From: Xiao Qimai <qimaix.xiao@intel.com>
To: dts@dpdk.org
Cc: Xiao Qimai <qimaix.xiao@intel.com>
Subject: [dts] [PATCH V3]pvp_vhost_user_reconnect: check perf data in each reconnect loop
Date: Fri, 19 Jun 2020 19:10:07 +0800 [thread overview]
Message-ID: <1592565007-21251-1-git-send-email-qimaix.xiao@intel.com> (raw)
* 1.check perf data in each reconnect loop
* 2.increase perf descend tolerance from 5% to 15% because of network fluctuations
Signed-off-by: Xiao Qimai <qimaix.xiao@intel.com>
---
tests/TestSuite_pvp_vhost_user_reconnect.py | 115 +++++++++++++++++-----------
1 file changed, 72 insertions(+), 43 deletions(-)
diff --git a/tests/TestSuite_pvp_vhost_user_reconnect.py b/tests/TestSuite_pvp_vhost_user_reconnect.py
index 2ddc454..b609115 100644
--- a/tests/TestSuite_pvp_vhost_user_reconnect.py
+++ b/tests/TestSuite_pvp_vhost_user_reconnect.py
@@ -71,7 +71,7 @@ class TestPVPVhostUserReconnect(TestCase):
self.frame_sizes = [64, 1518]
self.virtio_ip = ["1.1.1.2", "1.1.1.3"]
self.virtio_mac = ["52:54:00:00:00:01",
- "52:54:00:00:00:02"]
+ "52:54:00:00:00:02"]
self.src1 = "192.168.4.1"
self.dst1 = "192.168.3.1"
self.checked_vm = False
@@ -116,7 +116,8 @@ class TestPVPVhostUserReconnect(TestCase):
for i in range(self.vm_num):
vdev_info += "--vdev 'net_vhost%d,iface=vhost-net%d,client=1,queues=1' " % (i, i)
testcmd = self.dut.base_dir + "/%s/app/testpmd " % self.target
- eal_params = self.dut.create_eal_parameters(cores=self.cores, no_pci=True, prefix='vhost', ports=[self.pci_info])
+ eal_params = self.dut.create_eal_parameters(cores=self.cores, no_pci=True, prefix='vhost',
+ ports=[self.pci_info])
para = " -- -i --nb-cores=1 --txd=1024 --rxd=1024"
self.vhostapp_testcmd = testcmd + eal_params + vdev_info + para
self.vhost_user.send_expect(self.vhostapp_testcmd, "testpmd> ", 40)
@@ -127,10 +128,10 @@ class TestPVPVhostUserReconnect(TestCase):
check the link status is up after testpmd start
"""
loop = 1
- while(loop <= 5):
+ while (loop <= 5):
out = dut_info.send_expect("show port info all", "testpmd> ", 120)
port_status = re.findall("Link\s*status:\s*([a-z]*)", out)
- if("down" not in port_status):
+ if ("down" not in port_status):
break
time.sleep(3)
loop = loop + 1
@@ -153,11 +154,11 @@ class TestPVPVhostUserReconnect(TestCase):
out = self.dut.send_expect("%s --version" % self.vm_qemu_version, "#")
result = re.search("QEMU\s*emulator\s*version\s*(\d*.\d*)", out)
self.verify(result is not None,
- 'the qemu path may be not right: %s' % self.vm_qemu_version)
+ 'the qemu path may be not right: %s' % self.vm_qemu_version)
version = result.group(1)
index = version.find('.')
self.verify(int(version[:index]) > 2 or
- (int(version[:index]) == 2 and int(version[index+1:]) >= 7),
+ (int(version[:index]) == 2 and int(version[index + 1:]) >= 7),
'This qemu version should greater than 2.7 ' + \
'in this suite, please config it in vhost_sample.cfg file')
self.checked_vm = True
@@ -176,7 +177,7 @@ class TestPVPVhostUserReconnect(TestCase):
vm_params = {}
vm_params['driver'] = 'vhost-user'
vm_params['opt_path'] = './vhost-net%d' % (i)
- vm_params['opt_mac'] = '52:54:00:00:00:0%d' % (i+1)
+ vm_params['opt_mac'] = '52:54:00:00:00:0%d' % (i + 1)
vm_params['opt_server'] = 'server'
vm_params['opt_settings'] = setting_args
vm_info.set_vm_device(**vm_params)
@@ -198,7 +199,7 @@ class TestPVPVhostUserReconnect(TestCase):
start testpmd in vm
"""
vm_testpmd = self.dut.target + "/app/testpmd -c 0x3 -n 4 " + \
- "-- -i --port-topology=chained --txd=1024 --rxd=1024 "
+ "-- -i --port-topology=chained --txd=1024 --rxd=1024 "
for i in range(len(self.vm_dut)):
self.vm_dut[i].send_expect(vm_testpmd, "testpmd> ", 20)
self.vm_dut[i].send_expect("set fwd mac", "testpmd> ")
@@ -225,24 +226,24 @@ class TestPVPVhostUserReconnect(TestCase):
time.sleep(5)
vm_intf = self.vm_dut[i].ports_info[0]['intf']
self.vm_dut[i].send_expect("ifconfig %s %s" %
- (vm_intf, self.virtio_ip[i]), "#", 10)
+ (vm_intf, self.virtio_ip[i]), "#", 10)
self.vm_dut[i].send_expect("ifconfig %s up" % vm_intf, "#", 10)
self.vm_dut[0].send_expect('arp -s %s %s' %
- (self.virtio_ip[1], self.virtio_mac[1]), '#', 10)
+ (self.virtio_ip[1], self.virtio_mac[1]), '#', 10)
self.vm_dut[1].send_expect('arp -s %s %s' %
- (self.virtio_ip[0], self.virtio_mac[0]), '#', 10)
+ (self.virtio_ip[0], self.virtio_mac[0]), '#', 10)
def start_iperf(self):
"""
start iperf
"""
self.vm_dut[0].send_expect(
- 'iperf -s -p 12345 -i 1 > iperf_server.log &', '', 10)
+ 'iperf -s -p 12345 -i 1 > iperf_server.log &', '', 10)
self.vm_dut[1].send_expect(
- 'iperf -c %s -p 12345 -i 1 -t 5 > iperf_client.log &' %
- self.virtio_ip[0], '', 60)
- time.sleep(20)
+ 'iperf -c %s -p 12345 -i 1 -t 10 > iperf_client.log &' %
+ self.virtio_ip[0], '', 60)
+ time.sleep(15)
def iperf_result_verify(self, cycle, tinfo):
"""
@@ -250,7 +251,7 @@ class TestPVPVhostUserReconnect(TestCase):
"""
# copy iperf_client file from vm1
self.vm_dut[1].session.copy_file_from("%s/iperf_client.log" %
- self.dut.base_dir)
+ self.dut.base_dir)
fp = open("./iperf_client.log")
fmsg = fp.read()
fp.close()
@@ -261,12 +262,17 @@ class TestPVPVhostUserReconnect(TestCase):
else:
cinfo = tinfo
self.result_table_add(["vm2vm iperf", iperfdata[-1], cinfo])
+ data_li = iperfdata[-1].strip().split()
+ if self.nic in ['fortville_spirit']:
+ self.verify(data_li[1] == 'Gbits/sec', 'data unit not correct')
+ return float(data_li[0])
def send_and_verify(self, cycle=0, tinfo=""):
+ frame_data = dict().fromkeys(self.frame_sizes, 0)
for frame_size in self.frame_sizes:
- pkt = Packet(pkt_type = 'UDP', pkt_len = frame_size)
+ pkt = Packet(pkt_type='UDP', pkt_len=frame_size)
pkt.config_layers([('ether', {'dst': '%s' % self.dst_mac}),
- ('ipv4', {'dst': '%s' % self.dst1, 'src': '%s' % self.src1})])
+ ('ipv4', {'dst': '%s' % self.dst1, 'src': '%s' % self.src1})])
pkt.save_pcapfile(self.tester, "%s/reconnect.pcap" % self.out_path)
tgenInput = []
@@ -275,7 +281,7 @@ class TestPVPVhostUserReconnect(TestCase):
self.tester.pktgen.clear_streams()
streams = self.pktgen_helper.prepare_stream_from_tginput(tgenInput, 100,
- None, self.tester.pktgen)
+ None, self.tester.pktgen)
traffic_opt = {'delay': 30, }
_, pps = self.tester.pktgen.measure_throughput(stream_ids=streams, options=traffic_opt)
Mpps = pps / 1000000.0
@@ -285,7 +291,8 @@ class TestPVPVhostUserReconnect(TestCase):
check_speed = 5 if frame_size == 64 else 1
self.verify(Mpps > check_speed, "can not receive packets of frame size %d" % (frame_size))
pct = Mpps * 100 / \
- float(self.wirespeed(self.nic, frame_size, 1))
+ float(self.wirespeed(self.nic, frame_size, 1))
+ frame_data[frame_size] = Mpps
if cycle == 0:
data_row = [tinfo, frame_size, str(Mpps), str(pct),
"Before relaunch", "1"]
@@ -293,20 +300,30 @@ class TestPVPVhostUserReconnect(TestCase):
data_row = [tinfo, frame_size, str(Mpps), str(pct),
"After relaunch", "1"]
self.result_table_add(data_row)
+ return frame_data
+
+ def check_reconnect_perf(self):
+ if isinstance(self.before_data, dict):
+ for i in self.frame_sizes:
+ self.verify(
+ (self.before_data[i] - self.reconnect_data[i]) < self.before_data[i] * 0.15, 'verify reconnect speed failed')
+ else:
+ self.verify(
+ (self.before_data - self.reconnect_data) < self.before_data * 0.15, 'verify reconnect speed failed')
def test_perf_split_ring_reconnet_one_vm(self):
"""
test reconnect stability test of one vm
"""
self.header_row = ["Mode", "FrameSize(B)", "Throughput(Mpps)",
- "LineRate(%)", "Cycle", "Queue Number"]
+ "LineRate(%)", "Cycle", "Queue Number"]
self.result_table_create(self.header_row)
vm_cycle = 0
self.vm_num = 1
self.launch_testpmd_as_vhost_user()
self.start_vms()
self.vm_testpmd_start()
- self.send_and_verify(vm_cycle, "reconnet one vm")
+ self.before_data = self.send_and_verify(vm_cycle, "reconnet one vm")
vm_cycle = 1
# reconnet from vhost
@@ -314,7 +331,8 @@ class TestPVPVhostUserReconnect(TestCase):
for i in range(self.reconnect_times):
self.dut.send_expect("killall -s INT testpmd", "# ")
self.launch_testpmd_as_vhost_user()
- self.send_and_verify(vm_cycle, "reconnet from vhost")
+ self.reconnect_data = self.send_and_verify(vm_cycle, "reconnet from vhost")
+ self.check_reconnect_perf()
# reconnet from qemu
self.logger.info('now reconnect from vm')
@@ -322,7 +340,8 @@ class TestPVPVhostUserReconnect(TestCase):
self.dut.send_expect("killall -s INT qemu-system-x86_64", "# ")
self.start_vms()
self.vm_testpmd_start()
- self.send_and_verify(vm_cycle, "reconnet from VM")
+ self.reconnect_data = self.send_and_verify(vm_cycle, "reconnet from VM")
+ self.check_reconnect_perf()
self.result_table_print()
def test_perf_split_ring_reconnet_two_vms(self):
@@ -330,14 +349,14 @@ class TestPVPVhostUserReconnect(TestCase):
test reconnect stability test of two vms
"""
self.header_row = ["Mode", "FrameSize(B)", "Throughput(Mpps)",
- "LineRate(%)", "Cycle", "Queue Number"]
+ "LineRate(%)", "Cycle", "Queue Number"]
self.result_table_create(self.header_row)
vm_cycle = 0
self.vm_num = 2
self.launch_testpmd_as_vhost_user()
self.start_vms()
self.vm_testpmd_start()
- self.send_and_verify(vm_cycle, "reconnet two vm")
+ self.before_data = self.send_and_verify(vm_cycle, "reconnet two vm")
vm_cycle = 1
# reconnet from vhost
@@ -345,7 +364,8 @@ class TestPVPVhostUserReconnect(TestCase):
for i in range(self.reconnect_times):
self.dut.send_expect("killall -s INT testpmd", "# ")
self.launch_testpmd_as_vhost_user()
- self.send_and_verify(vm_cycle, "reconnet from vhost")
+ self.reconnect_data = self.send_and_verify(vm_cycle, "reconnet from vhost")
+ self.check_reconnect_perf()
# reconnet from qemu
self.logger.info('now reconnect from vm')
@@ -353,7 +373,8 @@ class TestPVPVhostUserReconnect(TestCase):
self.dut.send_expect("killall -s INT qemu-system-x86_64", "# ")
self.start_vms()
self.vm_testpmd_start()
- self.send_and_verify(vm_cycle, "reconnet from VM")
+ self.reconnect_data = self.send_and_verify(vm_cycle, "reconnet from VM")
+ self.check_reconnect_perf()
self.result_table_print()
def test_perf_split_ring_vm2vm_virtio_net_reconnet_two_vms(self):
@@ -368,7 +389,7 @@ class TestPVPVhostUserReconnect(TestCase):
self.start_vms()
self.config_vm_intf()
self.start_iperf()
- self.iperf_result_verify(vm_cycle, 'before reconnet')
+ self.before_data = self.iperf_result_verify(vm_cycle, 'before reconnet')
vm_cycle = 1
# reconnet from vhost
@@ -377,10 +398,12 @@ class TestPVPVhostUserReconnect(TestCase):
self.dut.send_expect("killall -s INT testpmd", "# ")
self.launch_testpmd_as_vhost_user_with_no_pci()
self.start_iperf()
- self.iperf_result_verify(vm_cycle, 'reconnet from vhost')
+ self.reconnect_data = self.iperf_result_verify(vm_cycle, 'reconnet from vhost')
+ self.check_reconnect_perf()
# reconnet from VM
self.logger.info('now reconnect from vm')
+ vm_tmp = list()
for i in range(self.reconnect_times):
self.vm_dut[0].send_expect('rm iperf_server.log', '# ', 10)
self.vm_dut[1].send_expect('rm iperf_client.log', '# ', 10)
@@ -388,7 +411,8 @@ class TestPVPVhostUserReconnect(TestCase):
self.start_vms()
self.config_vm_intf()
self.start_iperf()
- self.iperf_result_verify(vm_cycle, 'reconnet from vm')
+ self.reconnect_data = self.iperf_result_verify(vm_cycle, 'reconnet from vm')
+ self.check_reconnect_perf()
self.result_table_print()
def test_perf_packed_ring_reconnet_one_vm(self):
@@ -396,14 +420,14 @@ class TestPVPVhostUserReconnect(TestCase):
test reconnect stability test of one vm
"""
self.header_row = ["Mode", "FrameSize(B)", "Throughput(Mpps)",
- "LineRate(%)", "Cycle", "Queue Number"]
+ "LineRate(%)", "Cycle", "Queue Number"]
self.result_table_create(self.header_row)
vm_cycle = 0
self.vm_num = 1
self.launch_testpmd_as_vhost_user()
self.start_vms(packed=True)
self.vm_testpmd_start()
- self.send_and_verify(vm_cycle, "reconnet one vm")
+ self.before_data = self.send_and_verify(vm_cycle, "reconnet one vm")
vm_cycle = 1
# reconnet from vhost
@@ -411,7 +435,8 @@ class TestPVPVhostUserReconnect(TestCase):
for i in range(self.reconnect_times):
self.dut.send_expect("killall -s INT testpmd", "# ")
self.launch_testpmd_as_vhost_user()
- self.send_and_verify(vm_cycle, "reconnet from vhost")
+ self.reconnect_data = self.send_and_verify(vm_cycle, "reconnet from vhost")
+ self.check_reconnect_perf()
# reconnet from qemu
self.logger.info('now reconnect from vm')
@@ -419,7 +444,8 @@ class TestPVPVhostUserReconnect(TestCase):
self.dut.send_expect("killall -s INT qemu-system-x86_64", "# ")
self.start_vms(packed=True)
self.vm_testpmd_start()
- self.send_and_verify(vm_cycle, "reconnet from VM")
+ self.reconnect_data = self.send_and_verify(vm_cycle, "reconnet from VM")
+ self.check_reconnect_perf()
self.result_table_print()
def test_perf_packed_ring_reconnet_two_vms(self):
@@ -427,14 +453,14 @@ class TestPVPVhostUserReconnect(TestCase):
test reconnect stability test of two vms
"""
self.header_row = ["Mode", "FrameSize(B)", "Throughput(Mpps)",
- "LineRate(%)", "Cycle", "Queue Number"]
+ "LineRate(%)", "Cycle", "Queue Number"]
self.result_table_create(self.header_row)
vm_cycle = 0
self.vm_num = 2
self.launch_testpmd_as_vhost_user()
self.start_vms(packed=True)
self.vm_testpmd_start()
- self.send_and_verify(vm_cycle, "reconnet two vm")
+ self.before_data = self.send_and_verify(vm_cycle, "reconnet two vm")
vm_cycle = 1
# reconnet from vhost
@@ -442,15 +468,16 @@ class TestPVPVhostUserReconnect(TestCase):
for i in range(self.reconnect_times):
self.dut.send_expect("killall -s INT testpmd", "# ")
self.launch_testpmd_as_vhost_user()
- self.send_and_verify(vm_cycle, "reconnet from vhost")
-
+ self.reconnect_data = self.send_and_verify(vm_cycle, "reconnet from vhost")
+ self.check_reconnect_perf()
# reconnet from qemu
self.logger.info('now reconnect from vm')
for i in range(self.reconnect_times):
self.dut.send_expect("killall -s INT qemu-system-x86_64", "# ")
self.start_vms(packed=True)
self.vm_testpmd_start()
- self.send_and_verify(vm_cycle, "reconnet from VM")
+ self.reconnect_data = self.send_and_verify(vm_cycle, "reconnet from VM")
+ self.check_reconnect_perf()
self.result_table_print()
def test_perf_packed_ring_virtio_net_reconnet_two_vms(self):
@@ -465,7 +492,7 @@ class TestPVPVhostUserReconnect(TestCase):
self.start_vms(packed=True)
self.config_vm_intf()
self.start_iperf()
- self.iperf_result_verify(vm_cycle, 'before reconnet')
+ self.before_data = self.iperf_result_verify(vm_cycle, 'before reconnet')
vm_cycle = 1
# reconnet from vhost
@@ -474,7 +501,8 @@ class TestPVPVhostUserReconnect(TestCase):
self.dut.send_expect("killall -s INT testpmd", "# ")
self.launch_testpmd_as_vhost_user_with_no_pci()
self.start_iperf()
- self.iperf_result_verify(vm_cycle, 'reconnet from vhost')
+ self.reconnect_data = self.iperf_result_verify(vm_cycle, 'reconnet from vhost')
+ self.check_reconnect_perf()
# reconnet from VM
self.logger.info('now reconnect from vm')
@@ -485,7 +513,8 @@ class TestPVPVhostUserReconnect(TestCase):
self.start_vms(packed=True)
self.config_vm_intf()
self.start_iperf()
- self.iperf_result_verify(vm_cycle, 'reconnet from vm')
+ self.reconnect_data = self.iperf_result_verify(vm_cycle, 'reconnet from vm')
+ self.check_reconnect_perf()
self.result_table_print()
def tear_down(self):
--
1.8.3.1
next reply other threads:[~2020-06-19 11:18 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-06-19 11:10 Xiao Qimai [this message]
2020-06-19 12:07 ` Wang, Yinan
2020-06-28 6:36 ` Tu, Lijuan
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=1592565007-21251-1-git-send-email-qimaix.xiao@intel.com \
--to=qimaix.xiao@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).