test suite reviews and discussions
 help / color / mirror / Atom feed
From: Yulong Pei <yulong.pei@intel.com>
To: dts@dpdk.org
Subject: [dts] [PATCH] tests: Add VF packet drop and performance test
Date: Mon, 29 Feb 2016 16:48:18 +0800	[thread overview]
Message-ID: <1456735698-27253-1-git-send-email-yulong.pei@intel.com> (raw)

1.vf_perf.cfg: vm setting and qemu parameters.
2.vf_perf_test_plan.rst: test plan, describe test cases.
3.TestSuite_vf_perf.py: implement test cases according to the test plan.
4.using l3fwd to measure performance instead of testpmd.

Signed-off-by: Yulong Pei <yulong.pei@intel.com>
---
 conf/vf_perf.cfg                 | 107 +++++++++++++++++++
 test_plans/vf_perf_test_plan.rst | 182 ++++++++++++++++++++++++++++++++
 tests/TestSuite_vf_perf.py       | 222 +++++++++++++++++++++++++++++++++++++++
 3 files changed, 511 insertions(+)
 create mode 100644 conf/vf_perf.cfg
 create mode 100644 test_plans/vf_perf_test_plan.rst
 create mode 100644 tests/TestSuite_vf_perf.py

diff --git a/conf/vf_perf.cfg b/conf/vf_perf.cfg
new file mode 100644
index 0000000..36fac55
--- /dev/null
+++ b/conf/vf_perf.cfg
@@ -0,0 +1,107 @@
+# QEMU options
+# name
+#       name: vm0
+#
+# enable_kvm
+#       enable: [yes | no]
+#
+# cpu
+#       model: [host | core2duo | ...]
+#           usage:
+#               choose model value from the command
+#                   qemu-system-x86_64 -cpu help
+#       number: '4' #number of vcpus
+#       cpupin: '3 4 5 6' # host cpu list
+#
+# mem
+#       size: 1024
+#
+# disk
+#       file: /path/to/image/test.img
+#
+# net
+#        type: [nic | user | tap | bridge | ...]
+#           nic
+#               opt_vlan: 0
+#                   note: Default is 0.
+#               opt_macaddr: 00:00:00:00:01:01
+#                   note: if creating a nic, it`s better to specify a MAC,
+#                         else it will get a random number.
+#               opt_model:["e1000" | "virtio" | "i82551" | ...]
+#                   note: Default is e1000.
+#               opt_name: 'nic1'
+#               opt_addr: ''
+#                   note: PCI cards only.
+#               opt_vectors:
+#                   note: This option currently only affects virtio cards.
+#           user
+#               opt_vlan: 0
+#                   note: default is 0.
+#               opt_hostfwd: [tcp|udp]:[hostaddr]:hostport-[guestaddr]:guestport
+#                   note: If not specified, it will be setted automatically.
+#           tap
+#               opt_vlan: 0
+#                   note: default is 0.
+#               opt_br: br0
+#                   note: if choosing tap, need to specify bridge name,
+#                         else it will be br0.
+#               opt_script: QEMU_IFUP_PATH
+#                   note: if not specified, default is self.QEMU_IFUP_PATH.
+#               opt_downscript: QEMU_IFDOWN_PATH
+#                   note: if not specified, default is self.QEMU_IFDOWN_PATH.
+#
+# device
+#       driver: [pci-assign | virtio-net-pci | ...]
+#           pci-assign
+#               prop_host: 08:00.0
+#               prop_addr: 00:00:00:00:01:02
+#           virtio-net-pci
+#               prop_netdev: mynet1
+#               prop_id: net1
+#               prop_mac: 00:00:00:00:01:03
+#               prop_bus: pci.0
+#               prop_addr: 0x3
+#
+# monitor
+#       port: 6061   
+#           note: if adding monitor to vm, need to specicy
+#                 this port, else it will get a free port
+#                 on the host machine.
+#
+# qga
+#       enable: [yes | no]
+#
+# serial_port
+#       enable: [yes | no]
+#
+# vnc
+#       displayNum: 1
+#           note: you can choose a number not used on the host.
+#
+# daemon
+#       enable: 'yes'
+#           note:
+#               By default VM will start with the daemonize status.
+#               Not support starting it on the stdin now.
+
+# vm configuration for pmd sriov case
+[vm0]
+cpu =
+    model=host,number=4,cpupin=5 6 7 8;
+disk =
+    file=/home/image/sriov-fc20-1.img;
+login =
+    user=root,password=tester;
+net = 
+   type=nic,opt_vlan=0;
+   type=user,opt_vlan=0;
+monitor = 
+    port=;
+qga = 
+    enable=yes;
+vnc = 
+    displayNum=4;
+daemon =
+    enable=yes;
+qemu =
+    path=/usr/local/qemu-2.4.0/x86_64-softmmu/qemu-system-x86_64;
diff --git a/test_plans/vf_perf_test_plan.rst b/test_plans/vf_perf_test_plan.rst
new file mode 100644
index 0000000..53238ce
--- /dev/null
+++ b/test_plans/vf_perf_test_plan.rst
@@ -0,0 +1,182 @@
+.. Copyright (c) <2015>, Intel Corporation
+      All rights reserved.
+
+   Redistribution and use in source and binary forms, with or without
+   modification, are permitted provided that the following conditions
+   are met:
+
+   - Redistributions of source code must retain the above copyright
+     notice, this list of conditions and the following disclaimer.
+
+   - Redistributions in binary form must reproduce the above copyright
+     notice, this list of conditions and the following disclaimer in
+     the documentation and/or other materials provided with the
+     distribution.
+
+   - Neither the name of Intel Corporation nor the names of its
+     contributors may be used to endorse or promote products derived
+     from this software without specific prior written permission.
+
+   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+   FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
+   COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+   INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+   (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+   SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+   HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+   STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+   ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+   OF THE POSSIBILITY OF SUCH DAMAGE.
+
+Test Case 1: Measure packet loss with kernel PF & dpdk VF
+===================================
+
+1. got the pci device id of DUT, for example,
+
+./dpdk_nic_bind.py --st
+
+0000:81:00.0 'Ethernet Controller X710 for 10GbE SFP+' if=ens259f0 drv=i40e unused=
+
+2. create 2 VFs from 1 PF,
+
+echo 2 > /sys/bus/pci/devices/0000\:81\:00.0/sriov_numvfs
+./dpdk_nic_bind.py --st
+
+0000:81:00.0 'Ethernet Controller X710 for 10GbE SFP+' if=ens259f0 drv=i40e unused=igb_uio
+0000:81:02.0 'XL710/X710 Virtual Function' unused=
+0000:81:02.1 'XL710/X710 Virtual Function' unused=
+
+3. detach VFs from the host, bind them to pci-stub driver,
+
+virsh nodedev-detach pci_0000_81_02_0;
+virsh nodedev-detach pci_0000_81_02_1;
+
+./dpdk_nic_bind.py --st
+
+0000:81:02.0 'XL710/X710 Virtual Function' if= drv=pci-stub unused=igb_uio
+0000:81:02.1 'XL710/X710 Virtual Function' if= drv=pci-stub unused=igb_uio
+
+4. passthrough VFs 81:02.0 & 81:02.1 to vm0, start vm0,
+
+/usr/bin/qemu-system-x86_64  -name vm0 -enable-kvm \
+-cpu host -smp 4 -m 2048 -drive file=/home/image/sriov-fc20-1.img -vnc :1 \
+-device pci-assign,host=81:02.0,id=pt_0 \
+-device pci-assign,host=81:02.1,id=pt_1
+ 
+5. login vm0, got VFs pci device id in vm0, assume they are 00:06.0 & 00:07.0, bind them to igb_uio driver,
+and then start testpmd, set it in mac forward mode,
+
+./tools/dpdk_nic_bind.py --bind=igb_uio 00:06.0 00:07.0
+./x86_64-native-linuxapp-gcc/app/testpmd -c 0x0f -n 4 -w 00:06.0 -w 00:07.0 -- -i --portmask=0x3 --txqflags=0
+
+testpmd> set fwd mac
+testpmd> start
+
+6. using ixia traffic generator to send 64 bytes packet with 10% line rate to VF, verify packet loss < 0.0001.
+
+7. using ixia traffic generator to send 64 bytes packet with 100% line rate to VF, verify packet loss < 0.0001.
+
+Test Case 2: Measure performace with kernel PF & dpdk VF
+========================================================
+
+1. setup test environment as Test Case 1, step 1-5.
+
+2. Measure maximum RFC2544 performance throughput for the following packet sizes,
+
+frameSizes = [64, 128, 256, 512, 1024, 1280, 1518]
+
+The output format should be as below, with figures given in mpps:
+
++------------+--------+
+| Size\Cores |1S/4C/1T|
++------------+--------+
+| 64-byte    |        |
++------------+--------+
+| 128-byte   |        |
++------------+--------+
+| 256-byte   |        |
++------------+--------+
+| 512-byte   |        |
++------------+--------+
+| 1024-byte  |        |
++------------+--------+
+| 1280-byte  |        |
++------------+--------+
+| 1518-byte  |        |
++------------+--------+
+
+
+Test Case 3: Measure performace with dpdk PF & dpdk VF
+======================================================
+
+1. got the pci device id of DUT and bind it to igb_uio driver,  for example,
+
+./dpdk_nic_bind.py --st
+
+0000:81:00.0 'Ethernet Controller X710 for 10GbE SFP+' if=ens259f0 drv=i40e unused=
+
+./dpdk_nic_bind.py --bind=igb_uio 81:00.0
+
+2. create 2 VFs from 1 PF,
+
+echo 2 > /sys/bus/pci/devices/0000\:81\:00.0/max_vfs
+./dpdk_nic_bind.py --st
+
+0000:81:00.0 'Ethernet Controller X710 for 10GbE SFP+' if=ens259f0 drv=i40e unused=igb_uio
+0000:81:02.0 'XL710/X710 Virtual Function' unused=
+0000:81:02.1 'XL710/X710 Virtual Function' unused=
+
+3. detach VFs from the host, bind them to pci-stub driver,
+
+./dpdk_nic_bind.py --bind=pci-stub 81:02.0 81:02.1
+./dpdk_nic_bind.py --st
+
+0000:81:02.0 'XL710/X710 Virtual Function' if= drv=pci-stub unused=igb_uio
+0000:81:02.1 'XL710/X710 Virtual Function' if= drv=pci-stub unused=igb_uio
+
+4. bind PF 81:00.0 to testpmd and start it on the host,
+
+./x86_64-native-linuxapp-gcc/app/testpmd -c 0x0f -n 4 -w 81:00.0 -- -i --portmask=0x1 --txqflags=0
+
+5. passthrough VFs 81:02.0 & 81:02.1 to vm0, start vm0,
+
+/usr/bin/qemu-system-x86_64  -name vm0 -enable-kvm \
+-cpu host -smp 4 -m 2048 -drive file=/home/image/sriov-fc20-1.img -vnc :1 \
+-device pci-assign,host=81:02.0,id=pt_0 \
+-device pci-assign,host=81:02.1,id=pt_1
+
+6. login vm0, got VFs pci device id in vm0, assume they are 00:06.0 & 00:07.0, bind them to igb_uio driver,
+and then start testpmd, set it in mac forward mode,
+
+./tools/dpdk_nic_bind.py --bind=igb_uio 00:06.0 00:07.0
+./x86_64-native-linuxapp-gcc/app/testpmd -c 0x0f -n 4 -w 00:06.0 -w 00:07.0 -- -i --portmask=0x3 --txqflags=0
+
+testpmd> set fwd mac
+testpmd> start
+
+7. Measure maximum RFC2544 performance throughput for the following packet sizes,
+
+frameSizes = [64, 128, 256, 512, 1024, 1280, 1518]
+
+The output format should be as below, with figures given in mpps:
+
++------------+--------+
+| Size\Cores |1S/4C/1T|
++------------+--------+
+| 64-byte    |        |
++------------+--------+
+| 128-byte   |        |
++------------+--------+
+| 256-byte   |        |
++------------+--------+
+| 512-byte   |        |
++------------+--------+
+| 1024-byte  |        |
++------------+--------+
+| 1280-byte  |        |
++------------+--------+
+| 1518-byte  |        |
++------------+--------+
+
diff --git a/tests/TestSuite_vf_perf.py b/tests/TestSuite_vf_perf.py
new file mode 100644
index 0000000..2fccc59
--- /dev/null
+++ b/tests/TestSuite_vf_perf.py
@@ -0,0 +1,222 @@
+# <COPYRIGHT_TAG>
+
+import re
+import time
+
+import dts
+from qemu_kvm import QEMUKvm
+from test_case import TestCase
+from pmd_output import PmdOutput
+from etgen import IxiaPacketGenerator
+
+VM_CORES_MASK = 'all'
+
+class TestVfPerf(TestCase, IxiaPacketGenerator):
+
+    def set_up_all(self):
+
+        self.tester.extend_external_packet_generator(TestVfPerf, self)
+
+        self.dut_ports = self.dut.get_ports(self.nic)
+        self.verify(len(self.dut_ports) >= 1, "Insufficient ports")
+
+        self.core_configs = []
+        self.core_configs.append({'cores': '1S/4C/1T', 'pps': {}})
+
+        self.vm0 = None
+        self.vf0_mac = "00:12:34:56:78:01"
+
+
+    def set_up(self):
+
+        self.setup_2vf_1vm_env_flag = 0
+
+    def setup_2vf_1vm_env(self, driver='default'):
+        
+        self.used_dut_port = self.dut_ports[0]
+        self.dut.generate_sriov_vfs_by_port(self.used_dut_port, 2, driver=driver)
+        self.sriov_vfs_port = self.dut.ports_info[self.used_dut_port]['vfs_port']
+
+        if driver != 'igb_uio':
+            pf_itf = self.dut.ports_info[0]['port'].get_interface_name()
+            self.dut.send_expect("ip link set %s vf 0 mac %s" %(pf_itf, self.vf0_mac), "#")
+
+        try:
+
+            for port in self.sriov_vfs_port:
+                print port.pci
+                port.bind_driver('pci-stub')
+
+            time.sleep(1)
+            vf0_prop = {'opt_host': self.sriov_vfs_port[0].pci}
+            vf1_prop = {'opt_host': self.sriov_vfs_port[1].pci}
+
+            for port_id in self.dut_ports:
+                if port_id == self.used_dut_port:
+                    continue
+                port = self.dut.ports_info[port_id]['port']
+                port.bind_driver()
+
+            if driver == 'igb_uio':
+                self.host_testpmd = PmdOutput(self.dut)
+                eal_param = '-b %(vf0)s -b %(vf1)s' % {'vf0': self.sriov_vfs_port[0].pci,
+                                                       'vf1': self.sriov_vfs_port[1].pci}
+                self.host_testpmd.start_testpmd("1S/2C/2T", eal_param=eal_param)
+
+            # set up VM0 ENV
+            self.vm0 = QEMUKvm(self.dut, 'vm0', 'vf_perf')
+            self.vm0.set_vm_device(driver='pci-assign', **vf0_prop)
+            self.vm0.set_vm_device(driver='pci-assign', **vf1_prop)
+            self.vm_dut_0 = self.vm0.start()
+            if self.vm_dut_0 is None:
+                raise Exception("Set up VM0 ENV failed!")
+
+            self.setup_2vf_1vm_env_flag = 1
+        except Exception as e:
+            self.destroy_2vf_1vm_env()
+            raise Exception(e)
+
+    def destroy_2vf_1vm_env(self):
+
+        if getattr(self, 'vm0_testpmd', None):
+            self.vm0_testpmd.execute_cmd('stop')
+            self.vm0_testpmd.execute_cmd('quit', '# ')
+            self.vm0_testpmd = None
+
+        if getattr(self, 'vm0', None):
+            self.vm0_dut_ports = None
+            self.vm_dut_0 = None
+            self.vm0.stop()
+            self.vm0 = None
+
+        if getattr(self, 'host_testpmd', None):
+            self.host_testpmd.execute_cmd('quit', '# ')
+            self.host_testpmd = None
+
+        if getattr(self, 'used_dut_port', None):
+            self.dut.destroy_sriov_vfs_by_port(self.used_dut_port)
+            port = self.dut.ports_info[self.used_dut_port]['port']
+            port.bind_driver()
+            self.used_dut_port = None
+
+        for port_id in self.dut_ports:
+            self.dut.destroy_sriov_vfs_by_port(port_id)
+            port = self.dut.ports_info[port_id]['port']
+            port.bind_driver()
+
+        self.setup_2vf_1vm_env_flag = 0
+
+    def test_perf_kernel_pf_dpdk_vf_packet_loss(self):
+
+        self.setup_2vf_1vm_env(driver='')
+
+        self.vm0_dut_ports = self.vm_dut_0.get_ports('any')
+        port_id_0 = 0
+        self.vm0_testpmd = PmdOutput(self.vm_dut_0)
+        self.vm0_testpmd.start_testpmd(VM_CORES_MASK)
+        self.vm0_testpmd.execute_cmd('show port info all')
+        pmd0_vf0_mac = self.vm0_testpmd.get_port_mac(port_id_0)
+        self.vm0_testpmd.execute_cmd('set fwd mac')
+        self.vm0_testpmd.execute_cmd('start')
+
+        time.sleep(2)
+
+        tx_port = self.tester.get_local_port(self.dut_ports[0])
+        rx_port = tx_port
+        dst_mac = pmd0_vf0_mac
+        src_mac = self.tester.get_mac(tx_port)
+
+        self.tester.scapy_append('dmac="%s"' % dst_mac)
+        self.tester.scapy_append('smac="%s"' % src_mac)
+        self.tester.scapy_append('flows = [Ether(src=smac, dst=dmac)/IP(len=46)/UDP(len=26)/("X"*18)]')
+        self.tester.scapy_append('wrpcap("test.pcap", flows)')
+        self.tester.scapy_execute()
+
+        loss, _, _ = self.tester.traffic_generator_loss([(tx_port, rx_port, "test.pcap")], 10, delay=180)
+
+        self.verify(loss < 0.0001, "Excessive packet loss when sending 64 bytes packet with 10% line rate")
+
+        loss, _, _ = self.tester.traffic_generator_loss([(tx_port, rx_port, "test.pcap")], 100, delay=180)
+
+        self.verify(loss < 0.0001, "Excessive packet loss when sending 64 bytes packet with 100% line rate")
+
+
+    def measure_vf_performance(self, driver='default'):
+
+        if driver == 'igb_uio':
+            self.setup_2vf_1vm_env(driver='igb_uio')
+        else:
+            self.setup_2vf_1vm_env(driver='')
+
+        if driver == 'igb_uio':
+            self.vm0_dut_ports = self.vm_dut_0.get_ports('any')
+            port_id_0 = 0
+            self.vm0_testpmd = PmdOutput(self.vm_dut_0)
+            self.vm0_testpmd.start_testpmd(VM_CORES_MASK)
+            self.vm0_testpmd.execute_cmd('show port info all')
+            self.vf0_mac = self.vm0_testpmd.get_port_mac(port_id_0)
+            self.vm0_testpmd.execute_cmd('stop')
+            self.vm0_testpmd.execute_cmd('quit', '# ')
+            self.vm0_testpmd = None
+
+        out = self.vm_dut_0.build_dpdk_apps('examples/l3fwd')
+        self.verify("Error" not in out, "compilation error 1")
+        self.verify("No such file" not in out, "compilation error 2")
+
+        cmdline = "./examples/l3fwd/build/l3fwd -c 0xf -n 4 -- -p 0x3  --config '(0,0,0),(0,1,2),(1,0,1),(1,1,3)' "
+        self.vm_dut_0.send_expect(cmdline, "L3FWD:", 120)
+        
+        time.sleep(10)
+        
+        frameSizes = [64, 128, 256, 512, 1024, 1280, 1518]
+
+        for config in self.core_configs:
+            self.dut.kill_all()
+            cores = self.dut.get_core_list(config['cores'])
+
+            tx_port = self.tester.get_local_port(self.dut_ports[0])
+            rx_port = tx_port
+            dst_mac = self.vf0_mac
+            src_mac = self.tester.get_mac(tx_port)
+
+            global size
+            for size in frameSizes:
+                self.tester.scapy_append('dmac="%s"' % dst_mac)
+                self.tester.scapy_append('smac="%s"' % src_mac)
+                self.tester.scapy_append('flows = [Ether(src=smac, dst=dmac)/IP(src="1.2.3.4",dst="11.100.0.1")/UDP()/("X"*%d)]' % (size - 46))
+                self.tester.scapy_append('wrpcap("test.pcap", flows)')
+                self.tester.scapy_execute()
+                tgenInput = []
+                tgenInput.append((tx_port, rx_port, "test.pcap"))
+                _, pps = self.tester.traffic_generator_throughput(tgenInput)
+                config['pps'][size] = pps
+         
+        #Stop l3fwd
+        self.vm_dut_0.send_expect("^C", "#")
+
+        for n in range(len(self.core_configs)):
+            for size in frameSizes:
+                self.verify(
+                    self.core_configs[n]['pps'][size] is not 0, "No traffic detected")
+
+        # Print results
+        dts.results_table_add_header(['Frame size'] + [n['cores'] for n in self.core_configs])
+        for size in frameSizes:
+            dts.results_table_add_row([size] + [n['pps'][size] for n in self.core_configs])
+        dts.results_table_print()
+
+    def test_perf_kernel_pf_dpdk_vf_performance(self):
+        
+        self.measure_vf_performance(driver='')
+
+    def test_perf_dpdk_pf_dpdk_vf_performance(self):
+
+        self.measure_vf_performance(driver='igb_uio')
+
+    def tear_down(self):
+
+        if self.setup_2vf_1vm_env_flag == 1:
+            self.destroy_2vf_1vm_env()
+
+    def tear_down_all(self):
+        pass
-- 
2.1.0

             reply	other threads:[~2016-02-29  8:48 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-02-29  8:48 Yulong Pei [this message]
2016-03-02  5:39 ` Liu, Yong
  -- strict thread matches above, loose matches on Subject: below --
2016-01-19  8:07 Yulong Pei
2016-02-01  8:36 ` Liu, Yong

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1456735698-27253-1-git-send-email-yulong.pei@intel.com \
    --to=yulong.pei@intel.com \
    --cc=dts@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).