=8 --txq=8 \ + --lcore-dma=[lcore2@0000:00:04.0,lcore2@0000:00:04.1,lcore2@0000:00:04.2,lcore2@0000:00:04.3,lcore3@0000:00:04.0,lcore3@0000:00:04.2,lcore3@0000:00:04.4,lcore3@0000:00:04.5,lcore3@0000:00:04.6,lcore3@0000:00:04.7,lcore4@0000:00:04.1,lcore4@0000:00:04.3,lcore4@0000:80:04.0,lcore4@0000:80:04.1,lcore4@0000:80:04.2,lcore4@0000:80:04.3,lcore4@0000:80:04.4,lcore4@0000:80:04.5,lcore4@0000:80:04.6,lcore5@0000:80:04.7] + testpmd>start + +8. Rerun step 5-6. + +Test Case 8: vm2vm vhost/virtio-net split packed ring multi queues with 1G/4k-pages and cbdma enable +---------------------------------------------------------------------------------------------------- +This case uses iperf and scp to test the payload of large packet (larger than 1MB) is valid after packets forwarding in vm2vm vhost-user/virtio-net +split and packed ring mergeable path when vhost uses the asynchronous operations with CBDMA channels,the back-end is in 1G-pages memory environment and the front-end is in 4k-pages memory environment. + +1. Bind 16 CBDMA channel to vfio-pci, launch vhost:: + + ./usertools/dpdk-devbind.py --bind=vfio-pci 0000:80:04.0 0000:80:04.1 0000:80:04.2 0000:80:04.3 0000:80:04.4 0000:80:04.5 0000:80:04.6 0000:80:04.7 \ + 0000:00:04.0 0000:00:04.1 0000:00:04.2 0000:00:04.3 0000:00:04.4 0000:00:04.5 0000:00:04.6 0000:00:04.7 + + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -l 29-33 -n 4 -m 1024 --file-prefix=vhost \ + -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:00:04.2 -a 0000:00:04.3 -a 0000:00:04.4 -a 0000:00:04.5 -a 0000:00:04.6 -a 0000:00:04.7 \ + -a 0000:80:04.0 -a 0000:80:04.1 -a 0000:80:04.2 -a 0000:80:04.3 -a 0000:80:04.4 -a 0000:80:04.5 -a 0000:80:04.6 -a 0000:80:04.7 \ + --vdev 'net_vhost0,iface=vhost-net0,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;rxq2;rxq3;rxq4;rxq5;rxq6;rxq7]' \ + --vdev 'net_vhost1,iface=vhost-net1,client=1,queues=8,dmas=[txq0;txq1;txq2;txq3;txq4;txq5;rxq2;rxq3;rxq4;rxq5;rxq6;rxq7]' \ + --iova=va -- -i --nb-cores=4 --txd=1024 --rxd=1024 --rxq=8 --txq=8 \ + --lcore-dma=[lcore30@0000:80:04.0,lcore30@0000:80:04.1,lcore30@0000:00:04.2,lcore30@0000:00:04.3,lcore31@0000:00:04.4,lcore31@0000:00:04.5,lcore31@0000:00:04.6,lcore31@0000:00:04.7,lcore32@0000:80:04.0,lcore32@0000:80:04.1,lcore32@0000:80:04.2,lcore32@0000:80:04.3,lcore33@0000:80:04.4,lcore33@0000:80:04.5,lcore33@0000:80:04.6,lcore33@0000:80:04.7] + testpmd>start + +2. Launch VM qemu:: + + taskset -c 20,21,22,23,24,25,26,27 /home/QEMU/qemu-6.2.0/bin/qemu-system-x86_64 -name vm1 -enable-kvm -cpu host -smp 8 -m 4096 \ + -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge0,share=on \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004.img \ + -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ + -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ + -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ + -netdev user,id=nttsip1,hostfwd=tcp:10.239.251.220:6000-:22 \ + -chardev socket,id=char0,path=./vhost-net0,server \ + -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce,queues=8 \ + -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:01,disable-modern=false,mrg_rxbuf=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on -vnc :10 + + taskset -c 48,49,50,51,52,53,54,55 /home/QEMU/qemu-6.2.0/bin/qemu-system-x86_64 -name vm2 -enable-kvm -cpu host -smp 8 -m 4096 \ + -object memory-backend-file,id=mem,size=4096M,mem-path=/mnt/tmpfs_nohuge1,share=on \ + -numa node,memdev=mem -mem-prealloc -drive file=/home/image/ubuntu2004_2.img \ + -chardev socket,path=/tmp/vm2_qga0.sock,server,nowait,id=vm2_qga0 -device virtio-serial \ + -device virtserialport,chardev=vm2_qga0,name=org.qemu.guest_agent.2 -daemonize \ + -monitor unix:/tmp/vm2_monitor.sock,server,nowait -device e1000,netdev=nttsip1 \ + -netdev user,id=nttsip1,hostfwd=tcp:10.239.251.220:6001-:22 \ + -chardev socket,id=char0,path=./vhost-net1,server \ + -netdev type=vhost-user,id=netdev0,chardev=char0,vhostforce,queues=8 \ + -device virtio-net-pci,netdev=netdev0,mac=52:54:00:00:00:02,disable-modern=false,mrg_rxbuf=on,mq=on,vectors=40,csum=on,guest_csum=on,host_tso4=on,guest_tso4=on,guest_ecn=on,guest_ufo=on,host_ufo=on,packed=on -vnc :12 + +3. On VM1, set virtio device IP and run arp protocal:: + + ethtool -L ens5 combined 8 + ifconfig ens5 1.1.1.2 + arp -s 1.1.1.8 52:54:00:00:00:02 + +4. On VM2, set virtio device IP and run arp protocal:: + + ethtool -L ens5 combined 8 + ifconfig ens5 1.1.1.8 + arp -s 1.1.1.2 52:54:00:00:00:01 + +5. Scp 1MB file form VM1 to VM2:: + + Under VM1, run: `scp [xxx] root@1.1.1.8:/` [xxx] is the file name + +6. Check the iperf performance between two VMs by below commands:: + + Under VM1, run: `iperf -s -i 1` + Under VM2, run: `iperf -c 1.1.1.2 -i 1 -t 60` + +7. Relaunch VM1, and rerun step 3. + +8. Rerun step 5-6.